# HG changeset patch
# User emellor@xxxxxxxxxxxxxxxxxxxxxx
# Node ID cf89e8f0831b3a483c1b8c614d1b8eb02180548b
# Parent 516cf6553011ac4263814c7f7c0cf358f629a906
# Parent 1ecb7f1ddc7900f34a7d7d962e5be3df3d850d61
Merged.
diff -r 516cf6553011 -r cf89e8f0831b buildconfigs/mk.linux-2.6-xen
--- a/buildconfigs/mk.linux-2.6-xen Mon Mar 20 08:56:46 2006
+++ b/buildconfigs/mk.linux-2.6-xen Mon Mar 20 08:56:54 2006
@@ -2,8 +2,8 @@
OS = linux
LINUX_SERIES = 2.6
-LINUX_VER = 2.6.16-rc5
-LINUX_SRCS = linux-2.6.15.tar.bz2 patch-2.6.16-rc5.bz2
+LINUX_VER = 2.6.16-rc6
+LINUX_SRCS = linux-2.6.15.tar.bz2 patch-2.6.16-rc6.bz2
LINUX_PDIR = linux-$(LINUX_VER)
EXTRAVERSION ?= xen
@@ -34,7 +34,7 @@
touch $(@D)/.hgskip
touch $@
-pristine-linux-%.16-rc5/.valid-pristine: pristine-$(LINUX_PDIR)/.valid-srcs
+pristine-linux-%.16-rc6/.valid-pristine: pristine-$(LINUX_PDIR)/.valid-srcs
touch $@ # update timestamp to avoid rebuild
$(LINUX_DIR)/include/linux/autoconf.h: ref-$(OS)-$(LINUX_VER)/.valid-ref
diff -r 516cf6553011 -r cf89e8f0831b
linux-2.6-xen-sparse/arch/i386/kernel/Makefile
--- a/linux-2.6-xen-sparse/arch/i386/kernel/Makefile Mon Mar 20 08:56:46 2006
+++ b/linux-2.6-xen-sparse/arch/i386/kernel/Makefile Mon Mar 20 08:56:54 2006
@@ -11,7 +11,7 @@
obj-y += cpu/
obj-y += timers/
-obj-$(CONFIG_ACPI) += acpi/
+obj-y += acpi/
obj-$(CONFIG_X86_BIOS_REBOOT) += reboot.o
obj-$(CONFIG_MCA) += mca.o
obj-$(CONFIG_X86_MSR) += msr.o
diff -r 516cf6553011 -r cf89e8f0831b
linux-2.6-xen-sparse/arch/i386/kernel/acpi/Makefile
--- a/linux-2.6-xen-sparse/arch/i386/kernel/acpi/Makefile Mon Mar 20
08:56:46 2006
+++ b/linux-2.6-xen-sparse/arch/i386/kernel/acpi/Makefile Mon Mar 20
08:56:54 2006
@@ -1,4 +1,4 @@
-obj-y := boot.o
+obj-$(CONFIG_ACPI) += boot.o
obj-$(CONFIG_X86_IO_APIC) += earlyquirk.o
obj-$(CONFIG_ACPI_SLEEP) += sleep.o wakeup.o
diff -r 516cf6553011 -r cf89e8f0831b
linux-2.6-xen-sparse/arch/i386/kernel/acpi/boot-xen.c
--- a/linux-2.6-xen-sparse/arch/i386/kernel/acpi/boot-xen.c Mon Mar 20
08:56:46 2006
+++ b/linux-2.6-xen-sparse/arch/i386/kernel/acpi/boot-xen.c Mon Mar 20
08:56:54 2006
@@ -44,6 +44,9 @@
extern int gsi_irq_sharing(int gsi);
#include <asm/proto.h>
+static inline int acpi_madt_oem_check(char *oem_id, char *oem_table_id) {
return 0; }
+
+
#else /* X86 */
#ifdef CONFIG_X86_LOCAL_APIC
@@ -1111,9 +1114,6 @@
disable_acpi();
return error;
}
-#ifdef __i386__
- check_acpi_pci();
-#endif
acpi_table_parse(ACPI_BOOT, acpi_parse_sbf);
diff -r 516cf6553011 -r cf89e8f0831b
linux-2.6-xen-sparse/arch/i386/kernel/cpu/common-xen.c
--- a/linux-2.6-xen-sparse/arch/i386/kernel/cpu/common-xen.c Mon Mar 20
08:56:46 2006
+++ b/linux-2.6-xen-sparse/arch/i386/kernel/cpu/common-xen.c Mon Mar 20
08:56:54 2006
@@ -283,10 +283,10 @@
c->x86_capability[4] = excap;
c->x86 = (tfms >> 8) & 15;
c->x86_model = (tfms >> 4) & 15;
- if (c->x86 == 0xf) {
+ if (c->x86 == 0xf)
c->x86 += (tfms >> 20) & 0xff;
+ if (c->x86 >= 0x6)
c->x86_model += ((tfms >> 16) & 0xF) << 4;
- }
c->x86_mask = tfms & 15;
} else {
/* Have CPUID level 0 only - unheard of */
diff -r 516cf6553011 -r cf89e8f0831b
linux-2.6-xen-sparse/arch/i386/kernel/io_apic-xen.c
--- a/linux-2.6-xen-sparse/arch/i386/kernel/io_apic-xen.c Mon Mar 20
08:56:46 2006
+++ b/linux-2.6-xen-sparse/arch/i386/kernel/io_apic-xen.c Mon Mar 20
08:56:54 2006
@@ -92,6 +92,8 @@
static struct { int pin, apic; } ioapic_i8259 = { -1, -1 };
static DEFINE_SPINLOCK(ioapic_lock);
+
+int timer_over_8254 __initdata = 1;
/*
* Is the SiS APIC rmw bug present ?
@@ -2329,7 +2331,8 @@
apic_write_around(APIC_LVT0, APIC_LVT_MASKED | APIC_DM_EXTINT);
init_8259A(1);
timer_ack = 1;
- enable_8259A_irq(0);
+ if (timer_over_8254 > 0)
+ enable_8259A_irq(0);
pin1 = find_isa_irq_pin(0, mp_INT);
apic1 = find_isa_irq_apic(0, mp_INT);
@@ -2459,6 +2462,20 @@
print_IO_APIC();
}
+static int __init setup_disable_8254_timer(char *s)
+{
+ timer_over_8254 = -1;
+ return 1;
+}
+static int __init setup_enable_8254_timer(char *s)
+{
+ timer_over_8254 = 2;
+ return 1;
+}
+
+__setup("disable_8254_timer", setup_disable_8254_timer);
+__setup("enable_8254_timer", setup_enable_8254_timer);
+
/*
* Called after all the initialization is done. If we didnt find any
* APIC bugs then we can allow the modify fast path
diff -r 516cf6553011 -r cf89e8f0831b
linux-2.6-xen-sparse/arch/i386/kernel/microcode-xen.c
--- a/linux-2.6-xen-sparse/arch/i386/kernel/microcode-xen.c Mon Mar 20
08:56:46 2006
+++ b/linux-2.6-xen-sparse/arch/i386/kernel/microcode-xen.c Mon Mar 20
08:56:54 2006
@@ -25,6 +25,7 @@
#include <linux/kernel.h>
#include <linux/init.h>
#include <linux/sched.h>
+#include <linux/cpumask.h>
#include <linux/module.h>
#include <linux/slab.h>
#include <linux/vmalloc.h>
diff -r 516cf6553011 -r cf89e8f0831b
linux-2.6-xen-sparse/arch/i386/kernel/setup-xen.c
--- a/linux-2.6-xen-sparse/arch/i386/kernel/setup-xen.c Mon Mar 20 08:56:46 2006
+++ b/linux-2.6-xen-sparse/arch/i386/kernel/setup-xen.c Mon Mar 20 08:56:54 2006
@@ -1819,6 +1819,10 @@
op.u.set_iopl.iopl = 1;
HYPERVISOR_physdev_op(&op);
+#ifdef CONFIG_X86_IO_APIC
+ check_acpi_pci(); /* Checks more than just ACPI actually */
+#endif
+
#ifdef CONFIG_ACPI
if (!(xen_start_info->flags & SIF_INITDOMAIN)) {
printk(KERN_INFO "ACPI in unprivileged domain disabled\n");
diff -r 516cf6553011 -r cf89e8f0831b
linux-2.6-xen-sparse/arch/i386/kernel/time-xen.c
--- a/linux-2.6-xen-sparse/arch/i386/kernel/time-xen.c Mon Mar 20 08:56:46 2006
+++ b/linux-2.6-xen-sparse/arch/i386/kernel/time-xen.c Mon Mar 20 08:56:54 2006
@@ -840,9 +840,9 @@
write_seqlock_irqsave(&xtime_lock, flags);
xtime.tv_sec = sec;
xtime.tv_nsec = 0;
+ jiffies_64 += sleep_length;
+ wall_jiffies += sleep_length;
write_sequnlock_irqrestore(&xtime_lock, flags);
- jiffies += sleep_length;
- wall_jiffies += sleep_length;
if (last_timer->resume)
last_timer->resume();
cur_timer = last_timer;
diff -r 516cf6553011 -r cf89e8f0831b
linux-2.6-xen-sparse/arch/ia64/xen/hypercall.S
--- a/linux-2.6-xen-sparse/arch/ia64/xen/hypercall.S Mon Mar 20 08:56:46 2006
+++ b/linux-2.6-xen-sparse/arch/ia64/xen/hypercall.S Mon Mar 20 08:56:54 2006
@@ -254,7 +254,6 @@
st8 [r11]=r10
;;
br.ret.sptk.many rp
- ;;
END(xen_set_rr)
GLOBAL_ENTRY(xen_fc)
@@ -264,7 +263,16 @@
(p7) fc r32;;
(p7) br.ret.sptk.many rp
;;
- ptc.e r96 // this is a "privified" fc r32
+ movl r9=XSI_PSR_IC
+ mov r8=r32
+ ;;
+ ld8 r10=[r9]
+ ;;
+ st8 [r9]=r0
+ ;;
+ XEN_HYPER_FC
+ ;;
+ st8 [r9]=r10
;;
br.ret.sptk.many rp
END(xen_fc)
@@ -276,7 +284,16 @@
(p7) mov r8=cpuid[r32];;
(p7) br.ret.sptk.many rp
;;
- mov r72=rr[r32] // this is a "privified" mov r8=cpuid[r32]
+ movl r9=XSI_PSR_IC
+ mov r8=r32
+ ;;
+ ld8 r10=[r9]
+ ;;
+ st8 [r9]=r0
+ ;;
+ XEN_HYPER_GET_CPUID
+ ;;
+ st8 [r9]=r10
;;
br.ret.sptk.many rp
END(xen_get_cpuid)
@@ -288,7 +305,16 @@
(p7) mov r8=pmd[r32];;
(p7) br.ret.sptk.many rp
;;
- mov r72=pmc[r32] // this is a "privified" mov r8=pmd[r32]
+ movl r9=XSI_PSR_IC
+ mov r8=r32
+ ;;
+ ld8 r10=[r9]
+ ;;
+ st8 [r9]=r0
+ ;;
+ XEN_HYPER_GET_PMD
+ ;;
+ st8 [r9]=r10
;;
br.ret.sptk.many rp
END(xen_get_pmd)
@@ -301,10 +327,20 @@
(p7) mov r8=ar24;;
(p7) br.ret.sptk.many rp
;;
- mov ar24=r72 // this is a "privified" mov r8=ar.eflg
+ movl r9=XSI_PSR_IC
+ mov r8=r32
+ ;;
+ ld8 r10=[r9]
+ ;;
+ st8 [r9]=r0
+ ;;
+ XEN_HYPER_GET_EFLAG
+ ;;
+ st8 [r9]=r10
;;
br.ret.sptk.many rp
END(xen_get_eflag)
+
// some bits aren't set if pl!=0, see SDM vol1 3.1.8
GLOBAL_ENTRY(xen_set_eflag)
movl r8=running_on_xen;;
@@ -313,11 +349,17 @@
(p7) mov ar24=r32
(p7) br.ret.sptk.many rp
;;
- // FIXME: this remains no-op'd because it generates
- // a privileged register (general exception) trap rather than
- // a privileged operation fault
- //mov ar24=r32
- ;;
- br.ret.sptk.many rp
-END(xen_get_eflag)
+ movl r9=XSI_PSR_IC
+ mov r8=r32
+ ;;
+ ld8 r10=[r9]
+ ;;
+ st8 [r9]=r0
+ ;;
+ XEN_HYPER_SET_EFLAG
+ ;;
+ st8 [r9]=r10
+ ;;
+ br.ret.sptk.many rp
+END(xen_set_eflag)
#endif
diff -r 516cf6553011 -r cf89e8f0831b linux-2.6-xen-sparse/arch/ia64/xen/xenivt.S
--- a/linux-2.6-xen-sparse/arch/ia64/xen/xenivt.S Mon Mar 20 08:56:46 2006
+++ b/linux-2.6-xen-sparse/arch/ia64/xen/xenivt.S Mon Mar 20 08:56:54 2006
@@ -723,16 +723,12 @@
movl r30=1f // load continuation point in
case of nested fault
;;
#ifdef CONFIG_XEN
-#if 1
mov r18=r8;
mov r8=r16;
XEN_HYPER_THASH;;
mov r17=r8;
mov r8=r18;;
#else
- tak r17=r80 // "privified" thash
-#endif
-#else
thash r17=r16 // compute virtual address of
L3 PTE
#endif
mov r29=b0 // save b0 in case of nested
fault
@@ -812,16 +808,12 @@
#endif /* CONFIG_ITANIUM */
;;
#ifdef CONFIG_XEN
-#if 1
mov r18=r8;
mov r8=r16;
XEN_HYPER_THASH;;
mov r17=r8;
mov r8=r18;;
#else
- tak r17=r80 // "privified" thash
-#endif
-#else
thash r17=r16 // compute virtual address of
L3 PTE
#endif
mov r29=b0 // save b0 in case of nested
fault)
@@ -898,15 +890,11 @@
movl r30=1f // load continuation point in
case of nested fault
;;
#ifdef CONFIG_XEN
-#if 1
mov r18=r8;
mov r8=r16;
XEN_HYPER_THASH;;
mov r17=r8;
mov r8=r18;;
-#else
- tak r17=r80 // "privified" thash
-#endif
#else
thash r17=r16 // compute virtual address of
L3 PTE
#endif
diff -r 516cf6553011 -r cf89e8f0831b
linux-2.6-xen-sparse/arch/x86_64/kernel/entry-xen.S
--- a/linux-2.6-xen-sparse/arch/x86_64/kernel/entry-xen.S Mon Mar 20
08:56:46 2006
+++ b/linux-2.6-xen-sparse/arch/x86_64/kernel/entry-xen.S Mon Mar 20
08:56:54 2006
@@ -818,7 +818,7 @@
jmp error_exit
#ifdef CONFIG_X86_LOCAL_APIC
-ENTRY(nmi)
+KPROBE_ENTRY(nmi)
zeroentry do_nmi_callback
ENTRY(do_nmi_callback)
addq $8, %rsp
@@ -828,6 +828,7 @@
XEN_BLOCK_EVENTS(%rsi)
GET_THREAD_INFO(%rcx)
jmp retint_restore_args
+ .previous .text
#endif
ALIGN
diff -r 516cf6553011 -r cf89e8f0831b linux-2.6-xen-sparse/drivers/char/tty_io.c
--- a/linux-2.6-xen-sparse/drivers/char/tty_io.c Mon Mar 20 08:56:46 2006
+++ b/linux-2.6-xen-sparse/drivers/char/tty_io.c Mon Mar 20 08:56:54 2006
@@ -305,7 +305,7 @@
t->commit = 0;
t->read = 0;
/* DEBUG ONLY */
- memset(t->data, '*', size);
+/* memset(t->data, '*', size); */
/* printk("Flip recycle %p\n", t); */
return t;
}
diff -r 516cf6553011 -r cf89e8f0831b
linux-2.6-xen-sparse/drivers/xen/xenbus/xenbus_probe.c
--- a/linux-2.6-xen-sparse/drivers/xen/xenbus/xenbus_probe.c Mon Mar 20
08:56:46 2006
+++ b/linux-2.6-xen-sparse/drivers/xen/xenbus/xenbus_probe.c Mon Mar 20
08:56:54 2006
@@ -1049,6 +1049,8 @@
if (xsd_port_intf)
xsd_port_intf->read_proc = xsd_port_read;
}
+ else
+ xenstored_ready = 1;
/* Initialize the interface to xenstore. */
err = xs_init();
@@ -1058,10 +1060,8 @@
return err;
}
- if (!dom0) {
- xenstored_ready = 1;
+ if (!dom0)
xenbus_probe(NULL);
- }
return 0;
}
diff -r 516cf6553011 -r cf89e8f0831b
linux-2.6-xen-sparse/include/asm-i386/apic.h
--- a/linux-2.6-xen-sparse/include/asm-i386/apic.h Mon Mar 20 08:56:46 2006
+++ b/linux-2.6-xen-sparse/include/asm-i386/apic.h Mon Mar 20 08:56:54 2006
@@ -139,6 +139,8 @@
#define ARCH_APICTIMER_STOPS_ON_C3 1
#endif
+extern int timer_over_8254;
+
#else /* !CONFIG_X86_LOCAL_APIC */
static inline void lapic_shutdown(void) { }
diff -r 516cf6553011 -r cf89e8f0831b
linux-2.6-xen-sparse/include/asm-ia64/hypercall.h
--- a/linux-2.6-xen-sparse/include/asm-ia64/hypercall.h Mon Mar 20 08:56:46 2006
+++ b/linux-2.6-xen-sparse/include/asm-ia64/hypercall.h Mon Mar 20 08:56:54 2006
@@ -32,6 +32,7 @@
#include <xen/interface/xen.h>
#include <xen/interface/sched.h>
+#include <xen/interface/dom0_ops.h>
#include <linux/errno.h>
/* FIXME: temp place to hold these page related macros */
diff -r 516cf6553011 -r cf89e8f0831b
linux-2.6-xen-sparse/include/asm-ia64/hypervisor.h
--- a/linux-2.6-xen-sparse/include/asm-ia64/hypervisor.h Mon Mar 20
08:56:46 2006
+++ b/linux-2.6-xen-sparse/include/asm-ia64/hypervisor.h Mon Mar 20
08:56:54 2006
@@ -44,6 +44,8 @@
void force_evtchn_callback(void);
+int xen_init(void);
+
/* Turn jiffies into Xen system time. XXX Implement me. */
#define jiffies_to_st(j) 0
diff -r 516cf6553011 -r cf89e8f0831b
linux-2.6-xen-sparse/include/asm-ia64/xen/privop.h
--- a/linux-2.6-xen-sparse/include/asm-ia64/xen/privop.h Mon Mar 20
08:56:46 2006
+++ b/linux-2.6-xen-sparse/include/asm-ia64/xen/privop.h Mon Mar 20
08:56:54 2006
@@ -33,6 +33,11 @@
#define XEN_HYPER_GET_RR break 0x10
#define XEN_HYPER_SET_RR break 0x11
#define XEN_HYPER_SET_KR break 0x12
+#define XEN_HYPER_FC break 0x13
+#define XEN_HYPER_GET_CPUID break 0x14
+#define XEN_HYPER_GET_PMD break 0x15
+#define XEN_HYPER_GET_EFLAG break 0x16
+#define XEN_HYPER_SET_EFLAG break 0x17
#endif
#ifndef __ASSEMBLY__
diff -r 516cf6553011 -r cf89e8f0831b linux-2.6-xen-sparse/include/linux/gfp.h
--- a/linux-2.6-xen-sparse/include/linux/gfp.h Mon Mar 20 08:56:46 2006
+++ b/linux-2.6-xen-sparse/include/linux/gfp.h Mon Mar 20 08:56:54 2006
@@ -161,9 +161,9 @@
void page_alloc_init(void);
#ifdef CONFIG_NUMA
-void drain_remote_pages(void);
+void drain_node_pages(int node);
#else
-static inline void drain_remote_pages(void) { };
+static inline void drain_node_pages(int node) { };
#endif
#endif /* __LINUX_GFP_H */
diff -r 516cf6553011 -r cf89e8f0831b linux-2.6-xen-sparse/mm/page_alloc.c
--- a/linux-2.6-xen-sparse/mm/page_alloc.c Mon Mar 20 08:56:46 2006
+++ b/linux-2.6-xen-sparse/mm/page_alloc.c Mon Mar 20 08:56:54 2006
@@ -591,20 +591,19 @@
}
#ifdef CONFIG_NUMA
-/* Called from the slab reaper to drain remote pagesets */
-void drain_remote_pages(void)
-{
- struct zone *zone;
- int i;
+/*
+ * Called from the slab reaper to drain pagesets on a particular node that
+ * belong to the currently executing processor.
+ */
+void drain_node_pages(int nodeid)
+{
+ int i, z;
unsigned long flags;
local_irq_save(flags);
- for_each_zone(zone) {
+ for (z = 0; z < MAX_NR_ZONES; z++) {
+ struct zone *zone = NODE_DATA(nodeid)->node_zones + z;
struct per_cpu_pageset *pset;
-
- /* Do not drain local pagesets */
- if (zone->zone_pgdat->node_id == numa_node_id())
- continue;
pset = zone_pcp(zone, smp_processor_id());
for (i = 0; i < ARRAY_SIZE(pset->pcp); i++) {
diff -r 516cf6553011 -r cf89e8f0831b tools/examples/xmexample.vti
--- a/tools/examples/xmexample.vti Mon Mar 20 08:56:46 2006
+++ b/tools/examples/xmexample.vti Mon Mar 20 08:56:54 2006
@@ -14,8 +14,8 @@
# Kernel image file.
kernel = "/boot/Flash.fd"
-# The domain build function. VMX domain uses 'vmx'.
-builder='vmx'
+# The domain build function. VTI domain uses 'hvm'.
+builder='hvm'
# Initial memory allocation (in megabytes) for the new domain.
memory = 256
diff -r 516cf6553011 -r cf89e8f0831b tools/firmware/rombios/apmbios.S
--- a/tools/firmware/rombios/apmbios.S Mon Mar 20 08:56:46 2006
+++ b/tools/firmware/rombios/apmbios.S Mon Mar 20 08:56:54 2006
@@ -217,14 +217,22 @@
; APM interface disconnect
APMSYM(04):
cmp al, #0x04
+ jne APMSYM(05)
+ jmp APMSYM(ok)
+
+;-----------------
+; APM cpu idle
+APMSYM(05):
+ cmp al, #0x05
jne APMSYM(07)
+ hlt
jmp APMSYM(ok)
;-----------------
; APM Set Power State
APMSYM(07):
cmp al, #0x07
- jne APMSYM(0a)
+ jne APMSYM(08)
cmp bx, #1
jne APMSYM(ok)
@@ -265,6 +273,14 @@
mov ax, #APMSYM(07_standby_str)
call APMSYM(out_str)
pop edx
+ jmp APMSYM(ok)
+
+;-----------------
+; APM Enable / Disable
+APMSYM(08):
+ cmp al, #0x08
+ jne APMSYM(0a)
+
jmp APMSYM(ok)
;-----------------
@@ -297,11 +313,30 @@
; APM Driver Version
APMSYM(0e):
cmp al, #0x0e
- jne APMSYM(unimplemented)
+ jne APMSYM(0f)
mov ah, #1
mov al, #2
+ jmp APMSYM(ok)
+
+;-----------------
+; APM Engage / Disengage
+APMSYM(0f):
+ cmp al, #0x0f
+ jne APMSYM(10)
+
+ jmp APMSYM(ok)
+
+;-----------------
+; APM Get Capabilities
+APMSYM(10):
+ cmp al, #0x10
+ jne APMSYM(unimplemented)
+
+ mov bl, #0
+ mov cx, #0
+
jmp APMSYM(ok)
;-----------------
diff -r 516cf6553011 -r cf89e8f0831b tools/firmware/rombios/rombios.c
--- a/tools/firmware/rombios/rombios.c Mon Mar 20 08:56:46 2006
+++ b/tools/firmware/rombios/rombios.c Mon Mar 20 08:56:54 2006
@@ -149,7 +149,7 @@
#define BX_SUPPORT_FLOPPY 1
#define BX_FLOPPY_ON_CNT 37 /* 2 seconds */
#define BX_PCIBIOS 1
-#define BX_APM 0
+#define BX_APM 1
#define BX_USE_ATADRV 1
#define BX_ELTORITO_BOOT 1
diff -r 516cf6553011 -r cf89e8f0831b tools/libxc/xc_hvm_build.c
--- a/tools/libxc/xc_hvm_build.c Mon Mar 20 08:56:46 2006
+++ b/tools/libxc/xc_hvm_build.c Mon Mar 20 08:56:54 2006
@@ -51,7 +51,7 @@
char *elfbase, int xch, uint32_t dom, unsigned long *parray,
struct domain_setup_info *dsi);
-static unsigned char build_e820map(void *e820_page, unsigned long mem_size)
+static unsigned char build_e820map(void *e820_page, unsigned long long
mem_size)
{
struct e820entry *e820entry =
(struct e820entry *)(((unsigned char *)e820_page) + E820_MAP_OFFSET);
@@ -81,11 +81,17 @@
#define STATIC_PAGES 2 /* for ioreq_t and store_mfn */
/* Most of the ram goes here */
e820entry[nr_map].addr = 0x100000;
- e820entry[nr_map].size = mem_size - 0x100000 - STATIC_PAGES*PAGE_SIZE;
+ e820entry[nr_map].size = mem_size - 0x100000 - STATIC_PAGES * PAGE_SIZE;
e820entry[nr_map].type = E820_RAM;
nr_map++;
/* Statically allocated special pages */
+
+ /* For xenstore */
+ e820entry[nr_map].addr = mem_size - 2 * PAGE_SIZE;
+ e820entry[nr_map].size = PAGE_SIZE;
+ e820entry[nr_map].type = E820_XENSTORE;
+ nr_map++;
/* Shared ioreq_t page */
e820entry[nr_map].addr = mem_size - PAGE_SIZE;
@@ -93,12 +99,6 @@
e820entry[nr_map].type = E820_SHARED_PAGE;
nr_map++;
- /* For xenstore */
- e820entry[nr_map].addr = mem_size - 2*PAGE_SIZE;
- e820entry[nr_map].size = PAGE_SIZE;
- e820entry[nr_map].type = E820_XENSTORE;
- nr_map++;
-
e820entry[nr_map].addr = mem_size;
e820entry[nr_map].size = 0x3 * PAGE_SIZE;
e820entry[nr_map].type = E820_NVS;
@@ -117,8 +117,7 @@
return (*(((unsigned char *)e820_page) + E820_MAP_NR_OFFSET) = nr_map);
}
-static void
-set_hvm_info_checksum(struct hvm_info_table *t)
+static void set_hvm_info_checksum(struct hvm_info_table *t)
{
uint8_t *ptr = (uint8_t *)t, sum = 0;
unsigned int i;
@@ -142,19 +141,16 @@
char *va_map;
struct hvm_info_table *va_hvm;
-
- va_map = xc_map_foreign_range(
- xc_handle,
- dom,
- PAGE_SIZE,
- PROT_READ|PROT_WRITE,
- pfn_list[HVM_INFO_PFN]);
+ va_map = xc_map_foreign_range(xc_handle, dom, PAGE_SIZE,
+ PROT_READ | PROT_WRITE,
+ pfn_list[HVM_INFO_PFN]);
if ( va_map == NULL )
return -1;
va_hvm = (struct hvm_info_table *)(va_map + HVM_INFO_OFFSET);
memset(va_hvm, 0, sizeof(*va_hvm));
+
strncpy(va_hvm->signature, "HVM INFO", 8);
va_hvm->length = sizeof(struct hvm_info_table);
va_hvm->acpi_enabled = acpi;
@@ -183,58 +179,59 @@
unsigned long *store_mfn)
{
unsigned long *page_array = NULL;
-
unsigned long count, i;
+ unsigned long long ptr;
+ xc_mmu_t *mmu = NULL;
+
shared_info_t *shared_info;
void *e820_page;
unsigned char e820_map_nr;
- xc_mmu_t *mmu = NULL;
- int rc;
struct domain_setup_info dsi;
- unsigned long v_end;
+ unsigned long long v_end;
unsigned long shared_page_frame = 0;
shared_iopage_t *sp;
memset(&dsi, 0, sizeof(struct domain_setup_info));
- if ( (rc = parseelfimage(image, image_size, &dsi)) != 0 )
- goto error_out;
-
- if ( (dsi.v_start & (PAGE_SIZE-1)) != 0 )
+ if ( (parseelfimage(image, image_size, &dsi)) != 0 )
+ goto error_out;
+
+ if ( (dsi.v_kernstart & (PAGE_SIZE - 1)) != 0 )
{
PERROR("Guest OS must load to a page boundary.\n");
goto error_out;
}
/* memsize is in megabytes */
- v_end = (unsigned long)memsize << 20;
+ v_end = (unsigned long long)memsize << 20;
printf("VIRTUAL MEMORY ARRANGEMENT:\n"
- " Loaded HVM loader: %08lx->%08lx\n"
- " TOTAL: %08lx->%08lx\n",
+ " Loaded HVM loader: %08lx->%08lx\n"
+ " TOTAL: %08lx->%016llx\n",
dsi.v_kernstart, dsi.v_kernend,
dsi.v_start, v_end);
- printf(" ENTRY ADDRESS: %08lx\n", dsi.v_kernentry);
-
- if ( (v_end - dsi.v_start) > (nr_pages * PAGE_SIZE) )
- {
- ERROR("Initial guest OS requires too much space\n"
- "(%luMB is greater than %luMB limit)\n",
- (v_end-dsi.v_start)>>20, (nr_pages<<PAGE_SHIFT)>>20);
+ printf(" ENTRY ADDRESS: %08lx\n", dsi.v_kernentry);
+
+ if ( (v_end - dsi.v_start) > ((unsigned long long)nr_pages << PAGE_SHIFT) )
+ {
+ PERROR("Initial guest OS requires too much space: "
+ "(%lluMB is greater than %lluMB limit)\n",
+ (unsigned long long)(v_end - dsi.v_start) >> 20,
+ ((unsigned long long)nr_pages << PAGE_SHIFT) >> 20);
goto error_out;
}
if ( (page_array = malloc(nr_pages * sizeof(unsigned long))) == NULL )
{
- PERROR("Could not allocate memory");
+ PERROR("Could not allocate memory.\n");
goto error_out;
}
if ( xc_get_pfn_list(xc_handle, dom, page_array, nr_pages) != nr_pages )
{
- PERROR("Could not get the page frame list");
+ PERROR("Could not get the page frame list.\n");
goto error_out;
}
@@ -246,20 +243,21 @@
/* Write the machine->phys table entries. */
for ( count = 0; count < nr_pages; count++ )
{
+ ptr = (unsigned long long)page_array[count] << PAGE_SHIFT;
if ( xc_add_mmu_update(xc_handle, mmu,
- (page_array[count] << PAGE_SHIFT) |
- MMU_MACHPHYS_UPDATE, count) )
+ ptr | MMU_MACHPHYS_UPDATE, count) )
goto error_out;
}
- if ( set_hvm_info(xc_handle, dom, page_array, vcpus, pae, acpi, apic) ) {
- fprintf(stderr, "Couldn't set hvm info for HVM guest.\n");
+ if ( set_hvm_info(xc_handle, dom, page_array, vcpus, pae, acpi, apic) )
+ {
+ ERROR("Couldn't set hvm info for HVM guest.\n");
goto error_out;
}
if ( (e820_page = xc_map_foreign_range(
- xc_handle, dom, PAGE_SIZE, PROT_READ|PROT_WRITE,
- page_array[E820_MAP_PAGE >> PAGE_SHIFT])) == 0 )
+ xc_handle, dom, PAGE_SIZE, PROT_READ | PROT_WRITE,
+ page_array[E820_MAP_PAGE >> PAGE_SHIFT])) == 0 )
goto error_out;
memset(e820_page, 0, PAGE_SIZE);
e820_map_nr = build_e820map(e820_page, v_end);
@@ -267,8 +265,8 @@
/* shared_info page starts its life empty. */
if ( (shared_info = xc_map_foreign_range(
- xc_handle, dom, PAGE_SIZE, PROT_READ|PROT_WRITE,
- shared_info_frame)) == 0 )
+ xc_handle, dom, PAGE_SIZE, PROT_READ | PROT_WRITE,
+ shared_info_frame)) == 0 )
goto error_out;
memset(shared_info, 0, sizeof(shared_info_t));
/* Mask all upcalls... */
@@ -279,8 +277,8 @@
/* Populate the event channel port in the shared page */
shared_page_frame = page_array[(v_end >> PAGE_SHIFT) - 1];
if ( (sp = (shared_iopage_t *) xc_map_foreign_range(
- xc_handle, dom, PAGE_SIZE, PROT_READ|PROT_WRITE,
- shared_page_frame)) == 0 )
+ xc_handle, dom, PAGE_SIZE, PROT_READ | PROT_WRITE,
+ shared_page_frame)) == 0 )
goto error_out;
memset(sp, 0, PAGE_SIZE);
@@ -290,7 +288,7 @@
vp_eport = xc_evtchn_alloc_unbound(xc_handle, dom, 0);
if ( vp_eport < 0 ) {
- fprintf(stderr, "Couldn't get unbound port from VMX guest.\n");
+ PERROR("Couldn't get unbound port from VMX guest.\n");
goto error_out;
}
sp->vcpu_iodata[i].vp_eport = vp_eport;
diff -r 516cf6553011 -r cf89e8f0831b tools/libxc/xc_linux_build.c
--- a/tools/libxc/xc_linux_build.c Mon Mar 20 08:56:46 2006
+++ b/tools/libxc/xc_linux_build.c Mon Mar 20 08:56:54 2006
@@ -536,8 +536,12 @@
*store_mfn = page_array[1];
*console_mfn = page_array[2];
- printf("store_mfn: 0x%lx, console_mfn: 0x%lx\n",
- (uint64_t)store_mfn, (uint64_t)console_mfn);
+ printf("start_info: 0x%lx at 0x%lx, "
+ "store_mfn: 0x%lx at 0x%lx, "
+ "console_mfn: 0x%lx at 0x%lx\n",
+ page_array[0], nr_pages,
+ *store_mfn, nr_pages - 2,
+ *console_mfn, nr_pages - 1);
start_info = xc_map_foreign_range(
xc_handle, dom, PAGE_SIZE, PROT_READ|PROT_WRITE, page_array[0]);
diff -r 516cf6553011 -r cf89e8f0831b xen/Makefile
--- a/xen/Makefile Mon Mar 20 08:56:46 2006
+++ b/xen/Makefile Mon Mar 20 08:56:54 2006
@@ -61,6 +61,7 @@
$(MAKE) -C acm
endif
$(MAKE) -C arch/$(TARGET_ARCH)
+ $(MAKE) -C arch/$(TARGET_ARCH) $(TARGET)
# drivers/char/console.o contains static banner/compile info. Blow it away.
# Don't refresh these files during e.g., 'sudo make install'
diff -r 516cf6553011 -r cf89e8f0831b xen/Rules.mk
--- a/xen/Rules.mk Mon Mar 20 08:56:46 2006
+++ b/xen/Rules.mk Mon Mar 20 08:56:54 2006
@@ -7,6 +7,15 @@
perfc ?= n
perfc_arrays?= n
crash_debug ?= n
+
+# Hardcoded configuration implications and dependencies.
+# Do this is a neater way if it becomes unwieldy.
+ifeq ($(debug),y)
+verbose := y
+endif
+ifeq ($(perfc_arrays),y)
+perfc := y
+endif
XEN_ROOT=$(BASEDIR)/..
include $(XEN_ROOT)/Config.mk
@@ -27,41 +36,23 @@
HDRS := $(subst $(BASEDIR)/include/xen/banner.h,,$(HDRS))
HDRS := $(subst $(BASEDIR)/include/xen/compile.h,,$(HDRS))
-C_SRCS := $(wildcard *.c)
-S_SRCS := $(wildcard *.S)
-OBJS := $(patsubst %.S,%.o,$(S_SRCS))
-OBJS += $(patsubst %.c,%.o,$(C_SRCS))
+include $(BASEDIR)/arch/$(TARGET_ARCH)/Rules.mk
# Note that link order matters!
-ALL_OBJS := $(BASEDIR)/common/common.o
-ALL_OBJS += $(BASEDIR)/drivers/char/driver.o
-ALL_OBJS += $(BASEDIR)/drivers/acpi/driver.o
-ifeq ($(ACM_SECURITY),y)
-ALL_OBJS += $(BASEDIR)/acm/acm.o
-CFLAGS += -DACM_SECURITY
-endif
-ALL_OBJS += $(BASEDIR)/arch/$(TARGET_ARCH)/arch.o
+ALL_OBJS-y += $(BASEDIR)/common/built_in.o
+ALL_OBJS-y += $(BASEDIR)/drivers/built_in.o
+ALL_OBJS-$(ACM_SECURITY) += $(BASEDIR)/acm/built_in.o
+ALL_OBJS-y += $(BASEDIR)/arch/$(TARGET_ARCH)/built_in.o
-include $(BASEDIR)/arch/$(TARGET_ARCH)/Rules.mk
+CFLAGS-y += -g -D__XEN__
+CFLAGS-$(ACM_SECURITY) += -DACM_SECURITY
+CFLAGS-$(verbose) += -DVERBOSE
+CFLAGS-$(crash_debug) += -DCRASH_DEBUG
+CFLAGS-$(perfc) += -DPERF_COUNTERS
+CFLAGS-$(perfc_arrays) += -DPERF_ARRAYS
-CFLAGS += -g -D__XEN__
-
-ifneq ($(debug)$(verbose),nn)
-CFLAGS += -DVERBOSE
-endif
-
-ifeq ($(crash_debug),y)
-CFLAGS += -DCRASH_DEBUG
-endif
-
-ifeq ($(perfc),y)
-CFLAGS += -DPERF_COUNTERS
-ifeq ($(perfc_arrays),y)
-CFLAGS += -DPERF_ARRAYS
-endif
-endif
-
-CFLAGS := $(strip $(CFLAGS))
+ALL_OBJS := $(ALL_OBJS-y)
+CFLAGS := $(strip $(CFLAGS) $(CFLAGS-y))
%.o: %.c $(HDRS) Makefile
$(CC) $(CFLAGS) -c $< -o $@
diff -r 516cf6553011 -r cf89e8f0831b xen/acm/Makefile
--- a/xen/acm/Makefile Mon Mar 20 08:56:46 2006
+++ b/xen/acm/Makefile Mon Mar 20 08:56:54 2006
@@ -1,15 +1,9 @@
+include $(BASEDIR)/Rules.mk
-include $(BASEDIR)/Rules.mk
-OBJS = acm_core.o
-OBJS += acm_policy.o
-OBJS += acm_simple_type_enforcement_hooks.o
-OBJS += acm_chinesewall_hooks.o
-OBJS += acm_null_hooks.o
+obj-y += acm_core.o
+obj-y += acm_policy.o
+obj-y += acm_simple_type_enforcement_hooks.o
+obj-y += acm_chinesewall_hooks.o
+obj-y += acm_null_hooks.o
-default: acm.o
-
-acm.o: $(OBJS)
- $(LD) $(LDFLAGS) -r -o acm.o $(OBJS)
-
-clean:
- rm -f *.o *~ core
+include $(BASEDIR)/Post.mk
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/ia64/Rules.mk
--- a/xen/arch/ia64/Rules.mk Mon Mar 20 08:56:46 2006
+++ b/xen/arch/ia64/Rules.mk Mon Mar 20 08:56:54 2006
@@ -1,11 +1,13 @@
########################################
# ia64-specific definitions
+HAS_ACPI := y
VALIDATE_VT ?= n
ifneq ($(COMPILE_ARCH),$(TARGET_ARCH))
CROSS_COMPILE ?= /usr/local/sp_env/v2.2.5/i686/bin/ia64-unknown-linux-
endif
-AFLAGS += -D__ASSEMBLY__
+AFLAGS += -D__ASSEMBLY__ -nostdinc $(CPPFLAGS)
+AFLAGS += -mconstant-gp
CPPFLAGS += -I$(BASEDIR)/include -I$(BASEDIR)/include/asm-ia64 \
-I$(BASEDIR)/include/asm-ia64/linux \
-I$(BASEDIR)/include/asm-ia64/linux-xen \
@@ -13,6 +15,7 @@
-I$(BASEDIR)/arch/ia64/linux -I$(BASEDIR)/arch/ia64/linux-xen
CFLAGS += -nostdinc -fno-builtin -fno-common -fno-strict-aliasing
+CFLAGS += -mconstant-gp
#CFLAGS += -O3 # -O3 over-inlines making debugging tough!
CFLAGS += -O2 # but no optimization causes compile errors!
#CFLAGS += -iwithprefix include -Wall -DMONITOR_BASE=$(MONITOR_BASE)
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/ia64/asm-offsets.c
--- a/xen/arch/ia64/asm-offsets.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/ia64/asm-offsets.c Mon Mar 20 08:56:54 2006
@@ -84,7 +84,6 @@
//DEFINE(IA64_TASK_SIGHAND_OFFSET,offsetof (struct task_struct,
sighand));
//DEFINE(IA64_TASK_SIGNAL_OFFSET,offsetof (struct task_struct, signal));
//DEFINE(IA64_TASK_TGID_OFFSET, offsetof (struct task_struct, tgid));
- DEFINE(IA64_PGD, offsetof(struct domain, arch.mm));
DEFINE(IA64_TASK_THREAD_KSP_OFFSET, offsetof (struct vcpu,
arch._thread.ksp));
DEFINE(IA64_TASK_THREAD_ON_USTACK_OFFSET, offsetof (struct vcpu,
arch._thread.on_ustack));
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/ia64/linux-xen/efi.c
--- a/xen/arch/ia64/linux-xen/efi.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/ia64/linux-xen/efi.c Mon Mar 20 08:56:54 2006
@@ -37,11 +37,17 @@
#define EFI_DEBUG 0
extern efi_status_t efi_call_phys (void *, ...);
+extern unsigned long long memparse (char *ptr, char **retptr);
struct efi efi;
EXPORT_SYMBOL(efi);
static efi_runtime_services_t *runtime;
+#ifdef XEN
+// this is a temporary hack to avoid CONFIG_VIRTUAL_MEM_MAP
+static unsigned long mem_limit = ~0UL, max_addr = 0x100000000;
+#else
static unsigned long mem_limit = ~0UL, max_addr = ~0UL;
+#endif
#define efi_call_virt(f, args...) (*(f))(args)
@@ -328,8 +334,6 @@
if (running_on_sim && md->type != EFI_CONVENTIONAL_MEMORY)
continue;
}
-// this is a temporary hack to avoid CONFIG_VIRTUAL_MEM_MAP
- if (md->phys_addr >= 0x100000000) continue;
#endif
/*
* granule_addr is the base of md's first granule.
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/ia64/linux-xen/setup.c
--- a/xen/arch/ia64/linux-xen/setup.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/ia64/linux-xen/setup.c Mon Mar 20 08:56:54 2006
@@ -384,6 +384,9 @@
*cmdline_p = __va(ia64_boot_param->command_line);
#ifndef XEN
strlcpy(saved_command_line, *cmdline_p, COMMAND_LINE_SIZE);
+#else
+ early_cmdline_parse(cmdline_p);
+ cmdline_parse(*cmdline_p);
#endif
efi_init();
@@ -414,10 +417,6 @@
}
#endif
-#ifdef XEN
- early_cmdline_parse(cmdline_p);
- cmdline_parse(*cmdline_p);
-#endif
if (early_console_setup(*cmdline_p) == 0)
mark_bsp_online();
@@ -842,7 +841,9 @@
| IA64_DCR_DA | IA64_DCR_DD |
IA64_DCR_LC));
#endif
atomic_inc(&init_mm.mm_count);
+#ifndef XEN
current->active_mm = &init_mm;
+#endif
#ifdef XEN
if (current->domain->arch.mm)
#else
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/ia64/linux-xen/smp.c
--- a/xen/arch/ia64/linux-xen/smp.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/ia64/linux-xen/smp.c Mon Mar 20 08:56:54 2006
@@ -296,7 +296,12 @@
{
preempt_disable();
/* this happens for the common case of a single-threaded fork(): */
+#ifdef XEN
+ if (likely(mm == current->domain->arch.mm
+ && atomic_read(&mm->mm_users) == 1))
+#else
if (likely(mm == current->active_mm && atomic_read(&mm->mm_users) == 1))
+#endif
{
local_finish_flush_tlb_mm(mm);
preempt_enable();
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/ia64/linux-xen/smpboot.c
--- a/xen/arch/ia64/linux-xen/smpboot.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/ia64/linux-xen/smpboot.c Mon Mar 20 08:56:54 2006
@@ -72,8 +72,7 @@
#ifdef CONFIG_SMP /* ifdef XEN */
-//#define SMP_DEBUG 0
-#define SMP_DEBUG 1
+#define SMP_DEBUG 0
#if SMP_DEBUG
#define Dprintk(x...) printk(x)
@@ -482,16 +481,14 @@
do_rest:
task_for_booting_cpu = c_idle.idle;
#else
- struct domain *idle;
struct vcpu *v;
- void *stack;
v = idle_vcpu[cpu] = alloc_vcpu(idle_vcpu[0]->domain, cpu, cpu);
BUG_ON(v == NULL);
- printf ("do_boot_cpu: cpu=%d, domain=%p, vcpu=%p\n", cpu, idle, v);
-
- task_for_booting_cpu = v;
+ //printf ("do_boot_cpu: cpu=%d, domain=%p, vcpu=%p\n", cpu, idle, v);
+
+ task_for_booting_cpu = (task_t *)v;
/* Set cpu number. */
get_thread_info(v)->cpu = cpu;
@@ -522,6 +519,7 @@
return 0;
}
+#ifndef XEN
static int __init
decay (char *str)
{
@@ -531,6 +529,7 @@
}
__setup("decay=", decay);
+#endif
/*
* Initialize the logical CPU number to SAPICID mapping
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/ia64/vmx/mmio.c
--- a/xen/arch/ia64/vmx/mmio.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/ia64/vmx/mmio.c Mon Mar 20 08:56:54 2006
@@ -33,6 +33,7 @@
#include <asm/mm.h>
#include <asm/vmx.h>
#include <public/event_channel.h>
+#include <linux/event.h>
/*
struct mmio_list *lookup_mmio(u64 gpa, struct mmio_list *mio_base)
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/ia64/vmx/pal_emul.c
--- a/xen/arch/ia64/vmx/pal_emul.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/ia64/vmx/pal_emul.c Mon Mar 20 08:56:54 2006
@@ -20,6 +20,7 @@
#include <asm/vmx_vcpu.h>
#include <asm/pal.h>
+#include <asm/sal.h>
static void
get_pal_parameters (VCPU *vcpu, UINT64 *gr29,
@@ -68,7 +69,6 @@
static struct ia64_pal_retval
pal_vm_tr_read (VCPU *vcpu ) {
-#warning pal_vm_tr_read: to be implemented
struct ia64_pal_retval result;
result.status= -1; //unimplemented
@@ -101,7 +101,6 @@
static struct ia64_pal_retval
pal_halt (VCPU *vcpu) {
-#warning pal_halt: to be implemented
//bugbug: to be implement.
struct ia64_pal_retval result;
@@ -140,12 +139,18 @@
static struct ia64_pal_retval
pal_bus_get_features(VCPU *vcpu){
-
+ struct ia64_pal_retval result;
+
+ result.status= -1; //unimplemented
+ return result;
}
static struct ia64_pal_retval
pal_cache_summary(VCPU *vcpu){
-
+ struct ia64_pal_retval result;
+
+ result.status= -1; //unimplemented
+ return result;
}
static struct ia64_pal_retval
@@ -157,22 +162,34 @@
static struct ia64_pal_retval
pal_cache_info(VCPU *vcpu){
+ struct ia64_pal_retval result;
+
+ result.status= -1; //unimplemented
+ return result;
}
static struct ia64_pal_retval
pal_cache_prot_info(VCPU *vcpu){
-}
-
-static struct ia64_pal_retval
-pal_cache_shared_info(VCPU *vcpu){
+ struct ia64_pal_retval result;
+
+ result.status= -1; //unimplemented
+ return result;
}
static struct ia64_pal_retval
pal_mem_attrib(VCPU *vcpu){
+ struct ia64_pal_retval result;
+
+ result.status= -1; //unimplemented
+ return result;
}
static struct ia64_pal_retval
pal_debug_info(VCPU *vcpu){
+ struct ia64_pal_retval result;
+
+ result.status= -1; //unimplemented
+ return result;
}
static struct ia64_pal_retval
@@ -182,8 +199,16 @@
static struct ia64_pal_retval
pal_freq_base(VCPU *vcpu){
struct ia64_pal_retval result;
+ struct ia64_sal_retval isrv;
PAL_CALL(result,PAL_FREQ_BASE, 0, 0, 0);
+ if(result.v0 == 0){ //PAL_FREQ_BASE may not be implemented in some
platforms, call SAL instead.
+ SAL_CALL(isrv, SAL_FREQ_BASE,
+ SAL_FREQ_BASE_PLATFORM, 0, 0, 0, 0, 0, 0);
+ result.status = isrv.status;
+ result.v0 = isrv.v0;
+ result.v1 = result.v2 =0;
+ }
return result;
}
@@ -197,46 +222,89 @@
static struct ia64_pal_retval
pal_halt_info(VCPU *vcpu){
+ struct ia64_pal_retval result;
+
+ result.status= -1; //unimplemented
+ return result;
}
static struct ia64_pal_retval
pal_logical_to_physica(VCPU *vcpu){
+ struct ia64_pal_retval result;
+
+ result.status= -1; //unimplemented
+ return result;
}
static struct ia64_pal_retval
pal_perf_mon_info(VCPU *vcpu){
+ struct ia64_pal_retval result;
+
+ result.status= -1; //unimplemented
+ return result;
}
static struct ia64_pal_retval
pal_proc_get_features(VCPU *vcpu){
+ struct ia64_pal_retval result;
+
+ result.status= -1; //unimplemented
+ return result;
}
static struct ia64_pal_retval
pal_ptce_info(VCPU *vcpu){
+ struct ia64_pal_retval result;
+
+ result.status= -1; //unimplemented
+ return result;
}
static struct ia64_pal_retval
pal_register_info(VCPU *vcpu){
+ struct ia64_pal_retval result;
+
+ result.status= -1; //unimplemented
+ return result;
}
static struct ia64_pal_retval
pal_rse_info(VCPU *vcpu){
-}
-
+ struct ia64_pal_retval result;
+
+ result.status= -1; //unimplemented
+ return result;
+}
static struct ia64_pal_retval
pal_test_info(VCPU *vcpu){
+ struct ia64_pal_retval result;
+
+ result.status= -1; //unimplemented
+ return result;
}
static struct ia64_pal_retval
pal_vm_summary(VCPU *vcpu){
+ struct ia64_pal_retval result;
+
+ result.status= -1; //unimplemented
+ return result;
}
static struct ia64_pal_retval
pal_vm_info(VCPU *vcpu){
+ struct ia64_pal_retval result;
+
+ result.status= -1; //unimplemented
+ return result;
}
static struct ia64_pal_retval
pal_vm_page_size(VCPU *vcpu){
+ struct ia64_pal_retval result;
+
+ result.status= -1; //unimplemented
+ return result;
}
void
pal_emul( VCPU *vcpu) {
@@ -285,6 +353,82 @@
case PAL_FREQ_BASE:
result = pal_freq_base (vcpu);
+ break;
+
+ case PAL_BUS_GET_FEATURES :
+ result = pal_bus_get_features (vcpu);
+ break;
+
+ case PAL_CACHE_SUMMARY :
+ result = pal_cache_summary (vcpu);
+ break;
+
+ case PAL_CACHE_INIT :
+ result = pal_cache_init(vcpu);
+ break;
+
+ case PAL_CACHE_INFO :
+ result = pal_cache_info(vcpu);
+ break;
+
+ case PAL_CACHE_PROT_INFO :
+ result = pal_cache_prot_info(vcpu);
+ break;
+
+ case PAL_MEM_ATTRIB :
+ result = pal_mem_attrib(vcpu);
+ break;
+
+ case PAL_DEBUG_INFO :
+ result = pal_debug_info(vcpu);
+ break;
+
+ case PAL_FIXED_ADDR :
+ result = pal_fixed_addr(vcpu);
+ break;
+
+ case PAL_HALT_INFO :
+ result = pal_halt_info(vcpu);
+ break;
+
+ case PAL_LOGICAL_TO_PHYSICAL :
+ result = pal_logical_to_physica(vcpu);
+ break;
+
+ case PAL_PERF_MON_INFO :
+ result = pal_perf_mon_info(vcpu);
+ break;
+
+ case PAL_PROC_GET_FEATURES:
+ result = pal_proc_get_features(vcpu);
+ break;
+
+ case PAL_PTCE_INFO :
+ result = pal_ptce_info(vcpu);
+ break;
+
+ case PAL_REGISTER_INFO :
+ result = pal_register_info(vcpu);
+ break;
+
+ case PAL_RSE_INFO :
+ result = pal_rse_info(vcpu);
+ break;
+
+ case PAL_TEST_PROC :
+ result = pal_test_info(vcpu);
+ break;
+
+ case PAL_VM_SUMMARY :
+ result = pal_vm_summary(vcpu);
+ break;
+
+ case PAL_VM_INFO :
+ result = pal_vm_info(vcpu);
+ break;
+
+ case PAL_VM_PAGE_SIZE :
+ result = pal_vm_page_size(vcpu);
break;
default:
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/ia64/vmx/vlsapic.c
--- a/xen/arch/ia64/vmx/vlsapic.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/ia64/vmx/vlsapic.c Mon Mar 20 08:56:54 2006
@@ -97,16 +97,15 @@
{
vtime_t *vtm;
VCPU *vcpu = data;
- u64 cur_itc,vitm;
-
- UINT64 vec;
-
- vec = VCPU(vcpu, itv) & 0xff;
- vmx_vcpu_pend_interrupt(vcpu, vec);
-
+ u64 cur_itc,vitv;
+
+ vitv = VCPU(vcpu, itv);
+ if ( !ITV_IRQ_MASK(vitv) ){
+ vmx_vcpu_pend_interrupt(vcpu, vitv & 0xff);
+ }
vtm=&(vcpu->arch.arch_vmx.vtm);
cur_itc = now_itc(vtm);
- vitm =VCPU(vcpu, itm);
+ // vitm =VCPU(vcpu, itm);
//fire_itc2 = cur_itc;
//fire_itm2 = vitm;
update_last_itc(vtm,cur_itc); // pseudo read to update vITC
@@ -131,55 +130,76 @@
*/
uint64_t vtm_get_itc(VCPU *vcpu)
{
- uint64_t guest_itc, spsr;
+ uint64_t guest_itc;
vtime_t *vtm;
vtm=&(vcpu->arch.arch_vmx.vtm);
- // FIXME: should use local_irq_disable & local_irq_enable ??
- local_irq_save(spsr);
guest_itc = now_itc(vtm);
-// update_last_itc(vtm, guest_itc);
-
- local_irq_restore(spsr);
return guest_itc;
}
+
+
+
void vtm_set_itc(VCPU *vcpu, uint64_t new_itc)
{
- uint64_t spsr;
+ uint64_t vitm, vitv;
vtime_t *vtm;
-
+ vitm = VCPU(vcpu,itm);
+ vitv = VCPU(vcpu,itv);
vtm=&(vcpu->arch.arch_vmx.vtm);
- local_irq_save(spsr);
vtm->vtm_offset = new_itc - ia64_get_itc();
vtm->last_itc = new_itc;
- vtm_interruption_update(vcpu, vtm);
- local_irq_restore(spsr);
-}
-
-void vtm_set_itv(VCPU *vcpu)
-{
- uint64_t spsr,itv;
- vtime_t *vtm;
-
+ if(vitm < new_itc){
+ clear_bit(ITV_VECTOR(vitv), &VCPU(vcpu, irr[0]));
+ stop_timer(&vtm->vtm_timer);
+ }
+}
+
+
+#define TIMER_SLOP (50*1000) /* ns */ /* copy from timer.c */
+extern u64 cycle_to_ns(u64 cyle);
+
+
+void vtm_set_itm(VCPU *vcpu, uint64_t val)
+{
+ vtime_t *vtm;
+ uint64_t vitv, cur_itc, expires;
+ vitv = VCPU(vcpu, itv);
vtm=&(vcpu->arch.arch_vmx.vtm);
- local_irq_save(spsr);
- itv = VCPU(vcpu, itv);
- if ( ITV_IRQ_MASK(itv) )
+ // TODO; need to handle VHPI in future
+ clear_bit(ITV_VECTOR(vitv), &VCPU(vcpu, irr[0]));
+ VCPU(vcpu,itm)=val;
+ cur_itc =now_itc(vtm);
+ if(val > vtm->last_itc){
+ expires = NOW() + cycle_to_ns(val-cur_itc) + TIMER_SLOP;
+ set_timer(&vtm->vtm_timer, expires);
+ }else{
stop_timer(&vtm->vtm_timer);
- vtm_interruption_update(vcpu, vtm);
- local_irq_restore(spsr);
-}
-
-
-/*
- * Update interrupt or hook the vtm timer for fire
+ }
+}
+
+
+void vtm_set_itv(VCPU *vcpu, uint64_t val)
+{
+ uint64_t olditv;
+ olditv = VCPU(vcpu, itv);
+ VCPU(vcpu, itv) = val;
+ if(ITV_IRQ_MASK(val)){
+ clear_bit(ITV_VECTOR(olditv), &VCPU(vcpu, irr[0]));
+ }else if(ITV_VECTOR(olditv)!=ITV_VECTOR(val)){
+ if(test_and_clear_bit(ITV_VECTOR(olditv), &VCPU(vcpu, irr[0])))
+ set_bit(ITV_VECTOR(val), &VCPU(vcpu, irr[0]));
+ }
+}
+
+
+/*
+ * Update interrupt or hook the vtm timer for fire
* At this point vtm_timer should be removed if itv is masked.
*/
/* Interrupt must be disabled at this point */
-
-extern u64 cycle_to_ns(u64 cyle);
-#define TIMER_SLOP (50*1000) /* ns */ /* copy from timer.c */
+/*
void vtm_interruption_update(VCPU *vcpu, vtime_t* vtm)
{
uint64_t cur_itc,vitm,vitv;
@@ -197,8 +217,7 @@
cur_itc =now_itc(vtm);
diff_last = vtm->last_itc - vitm;
diff_now = cur_itc - vitm;
- update_last_itc (vtm,cur_itc);
-
+
if ( diff_last >= 0 ) {
// interrupt already fired.
stop_timer(&vtm->vtm_timer);
@@ -207,28 +226,32 @@
// ITV is fired.
vmx_vcpu_pend_interrupt(vcpu, vitv&0xff);
}
+*/
/* Both last_itc & cur_itc < itm, wait for fire condition */
- else {
+/* else {
expires = NOW() + cycle_to_ns(0-diff_now) + TIMER_SLOP;
set_timer(&vtm->vtm_timer, expires);
}
local_irq_restore(spsr);
}
+ */
/*
* Action for vtm when the domain is scheduled out.
* Remove the timer for vtm.
*/
+/*
void vtm_domain_out(VCPU *vcpu)
{
if(!is_idle_domain(vcpu->domain))
stop_timer(&vcpu->arch.arch_vmx.vtm.vtm_timer);
}
-
+ */
/*
* Action for vtm when the domain is scheduled in.
* Fire vtm IRQ or add the timer for vtm.
*/
+/*
void vtm_domain_in(VCPU *vcpu)
{
vtime_t *vtm;
@@ -238,6 +261,7 @@
vtm_interruption_update(vcpu, vtm);
}
}
+ */
/*
* Next for vLSapic
@@ -413,11 +437,12 @@
{
return ( (pending >> 4) > mic );
}
-
+#if 0
static int is_invalid_irq(int vec)
{
return (vec == 1 || ((vec <= 14 && vec >= 3)));
}
+#endif //shadow it due to no use currently
#define IRQ_NO_MASKED 0
#define IRQ_MASKED_BY_VTPR 1
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/ia64/vmx/vmmu.c
--- a/xen/arch/ia64/vmx/vmmu.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/ia64/vmx/vmmu.c Mon Mar 20 08:56:54 2006
@@ -31,6 +31,7 @@
#include <asm/hw_irq.h>
#include <asm/vmx_pal_vsa.h>
#include <asm/kregs.h>
+#include <asm/vcpu.h>
#include <xen/irq.h>
/*
@@ -68,14 +69,14 @@
/*
* The VRN bits of va stand for which rr to get.
*/
-ia64_rr vmmu_get_rr(VCPU *vcpu, u64 va)
-{
- ia64_rr vrr;
- vmx_vcpu_get_rr(vcpu, va, &vrr.rrval);
- return vrr;
-}
-
-
+//ia64_rr vmmu_get_rr(struct vcpu *vcpu, u64 va)
+//{
+// ia64_rr vrr;
+// vcpu_get_rr(vcpu, va, &vrr.rrval);
+// return vrr;
+//}
+
+/*
void recycle_message(thash_cb_t *hcb, u64 para)
{
if(hcb->ht == THASH_VHPT)
@@ -84,7 +85,7 @@
}
printk("hcb=%p recycled with %lx\n",hcb,para);
}
-
+ */
/*
* Purge all guest TCs in logical processor.
@@ -102,7 +103,6 @@
u32 stride1,stride2;
u32 i,j;
u64 psr;
-
result = ia64_pal_call_static(PAL_PTCE_INFO,0,0,0, 0);
if ( result.status != 0 ) {
@@ -113,7 +113,7 @@
count2 = LOW_32BITS (result.v1);
stride1 = HIGH_32BITS(result.v2);
stride2 = LOW_32BITS (result.v2);
-
+
local_irq_save(psr);
for (i=0; i<count1; i++) {
for (j=0; j<count2; j++) {
@@ -133,24 +133,10 @@
// struct page_info *page;
thash_cb_t *vhpt;
PTA pta_value;
-/*
- page = alloc_domheap_pages (NULL, VCPU_VHPT_ORDER, 0);
- if ( page == NULL ) {
- panic("No enough contiguous memory for init_domain_mm\n");
- }
- vbase = page_to_virt(page);
- printk("Allocate domain vhpt at 0x%lx\n", (u64)vbase);
- memset(vbase, 0, VCPU_VHPT_SIZE);
- */
-// vcur = (void*)((u64)vbase + VCPU_VHPT_SIZE);
vcur -= sizeof (thash_cb_t);
vhpt = vcur;
vhpt->ht = THASH_VHPT;
vhpt->vcpu = d;
-// vhpt->hash_func = machine_thash;
-// vcur -= sizeof (vhpt_special);
-// vs = vcur;
-
/* Setup guest pta */
pta_value.val = 0;
pta_value.ve = 1;
@@ -159,14 +145,10 @@
pta_value.base = ((u64)vbase) >> PTA_BASE_SHIFT;
d->arch.arch_vmx.mpta = pta_value.val;
-// vhpt->vs = vs;
-// vhpt->vs->get_mfn = __gpfn_to_mfn_foreign;
-// vhpt->vs->tag_func = machine_ttag;
vhpt->hash = vbase;
vhpt->hash_sz = VCPU_VHPT_SIZE/2;
vhpt->cch_buf = (void *)(vbase + vhpt->hash_sz);
vhpt->cch_sz = (u64)vcur - (u64)vhpt->cch_buf;
-// vhpt->recycle_notifier = recycle_message;
thash_init(vhpt,VCPU_VHPT_SHIFT-1);
return vhpt;
}
@@ -177,9 +159,8 @@
{
struct page_info *page;
void *vbase, *vhptbase, *vcur;
- tlb_special_t *ts;
thash_cb_t *tlb;
-
+
page = alloc_domheap_pages (NULL, VCPU_VHPT_ORDER, 0);
if ( page == NULL ) {
panic("No enough contiguous memory for init_domain_mm\n");
@@ -193,10 +174,7 @@
tlb = vcur;
tlb->ht = THASH_TLB;
tlb->vcpu = d;
- vcur -= sizeof (tlb_special_t);
- ts = vcur;
- tlb->ts = ts;
- tlb->ts->vhpt = init_domain_vhpt(d,vhptbase,vbase);
+ tlb->vhpt = init_domain_vhpt(d,vhptbase,vbase);
// tlb->hash_func = machine_thash;
tlb->hash = vbase;
tlb->hash_sz = VCPU_VTLB_SIZE/2;
@@ -207,27 +185,6 @@
return tlb;
}
-/* Allocate physical to machine mapping table for domN
- * FIXME: Later this interface may be removed, if that table is provided
- * by control panel. Dom0 has gpfn identical to mfn, which doesn't need
- * this interface at all.
- */
-#if 0
-void
-alloc_pmt(struct domain *d)
-{
- struct page_info *page;
-
- /* Only called once */
- ASSERT(d->arch.pmt);
-
- page = alloc_domheap_pages(NULL, get_order(d->max_pages), 0);
- ASSERT(page);
-
- d->arch.pmt = page_to_virt(page);
- memset(d->arch.pmt, 0x55, d->max_pages * 8);
-}
-#endif
/*
* Insert guest TLB to machine TLB.
* data: In TLB format
@@ -240,7 +197,6 @@
unsigned long mtlb_ppn;
mtlb.ifa = tlb->vadr;
mtlb.itir = tlb->itir & ~ITIR_RV_MASK;
- //vmx_vcpu_get_rr(d, mtlb.ifa, &vrr.value);
mtlb.page_flags = tlb->page_flags & ~PAGE_FLAGS_RV_MASK;
mtlb.ppn = get_mfn(d->domain,tlb->ppn);
mtlb_ppn=mtlb.ppn;
@@ -311,7 +267,7 @@
IA64_PSR vpsr;
vpsr.val = vmx_vcpu_get_psr(vcpu);
- vrr = vmx_vcpu_rr(vcpu, vadr);
+ vcpu_get_rr(vcpu, vadr, &vrr.rrval);
vmx_vcpu_get_pta(vcpu,&vpta.val);
if ( vrr.ve & vpta.ve ) {
@@ -355,21 +311,18 @@
u64 *vpa;
thash_data_t *tlb;
thash_cb_t *hcb;
- ia64_rr vrr;
u64 mfn;
if ( !(VCPU(vcpu, vpsr) & IA64_PSR_IT) ) { // I-side physical mode
gpip = gip;
}
else {
- vmx_vcpu_get_rr(vcpu, gip, &vrr.rrval);
- hcb = vmx_vcpu_get_vtlb(vcpu);
- tlb = vtlb_lookup_ex (hcb, vrr.rid, gip, ISIDE_TLB );
- if( tlb == NULL )
- tlb = vtlb_lookup_ex (hcb,
- vrr.rid, gip, DSIDE_TLB );
- if (tlb)
- gpip = (tlb->ppn << 12) | ( gip & (PSIZE(tlb->ps)-1) );
+ hcb = vmx_vcpu_get_vtlb(vcpu);
+ tlb = vtlb_lookup(hcb, gip, ISIDE_TLB);
+// if( tlb == NULL )
+// tlb = vtlb_lookup(hcb, gip, DSIDE_TLB );
+ if (tlb)
+ gpip = (tlb->ppn >>(tlb->ps-12)<<tlb->ps) | ( gip &
(PSIZE(tlb->ps)-1) );
}
if( gpip){
mfn = gmfn_to_mfn(vcpu->domain, gpip >>PAGE_SHIFT);
@@ -388,236 +341,146 @@
IA64FAULT vmx_vcpu_itc_i(VCPU *vcpu, UINT64 pte, UINT64 itir, UINT64 ifa)
{
-
- thash_data_t data, *ovl;
- thash_cb_t *hcb;
- search_section_t sections;
- ia64_rr vrr;
-
- hcb = vmx_vcpu_get_vtlb(vcpu);
- data.page_flags=pte & ~PAGE_FLAGS_RV_MASK;
- data.itir=itir;
- data.vadr=PAGEALIGN(ifa,data.ps);
- data.tc = 1;
- data.cl=ISIDE_TLB;
- vmx_vcpu_get_rr(vcpu, ifa, (UINT64 *)&vrr);
- data.rid = vrr.rid;
-
- sections.tr = 1;
- sections.tc = 0;
-
- ovl = vtr_find_overlap(hcb, &data, ISIDE_TLB);
- while (ovl) {
+ int slot;
+ u64 ps, va;
+ thash_cb_t *hcb;
+
+ ps = itir_ps(itir);
+ va = PAGEALIGN(ifa, ps);
+ slot = vtr_find_overlap(vcpu, va, ps, ISIDE_TLB);
+ if (slot >=0) {
// generate MCA.
panic("Tlb conflict!!");
return IA64_FAULT;
}
- thash_purge_and_insert(hcb, &data, ifa);
- return IA64_NO_FAULT;
-}
-
-
-
+ hcb = vmx_vcpu_get_vtlb(vcpu);
+ thash_purge_and_insert(hcb, pte, itir, ifa);
+ return IA64_NO_FAULT;
+}
IA64FAULT vmx_vcpu_itc_d(VCPU *vcpu, UINT64 pte, UINT64 itir, UINT64 ifa)
{
-
- thash_data_t data, *ovl;
- thash_cb_t *hcb;
- search_section_t sections;
- ia64_rr vrr;
-
- hcb = vmx_vcpu_get_vtlb(vcpu);
- data.page_flags=pte & ~PAGE_FLAGS_RV_MASK;
- data.itir=itir;
- data.vadr=PAGEALIGN(ifa,data.ps);
- data.tc = 1;
- data.cl=DSIDE_TLB;
- vmx_vcpu_get_rr(vcpu, ifa,(UINT64 *)&vrr);
- data.rid = vrr.rid;
- sections.tr = 1;
- sections.tc = 0;
-
- ovl = vtr_find_overlap(hcb, &data, DSIDE_TLB);
- if (ovl) {
- // generate MCA.
- panic("Tlb conflict!!");
- return IA64_FAULT;
- }
- thash_purge_and_insert(hcb, &data, ifa);
- return IA64_NO_FAULT;
-}
-
-/*
- * Return TRUE/FALSE for success of lock operation
- */
-
-/*
-int vmx_lock_guest_dtc (VCPU *vcpu, UINT64 va, int lock)
-{
-
- thash_cb_t *hcb;
- ia64_rr vrr;
- u64 preferred_size;
-
- vmx_vcpu_get_rr(vcpu, va, &vrr);
- hcb = vmx_vcpu_get_vtlb(vcpu);
- va = PAGEALIGN(va,vrr.ps);
- preferred_size = PSIZE(vrr.ps);
- return thash_lock_tc(hcb, va, preferred_size, vrr.rid, DSIDE_TLB, lock);
-}
- */
-
-
-
-IA64FAULT vmx_vcpu_itr_i(VCPU *vcpu, UINT64 pte, UINT64 itir, UINT64 ifa,
UINT64 idx)
-{
-
- thash_data_t data, *ovl;
- thash_cb_t *hcb;
- search_section_t sections;
- ia64_rr vrr;
- /* u64 mfn,psr; */
-
- hcb = vmx_vcpu_get_vtlb(vcpu);
- data.page_flags=pte & ~PAGE_FLAGS_RV_MASK;
- data.itir=itir;
- data.vadr=PAGEALIGN(ifa,data.ps);
- data.tc = 0;
- data.cl=ISIDE_TLB;
- vmx_vcpu_get_rr(vcpu, ifa, (UINT64 *)&vrr);
- data.rid = vrr.rid;
- sections.tr = 1;
- sections.tc = 0;
-
-
- ovl = vtr_find_overlap(hcb, &data, ISIDE_TLB);
- if (ovl) {
+ int slot;
+ u64 ps, va, gpfn;
+ thash_cb_t *hcb;
+
+ ps = itir_ps(itir);
+ va = PAGEALIGN(ifa, ps);
+ slot = vtr_find_overlap(vcpu, va, ps, DSIDE_TLB);
+ if (slot >=0) {
// generate MCA.
panic("Tlb conflict!!");
return IA64_FAULT;
}
- sections.tr = 0;
- sections.tc = 1;
- thash_purge_entries(hcb, &data, sections);
-/* if((idx==IA64_TR_KERNEL)&&(data.ps == KERNEL_TR_PAGE_SHIFT)){
- data.contiguous=1;
- }
- */
- thash_tr_insert(hcb, &data, ifa, idx);
-/*
- if((idx==IA64_TR_KERNEL)&&(data.ps == KERNEL_TR_PAGE_SHIFT)){
- mfn = __gpfn_to_mfn_foreign(vcpu->domain,arch_to_xen_ppn(data.ppn));
- data.page_flags=pte&~PAGE_FLAGS_RV_MASK;
- data.ppn = xen_to_arch_ppn(mfn);
- psr = ia64_clear_ic();
- ia64_itr(0x1, IA64_ITR_GUEST_KERNEL, data.vadr, data.page_flags,
data.ps);
- ia64_set_psr(psr); // restore psr
- ia64_srlz_i();
-// return IA64_NO_FAULT;
- }
-*/
- return IA64_NO_FAULT;
-}
-
-IA64FAULT vmx_vcpu_itr_d(VCPU *vcpu, UINT64 pte, UINT64 itir, UINT64 ifa,
UINT64 idx)
-{
-
- thash_data_t data, *ovl;
- thash_cb_t *hcb;
- search_section_t sections;
- ia64_rr vrr;
- /* u64 mfn,psr; */
-
- hcb = vmx_vcpu_get_vtlb(vcpu);
- data.page_flags=pte & ~PAGE_FLAGS_RV_MASK;
- data.itir=itir;
- data.vadr=PAGEALIGN(ifa,data.ps);
- data.tc = 0;
- data.cl=DSIDE_TLB;
- vmx_vcpu_get_rr(vcpu, ifa,(UINT64 *)&vrr);
- data.rid = vrr.rid;
- sections.tr = 1;
- sections.tc = 0;
-
- ovl = vtr_find_overlap(hcb, &data, DSIDE_TLB);
- while (ovl) {
+ hcb = vmx_vcpu_get_vtlb(vcpu);
+ gpfn = (pte & _PAGE_PPN_MASK)>> PAGE_SHIFT;
+ if(__gpfn_is_io(vcpu->domain,gpfn))
+ pte |= VTLB_PTE_IO;
+ thash_purge_and_insert(hcb, pte, itir, ifa);
+ return IA64_NO_FAULT;
+
+}
+
+
+
+
+IA64FAULT vmx_vcpu_itr_i(VCPU *vcpu, u64 slot, u64 pte, u64 itir, u64 ifa)
+{
+ int index;
+ u64 ps, va, rid;
+ thash_cb_t *hcb;
+
+ ps = itir_ps(itir);
+ va = PAGEALIGN(ifa, ps);
+ index = vtr_find_overlap(vcpu, va, ps, ISIDE_TLB);
+ if (index >=0) {
// generate MCA.
panic("Tlb conflict!!");
return IA64_FAULT;
}
- sections.tr = 0;
- sections.tc = 1;
- thash_purge_entries(hcb, &data, sections);
-/*
- if((idx==IA64_TR_KERNEL)&&(data.ps == KERNEL_TR_PAGE_SHIFT)){
- data.contiguous=1;
- }
- */
- thash_tr_insert(hcb, &data, ifa, idx);
-/*
- if((idx==IA64_TR_KERNEL)&&(data.ps == KERNEL_TR_PAGE_SHIFT)){
- mfn = __gpfn_to_mfn_foreign(vcpu->domain,arch_to_xen_ppn(data.ppn));
- data.page_flags=pte&~PAGE_FLAGS_RV_MASK;
- data.ppn = xen_to_arch_ppn(mfn);
- psr = ia64_clear_ic();
- ia64_itr(0x2,IA64_DTR_GUEST_KERNEL , data.vadr, data.page_flags,
data.ps);
- ia64_set_psr(psr); // restore psr
- ia64_srlz_i();
-// return IA64_NO_FAULT;
- }
-*/
-
- return IA64_NO_FAULT;
-}
-
-
-
-IA64FAULT vmx_vcpu_ptr_d(VCPU *vcpu,UINT64 vadr,UINT64 ps)
-{
- thash_cb_t *hcb;
- ia64_rr rr;
- search_section_t sections;
-
- hcb = vmx_vcpu_get_vtlb(vcpu);
- rr=vmx_vcpu_rr(vcpu,vadr);
- sections.tr = 1;
- sections.tc = 1;
- thash_purge_entries_ex(hcb,rr.rid,vadr,ps,sections,DSIDE_TLB);
- return IA64_NO_FAULT;
-}
-
-IA64FAULT vmx_vcpu_ptr_i(VCPU *vcpu,UINT64 vadr,UINT64 ps)
-{
- thash_cb_t *hcb;
- ia64_rr rr;
- search_section_t sections;
- hcb = vmx_vcpu_get_vtlb(vcpu);
- rr=vmx_vcpu_rr(vcpu,vadr);
- sections.tr = 1;
- sections.tc = 1;
- thash_purge_entries_ex(hcb,rr.rid,vadr,ps,sections,ISIDE_TLB);
- return IA64_NO_FAULT;
-}
-
-IA64FAULT vmx_vcpu_ptc_l(VCPU *vcpu, UINT64 vadr, UINT64 ps)
-{
- thash_cb_t *hcb;
- ia64_rr vrr;
- search_section_t sections;
- hcb = vmx_vcpu_get_vtlb(vcpu);
- vrr=vmx_vcpu_rr(vcpu,vadr);
- sections.tr = 0;
- sections.tc = 1;
- vadr = PAGEALIGN(vadr, ps);
-
- thash_purge_entries_ex(hcb,vrr.rid,vadr,ps,sections,DSIDE_TLB);
- thash_purge_entries_ex(hcb,vrr.rid,vadr,ps,sections,ISIDE_TLB);
- return IA64_NO_FAULT;
-}
-
-
-IA64FAULT vmx_vcpu_ptc_e(VCPU *vcpu, UINT64 vadr)
+ hcb = vmx_vcpu_get_vtlb(vcpu);
+ thash_purge_entries(hcb, va, ps);
+ vcpu_get_rr(vcpu, va, &rid);
+ rid = rid& RR_RID_MASK;
+ vmx_vcpu_set_tr((thash_data_t *)&vcpu->arch.itrs[slot], pte, itir, va,
rid);
+ vcpu_quick_region_set(PSCBX(vcpu,itr_regions),va);
+ return IA64_NO_FAULT;
+}
+
+
+IA64FAULT vmx_vcpu_itr_d(VCPU *vcpu, u64 slot, u64 pte, u64 itir, u64 ifa)
+{
+ int index;
+ u64 ps, va, gpfn, rid;
+ thash_cb_t *hcb;
+
+ ps = itir_ps(itir);
+ va = PAGEALIGN(ifa, ps);
+ index = vtr_find_overlap(vcpu, va, ps, DSIDE_TLB);
+ if (index>=0) {
+ // generate MCA.
+ panic("Tlb conflict!!");
+ return IA64_FAULT;
+ }
+ hcb = vmx_vcpu_get_vtlb(vcpu);
+ thash_purge_entries(hcb, va, ps);
+ gpfn = (pte & _PAGE_PPN_MASK)>> PAGE_SHIFT;
+ if(__gpfn_is_io(vcpu->domain,gpfn))
+ pte |= VTLB_PTE_IO;
+ vcpu_get_rr(vcpu, va, &rid);
+ rid = rid& RR_RID_MASK;
+ vmx_vcpu_set_tr((thash_data_t *)&vcpu->arch.dtrs[slot], pte, itir, va,
rid);
+ vcpu_quick_region_set(PSCBX(vcpu,dtr_regions),va);
+ return IA64_NO_FAULT;
+}
+
+
+
+IA64FAULT vmx_vcpu_ptr_d(VCPU *vcpu,UINT64 ifa,UINT64 ps)
+{
+ int index;
+ u64 va;
+ thash_cb_t *hcb;
+
+ va = PAGEALIGN(ifa, ps);
+ index = vtr_find_overlap(vcpu, va, ps, DSIDE_TLB);
+ if (index>=0) {
+ vcpu->arch.dtrs[index].p=0;
+ index = vtr_find_overlap(vcpu, va, ps, DSIDE_TLB);
+ }
+ hcb = vmx_vcpu_get_vtlb(vcpu);
+ thash_purge_entries(hcb, va, ps);
+ return IA64_NO_FAULT;
+}
+
+IA64FAULT vmx_vcpu_ptr_i(VCPU *vcpu,UINT64 ifa,UINT64 ps)
+{
+ int index;
+ u64 va;
+ thash_cb_t *hcb;
+
+ va = PAGEALIGN(ifa, ps);
+ index = vtr_find_overlap(vcpu, va, ps, ISIDE_TLB);
+ if (index>=0) {
+ vcpu->arch.itrs[index].p=0;
+ index = vtr_find_overlap(vcpu, va, ps, ISIDE_TLB);
+ }
+ hcb = vmx_vcpu_get_vtlb(vcpu);
+ thash_purge_entries(hcb, va, ps);
+ return IA64_NO_FAULT;
+}
+
+IA64FAULT vmx_vcpu_ptc_l(VCPU *vcpu, UINT64 va, UINT64 ps)
+{
+ thash_cb_t *hcb;
+ va = PAGEALIGN(va, ps);
+ hcb = vmx_vcpu_get_vtlb(vcpu);
+ thash_purge_entries(hcb, va, ps);
+ return IA64_NO_FAULT;
+}
+
+
+IA64FAULT vmx_vcpu_ptc_e(VCPU *vcpu, UINT64 va)
{
thash_cb_t *hcb;
hcb = vmx_vcpu_get_vtlb(vcpu);
@@ -625,15 +488,15 @@
return IA64_NO_FAULT;
}
-IA64FAULT vmx_vcpu_ptc_g(VCPU *vcpu, UINT64 vadr, UINT64 ps)
-{
- vmx_vcpu_ptc_l(vcpu, vadr, ps);
+IA64FAULT vmx_vcpu_ptc_g(VCPU *vcpu, UINT64 va, UINT64 ps)
+{
+ vmx_vcpu_ptc_l(vcpu, va, ps);
return IA64_ILLOP_FAULT;
}
-IA64FAULT vmx_vcpu_ptc_ga(VCPU *vcpu,UINT64 vadr,UINT64 ps)
-{
- vmx_vcpu_ptc_l(vcpu, vadr, ps);
+IA64FAULT vmx_vcpu_ptc_ga(VCPU *vcpu,UINT64 va,UINT64 ps)
+{
+ vmx_vcpu_ptc_l(vcpu, va, ps);
return IA64_NO_FAULT;
}
@@ -644,7 +507,7 @@
ia64_rr vrr;
u64 vhpt_offset;
vmx_vcpu_get_pta(vcpu, &vpta.val);
- vrr=vmx_vcpu_rr(vcpu, vadr);
+ vcpu_get_rr(vcpu, vadr, &vrr.rrval);
if(vpta.vf){
panic("THASH,Don't support long format VHPT");
*pval = ia64_call_vsa(PAL_VPS_THASH,vadr,vrr.rrval,vpta.val,0,0,0,0);
@@ -663,7 +526,7 @@
ia64_rr vrr;
PTA vpta;
vmx_vcpu_get_pta(vcpu, &vpta.val);
- vrr=vmx_vcpu_rr(vcpu, vadr);
+ vcpu_get_rr(vcpu, vadr, &vrr.rrval);
if(vpta.vf){
panic("THASH,Don't support long format VHPT");
*pval = ia64_call_vsa(PAL_VPS_TTAG,vadr,vrr.rrval,0,0,0,0,0);
@@ -679,13 +542,11 @@
{
thash_data_t *data;
thash_cb_t *hcb;
- ia64_rr vrr;
ISR visr,pt_isr;
REGS *regs;
u64 vhpt_adr;
IA64_PSR vpsr;
hcb = vmx_vcpu_get_vtlb(vcpu);
- vrr=vmx_vcpu_rr(vcpu,vadr);
regs=vcpu_regs(vcpu);
pt_isr.val=VMX(vcpu,cr_isr);
visr.val=0;
@@ -696,7 +557,7 @@
visr.ni=1;
}
visr.na=1;
- data = vtlb_lookup_ex(hcb, vrr.rid, vadr, DSIDE_TLB);
+ data = vtlb_lookup(hcb, vadr, DSIDE_TLB);
if(data){
if(data->p==0){
visr.na=1;
@@ -744,8 +605,7 @@
}
else{
vmx_vcpu_thash(vcpu, vadr, &vhpt_adr);
- vrr=vmx_vcpu_rr(vcpu,vhpt_adr);
- data = vtlb_lookup_ex(hcb, vrr.rid, vhpt_adr, DSIDE_TLB);
+ data = vtlb_lookup(hcb, vhpt_adr, DSIDE_TLB);
if(data){
if(vpsr.ic){
vcpu_set_isr(vcpu, visr.val);
@@ -776,7 +636,6 @@
{
thash_data_t *data;
thash_cb_t *hcb;
- ia64_rr rr;
PTA vpta;
vmx_vcpu_get_pta(vcpu, &vpta.val);
if(vpta.vf==0 || unimplemented_gva(vcpu, vadr)){
@@ -784,8 +643,7 @@
return IA64_NO_FAULT;
}
hcb = vmx_vcpu_get_vtlb(vcpu);
- rr=vmx_vcpu_rr(vcpu,vadr);
- data = vtlb_lookup_ex(hcb, rr.rid, vadr, DSIDE_TLB);
+ data = vtlb_lookup(hcb, vadr, DSIDE_TLB);
if(!data||!data->p){
*key=1;
}else{
@@ -821,11 +679,9 @@
unsigned long end; /* end of the area mapped by current entry */
thash_data_t *entry;
struct vcpu *v = current;
- ia64_rr vrr;
vtlb = vmx_vcpu_get_vtlb(v);
- vrr = vmx_vcpu_rr(v, va);
- entry = vtlb_lookup_ex(vtlb, vrr.rid, va, DSIDE_TLB);
+ entry = vtlb_lookup(vtlb, va, DSIDE_TLB);
if (entry == NULL)
return -EFAULT;
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/ia64/vmx/vmx_hypercall.c
--- a/xen/arch/ia64/vmx/vmx_hypercall.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/ia64/vmx/vmx_hypercall.c Mon Mar 20 08:56:54 2006
@@ -35,6 +35,9 @@
#include <asm/dom_fw.h>
#include <xen/domain.h>
+extern long do_sched_op(int cmd, unsigned long arg);
+extern unsigned long domain_mpa_to_imva(struct domain *,unsigned long mpaddr);
+
void hyper_not_support(void)
{
VCPU *vcpu=current;
@@ -123,7 +126,7 @@
vcpu_set_gr(vcpu, 8, ret, 0);
vmx_vcpu_increment_iip(vcpu);
}
-
+/*
static int do_lock_page(VCPU *vcpu, u64 va, u64 lock)
{
ia64_rr rr;
@@ -132,7 +135,7 @@
rr = vmx_vcpu_rr(vcpu, va);
return thash_lock_tc(hcb, va ,1U<<rr.ps, rr.rid, DSIDE_TLB, lock);
}
-
+ */
/*
* Lock guest page in vTLB, so that it's not relinquished by recycle
* session when HV is servicing that hypercall.
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/ia64/vmx/vmx_init.c
--- a/xen/arch/ia64/vmx/vmx_init.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/ia64/vmx/vmx_init.c Mon Mar 20 08:56:54 2006
@@ -50,6 +50,8 @@
#include <public/arch-ia64.h>
#include <asm/hvm/vioapic.h>
#include <public/event_channel.h>
+#include <xen/event.h>
+#include <asm/vlsapic.h>
/* Global flag to identify whether Intel vmx feature is on */
u32 vmx_enabled = 0;
@@ -96,7 +98,7 @@
if (!(vp_env_info & VP_OPCODE))
printk("WARNING: no opcode provided from hardware(%lx)!!!\n",
vp_env_info);
vm_order = get_order(buffer_size);
- printk("vm buffer size: %ld, order: %ld\n", buffer_size, vm_order);
+ printk("vm buffer size: %ld, order: %d\n", buffer_size, vm_order);
vmx_enabled = 1;
no_vti:
@@ -161,7 +163,7 @@
return NULL;
}
- printk("vpd base: 0x%lx, vpd size:%d\n", vpd, sizeof(vpd_t));
+ printk("vpd base: 0x%p, vpd size:%ld\n", vpd, sizeof(vpd_t));
memset(vpd, 0, VPD_SIZE);
/* CPUID init */
for (i = 0; i < 5; i++)
@@ -234,7 +236,7 @@
{
u64 status;
- status = ia64_pal_vp_restore(v->arch.privregs, 0);
+ status = ia64_pal_vp_restore((u64 *)v->arch.privregs, 0);
if (status != PAL_STATUS_SUCCESS)
panic("Restore vp status failed\n");
@@ -307,7 +309,6 @@
int vmx_alloc_contig_pages(struct domain *d)
{
- unsigned int order;
unsigned long i, j, start,tmp, end, pgnr, conf_nr;
struct page_info *page;
struct vcpu *v = d->vcpu[0];
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/ia64/vmx/vmx_interrupt.c
--- a/xen/arch/ia64/vmx/vmx_interrupt.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/ia64/vmx/vmx_interrupt.c Mon Mar 20 08:56:54 2006
@@ -334,12 +334,13 @@
* @ Nat Consumption Vector
* Refer to SDM Vol2 Table 5-6 & 8-1
*/
-
+#if 0
static void
ir_nat_page_consumption (VCPU *vcpu, u64 vadr)
{
_nat_consumption_fault(vcpu, vadr, DATA);
}
+#endif //shadow it due to no use currently
/*
* Instruction Nat Page Consumption Fault
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/ia64/vmx/vmx_irq_ia64.c
--- a/xen/arch/ia64/vmx/vmx_irq_ia64.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/ia64/vmx/vmx_irq_ia64.c Mon Mar 20 08:56:54 2006
@@ -128,6 +128,6 @@
* come through until ia64_eoi() has been done.
*/
vmx_irq_exit();
- if (current && wake_dom0 != dom0 )
+ if (wake_dom0 && current->domain != dom0 )
vcpu_wake(dom0->vcpu[0]);
}
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/ia64/vmx/vmx_phy_mode.c
--- a/xen/arch/ia64/vmx/vmx_phy_mode.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/ia64/vmx/vmx_phy_mode.c Mon Mar 20 08:56:54 2006
@@ -104,57 +104,51 @@
vcpu->arch.mode_flags = GUEST_IN_PHY;
}
-extern u64 get_mfn(struct domain *d, u64 gpfn);
extern void vmx_switch_rr7(unsigned long ,shared_info_t*,void *,void *,void *);
-void
-physical_itlb_miss_dom0(VCPU *vcpu, u64 vadr)
+/*void
+physical_itlb_miss(VCPU *vcpu, u64 vadr)
{
u64 psr;
IA64_PSR vpsr;
- u64 mppn,gppn;
+ u64 xen_mppn,xen_gppn;
vpsr.val=vmx_vcpu_get_psr(vcpu);
- gppn=(vadr<<1)>>13;
- mppn = get_mfn(vcpu->domain,gppn);
- mppn=(mppn<<12)|(vpsr.cpl<<7);
-// if(vadr>>63)
-// mppn |= PHY_PAGE_UC;
-// else
- mppn |= PHY_PAGE_WB;
+ xen_gppn=(vadr<<1)>>(PAGE_SHIFT+1);
+ xen_mppn = gmfn_to_mfn(vcpu->domain, xen_gppn);
+ xen_mppn=(xen_mppn<<PAGE_SHIFT)|(vpsr.cpl<<7);
+ if(vadr>>63)
+ xen_mppn |= PHY_PAGE_UC;
+ else
+ xen_mppn |= PHY_PAGE_WB;
psr=ia64_clear_ic();
- ia64_itc(1,vadr&(~0xfff),mppn,EMUL_PHY_PAGE_SHIFT);
+ ia64_itc(1,vadr&PAGE_MASK,xen_mppn,PAGE_SHIFT);
ia64_set_psr(psr);
ia64_srlz_i();
return;
}
-
-void
-physical_itlb_miss(VCPU *vcpu, u64 vadr)
-{
- physical_itlb_miss_dom0(vcpu, vadr);
-}
-
-
-void
-physical_dtlb_miss(VCPU *vcpu, u64 vadr)
+*/
+/*
+ * vec=1, itlb miss
+ * vec=2, dtlb miss
+ */
+void
+physical_tlb_miss(VCPU *vcpu, u64 vadr, u64 vec)
{
u64 psr;
IA64_PSR vpsr;
- u64 mppn,gppn;
-// if(vcpu->domain!=dom0)
-// panic("dom n physical dtlb miss happen\n");
+ u64 xen_mppn,xen_gppn;
vpsr.val=vmx_vcpu_get_psr(vcpu);
- gppn=(vadr<<1)>>13;
- mppn = get_mfn(vcpu->domain, gppn);
- mppn=(mppn<<12)|(vpsr.cpl<<7);
+ xen_gppn=(vadr<<1)>>(PAGE_SHIFT+1);
+ xen_mppn = gmfn_to_mfn(vcpu->domain, xen_gppn);
+ xen_mppn=(xen_mppn<<PAGE_SHIFT)|(vpsr.cpl<<7);
if(vadr>>63)
- mppn |= PHY_PAGE_UC;
+ xen_mppn |= PHY_PAGE_UC;
else
- mppn |= PHY_PAGE_WB;
+ xen_mppn |= PHY_PAGE_WB;
psr=ia64_clear_ic();
- ia64_itc(2,vadr&(~0xfff),mppn,EMUL_PHY_PAGE_SHIFT);
+ ia64_itc(vec,vadr&PAGE_MASK,xen_mppn,PAGE_SHIFT);
ia64_set_psr(psr);
ia64_srlz_i();
return;
@@ -193,13 +187,13 @@
if (is_physical_mode(vcpu)) {
if (vcpu->arch.mode_flags & GUEST_PHY_EMUL)
panic("Unexpected domain switch in phy emul\n");
- phy_rr.rrval = vcpu->domain->arch.metaphysical_rr0;
- phy_rr.ps = EMUL_PHY_PAGE_SHIFT;
+ phy_rr.rrval = vcpu->arch.metaphysical_rr0;
+ // phy_rr.ps = PAGE_SHIFT;
phy_rr.ve = 1;
ia64_set_rr((VRN0 << VRN_SHIFT), phy_rr.rrval);
- phy_rr.rrval = vcpu->domain->arch.metaphysical_rr4;
- phy_rr.ps = EMUL_PHY_PAGE_SHIFT;
+ phy_rr.rrval = vcpu->arch.metaphysical_rr4;
+// phy_rr.ps = PAGE_SHIFT;
phy_rr.ve = 1;
ia64_set_rr((VRN4 << VRN_SHIFT), phy_rr.rrval);
@@ -224,7 +218,7 @@
extern void * pal_vaddr;
vmx_switch_rr7(vmx_vrrtomrr(vcpu,VMX(vcpu, vrr[VRN7])),(void
*)vcpu->domain->shared_info,
(void *)vcpu->arch.privregs,
- ( void *)vcpu->arch.vtlb->ts->vhpt->hash, pal_vaddr );
+ (void *)vcpu->arch.vtlb->vhpt->hash, pal_vaddr );
ia64_set_pta(vcpu->arch.arch_vmx.mpta);
ia64_srlz_d();
@@ -242,12 +236,12 @@
/* Save original virtual mode rr[0] and rr[4] */
psr=ia64_clear_ic();
phy_rr.rrval = vcpu->domain->arch.metaphysical_rr0;
- phy_rr.ps = EMUL_PHY_PAGE_SHIFT;
+// phy_rr.ps = EMUL_PHY_PAGE_SHIFT;
phy_rr.ve = 1;
ia64_set_rr(VRN0<<VRN_SHIFT, phy_rr.rrval);
ia64_srlz_d();
phy_rr.rrval = vcpu->domain->arch.metaphysical_rr4;
- phy_rr.ps = EMUL_PHY_PAGE_SHIFT;
+// phy_rr.ps = EMUL_PHY_PAGE_SHIFT;
phy_rr.ve = 1;
ia64_set_rr(VRN4<<VRN_SHIFT, phy_rr.rrval);
ia64_srlz_d();
@@ -266,10 +260,10 @@
psr=ia64_clear_ic();
- mrr=vmx_vcpu_rr(vcpu,VRN0<<VRN_SHIFT);
+ vcpu_get_rr(vcpu,VRN0<<VRN_SHIFT,&mrr.rrval);
ia64_set_rr(VRN0<<VRN_SHIFT, vmx_vrrtomrr(vcpu, mrr.rrval));
ia64_srlz_d();
- mrr=vmx_vcpu_rr(vcpu,VRN4<<VRN_SHIFT);
+ vcpu_get_rr(vcpu,VRN4<<VRN_SHIFT,&mrr.rrval);
ia64_set_rr(VRN4<<VRN_SHIFT, vmx_vrrtomrr(vcpu, mrr.rrval));
ia64_srlz_d();
ia64_set_psr(psr);
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/ia64/vmx/vmx_process.c
--- a/xen/arch/ia64/vmx/vmx_process.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/ia64/vmx/vmx_process.c Mon Mar 20 08:56:54 2006
@@ -58,6 +58,11 @@
extern void die_if_kernel(char *str, struct pt_regs *regs, long err);
extern void rnat_consumption (VCPU *vcpu);
+extern unsigned long translate_domain_mpaddr(unsigned long mpaddr);
+extern void alt_itlb (VCPU *vcpu, u64 vadr);
+extern void itlb_fault (VCPU *vcpu, u64 vadr);
+extern void ivhpt_fault (VCPU *vcpu, u64 vadr);
+
#define DOMN_PAL_REQUEST 0x110000
static UINT64 vec2off[68] = {0x0,0x400,0x800,0xc00,0x1000, 0x1400,0x1800,
@@ -292,10 +297,9 @@
vmx_hpw_miss(u64 vadr , u64 vec, REGS* regs)
{
IA64_PSR vpsr;
- CACHE_LINE_TYPE type=ISIDE_TLB;
+ int type=ISIDE_TLB;
u64 vhpt_adr, gppa;
ISR misr;
- ia64_rr vrr;
// REGS *regs;
thash_cb_t *vtlb;
thash_data_t *data;
@@ -315,34 +319,32 @@
return;
}
*/
- if(vadr == 0x1ea18c00 ){
+/* if(vadr == 0x1ea18c00 ){
ia64_clear_ic();
while(1);
}
+ */
if(is_physical_mode(v)&&(!(vadr<<1>>62))){
- if(vec==1){
- physical_itlb_miss(v, vadr);
- return IA64_FAULT;
- }
if(vec==2){
if(v->domain!=dom0&&__gpfn_is_io(v->domain,(vadr<<1)>>(PAGE_SHIFT+1))){
emulate_io_inst(v,((vadr<<1)>>1),4); // UC
- }else{
- physical_dtlb_miss(v, vadr);
+ return IA64_FAULT;
}
- return IA64_FAULT;
}
- }
- vrr = vmx_vcpu_rr(v, vadr);
+ physical_tlb_miss(v, vadr, vec);
+ return IA64_FAULT;
+ }
if(vec == 1) type = ISIDE_TLB;
else if(vec == 2) type = DSIDE_TLB;
else panic("wrong vec\n");
// prepare_if_physical_mode(v);
- if((data=vtlb_lookup_ex(vtlb, vrr.rid, vadr,type))!=0){
- gppa = (vadr&((1UL<<data->ps)-1))+(data->ppn>>(data->ps-12)<<data->ps);
- if(v->domain!=dom0&&type==DSIDE_TLB &&
__gpfn_is_io(v->domain,gppa>>PAGE_SHIFT)){
+ if((data=vtlb_lookup(vtlb, vadr,type))!=0){
+// gppa = (vadr&((1UL<<data->ps)-1))+(data->ppn>>(data->ps-12)<<data->ps);
+// if(v->domain!=dom0&&type==DSIDE_TLB &&
__gpfn_is_io(v->domain,gppa>>PAGE_SHIFT)){
+ if(v->domain!=dom0 && data->io && type==DSIDE_TLB ){
+ gppa =
(vadr&((1UL<<data->ps)-1))+(data->ppn>>(data->ps-12)<<data->ps);
emulate_io_inst(v, gppa, data->ma);
return IA64_FAULT;
}
@@ -356,7 +358,7 @@
}
else{
*/
- thash_vhpt_insert(vtlb->ts->vhpt,data,vadr);
+ thash_vhpt_insert(vtlb->vhpt,data->page_flags, data->itir ,vadr);
// }
// }
}else if(type == DSIDE_TLB){
@@ -377,8 +379,7 @@
}
} else{
vmx_vcpu_thash(v, vadr, &vhpt_adr);
- vrr=vmx_vcpu_rr(v,vhpt_adr);
- if(vhpt_lookup(vhpt_adr) || vtlb_lookup_ex(vtlb, vrr.rid,
vhpt_adr, DSIDE_TLB)){
+ if(vhpt_lookup(vhpt_adr) || vtlb_lookup(vtlb, vhpt_adr,
DSIDE_TLB)){
if(vpsr.ic){
vcpu_set_isr(v, misr.val);
dtlb_fault(v, vadr);
@@ -420,8 +421,7 @@
return IA64_FAULT;
} else{
vmx_vcpu_thash(v, vadr, &vhpt_adr);
- vrr=vmx_vcpu_rr(v,vhpt_adr);
- if(vhpt_lookup(vhpt_adr) || vtlb_lookup_ex(vtlb, vrr.rid,
vhpt_adr, DSIDE_TLB)){
+ if(vhpt_lookup(vhpt_adr) || vtlb_lookup(vtlb, vhpt_adr,
DSIDE_TLB)){
if(!vpsr.ic){
misr.ni=1;
}
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/ia64/vmx/vmx_vcpu.c
--- a/xen/arch/ia64/vmx/vmx_vcpu.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/ia64/vmx/vmx_vcpu.c Mon Mar 20 08:56:54 2006
@@ -204,32 +204,24 @@
}
-ia64_rr vmx_vcpu_rr(VCPU *vcpu,UINT64 vadr)
-{
- return (ia64_rr)VMX(vcpu,vrr[vadr>>61]);
-}
-
IA64FAULT vmx_vcpu_set_rr(VCPU *vcpu, UINT64 reg, UINT64 val)
{
ia64_rr oldrr,newrr;
thash_cb_t *hcb;
extern void * pal_vaddr;
- oldrr=vmx_vcpu_rr(vcpu,reg);
+ vcpu_get_rr(vcpu, reg, &oldrr.rrval);
newrr.rrval=val;
-#if 1
if(oldrr.ps!=newrr.ps){
hcb = vmx_vcpu_get_vtlb(vcpu);
thash_purge_all(hcb);
}
-#endif
VMX(vcpu,vrr[reg>>61]) = val;
-
switch((u64)(reg>>61)) {
case VRN7:
- vmx_switch_rr7(vmx_vrrtomrr(vcpu,val),vcpu->domain->shared_info,
+ vmx_switch_rr7(vmx_vrrtomrr(vcpu,val),vcpu->domain->shared_info,
(void *)vcpu->arch.privregs,
- ( void *)vcpu->arch.vtlb->ts->vhpt->hash, pal_vaddr );
+ (void *)vcpu->arch.vtlb->vhpt->hash, pal_vaddr );
break;
default:
ia64_set_rr(reg,vmx_vrrtomrr(vcpu,val));
@@ -275,7 +267,7 @@
u64 vmx_vcpu_get_itir_on_fault(VCPU *vcpu, u64 ifa)
{
ia64_rr rr,rr1;
- rr=vmx_vcpu_rr(vcpu,ifa);
+ vcpu_get_rr(vcpu,ifa,&rr.rrval);
rr1.rrval=0;
rr1.ps=rr.ps;
rr1.rid=rr.rid;
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/ia64/vmx/vmx_virt.c
--- a/xen/arch/ia64/vmx/vmx_virt.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/ia64/vmx/vmx_virt.c Mon Mar 20 08:56:54 2006
@@ -34,6 +34,7 @@
#include <asm/virt_event.h>
#include <asm/vmx_phy_mode.h>
extern UINT64 privop_trace;
+extern void vhpi_detection(VCPU *vcpu);//temporarily place here,need a header
file.
void
ia64_priv_decoder(IA64_SLOT_TYPE slot_type, INST64 inst, UINT64 * cause)
@@ -572,7 +573,7 @@
}
#endif // VMAL_NO_FAULT_CHECK
- return (vmx_vcpu_itr_d(vcpu,pte,itir,ifa,slot));
+ return (vmx_vcpu_itr_d(vcpu,slot,pte,itir,ifa));
}
IA64FAULT vmx_emul_itr_i(VCPU *vcpu, INST64 inst)
@@ -631,7 +632,7 @@
}
#endif // VMAL_NO_FAULT_CHECK
- return (vmx_vcpu_itr_i(vcpu,pte,itir,ifa,slot));
+ return (vmx_vcpu_itr_i(vcpu,slot,pte,itir,ifa));
}
IA64FAULT itc_fault_check(VCPU *vcpu, INST64 inst, u64 *itir, u64 *ifa,u64
*pte)
@@ -972,7 +973,7 @@
rsv_reg_field(vcpu);
}
#endif //CHECK_FAULT
- vmx_vcpu_get_rr(vcpu,r3,&r1);
+ vcpu_get_rr(vcpu,r3,&r1);
return vcpu_set_gr(vcpu, inst.M43.r1, r1,0);
}
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/ia64/vmx/vtlb.c
--- a/xen/arch/ia64/vmx/vtlb.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/ia64/vmx/vtlb.c Mon Mar 20 08:56:54 2006
@@ -32,7 +32,7 @@
#include <asm/tlbflush.h>
#define MAX_CCH_LENGTH 40
-thash_data_t *__alloc_chain(thash_cb_t *, thash_data_t *);
+thash_data_t *__alloc_chain(thash_cb_t *);
static void cch_mem_init(thash_cb_t *hcb)
{
@@ -71,126 +71,31 @@
* Check to see if the address rid:va is translated by the TLB
*/
-static int __is_tr_translated(thash_data_t *tlb, u64 rid, u64 va,
CACHE_LINE_TYPE cl)
-{
- u64 size;
- size = PSIZE(tlb->ps);
- if(tlb->vadr&(size-1))
- while(1);
- if ((tlb->rid == rid) && ((va-tlb->vadr)<size))
- return 1;
- else
+static inline int __is_tr_translated(thash_data_t *trp, u64 rid, u64 va)
+{
+ return ((trp->p) && (trp->rid == rid) && ((va-trp->vadr)<PSIZE(trp->ps)));
+}
+
+/*
+ * Only for GUEST TR format.
+ */
+static int
+__is_tr_overlap(thash_data_t *trp, u64 rid, u64 sva, u64 eva)
+{
+ uint64_t sa1, ea1;
+
+ if (!trp->p || trp->rid != rid ) {
return 0;
-}
-
-/*
- * Only for GUEST TR format.
- */
-static int
-__is_tr_overlap(thash_cb_t *hcb,thash_data_t *entry,int rid, char cl, u64 sva,
u64 eva)
-{
- uint64_t size, sa1, ea1;
-
-// if ( entry->invalid || entry->rid != rid || (entry->cl != cl ) ) {
- if ( entry->invalid || entry->rid != rid ) {
- return 0;
- }
- size = PSIZE(entry->ps);
- sa1 = entry->vadr;
- ea1 = sa1 + size -1;
+ }
+ sa1 = trp->vadr;
+ ea1 = sa1 + PSIZE(trp->ps) -1;
eva -= 1;
- if(sa1&(size-1))
- while(1);
if ( (sva>ea1) || (sa1>eva) )
return 0;
else
return 1;
}
-
-static void __rem_tr (thash_cb_t *hcb, thash_data_t *tr)
-{
-/*
- if ( hcb->remove_notifier ) {
- (hcb->remove_notifier)(hcb,tr);
- }
-*/
- tr->invalid = 1;
-}
-
-static inline void __set_tr (thash_data_t *tr, thash_data_t *data, int idx)
-{
- *tr = *data;
- tr->tr_idx = idx;
-}
-
-
-static void __init_tr(thash_cb_t *hcb)
-{
- int i;
- thash_data_t *tr;
-
- for ( i=0, tr = &ITR(hcb,0); i<NITRS; i++ ) {
- tr[i].invalid = 1;
- }
- for ( i=0, tr = &DTR(hcb,0); i<NDTRS; i++ ) {
- tr[i].invalid = 1;
- }
-}
-
-/*
- * Replace TR entry.
- */
-static void rep_tr(thash_cb_t *hcb,thash_data_t *insert, int idx)
-{
- thash_data_t *tr;
-
- if ( insert->cl == ISIDE_TLB ) {
- tr = &ITR(hcb,idx);
- }
- else {
- tr = &DTR(hcb,idx);
- }
- if ( !INVALID_TR(tr) ) {
- __rem_tr(hcb, tr);
- }
- __set_tr (tr, insert, idx);
-}
-
-/*
- * remove TR entry.
- */
-/*
-static void rem_tr(thash_cb_t *hcb,CACHE_LINE_TYPE cl, int idx)
-{
- thash_data_t *tr;
-
- if ( cl == ISIDE_TLB ) {
- tr = &ITR(hcb,idx);
- }
- else {
- tr = &DTR(hcb,idx);
- }
- if ( !INVALID_TR(tr) ) {
- __rem_tr(hcb, tr);
- }
-}
- */
-/*
- * Delete an thash entry in collision chain.
- * prev: the previous entry.
- * rem: the removed entry.
- */
-/*
-static void __rem_chain(thash_cb_t *hcb, thash_data_t *prev, thash_data_t *rem)
-{
- //prev->next = rem->next;
- if ( hcb->remove_notifier ) {
- (hcb->remove_notifier)(hcb,rem);
- }
- cch_free (hcb, rem);
-}
- */
/*
* Delete an thash entry leading collision chain.
@@ -212,69 +117,35 @@
}
}
-thash_data_t *__vtr_lookup(thash_cb_t *hcb,
- u64 rid, u64 va,
- CACHE_LINE_TYPE cl)
-{
- thash_data_t *tr;
- int num,i;
-
- if ( cl == ISIDE_TLB ) {
- tr = &ITR(hcb,0);
- num = NITRS;
+thash_data_t *__vtr_lookup(VCPU *vcpu, u64 va, int is_data)
+{
+
+ thash_data_t *trp;
+ int i;
+ u64 rid;
+ vcpu_get_rr(vcpu, va, &rid);
+ rid = rid&RR_RID_MASK;;
+ if (is_data) {
+ if (vcpu_quick_region_check(vcpu->arch.dtr_regions,va)) {
+ for (trp =(thash_data_t *) vcpu->arch.dtrs,i=0; i<NDTRS; i++,
trp++) {
+ if (__is_tr_translated(trp, rid, va)) {
+ return trp;
+ }
+ }
+ }
}
else {
- tr = &DTR(hcb,0);
- num = NDTRS;
- }
- for ( i=0; i<num; i++ ) {
- if ( !INVALID_TR(&tr[i]) &&
- __is_tr_translated(&tr[i], rid, va, cl) )
- return &tr[i];
+ if (vcpu_quick_region_check(vcpu->arch.itr_regions,va)) {
+ for (trp =(thash_data_t *) vcpu->arch.itrs,i=0; i<NITRS; i++,
trp++) {
+ if (__is_tr_translated(trp, rid, va)) {
+ return trp;
+ }
+ }
+ }
}
return NULL;
}
-
-/*
- * Find overlap VHPT entry within current collision chain
- * base on internal priv info.
- */
-/*
-static inline thash_data_t* _vhpt_next_overlap_in_chain(thash_cb_t *hcb)
-{
- thash_data_t *cch;
- thash_internal_t *priv = &hcb->priv;
-
-
- for (cch=priv->cur_cch; cch; cch = cch->next) {
- if ( priv->tag == cch->etag ) {
- return cch;
- }
- }
- return NULL;
-}
-*/
-/*
- * Find overlap TLB/VHPT entry within current collision chain
- * base on internal priv info.
- */
-/*
-static thash_data_t *_vtlb_next_overlap_in_chain(thash_cb_t *hcb)
-{
- thash_data_t *cch;
- thash_internal_t *priv = &hcb->priv;
-
- // Find overlap TLB entry
- for (cch=priv->cur_cch; cch; cch = cch->next) {
- if ( ( cch->tc ? priv->s_sect.tc : priv->s_sect.tr ) &&
- __is_translated( cch, priv->rid, priv->_curva, priv->cl)) {
- return cch;
- }
- }
- return NULL;
-}
- */
/*
* Get the machine format of VHPT entry.
@@ -292,24 +163,16 @@
* 0/1: means successful or fail.
*
*/
-int __tlb_to_vhpt(thash_cb_t *hcb,
- thash_data_t *tlb, u64 va,
- thash_data_t *vhpt)
+int __tlb_to_vhpt(thash_cb_t *hcb, thash_data_t *vhpt, u64 va)
{
u64 padr,pte;
-// ia64_rr vrr;
ASSERT ( hcb->ht == THASH_VHPT );
-// vrr = (hcb->get_rr_fn)(hcb->vcpu,va);
- padr = tlb->ppn >>(tlb->ps-ARCH_PAGE_SHIFT)<<tlb->ps;
- padr += va&((1UL<<tlb->ps)-1);
+ padr = vhpt->ppn >>(vhpt->ps-ARCH_PAGE_SHIFT)<<vhpt->ps;
+ padr += va&((1UL<<vhpt->ps)-1);
pte=lookup_domain_mpa(current->domain,padr);
if((pte>>56))
return 0;
- // TODO with machine discontinuous address space issue.
vhpt->etag = ia64_ttag(va);
- //vhpt->ti = 0;
- vhpt->itir = tlb->itir & ~ITIR_RV_MASK;
- vhpt->page_flags = tlb->page_flags & ~PAGE_FLAGS_RV_MASK;
vhpt->ps = PAGE_SHIFT;
vhpt->ppn =
(pte&((1UL<<IA64_MAX_PHYS_BITS)-(1UL<<PAGE_SHIFT)))>>ARCH_PAGE_SHIFT;
vhpt->next = 0;
@@ -331,17 +194,20 @@
/* vhpt only has entries with PAGE_SIZE page size */
-void thash_vhpt_insert(thash_cb_t *hcb, thash_data_t *entry, u64 va)
+void thash_vhpt_insert(thash_cb_t *hcb, u64 pte, u64 itir, u64 ifa)
{
thash_data_t vhpt_entry, *hash_table, *cch;
+ vhpt_entry.page_flags = pte & ~PAGE_FLAGS_RV_MASK;
+ vhpt_entry.itir=itir;
+
// ia64_rr vrr;
- if ( !__tlb_to_vhpt(hcb, entry, va, &vhpt_entry) ) {
+ if ( !__tlb_to_vhpt(hcb, &vhpt_entry, ifa) ) {
return;
//panic("Can't convert to machine VHPT entry\n");
}
- hash_table = (thash_data_t *)ia64_thash(va);
+ hash_table = (thash_data_t *)ia64_thash(ifa);
if( INVALID_VHPT(hash_table) ) {
*hash_table = vhpt_entry;
hash_table->next = 0;
@@ -358,6 +224,7 @@
}
cch = cch->next;
}
+
if(hash_table->len>=MAX_CCN_DEPTH){
thash_remove_cch(hcb, hash_table);
cch = cch_alloc(hcb);
@@ -367,9 +234,9 @@
hash_table->next = cch;
return;
}
-
+
// TODO: Add collision chain length limitation.
- cch = __alloc_chain(hcb,entry);
+ cch = __alloc_chain(hcb);
if(cch == NULL){
*hash_table = vhpt_entry;
hash_table->next = 0;
@@ -377,10 +244,8 @@
*cch = *hash_table;
*hash_table = vhpt_entry;
hash_table->next = cch;
- hash_table->len = cch->len + 1;
- cch->len = 0;
-// if(hash_table->tag==hash_table->next->tag)
-// while(1);
+ hash_table->len = cch->len + 1;
+ cch->len = 0;
}
return /*hash_table*/;
@@ -414,7 +279,7 @@
thash_data_t *hash_table, *prev, *next;
u64 start, end, size, tag, rid;
ia64_rr vrr;
- vrr=vmx_vcpu_rr(current, va);
+ vcpu_get_rr(current, va, &vrr.rrval);
rid = vrr.rid;
size = PSIZE(ps);
start = va & (-size);
@@ -480,36 +345,6 @@
}
machine_tlb_purge(va, ps);
}
-/*
- * Insert an entry to hash table.
- * NOTES:
- * 1: TLB entry may be TR, TC or Foreign Map. For TR entry,
- * itr[]/dtr[] need to be updated too.
- * 2: Inserting to collision chain may trigger recycling if
- * the buffer for collision chain is empty.
- * 3: The new entry is inserted at the next of hash table.
- * (I.e. head of the collision chain)
- * 4: The buffer holding the entry is allocated internally
- * from cch_buf or just in the hash table.
- * 5: Return the entry in hash table or collision chain.
- * 6: Input parameter, entry, should be in TLB format.
- * I.e. Has va, rid, ps...
- * 7: This API is invoked by emulating ITC/ITR and tlb_miss.
- *
- */
-
-void thash_tr_insert(thash_cb_t *hcb, thash_data_t *entry, u64 va, int idx)
-{
- if ( hcb->ht != THASH_TLB || entry->tc ) {
- panic("wrong parameter\n");
- }
- entry->vadr = PAGEALIGN(entry->vadr,entry->ps);
- entry->ppn = PAGEALIGN(entry->ppn, entry->ps-12);
- rep_tr(hcb, entry, idx);
-// thash_vhpt_insert(hcb->ts->vhpt, entry, va);
- return ;
-}
-
/*
* Recycle all collisions chain in VTLB or VHPT.
@@ -525,30 +360,13 @@
thash_remove_cch(hcb,hash_table);
}
}
-/*
-thash_data_t *vtlb_alloc_chain(thash_cb_t *hcb,thash_data_t *entry)
+
+thash_data_t *__alloc_chain(thash_cb_t *hcb)
{
thash_data_t *cch;
cch = cch_alloc(hcb);
if(cch == NULL){
- thash_recycle_cch(hcb);
- cch = cch_alloc(hcb);
- }
- return cch;
-}
-*/
-
-thash_data_t *__alloc_chain(thash_cb_t *hcb,thash_data_t *entry)
-{
- thash_data_t *cch;
-
- cch = cch_alloc(hcb);
- if(cch == NULL){
- // recycle
-// if ( hcb->recycle_notifier ) {
-// hcb->recycle_notifier(hcb,(u64)entry);
-// }
thash_recycle_cch(hcb);
cch = cch_alloc(hcb);
}
@@ -564,474 +382,117 @@
* 3: The caller need to make sure the new entry will not overlap
* with any existed entry.
*/
-void vtlb_insert(thash_cb_t *hcb, thash_data_t *entry, u64 va)
+void vtlb_insert(thash_cb_t *hcb, u64 pte, u64 itir, u64 va)
{
thash_data_t *hash_table, *cch;
/* int flag; */
ia64_rr vrr;
/* u64 gppn, ppns, ppne; */
- u64 tag;
- vrr=vmx_vcpu_rr(current, va);
- if (vrr.ps != entry->ps) {
+ u64 tag, ps;
+ ps = itir_ps(itir);
+ vcpu_get_rr(current, va, &vrr.rrval);
+ if (vrr.ps != ps) {
// machine_tlb_insert(hcb->vcpu, entry);
panic("not preferred ps with va: 0x%lx\n", va);
return;
}
- entry->vadr = PAGEALIGN(entry->vadr,entry->ps);
- entry->ppn = PAGEALIGN(entry->ppn, entry->ps-12);
hash_table = vsa_thash(hcb->pta, va, vrr.rrval, &tag);
- entry->etag = tag;
if( INVALID_TLB(hash_table) ) {
- *hash_table = *entry;
+ hash_table->page_flags = pte;
+ hash_table->itir=itir;
+ hash_table->etag=tag;
hash_table->next = 0;
}
else if (hash_table->len>=MAX_CCN_DEPTH){
thash_remove_cch(hcb, hash_table);
cch = cch_alloc(hcb);
*cch = *hash_table;
- *hash_table = *entry;
+ hash_table->page_flags = pte;
+ hash_table->itir=itir;
+ hash_table->etag=tag;
hash_table->len = 1;
hash_table->next = cch;
}
+
else {
// TODO: Add collision chain length limitation.
- cch = __alloc_chain(hcb,entry);
+ cch = __alloc_chain(hcb);
if(cch == NULL){
- *hash_table = *entry;
+ hash_table->page_flags = pte;
+ hash_table->itir=itir;
+ hash_table->etag=tag;
hash_table->next = 0;
}else{
*cch = *hash_table;
- *hash_table = *entry;
+ hash_table->page_flags = pte;
+ hash_table->itir=itir;
+ hash_table->etag=tag;
hash_table->next = cch;
hash_table->len = cch->len + 1;
cch->len = 0;
}
}
-#if 0
- if(hcb->vcpu->domain->domain_id==0){
- thash_insert(hcb->ts->vhpt, entry, va);
- return;
- }
-#endif
-/*
- flag = 1;
- gppn =
(POFFSET(va,entry->ps)|PAGEALIGN((entry->ppn<<12),entry->ps))>>PAGE_SHIFT;
- ppns = PAGEALIGN((entry->ppn<<12),entry->ps);
- ppne = ppns + PSIZE(entry->ps);
- if(((ppns<=0xa0000)&&(ppne>0xa0000))||((ppne>0xc0000)&&(ppns<=0xc0000)))
- flag = 0;
- if((__gpfn_is_mem(hcb->vcpu->domain, gppn)&&flag))
- thash_insert(hcb->ts->vhpt, entry, va);
-*/
return ;
}
-/*
-void thash_insert(thash_cb_t *hcb, thash_data_t *entry, u64 va)
-{
- thash_data_t *hash_table;
- ia64_rr vrr;
-
- vrr = vmx_vcpu_rr(hcb->vcpu,entry->vadr);
- if ( entry->ps != vrr.ps && entry->tc ) {
- panic("Not support for multiple page size now\n");
- }
- entry->vadr = PAGEALIGN(entry->vadr,entry->ps);
- entry->ppn = PAGEALIGN(entry->ppn, entry->ps-12);
- (hcb->ins_hash)(hcb, entry, va);
-
-}
-*/
-/*
-static void rem_thash(thash_cb_t *hcb, thash_data_t *entry)
-{
- thash_data_t *hash_table, *p, *q;
- thash_internal_t *priv = &hcb->priv;
- int idx;
-
- hash_table = priv->hash_base;
- if ( hash_table == entry ) {
-// if ( PURGABLE_ENTRY(hcb, entry) ) {
- __rem_hash_head (hcb, entry);
-// }
- return ;
- }
- // remove from collision chain
- p = hash_table;
- for ( q=p->next; q; q = p->next ) {
- if ( q == entry ){
-// if ( PURGABLE_ENTRY(hcb,q ) ) {
- p->next = q->next;
- __rem_chain(hcb, entry);
- hash_table->len--;
-// }
- return ;
- }
- p = q;
- }
- panic("Entry not existed or bad sequence\n");
-}
-*/
-/*
-static void rem_vtlb(thash_cb_t *hcb, thash_data_t *entry)
-{
- thash_data_t *hash_table, *p, *q;
- thash_internal_t *priv = &hcb->priv;
- int idx;
-
- if ( !entry->tc ) {
- return rem_tr(hcb, entry->cl, entry->tr_idx);
- }
- rem_thash(hcb, entry);
-}
-*/
-int cch_depth=0;
-/*
- * Purge the collision chain starting from cch.
- * NOTE:
- * For those UN-Purgable entries(FM), this function will return
- * the head of left collision chain.
- */
-/*
-static thash_data_t *thash_rem_cch(thash_cb_t *hcb, thash_data_t *cch)
-{
- thash_data_t *next;
-
-// if ( ++cch_depth > MAX_CCH_LENGTH ) {
-// printf ("cch length > MAX_CCH_LENGTH, exceed the expected length\n");
-// while(1);
-// }
- if ( cch -> next ) {
- next = thash_rem_cch(hcb, cch->next);
+int vtr_find_overlap(VCPU *vcpu, u64 va, u64 ps, int is_data)
+{
+ thash_data_t *trp;
+ int i;
+ u64 end, rid;
+ vcpu_get_rr(vcpu, va, &rid);
+ rid = rid&RR_RID_MASK;;
+ end = va + PSIZE(ps);
+ if (is_data) {
+ if (vcpu_quick_region_check(vcpu->arch.dtr_regions,va)) {
+ for (trp =(thash_data_t *) vcpu->arch.dtrs,i=0; i<NDTRS; i++,
trp++) {
+ if (__is_tr_overlap(trp, rid, va, end )) {
+ return i;
+ }
+ }
+ }
}
else {
- next = NULL;
- }
- if ( PURGABLE_ENTRY(hcb, cch) ) {
- __rem_chain(hcb, cch);
- return next;
- }
- else {
- cch->next = next;
- return cch;
- }
-}
- */
-
-/*
- * Purge one hash line (include the entry in hash table).
- * Can only be called by thash_purge_all.
- * Input:
- * hash: The head of collision chain (hash table)
- *
- */
-/*
-static void thash_rem_line(thash_cb_t *hcb, thash_data_t *hash)
-{
- if ( INVALID_ENTRY(hcb, hash) ) return;
-
- if ( hash->next ) {
- cch_depth = 0;
- hash->next = thash_rem_cch(hcb, hash->next);
- }
- // Then hash table itself.
- if ( PURGABLE_ENTRY(hcb, hash) ) {
- __rem_hash_head(hcb, hash);
- }
-}
- */
-
-/*
- * Find an overlap entry in hash table and its collision chain.
- * Refer to SDM2 4.1.1.4 for overlap definition.
- * PARAS:
- * 1: in: TLB format entry, rid:ps must be same with vrr[].
- * va & ps identify the address space for overlap lookup
- * 2: section can be combination of TR, TC and FM. (THASH_SECTION_XX)
- * 3: cl means I side or D side.
- * RETURNS:
- * NULL to indicate the end of findings.
- * NOTES:
- *
- */
-
-/*
-thash_data_t *thash_find_overlap(thash_cb_t *hcb,
- thash_data_t *in, search_section_t s_sect)
-{
- return (hcb->find_overlap)(hcb, in->vadr,
- PSIZE(in->ps), in->rid, in->cl, s_sect);
-}
-*/
-
-/*
-static thash_data_t *vtlb_find_overlap(thash_cb_t *hcb,
- u64 va, u64 size, int rid, char cl, search_section_t s_sect)
-{
- thash_data_t *hash_table;
- thash_internal_t *priv = &hcb->priv;
- u64 tag;
- ia64_rr vrr;
-
- priv->_curva = va & ~(size-1);
- priv->_eva = priv->_curva + size;
- priv->rid = rid;
- vrr = vmx_vcpu_rr(hcb->vcpu,va);
- priv->ps = vrr.ps;
- hash_table = vsa_thash(hcb->pta, priv->_curva, vrr.rrval, &tag);
- priv->s_sect = s_sect;
- priv->cl = cl;
- priv->_tr_idx = 0;
- priv->hash_base = hash_table;
- priv->cur_cch = hash_table;
- return (hcb->next_overlap)(hcb);
-}
-*/
-
-/*
-static thash_data_t *vhpt_find_overlap(thash_cb_t *hcb,
- u64 va, u64 size, int rid, char cl, search_section_t s_sect)
-{
- thash_data_t *hash_table;
- thash_internal_t *priv = &hcb->priv;
- u64 tag;
- ia64_rr vrr;
-
- priv->_curva = va & ~(size-1);
- priv->_eva = priv->_curva + size;
- priv->rid = rid;
- vrr = vmx_vcpu_rr(hcb->vcpu,va);
- priv->ps = vrr.ps;
- hash_table = ia64_thash(priv->_curva);
- tag = ia64_ttag(priv->_curva);
- priv->tag = tag;
- priv->hash_base = hash_table;
- priv->cur_cch = hash_table;
- return (hcb->next_overlap)(hcb);
-}
-*/
-
-
-thash_data_t *vtr_find_overlap(thash_cb_t *hcb, thash_data_t *data, char cl)
-{
- thash_data_t *tr;
- int i,num;
- u64 end;
-
- if (cl == ISIDE_TLB ) {
- num = NITRS;
- tr = &ITR(hcb,0);
- }
- else {
- num = NDTRS;
- tr = &DTR(hcb,0);
- }
- end=data->vadr + PSIZE(data->ps);
- for (i=0; i<num; i++ ) {
- if ( __is_tr_overlap(hcb, &tr[i], data->rid, cl, data->vadr, end )) {
- return &tr[i];
- }
- }
- return NULL;
-}
-
-
-/*
-static thash_data_t *vtr_find_next_overlap(thash_cb_t *hcb)
-{
- thash_data_t *tr;
- thash_internal_t *priv = &hcb->priv;
- int num;
-
- if ( priv->cl == ISIDE_TLB ) {
- num = NITRS;
- tr = &ITR(hcb,0);
- }
- else {
- num = NDTRS;
- tr = &DTR(hcb,0);
- }
- for (; priv->_tr_idx < num; priv->_tr_idx ++ ) {
- if ( __is_tr_overlap(hcb, &tr[priv->_tr_idx],
- priv->rid, priv->cl,
- priv->_curva, priv->_eva) ) {
- return &tr[priv->_tr_idx++];
- }
- }
- return NULL;
-}
-*/
-
-/*
- * Similar with vtlb_next_overlap but find next entry.
- * NOTES:
- * Intermediate position information is stored in hcb->priv.
- */
-/*
-static thash_data_t *vtlb_next_overlap(thash_cb_t *hcb)
-{
- thash_data_t *ovl;
- thash_internal_t *priv = &hcb->priv;
- u64 addr,rr_psize,tag;
- ia64_rr vrr;
-
- if ( priv->s_sect.tr ) {
- ovl = vtr_find_next_overlap (hcb);
- if ( ovl ) return ovl;
- priv->s_sect.tr = 0;
- }
- if ( priv->s_sect.v == 0 ) return NULL;
- vrr = vmx_vcpu_rr(hcb->vcpu,priv->_curva);
- rr_psize = PSIZE(vrr.ps);
-
- while ( priv->_curva < priv->_eva ) {
- if ( !INVALID_ENTRY(hcb, priv->hash_base) ) {
- ovl = _vtlb_next_overlap_in_chain(hcb);
- if ( ovl ) {
- priv->cur_cch = ovl->next;
- return ovl;
+ if (vcpu_quick_region_check(vcpu->arch.itr_regions,va)) {
+ for (trp =(thash_data_t *) vcpu->arch.itrs,i=0; i<NITRS; i++,
trp++) {
+ if (__is_tr_overlap(trp, rid, va, end )) {
+ return i;
+ }
}
}
- priv->_curva += rr_psize;
- priv->hash_base = vsa_thash( hcb->pta, priv->_curva, vrr.rrval, &tag);
- priv->cur_cch = priv->hash_base;
- }
- return NULL;
-}
- */
-
-
-/*
-static thash_data_t *vhpt_next_overlap(thash_cb_t *hcb)
-{
- thash_data_t *ovl;
- thash_internal_t *priv = &hcb->priv;
- u64 addr,rr_psize;
- ia64_rr vrr;
-
- vrr = vmx_vcpu_rr(hcb->vcpu,priv->_curva);
- rr_psize = PSIZE(vrr.ps);
-
- while ( priv->_curva < priv->_eva ) {
- if ( !INVALID_ENTRY(hcb, priv->hash_base) ) {
- ovl = _vhpt_next_overlap_in_chain(hcb);
- if ( ovl ) {
- priv->cur_cch = ovl->next;
- return ovl;
- }
- }
- priv->_curva += rr_psize;
- priv->hash_base = ia64_thash(priv->_curva);
- priv->tag = ia64_ttag(priv->_curva);
- priv->cur_cch = priv->hash_base;
- }
- return NULL;
-}
-*/
-
-/*
- * Find and purge overlap entries in hash table and its collision chain.
- * PARAS:
- * 1: in: TLB format entry, rid:ps must be same with vrr[].
- * rid, va & ps identify the address space for purge
- * 2: section can be combination of TR, TC and FM. (thash_SECTION_XX)
- * 3: cl means I side or D side.
- * NOTES:
- *
- */
-void thash_purge_entries(thash_cb_t *hcb,
- thash_data_t *in, search_section_t p_sect)
-{
- return thash_purge_entries_ex(hcb, in->rid, in->vadr,
- in->ps, p_sect, in->cl);
-}
-
-void thash_purge_entries_ex(thash_cb_t *hcb,
- u64 rid, u64 va, u64 ps,
- search_section_t p_sect,
- CACHE_LINE_TYPE cl)
-{
-/*
- thash_data_t *ovl;
-
- ovl = (hcb->find_overlap)(hcb, va, PSIZE(ps), rid, cl, p_sect);
- while ( ovl != NULL ) {
- (hcb->rem_hash)(hcb, ovl);
- ovl = (hcb->next_overlap)(hcb);
- };
- */
+ }
+ return -1;
+}
+
+/*
+ * Purge entries in VTLB and VHPT
+ */
+void thash_purge_entries(thash_cb_t *hcb, u64 va, u64 ps)
+{
vtlb_purge(hcb, va, ps);
- vhpt_purge(hcb->ts->vhpt, va, ps);
-}
+ vhpt_purge(hcb->vhpt, va, ps);
+}
+
/*
* Purge overlap TCs and then insert the new entry to emulate itc ops.
* Notes: Only TC entry can purge and insert.
*/
-void thash_purge_and_insert(thash_cb_t *hcb, thash_data_t *in, u64 va)
-{
- /* thash_data_t *ovl; */
- search_section_t sections;
-
-#ifdef XEN_DEBUGGER
- vrr = vmx_vcpu_rr(hcb->vcpu,in->vadr);
- if ( in->ps != vrr.ps || hcb->ht != THASH_TLB || !in->tc ) {
- panic ("Oops, wrong call for purge_and_insert\n");
- return;
- }
-#endif
- in->vadr = PAGEALIGN(in->vadr,in->ps);
- in->ppn = PAGEALIGN(in->ppn, in->ps-12);
- sections.tr = 0;
- sections.tc = 1;
-/*
- ovl = (hcb->find_overlap)(hcb, in->vadr, PSIZE(in->ps),
- in->rid, in->cl, sections);
- if(ovl)
- (hcb->rem_hash)(hcb, ovl);
- */
- vtlb_purge(hcb, va, in->ps);
- vhpt_purge(hcb->ts->vhpt, va, in->ps);
-#ifdef XEN_DEBUGGER
- ovl = (hcb->next_overlap)(hcb);
- if ( ovl ) {
- panic ("Oops, 2+ overlaps for purge_and_insert\n");
- return;
- }
-#endif
- if(in->ps!=PAGE_SHIFT)
- vtlb_insert(hcb, in, va);
- thash_vhpt_insert(hcb->ts->vhpt, in, va);
-}
-/*
- * Purge one hash line (include the entry in hash table).
- * Can only be called by thash_purge_all.
- * Input:
- * hash: The head of collision chain (hash table)
- *
- */
-/*
-static void thash_purge_line(thash_cb_t *hcb, thash_data_t *hash)
-{
- if ( INVALID_ENTRY(hcb, hash) ) return;
- thash_data_t *prev, *next;
- next=hash->next;
- while ( next ) {
- prev=next;
- next=next->next;
- cch_free(hcb, prev);
- }
- // Then hash table itself.
- INVALIDATE_HASH(hcb, hash);
-}
-*/
-
-
-
-
-
-
+void thash_purge_and_insert(thash_cb_t *hcb, u64 pte, u64 itir, u64 ifa)
+{
+ u64 ps, va;
+ ps = itir_ps(itir);
+ va = PAGEALIGN(ifa,ps);
+ vtlb_purge(hcb, va, ps);
+ vhpt_purge(hcb->vhpt, va, ps);
+ if((ps!=PAGE_SHIFT)||(pte&VTLB_PTE_IO))
+ vtlb_insert(hcb, pte, itir, va);
+ if(!(pte&VTLB_PTE_IO)){
+ va = PAGEALIGN(ifa,PAGE_SHIFT);
+ thash_vhpt_insert(hcb->vhpt, pte, itir, va);
+ }
+}
@@ -1064,27 +525,12 @@
}
cch_mem_init (hcb);
- vhpt = hcb->ts->vhpt;
+ vhpt = hcb->vhpt;
hash_table = (thash_data_t*)((u64)vhpt->hash + vhpt->hash_sz);
for (--hash_table;(u64)hash_table >= (u64)vhpt->hash;hash_table--) {
INVALIDATE_VHPT_HEADER(hash_table);
}
cch_mem_init (vhpt);
-
-/*
- entry = &hcb->ts->itr[0];
- for(i=0; i< (NITRS+NDTRS); i++){
- if(!INVALID_TLB(entry)){
- start=entry->vadr & (-PSIZE(entry->ps));
- end = start + PSIZE(entry->ps);
- while(start<end){
- thash_vhpt_insert(vhpt, entry, start);
- start += PAGE_SIZE;
- }
- }
- entry++;
- }
-*/
local_flush_tlb_all();
}
@@ -1096,100 +542,32 @@
* INPUT:
* in: TLB format for both VHPT & TLB.
*/
-thash_data_t *vtlb_lookup(thash_cb_t *hcb,
- thash_data_t *in)
-{
- return vtlb_lookup_ex(hcb, in->rid, in->vadr, in->cl);
-}
-
-thash_data_t *vtlb_lookup_ex(thash_cb_t *hcb,
- u64 rid, u64 va,
- CACHE_LINE_TYPE cl)
+
+thash_data_t *vtlb_lookup(thash_cb_t *hcb, u64 va,int is_data)
{
thash_data_t *hash_table, *cch;
u64 tag;
ia64_rr vrr;
-
+
ASSERT ( hcb->ht == THASH_TLB );
-
- cch = __vtr_lookup(hcb, rid, va, cl);;
+
+ cch = __vtr_lookup(hcb->vcpu, va, is_data);;
if ( cch ) return cch;
- vrr = vmx_vcpu_rr(hcb->vcpu,va);
+ vcpu_get_rr(hcb->vcpu,va,&vrr.rrval);
hash_table = vsa_thash( hcb->pta, va, vrr.rrval, &tag);
if ( INVALID_ENTRY(hcb, hash_table ) )
return NULL;
-
+
for (cch=hash_table; cch; cch = cch->next) {
-// if ( __is_translated(cch, rid, va, cl) )
if(cch->etag == tag)
return cch;
}
return NULL;
}
-/*
- * Lock/Unlock TC if found.
- * NOTES: Only the page in prefered size can be handled.
- * return:
- * 1: failure
- * 0: success
- */
-/*
-int thash_lock_tc(thash_cb_t *hcb, u64 va, u64 size, int rid, char cl, int
lock)
-{
- thash_data_t *ovl;
- search_section_t sections;
-
- sections.tr = 1;
- sections.tc = 1;
- ovl = (hcb->find_overlap)(hcb, va, size, rid, cl, sections);
- if ( ovl ) {
- if ( !ovl->tc ) {
-// panic("Oops, TR for lock\n");
- return 0;
- }
- else if ( lock ) {
- if ( ovl->locked ) {
- DPRINTK("Oops, already locked entry\n");
- }
- ovl->locked = 1;
- }
- else if ( !lock ) {
- if ( !ovl->locked ) {
- DPRINTK("Oops, already unlocked entry\n");
- }
- ovl->locked = 0;
- }
- return 0;
- }
- return 1;
-}
-*/
-
-/*
- * Notifier when TLB is deleted from hash table and its collision chain.
- * NOTES:
- * The typical situation is that TLB remove needs to inform
- * VHPT to remove too.
- * PARAS:
- * 1: hcb is TLB object.
- * 2: The format of entry is always in TLB.
- *
- */
-//void tlb_remove_notifier(thash_cb_t *hcb, thash_data_t *entry)
-//{
-// vhpt_purge(hcb->ts->vhpt,entry->vadr,entry->ps);
-// thash_cb_t *vhpt;
-
-// search_section_t s_sect;
-
-// s_sect.v = 0;
-// thash_purge_entries(hcb->ts->vhpt, entry, s_sect);
-// machine_tlb_purge(entry->vadr, entry->ps);
-//}
/*
* Initialize internal control data before service.
@@ -1206,28 +584,15 @@
hcb->pta.size = sz;
// hcb->get_rr_fn = vmmu_get_rr;
ASSERT ( hcb->hash_sz % sizeof(thash_data_t) == 0 );
- if ( hcb->ht == THASH_TLB ) {
-// hcb->remove_notifier = NULL; //tlb_remove_notifier;
-// hcb->find_overlap = vtlb_find_overlap;
-// hcb->next_overlap = vtlb_next_overlap;
-// hcb->rem_hash = rem_vtlb;
-// hcb->ins_hash = vtlb_insert;
- __init_tr(hcb);
- }
- else {
-// hcb->remove_notifier = NULL;
-// hcb->find_overlap = vhpt_find_overlap;
-// hcb->next_overlap = vhpt_next_overlap;
-// hcb->rem_hash = rem_thash;
-// hcb->ins_hash = thash_vhpt_insert;
- }
hash_table = (thash_data_t*)((u64)hcb->hash + hcb->hash_sz);
for (--hash_table;(u64)hash_table >= (u64)hcb->hash;hash_table--) {
INVALIDATE_HASH_HEADER(hcb,hash_table);
}
}
+
#ifdef VTLB_DEBUG
+/*
static u64 cch_length_statistics[MAX_CCH_LENGTH+1];
u64 sanity_check=0;
u64 vtlb_chain_sanity(thash_cb_t *vtlb, thash_cb_t *vhpt, thash_data_t *hash)
@@ -1264,7 +629,7 @@
thash_data_t *hash, *cch;
thash_data_t *ovl;
search_section_t s_sect;
- thash_cb_t *vhpt = vtlb->ts->vhpt;
+ thash_cb_t *vhpt = vtlb->vhpt;
u64 invalid_ratio;
if ( sanity_check == 0 ) return;
@@ -1403,4 +768,5 @@
}
printf("End of vTLB dump\n");
}
+*/
#endif
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/ia64/xen/dom_fw.c
--- a/xen/arch/ia64/xen/dom_fw.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/ia64/xen/dom_fw.c Mon Mar 20 08:56:54 2006
@@ -47,7 +47,7 @@
}
// builds a hypercall bundle at domain physical address
-void dom_efi_hypercall_patch(struct domain *d, unsigned long paddr, unsigned
long hypercall)
+static void dom_efi_hypercall_patch(struct domain *d, unsigned long paddr,
unsigned long hypercall)
{
unsigned long *imva;
@@ -96,122 +96,13 @@
# define NUM_MEM_DESCS 5
-#define SECS_PER_HOUR (60 * 60)
-#define SECS_PER_DAY (SECS_PER_HOUR * 24)
-
-/* Compute the `struct tm' representation of *T,
- offset OFFSET seconds east of UTC,
- and store year, yday, mon, mday, wday, hour, min, sec into *TP.
- Return nonzero if successful. */
-int
-offtime (unsigned long t, efi_time_t *tp)
-{
- const unsigned short int __mon_yday[2][13] =
- {
- /* Normal years. */
- { 0, 31, 59, 90, 120, 151, 181, 212, 243, 273, 304, 334, 365 },
- /* Leap years. */
- { 0, 31, 60, 91, 121, 152, 182, 213, 244, 274, 305, 335, 366 }
- };
- long int days, rem, y;
- const unsigned short int *ip;
-
- days = t / SECS_PER_DAY;
- rem = t % SECS_PER_DAY;
- while (rem < 0) {
- rem += SECS_PER_DAY;
- --days;
- }
- while (rem >= SECS_PER_DAY) {
- rem -= SECS_PER_DAY;
- ++days;
- }
- tp->hour = rem / SECS_PER_HOUR;
- rem %= SECS_PER_HOUR;
- tp->minute = rem / 60;
- tp->second = rem % 60;
- /* January 1, 1970 was a Thursday. */
- y = 1970;
-
-# define DIV(a, b) ((a) / (b) - ((a) % (b) < 0))
-# define LEAPS_THRU_END_OF(y) (DIV (y, 4) - DIV (y, 100) + DIV (y, 400))
-# define __isleap(year) \
- ((year) % 4 == 0 && ((year) % 100 != 0 || (year) % 400 == 0))
-
- while (days < 0 || days >= (__isleap (y) ? 366 : 365)) {
- /* Guess a corrected year, assuming 365 days per year. */
- long int yg = y + days / 365 - (days % 365 < 0);
-
- /* Adjust DAYS and Y to match the guessed year. */
- days -= ((yg - y) * 365 + LEAPS_THRU_END_OF (yg - 1)
- - LEAPS_THRU_END_OF (y - 1));
- y = yg;
- }
- tp->year = y;
- ip = __mon_yday[__isleap(y)];
- for (y = 11; days < (long int) ip[y]; --y)
- continue;
- days -= ip[y];
- tp->month = y + 1;
- tp->day = days + 1;
- return 1;
-}
-
-/* Macro to emulate SAL call using legacy IN and OUT calls to CF8, CFC etc.. */
-
-#define BUILD_CMD(addr) ((0x80000000 | (addr)) & ~3)
-
-#define REG_OFFSET(addr) (0x00000000000000FF & (addr))
-#define DEVICE_FUNCTION(addr) (0x000000000000FF00 & (addr))
-#define BUS_NUMBER(addr) (0x0000000000FF0000 & (addr))
-
-#ifndef XEN
-static efi_status_t
-fw_efi_get_time (efi_time_t *tm, efi_time_cap_t *tc)
-{
-#if defined(CONFIG_IA64_HP_SIM) || defined(CONFIG_IA64_GENERIC)
- struct {
- int tv_sec; /* must be 32bits to work */
- int tv_usec;
- } tv32bits;
-
- ssc((unsigned long) &tv32bits, 0, 0, 0, SSC_GET_TOD);
-
- memset(tm, 0, sizeof(*tm));
- offtime(tv32bits.tv_sec, tm);
-
- if (tc)
- memset(tc, 0, sizeof(*tc));
-#else
-# error Not implemented yet...
-#endif
- return EFI_SUCCESS;
-}
-
-static void
-efi_reset_system (int reset_type, efi_status_t status, unsigned long
data_size, efi_char16_t *data)
-{
-#if defined(CONFIG_IA64_HP_SIM) || defined(CONFIG_IA64_GENERIC)
- ssc(status, 0, 0, 0, SSC_EXIT);
-#else
-# error Not implemented yet...
-#endif
-}
-
-static efi_status_t
-efi_unimplemented (void)
-{
- return EFI_UNSUPPORTED;
-}
-#endif /* !XEN */
-
struct sal_ret_values
sal_emulator (long index, unsigned long in1, unsigned long in2,
unsigned long in3, unsigned long in4, unsigned long in5,
unsigned long in6, unsigned long in7)
{
- long r9 = 0;
- long r10 = 0;
+ unsigned long r9 = 0;
+ unsigned long r10 = 0;
long r11 = 0;
long status;
@@ -285,12 +176,11 @@
}
struct ia64_pal_retval
-xen_pal_emulator(unsigned long index, unsigned long in1,
- unsigned long in2, unsigned long in3)
-{
- long r9 = 0;
- long r10 = 0;
- long r11 = 0;
+xen_pal_emulator(unsigned long index, u64 in1, u64 in2, u64 in3)
+{
+ unsigned long r9 = 0;
+ unsigned long r10 = 0;
+ unsigned long r11 = 0;
long status = -1;
if (running_on_sim) return pal_emulator_static(index);
@@ -364,7 +254,7 @@
&r10);
break;
case PAL_REGISTER_INFO:
- status = ia64_pal_register_info(in1,&r9,&r10);
+ status = ia64_pal_register_info(in1, &r9, &r10);
break;
case PAL_CACHE_FLUSH:
/* FIXME */
@@ -434,7 +324,7 @@
#define NFUNCPTRS 20
-void print_md(efi_memory_desc_t *md)
+static void print_md(efi_memory_desc_t *md)
{
#if 1
printk("domain mem: type=%u, attr=0x%lx, range=[0x%016lx-0x%016lx)
(%luMB)\n",
@@ -496,7 +386,7 @@
}
/* base is physical address of acpi table */
-void touch_acpi_table(void)
+static void touch_acpi_table(void)
{
if (acpi_table_parse_madt(ACPI_MADT_LSAPIC, acpi_update_lsapic, 0) < 0)
printk("Error parsing MADT - no LAPIC entires\n");
@@ -514,15 +404,15 @@
struct acpi_table_header dsdt;
u8 aml[16];
struct acpi_table_madt madt;
- struct acpi_table_lsapic lsapic;
+ struct acpi_table_lsapic lsapic[MAX_VIRT_CPUS];
u8 pm1a_evt_blk[4];
u8 pm1a_cnt_blk[1];
u8 pm_tmr_blk[4];
};
/* Create enough of an ACPI structure to make the guest OS ACPI happy. */
-void
-dom_fw_fake_acpi(struct fake_acpi_tables *tables)
+static void
+dom_fw_fake_acpi(struct domain *d, struct fake_acpi_tables *tables)
{
struct acpi20_table_rsdp *rsdp = &tables->rsdp;
struct xsdt_descriptor_rev2 *xsdt = &tables->xsdt;
@@ -530,7 +420,8 @@
struct facs_descriptor_rev2 *facs = &tables->facs;
struct acpi_table_header *dsdt = &tables->dsdt;
struct acpi_table_madt *madt = &tables->madt;
- struct acpi_table_lsapic *lsapic = &tables->lsapic;
+ struct acpi_table_lsapic *lsapic = tables->lsapic;
+ int i;
memset(tables, 0, sizeof(struct fake_acpi_tables));
@@ -608,13 +499,13 @@
/* Trivial namespace, avoids ACPI CA complaints */
tables->aml[0] = 0x10; /* Scope */
tables->aml[1] = 0x12; /* length/offset to next object */
- strncpy(&tables->aml[2], "_SB_", 4);
+ strncpy((char *)&tables->aml[2], "_SB_", 4);
/* The processor object isn't absolutely necessary, revist for SMP */
tables->aml[6] = 0x5b; /* processor object */
tables->aml[7] = 0x83;
tables->aml[8] = 0x0b; /* next */
- strncpy(&tables->aml[9], "CPU0", 4);
+ strncpy((char *)&tables->aml[9], "CPU0", 4);
dsdt->checksum = generate_acpi_checksum(dsdt, dsdt->length);
@@ -622,16 +513,20 @@
strncpy(madt->header.signature, APIC_SIG, 4);
madt->header.revision = 2;
madt->header.length = sizeof(struct acpi_table_madt) +
- sizeof(struct acpi_table_lsapic);
+ MAX_VIRT_CPUS * sizeof(struct acpi_table_lsapic);
strcpy(madt->header.oem_id, "XEN");
strcpy(madt->header.oem_table_id, "Xen/ia64");
strcpy(madt->header.asl_compiler_id, "XEN");
madt->header.asl_compiler_revision = (XEN_VERSION<<16)|(XEN_SUBVERSION);
- /* A single LSAPIC entry describes the CPU. Revisit for SMP guests */
- lsapic->header.type = ACPI_MADT_LSAPIC;
- lsapic->header.length = sizeof(struct acpi_table_lsapic);
- lsapic->flags.enabled = 1;
+ /* An LSAPIC entry describes a CPU. */
+ for (i = 0; i < MAX_VIRT_CPUS; i++) {
+ lsapic[i].header.type = ACPI_MADT_LSAPIC;
+ lsapic[i].header.length = sizeof(struct acpi_table_lsapic);
+ lsapic[i].id = i;
+ lsapic[i].eid = 0;
+ lsapic[i].flags.enabled = (d->vcpu[i] != NULL);
+ }
madt->header.checksum = generate_acpi_checksum(madt,
madt->header.length);
@@ -785,7 +680,7 @@
acpi_tables = (void *)cp;
cp += sizeof(struct fake_acpi_tables);
- dom_fw_fake_acpi(acpi_tables);
+ dom_fw_fake_acpi(d, acpi_tables);
efi_tables[i].guid = ACPI_20_TABLE_GUID;
efi_tables[i].table = dom_pa((unsigned long)
acpi_tables);
@@ -801,8 +696,8 @@
sal_systab->sal_rev_major = 0;
sal_systab->entry_count = 1;
- strcpy(sal_systab->oem_id, "Xen/ia64");
- strcpy(sal_systab->product_id, "Xen/ia64");
+ strcpy((char *)sal_systab->oem_id, "Xen/ia64");
+ strcpy((char *)sal_systab->product_id, "Xen/ia64");
/* fill in an entry point: */
sal_ed->type = SAL_DESC_ENTRY_POINT;
@@ -861,7 +756,10 @@
/* hypercall patches live here, masquerade as reserved PAL
memory */
MAKE_MD(EFI_PAL_CODE,EFI_MEMORY_WB,HYPERCALL_START,HYPERCALL_END, 1);
MAKE_MD(EFI_CONVENTIONAL_MEMORY,EFI_MEMORY_WB,HYPERCALL_END,maxmem, 1);
- MAKE_MD(EFI_RESERVED_TYPE,0,0,0,0);
+ /* Create a dummy entry for IO ports, so that IO accesses are
+ trapped by Xen. */
+ MAKE_MD(EFI_MEMORY_MAPPED_IO_PORT_SPACE,EFI_MEMORY_UC,
+ 0x00000ffffc000000, 0x00000fffffffffff, 1);
MAKE_MD(EFI_RESERVED_TYPE,0,0,0,0);
}
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/ia64/xen/domain.c
--- a/xen/arch/ia64/xen/domain.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/ia64/xen/domain.c Mon Mar 20 08:56:54 2006
@@ -26,6 +26,7 @@
#include <asm/processor.h>
#include <asm/desc.h>
#include <asm/hw_irq.h>
+#include <asm/setup.h>
//#include <asm/mpspec.h>
#include <xen/irq.h>
#include <xen/event.h>
@@ -36,7 +37,6 @@
#include <xen/elf.h>
//#include <asm/page.h>
#include <asm/pgalloc.h>
-#include <asm/dma.h> /* for MAX_DMA_ADDRESS */
#include <asm/asm-offsets.h> /* for IA64_THREAD_INFO_SIZE */
@@ -49,6 +49,9 @@
#include <asm/pal.h>
#include <asm/vhpt.h>
#include <public/hvm/ioreq.h>
+#include <public/arch-ia64.h>
+#include <asm/tlbflush.h>
+#include <asm/regionreg.h>
#define CONFIG_DOMAIN0_CONTIGUOUS
unsigned long dom0_start = -1L;
@@ -69,10 +72,7 @@
/* FIXME: where these declarations should be there ? */
extern void domain_pend_keyboard_interrupt(int);
extern long platform_is_hp_ski(void);
-extern unsigned long allocate_metaphysical_rr(void);
-extern int allocate_rid_range(struct domain *, unsigned long);
extern void sync_split_caches(void);
-extern void init_all_rr(struct vcpu *);
extern void serial_input_init(void);
static void init_switch_stack(struct vcpu *v);
@@ -80,9 +80,33 @@
/* this belongs in include/asm, but there doesn't seem to be a suitable place
*/
void arch_domain_destroy(struct domain *d)
{
- printf("arch_domain_destroy: not implemented\n");
- //free_page((unsigned long)d->mm.perdomain_pt);
- free_xenheap_page(d->shared_info);
+ struct page_info *page;
+ struct list_head *ent, *prev;
+
+ if (d->arch.mm->pgd != NULL)
+ {
+ list_for_each ( ent, &d->arch.mm->pt_list )
+ {
+ page = list_entry(ent, struct page_info, list);
+ prev = ent->prev;
+ list_del(ent);
+ free_xenheap_page(page_to_virt(page));
+ ent = prev;
+ }
+ pgd_free(d->arch.mm->pgd);
+ }
+ if (d->arch.mm != NULL)
+ xfree(d->arch.mm);
+ if (d->shared_info != NULL)
+ free_xenheap_page(d->shared_info);
+
+ deallocate_rid_range(d);
+
+ /* It is really good in this? */
+ flush_tlb_all();
+
+ /* It is really good in this? */
+ vhpt_flush();
}
static void default_idle(void)
@@ -115,23 +139,9 @@
void startup_cpu_idle_loop(void)
{
- int cpu = smp_processor_id ();
/* Just some sanity to ensure that the scheduler is set up okay. */
ASSERT(current->domain == IDLE_DOMAIN_ID);
- printf ("idle%dA\n", cpu);
raise_softirq(SCHEDULE_SOFTIRQ);
-#if 0 /* All this work is done within continue_cpu_idle_loop */
- printf ("idle%dB\n", cpu);
- asm volatile ("mov ar.k2=r0");
- do_softirq();
- printf ("idle%dC\n", cpu);
-
- /*
- * Declares CPU setup done to the boot processor.
- * Therefore memory barrier to ensure state is visible.
- */
- smp_mb();
-#endif
#if 0
//do we have to ensure the idle task has a shared page so that, for example,
//region registers can be loaded from it. Apparently not...
@@ -201,6 +211,8 @@
void free_vcpu_struct(struct vcpu *v)
{
+ if (v->arch.privregs != NULL)
+ free_xenheap_pages(v->arch.privregs,
get_order(sizeof(mapped_regs_t)));
free_xenheap_pages(v, KERNEL_STACK_SIZE_ORDER);
}
@@ -253,6 +265,7 @@
if ((d->arch.mm = xmalloc(struct mm_struct)) == NULL)
goto fail_nomem;
memset(d->arch.mm, 0, sizeof(*d->arch.mm));
+ INIT_LIST_HEAD(&d->arch.mm->pt_list);
if ((d->arch.mm->pgd = pgd_alloc(d->arch.mm)) == NULL)
goto fail_nomem;
@@ -324,10 +337,74 @@
return 0;
}
+static void relinquish_memory(struct domain *d, struct list_head *list)
+{
+ struct list_head *ent;
+ struct page_info *page;
+#ifndef __ia64__
+ unsigned long x, y;
+#endif
+
+ /* Use a recursive lock, as we may enter 'free_domheap_page'. */
+ spin_lock_recursive(&d->page_alloc_lock);
+ ent = list->next;
+ while ( ent != list )
+ {
+ page = list_entry(ent, struct page_info, list);
+ /* Grab a reference to the page so it won't disappear from under us. */
+ if ( unlikely(!get_page(page, d)) )
+ {
+ /* Couldn't get a reference -- someone is freeing this page. */
+ ent = ent->next;
+ continue;
+ }
+
+ if ( test_and_clear_bit(_PGT_pinned, &page->u.inuse.type_info) )
+ put_page_and_type(page);
+
+ if ( test_and_clear_bit(_PGC_allocated, &page->count_info) )
+ put_page(page);
+
+#ifndef __ia64__
+ /*
+ * Forcibly invalidate base page tables at this point to break circular
+ * 'linear page table' references. This is okay because MMU structures
+ * are not shared across domains and this domain is now dead. Thus base
+ * tables are not in use so a non-zero count means circular reference.
+ */
+ y = page->u.inuse.type_info;
+ for ( ; ; )
+ {
+ x = y;
+ if ( likely((x & (PGT_type_mask|PGT_validated)) !=
+ (PGT_base_page_table|PGT_validated)) )
+ break;
+
+ y = cmpxchg(&page->u.inuse.type_info, x, x & ~PGT_validated);
+ if ( likely(y == x) )
+ {
+ free_page_type(page, PGT_base_page_table);
+ break;
+ }
+ }
+#endif
+
+ /* Follow the list chain and /then/ potentially free the page. */
+ ent = ent->next;
+ put_page(page);
+ }
+
+ spin_unlock_recursive(&d->page_alloc_lock);
+}
+
void domain_relinquish_resources(struct domain *d)
{
- /* FIXME */
- printf("domain_relinquish_resources: not implemented\n");
+ /* Relinquish every page of memory. */
+
+ /* xenheap_list is not used in ia64. */
+ BUG_ON(!list_empty(&d->xenpage_list));
+
+ relinquish_memory(d, &d->page_list);
}
// heavily leveraged from linux/arch/ia64/kernel/process.c:copy_thread()
@@ -339,7 +416,7 @@
{
struct domain *d = v->domain;
struct pt_regs *regs;
- extern char saved_command_line[];
+ extern char dom0_command_line[];
#ifdef CONFIG_DOMAIN0_CONTIGUOUS
if (d == dom0) start_pc += dom0_start;
@@ -351,8 +428,9 @@
regs->cr_ipsr = 0x501008826008; /* Need to be expanded as macro
*/
} else {
regs->cr_ipsr = ia64_getreg(_IA64_REG_PSR)
- | IA64_PSR_BITS_TO_SET | IA64_PSR_BN
- & ~(IA64_PSR_BITS_TO_CLEAR | IA64_PSR_RI | IA64_PSR_IS);
+ | IA64_PSR_BITS_TO_SET | IA64_PSR_BN;
+ regs->cr_ipsr &= ~(IA64_PSR_BITS_TO_CLEAR
+ | IA64_PSR_RI | IA64_PSR_IS);
regs->cr_ipsr |= 2UL << IA64_PSR_CPL0_BIT; // domain runs at PL2
}
regs->cr_iip = start_pc;
@@ -362,24 +440,27 @@
if (VMX_DOMAIN(v)) {
vmx_init_all_rr(v);
if (d == dom0)
-// VCPU(v,vgr[12]) = dom_fw_setup(d,saved_command_line,256L);
- regs->r28 = dom_fw_setup(d,saved_command_line,256L);
+ regs->r28 = dom_fw_setup(d,dom0_command_line,
+ COMMAND_LINE_SIZE);
/* Virtual processor context setup */
VCPU(v, vpsr) = IA64_PSR_BN;
VCPU(v, dcr) = 0;
} else {
init_all_rr(v);
if (d == dom0)
- regs->r28 = dom_fw_setup(d,saved_command_line,256L);
+ regs->r28 = dom_fw_setup(d,dom0_command_line,
+ COMMAND_LINE_SIZE);
else {
regs->ar_rsc |= (2 << 2); /* force PL2/3 */
if (*d->arch.cmdline == '\0') {
#define DEFAULT_CMDLINE "nomca nosmp xencons=tty0 console=tty0 root=/dev/hda1"
- regs->r28 = dom_fw_setup(d,DEFAULT_CMDLINE,256L);
+ regs->r28 = dom_fw_setup(d,DEFAULT_CMDLINE,
+ sizeof (DEFAULT_CMDLINE));
printf("domU command line defaulted to"
DEFAULT_CMDLINE "\n");
}
- else regs->r28 = dom_fw_setup(d,d->arch.cmdline,256L);
+ else regs->r28 = dom_fw_setup(d,d->arch.cmdline,
+ IA64_COMMAND_LINE_SIZE);
}
VCPU(v, banknum) = 1;
VCPU(v, metaphysical_mode) = 1;
@@ -387,7 +468,7 @@
}
}
-static struct page * assign_new_domain0_page(unsigned long mpaddr)
+static struct page_info * assign_new_domain0_page(unsigned long mpaddr)
{
if (mpaddr < dom0_start || mpaddr >= dom0_start + dom0_size) {
printk("assign_new_domain0_page: bad domain0 mpaddr
0x%lx!\n",mpaddr);
@@ -399,10 +480,10 @@
}
/* allocate new page for domain and map it to the specified metaphysical addr
*/
-struct page * assign_new_domain_page(struct domain *d, unsigned long mpaddr)
+struct page_info * assign_new_domain_page(struct domain *d, unsigned long
mpaddr)
{
struct mm_struct *mm = d->arch.mm;
- struct page *p = (struct page *)0;
+ struct page_info *pt, *p = (struct page_info *)0;
pgd_t *pgd;
pud_t *pud;
pmd_t *pmd;
@@ -414,16 +495,28 @@
}
pgd = pgd_offset(mm,mpaddr);
if (pgd_none(*pgd))
+ {
pgd_populate(mm, pgd, pud_alloc_one(mm,mpaddr));
+ pt = maddr_to_page(pgd_val(*pgd));
+ list_add_tail(&pt->list, &d->arch.mm->pt_list);
+ }
pud = pud_offset(pgd, mpaddr);
if (pud_none(*pud))
+ {
pud_populate(mm, pud, pmd_alloc_one(mm,mpaddr));
+ pt = maddr_to_page(pud_val(*pud));
+ list_add_tail(&pt->list, &d->arch.mm->pt_list);
+ }
pmd = pmd_offset(pud, mpaddr);
if (pmd_none(*pmd))
+ {
pmd_populate_kernel(mm, pmd, pte_alloc_one_kernel(mm,mpaddr));
// pmd_populate(mm, pmd, pte_alloc_one(mm,mpaddr));
+ pt = maddr_to_page(pmd_val(*pmd));
+ list_add_tail(&pt->list, &d->arch.mm->pt_list);
+ }
pte = pte_offset_map(pmd, mpaddr);
if (pte_none(*pte)) {
@@ -456,6 +549,7 @@
void assign_domain_page(struct domain *d, unsigned long mpaddr, unsigned long
physaddr)
{
struct mm_struct *mm = d->arch.mm;
+ struct page_info *pt;
pgd_t *pgd;
pud_t *pud;
pmd_t *pmd;
@@ -467,16 +561,28 @@
}
pgd = pgd_offset(mm,mpaddr);
if (pgd_none(*pgd))
+ {
pgd_populate(mm, pgd, pud_alloc_one(mm,mpaddr));
+ pt = maddr_to_page(pgd_val(*pgd));
+ list_add_tail(&pt->list, &d->arch.mm->pt_list);
+ }
pud = pud_offset(pgd, mpaddr);
if (pud_none(*pud))
+ {
pud_populate(mm, pud, pmd_alloc_one(mm,mpaddr));
+ pt = maddr_to_page(pud_val(*pud));
+ list_add_tail(&pt->list, &d->arch.mm->pt_list);
+ }
pmd = pmd_offset(pud, mpaddr);
if (pmd_none(*pmd))
+ {
pmd_populate_kernel(mm, pmd, pte_alloc_one_kernel(mm,mpaddr));
// pmd_populate(mm, pmd, pte_alloc_one(mm,mpaddr));
+ pt = maddr_to_page(pmd_val(*pmd));
+ list_add_tail(&pt->list, &d->arch.mm->pt_list);
+ }
pte = pte_offset_map(pmd, mpaddr);
if (pte_none(*pte)) {
@@ -543,12 +649,13 @@
#ifdef CONFIG_DOMAIN0_CONTIGUOUS
if (d == dom0) {
+ pte_t pteval;
if (mpaddr < dom0_start || mpaddr >= dom0_start + dom0_size) {
//printk("lookup_domain_mpa: bad dom0 mpaddr
0x%lx!\n",mpaddr);
//printk("lookup_domain_mpa:
start=0x%lx,end=0x%lx!\n",dom0_start,dom0_start+dom0_size);
mpafoo(mpaddr);
}
- pte_t pteval = pfn_pte(mpaddr >> PAGE_SHIFT,
+ pteval = pfn_pte(mpaddr >> PAGE_SHIFT,
__pgprot(__DIRTY_BITS | _PAGE_PL_2 | _PAGE_AR_RWX));
pte = &pteval;
return *(unsigned long *)pte;
@@ -639,7 +746,7 @@
Elf_Phdr phdr;
int h, filesz, memsz;
unsigned long elfaddr, dom_mpaddr, dom_imva;
- struct page *p;
+ struct page_info *p;
copy_memory(&ehdr, (void *) image_start, sizeof(Elf_Ehdr));
for ( h = 0; h < ehdr.e_phnum; h++ ) {
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/ia64/xen/gdbstub.c
--- a/xen/arch/ia64/xen/gdbstub.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/ia64/xen/gdbstub.c Mon Mar 20 08:56:54 2006
@@ -32,6 +32,7 @@
#include <xen/lib.h>
+#include <xen/mm.h>
#include <asm/byteorder.h>
#include <asm/debugger.h>
#include <asm/uaccess.h>
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/ia64/xen/irq.c
--- a/xen/arch/ia64/xen/irq.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/ia64/xen/irq.c Mon Mar 20 08:56:54 2006
@@ -1338,6 +1338,7 @@
struct domain *guest[IRQ_MAX_GUESTS];
} irq_guest_action_t;
+/*
static void __do_IRQ_guest(int irq)
{
irq_desc_t *desc = &irq_desc[irq];
@@ -1353,7 +1354,7 @@
send_guest_pirq(d, irq);
}
}
-
+ */
int pirq_guest_unmask(struct domain *d)
{
irq_desc_t *desc;
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/ia64/xen/mm_init.c
--- a/xen/arch/ia64/xen/mm_init.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/ia64/xen/mm_init.c Mon Mar 20 08:56:54 2006
@@ -60,16 +60,16 @@
#ifdef CONFIG_VIRTUAL_MEM_MAP
unsigned long vmalloc_end = VMALLOC_END_INIT;
EXPORT_SYMBOL(vmalloc_end);
-struct page *vmem_map;
+struct page_info *vmem_map;
EXPORT_SYMBOL(vmem_map);
#endif
// static int pgt_cache_water[2] = { 25, 50 };
-struct page *zero_page_memmap_ptr; /* map entry for zero page */
+#ifndef XEN
+struct page_info *zero_page_memmap_ptr; /* map entry for zero page */
EXPORT_SYMBOL(zero_page_memmap_ptr);
-#ifdef XEN
void *high_memory;
EXPORT_SYMBOL(high_memory);
@@ -172,7 +172,7 @@
pte_t fastcall * pte_alloc_map(struct mm_struct *mm, pmd_t *pmd, unsigned long
address)
{
if (!pmd_present(*pmd)) {
- struct page *new;
+ struct page_info *new;
spin_unlock(&mm->page_table_lock);
new = pte_alloc_one(mm, address);
@@ -202,7 +202,7 @@
update_mmu_cache (struct vm_area_struct *vma, unsigned long vaddr, pte_t pte)
{
unsigned long addr;
- struct page *page;
+ struct page_info *page;
if (!pte_exec(pte))
return; /* not an executable page... */
@@ -386,7 +386,7 @@
create_mem_map_page_table (u64 start, u64 end, void *arg)
{
unsigned long address, start_page, end_page;
- struct page *map_start, *map_end;
+ struct page_info *map_start, *map_end;
int node;
pgd_t *pgd;
pmd_t *pmd;
@@ -417,8 +417,8 @@
}
struct memmap_init_callback_data {
- struct page *start;
- struct page *end;
+ struct page_info *start;
+ struct page_info *end;
int nid;
unsigned long zone;
};
@@ -427,7 +427,7 @@
virtual_memmap_init (u64 start, u64 end, void *arg)
{
struct memmap_init_callback_data *args;
- struct page *map_start, *map_end;
+ struct page_info *map_start, *map_end;
args = (struct memmap_init_callback_data *) arg;
@@ -440,13 +440,13 @@
map_end = args->end;
/*
- * We have to initialize "out of bounds" struct page elements that fit
completely
+ * We have to initialize "out of bounds" struct page_info elements that
fit completely
* on the same pages that were allocated for the "in bounds" elements
because they
* may be referenced later (and found to be "reserved").
*/
- map_start -= ((unsigned long) map_start & (PAGE_SIZE - 1)) /
sizeof(struct page);
+ map_start -= ((unsigned long) map_start & (PAGE_SIZE - 1)) /
sizeof(struct page_info);
map_end += ((PAGE_ALIGN((unsigned long) map_end) - (unsigned long)
map_end)
- / sizeof(struct page));
+ / sizeof(struct page_info));
if (map_start < map_end)
memmap_init_zone(map_start, (unsigned long) (map_end -
map_start),
@@ -455,7 +455,7 @@
}
void
-memmap_init (struct page *start, unsigned long size, int nid,
+memmap_init (struct page_info *start, unsigned long size, int nid,
unsigned long zone, unsigned long start_pfn)
{
if (!vmem_map)
@@ -476,7 +476,7 @@
ia64_mfn_valid (unsigned long pfn)
{
char byte;
- struct page *pg = mfn_to_page(pfn);
+ struct page_info *pg = mfn_to_page(pfn);
return (__get_user(byte, (char *) pg) == 0)
&& ((((u64)pg & PAGE_MASK) == (((u64)(pg + 1) - 1) & PAGE_MASK))
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/ia64/xen/privop.c
--- a/xen/arch/ia64/xen/privop.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/ia64/xen/privop.c Mon Mar 20 08:56:54 2006
@@ -19,6 +19,9 @@
extern void zero_reflect_counts(void);
long priv_verbose=0;
+
+/* Set to 1 to handle privified instructions from the privify tool. */
+static const int privify_en = 0;
/**************************************************************************
Hypercall bundle creation
@@ -131,7 +134,8 @@
UINT src = inst.M28.r3;
// NOTE: ptc_e with source gr > 63 is emulated as a fc r(y-64)
- if (src > 63) return(vcpu_fc(vcpu,vcpu_get_gr(vcpu,src - 64)));
+ if (privify_en && src > 63)
+ return(vcpu_fc(vcpu,vcpu_get_gr(vcpu,src - 64)));
return vcpu_ptc_e(vcpu,vcpu_get_gr(vcpu,src));
}
@@ -178,7 +182,7 @@
UINT src = inst.M46.r3;
// NOTE: tpa with source gr > 63 is emulated as a ttag rx=r(y-64)
- if (src > 63)
+ if (privify_en && src > 63)
fault = vcpu_ttag(vcpu,vcpu_get_gr(vcpu,src-64),&padr);
else fault = vcpu_tpa(vcpu,vcpu_get_gr(vcpu,src),&padr);
if (fault == IA64_NO_FAULT)
@@ -193,7 +197,7 @@
UINT src = inst.M46.r3;
// NOTE: tak with source gr > 63 is emulated as a thash rx=r(y-64)
- if (src > 63)
+ if (privify_en && src > 63)
fault = vcpu_thash(vcpu,vcpu_get_gr(vcpu,src-64),&key);
else fault = vcpu_tak(vcpu,vcpu_get_gr(vcpu,src),&key);
if (fault == IA64_NO_FAULT)
@@ -280,7 +284,8 @@
// I26 and M29 are identical for these fields
UINT64 ar3 = inst.M29.ar3;
- if (inst.M29.r2 > 63 && inst.M29.ar3 < 8) { // privified mov from kr
+ if (privify_en && inst.M29.r2 > 63 && inst.M29.ar3 < 8) {
+ // privified mov from kr
UINT64 val;
if (vcpu_get_ar(vcpu,ar3,&val) != IA64_ILLOP_FAULT)
return vcpu_set_gr(vcpu, inst.M29.r2-64, val,0);
@@ -404,14 +409,17 @@
{
UINT64 val;
IA64FAULT fault;
+ int reg;
- if (inst.M43.r1 > 63) { // privified mov from cpuid
- fault = vcpu_get_cpuid(vcpu,vcpu_get_gr(vcpu,inst.M43.r3),&val);
+ reg = vcpu_get_gr(vcpu,inst.M43.r3);
+ if (privify_en && inst.M43.r1 > 63) {
+ // privified mov from cpuid
+ fault = vcpu_get_cpuid(vcpu,reg,&val);
if (fault == IA64_NO_FAULT)
return vcpu_set_gr(vcpu, inst.M43.r1-64, val, 0);
}
else {
- fault = vcpu_get_rr(vcpu,vcpu_get_gr(vcpu,inst.M43.r3),&val);
+ fault = vcpu_get_rr(vcpu,reg,&val);
if (fault == IA64_NO_FAULT)
return vcpu_set_gr(vcpu, inst.M43.r1, val, 0);
}
@@ -455,14 +463,17 @@
{
UINT64 val;
IA64FAULT fault;
+ int reg;
- if (inst.M43.r1 > 63) { // privified mov from pmd
- fault = vcpu_get_pmd(vcpu,vcpu_get_gr(vcpu,inst.M43.r3),&val);
+ reg = vcpu_get_gr(vcpu,inst.M43.r3);
+ if (privify_en && inst.M43.r1 > 63) {
+ // privified mov from pmd
+ fault = vcpu_get_pmd(vcpu,reg,&val);
if (fault == IA64_NO_FAULT)
return vcpu_set_gr(vcpu, inst.M43.r1-64, val, 0);
}
else {
- fault = vcpu_get_pmc(vcpu,vcpu_get_gr(vcpu,inst.M43.r3),&val);
+ fault = vcpu_get_pmc(vcpu,reg,&val);
if (fault == IA64_NO_FAULT)
return vcpu_set_gr(vcpu, inst.M43.r1, val, 0);
}
@@ -666,7 +677,7 @@
else if (inst.generic.major != 1) break;
x6 = inst.M29.x6;
if (x6 == 0x2a) {
- if (inst.M29.r2 > 63 && inst.M29.ar3 < 8)
+ if (privify_en && inst.M29.r2 > 63 && inst.M29.ar3 < 8)
privcnt.mov_from_ar++; // privified mov from kr
else privcnt.mov_to_ar_reg++;
return priv_mov_to_ar_reg(vcpu,inst);
@@ -674,14 +685,14 @@
if (inst.M29.x3 != 0) break;
if (!(pfunc = Mpriv_funcs[x6])) break;
if (x6 == 0x1e || x6 == 0x1f) { // tpa or tak are "special"
- if (inst.M46.r3 > 63) {
+ if (privify_en && inst.M46.r3 > 63) {
if (x6 == 0x1e) x6 = 0x1b;
else x6 = 0x1a;
}
}
- if (x6 == 52 && inst.M28.r3 > 63)
+ if (privify_en && x6 == 52 && inst.M28.r3 > 63)
privcnt.fc++;
- else if (x6 == 16 && inst.M43.r3 > 63)
+ else if (privify_en && x6 == 16 && inst.M43.r3 > 63)
privcnt.cpuid++;
else privcnt.Mpriv_cnt[x6]++;
return (*pfunc)(vcpu,inst);
@@ -718,7 +729,7 @@
#endif
if (inst.I26.x3 != 0) break; // I26.x3 == I27.x3
if (inst.I26.x6 == 0x2a) {
- if (inst.I26.r2 > 63 && inst.I26.ar3 < 8)
+ if (privify_en && inst.I26.r2 > 63 && inst.I26.ar3 < 8)
privcnt.mov_from_ar++; // privified mov from kr
else privcnt.mov_to_ar_reg++;
return priv_mov_to_ar_reg(vcpu,inst);
@@ -797,12 +808,17 @@
#define HYPERPRIVOP_GET_RR 0x10
#define HYPERPRIVOP_SET_RR 0x11
#define HYPERPRIVOP_SET_KR 0x12
-#define HYPERPRIVOP_MAX 0x12
+#define HYPERPRIVOP_FC 0x13
+#define HYPERPRIVOP_GET_CPUID 0x14
+#define HYPERPRIVOP_GET_PMD 0x15
+#define HYPERPRIVOP_GET_EFLAG 0x16
+#define HYPERPRIVOP_SET_EFLAG 0x17
+#define HYPERPRIVOP_MAX 0x17
static const char * const hyperpriv_str[HYPERPRIVOP_MAX+1] = {
0, "rfi", "rsm.dt", "ssm.dt", "cover", "itc.d", "itc.i", "ssm.i",
"=ivr", "=tpr", "tpr=", "eoi", "itm=", "thash", "ptc.ga", "itr.d",
- "=rr", "rr=", "kr="
+ "=rr", "rr=", "kr=", "fc", "=cpuid", "=pmd", "=ar.eflg", "ar.eflg="
};
unsigned long slow_hyperpriv_cnt[HYPERPRIVOP_MAX+1] = { 0 };
@@ -888,6 +904,24 @@
return 1;
case HYPERPRIVOP_SET_KR:
(void)vcpu_set_ar(v,regs->r8,regs->r9);
+ return 1;
+ case HYPERPRIVOP_FC:
+ (void)vcpu_fc(v,regs->r8);
+ return 1;
+ case HYPERPRIVOP_GET_CPUID:
+ (void)vcpu_get_cpuid(v,regs->r8,&val);
+ regs->r8 = val;
+ return 1;
+ case HYPERPRIVOP_GET_PMD:
+ (void)vcpu_get_pmd(v,regs->r8,&val);
+ regs->r8 = val;
+ return 1;
+ case HYPERPRIVOP_GET_EFLAG:
+ (void)vcpu_get_ar(v,24,&val);
+ regs->r8 = val;
+ return 1;
+ case HYPERPRIVOP_SET_EFLAG:
+ (void)vcpu_set_ar(v,24,regs->r8);
return 1;
}
return 0;
@@ -934,7 +968,7 @@
};
// FIXME: should use snprintf to ensure no buffer overflow
-int dump_privop_counts(char *buf)
+static int dump_privop_counts(char *buf)
{
int i, j;
UINT64 sum = 0;
@@ -1007,7 +1041,7 @@
return s - buf;
}
-int zero_privop_counts(char *buf)
+static int zero_privop_counts(char *buf)
{
int i, j;
char *s = buf;
@@ -1043,7 +1077,7 @@
v->overflow++;;
}
-int dump_privop_addrs(char *buf)
+static int dump_privop_addrs(char *buf)
{
int i,j;
char *s = buf;
@@ -1061,7 +1095,7 @@
return s - buf;
}
-void zero_privop_addrs(void)
+static void zero_privop_addrs(void)
{
int i,j;
for (i = 0; i < PRIVOP_COUNT_NINSTS; i++) {
@@ -1085,7 +1119,7 @@
extern unsigned long pal_halt_light_count;
extern unsigned long context_switch_count;
-int dump_misc_stats(char *buf)
+static int dump_misc_stats(char *buf)
{
char *s = buf;
s += sprintf(s,"Virtual TR translations: %ld\n",tr_translate_count);
@@ -1102,7 +1136,7 @@
return s - buf;
}
-void zero_misc_stats(void)
+static void zero_misc_stats(void)
{
dtlb_translate_count = 0;
tr_translate_count = 0;
@@ -1117,7 +1151,7 @@
context_switch_count = 0;
}
-int dump_hyperprivop_counts(char *buf)
+static int dump_hyperprivop_counts(char *buf)
{
int i;
char *s = buf;
@@ -1138,7 +1172,7 @@
return s - buf;
}
-void zero_hyperprivop_counts(void)
+static void zero_hyperprivop_counts(void)
{
int i;
for (i = 0; i <= HYPERPRIVOP_MAX; i++) slow_hyperpriv_cnt[i] = 0;
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/ia64/xen/process.c
--- a/xen/arch/ia64/xen/process.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/ia64/xen/process.c Mon Mar 20 08:56:54 2006
@@ -1,3 +1,4 @@
+
/*
* Miscellaneous process/domain related routines
*
@@ -41,6 +42,7 @@
extern int ia64_hyperprivop(unsigned long, REGS *);
extern int ia64_hypercall(struct pt_regs *regs);
extern void vmx_do_launch(struct vcpu *);
+extern unsigned long lookup_domain_mpa(struct domain *,unsigned long);
extern unsigned long dom0_start, dom0_size;
@@ -57,20 +59,7 @@
IA64_PSR_ID | IA64_PSR_DA | IA64_PSR_DD | \
IA64_PSR_SS | IA64_PSR_RI | IA64_PSR_ED | IA64_PSR_IA)
-#define PSCB(x,y) VCPU(x,y)
-#define PSCBX(x,y) x->arch.y
-
-extern unsigned long vcpu_verbose;
-
-long do_iopl(domid_t domain, unsigned int new_io_pl)
-{
- dummy();
- return 0;
-}
-
#include <xen/sched-if.h>
-
-extern struct schedule_data schedule_data[NR_CPUS];
void schedule_tail(struct vcpu *prev)
{
@@ -95,9 +84,6 @@
{
struct domain *d = current->domain;
unsigned long mask, pteval2, mpaddr;
- unsigned long lookup_domain_mpa(struct domain *,unsigned long);
- extern struct domain *dom0;
- extern unsigned long dom0_start, dom0_size;
// FIXME address had better be pre-validated on insert
mask = ~itir_mask(itir);
@@ -127,7 +113,6 @@
// given a current domain metaphysical address, return the physical address
unsigned long translate_domain_mpaddr(unsigned long mpaddr)
{
- extern unsigned long lookup_domain_mpa(struct domain *,unsigned long);
unsigned long pteval;
if (current->domain == dom0) {
@@ -224,7 +209,7 @@
void foodpi(void) {}
-unsigned long pending_false_positive = 0;
+static unsigned long pending_false_positive = 0;
void reflect_extint(struct pt_regs *regs)
{
@@ -293,13 +278,14 @@
return;
}
- fault = vcpu_translate(current,address,is_data,&pteval,&itir,&iha);
+ fault = vcpu_translate(current,address,is_data,0,&pteval,&itir,&iha);
if (fault == IA64_NO_FAULT) {
pteval = translate_domain_pte(pteval,address,itir);
vcpu_itc_no_srlz(current,is_data?2:1,address,pteval,-1UL,(itir>>2)&0x3f);
return;
}
- if (IS_VMM_ADDRESS(iip)) {
+ if (!user_mode (regs)) {
+ /* The fault occurs inside Xen. */
if (!ia64_done_with_exception(regs)) {
// should never happen. If it does, region 0 addr may
// indicate a bad xen pointer
@@ -543,7 +529,6 @@
void
do_ssc(unsigned long ssc, struct pt_regs *regs)
{
- extern unsigned long lookup_domain_mpa(struct domain *,unsigned long);
unsigned long arg0, arg1, arg2, arg3, retval;
char buf[2];
/**/ static int last_fd, last_count; // FIXME FIXME FIXME
@@ -653,14 +638,14 @@
vcpu_increment_iip(current);
}
+/* Also read in hyperprivop.S */
int first_break = 1;
void
ia64_handle_break (unsigned long ifa, struct pt_regs *regs, unsigned long isr,
unsigned long iim)
{
- struct domain *d = (struct domain *) current->domain;
+ struct domain *d = current->domain;
struct vcpu *v = current;
- extern unsigned long running_on_sim;
if (first_break) {
if (platform_is_hp_ski()) running_on_sim = 1;
@@ -668,8 +653,7 @@
first_break = 0;
}
if (iim == 0x80001 || iim == 0x80002) { //FIXME: don't hardcode constant
- if (running_on_sim) do_ssc(vcpu_get_gr(current,36), regs);
- else do_ssc(vcpu_get_gr(current,36), regs);
+ do_ssc(vcpu_get_gr(current,36), regs);
}
#ifdef CRASH_DEBUG
else if ((iim == 0 || iim == CDB_BREAK_NUM) && !user_mode(regs)) {
@@ -711,6 +695,7 @@
}
}
+/* Used in vhpt.h. */
#define INTR_TYPE_MAX 10
UINT64 int_counts[INTR_TYPE_MAX];
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/ia64/xen/regionreg.c
--- a/xen/arch/ia64/xen/regionreg.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/ia64/xen/regionreg.c Mon Mar 20 08:56:54 2006
@@ -157,7 +157,6 @@
int rid_block_end = d->arch.ending_rid >> IA64_MIN_IMPL_RID_BITS;
int rid_block_start = d->arch.starting_rid >> IA64_MIN_IMPL_RID_BITS;
- return 1; // KLUDGE ALERT
//
// not all domains will have allocated RIDs (physical mode loaders for
instance)
//
@@ -250,13 +249,18 @@
newrrv.rid = newrid;
newrrv.ve = 1; // VHPT now enabled for region 7!!
newrrv.ps = PAGE_SHIFT;
- if (rreg == 0)
+
+ if (rreg == 0) {
v->arch.metaphysical_saved_rr0 = vmMangleRID(newrrv.rrval);
- else if (rreg == 7)
+ if (!PSCB(v,metaphysical_mode))
+ set_rr(rr,newrrv.rrval);
+ } else if (rreg == 7) {
ia64_new_rr7(vmMangleRID(newrrv.rrval),v->vcpu_info,
v->arch.privregs, __get_cpu_var(vhpt_paddr),
(unsigned long) pal_vaddr);
- else set_rr(rr,newrrv.rrval);
+ } else {
+ set_rr(rr,newrrv.rrval);
+ }
#endif
return 1;
}
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/ia64/xen/vcpu.c
--- a/xen/arch/ia64/xen/vcpu.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/ia64/xen/vcpu.c Mon Mar 20 08:56:54 2006
@@ -5,12 +5,6 @@
* Dan Magenheimer (dan.magenheimer@xxxxxx)
*
*/
-
-#if 1
-// TEMPORARY PATCH for match_dtlb uses this, can be removed later
-// FIXME SMP
-int in_tpa = 0;
-#endif
#include <linux/sched.h>
#include <public/arch-ia64.h>
@@ -30,19 +24,18 @@
extern void setreg(unsigned long regnum, unsigned long val, int nat, struct
pt_regs *regs);
extern void panic_domain(struct pt_regs *, const char *, ...);
extern int set_metaphysical_rr0(void);
+extern unsigned long translate_domain_pte(UINT64,UINT64,UINT64);
+extern unsigned long translate_domain_mpaddr(unsigned long);
+extern void ia64_global_tlb_purge(UINT64 start, UINT64 end, UINT64 nbits);
+
typedef union {
struct ia64_psr ia64_psr;
unsigned long i64;
} PSR;
-//typedef struct pt_regs REGS;
-//typedef struct domain VCPU;
-
// this def for vcpu_regs won't work if kernel stack is present
//#define vcpu_regs(vcpu) ((struct pt_regs *) vcpu->arch.regs
-#define PSCB(x,y) VCPU(x,y)
-#define PSCBX(x,y) x->arch.y
#define TRUE 1
#define FALSE 0
@@ -71,18 +64,6 @@
unsigned long phys_translate_count = 0;
unsigned long vcpu_verbose = 0;
-#define verbose(a...) do {if (vcpu_verbose) printf(a);} while(0)
-
-//#define vcpu_quick_region_check(_tr_regions,_ifa) 1
-#define vcpu_quick_region_check(_tr_regions,_ifa) \
- (_tr_regions & (1 << ((unsigned long)_ifa >> 61)))
-#define vcpu_quick_region_set(_tr_regions,_ifa)
\
- do {_tr_regions |= (1 << ((unsigned long)_ifa >> 61)); } while (0)
-
-// FIXME: also need to check && (!trp->key || vcpu_pkr_match(trp->key))
-#define vcpu_match_tr_entry(_trp,_ifa,_rid) \
- ((_trp->p && (_trp->rid==_rid) && (_ifa >= _trp->vadr) && \
- (_ifa < (_trp->vadr + (1L<< _trp->ps)) - 1)))
/**************************************************************************
VCPU general register access routines
@@ -238,7 +219,6 @@
return IA64_NO_FAULT;
}
-extern UINT64 vcpu_check_pending_interrupts(VCPU *vcpu);
#define SPURIOUS_VECTOR 0xf
IA64FAULT vcpu_set_psr_dt(VCPU *vcpu)
@@ -659,13 +639,6 @@
}
}
-void early_tick(VCPU *vcpu)
-{
- UINT64 *p = &PSCBX(vcpu,irr[3]);
- printf("vcpu_check_pending: about to deliver early tick\n");
- printf("&irr[0]=%p, irr[0]=0x%lx\n",p,*p);
-}
-
#define IA64_TPR_MMI 0x10000
#define IA64_TPR_MIC 0x000f0
@@ -677,7 +650,7 @@
* and this must be checked independently; see vcpu_deliverable interrupts() */
UINT64 vcpu_check_pending_interrupts(VCPU *vcpu)
{
- UINT64 *p, *q, *r, bits, bitnum, mask, i, vector;
+ UINT64 *p, *r, bits, bitnum, mask, i, vector;
/* Always check pending event, since guest may just ack the
* event injection without handle. Later guest may throw out
@@ -691,8 +664,8 @@
p = &PSCBX(vcpu,irr[3]);
r = &PSCBX(vcpu,insvc[3]);
- for (i = 3; ; p--, q--, r--, i--) {
- bits = *p /* & *q */;
+ for (i = 3; ; p--, r--, i--) {
+ bits = *p ;
if (bits) break; // got a potential interrupt
if (*r) {
// nothing in this word which is pending+inservice
@@ -713,7 +686,7 @@
if (vector == (PSCB(vcpu,itv) & 0xff)) {
uint64_t now = ia64_get_itc();
if (now < PSCBX(vcpu,domain_itm)) {
- printk("Ooops, pending guest timer before its due\n");
+// printk("Ooops, pending guest timer before its due\n");
PSCBX(vcpu,irr[i]) &= ~mask;
goto check_start;
}
@@ -753,12 +726,12 @@
IA64FAULT vcpu_get_lid(VCPU *vcpu, UINT64 *pval)
{
-//extern unsigned long privop_trace;
-//privop_trace=1;
- //TODO: Implement this
- printf("vcpu_get_lid: WARNING: Getting cr.lid always returns zero\n");
- //*pval = 0;
- *pval = ia64_getreg(_IA64_REG_CR_LID);
+ /* Use real LID for domain0 until vIOSAPIC is present.
+ Use EID=0, ID=vcpu_id for domU. */
+ if (vcpu->domain == dom0)
+ *pval = ia64_getreg(_IA64_REG_CR_LID);
+ else
+ *pval = vcpu->vcpu_id << 24;
return IA64_NO_FAULT;
}
@@ -932,6 +905,7 @@
{
if (val & 0xff00) return IA64_RSVDREG_FAULT;
PSCB(vcpu,tpr) = val;
+ /* This can unmask interrupts. */
if (vcpu_check_pending_interrupts(vcpu) != SPURIOUS_VECTOR)
PSCB(vcpu,pending_interruption) = 1;
return (IA64_NO_FAULT);
@@ -945,8 +919,8 @@
p = &PSCBX(vcpu,insvc[3]);
for (i = 3; (i >= 0) && !(bits = *p); i--, p--);
if (i < 0) {
- printf("Trying to EOI interrupt when none are in-service.\r\n");
- return;
+ printf("Trying to EOI interrupt when none are in-service.\n");
+ return IA64_NO_FAULT;
}
bitnum = ia64_fls(bits);
vec = bitnum + (i*64);
@@ -957,7 +931,7 @@
if (PSCB(vcpu,interrupt_delivery_enabled)) { // but only if enabled...
// worry about this later... Linux only calls eoi
// with interrupts disabled
- printf("Trying to EOI interrupt with interrupts enabled\r\n");
+ printf("Trying to EOI interrupt with interrupts enabled\n");
}
if (vcpu_check_pending_interrupts(vcpu) != SPURIOUS_VECTOR)
PSCB(vcpu,pending_interruption) = 1;
@@ -1296,7 +1270,7 @@
int warn_region0_address = 0; // FIXME later: tie to a boot parameter?
-IA64FAULT vcpu_translate(VCPU *vcpu, UINT64 address, BOOLEAN is_data, UINT64
*pteval, UINT64 *itir, UINT64 *iha)
+IA64FAULT vcpu_translate(VCPU *vcpu, UINT64 address, BOOLEAN is_data, BOOLEAN
in_tpa, UINT64 *pteval, UINT64 *itir, UINT64 *iha)
{
unsigned long region = address >> 61;
unsigned long pta, pte, rid, rr;
@@ -1309,13 +1283,23 @@
// FIXME: This seems to happen even though it shouldn't. Need to track
// this down, but since it has been apparently harmless, just flag it for now
// panic_domain(vcpu_regs(vcpu),
- printk(
- "vcpu_translate: bad physical address:
0x%lx\n",address);
+
+ /*
+ * Guest may execute itc.d and rfi with psr.dt=0
+ * When VMM try to fetch opcode, tlb miss may happen,
+ * At this time PSCB(vcpu,metaphysical_mode)=1,
+ * region=5,VMM need to handle this tlb miss as if
+ * PSCB(vcpu,metaphysical_mode)=0
+ */
+ printk("vcpu_translate: bad physical address: 0x%lx\n",
+ address);
+ } else {
+ *pteval = (address & _PAGE_PPN_MASK) | __DIRTY_BITS |
+ _PAGE_PL_2 | _PAGE_AR_RWX;
+ *itir = PAGE_SHIFT << 2;
+ phys_translate_count++;
+ return IA64_NO_FAULT;
}
- *pteval = (address & _PAGE_PPN_MASK) | __DIRTY_BITS |
_PAGE_PL_2 | _PAGE_AR_RWX;
- *itir = PAGE_SHIFT << 2;
- phys_translate_count++;
- return IA64_NO_FAULT;
}
else if (!region && warn_region0_address) {
REGS *regs = vcpu_regs(vcpu);
@@ -1408,9 +1392,7 @@
UINT64 pteval, itir, mask, iha;
IA64FAULT fault;
- in_tpa = 1;
- fault = vcpu_translate(vcpu, vadr, 1, &pteval, &itir, &iha);
- in_tpa = 0;
+ fault = vcpu_translate(vcpu, vadr, TRUE, TRUE, &pteval, &itir, &iha);
if (fault == IA64_NO_FAULT)
{
mask = itir_mask(itir);
@@ -1655,8 +1637,11 @@
IA64FAULT vcpu_get_rr(VCPU *vcpu, UINT64 reg, UINT64 *pval)
{
- UINT val = PSCB(vcpu,rrs)[reg>>61];
- *pval = val;
+ if(VMX_DOMAIN(vcpu)){
+ *pval = VMX(vcpu,vrr[reg>>61]);
+ }else{
+ *pval = PSCB(vcpu,rrs)[reg>>61];
+ }
return (IA64_NO_FAULT);
}
@@ -1693,7 +1678,7 @@
VCPU translation register access routines
**************************************************************************/
-static void vcpu_purge_tr_entry(TR_ENTRY *trp)
+static inline void vcpu_purge_tr_entry(TR_ENTRY *trp)
{
trp->p = 0;
}
@@ -1746,8 +1731,6 @@
**************************************************************************/
void foobar(void) { /*vcpu_verbose = 1;*/ }
-
-extern struct domain *dom0;
void vcpu_itc_no_srlz(VCPU *vcpu, UINT64 IorD, UINT64 vaddr, UINT64 pte,
UINT64 mp_pte, UINT64 logps)
{
@@ -1793,7 +1776,6 @@
IA64FAULT vcpu_itc_d(VCPU *vcpu, UINT64 pte, UINT64 itir, UINT64 ifa)
{
unsigned long pteval, logps = itir_ps(itir);
- unsigned long translate_domain_pte(UINT64,UINT64,UINT64);
BOOLEAN swap_rr0 = (!(ifa>>61) && PSCB(vcpu,metaphysical_mode));
if (logps < PAGE_SHIFT) {
@@ -1813,7 +1795,6 @@
IA64FAULT vcpu_itc_i(VCPU *vcpu, UINT64 pte, UINT64 itir, UINT64 ifa)
{
unsigned long pteval, logps = itir_ps(itir);
- unsigned long translate_domain_pte(UINT64,UINT64,UINT64);
BOOLEAN swap_rr0 = (!(ifa>>61) && PSCB(vcpu,metaphysical_mode));
// FIXME: validate ifa here (not in Xen space), COULD MACHINE CHECK!
@@ -1849,8 +1830,6 @@
// TODO: Only allowed for current vcpu
UINT64 mpaddr, paddr;
IA64FAULT fault;
- unsigned long translate_domain_mpaddr(unsigned long);
- IA64FAULT vcpu_tpa(VCPU *, UINT64, UINT64 *);
fault = vcpu_tpa(vcpu, vadr, &mpaddr);
if (fault == IA64_NO_FAULT) {
@@ -1885,7 +1864,6 @@
IA64FAULT vcpu_ptc_ga(VCPU *vcpu,UINT64 vadr,UINT64 addr_range)
{
- extern void ia64_global_tlb_purge(UINT64 start, UINT64 end, UINT64
nbits);
// FIXME: validate not flushing Xen addresses
// if (Xen address) return(IA64_ILLOP_FAULT);
// FIXME: ??breaks if domain PAGE_SIZE < Xen PAGE_SIZE
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/ia64/xen/xenirq.c
--- a/xen/arch/ia64/xen/xenirq.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/ia64/xen/xenirq.c Mon Mar 20 08:56:54 2006
@@ -68,7 +68,7 @@
* ONLY gets called from ia64_leave_kernel
* ONLY call with interrupts enabled
*/
-void process_soft_irq()
+void process_soft_irq(void)
{
if (!in_interrupt() && local_softirq_pending()) {
add_preempt_count(SOFTIRQ_OFFSET);
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/ia64/xen/xenmem.c
--- a/xen/arch/ia64/xen/xenmem.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/ia64/xen/xenmem.c Mon Mar 20 08:56:54 2006
@@ -13,12 +13,11 @@
#include <asm/pgtable.h>
#include <xen/mm.h>
-extern struct page *zero_page_memmap_ptr;
struct page_info *frame_table;
unsigned long frame_table_size;
unsigned long max_page;
-struct page *mem_map;
+struct page_info *mem_map;
#define MAX_DMA_ADDRESS ~0UL // FIXME???
#ifdef CONFIG_VIRTUAL_MEM_MAP
@@ -35,6 +34,8 @@
paging_init (void)
{
unsigned int mpt_order;
+ unsigned long i;
+
/* Create machine to physical mapping table
* NOTE: similar to frame table, later we may need virtually
* mapped mpt table if large hole exists. Also MAX_ORDER needs
@@ -47,10 +48,9 @@
panic("Not enough memory to bootstrap Xen.\n");
printk("machine to physical table: 0x%lx\n", (u64)mpt_table);
- memset(mpt_table, INVALID_M2P_ENTRY, mpt_table_size);
- /* Other mapping setup */
-
- zero_page_memmap_ptr = virt_to_page(ia64_imva(empty_zero_page));
+ for (i = 0; i < (1UL << mpt_order); i++) {
+ mpt_table[i] = INVALID_M2P_ENTRY;
+ }
}
/* FIXME: postpone support to machines with big holes between physical memorys.
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/ia64/xen/xenmisc.c
--- a/xen/arch/ia64/xen/xenmisc.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/ia64/xen/xenmisc.c Mon Mar 20 08:56:54 2006
@@ -114,23 +114,6 @@
#endif
///////////////////////////////
-// from arch/ia64/page_alloc.c
-///////////////////////////////
-DEFINE_PER_CPU(struct page_state, page_states) = {0};
-unsigned long totalram_pages;
-
-void __mod_page_state(unsigned long offset, unsigned long delta)
-{
- unsigned long flags;
- void* ptr;
-
- local_irq_save(flags);
- ptr = &__get_cpu_var(page_states);
- *(unsigned long*)(ptr + offset) += delta;
- local_irq_restore(flags);
-}
-
-///////////////////////////////
// from arch/x86/flushtlb.c
///////////////////////////////
@@ -147,12 +130,17 @@
//memset(percpu_info, 0, sizeof(percpu_info));
}
-#if 0
-void free_page_type(struct page_info *page, unsigned int type)
-{
- dummy();
-}
-#endif
+void free_page_type(struct page_info *page, u32 type)
+{
+// dummy();
+ return;
+}
+
+int alloc_page_type(struct page_info *page, u32 type)
+{
+// dummy();
+ return 1;
+}
///////////////////////////////
//// misc memory stuff
@@ -166,7 +154,7 @@
return (unsigned long)p;
}
-void __free_pages(struct page *page, unsigned int order)
+void __free_pages(struct page_info *page, unsigned int order)
{
if (order) BUG();
free_xenheap_page(page);
@@ -306,9 +294,9 @@
uint64_t pta;
local_irq_save(spsr);
- if(VMX_DOMAIN(prev)){
- vtm_domain_out(prev);
- }
+// if(VMX_DOMAIN(prev)){
+// vtm_domain_out(prev);
+// }
context_switch_count++;
switch_to(prev,next,prev);
// if(VMX_DOMAIN(current)){
@@ -326,7 +314,7 @@
}
if (VMX_DOMAIN(current)){
- vtm_domain_in(current);
+// vtm_domain_in(current);
vmx_load_all_rr(current);
}else{
extern char ia64_ivt;
@@ -415,3 +403,203 @@
}
else printk("sync_split_caches ignored for CPU with no split cache\n");
}
+
+///////////////////////////////
+// from arch/x86/mm.c
+///////////////////////////////
+
+#ifdef VERBOSE
+#define MEM_LOG(_f, _a...) \
+ printk("DOM%u: (file=mm.c, line=%d) " _f "\n", \
+ current->domain->domain_id , __LINE__ , ## _a )
+#else
+#define MEM_LOG(_f, _a...) ((void)0)
+#endif
+
+void cleanup_writable_pagetable(struct domain *d)
+{
+ return;
+}
+
+void put_page_type(struct page_info *page)
+{
+ u32 nx, x, y = page->u.inuse.type_info;
+
+ again:
+ do {
+ x = y;
+ nx = x - 1;
+
+ ASSERT((x & PGT_count_mask) != 0);
+
+ /*
+ * The page should always be validated while a reference is held. The
+ * exception is during domain destruction, when we forcibly invalidate
+ * page-table pages if we detect a referential loop.
+ * See domain.c:relinquish_list().
+ */
+ ASSERT((x & PGT_validated) ||
+ test_bit(_DOMF_dying, &page_get_owner(page)->domain_flags));
+
+ if ( unlikely((nx & PGT_count_mask) == 0) )
+ {
+ /* Record TLB information for flush later. Races are harmless. */
+ page->tlbflush_timestamp = tlbflush_current_time();
+
+ if ( unlikely((nx & PGT_type_mask) <= PGT_l4_page_table) &&
+ likely(nx & PGT_validated) )
+ {
+ /*
+ * Page-table pages must be unvalidated when count is zero. The
+ * 'free' is safe because the refcnt is non-zero and validated
+ * bit is clear => other ops will spin or fail.
+ */
+ if ( unlikely((y = cmpxchg(&page->u.inuse.type_info, x,
+ x & ~PGT_validated)) != x) )
+ goto again;
+ /* We cleared the 'valid bit' so we do the clean up. */
+ free_page_type(page, x);
+ /* Carry on, but with the 'valid bit' now clear. */
+ x &= ~PGT_validated;
+ nx &= ~PGT_validated;
+ }
+ }
+ else if ( unlikely(((nx & (PGT_pinned | PGT_count_mask)) ==
+ (PGT_pinned | 1)) &&
+ ((nx & PGT_type_mask) != PGT_writable_page)) )
+ {
+ /* Page is now only pinned. Make the back pointer mutable again. */
+ nx |= PGT_va_mutable;
+ }
+ }
+ while ( unlikely((y = cmpxchg(&page->u.inuse.type_info, x, nx)) != x) );
+}
+
+
+int get_page_type(struct page_info *page, u32 type)
+{
+ u32 nx, x, y = page->u.inuse.type_info;
+
+ again:
+ do {
+ x = y;
+ nx = x + 1;
+ if ( unlikely((nx & PGT_count_mask) == 0) )
+ {
+ MEM_LOG("Type count overflow on pfn %lx", page_to_mfn(page));
+ return 0;
+ }
+ else if ( unlikely((x & PGT_count_mask) == 0) )
+ {
+ if ( (x & (PGT_type_mask|PGT_va_mask)) != type )
+ {
+ if ( (x & PGT_type_mask) != (type & PGT_type_mask) )
+ {
+ /*
+ * On type change we check to flush stale TLB
+ * entries. This may be unnecessary (e.g., page
+ * was GDT/LDT) but those circumstances should be
+ * very rare.
+ */
+ cpumask_t mask =
+ page_get_owner(page)->domain_dirty_cpumask;
+ tlbflush_filter(mask, page->tlbflush_timestamp);
+
+ if ( unlikely(!cpus_empty(mask)) )
+ {
+ perfc_incrc(need_flush_tlb_flush);
+ flush_tlb_mask(mask);
+ }
+ }
+
+ /* We lose existing type, back pointer, and validity. */
+ nx &= ~(PGT_type_mask | PGT_va_mask | PGT_validated);
+ nx |= type;
+
+ /* No special validation needed for writable pages. */
+ /* Page tables and GDT/LDT need to be scanned for validity. */
+ if ( type == PGT_writable_page )
+ nx |= PGT_validated;
+ }
+ }
+ else
+ {
+ if ( unlikely((x & (PGT_type_mask|PGT_va_mask)) != type) )
+ {
+ if ( unlikely((x & PGT_type_mask) != (type & PGT_type_mask) ) )
+ {
+ if ( current->domain == page_get_owner(page) )
+ {
+ /*
+ * This ensures functions like set_gdt() see up-to-date
+ * type info without needing to clean up writable p.t.
+ * state on the fast path.
+ */
+ LOCK_BIGLOCK(current->domain);
+ cleanup_writable_pagetable(current->domain);
+ y = page->u.inuse.type_info;
+ UNLOCK_BIGLOCK(current->domain);
+ /* Can we make progress now? */
+ if ( ((y & PGT_type_mask) == (type & PGT_type_mask)) ||
+ ((y & PGT_count_mask) == 0) )
+ goto again;
+ }
+ if ( ((x & PGT_type_mask) != PGT_l2_page_table) ||
+ ((type & PGT_type_mask) != PGT_l1_page_table) )
+ MEM_LOG("Bad type (saw %" PRtype_info
+ " != exp %" PRtype_info ") "
+ "for mfn %lx (pfn %lx)",
+ x, type, page_to_mfn(page),
+ get_gpfn_from_mfn(page_to_mfn(page)));
+ return 0;
+ }
+ else if ( (x & PGT_va_mask) == PGT_va_mutable )
+ {
+ /* The va backpointer is mutable, hence we update it. */
+ nx &= ~PGT_va_mask;
+ nx |= type; /* we know the actual type is correct */
+ }
+ else if ( ((type & PGT_va_mask) != PGT_va_mutable) &&
+ ((type & PGT_va_mask) != (x & PGT_va_mask)) )
+ {
+#ifdef CONFIG_X86_PAE
+ /* We use backptr as extra typing. Cannot be unknown. */
+ if ( (type & PGT_type_mask) == PGT_l2_page_table )
+ return 0;
+#endif
+ /* This table is possibly mapped at multiple locations. */
+ nx &= ~PGT_va_mask;
+ nx |= PGT_va_unknown;
+ }
+ }
+ if ( unlikely(!(x & PGT_validated)) )
+ {
+ /* Someone else is updating validation of this page. Wait... */
+ while ( (y = page->u.inuse.type_info) == x )
+ cpu_relax();
+ goto again;
+ }
+ }
+ }
+ while ( unlikely((y = cmpxchg(&page->u.inuse.type_info, x, nx)) != x) );
+
+ if ( unlikely(!(nx & PGT_validated)) )
+ {
+ /* Try to validate page type; drop the new reference on failure. */
+ if ( unlikely(!alloc_page_type(page, type)) )
+ {
+ MEM_LOG("Error while validating mfn %lx (pfn %lx) for type %"
+ PRtype_info ": caf=%08x taf=%" PRtype_info,
+ page_to_mfn(page), get_gpfn_from_mfn(page_to_mfn(page)),
+ type, page->count_info, page->u.inuse.type_info);
+ /* Noone else can get a reference. We hold the only ref. */
+ page->u.inuse.type_info = 0;
+ return 0;
+ }
+
+ /* Noone else is updating simultaneously. */
+ __set_bit(_PGT_validated, &page->u.inuse.type_info);
+ }
+
+ return 1;
+}
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/ia64/xen/xensetup.c
--- a/xen/arch/ia64/xen/xensetup.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/ia64/xen/xensetup.c Mon Mar 20 08:56:54 2006
@@ -27,6 +27,7 @@
unsigned long xenheap_phys_end;
char saved_command_line[COMMAND_LINE_SIZE];
+char dom0_command_line[COMMAND_LINE_SIZE];
struct vcpu *idle_vcpu[NR_CPUS];
@@ -119,11 +120,12 @@
void early_cmdline_parse(char **cmdline_p)
{
char *guest_cmd;
- char *split = "--";
+ static const char * const split = "--";
if (*cmdline_p == NULL) {
*cmdline_p = &null[0];
saved_command_line[0] = '\0';
+ dom0_command_line[0] = '\0';
return;
}
@@ -138,7 +140,8 @@
while (*guest_cmd == ' ') guest_cmd++;
}
- strlcpy(saved_command_line, guest_cmd, COMMAND_LINE_SIZE);
+ strlcpy(saved_command_line, *cmdline_p, COMMAND_LINE_SIZE);
+ strlcpy(dom0_command_line, guest_cmd, COMMAND_LINE_SIZE);
return;
}
@@ -155,24 +158,6 @@
.parity = 'n',
.stop_bits = 1
};
-/* This is a wrapper function of init_domheap_pages,
- * memory exceeds (max_page<<PAGE_SHIFT) will not be reclaimed.
- * This function will go away when the virtual memmap/discontig
- * memory issues are solved
- */
-void init_domheap_pages_wrapper(unsigned long ps, unsigned long pe)
-{
- unsigned long s_nrm, e_nrm, max_mem;
- max_mem = (max_page+1)<<PAGE_SHIFT;
- s_nrm = (ps+PAGE_SIZE-1)&PAGE_MASK;
- e_nrm = pe&PAGE_MASK;
- s_nrm = min(s_nrm, max_mem);
- e_nrm = min(e_nrm, max_mem);
- if(s_nrm < e_nrm)
- init_domheap_pages(s_nrm, e_nrm);
-}
-
-
void start_kernel(void)
{
@@ -349,7 +334,6 @@
if ( num_online_cpus() >= max_cpus )
break;
if ( !cpu_online(i) ) {
-printk("About to call __cpu_up(%d)\n",i);
__cpu_up(i);
}
}
@@ -395,17 +379,6 @@
/* PIN domain0 on CPU 0. */
dom0->vcpu[0]->cpu_affinity = cpumask_of_cpu(0);
- /* The stash space for the initial kernel image can now be freed up. */
- /* init_domheap_pages_wrapper is temporary solution, please refer to the
- * descriptor of this function */
- init_domheap_pages_wrapper(ia64_boot_param->domain_start,
- ia64_boot_param->domain_start+ia64_boot_param->domain_size);
- /* throw away initrd area passed from elilo */
- if (ia64_boot_param->initrd_size) {
- init_domheap_pages_wrapper(ia64_boot_param->initrd_start,
- ia64_boot_param->initrd_start+ia64_boot_param->initrd_size);
- }
-
if (!running_on_sim) // slow on ski and pages are pre-initialized to zero
scrub_heap_pages();
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/ia64/xen/xentime.c
--- a/xen/arch/ia64/xen/xentime.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/ia64/xen/xentime.c Mon Mar 20 08:56:54 2006
@@ -116,6 +116,12 @@
}
#endif
#endif
+
+#if 0
+ /* Nobody seems to be able to explain this code.
+ It seems to be accumulated tricks, which are not required anymore.
+ Also I have made many tests, I'd like to get confirmation from
+ other site (TG). */
if (current->domain == dom0) {
// FIXME: there's gotta be a better way of doing this...
// We have to ensure that domain0 is launched before we
@@ -130,6 +136,7 @@
vcpu_wake(dom0->vcpu[0]);
}
}
+#endif
if (!is_idle_domain(current->domain)) {
if (vcpu_timer_expired(current)) {
vcpu_pend_timer(current);
@@ -141,7 +148,7 @@
new_itm = local_cpu_data->itm_next;
if (!VMX_DOMAIN(current) && !time_after(ia64_get_itc(), new_itm))
- return;
+ return IRQ_HANDLED;
if (VMX_DOMAIN(current))
vcpu_wake(current);
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/x86/Makefile
--- a/xen/arch/x86/Makefile Mon Mar 20 08:56:46 2006
+++ b/xen/arch/x86/Makefile Mon Mar 20 08:56:54 2006
@@ -1,59 +1,61 @@
-
include $(BASEDIR)/Rules.mk
-OBJS += $(patsubst %.S,%.o,$(wildcard $(TARGET_SUBARCH)/*.S))
-OBJS += $(patsubst %.c,%.o,$(wildcard $(TARGET_SUBARCH)/*.c))
-OBJS += $(patsubst %.c,%.o,$(wildcard acpi/*.c))
-OBJS += $(patsubst %.c,%.o,$(wildcard genapic/*.c))
-OBJS += $(patsubst %.c,%.o,$(wildcard cpu/*.c))
-OBJS += $(patsubst %.c,%.o,$(wildcard cpu/mcheck/*.c))
-OBJS += $(patsubst %.c,%.o,$(wildcard cpu/mtrr/*.c))
-OBJS += $(patsubst %.c,%.o,$(wildcard hvm/*.c))
-OBJS += $(patsubst %.c,%.o,$(wildcard hvm/vmx/*.c))
-OBJS += $(patsubst %.S,%.o,$(wildcard hvm/vmx/$(TARGET_SUBARCH)/*.S))
-OBJS += $(patsubst %.c,%.o,$(wildcard hvm/svm/*.c))
-OBJS += $(patsubst %.S,%.o,$(wildcard hvm/svm/$(TARGET_SUBARCH)/*.S))
+subdirs-y += acpi
+subdirs-y += cpu
+subdirs-y += genapic
+subdirs-y += hvm
-ifeq ($(TARGET_SUBARCH),x86_64)
-OBJS := $(subst cpu/centaur.o,,$(OBJS))
-OBJS := $(subst cpu/cyrix.o,,$(OBJS))
-OBJS := $(subst cpu/rise.o,,$(OBJS))
-OBJS := $(subst cpu/transmeta.o,,$(OBJS))
+subdirs-$(x86_32) += x86_32
+subdirs-$(x86_64) += x86_64
+
+obj-y += apic.o
+obj-y += audit.o
+obj-y += bitops.o
+obj-y += delay.o
+obj-y += dmi_scan.o
+obj-y += dom0_ops.o
+obj-y += domain.o
+obj-y += domain_build.o
+obj-y += e820.o
+obj-y += extable.o
+obj-y += flushtlb.o
+obj-y += i387.o
+obj-y += i8259.o
+obj-y += io_apic.o
+obj-y += irq.o
+obj-y += microcode.o
+obj-y += mm.o
+obj-y += mpparse.o
+obj-y += nmi.o
+obj-y += physdev.o
+obj-y += rwlock.o
+obj-y += setup.o
+obj-y += smp.o
+obj-y += smpboot.o
+obj-y += string.o
+obj-y += time.o
+obj-y += trampoline.o
+obj-y += traps.o
+obj-y += usercopy.o
+obj-y += x86_emulate.o
+
+ifneq ($(pae),n)
+obj-$(x86_32) += shadow.o shadow_public.o shadow_guest32.o
+else
+obj-$(x86_32) += shadow32.o
endif
-OBJS := $(patsubst shadow%.o,,$(OBJS)) # drop all
-ifeq ($(TARGET_SUBARCH),x86_64)
- OBJS += shadow.o shadow_public.o shadow_guest32.o shadow_guest32pae.o #
x86_64: new code
-endif
-ifeq ($(TARGET_SUBARCH),x86_32)
- ifneq ($(pae),n)
- OBJS += shadow.o shadow_public.o shadow_guest32.o # x86_32p: new code
- else
- OBJS += shadow32.o # x86_32: old code
- endif
-endif
+obj-$(x86_64) += shadow.o shadow_public.o shadow_guest32.o shadow_guest32pae.o
-ifneq ($(supervisor_mode_kernel),y)
-OBJS := $(subst x86_32/supervisor_mode_kernel.o,,$(OBJS))
-endif
+obj-$(crash_debug) += gdbstub.o
-OBJS := $(subst $(TARGET_SUBARCH)/asm-offsets.o,,$(OBJS))
-OBJS := $(subst $(TARGET_SUBARCH)/xen.lds.o,,$(OBJS))
-
-ifneq ($(crash_debug),y)
-OBJS := $(patsubst gdbstub%.o,,$(OBJS))
-endif
-
-default: $(TARGET)
+include $(BASEDIR)/Post.mk
$(TARGET): $(TARGET)-syms boot/mkelf32
./boot/mkelf32 $(TARGET)-syms $(TARGET) 0x100000 \
`$(NM) $(TARGET)-syms | sort | tail -n 1 | sed -e 's/^\([^
]*\).*/0x\1/'`
-$(CURDIR)/arch.o: $(OBJS)
- $(LD) $(LDFLAGS) -r -o $@ $(OBJS)
-
-$(TARGET)-syms: boot/$(TARGET_SUBARCH).o $(ALL_OBJS) xen.lds
+$(TARGET)-syms: boot/$(TARGET_SUBARCH).o xen.lds
$(LD) $(LDFLAGS) -T xen.lds -N \
boot/$(TARGET_SUBARCH).o $(ALL_OBJS) -o $@
$(NM) -n $@ | $(BASEDIR)/tools/symbols >$(BASEDIR)/xen-syms.S
@@ -77,21 +79,5 @@
shadow_guest32.o: shadow.c
-clean:
- rm -f *.o *.s *~ core boot/*.o boot/*~ boot/core boot/mkelf32
- rm -f x86_32/*.o x86_32/*~ x86_32/core
- rm -f x86_64/*.o x86_64/*~ x86_64/core
- rm -f mtrr/*.o mtrr/*~ mtrr/core
- rm -f acpi/*.o acpi/*~ acpi/core
- rm -f genapic/*.o genapic/*~ genapic/core
- rm -f cpu/*.o cpu/*~ cpu/core
- rm -f hvm/*.o hvm/*~ hvm/core
- rm -f hvm/vmx/*.o hvm/vmx/*~ hvm/vmx/core
- rm -f hvm/vmx/x86_32/*.o hvm/vmx/x86_32/*~ hvm/vmx/x86_32/core
- rm -f hvm/vmx/x86_64/*.o hvm/vmx/x86_64/*~ hvm/vmx/x86_64/core
- rm -f hvm/svm/*.o hvm/svm/*~ hvm/svm/core
- rm -f hvm/svm/x86_32/*.o hvm/svm/x86_32/*~ hvm/svm/x86_32/core
- rm -f hvm/svm/x86_64/*.o hvm/svm/x86_64/*~ hvm/svm/x86_64/core
- rm -f xen.lds
-
-.PHONY: default clean
+clean:: FORCE
+ rm -f asm-offsets.s xen.lds boot/*.o boot/*~ boot/core boot/mkelf32
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/x86/Rules.mk
--- a/xen/arch/x86/Rules.mk Mon Mar 20 08:56:46 2006
+++ b/xen/arch/x86/Rules.mk Mon Mar 20 08:56:54 2006
@@ -1,5 +1,7 @@
########################################
# x86-specific definitions
+
+HAS_ACPI := y
#
# If you change any of these configuration options then you must
@@ -31,13 +33,17 @@
endif
ifeq ($(XEN_TARGET_ARCH),x86_32)
-LDFLAGS += -m elf_i386
+LDFLAGS += -m elf_i386
+x86_32 := y
+x86_64 := n
endif
ifeq ($(TARGET_SUBARCH),x86_64)
CFLAGS += -mno-red-zone -fpic -fno-reorder-blocks
CFLAGS += -fno-asynchronous-unwind-tables
LDFLAGS += -m elf_x86_64
+x86_32 := n
+x86_64 := y
endif
# Test for at least GCC v3.2.x.
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/x86/domain.c
--- a/xen/arch/x86/domain.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/x86/domain.c Mon Mar 20 08:56:54 2006
@@ -51,6 +51,9 @@
} __cacheline_aligned;
static struct percpu_ctxt percpu_ctxt[NR_CPUS];
+static void paravirt_ctxt_switch_from(struct vcpu *v);
+static void paravirt_ctxt_switch_to(struct vcpu *v);
+
static void continue_idle_domain(struct vcpu *v)
{
reset_stack_and_jump(idle_loop);
@@ -225,6 +228,9 @@
{
v->arch.schedule_tail = continue_nonidle_domain;
}
+
+ v->arch.ctxt_switch_from = paravirt_ctxt_switch_from;
+ v->arch.ctxt_switch_to = paravirt_ctxt_switch_to;
v->arch.perdomain_ptes =
d->arch.mm_perdomain_pt + (vcpu_id << GDT_LDT_VCPU_SHIFT);
@@ -685,21 +691,32 @@
percpu_ctxt[smp_processor_id()].dirty_segment_mask = dirty_segment_mask;
}
-#define switch_kernel_stack(_n,_c) ((void)0)
+#define switch_kernel_stack(v) ((void)0)
#elif defined(__i386__)
#define load_segments(n) ((void)0)
#define save_segments(p) ((void)0)
-static inline void switch_kernel_stack(struct vcpu *n, unsigned int cpu)
-{
- struct tss_struct *tss = &init_tss[cpu];
- tss->esp1 = n->arch.guest_context.kernel_sp;
- tss->ss1 = n->arch.guest_context.kernel_ss;
-}
-
-#endif
+static inline void switch_kernel_stack(struct vcpu *v)
+{
+ struct tss_struct *tss = &init_tss[smp_processor_id()];
+ tss->esp1 = v->arch.guest_context.kernel_sp;
+ tss->ss1 = v->arch.guest_context.kernel_ss;
+}
+
+#endif /* __i386__ */
+
+static void paravirt_ctxt_switch_from(struct vcpu *v)
+{
+ save_segments(v);
+}
+
+static void paravirt_ctxt_switch_to(struct vcpu *v)
+{
+ set_int80_direct_trap(v);
+ switch_kernel_stack(v);
+}
#define loaddebug(_v,_reg) \
__asm__ __volatile__ ("mov %0,%%db" #_reg : : "r" ((_v)->debugreg[_reg]))
@@ -720,15 +737,7 @@
stack_regs,
CTXT_SWITCH_STACK_BYTES);
unlazy_fpu(p);
- if ( !hvm_guest(p) )
- {
- save_segments(p);
- }
- else
- {
- hvm_save_segments(p);
- hvm_load_msrs();
- }
+ p->arch.ctxt_switch_from(p);
}
if ( !is_idle_vcpu(n) )
@@ -749,15 +758,7 @@
loaddebug(&n->arch.guest_context, 7);
}
- if ( !hvm_guest(n) )
- {
- set_int80_direct_trap(n);
- switch_kernel_stack(n, cpu);
- }
- else
- {
- hvm_restore_msrs(n);
- }
+ n->arch.ctxt_switch_to(n);
}
if ( p->domain != n->domain )
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/x86/hvm/intercept.c
--- a/xen/arch/x86/hvm/intercept.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/x86/hvm/intercept.c Mon Mar 20 08:56:54 2006
@@ -338,10 +338,10 @@
static __inline__ void missed_ticks(struct hvm_virpit*vpit)
{
- int missed_ticks;
+ int missed_ticks;
missed_ticks = (NOW() - vpit->scheduled)/(s_time_t) vpit->period;
- if ( missed_ticks > 0 ) {
+ if ( missed_ticks++ >= 0 ) {
vpit->pending_intr_nr += missed_ticks;
vpit->scheduled += missed_ticks * vpit->period;
}
@@ -355,22 +355,16 @@
/* pick up missed timer tick */
missed_ticks(vpit);
-
- vpit->pending_intr_nr++;
if ( test_bit(_VCPUF_running, &v->vcpu_flags) ) {
- vpit->scheduled += vpit->period;
set_timer(&vpit->pit_timer, vpit->scheduled);
}
}
+/* pick up missed timer ticks at deactive time */
void pickup_deactive_ticks(struct hvm_virpit *vpit)
{
-
if ( !active_timer(&(vpit->pit_timer)) ) {
- /* pick up missed timer tick */
missed_ticks(vpit);
-
- vpit->scheduled += vpit->period;
set_timer(&vpit->pit_timer, vpit->scheduled);
}
}
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/x86/hvm/svm/svm.c
--- a/xen/arch/x86/hvm/svm/svm.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/x86/hvm/svm/svm.c Mon Mar 20 08:56:54 2006
@@ -200,7 +200,8 @@
return 1;
}
-void svm_store_cpu_guest_regs(struct vcpu *v, struct cpu_user_regs *regs)
+static void svm_store_cpu_guest_regs(
+ struct vcpu *v, struct cpu_user_regs *regs)
{
struct vmcb_struct *vmcb = v->arch.hvm_svm.vmcb;
@@ -227,23 +228,11 @@
#endif
}
-void svm_load_cpu_guest_regs(struct vcpu *v, struct cpu_user_regs *regs)
+static void svm_load_cpu_guest_regs(
+ struct vcpu *v, struct cpu_user_regs *regs)
{
svm_load_cpu_user_regs(v, regs);
}
-
-#ifdef __x86_64__
-
-void svm_save_segments(struct vcpu *v)
-{
-}
-void svm_load_msrs(void)
-{
-}
-void svm_restore_msrs(struct vcpu *v)
-{
-}
-#endif
#define IS_CANO_ADDRESS(add) 1
@@ -458,12 +447,6 @@
hvm_funcs.store_cpu_guest_regs = svm_store_cpu_guest_regs;
hvm_funcs.load_cpu_guest_regs = svm_load_cpu_guest_regs;
-
-#ifdef __x86_64__
- hvm_funcs.save_segments = svm_save_segments;
- hvm_funcs.load_msrs = svm_load_msrs;
- hvm_funcs.restore_msrs = svm_restore_msrs;
-#endif
hvm_funcs.store_cpu_guest_ctrl_regs = svm_store_cpu_guest_ctrl_regs;
hvm_funcs.modify_guest_state = svm_modify_guest_state;
@@ -687,9 +670,19 @@
reset_stack_and_jump(svm_asm_do_launch);
}
+static void svm_ctxt_switch_from(struct vcpu *v)
+{
+}
+
+static void svm_ctxt_switch_to(struct vcpu *v)
+{
+}
+
void svm_final_setup_guest(struct vcpu *v)
{
- v->arch.schedule_tail = arch_svm_do_launch;
+ v->arch.schedule_tail = arch_svm_do_launch;
+ v->arch.ctxt_switch_from = svm_ctxt_switch_from;
+ v->arch.ctxt_switch_to = svm_ctxt_switch_to;
if (v == v->domain->vcpu[0])
{
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/x86/hvm/vmx/io.c
--- a/xen/arch/x86/hvm/vmx/io.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/x86/hvm/vmx/io.c Mon Mar 20 08:56:54 2006
@@ -40,20 +40,33 @@
#define BSP_CPU(v) (!(v->vcpu_id))
-void vmx_set_tsc_shift(struct vcpu *v, struct hvm_virpit *vpit)
-{
- u64 drift;
-
- if ( vpit->first_injected )
- drift = vpit->period_cycles * vpit->pending_intr_nr;
- else
- drift = 0;
- vpit->shift = v->arch.hvm_vmx.tsc_offset - drift;
- __vmwrite(TSC_OFFSET, vpit->shift);
-
+static inline
+void __set_tsc_offset(u64 offset)
+{
+ __vmwrite(TSC_OFFSET, offset);
#if defined (__i386__)
- __vmwrite(TSC_OFFSET_HIGH, ((vpit->shift)>> 32));
+ __vmwrite(TSC_OFFSET_HIGH, offset >> 32);
#endif
+}
+
+u64 get_guest_time(struct vcpu *v)
+{
+ struct hvm_virpit *vpit = &(v->domain->arch.hvm_domain.vpit);
+ u64 host_tsc;
+
+ rdtscll(host_tsc);
+ return host_tsc + vpit->cache_tsc_offset;
+}
+
+void set_guest_time(struct vcpu *v, u64 gtime)
+{
+ struct hvm_virpit *vpit = &(v->domain->arch.hvm_domain.vpit);
+ u64 host_tsc;
+
+ rdtscll(host_tsc);
+
+ vpit->cache_tsc_offset = gtime - host_tsc;
+ __set_tsc_offset(vpit->cache_tsc_offset);
}
static inline void
@@ -64,6 +77,7 @@
if ( is_pit_irq(v, vector, type) ) {
if ( !vpit->first_injected ) {
vpit->pending_intr_nr = 0;
+ vpit->last_pit_gtime = get_guest_time(v);
vpit->scheduled = NOW() + vpit->period;
set_timer(&vpit->pit_timer, vpit->scheduled);
vpit->first_injected = 1;
@@ -71,7 +85,9 @@
vpit->pending_intr_nr--;
}
vpit->inject_point = NOW();
- vmx_set_tsc_shift (v, vpit);
+
+ vpit->last_pit_gtime += vpit->period;
+ set_guest_time(v, vpit->last_pit_gtime);
}
switch(type)
@@ -189,14 +205,15 @@
vmx_stts();
+ /* pick up the elapsed PIT ticks and re-enable pit_timer */
+ if ( vpit->first_injected) {
+ set_guest_time(v, v->domain->arch.hvm_domain.guest_time);
+ pickup_deactive_ticks(vpit);
+ }
+
if ( test_bit(iopacket_port(v), &d->shared_info->evtchn_pending[0]) ||
test_bit(ARCH_HVM_IO_WAIT, &v->arch.hvm_vcpu.ioflags) )
hvm_wait_io();
-
- /* pick up the elapsed PIT ticks and re-enable pit_timer */
- if ( vpit->first_injected )
- pickup_deactive_ticks(vpit);
- vmx_set_tsc_shift(v, vpit);
/* We can't resume the guest if we're waiting on I/O */
ASSERT(!test_bit(ARCH_HVM_IO_WAIT, &v->arch.hvm_vcpu.ioflags));
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/x86/hvm/vmx/vmcs.c
--- a/xen/arch/x86/hvm/vmx/vmcs.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/x86/hvm/vmx/vmcs.c Mon Mar 20 08:56:54 2006
@@ -195,7 +195,6 @@
/* Update CR3, GDT, LDT, TR */
unsigned int error = 0;
unsigned long cr0, cr4;
- u64 host_tsc;
if (v->vcpu_id == 0)
hvm_setup_platform(v->domain);
@@ -250,9 +249,7 @@
v->arch.hvm_vmx.launch_cpu = smp_processor_id();
/* init guest tsc to start from 0 */
- rdtscll(host_tsc);
- v->arch.hvm_vmx.tsc_offset = 0 - host_tsc;
- vmx_set_tsc_shift(v, &v->domain->arch.hvm_domain.vpit);
+ set_guest_time(v, 0);
}
/*
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/x86/hvm/vmx/vmx.c
--- a/xen/arch/x86/hvm/vmx/vmx.c Mon Mar 20 08:56:46 2006
+++ b/xen/arch/x86/hvm/vmx/vmx.c Mon Mar 20 08:56:54 2006
@@ -50,9 +50,14 @@
static unsigned long trace_values[NR_CPUS][4];
#define TRACE_VMEXIT(index,value) trace_values[smp_processor_id()][index]=value
+static void vmx_ctxt_switch_from(struct vcpu *v);
+static void vmx_ctxt_switch_to(struct vcpu *v);
+
void vmx_final_setup_guest(struct vcpu *v)
{
- v->arch.schedule_tail = arch_vmx_do_launch;
+ v->arch.schedule_tail = arch_vmx_do_launch;
+ v->arch.ctxt_switch_from = vmx_ctxt_switch_from;
+ v->arch.ctxt_switch_to = vmx_ctxt_switch_to;
if ( v->vcpu_id == 0 )
{
@@ -105,6 +110,7 @@
}
#ifdef __x86_64__
+
static struct vmx_msr_state percpu_msr[NR_CPUS];
static u32 msr_data_index[VMX_MSR_COUNT] =
@@ -113,7 +119,7 @@
MSR_SYSCALL_MASK, MSR_EFER,
};
-void vmx_save_segments(struct vcpu *v)
+static void vmx_save_segments(struct vcpu *v)
{
rdmsrl(MSR_SHADOW_GS_BASE, v->arch.hvm_vmx.msr_content.shadow_gs);
}
@@ -124,7 +130,7 @@
* are not modified once set for generic domains, we don't save them,
* but simply reset them to the values set at percpu_traps_init().
*/
-void vmx_load_msrs(void)
+static void vmx_load_msrs(void)
{
struct vmx_msr_state *host_state = &percpu_msr[smp_processor_id()];
int i;
@@ -166,118 +172,143 @@
#define IS_CANO_ADDRESS(add) 1
static inline int long_mode_do_msr_read(struct cpu_user_regs *regs)
{
- u64 msr_content = 0;
- struct vcpu *vc = current;
- struct vmx_msr_state * msr = &vc->arch.hvm_vmx.msr_content;
- switch(regs->ecx){
+ u64 msr_content = 0;
+ struct vcpu *v = current;
+ struct vmx_msr_state *msr = &v->arch.hvm_vmx.msr_content;
+
+ switch ( regs->ecx ) {
case MSR_EFER:
+ HVM_DBG_LOG(DBG_LEVEL_2, "EFER msr_content 0x%"PRIx64, msr_content);
msr_content = msr->msr_items[VMX_INDEX_MSR_EFER];
- HVM_DBG_LOG(DBG_LEVEL_2, "EFER msr_content %"PRIx64"\n", msr_content);
- if (test_bit(VMX_CPU_STATE_LME_ENABLED,
- &vc->arch.hvm_vmx.cpu_state))
- msr_content |= 1 << _EFER_LME;
-
- if (VMX_LONG_GUEST(vc))
- msr_content |= 1 << _EFER_LMA;
- break;
+
+ /* the following code may be not needed */
+ if ( test_bit(VMX_CPU_STATE_LME_ENABLED, &v->arch.hvm_vmx.cpu_state) )
+ msr_content |= EFER_LME;
+ else
+ msr_content &= ~EFER_LME;
+
+ if ( VMX_LONG_GUEST(v) )
+ msr_content |= EFER_LMA;
+ else
+ msr_content &= ~EFER_LMA;
+ break;
+
case MSR_FS_BASE:
- if (!(VMX_LONG_GUEST(vc)))
+ if ( !(VMX_LONG_GUEST(v)) )
/* XXX should it be GP fault */
domain_crash_synchronous();
+
__vmread(GUEST_FS_BASE, &msr_content);
break;
+
case MSR_GS_BASE:
- if (!(VMX_LONG_GUEST(vc)))
+ if ( !(VMX_LONG_GUEST(v)) )
domain_crash_synchronous();
+
__vmread(GUEST_GS_BASE, &msr_content);
break;
+
case MSR_SHADOW_GS_BASE:
msr_content = msr->shadow_gs;
break;
- CASE_READ_MSR(STAR);
- CASE_READ_MSR(LSTAR);
- CASE_READ_MSR(CSTAR);
- CASE_READ_MSR(SYSCALL_MASK);
+ CASE_READ_MSR(STAR);
+ CASE_READ_MSR(LSTAR);
+ CASE_READ_MSR(CSTAR);
+ CASE_READ_MSR(SYSCALL_MASK);
+
default:
return 0;
}
- HVM_DBG_LOG(DBG_LEVEL_2, "mode_do_msr_read: msr_content: %"PRIx64"\n",
- msr_content);
+
+ HVM_DBG_LOG(DBG_LEVEL_2, "msr_content: 0x%"PRIx64, msr_content);
+
regs->eax = msr_content & 0xffffffff;
regs->edx = msr_content >> 32;
+
return 1;
}
static inline int long_mode_do_msr_write(struct cpu_user_regs *regs)
{
- u64 msr_content = regs->eax | ((u64)regs->edx << 32);
- struct vcpu *vc = current;
- struct vmx_msr_state * msr = &vc->arch.hvm_vmx.msr_content;
- struct vmx_msr_state * host_state =
- &percpu_msr[smp_processor_id()];
-
- HVM_DBG_LOG(DBG_LEVEL_1, " mode_do_msr_write msr %lx "
- "msr_content %"PRIx64"\n",
+ u64 msr_content = regs->eax | ((u64)regs->edx << 32);
+ struct vcpu *v = current;
+ struct vmx_msr_state *msr = &v->arch.hvm_vmx.msr_content;
+ struct vmx_msr_state *host_state = &percpu_msr[smp_processor_id()];
+
+ HVM_DBG_LOG(DBG_LEVEL_1, "msr 0x%lx msr_content 0x%"PRIx64"\n",
(unsigned long)regs->ecx, msr_content);
- switch (regs->ecx){
+ switch ( regs->ecx ) {
case MSR_EFER:
/* offending reserved bit will cause #GP */
- if ( msr_content &
- ~( EFER_LME | EFER_LMA | EFER_NX | EFER_SCE ) )
- vmx_inject_exception(vc, TRAP_gp_fault, 0);
-
- if ((msr_content & EFER_LME) ^
- test_bit(VMX_CPU_STATE_LME_ENABLED,
- &vc->arch.hvm_vmx.cpu_state)){
- if ( vmx_paging_enabled(vc) ||
+ if ( msr_content & ~(EFER_LME | EFER_LMA | EFER_NX | EFER_SCE) )
+ {
+ printk("trying to set reserved bit in EFER\n");
+ vmx_inject_exception(v, TRAP_gp_fault, 0);
+ return 0;
+ }
+
+ /* LME: 0 -> 1 */
+ if ( msr_content & EFER_LME &&
+ !test_bit(VMX_CPU_STATE_LME_ENABLED, &v->arch.hvm_vmx.cpu_state) )
+ {
+ if ( vmx_paging_enabled(v) ||
!test_bit(VMX_CPU_STATE_PAE_ENABLED,
- &vc->arch.hvm_vmx.cpu_state)) {
- vmx_inject_exception(vc, TRAP_gp_fault, 0);
+ &v->arch.hvm_vmx.cpu_state) )
+ {
+ printk("trying to set LME bit when "
+ "in paging mode or PAE bit is not set\n");
+ vmx_inject_exception(v, TRAP_gp_fault, 0);
+ return 0;
}
- }
- if (msr_content & EFER_LME)
- set_bit(VMX_CPU_STATE_LME_ENABLED,
- &vc->arch.hvm_vmx.cpu_state);
-
- msr->msr_items[VMX_INDEX_MSR_EFER] =
- msr_content;
+
+ set_bit(VMX_CPU_STATE_LME_ENABLED, &v->arch.hvm_vmx.cpu_state);
+ }
+
+ msr->msr_items[VMX_INDEX_MSR_EFER] = msr_content;
break;
case MSR_FS_BASE:
case MSR_GS_BASE:
- if (!(VMX_LONG_GUEST(vc)))
+ if ( !(VMX_LONG_GUEST(v)) )
domain_crash_synchronous();
- if (!IS_CANO_ADDRESS(msr_content)){
+
+ if ( !IS_CANO_ADDRESS(msr_content) )
+ {
HVM_DBG_LOG(DBG_LEVEL_1, "Not cano address of msr write\n");
- vmx_inject_exception(vc, TRAP_gp_fault, 0);
- }
- if (regs->ecx == MSR_FS_BASE)
+ vmx_inject_exception(v, TRAP_gp_fault, 0);
+ return 0;
+ }
+
+ if ( regs->ecx == MSR_FS_BASE )
__vmwrite(GUEST_FS_BASE, msr_content);
else
__vmwrite(GUEST_GS_BASE, msr_content);
+
break;
case MSR_SHADOW_GS_BASE:
- if (!(VMX_LONG_GUEST(vc)))
+ if ( !(VMX_LONG_GUEST(v)) )
domain_crash_synchronous();
- vc->arch.hvm_vmx.msr_content.shadow_gs = msr_content;
+
+ v->arch.hvm_vmx.msr_content.shadow_gs = msr_content;
wrmsrl(MSR_SHADOW_GS_BASE, msr_content);
break;
- CASE_WRITE_MSR(STAR);
- CASE_WRITE_MSR(LSTAR);
- CASE_WRITE_MSR(CSTAR);
- CASE_WRITE_MSR(SYSCALL_MASK);
+ CASE_WRITE_MSR(STAR);
+ CASE_WRITE_MSR(LSTAR);
+ CASE_WRITE_MSR(CSTAR);
+ CASE_WRITE_MSR(SYSCALL_MASK);
+
default:
return 0;
}
+
return 1;
}
-void
-vmx_restore_msrs(struct vcpu *v)
+static void vmx_restore_msrs(struct vcpu *v)
{
int i = 0;
struct vmx_msr_state *guest_state;
@@ -297,22 +328,52 @@
HVM_DBG_LOG(DBG_LEVEL_2,
"restore guest's index %d msr %lx with %lx\n",
- i, (unsigned long) msr_data_index[i], (unsigned long)
guest_state->msr_items[i]);
+ i, (unsigned long)msr_data_index[i],
+ (unsigned long)guest_state->msr_items[i]);
set_bit(i, &host_state->flags);
wrmsrl(msr_data_index[i], guest_state->msr_items[i]);
clear_bit(i, &guest_flags);
}
}
#else /* __i386__ */
-#define vmx_save_init_msrs() ((void)0)
-
-static inline int long_mode_do_msr_read(struct cpu_user_regs *regs){
+
+#define vmx_save_segments(v) ((void)0)
+#define vmx_load_msrs() ((void)0)
+#define vmx_restore_msrs(v) ((void)0)
+#define vmx_save_init_msrs() ((void)0)
+
+static inline int long_mode_do_msr_read(struct cpu_user_regs *regs)
+{
return 0;
}
-static inline int long_mode_do_msr_write(struct cpu_user_regs *regs){
+
+static inline int long_mode_do_msr_write(struct cpu_user_regs *regs)
+{
return 0;
}
-#endif
+
+#endif /* __i386__ */
+
+static void vmx_freeze_time(struct vcpu *v)
+{
+ struct hvm_virpit *vpit = &v->domain->arch.hvm_domain.vpit;
+
+ v->domain->arch.hvm_domain.guest_time = get_guest_time(v);
+ if ( vpit->first_injected )
+ stop_timer(&(vpit->pit_timer));
+}
+
+static void vmx_ctxt_switch_from(struct vcpu *v)
+{
+ vmx_freeze_time(v);
+ vmx_save_segments(v);
+ vmx_load_msrs();
+}
+
+static void vmx_ctxt_switch_to(struct vcpu *v)
+{
+ vmx_restore_msrs(v);
+}
void stop_vmx(void)
{
@@ -554,12 +615,6 @@
hvm_funcs.store_cpu_guest_regs = vmx_store_cpu_guest_regs;
hvm_funcs.load_cpu_guest_regs = vmx_load_cpu_guest_regs;
-#ifdef __x86_64__
- hvm_funcs.save_segments = vmx_save_segments;
- hvm_funcs.load_msrs = vmx_load_msrs;
- hvm_funcs.restore_msrs = vmx_restore_msrs;
-#endif
-
hvm_funcs.store_cpu_guest_ctrl_regs = vmx_store_cpu_guest_ctrl_regs;
hvm_funcs.modify_guest_state = vmx_modify_guest_state;
@@ -670,27 +725,31 @@
/* Reserved bits: [31:15], [12:11], [9], [6], [2:1] */
#define VMX_VCPU_CPUID_L1_RESERVED 0xffff9a46
-static void vmx_vmexit_do_cpuid(unsigned long input, struct cpu_user_regs
*regs)
-{
+static void vmx_vmexit_do_cpuid(struct cpu_user_regs *regs)
+{
+ unsigned int input = (unsigned int)regs->eax;
+ unsigned int count = (unsigned int)regs->ecx;
unsigned int eax, ebx, ecx, edx;
unsigned long eip;
struct vcpu *v = current;
__vmread(GUEST_RIP, &eip);
- HVM_DBG_LOG(DBG_LEVEL_1,
- "do_cpuid: (eax) %lx, (ebx) %lx, (ecx) %lx, (edx) %lx,"
- " (esi) %lx, (edi) %lx",
+ HVM_DBG_LOG(DBG_LEVEL_3, "(eax) 0x%08lx, (ebx) 0x%08lx, "
+ "(ecx) 0x%08lx, (edx) 0x%08lx, (esi) 0x%08lx, (edi) 0x%08lx",
(unsigned long)regs->eax, (unsigned long)regs->ebx,
(unsigned long)regs->ecx, (unsigned long)regs->edx,
(unsigned long)regs->esi, (unsigned long)regs->edi);
- cpuid(input, &eax, &ebx, &ecx, &edx);
+ if ( input == 4 )
+ cpuid_count(input, count, &eax, &ebx, &ecx, &edx);
+ else
+ cpuid(input, &eax, &ebx, &ecx, &edx);
if ( input == 1 )
{
if ( hvm_apic_support(v->domain) &&
- !vlapic_global_enabled((VLAPIC(v))) )
+ !vlapic_global_enabled((VLAPIC(v))) )
clear_bit(X86_FEATURE_APIC, &edx);
#if CONFIG_PAGING_LEVELS < 3
@@ -725,10 +784,12 @@
regs->ecx = (unsigned long) ecx;
regs->edx = (unsigned long) edx;
- HVM_DBG_LOG(DBG_LEVEL_1,
- "vmx_vmexit_do_cpuid: eip: %lx, input: %lx, out:eax=%x,
ebx=%x, ecx=%x, edx=%x",
- eip, input, eax, ebx, ecx, edx);
-
+ HVM_DBG_LOG(DBG_LEVEL_3, "eip@%lx, input: 0x%lx, "
+ "output: eax = 0x%08lx, ebx = 0x%08lx, "
+ "ecx = 0x%08lx, edx = 0x%08lx",
+ (unsigned long)eip, (unsigned long)input,
+ (unsigned long)eax, (unsigned long)ebx,
+ (unsigned long)ecx, (unsigned long)edx);
}
#define CASE_GET_REG_P(REG, reg) \
@@ -1656,7 +1717,7 @@
rdtscll(msr_content);
vpit = &(v->domain->arch.hvm_domain.vpit);
- msr_content += vpit->shift;
+ msr_content += vpit->cache_tsc_offset;
break;
}
case MSR_IA32_SYSENTER_CS:
@@ -1700,22 +1761,8 @@
switch (regs->ecx) {
case MSR_IA32_TIME_STAMP_COUNTER:
- {
- struct hvm_virpit *vpit;
- u64 host_tsc, drift;
-
- rdtscll(host_tsc);
- vpit = &(v->domain->arch.hvm_domain.vpit);
- drift = v->arch.hvm_vmx.tsc_offset - vpit->shift;
- vpit->shift = msr_content - host_tsc;
- v->arch.hvm_vmx.tsc_offset = vpit->shift + drift;
- __vmwrite(TSC_OFFSET, vpit->shift);
-
-#if defined (__i386__)
- __vmwrite(TSC_OFFSET_HIGH, ((vpit->shift)>>32));
-#endif
- break;
- }
+ set_guest_time(v, msr_content);
+ break;
case MSR_IA32_SYSENTER_CS:
__vmwrite(GUEST_SYSENTER_CS, msr_content);
break;
@@ -2014,8 +2061,8 @@
__hvm_bug(®s);
break;
case EXIT_REASON_CPUID:
+ vmx_vmexit_do_cpuid(®s);
__get_instruction_length(inst_len);
- vmx_vmexit_do_cpuid(regs.eax, ®s);
__update_guest_eip(inst_len);
break;
case EXIT_REASON_HLT:
diff -r 516cf6553011 -r cf89e8f0831b xen/common/Makefile
--- a/xen/common/Makefile Mon Mar 20 08:56:46 2006
+++ b/xen/common/Makefile Mon Mar 20 08:56:54 2006
@@ -1,19 +1,34 @@
-
include $(BASEDIR)/Rules.mk
-ifneq ($(perfc),y)
-OBJS := $(subst perfc.o,,$(OBJS))
-endif
-ifneq ($(crash_debug),y)
-OBJS := $(patsubst gdbstub.o,,$(OBJS))
-endif
+obj-y += acm_ops.o
+obj-y += bitmap.o
+obj-y += dom0_ops.o
+obj-y += domain.o
+obj-y += elf.o
+obj-y += event_channel.o
+obj-y += grant_table.o
+obj-y += kernel.o
+obj-y += keyhandler.o
+obj-y += lib.o
+obj-y += memory.o
+obj-y += multicall.o
+obj-y += page_alloc.o
+obj-y += rangeset.o
+obj-y += sched_bvt.o
+obj-y += sched_sedf.o
+obj-y += schedule.o
+obj-y += softirq.o
+obj-y += string.o
+obj-y += symbols.o
+obj-y += trace.o
+obj-y += timer.o
+obj-y += vsprintf.o
+obj-y += xmalloc.o
-default: common.o
-common.o: $(OBJS)
- $(LD) $(LDFLAGS) -r -o common.o $(OBJS)
+obj-$(perfc) += perfc.o
+obj-$(crash_debug) += gdbstub.o
-clean:
- rm -f *.o *~ core
+include $(BASEDIR)/Post.mk
# Object file contains changeset and compiler information.
kernel.o: $(BASEDIR)/include/xen/compile.h
diff -r 516cf6553011 -r cf89e8f0831b xen/drivers/Makefile
--- a/xen/drivers/Makefile Mon Mar 20 08:56:46 2006
+++ b/xen/drivers/Makefile Mon Mar 20 08:56:54 2006
@@ -1,8 +1,6 @@
+include $(BASEDIR)/Rules.mk
-default:
- $(MAKE) -C char
- $(MAKE) -C acpi
+subdirs-y := char/
+subdirs-$(HAS_ACPI) += acpi/
-clean:
- $(MAKE) -C char clean
- $(MAKE) -C acpi clean
+include $(BASEDIR)/Post.mk
diff -r 516cf6553011 -r cf89e8f0831b xen/drivers/acpi/Makefile
--- a/xen/drivers/acpi/Makefile Mon Mar 20 08:56:46 2006
+++ b/xen/drivers/acpi/Makefile Mon Mar 20 08:56:54 2006
@@ -1,11 +1,5 @@
-
include $(BASEDIR)/Rules.mk
-OBJS := tables.o
+obj-y += tables.o
-default: driver.o
-driver.o: $(OBJS)
- $(LD) $(LDFLAGS) -r -o driver.o $(OBJS)
-
-clean:
- rm -f *.o *~ core
+include $(BASEDIR)/Post.mk
diff -r 516cf6553011 -r cf89e8f0831b xen/drivers/char/Makefile
--- a/xen/drivers/char/Makefile Mon Mar 20 08:56:46 2006
+++ b/xen/drivers/char/Makefile Mon Mar 20 08:56:54 2006
@@ -1,12 +1,10 @@
-
include $(BASEDIR)/Rules.mk
-default: driver.o
-driver.o: $(OBJS)
- $(LD) $(LDFLAGS) -r -o driver.o $(OBJS)
+obj-y += console.o
+obj-y += ns16550.o
+obj-y += serial.o
-clean:
- rm -f *.o *~ core
+include $(BASEDIR)/Post.mk
# Object file contains changeset and compiler information.
console.o: $(BASEDIR)/include/xen/compile.h
diff -r 516cf6553011 -r cf89e8f0831b xen/include/asm-ia64/config.h
--- a/xen/include/asm-ia64/config.h Mon Mar 20 08:56:46 2006
+++ b/xen/include/asm-ia64/config.h Mon Mar 20 08:56:54 2006
@@ -74,7 +74,7 @@
extern unsigned long dom0_size;
// from linux/include/linux/mm.h
-extern struct page *mem_map;
+extern struct page_info *mem_map;
// xen/include/asm/config.h
extern char _end[]; /* standard ELF symbol */
@@ -134,9 +134,6 @@
#define smp_num_siblings 1
#endif
-// from linux/include/linux/mm.h
-struct page;
-
// function calls; see decl in xen/include/xen/sched.h
#undef free_task_struct
#undef alloc_task_struct
@@ -206,8 +203,6 @@
#define _atomic_read(v) ((v).counter)
#define atomic_compareandswap(old, new, v) ((atomic_t){ cmpxchg(v,
_atomic_read(old), _atomic_read(new)) })
-// see include/asm-ia64/mm.h, handle remaining page_info uses until gone
-#define page_info page
// Deprivated linux inf and put here for short time compatibility
#define kmalloc(s, t) xmalloc_bytes((s))
#define kfree(s) xfree((s))
@@ -249,7 +244,7 @@
extern char saved_command_line[];
struct screen_info { };
#define seq_printf(a,b...) printf(b)
-#define CONFIG_BLK_DEV_INITRD // needed to reserve memory for domain0
+//#define CONFIG_BLK_DEV_INITRD // needed to reserve memory for domain0
void dummy_called(char *function);
#define dummy() dummy_called((char *) __FUNCTION__)
diff -r 516cf6553011 -r cf89e8f0831b xen/include/asm-ia64/dom_fw.h
--- a/xen/include/asm-ia64/dom_fw.h Mon Mar 20 08:56:46 2006
+++ b/xen/include/asm-ia64/dom_fw.h Mon Mar 20 08:56:54 2006
@@ -119,7 +119,7 @@
#define FW_HYPERCALL_EFI_GET_NEXT_HIGH_MONO_COUNT_PADDR
FW_HYPERCALL_PADDR(FW_HYPERCALL_EFI_GET_NEXT_HIGH_MONO_COUNT_INDEX)
#define FW_HYPERCALL_EFI_RESET_SYSTEM_PADDR
FW_HYPERCALL_PADDR(FW_HYPERCALL_EFI_RESET_SYSTEM_INDEX)
-extern struct ia64_pal_retval xen_pal_emulator(UINT64,UINT64,UINT64,UINT64);
+extern struct ia64_pal_retval xen_pal_emulator(UINT64, u64, u64, u64);
extern struct sal_ret_values sal_emulator (long index, unsigned long in1,
unsigned long in2, unsigned long in3, unsigned long in4, unsigned long in5,
unsigned long in6, unsigned long in7);
extern struct ia64_pal_retval pal_emulator_static (unsigned long);
diff -r 516cf6553011 -r cf89e8f0831b xen/include/asm-ia64/domain.h
--- a/xen/include/asm-ia64/domain.h Mon Mar 20 08:56:46 2006
+++ b/xen/include/asm-ia64/domain.h Mon Mar 20 08:56:54 2006
@@ -14,7 +14,6 @@
extern void domain_relinquish_resources(struct domain *);
struct arch_domain {
- struct mm_struct *active_mm;
struct mm_struct *mm;
unsigned long metaphysical_rr0;
unsigned long metaphysical_rr4;
@@ -68,7 +67,6 @@
int breakimm; // from arch_domain (so is pinned)
int starting_rid; /* first RID assigned to domain */
int ending_rid; /* one beyond highest RID assigned to domain */
- struct mm_struct *active_mm;
struct thread_struct _thread; // this must be last
thash_cb_t *vtlb;
@@ -81,7 +79,6 @@
struct arch_vmx_struct arch_vmx; /* Virtual Machine Extensions */
};
-#define active_mm arch.active_mm
//#define thread arch._thread
// FOLLOWING FROM linux-2.6.7/include/sched.h
@@ -101,6 +98,8 @@
struct rw_semaphore mmap_sem;
#endif
spinlock_t page_table_lock; /* Protects task page tables
and mm->rss */
+
+ struct list_head pt_list; /* List of pagetable */
struct list_head mmlist; /* List of all active mm's.
These are globally strung
* together off init_mm.mmlist,
and are protected
diff -r 516cf6553011 -r cf89e8f0831b xen/include/asm-ia64/flushtlb.h
--- a/xen/include/asm-ia64/flushtlb.h Mon Mar 20 08:56:46 2006
+++ b/xen/include/asm-ia64/flushtlb.h Mon Mar 20 08:56:54 2006
@@ -1,5 +1,7 @@
#ifndef __FLUSHTLB_H__
#define __FLUSHTLB_H__
+
+#include <asm/tlbflush.h>
/* The current time as shown by the virtual TLB clock. */
extern u32 tlbflush_clock;
diff -r 516cf6553011 -r cf89e8f0831b xen/include/asm-ia64/linux-xen/asm/page.h
--- a/xen/include/asm-ia64/linux-xen/asm/page.h Mon Mar 20 08:56:46 2006
+++ b/xen/include/asm-ia64/linux-xen/asm/page.h Mon Mar 20 08:56:54 2006
@@ -75,7 +75,7 @@
flush_dcache_page(page); \
} while (0)
-
+#ifndef XEN
#define alloc_zeroed_user_highpage(vma, vaddr) \
({ \
struct page *page = alloc_page_vma(GFP_HIGHUSER | __GFP_ZERO, vma,
vaddr); \
@@ -83,6 +83,7 @@
flush_dcache_page(page); \
page; \
})
+#endif
#define __HAVE_ARCH_ALLOC_ZEROED_USER_HIGHPAGE
diff -r 516cf6553011 -r cf89e8f0831b
xen/include/asm-ia64/linux-xen/asm/pgalloc.h
--- a/xen/include/asm-ia64/linux-xen/asm/pgalloc.h Mon Mar 20 08:56:46 2006
+++ b/xen/include/asm-ia64/linux-xen/asm/pgalloc.h Mon Mar 20 08:56:54 2006
@@ -106,11 +106,13 @@
#define __pmd_free_tlb(tlb, pmd) pmd_free(pmd)
+#ifndef XEN
static inline void
pmd_populate(struct mm_struct *mm, pmd_t * pmd_entry, struct page *pte)
{
pmd_val(*pmd_entry) = page_to_maddr(pte);
}
+#endif
static inline void
pmd_populate_kernel(struct mm_struct *mm, pmd_t * pmd_entry, pte_t * pte)
@@ -118,11 +120,13 @@
pmd_val(*pmd_entry) = __pa(pte);
}
+#ifndef XEN
static inline struct page *pte_alloc_one(struct mm_struct *mm,
unsigned long addr)
{
return virt_to_page(pgtable_quicklist_alloc());
}
+#endif
static inline pte_t *pte_alloc_one_kernel(struct mm_struct *mm,
unsigned long addr)
@@ -130,6 +134,7 @@
return pgtable_quicklist_alloc();
}
+#ifndef XEN
static inline void pte_free(struct page *pte)
{
pgtable_quicklist_free(page_address(pte));
@@ -141,6 +146,7 @@
}
#define __pte_free_tlb(tlb, pte) pte_free(pte)
+#endif
extern void check_pgt_cache(void);
diff -r 516cf6553011 -r cf89e8f0831b
xen/include/asm-ia64/linux-xen/asm/pgtable.h
--- a/xen/include/asm-ia64/linux-xen/asm/pgtable.h Mon Mar 20 08:56:46 2006
+++ b/xen/include/asm-ia64/linux-xen/asm/pgtable.h Mon Mar 20 08:56:54 2006
@@ -467,8 +467,10 @@
* for zero-mapped memory areas etc..
*/
extern unsigned long empty_zero_page[PAGE_SIZE/sizeof(unsigned long)];
+#ifndef XEN
extern struct page *zero_page_memmap_ptr;
#define ZERO_PAGE(vaddr) (zero_page_memmap_ptr)
+#endif
/* We provide our own get_unmapped_area to cope with VA holes for userland */
#define HAVE_ARCH_UNMAPPED_AREA
diff -r 516cf6553011 -r cf89e8f0831b xen/include/asm-ia64/linux-xen/asm/system.h
--- a/xen/include/asm-ia64/linux-xen/asm/system.h Mon Mar 20 08:56:46 2006
+++ b/xen/include/asm-ia64/linux-xen/asm/system.h Mon Mar 20 08:56:54 2006
@@ -290,6 +290,9 @@
#ifdef XEN
#include <asm/xensystem.h>
+#ifndef __ASSEMBLY__
+struct resource;
+#endif
#endif
#endif /* _ASM_IA64_SYSTEM_H */
diff -r 516cf6553011 -r cf89e8f0831b
xen/include/asm-ia64/linux-xen/asm/tlbflush.h
--- a/xen/include/asm-ia64/linux-xen/asm/tlbflush.h Mon Mar 20 08:56:46 2006
+++ b/xen/include/asm-ia64/linux-xen/asm/tlbflush.h Mon Mar 20 08:56:54 2006
@@ -80,7 +80,11 @@
#ifdef CONFIG_SMP
flush_tlb_range(vma, (addr & PAGE_MASK), (addr & PAGE_MASK) +
PAGE_SIZE);
#else
+#ifdef XEN
+ if (vma->vm_mm == current->domain->arch.mm)
+#else
if (vma->vm_mm == current->active_mm)
+#endif
ia64_ptcl(addr, (PAGE_SHIFT << 2));
#ifndef XEN
// FIXME SMP?
diff -r 516cf6553011 -r cf89e8f0831b xen/include/asm-ia64/linux-xen/linux/gfp.h
--- a/xen/include/asm-ia64/linux-xen/linux/gfp.h Mon Mar 20 08:56:46 2006
+++ b/xen/include/asm-ia64/linux-xen/linux/gfp.h Mon Mar 20 08:56:54 2006
@@ -3,6 +3,7 @@
#ifdef XEN
#include <asm/bitops.h>
+#include <linux/topology.h>
#endif
#include <linux/mmzone.h>
#include <linux/stddef.h>
@@ -81,6 +82,7 @@
* optimized to &contig_page_data at compile-time.
*/
+#ifndef XEN
#ifndef HAVE_ARCH_FREE_PAGE
static inline void arch_free_page(struct page *page, int order) { }
#endif
@@ -134,6 +136,7 @@
#define __free_page(page) __free_pages((page), 0)
#define free_page(addr) free_pages((addr),0)
+#endif /* XEN */
void page_alloc_init(void);
#ifdef CONFIG_NUMA
diff -r 516cf6553011 -r cf89e8f0831b xen/include/asm-ia64/linux/README.origin
--- a/xen/include/asm-ia64/linux/README.origin Mon Mar 20 08:56:46 2006
+++ b/xen/include/asm-ia64/linux/README.origin Mon Mar 20 08:56:54 2006
@@ -14,13 +14,10 @@
jiffies.h -> linux/include/linux/jiffies.h
kmalloc_sizes.h -> linux/include/linux/kmalloc_sizes.h
linkage.h -> linux/include/linux/linkage.h
-mmzone.h -> linux/include/linux/mmzone.h
notifier.h -> linux/include/linux/notifier.h
numa.h -> linux/include/linux/numa.h
-page-flags.h -> linux/include/linux/page-flags.h
percpu.h -> linux/include/linux/percpu.h
preempt.h -> linux/include/linux/preempt.h
-rbtree.h -> linux/include/linux/rbtree.h
rwsem.h -> linux/include/linux/rwsem.h
seqlock.h -> linux/include/linux/seqlock.h
sort.h -> linux/include/linux/sort.h
diff -r 516cf6553011 -r cf89e8f0831b
xen/include/asm-ia64/linux/asm-generic/README.origin
--- a/xen/include/asm-ia64/linux/asm-generic/README.origin Mon Mar 20
08:56:46 2006
+++ b/xen/include/asm-ia64/linux/asm-generic/README.origin Mon Mar 20
08:56:54 2006
@@ -10,7 +10,6 @@
errno.h -> linux/include/asm-generic/errno.h
ide_iops.h -> linux/include/asm-generic/ide_iops.h
iomap.h -> linux/include/asm-generic/iomap.h
-pci-dma-compat.h -> linux/include/asm-generic/pci-dma-compat.h
pci.h -> linux/include/asm-generic/pci.h
pgtable.h -> linux/include/asm-generic/pgtable.h
pgtable-nopud.h -> linux/include/asm-generic/pgtable-nopud.h
diff -r 516cf6553011 -r cf89e8f0831b
xen/include/asm-ia64/linux/asm/README.origin
--- a/xen/include/asm-ia64/linux/asm/README.origin Mon Mar 20 08:56:46 2006
+++ b/xen/include/asm-ia64/linux/asm/README.origin Mon Mar 20 08:56:54 2006
@@ -42,7 +42,6 @@
rse.h -> linux/include/asm-ia64/rse.h
rwsem.h -> linux/include/asm-ia64/rwsem.h
sal.h -> linux/include/asm-ia64/sal.h
-scatterlist.h -> linux/include/asm-ia64/scatterlist.h
sections.h -> linux/include/asm-ia64/sections.h
semaphore.h -> linux/include/asm-ia64/semaphore.h
setup.h -> linux/include/asm-ia64/setup.h
diff -r 516cf6553011 -r cf89e8f0831b xen/include/asm-ia64/linux/asm/hw_irq.h
--- a/xen/include/asm-ia64/linux/asm/hw_irq.h Mon Mar 20 08:56:46 2006
+++ b/xen/include/asm-ia64/linux/asm/hw_irq.h Mon Mar 20 08:56:54 2006
@@ -85,6 +85,7 @@
extern void free_irq_vector (int vector);
extern void ia64_send_ipi (int cpu, int vector, int delivery_mode, int
redirect);
extern void register_percpu_irq (ia64_vector vec, struct irqaction *action);
+extern int xen_do_IRQ(ia64_vector vector);
static inline void
hw_resend_irq (struct hw_interrupt_type *h, unsigned int vector)
diff -r 516cf6553011 -r cf89e8f0831b xen/include/asm-ia64/linux/asm/irq.h
--- a/xen/include/asm-ia64/linux/asm/irq.h Mon Mar 20 08:56:46 2006
+++ b/xen/include/asm-ia64/linux/asm/irq.h Mon Mar 20 08:56:54 2006
@@ -40,4 +40,6 @@
struct pt_regs;
int handle_IRQ_event(unsigned int, struct pt_regs *, struct irqaction *);
+extern fastcall unsigned int __do_IRQ(unsigned int irq, struct pt_regs *regs);
+
#endif /* _ASM_IA64_IRQ_H */
diff -r 516cf6553011 -r cf89e8f0831b xen/include/asm-ia64/mm.h
--- a/xen/include/asm-ia64/mm.h Mon Mar 20 08:56:46 2006
+++ b/xen/include/asm-ia64/mm.h Mon Mar 20 08:56:54 2006
@@ -9,8 +9,6 @@
#include <xen/spinlock.h>
#include <xen/perfc.h>
#include <xen/sched.h>
-
-#include <linux/rbtree.h>
#include <asm/processor.h>
#include <asm/atomic.h>
@@ -36,37 +34,38 @@
#define PRtype_info "08x"
-struct page
+struct page_info
{
/* Each frame can be threaded onto a doubly-linked list. */
struct list_head list;
+ /* Reference count and various PGC_xxx flags and fields. */
+ u32 count_info;
+
+ /* Context-dependent fields follow... */
+ union {
+
+ /* Page is in use: ((count_info & PGC_count_mask) != 0). */
+ struct {
+ /* Owner of this page (NULL if page is anonymous). */
+ u32 _domain; /* pickled format */
+ /* Type reference count and various PGT_xxx flags and fields. */
+ unsigned long type_info;
+ } __attribute__ ((packed)) inuse;
+
+ /* Page is on a free list: ((count_info & PGC_count_mask) == 0). */
+ struct {
+ /* Order-size of the free chunk this page is the head of. */
+ u32 order;
+ /* Mask of possibly-tainted TLBs. */
+ cpumask_t cpumask;
+ } __attribute__ ((packed)) free;
+
+ } u;
+
/* Timestamp from 'TLB clock', used to reduce need for safety flushes. */
u32 tlbflush_timestamp;
- /* Reference count and various PGC_xxx flags and fields. */
- u32 count_info;
-
- /* Context-dependent fields follow... */
- union {
-
- /* Page is in use by a domain. */
- struct {
- /* Owner of this page. */
- u32 _domain;
- /* Type reference count and various PGT_xxx flags and fields. */
- u32 type_info;
- } inuse;
-
- /* Page is on a free list. */
- struct {
- /* Mask of possibly-tainted TLBs. */
- cpumask_t cpumask;
- /* Order-size of the free chunk this page is the head of. */
- u8 order;
- } free;
-
- } u;
#if 0
// following added for Linux compiling
page_flags_t flags;
@@ -77,34 +76,46 @@
#define set_page_count(p,v) atomic_set(&(p)->_count, v - 1)
-/* Still small set of flags defined by far on IA-64 */
+/*
+ * Still small set of flags defined by far on IA-64.
+ * IA-64 should make it a definition same as x86_64.
+ */
/* The following page types are MUTUALLY EXCLUSIVE. */
#define PGT_none (0<<29) /* no special uses of this page */
#define PGT_l1_page_table (1<<29) /* using this page as an L1 page table? */
#define PGT_l2_page_table (2<<29) /* using this page as an L2 page table? */
#define PGT_l3_page_table (3<<29) /* using this page as an L3 page table? */
#define PGT_l4_page_table (4<<29) /* using this page as an L4 page table? */
-#define PGT_writable_page (5<<29) /* has writable mappings of this page? */
-#define PGT_type_mask (5<<29) /* Bits 29-31. */
+ /* Value 5 reserved. See asm-x86/mm.h */
+ /* Value 6 reserved. See asm-x86/mm.h */
+#define PGT_writable_page (7<<29) /* has writable mappings of this page? */
+#define PGT_type_mask (7<<29) /* Bits 29-31. */
/* Has this page been validated for use as its current type? */
#define _PGT_validated 28
#define PGT_validated (1<<_PGT_validated)
-/* Owning guest has pinned this page to its current type? */
+ /* Owning guest has pinned this page to its current type? */
#define _PGT_pinned 27
#define PGT_pinned (1U<<_PGT_pinned)
-/* 27-bit count of uses of this frame as its current type. */
-#define PGT_count_mask ((1U<<27)-1)
-
-/* Cleared when the owning guest 'frees' this page. */
+ /* The 27 most significant bits of virt address if this is a page table. */
+#define PGT_va_shift 32
+#define PGT_va_mask ((unsigned long)((1U<<28)-1)<<PGT_va_shift)
+ /* Is the back pointer still mutable (i.e. not fixed yet)? */
+#define PGT_va_mutable ((unsigned long)((1U<<28)-1)<<PGT_va_shift)
+ /* Is the back pointer unknown (e.g., p.t. is mapped at multiple VAs)? */
+#define PGT_va_unknown ((unsigned long)((1U<<28)-2)<<PGT_va_shift)
+
+ /* 16-bit count of uses of this frame as its current type. */
+#define PGT_count_mask ((1U<<16)-1)
+
+ /* Cleared when the owning guest 'frees' this page. */
#define _PGC_allocated 31
#define PGC_allocated (1U<<_PGC_allocated)
-/* Set when the page is used as a page table */
-#define _PGC_page_table 30
-#define PGC_page_table (1U<<_PGC_page_table)
-/* 30-bit count of references to this frame. */
-#define PGC_count_mask ((1U<<30)-1)
+ /* Bit 30 reserved. See asm-x86/mm.h */
+ /* Bit 29 reserved. See asm-x86/mm.h */
+ /* 29-bit count of references to this frame. */
+#define PGC_count_mask ((1U<<29)-1)
#define IS_XEN_HEAP_FRAME(_pfn) ((page_to_maddr(_pfn) < xenheap_phys_end) \
&& (page_to_maddr(_pfn) >= xen_pstart))
@@ -139,7 +150,6 @@
static inline void put_page(struct page_info *page)
{
-#ifdef VALIDATE_VT // doesn't work with non-VTI in grant tables yet
u32 nx, x, y = page->count_info;
do {
@@ -150,14 +160,12 @@
if (unlikely((nx & PGC_count_mask) == 0))
free_domheap_page(page);
-#endif
}
/* count_info and ownership are checked atomically. */
static inline int get_page(struct page_info *page,
struct domain *domain)
{
-#ifdef VALIDATE_VT
u64 x, nx, y = *((u64*)&page->count_info);
u32 _domain = pickle_domptr(domain);
@@ -173,14 +181,13 @@
return 0;
}
}
- while(unlikely(y = cmpxchg(&page->count_info, x, nx)) != x);
-#endif
+ while(unlikely((y = cmpxchg((u64*)&page->count_info, x, nx)) != x));
return 1;
}
-/* No type info now */
-#define put_page_type(page)
-#define get_page_type(page, type) 1
+extern void put_page_type(struct page_info *page);
+extern int get_page_type(struct page_info *page, u32 type);
+
static inline void put_page_and_type(struct page_info *page)
{
put_page_type(page);
@@ -219,7 +226,7 @@
// prototype of misc memory stuff
//unsigned long __get_free_pages(unsigned int mask, unsigned int order);
-//void __free_pages(struct page *page, unsigned int order);
+//void __free_pages(struct page_info *page, unsigned int order);
void *pgtable_quicklist_alloc(void);
void pgtable_quicklist_free(void *pgtable_entry);
@@ -339,11 +346,11 @@
#define NODEZONE_SHIFT (sizeof(page_flags_t)*8 - MAX_NODES_SHIFT -
MAX_ZONES_SHIFT)
#define NODEZONE(node, zone) ((node << ZONES_SHIFT) | zone)
-static inline unsigned long page_zonenum(struct page *page)
+static inline unsigned long page_zonenum(struct page_info *page)
{
return (page->flags >> NODEZONE_SHIFT) & (~(~0UL << ZONES_SHIFT));
}
-static inline unsigned long page_to_nid(struct page *page)
+static inline unsigned long page_to_nid(struct page_info *page)
{
return (page->flags >> (NODEZONE_SHIFT + ZONES_SHIFT));
}
@@ -351,12 +358,12 @@
struct zone;
extern struct zone *zone_table[];
-static inline struct zone *page_zone(struct page *page)
+static inline struct zone *page_zone(struct page_info *page)
{
return zone_table[page->flags >> NODEZONE_SHIFT];
}
-static inline void set_page_zone(struct page *page, unsigned long nodezone_num)
+static inline void set_page_zone(struct page_info *page, unsigned long
nodezone_num)
{
page->flags &= ~(~0UL << NODEZONE_SHIFT);
page->flags |= nodezone_num << NODEZONE_SHIFT;
@@ -367,7 +374,7 @@
extern unsigned long max_mapnr;
#endif
-static inline void *lowmem_page_address(struct page *page)
+static inline void *lowmem_page_address(struct page_info *page)
{
return __va(page_to_mfn(page) << PAGE_SHIFT);
}
@@ -386,8 +393,8 @@
#endif
#if defined(HASHED_PAGE_VIRTUAL)
-void *page_address(struct page *page);
-void set_page_address(struct page *page, void *virtual);
+void *page_address(struct page_info *page);
+void set_page_address(struct page_info *page, void *virtual);
void page_address_init(void);
#endif
@@ -400,7 +407,7 @@
#ifndef CONFIG_DEBUG_PAGEALLOC
static inline void
-kernel_map_pages(struct page *page, int numpages, int enable)
+kernel_map_pages(struct page_info *page, int numpages, int enable)
{
}
#endif
@@ -415,8 +422,8 @@
#undef machine_to_phys_mapping
#define machine_to_phys_mapping mpt_table
-#define INVALID_M2P_ENTRY (~0U)
-#define VALID_M2P(_e) (!((_e) & (1U<<63)))
+#define INVALID_M2P_ENTRY (~0UL)
+#define VALID_M2P(_e) (!((_e) & (1UL<<63)))
#define IS_INVALID_M2P_ENTRY(_e) (!VALID_M2P(_e))
#define set_gpfn_from_mfn(mfn, pfn) (machine_to_phys_mapping[(mfn)] = (pfn))
@@ -463,4 +470,6 @@
/* Arch-specific portion of memory_op hypercall. */
#define arch_memory_op(op, arg) (-ENOSYS)
+extern void assign_domain_page(struct domain *d, unsigned long mpaddr,
+ unsigned long physaddr);
#endif /* __ASM_IA64_MM_H__ */
diff -r 516cf6553011 -r cf89e8f0831b xen/include/asm-ia64/regionreg.h
--- a/xen/include/asm-ia64/regionreg.h Mon Mar 20 08:56:46 2006
+++ b/xen/include/asm-ia64/regionreg.h Mon Mar 20 08:56:54 2006
@@ -64,4 +64,13 @@
// since vmMangleRID is symmetric, use it for unmangling also
#define vmUnmangleRID(x) vmMangleRID(x)
+extern unsigned long allocate_metaphysical_rr(void);
+
+struct domain;
+extern int allocate_rid_range(struct domain *d, unsigned long ridbits);
+extern int deallocate_rid_range(struct domain *d);
+
+struct vcpu;
+extern void init_all_rr(struct vcpu *v);
+
#endif /* !_REGIONREG_H_ */
diff -r 516cf6553011 -r cf89e8f0831b xen/include/asm-ia64/vcpu.h
--- a/xen/include/asm-ia64/vcpu.h Mon Mar 20 08:56:46 2006
+++ b/xen/include/asm-ia64/vcpu.h Mon Mar 20 08:56:54 2006
@@ -7,7 +7,6 @@
//#include "thread.h"
#include <asm/ia64_int.h>
#include <public/arch-ia64.h>
-
typedef unsigned long UINT64;
typedef unsigned int UINT;
typedef int BOOLEAN;
@@ -16,7 +15,10 @@
typedef cpu_user_regs_t REGS;
-#define VCPU(_v,_x) _v->arch.privregs->_x
+
+#define VCPU(_v,_x) (_v->arch.privregs->_x)
+#define PSCB(_v,_x) VCPU(_v,_x)
+#define PSCBX(_v,_x) (_v->arch._x)
#define PRIVOP_ADDR_COUNT
#ifdef PRIVOP_ADDR_COUNT
@@ -140,7 +142,9 @@
extern IA64FAULT vcpu_ptc_ga(VCPU *vcpu, UINT64 vadr, UINT64 addr_range);
extern IA64FAULT vcpu_ptr_d(VCPU *vcpu,UINT64 vadr, UINT64 addr_range);
extern IA64FAULT vcpu_ptr_i(VCPU *vcpu,UINT64 vadr, UINT64 addr_range);
-extern IA64FAULT vcpu_translate(VCPU *vcpu, UINT64 address, BOOLEAN is_data,
UINT64 *pteval, UINT64 *itir, UINT64 *iha);
+extern IA64FAULT vcpu_translate(VCPU *vcpu, UINT64 address,
+ BOOLEAN is_data, BOOLEAN in_tpa,
+ UINT64 *pteval, UINT64 *itir, UINT64 *iha);
extern IA64FAULT vcpu_tpa(VCPU *vcpu, UINT64 vadr, UINT64 *padr);
extern IA64FAULT vcpu_force_data_miss(VCPU *vcpu, UINT64 ifa);
extern IA64FAULT vcpu_fc(VCPU *vcpu, UINT64 vadr);
@@ -173,4 +177,18 @@
return (~((1UL << itir_ps(itir)) - 1));
}
+#define verbose(a...) do {if (vcpu_verbose) printf(a);} while(0)
+
+//#define vcpu_quick_region_check(_tr_regions,_ifa) 1
+#define vcpu_quick_region_check(_tr_regions,_ifa) \
+ (_tr_regions & (1 << ((unsigned long)_ifa >> 61)))
+#define vcpu_quick_region_set(_tr_regions,_ifa) \
+ do {_tr_regions |= (1 << ((unsigned long)_ifa >> 61)); } while (0)
+
+// FIXME: also need to check && (!trp->key || vcpu_pkr_match(trp->key))
+#define vcpu_match_tr_entry(_trp,_ifa,_rid) \
+ ((_trp->p && (_trp->rid==_rid) && (_ifa >= _trp->vadr) && \
+ (_ifa < (_trp->vadr + (1L<< _trp->ps)) - 1)))
+
+
#endif
diff -r 516cf6553011 -r cf89e8f0831b xen/include/asm-ia64/vmmu.h
--- a/xen/include/asm-ia64/vmmu.h Mon Mar 20 08:56:46 2006
+++ b/xen/include/asm-ia64/vmmu.h Mon Mar 20 08:56:54 2006
@@ -68,11 +68,14 @@
} search_section_t;
-typedef enum {
+enum {
ISIDE_TLB=0,
DSIDE_TLB=1
-} CACHE_LINE_TYPE;
-
+};
+#define VTLB_PTE_P_BIT 0
+#define VTLB_PTE_IO_BIT 60
+#define VTLB_PTE_IO (1UL<<VTLB_PTE_IO_BIT)
+#define VTLB_PTE_P (1UL<<VTLB_PTE_P_BIT)
typedef struct thash_data {
union {
struct {
@@ -86,18 +89,16 @@
u64 ppn : 38; // 12-49
u64 rv2 : 2; // 50-51
u64 ed : 1; // 52
- u64 ig1 : 3; // 53-55
- u64 len : 4; // 56-59
- u64 ig2 : 3; // 60-63
+ u64 ig1 : 3; // 53-63
};
struct {
u64 __rv1 : 53; // 0-52
u64 contiguous : 1; //53
u64 tc : 1; // 54 TR or TC
- CACHE_LINE_TYPE cl : 1; // 55 I side or D side cache line
+ u64 cl : 1; // 55 I side or D side cache line
// next extension to ig1, only for TLB instance
- u64 __ig1 : 4; // 56-59
- u64 locked : 1; // 60 entry locked or not
+ u64 len : 4; // 56-59
+ u64 io : 1; // 60 entry is for io or not
u64 nomap : 1; // 61 entry cann't be inserted into machine TLB.
u64 checked : 1; // 62 for VTLB/VHPT sanity check
u64 invalid : 1; // 63 invalid entry
@@ -112,12 +113,12 @@
u64 key : 24; // 8-31
u64 rv4 : 32; // 32-63
};
- struct {
- u64 __rv3 : 32; // 0-31
+// struct {
+// u64 __rv3 : 32; // 0-31
// next extension to rv4
- u64 rid : 24; // 32-55
- u64 __rv4 : 8; // 56-63
- };
+// u64 rid : 24; // 32-55
+// u64 __rv4 : 8; // 56-63
+// };
u64 itir;
};
union {
@@ -136,7 +137,8 @@
};
union {
struct thash_data *next;
- u64 tr_idx;
+ u64 rid; // only used in guest TR
+// u64 tr_idx;
};
} thash_data_t;
@@ -152,7 +154,7 @@
#define INVALID_VHPT(hdata) ((hdata)->ti)
#define INVALID_TLB(hdata) ((hdata)->ti)
-#define INVALID_TR(hdata) ((hdata)->invalid)
+#define INVALID_TR(hdata) (!(hdata)->p)
#define INVALID_ENTRY(hcb, hdata) INVALID_VHPT(hdata)
/* ((hcb)->ht==THASH_TLB ? INVALID_TLB(hdata) : INVALID_VHPT(hdata)) */
@@ -199,18 +201,18 @@
typedef void (REM_THASH_FN)(struct thash_cb *hcb, thash_data_t *entry);
typedef void (INS_THASH_FN)(struct thash_cb *hcb, thash_data_t *entry, u64 va);
-typedef struct tlb_special {
- thash_data_t itr[NITRS];
- thash_data_t dtr[NDTRS];
- struct thash_cb *vhpt;
-} tlb_special_t;
+//typedef struct tlb_special {
+// thash_data_t itr[NITRS];
+// thash_data_t dtr[NDTRS];
+// struct thash_cb *vhpt;
+//} tlb_special_t;
//typedef struct vhpt_cb {
//u64 pta; // pta value.
// GET_MFN_FN *get_mfn;
// TTAG_FN *tag_func;
//} vhpt_special;
-
+/*
typedef struct thash_internal {
thash_data_t *hash_base;
thash_data_t *cur_cch; // head of overlap search
@@ -227,7 +229,7 @@
u64 _curva; // current address to search
u64 _eva;
} thash_internal_t;
-
+ */
#define THASH_CB_MAGIC 0x55aa00aa55aa55aaUL
typedef struct thash_cb {
/* THASH base information */
@@ -243,6 +245,7 @@
thash_cch_mem_t *cch_freelist;
struct vcpu *vcpu;
PTA pta;
+ struct thash_cb *vhpt;
/* VTLB/VHPT common information */
// FIND_OVERLAP_FN *find_overlap;
// FIND_NEXT_OVL_FN *next_overlap;
@@ -251,15 +254,15 @@
// REM_NOTIFIER_FN *remove_notifier;
/* private information */
// thash_internal_t priv;
- union {
- tlb_special_t *ts;
+// union {
+// tlb_special_t *ts;
// vhpt_special *vs;
- };
+// };
// Internal positon information, buffer and storage etc. TBD
} thash_cb_t;
-#define ITR(hcb,id) ((hcb)->ts->itr[id])
-#define DTR(hcb,id) ((hcb)->ts->dtr[id])
+//#define ITR(hcb,id) ((hcb)->ts->itr[id])
+//#define DTR(hcb,id) ((hcb)->ts->dtr[id])
#define INVALIDATE_HASH_HEADER(hcb,hash) INVALIDATE_TLB_HEADER(hash)
/* \
{ if ((hcb)->ht==THASH_TLB){ \
@@ -290,10 +293,10 @@
* 4: Return the entry in hash table or collision chain.
*
*/
-extern void thash_vhpt_insert(thash_cb_t *hcb, thash_data_t *entry, u64 va);
+extern void thash_vhpt_insert(thash_cb_t *hcb, u64 pte, u64 itir, u64 ifa);
//extern void thash_insert(thash_cb_t *hcb, thash_data_t *entry, u64 va);
-extern void thash_tr_insert(thash_cb_t *hcb, thash_data_t *entry, u64 va, int
idx);
-extern thash_data_t *vtr_find_overlap(thash_cb_t *hcb, thash_data_t *data,
char cl);
+//extern void thash_tr_insert(thash_cb_t *hcb, thash_data_t *entry, u64 va,
int idx);
+extern int vtr_find_overlap(struct vcpu *vcpu, u64 va, u64 ps, int is_data);
extern u64 get_mfn(struct domain *d, u64 gpfn);
/*
* Force to delete a found entry no matter TR or foreign map for TLB.
@@ -344,13 +347,8 @@
* NOTES:
*
*/
-extern void thash_purge_entries(thash_cb_t *hcb,
- thash_data_t *in, search_section_t p_sect);
-extern void thash_purge_entries_ex(thash_cb_t *hcb,
- u64 rid, u64 va, u64 sz,
- search_section_t p_sect,
- CACHE_LINE_TYPE cl);
-extern void thash_purge_and_insert(thash_cb_t *hcb, thash_data_t *in, u64 va);
+extern void thash_purge_entries(thash_cb_t *hcb, u64 va, u64 ps);
+extern void thash_purge_and_insert(thash_cb_t *hcb, u64 pte, u64 itir, u64
ifa);
/*
* Purge all TCs or VHPT entries including those in Hash table.
@@ -363,10 +361,7 @@
* covering this address rid:va.
*
*/
-extern thash_data_t *vtlb_lookup(thash_cb_t *hcb,
- thash_data_t *in);
-extern thash_data_t *vtlb_lookup_ex(thash_cb_t *hcb,
- u64 rid, u64 va,CACHE_LINE_TYPE cl);
+extern thash_data_t *vtlb_lookup(thash_cb_t *hcb,u64 va,int is_data);
extern int thash_lock_tc(thash_cb_t *hcb, u64 va, u64 size, int rid, char cl,
int lock);
@@ -381,6 +376,18 @@
extern thash_data_t * vsa_thash(PTA vpta, u64 va, u64 vrr, u64 *tag);
extern thash_data_t * vhpt_lookup(u64 va);
extern void machine_tlb_purge(u64 va, u64 ps);
+extern int fetch_code(struct vcpu *vcpu, u64 gip, u64 *code1, u64 *code2);
+extern void emulate_io_inst(struct vcpu *vcpu, u64 padr, u64 ma);
+extern int vhpt_enabled(struct vcpu *vcpu, uint64_t vadr, vhpt_ref_t ref);
+
+static inline void vmx_vcpu_set_tr (thash_data_t *trp, u64 pte, u64 itir, u64
va, u64 rid)
+{
+ trp->page_flags = pte;
+ trp->itir = itir;
+ trp->vadr = va;
+ trp->rid = rid;
+}
+
//#define VTLB_DEBUG
#ifdef VTLB_DEBUG
diff -r 516cf6553011 -r cf89e8f0831b xen/include/asm-ia64/vmx_phy_mode.h
--- a/xen/include/asm-ia64/vmx_phy_mode.h Mon Mar 20 08:56:46 2006
+++ b/xen/include/asm-ia64/vmx_phy_mode.h Mon Mar 20 08:56:54 2006
@@ -75,11 +75,11 @@
#define PHY_PAGE_UC (_PAGE_A|_PAGE_D|_PAGE_P|_PAGE_MA_UC|_PAGE_AR_RWX)
#define PHY_PAGE_WB (_PAGE_A|_PAGE_D|_PAGE_P|_PAGE_MA_WB|_PAGE_AR_RWX)
-#ifdef PHY_16M /* 16M: large granule for test*/
-#define EMUL_PHY_PAGE_SHIFT 24
-#else /* 4K: emulated physical page granule */
-#define EMUL_PHY_PAGE_SHIFT 12
-#endif
+//#ifdef PHY_16M /* 16M: large granule for test*/
+//#define EMUL_PHY_PAGE_SHIFT 24
+//#else /* 4K: emulated physical page granule */
+//#define EMUL_PHY_PAGE_SHIFT 12
+//#endif
#define IA64_RSC_MODE 0x0000000000000003
#define XEN_RR7_RID (0xf00010)
#define GUEST_IN_PHY 0x1
@@ -96,8 +96,7 @@
extern void recover_if_physical_mode(VCPU *vcpu);
extern void vmx_init_all_rr(VCPU *vcpu);
extern void vmx_load_all_rr(VCPU *vcpu);
-extern void physical_itlb_miss(VCPU *vcpu, u64 vadr);
-extern void physical_dtlb_miss(VCPU *vcpu, u64 vadr);
+extern void physical_tlb_miss(VCPU *vcpu, u64 vadr, u64 vec);
/*
* No sanity check here, since all psr changes have been
* checked in switch_mm_mode().
diff -r 516cf6553011 -r cf89e8f0831b xen/include/asm-ia64/vmx_platform.h
--- a/xen/include/asm-ia64/vmx_platform.h Mon Mar 20 08:56:46 2006
+++ b/xen/include/asm-ia64/vmx_platform.h Mon Mar 20 08:56:54 2006
@@ -22,7 +22,6 @@
#include <public/xen.h>
#include <public/arch-ia64.h>
#include <asm/hvm/vioapic.h>
-
struct mmio_list;
typedef struct virtual_platform_def {
unsigned long shared_page_va;
@@ -51,9 +50,8 @@
} vlapic_t;
extern uint64_t dummy_tmr[];
-#define VCPU(_v,_x) _v->arch.privregs->_x
-#define VLAPIC_ID(l) (uint16_t)(VCPU((l)->vcpu, lid) >> 16)
-#define VLAPIC_IRR(l) VCPU((l)->vcpu, irr[0])
+#define VLAPIC_ID(l) (uint16_t)(((l)->vcpu->arch.privregs->lid) >> 16)
+#define VLAPIC_IRR(l) ((l)->vcpu->arch.privregs->irr[0])
struct vlapic* apic_round_robin(struct domain *d, uint8_t dest_mode, uint8_t
vector, uint32_t bitmap);
extern int vmx_vcpu_pend_interrupt(struct vcpu *vcpu, uint8_t vector);
static inline int vlapic_set_irq(struct vlapic *t, uint8_t vec, uint8_t trig)
diff -r 516cf6553011 -r cf89e8f0831b xen/include/asm-ia64/vmx_vcpu.h
--- a/xen/include/asm-ia64/vmx_vcpu.h Mon Mar 20 08:56:46 2006
+++ b/xen/include/asm-ia64/vmx_vcpu.h Mon Mar 20 08:56:54 2006
@@ -66,17 +66,13 @@
extern IA64FAULT vmx_vcpu_cover(VCPU *vcpu);
extern thash_cb_t *vmx_vcpu_get_vtlb(VCPU *vcpu);
extern thash_cb_t *vmx_vcpu_get_vhpt(VCPU *vcpu);
-extern ia64_rr vmx_vcpu_rr(VCPU *vcpu,UINT64 vadr);
extern IA64FAULT vmx_vcpu_set_rr(VCPU *vcpu, UINT64 reg, UINT64 val);
-#if 0
-extern IA64FAULT vmx_vcpu_get_rr(VCPU *vcpu, UINT64 reg, UINT64 *pval);
-#endif
extern IA64FAULT vmx_vcpu_get_pkr(VCPU *vcpu, UINT64 reg, UINT64 *pval);
IA64FAULT vmx_vcpu_set_pkr(VCPU *vcpu, UINT64 reg, UINT64 val);
extern IA64FAULT vmx_vcpu_itc_i(VCPU *vcpu, UINT64 pte, UINT64 itir, UINT64
ifa);
extern IA64FAULT vmx_vcpu_itc_d(VCPU *vcpu, UINT64 pte, UINT64 itir, UINT64
ifa);
-extern IA64FAULT vmx_vcpu_itr_i(VCPU *vcpu, UINT64 pte, UINT64 itir, UINT64
ifa, UINT64 idx);
-extern IA64FAULT vmx_vcpu_itr_d(VCPU *vcpu, UINT64 pte, UINT64 itir, UINT64
ifa, UINT64 idx);
+extern IA64FAULT vmx_vcpu_itr_i(VCPU *vcpu, UINT64 slot, UINT64 pte, UINT64
itir, UINT64 ifa);
+extern IA64FAULT vmx_vcpu_itr_d(VCPU *vcpu, UINT64 slot, UINT64 pte, UINT64
itir, UINT64 ifa);
extern IA64FAULT vmx_vcpu_ptr_d(VCPU *vcpu,UINT64 vadr,UINT64 ps);
extern IA64FAULT vmx_vcpu_ptr_i(VCPU *vcpu,UINT64 vadr,UINT64 ps);
extern IA64FAULT vmx_vcpu_ptc_l(VCPU *vcpu, UINT64 vadr, UINT64 ps);
@@ -102,10 +98,11 @@
extern void vtm_init(VCPU *vcpu);
extern uint64_t vtm_get_itc(VCPU *vcpu);
extern void vtm_set_itc(VCPU *vcpu, uint64_t new_itc);
-extern void vtm_set_itv(VCPU *vcpu);
+extern void vtm_set_itv(VCPU *vcpu, uint64_t val);
+extern void vtm_set_itm(VCPU *vcpu, uint64_t val);
extern void vtm_interruption_update(VCPU *vcpu, vtime_t* vtm);
-extern void vtm_domain_out(VCPU *vcpu);
-extern void vtm_domain_in(VCPU *vcpu);
+//extern void vtm_domain_out(VCPU *vcpu);
+//extern void vtm_domain_in(VCPU *vcpu);
extern void vlsapic_reset(VCPU *vcpu);
extern int vmx_check_pending_irq(VCPU *vcpu);
extern void guest_write_eoi(VCPU *vcpu);
@@ -255,10 +252,7 @@
IA64FAULT
vmx_vcpu_set_itm(VCPU *vcpu, u64 val)
{
- vtime_t *vtm;
- vtm=&(vcpu->arch.arch_vmx.vtm);
- VCPU(vcpu,itm)=val;
- vtm_interruption_update(vcpu, vtm);
+ vtm_set_itm(vcpu, val);
return IA64_NO_FAULT;
}
static inline
@@ -299,8 +293,7 @@
vmx_vcpu_set_itv(VCPU *vcpu, u64 val)
{
- VCPU(vcpu,itv)=val;
- vtm_set_itv(vcpu);
+ vtm_set_itv(vcpu, val);
return IA64_NO_FAULT;
}
static inline
@@ -350,12 +343,14 @@
*val = vtm_get_itc(vcpu);
return IA64_NO_FAULT;
}
+/*
static inline
IA64FAULT vmx_vcpu_get_rr(VCPU *vcpu, UINT64 reg, UINT64 *pval)
{
*pval = VMX(vcpu,vrr[reg>>61]);
return (IA64_NO_FAULT);
}
+ */
/**************************************************************************
VCPU debug breakpoint register access routines
**************************************************************************/
diff -r 516cf6553011 -r cf89e8f0831b xen/include/asm-ia64/vtm.h
--- a/xen/include/asm-ia64/vtm.h Mon Mar 20 08:56:46 2006
+++ b/xen/include/asm-ia64/vtm.h Mon Mar 20 08:56:54 2006
@@ -63,5 +63,4 @@
#define VTM_FIRED(vtm) ((vtm)->triggered)
-extern void vtm_init();
#endif /* _STATS_H_ */
diff -r 516cf6553011 -r cf89e8f0831b xen/include/asm-x86/domain.h
--- a/xen/include/asm-x86/domain.h Mon Mar 20 08:56:46 2006
+++ b/xen/include/asm-x86/domain.h Mon Mar 20 08:56:54 2006
@@ -124,6 +124,9 @@
void (*schedule_tail) (struct vcpu *);
+ void (*ctxt_switch_from) (struct vcpu *);
+ void (*ctxt_switch_to) (struct vcpu *);
+
/* Bounce information for propagating an exception to guest OS. */
struct trap_bounce trap_bounce;
diff -r 516cf6553011 -r cf89e8f0831b xen/include/asm-x86/hvm/domain.h
--- a/xen/include/asm-x86/hvm/domain.h Mon Mar 20 08:56:46 2006
+++ b/xen/include/asm-x86/hvm/domain.h Mon Mar 20 08:56:54 2006
@@ -37,6 +37,7 @@
unsigned int pae_enabled;
struct hvm_virpit vpit;
+ u64 guest_time;
struct hvm_virpic vpic;
struct hvm_vioapic vioapic;
struct hvm_io_handler io_handler;
diff -r 516cf6553011 -r cf89e8f0831b xen/include/asm-x86/hvm/hvm.h
--- a/xen/include/asm-x86/hvm/hvm.h Mon Mar 20 08:56:46 2006
+++ b/xen/include/asm-x86/hvm/hvm.h Mon Mar 20 08:56:54 2006
@@ -41,18 +41,11 @@
/*
* Store and load guest state:
* 1) load/store guest register state,
- * 2) load/store segment state (x86_64 only),
- * 3) load/store msr register state (x86_64 only),
- * 4) store guest control register state (used for panic dumps),
- * 5) modify guest state (e.g., set debug flags).
+ * 2) store guest control register state (used for panic dumps),
+ * 3) modify guest state (e.g., set debug flags).
*/
void (*store_cpu_guest_regs)(struct vcpu *v, struct cpu_user_regs *r);
void (*load_cpu_guest_regs)(struct vcpu *v, struct cpu_user_regs *r);
-#ifdef __x86_64__
- void (*save_segments)(struct vcpu *v);
- void (*load_msrs)(void);
- void (*restore_msrs)(struct vcpu *v);
-#endif
void (*store_cpu_guest_ctrl_regs)(struct vcpu *v, unsigned long crs[8]);
void (*modify_guest_state)(struct vcpu *v);
@@ -111,33 +104,6 @@
hvm_funcs.load_cpu_guest_regs(v, r);
}
-#ifdef __x86_64__
-static inline void
-hvm_save_segments(struct vcpu *v)
-{
- if (hvm_funcs.save_segments)
- hvm_funcs.save_segments(v);
-}
-
-static inline void
-hvm_load_msrs(void)
-{
- if (hvm_funcs.load_msrs)
- hvm_funcs.load_msrs();
-}
-
-static inline void
-hvm_restore_msrs(struct vcpu *v)
-{
- if (hvm_funcs.restore_msrs)
- hvm_funcs.restore_msrs(v);
-}
-#else
-#define hvm_save_segments(v) ((void)0)
-#define hvm_load_msrs(v) ((void)0)
-#define hvm_restore_msrs(v) ((void)0)
-#endif /* __x86_64__ */
-
static inline void
hvm_store_cpu_guest_ctrl_regs(struct vcpu *v, unsigned long crs[8])
{
diff -r 516cf6553011 -r cf89e8f0831b xen/include/asm-x86/hvm/vmx/vmcs.h
--- a/xen/include/asm-x86/hvm/vmx/vmcs.h Mon Mar 20 08:56:46 2006
+++ b/xen/include/asm-x86/hvm/vmx/vmcs.h Mon Mar 20 08:56:54 2006
@@ -77,7 +77,6 @@
unsigned long cpu_based_exec_control;
struct vmx_msr_state msr_content;
void *io_bitmap_a, *io_bitmap_b;
- u64 tsc_offset;
struct timer hlt_timer; /* hlt ins emulation wakeup timer */
};
diff -r 516cf6553011 -r cf89e8f0831b xen/include/asm-x86/hvm/vmx/vmx.h
--- a/xen/include/asm-x86/hvm/vmx/vmx.h Mon Mar 20 08:56:46 2006
+++ b/xen/include/asm-x86/hvm/vmx/vmx.h Mon Mar 20 08:56:54 2006
@@ -30,10 +30,11 @@
extern void vmx_asm_do_resume(void);
extern void vmx_asm_do_launch(void);
extern void vmx_intr_assist(void);
-extern void vmx_set_tsc_shift(struct vcpu *, struct hvm_virpit *);
extern void vmx_migrate_timers(struct vcpu *v);
extern void arch_vmx_do_launch(struct vcpu *);
extern void arch_vmx_do_resume(struct vcpu *);
+extern void set_guest_time(struct vcpu *v, u64 gtime);
+extern u64 get_guest_time(struct vcpu *v);
extern unsigned int cpu_rev;
diff -r 516cf6553011 -r cf89e8f0831b xen/include/asm-x86/hvm/vpit.h
--- a/xen/include/asm-x86/hvm/vpit.h Mon Mar 20 08:56:46 2006
+++ b/xen/include/asm-x86/hvm/vpit.h Mon Mar 20 08:56:54 2006
@@ -38,7 +38,6 @@
struct hvm_virpit {
/* for simulation of counter 0 in mode 2 */
u64 period_cycles; /* pit frequency in cpu cycles */
- u64 shift; /* save the value of offset - drift */
s_time_t inject_point; /* the time inject virt intr */
s_time_t scheduled; /* scheduled timer interrupt */
struct timer pit_timer; /* periodic timer for mode 2*/
@@ -46,6 +45,8 @@
unsigned int pending_intr_nr; /* the couner for pending timer interrupts */
u32 period; /* pit frequency in ns */
int first_injected; /* flag to prevent shadow window */
+ s64 cache_tsc_offset; /* cache of VMCS TSC_OFFSET offset */
+ u64 last_pit_gtime; /* guest time when last pit is injected */
/* virtual PIT state for handle related I/O */
int read_state;
diff -r 516cf6553011 -r cf89e8f0831b
patches/linux-2.6.16-rc6/i386-mach-io-check-nmi.patch
--- /dev/null Mon Mar 20 08:56:46 2006
+++ b/patches/linux-2.6.16-rc6/i386-mach-io-check-nmi.patch Mon Mar 20
08:56:54 2006
@@ -0,0 +1,45 @@
+diff -pruN ../pristine-linux-2.6.16-rc6/arch/i386/kernel/traps.c
./arch/i386/kernel/traps.c
+--- ../pristine-linux-2.6.16-rc6/arch/i386/kernel/traps.c 2006-03-17
22:59:01.000000000 +0000
++++ ./arch/i386/kernel/traps.c 2006-03-17 23:04:16.000000000 +0000
+@@ -567,18 +567,11 @@ static void mem_parity_error(unsigned ch
+
+ static void io_check_error(unsigned char reason, struct pt_regs * regs)
+ {
+- unsigned long i;
+-
+ printk(KERN_EMERG "NMI: IOCK error (debug interrupt?)\n");
+ show_registers(regs);
+
+ /* Re-enable the IOCK line, wait for a few seconds */
+- reason = (reason & 0xf) | 8;
+- outb(reason, 0x61);
+- i = 2000;
+- while (--i) udelay(1000);
+- reason &= ~8;
+- outb(reason, 0x61);
++ clear_io_check_error(reason);
+ }
+
+ static void unknown_nmi_error(unsigned char reason, struct pt_regs * regs)
+diff -pruN
../pristine-linux-2.6.16-rc6/include/asm-i386/mach-default/mach_traps.h
./include/asm-i386/mach-default/mach_traps.h
+--- ../pristine-linux-2.6.16-rc6/include/asm-i386/mach-default/mach_traps.h
2006-01-03 03:21:10.000000000 +0000
++++ ./include/asm-i386/mach-default/mach_traps.h 2006-03-17
23:04:16.000000000 +0000
+@@ -15,6 +15,18 @@ static inline void clear_mem_error(unsig
+ outb(reason, 0x61);
+ }
+
++static inline void clear_io_check_error(unsigned char reason)
++{
++ unsigned long i;
++
++ reason = (reason & 0xf) | 8;
++ outb(reason, 0x61);
++ i = 2000;
++ while (--i) udelay(1000);
++ reason &= ~8;
++ outb(reason, 0x61);
++}
++
+ static inline unsigned char get_nmi_reason(void)
+ {
+ return inb(0x61);
diff -r 516cf6553011 -r cf89e8f0831b patches/linux-2.6.16-rc6/net-csum.patch
--- /dev/null Mon Mar 20 08:56:46 2006
+++ b/patches/linux-2.6.16-rc6/net-csum.patch Mon Mar 20 08:56:54 2006
@@ -0,0 +1,41 @@
+diff -pruN ../pristine-linux-2.6.16-rc6/net/ipv4/netfilter/ip_nat_proto_tcp.c
./net/ipv4/netfilter/ip_nat_proto_tcp.c
+--- ../pristine-linux-2.6.16-rc6/net/ipv4/netfilter/ip_nat_proto_tcp.c
2006-03-17 22:59:16.000000000 +0000
++++ ./net/ipv4/netfilter/ip_nat_proto_tcp.c 2006-03-17 23:04:19.000000000
+0000
+@@ -129,10 +129,14 @@ tcp_manip_pkt(struct sk_buff **pskb,
+ if (hdrsize < sizeof(*hdr))
+ return 1;
+
+- hdr->check = ip_nat_cheat_check(~oldip, newip,
++ if ((*pskb)->proto_csum_blank) {
++ hdr->check = ip_nat_cheat_check(oldip, ~newip, hdr->check);
++ } else {
++ hdr->check = ip_nat_cheat_check(~oldip, newip,
+ ip_nat_cheat_check(oldport ^ 0xFFFF,
+ newport,
+ hdr->check));
++ }
+ return 1;
+ }
+
+diff -pruN ../pristine-linux-2.6.16-rc6/net/ipv4/netfilter/ip_nat_proto_udp.c
./net/ipv4/netfilter/ip_nat_proto_udp.c
+--- ../pristine-linux-2.6.16-rc6/net/ipv4/netfilter/ip_nat_proto_udp.c
2006-03-17 22:59:16.000000000 +0000
++++ ./net/ipv4/netfilter/ip_nat_proto_udp.c 2006-03-17 23:04:19.000000000
+0000
+@@ -113,11 +113,16 @@ udp_manip_pkt(struct sk_buff **pskb,
+ newport = tuple->dst.u.udp.port;
+ portptr = &hdr->dest;
+ }
+- if (hdr->check) /* 0 is a special case meaning no checksum */
+- hdr->check = ip_nat_cheat_check(~oldip, newip,
++ if (hdr->check) { /* 0 is a special case meaning no checksum */
++ if ((*pskb)->proto_csum_blank) {
++ hdr->check = ip_nat_cheat_check(oldip, ~newip,
hdr->check);
++ } else {
++ hdr->check = ip_nat_cheat_check(~oldip, newip,
+ ip_nat_cheat_check(*portptr ^ 0xFFFF,
+ newport,
+ hdr->check));
++ }
++ }
+ *portptr = newport;
+ return 1;
+ }
diff -r 516cf6553011 -r cf89e8f0831b patches/linux-2.6.16-rc6/pmd-shared.patch
--- /dev/null Mon Mar 20 08:56:46 2006
+++ b/patches/linux-2.6.16-rc6/pmd-shared.patch Mon Mar 20 08:56:54 2006
@@ -0,0 +1,111 @@
+diff -pruN ../pristine-linux-2.6.16-rc6/arch/i386/mm/pageattr.c
./arch/i386/mm/pageattr.c
+--- ../pristine-linux-2.6.16-rc6/arch/i386/mm/pageattr.c 2006-03-17
22:59:01.000000000 +0000
++++ ./arch/i386/mm/pageattr.c 2006-03-17 23:04:21.000000000 +0000
+@@ -78,7 +78,7 @@ static void set_pmd_pte(pte_t *kpte, uns
+ unsigned long flags;
+
+ set_pte_atomic(kpte, pte); /* change init_mm */
+- if (PTRS_PER_PMD > 1)
++ if (HAVE_SHARED_KERNEL_PMD)
+ return;
+
+ spin_lock_irqsave(&pgd_lock, flags);
+diff -pruN ../pristine-linux-2.6.16-rc6/arch/i386/mm/pgtable.c
./arch/i386/mm/pgtable.c
+--- ../pristine-linux-2.6.16-rc6/arch/i386/mm/pgtable.c 2006-01-03
03:21:10.000000000 +0000
++++ ./arch/i386/mm/pgtable.c 2006-03-17 23:04:21.000000000 +0000
+@@ -215,9 +215,10 @@ void pgd_ctor(void *pgd, kmem_cache_t *c
+ spin_lock_irqsave(&pgd_lock, flags);
+ }
+
+- clone_pgd_range((pgd_t *)pgd + USER_PTRS_PER_PGD,
+- swapper_pg_dir + USER_PTRS_PER_PGD,
+- KERNEL_PGD_PTRS);
++ if (PTRS_PER_PMD == 1 || HAVE_SHARED_KERNEL_PMD)
++ clone_pgd_range((pgd_t *)pgd + USER_PTRS_PER_PGD,
++ swapper_pg_dir + USER_PTRS_PER_PGD,
++ KERNEL_PGD_PTRS);
+ if (PTRS_PER_PMD > 1)
+ return;
+
+@@ -249,6 +250,30 @@ pgd_t *pgd_alloc(struct mm_struct *mm)
+ goto out_oom;
+ set_pgd(&pgd[i], __pgd(1 + __pa(pmd)));
+ }
++
++ if (!HAVE_SHARED_KERNEL_PMD) {
++ unsigned long flags;
++
++ for (i = USER_PTRS_PER_PGD; i < PTRS_PER_PGD; i++) {
++ pmd_t *pmd = kmem_cache_alloc(pmd_cache, GFP_KERNEL);
++ if (!pmd)
++ goto out_oom;
++ set_pgd(&pgd[USER_PTRS_PER_PGD], __pgd(1 + __pa(pmd)));
++ }
++
++ spin_lock_irqsave(&pgd_lock, flags);
++ for (i = USER_PTRS_PER_PGD; i < PTRS_PER_PGD; i++) {
++ unsigned long v = (unsigned long)i << PGDIR_SHIFT;
++ pgd_t *kpgd = pgd_offset_k(v);
++ pud_t *kpud = pud_offset(kpgd, v);
++ pmd_t *kpmd = pmd_offset(kpud, v);
++ pmd_t *pmd = (void *)__va(pgd_val(pgd[i])-1);
++ memcpy(pmd, kpmd, PAGE_SIZE);
++ }
++ pgd_list_add(pgd);
++ spin_unlock_irqrestore(&pgd_lock, flags);
++ }
++
+ return pgd;
+
+ out_oom:
+@@ -263,9 +288,23 @@ void pgd_free(pgd_t *pgd)
+ int i;
+
+ /* in the PAE case user pgd entries are overwritten before usage */
+- if (PTRS_PER_PMD > 1)
+- for (i = 0; i < USER_PTRS_PER_PGD; ++i)
+- kmem_cache_free(pmd_cache, (void
*)__va(pgd_val(pgd[i])-1));
++ if (PTRS_PER_PMD > 1) {
++ for (i = 0; i < USER_PTRS_PER_PGD; ++i) {
++ pmd_t *pmd = (void *)__va(pgd_val(pgd[i])-1);
++ kmem_cache_free(pmd_cache, pmd);
++ }
++ if (!HAVE_SHARED_KERNEL_PMD) {
++ unsigned long flags;
++ spin_lock_irqsave(&pgd_lock, flags);
++ pgd_list_del(pgd);
++ spin_unlock_irqrestore(&pgd_lock, flags);
++ for (i = USER_PTRS_PER_PGD; i < PTRS_PER_PGD; i++) {
++ pmd_t *pmd = (void *)__va(pgd_val(pgd[i])-1);
++ memset(pmd, 0, PTRS_PER_PMD*sizeof(pmd_t));
++ kmem_cache_free(pmd_cache, pmd);
++ }
++ }
++ }
+ /* in the non-PAE case, free_pgtables() clears user pgd entries */
+ kmem_cache_free(pgd_cache, pgd);
+ }
+diff -pruN ../pristine-linux-2.6.16-rc6/include/asm-i386/pgtable-2level-defs.h
./include/asm-i386/pgtable-2level-defs.h
+--- ../pristine-linux-2.6.16-rc6/include/asm-i386/pgtable-2level-defs.h
2006-01-03 03:21:10.000000000 +0000
++++ ./include/asm-i386/pgtable-2level-defs.h 2006-03-17 23:04:21.000000000
+0000
+@@ -1,6 +1,8 @@
+ #ifndef _I386_PGTABLE_2LEVEL_DEFS_H
+ #define _I386_PGTABLE_2LEVEL_DEFS_H
+
++#define HAVE_SHARED_KERNEL_PMD 0
++
+ /*
+ * traditional i386 two-level paging structure:
+ */
+diff -pruN ../pristine-linux-2.6.16-rc6/include/asm-i386/pgtable-3level-defs.h
./include/asm-i386/pgtable-3level-defs.h
+--- ../pristine-linux-2.6.16-rc6/include/asm-i386/pgtable-3level-defs.h
2006-01-03 03:21:10.000000000 +0000
++++ ./include/asm-i386/pgtable-3level-defs.h 2006-03-17 23:04:21.000000000
+0000
+@@ -1,6 +1,8 @@
+ #ifndef _I386_PGTABLE_3LEVEL_DEFS_H
+ #define _I386_PGTABLE_3LEVEL_DEFS_H
+
++#define HAVE_SHARED_KERNEL_PMD 1
++
+ /*
+ * PGDIR_SHIFT determines what a top-level page table entry can map
+ */
diff -r 516cf6553011 -r cf89e8f0831b patches/linux-2.6.16-rc6/smp-alts.patch
--- /dev/null Mon Mar 20 08:56:46 2006
+++ b/patches/linux-2.6.16-rc6/smp-alts.patch Mon Mar 20 08:56:54 2006
@@ -0,0 +1,591 @@
+diff -pruN ../pristine-linux-2.6.16-rc6/arch/i386/Kconfig ./arch/i386/Kconfig
+--- ../pristine-linux-2.6.16-rc6/arch/i386/Kconfig 2006-03-17
22:59:01.000000000 +0000
++++ ./arch/i386/Kconfig 2006-03-17 23:04:23.000000000 +0000
+@@ -202,6 +202,19 @@ config SMP
+
+ If you don't know what to do here, say N.
+
++config SMP_ALTERNATIVES
++ bool "SMP alternatives support (EXPERIMENTAL)"
++ depends on SMP && EXPERIMENTAL
++ help
++ Try to reduce the overhead of running an SMP kernel on a uniprocessor
++ host slightly by replacing certain key instruction sequences
++ according to whether we currently have more than one CPU available.
++ This should provide a noticeable boost to performance when
++ running SMP kernels on UP machines, and have negligible impact
++ when running on an true SMP host.
++
++ If unsure, say N.
++
+ config NR_CPUS
+ int "Maximum number of CPUs (2-255)"
+ range 2 255
+diff -pruN ../pristine-linux-2.6.16-rc6/arch/i386/kernel/Makefile
./arch/i386/kernel/Makefile
+--- ../pristine-linux-2.6.16-rc6/arch/i386/kernel/Makefile 2006-03-17
22:59:01.000000000 +0000
++++ ./arch/i386/kernel/Makefile 2006-03-17 23:04:23.000000000 +0000
+@@ -37,6 +37,7 @@ obj-$(CONFIG_EFI) += efi.o efi_stub.o
+ obj-$(CONFIG_DOUBLEFAULT) += doublefault.o
+ obj-$(CONFIG_VM86) += vm86.o
+ obj-$(CONFIG_EARLY_PRINTK) += early_printk.o
++obj-$(CONFIG_SMP_ALTERNATIVES) += smpalts.o
+
+ EXTRA_AFLAGS := -traditional
+
+diff -pruN ../pristine-linux-2.6.16-rc6/arch/i386/kernel/smpalts.c
./arch/i386/kernel/smpalts.c
+--- ../pristine-linux-2.6.16-rc6/arch/i386/kernel/smpalts.c 1970-01-01
01:00:00.000000000 +0100
++++ ./arch/i386/kernel/smpalts.c 2006-03-17 23:04:23.000000000 +0000
+@@ -0,0 +1,85 @@
++#include <linux/kernel.h>
++#include <asm/system.h>
++#include <asm/smp_alt.h>
++#include <asm/processor.h>
++#include <asm/string.h>
++
++struct smp_replacement_record {
++ unsigned char targ_size;
++ unsigned char smp1_size;
++ unsigned char smp2_size;
++ unsigned char up_size;
++ unsigned char feature;
++ unsigned char data[0];
++};
++
++struct smp_alternative_record {
++ void *targ_start;
++ struct smp_replacement_record *repl;
++};
++
++extern struct smp_alternative_record __start_smp_alternatives_table,
++ __stop_smp_alternatives_table;
++extern unsigned long __init_begin, __init_end;
++
++void prepare_for_smp(void)
++{
++ struct smp_alternative_record *r;
++ printk(KERN_INFO "Enabling SMP...\n");
++ for (r = &__start_smp_alternatives_table;
++ r != &__stop_smp_alternatives_table;
++ r++) {
++ BUG_ON(r->repl->targ_size < r->repl->smp1_size);
++ BUG_ON(r->repl->targ_size < r->repl->smp2_size);
++ BUG_ON(r->repl->targ_size < r->repl->up_size);
++ if (system_state == SYSTEM_RUNNING &&
++ r->targ_start >= (void *)&__init_begin &&
++ r->targ_start < (void *)&__init_end)
++ continue;
++ if (r->repl->feature != (unsigned char)-1 &&
++ boot_cpu_has(r->repl->feature)) {
++ memcpy(r->targ_start,
++ r->repl->data + r->repl->smp1_size,
++ r->repl->smp2_size);
++ memset(r->targ_start + r->repl->smp2_size,
++ 0x90,
++ r->repl->targ_size - r->repl->smp2_size);
++ } else {
++ memcpy(r->targ_start,
++ r->repl->data,
++ r->repl->smp1_size);
++ memset(r->targ_start + r->repl->smp1_size,
++ 0x90,
++ r->repl->targ_size - r->repl->smp1_size);
++ }
++ }
++ /* Paranoia */
++ asm volatile ("jmp 1f\n1:");
++ mb();
++}
++
++void unprepare_for_smp(void)
++{
++ struct smp_alternative_record *r;
++ printk(KERN_INFO "Disabling SMP...\n");
++ for (r = &__start_smp_alternatives_table;
++ r != &__stop_smp_alternatives_table;
++ r++) {
++ BUG_ON(r->repl->targ_size < r->repl->smp1_size);
++ BUG_ON(r->repl->targ_size < r->repl->smp2_size);
++ BUG_ON(r->repl->targ_size < r->repl->up_size);
++ if (system_state == SYSTEM_RUNNING &&
++ r->targ_start >= (void *)&__init_begin &&
++ r->targ_start < (void *)&__init_end)
++ continue;
++ memcpy(r->targ_start,
++ r->repl->data + r->repl->smp1_size + r->repl->smp2_size,
++ r->repl->up_size);
++ memset(r->targ_start + r->repl->up_size,
++ 0x90,
++ r->repl->targ_size - r->repl->up_size);
++ }
++ /* Paranoia */
++ asm volatile ("jmp 1f\n1:");
++ mb();
++}
+diff -pruN ../pristine-linux-2.6.16-rc6/arch/i386/kernel/smpboot.c
./arch/i386/kernel/smpboot.c
+--- ../pristine-linux-2.6.16-rc6/arch/i386/kernel/smpboot.c 2006-03-17
22:59:01.000000000 +0000
++++ ./arch/i386/kernel/smpboot.c 2006-03-17 23:04:23.000000000 +0000
+@@ -1208,6 +1208,11 @@ static void __init smp_boot_cpus(unsigne
+ if (max_cpus <= cpucount+1)
+ continue;
+
++#ifdef CONFIG_SMP_ALTERNATIVES
++ if (kicked == 1)
++ prepare_for_smp();
++#endif
++
+ if (((cpu = alloc_cpu_id()) <= 0) || do_boot_cpu(apicid, cpu))
+ printk("CPU #%d not responding - cannot use it.\n",
+ apicid);
+@@ -1386,6 +1391,11 @@ int __devinit __cpu_up(unsigned int cpu)
+ return -EIO;
+ }
+
++#ifdef CONFIG_SMP_ALTERNATIVES
++ if (num_online_cpus() == 1)
++ prepare_for_smp();
++#endif
++
+ local_irq_enable();
+ per_cpu(cpu_state, cpu) = CPU_UP_PREPARE;
+ /* Unleash the CPU! */
+diff -pruN ../pristine-linux-2.6.16-rc6/arch/i386/kernel/vmlinux.lds.S
./arch/i386/kernel/vmlinux.lds.S
+--- ../pristine-linux-2.6.16-rc6/arch/i386/kernel/vmlinux.lds.S
2006-01-03 03:21:10.000000000 +0000
++++ ./arch/i386/kernel/vmlinux.lds.S 2006-03-17 23:04:23.000000000 +0000
+@@ -34,6 +34,13 @@ SECTIONS
+ __ex_table : AT(ADDR(__ex_table) - LOAD_OFFSET) { *(__ex_table) }
+ __stop___ex_table = .;
+
++ . = ALIGN(16);
++ __start_smp_alternatives_table = .;
++ __smp_alternatives : { *(__smp_alternatives) }
++ __stop_smp_alternatives_table = .;
++
++ __smp_replacements : { *(__smp_replacements) }
++
+ RODATA
+
+ /* writeable */
+diff -pruN ../pristine-linux-2.6.16-rc6/include/asm-i386/atomic.h
./include/asm-i386/atomic.h
+--- ../pristine-linux-2.6.16-rc6/include/asm-i386/atomic.h 2006-03-17
22:59:05.000000000 +0000
++++ ./include/asm-i386/atomic.h 2006-03-17 23:04:23.000000000 +0000
+@@ -4,18 +4,13 @@
+ #include <linux/config.h>
+ #include <linux/compiler.h>
+ #include <asm/processor.h>
++#include <asm/smp_alt.h>
+
+ /*
+ * Atomic operations that C can't guarantee us. Useful for
+ * resource counting etc..
+ */
+
+-#ifdef CONFIG_SMP
+-#define LOCK "lock ; "
+-#else
+-#define LOCK ""
+-#endif
+-
+ /*
+ * Make sure gcc doesn't try to be clever and move things around
+ * on us. We need to use _exactly_ the address the user gave us,
+diff -pruN ../pristine-linux-2.6.16-rc6/include/asm-i386/bitops.h
./include/asm-i386/bitops.h
+--- ../pristine-linux-2.6.16-rc6/include/asm-i386/bitops.h 2006-03-17
22:59:05.000000000 +0000
++++ ./include/asm-i386/bitops.h 2006-03-17 23:04:23.000000000 +0000
+@@ -7,6 +7,7 @@
+
+ #include <linux/config.h>
+ #include <linux/compiler.h>
++#include <asm/smp_alt.h>
+
+ /*
+ * These have to be done with inline assembly: that way the bit-setting
+@@ -16,12 +17,6 @@
+ * bit 0 is the LSB of addr; bit 32 is the LSB of (addr+1).
+ */
+
+-#ifdef CONFIG_SMP
+-#define LOCK_PREFIX "lock ; "
+-#else
+-#define LOCK_PREFIX ""
+-#endif
+-
+ #define ADDR (*(volatile long *) addr)
+
+ /**
+@@ -41,7 +36,7 @@
+ */
+ static inline void set_bit(int nr, volatile unsigned long * addr)
+ {
+- __asm__ __volatile__( LOCK_PREFIX
++ __asm__ __volatile__( LOCK
+ "btsl %1,%0"
+ :"+m" (ADDR)
+ :"Ir" (nr));
+@@ -76,7 +71,7 @@ static inline void __set_bit(int nr, vol
+ */
+ static inline void clear_bit(int nr, volatile unsigned long * addr)
+ {
+- __asm__ __volatile__( LOCK_PREFIX
++ __asm__ __volatile__( LOCK
+ "btrl %1,%0"
+ :"+m" (ADDR)
+ :"Ir" (nr));
+@@ -121,7 +116,7 @@ static inline void __change_bit(int nr,
+ */
+ static inline void change_bit(int nr, volatile unsigned long * addr)
+ {
+- __asm__ __volatile__( LOCK_PREFIX
++ __asm__ __volatile__( LOCK
+ "btcl %1,%0"
+ :"+m" (ADDR)
+ :"Ir" (nr));
+@@ -140,7 +135,7 @@ static inline int test_and_set_bit(int n
+ {
+ int oldbit;
+
+- __asm__ __volatile__( LOCK_PREFIX
++ __asm__ __volatile__( LOCK
+ "btsl %2,%1\n\tsbbl %0,%0"
+ :"=r" (oldbit),"+m" (ADDR)
+ :"Ir" (nr) : "memory");
+@@ -180,7 +175,7 @@ static inline int test_and_clear_bit(int
+ {
+ int oldbit;
+
+- __asm__ __volatile__( LOCK_PREFIX
++ __asm__ __volatile__( LOCK
+ "btrl %2,%1\n\tsbbl %0,%0"
+ :"=r" (oldbit),"+m" (ADDR)
+ :"Ir" (nr) : "memory");
+@@ -231,7 +226,7 @@ static inline int test_and_change_bit(in
+ {
+ int oldbit;
+
+- __asm__ __volatile__( LOCK_PREFIX
++ __asm__ __volatile__( LOCK
+ "btcl %2,%1\n\tsbbl %0,%0"
+ :"=r" (oldbit),"+m" (ADDR)
+ :"Ir" (nr) : "memory");
+diff -pruN ../pristine-linux-2.6.16-rc6/include/asm-i386/futex.h
./include/asm-i386/futex.h
+--- ../pristine-linux-2.6.16-rc6/include/asm-i386/futex.h 2006-03-17
22:59:05.000000000 +0000
++++ ./include/asm-i386/futex.h 2006-03-17 23:04:23.000000000 +0000
+@@ -28,7 +28,7 @@
+ "1: movl %2, %0\n\
+ movl %0, %3\n" \
+ insn "\n" \
+-"2: " LOCK_PREFIX "cmpxchgl %3, %2\n\
++"2: " LOCK "cmpxchgl %3, %2\n\
+ jnz 1b\n\
+ 3: .section .fixup,\"ax\"\n\
+ 4: mov %5, %1\n\
+@@ -68,7 +68,7 @@ futex_atomic_op_inuser (int encoded_op,
+ #endif
+ switch (op) {
+ case FUTEX_OP_ADD:
+- __futex_atomic_op1(LOCK_PREFIX "xaddl %0, %2", ret,
++ __futex_atomic_op1(LOCK "xaddl %0, %2", ret,
+ oldval, uaddr, oparg);
+ break;
+ case FUTEX_OP_OR:
+diff -pruN ../pristine-linux-2.6.16-rc6/include/asm-i386/rwsem.h
./include/asm-i386/rwsem.h
+--- ../pristine-linux-2.6.16-rc6/include/asm-i386/rwsem.h 2006-01-03
03:21:10.000000000 +0000
++++ ./include/asm-i386/rwsem.h 2006-03-17 23:04:23.000000000 +0000
+@@ -40,6 +40,7 @@
+
+ #include <linux/list.h>
+ #include <linux/spinlock.h>
++#include <asm/smp_alt.h>
+
+ struct rwsem_waiter;
+
+@@ -99,7 +100,7 @@ static inline void __down_read(struct rw
+ {
+ __asm__ __volatile__(
+ "# beginning down_read\n\t"
+-LOCK_PREFIX " incl (%%eax)\n\t" /* adds 0x00000001, returns the old
value */
++LOCK " incl (%%eax)\n\t" /* adds 0x00000001, returns the old
value */
+ " js 2f\n\t" /* jump if we weren't granted the lock */
+ "1:\n\t"
+ LOCK_SECTION_START("")
+@@ -130,7 +131,7 @@ static inline int __down_read_trylock(st
+ " movl %1,%2\n\t"
+ " addl %3,%2\n\t"
+ " jle 2f\n\t"
+-LOCK_PREFIX " cmpxchgl %2,%0\n\t"
++LOCK " cmpxchgl %2,%0\n\t"
+ " jnz 1b\n\t"
+ "2:\n\t"
+ "# ending __down_read_trylock\n\t"
+@@ -150,7 +151,7 @@ static inline void __down_write(struct r
+ tmp = RWSEM_ACTIVE_WRITE_BIAS;
+ __asm__ __volatile__(
+ "# beginning down_write\n\t"
+-LOCK_PREFIX " xadd %%edx,(%%eax)\n\t" /* subtract 0x0000ffff, returns
the old value */
++LOCK " xadd %%edx,(%%eax)\n\t" /* subtract 0x0000ffff, returns
the old value */
+ " testl %%edx,%%edx\n\t" /* was the count 0 before? */
+ " jnz 2f\n\t" /* jump if we weren't granted the lock */
+ "1:\n\t"
+@@ -188,7 +189,7 @@ static inline void __up_read(struct rw_s
+ __s32 tmp = -RWSEM_ACTIVE_READ_BIAS;
+ __asm__ __volatile__(
+ "# beginning __up_read\n\t"
+-LOCK_PREFIX " xadd %%edx,(%%eax)\n\t" /* subtracts 1, returns the old
value */
++LOCK " xadd %%edx,(%%eax)\n\t" /* subtracts 1, returns the old
value */
+ " js 2f\n\t" /* jump if the lock is being waited upon */
+ "1:\n\t"
+ LOCK_SECTION_START("")
+@@ -214,7 +215,7 @@ static inline void __up_write(struct rw_
+ __asm__ __volatile__(
+ "# beginning __up_write\n\t"
+ " movl %2,%%edx\n\t"
+-LOCK_PREFIX " xaddl %%edx,(%%eax)\n\t" /* tries to transition
0xffff0001 -> 0x00000000 */
++LOCK " xaddl %%edx,(%%eax)\n\t" /* tries to transition
0xffff0001 -> 0x00000000 */
+ " jnz 2f\n\t" /* jump if the lock is being waited upon */
+ "1:\n\t"
+ LOCK_SECTION_START("")
+@@ -239,7 +240,7 @@ static inline void __downgrade_write(str
+ {
+ __asm__ __volatile__(
+ "# beginning __downgrade_write\n\t"
+-LOCK_PREFIX " addl %2,(%%eax)\n\t" /* transitions 0xZZZZ0001 ->
0xYYYY0001 */
++LOCK " addl %2,(%%eax)\n\t" /* transitions 0xZZZZ0001 ->
0xYYYY0001 */
+ " js 2f\n\t" /* jump if the lock is being waited upon */
+ "1:\n\t"
+ LOCK_SECTION_START("")
+@@ -263,7 +264,7 @@ LOCK_PREFIX " addl %2,(%%eax)\n\t"
+ static inline void rwsem_atomic_add(int delta, struct rw_semaphore *sem)
+ {
+ __asm__ __volatile__(
+-LOCK_PREFIX "addl %1,%0"
++LOCK "addl %1,%0"
+ : "=m"(sem->count)
+ : "ir"(delta), "m"(sem->count));
+ }
+@@ -276,7 +277,7 @@ static inline int rwsem_atomic_update(in
+ int tmp = delta;
+
+ __asm__ __volatile__(
+-LOCK_PREFIX "xadd %0,(%2)"
++LOCK "xadd %0,(%2)"
+ : "+r"(tmp), "=m"(sem->count)
+ : "r"(sem), "m"(sem->count)
+ : "memory");
+diff -pruN ../pristine-linux-2.6.16-rc6/include/asm-i386/smp_alt.h
./include/asm-i386/smp_alt.h
+--- ../pristine-linux-2.6.16-rc6/include/asm-i386/smp_alt.h 1970-01-01
01:00:00.000000000 +0100
++++ ./include/asm-i386/smp_alt.h 2006-03-17 23:04:23.000000000 +0000
+@@ -0,0 +1,32 @@
++#ifndef __ASM_SMP_ALT_H__
++#define __ASM_SMP_ALT_H__
++
++#include <linux/config.h>
++
++#ifdef CONFIG_SMP
++#if defined(CONFIG_SMP_ALTERNATIVES) && !defined(MODULE)
++#define LOCK \
++ "6677: nop\n" \
++ ".section __smp_alternatives,\"a\"\n" \
++ ".long 6677b\n" \
++ ".long 6678f\n" \
++ ".previous\n" \
++ ".section __smp_replacements,\"a\"\n" \
++ "6678: .byte 1\n" \
++ ".byte 1\n" \
++ ".byte 0\n" \
++ ".byte 1\n" \
++ ".byte -1\n" \
++ "lock\n" \
++ "nop\n" \
++ ".previous\n"
++void prepare_for_smp(void);
++void unprepare_for_smp(void);
++#else
++#define LOCK "lock ; "
++#endif
++#else
++#define LOCK ""
++#endif
++
++#endif /* __ASM_SMP_ALT_H__ */
+diff -pruN ../pristine-linux-2.6.16-rc6/include/asm-i386/spinlock.h
./include/asm-i386/spinlock.h
+--- ../pristine-linux-2.6.16-rc6/include/asm-i386/spinlock.h 2006-01-03
03:21:10.000000000 +0000
++++ ./include/asm-i386/spinlock.h 2006-03-17 23:04:23.000000000 +0000
+@@ -6,6 +6,7 @@
+ #include <asm/page.h>
+ #include <linux/config.h>
+ #include <linux/compiler.h>
++#include <asm/smp_alt.h>
+
+ /*
+ * Your basic SMP spinlocks, allowing only a single CPU anywhere
+@@ -23,7 +24,8 @@
+
+ #define __raw_spin_lock_string \
+ "\n1:\t" \
+- "lock ; decb %0\n\t" \
++ LOCK \
++ "decb %0\n\t" \
+ "jns 3f\n" \
+ "2:\t" \
+ "rep;nop\n\t" \
+@@ -34,7 +36,8 @@
+
+ #define __raw_spin_lock_string_flags \
+ "\n1:\t" \
+- "lock ; decb %0\n\t" \
++ LOCK \
++ "decb %0\n\t" \
+ "jns 4f\n\t" \
+ "2:\t" \
+ "testl $0x200, %1\n\t" \
+@@ -65,10 +68,34 @@ static inline void __raw_spin_lock_flags
+ static inline int __raw_spin_trylock(raw_spinlock_t *lock)
+ {
+ char oldval;
++#ifdef CONFIG_SMP_ALTERNATIVES
+ __asm__ __volatile__(
+- "xchgb %b0,%1"
++ "1:movb %1,%b0\n"
++ "movb $0,%1\n"
++ "2:"
++ ".section __smp_alternatives,\"a\"\n"
++ ".long 1b\n"
++ ".long 3f\n"
++ ".previous\n"
++ ".section __smp_replacements,\"a\"\n"
++ "3: .byte 2b - 1b\n"
++ ".byte 5f-4f\n"
++ ".byte 0\n"
++ ".byte 6f-5f\n"
++ ".byte -1\n"
++ "4: xchgb %b0,%1\n"
++ "5: movb %1,%b0\n"
++ "movb $0,%1\n"
++ "6:\n"
++ ".previous\n"
+ :"=q" (oldval), "=m" (lock->slock)
+ :"0" (0) : "memory");
++#else
++ __asm__ __volatile__(
++ "xchgb %b0,%1\n"
++ :"=q" (oldval), "=m" (lock->slock)
++ :"0" (0) : "memory");
++#endif
+ return oldval > 0;
+ }
+
+@@ -178,12 +205,12 @@ static inline int __raw_write_trylock(ra
+
+ static inline void __raw_read_unlock(raw_rwlock_t *rw)
+ {
+- asm volatile("lock ; incl %0" :"=m" (rw->lock) : : "memory");
++ asm volatile(LOCK "incl %0" :"=m" (rw->lock) : : "memory");
+ }
+
+ static inline void __raw_write_unlock(raw_rwlock_t *rw)
+ {
+- asm volatile("lock ; addl $" RW_LOCK_BIAS_STR ", %0"
++ asm volatile(LOCK "addl $" RW_LOCK_BIAS_STR ", %0"
+ : "=m" (rw->lock) : : "memory");
+ }
+
+diff -pruN ../pristine-linux-2.6.16-rc6/include/asm-i386/system.h
./include/asm-i386/system.h
+--- ../pristine-linux-2.6.16-rc6/include/asm-i386/system.h 2006-03-17
22:59:05.000000000 +0000
++++ ./include/asm-i386/system.h 2006-03-17 23:04:23.000000000 +0000
+@@ -5,7 +5,7 @@
+ #include <linux/kernel.h>
+ #include <asm/segment.h>
+ #include <asm/cpufeature.h>
+-#include <linux/bitops.h> /* for LOCK_PREFIX */
++#include <asm/smp_alt.h>
+
+ #ifdef __KERNEL__
+
+@@ -271,19 +271,19 @@ static inline unsigned long __cmpxchg(vo
+ unsigned long prev;
+ switch (size) {
+ case 1:
+- __asm__ __volatile__(LOCK_PREFIX "cmpxchgb %b1,%2"
++ __asm__ __volatile__(LOCK "cmpxchgb %b1,%2"
+ : "=a"(prev)
+ : "q"(new), "m"(*__xg(ptr)), "0"(old)
+ : "memory");
+ return prev;
+ case 2:
+- __asm__ __volatile__(LOCK_PREFIX "cmpxchgw %w1,%2"
++ __asm__ __volatile__(LOCK "cmpxchgw %w1,%2"
+ : "=a"(prev)
+ : "r"(new), "m"(*__xg(ptr)), "0"(old)
+ : "memory");
+ return prev;
+ case 4:
+- __asm__ __volatile__(LOCK_PREFIX "cmpxchgl %1,%2"
++ __asm__ __volatile__(LOCK "cmpxchgl %1,%2"
+ : "=a"(prev)
+ : "r"(new), "m"(*__xg(ptr)), "0"(old)
+ : "memory");
+@@ -336,7 +336,7 @@ static inline unsigned long long __cmpxc
+ unsigned long long new)
+ {
+ unsigned long long prev;
+- __asm__ __volatile__(LOCK_PREFIX "cmpxchg8b %3"
++ __asm__ __volatile__(LOCK "cmpxchg8b %3"
+ : "=A"(prev)
+ : "b"((unsigned long)new),
+ "c"((unsigned long)(new >> 32)),
+@@ -503,11 +503,55 @@ struct alt_instr {
+ #endif
+
+ #ifdef CONFIG_SMP
++#if defined(CONFIG_SMP_ALTERNATIVES) && !defined(MODULE)
++#define smp_alt_mb(instr) \
++__asm__ __volatile__("6667:\nnop\nnop\nnop\nnop\nnop\nnop\n6668:\n" \
++ ".section __smp_alternatives,\"a\"\n" \
++ ".long 6667b\n" \
++ ".long 6673f\n" \
++ ".previous\n" \
++ ".section __smp_replacements,\"a\"\n" \
++ "6673:.byte 6668b-6667b\n" \
++ ".byte 6670f-6669f\n" \
++ ".byte 6671f-6670f\n" \
++ ".byte 0\n" \
++ ".byte %c0\n" \
++ "6669:lock;addl $0,0(%%esp)\n" \
++ "6670:" instr "\n" \
++ "6671:\n" \
++ ".previous\n" \
++ : \
++ : "i" (X86_FEATURE_XMM2) \
++ : "memory")
++#define smp_rmb() smp_alt_mb("lfence")
++#define smp_mb() smp_alt_mb("mfence")
++#define set_mb(var, value) do { \
++unsigned long __set_mb_temp; \
++__asm__ __volatile__("6667:movl %1, %0\n6668:\n" \
++ ".section __smp_alternatives,\"a\"\n" \
++ ".long 6667b\n" \
++ ".long 6673f\n" \
++ ".previous\n" \
++ ".section __smp_replacements,\"a\"\n" \
++ "6673: .byte 6668b-6667b\n" \
++ ".byte 6670f-6669f\n" \
++ ".byte 0\n" \
++ ".byte 6671f-6670f\n" \
++ ".byte -1\n" \
++ "6669: xchg %1, %0\n" \
++ "6670:movl %1, %0\n" \
++ "6671:\n" \
++ ".previous\n" \
++ : "=m" (var), "=r" (__set_mb_temp) \
++ : "1" (value) \
++ : "memory"); } while (0)
++#else
+ #define smp_mb() mb()
+ #define smp_rmb() rmb()
++#define set_mb(var, value) do { (void) xchg(&var, value); } while (0)
++#endif
+ #define smp_wmb() wmb()
+ #define smp_read_barrier_depends() read_barrier_depends()
+-#define set_mb(var, value) do { (void) xchg(&var, value); } while (0)
+ #else
+ #define smp_mb() barrier()
+ #define smp_rmb() barrier()
diff -r 516cf6553011 -r cf89e8f0831b xen/Post.mk
--- /dev/null Mon Mar 20 08:56:46 2006
+++ b/xen/Post.mk Mon Mar 20 08:56:54 2006
@@ -0,0 +1,18 @@
+
+subdirs-all := $(subdirs-y) $(subdirs-n)
+
+obj-y += $(patsubst %,%/built_in.o,$(subdirs-y))
+
+built_in.o: $(obj-y)
+ $(LD) $(LDFLAGS) -r -o $@ $^
+
+.PHONY: FORCE
+FORCE:
+
+%/built_in.o: FORCE
+ $(MAKE) -C $*
+
+clean:: $(addprefix _clean_, $(subdirs-all)) FORCE
+ rm -f *.o *~ core
+_clean_%/: FORCE
+ $(MAKE) -C $* clean
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/ia64/tools/privop/Makefile
--- /dev/null Mon Mar 20 08:56:46 2006
+++ b/xen/arch/ia64/tools/privop/Makefile Mon Mar 20 08:56:54 2006
@@ -0,0 +1,11 @@
+CC=gcc
+CFLAGS=-O -Wall
+
+all: postat
+
+postat: postat.c pohcalls.o
+
+clean:
+ $(RM) -f *.o postat *.s *~
+
+
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/ia64/tools/privop/pohcalls.S
--- /dev/null Mon Mar 20 08:56:46 2006
+++ b/xen/arch/ia64/tools/privop/pohcalls.S Mon Mar 20 08:56:54 2006
@@ -0,0 +1,30 @@
+ .file "hypercall.S"
+ .pred.safe_across_calls p1-p5,p16-p63
+ .text
+ .align 16
+ .global dump_privop_counts#
+ .proc dump_privop_counts#
+dump_privop_counts:
+ .prologue
+ .body
+ mov r2 = 0xffff
+ ;;
+ break 0x1000
+ ;;
+ br.ret.sptk.many b0
+ ;;
+ .endp dump_privop_counts#
+ .align 16
+ .global zero_privop_counts#
+ .proc zero_privop_counts#
+zero_privop_counts:
+ .prologue
+ .body
+ mov r2 = 0xfffe
+ ;;
+ break 0x1000
+ ;;
+ br.ret.sptk.many b0
+ ;;
+ .endp zero_privop_counts#
+
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/ia64/tools/privop/postat.c
--- /dev/null Mon Mar 20 08:56:46 2006
+++ b/xen/arch/ia64/tools/privop/postat.c Mon Mar 20 08:56:54 2006
@@ -0,0 +1,27 @@
+#include <stdio.h>
+#include <string.h>
+
+extern int dump_privop_counts (char *buf, int len);
+
+extern int zero_privop_counts (char *buf, int len);
+
+int
+main (int argc, char *argv[])
+{
+ static char buf[8192];
+ int res;
+
+ if (argc == 1)
+ res = dump_privop_counts (buf, sizeof (buf));
+ else if (argc == 2 && strcmp (argv[1], "--clear") == 0)
+ res = zero_privop_counts (buf, sizeof (buf));
+ else
+ {
+ printf ("usage: %s [--clear]\n", argv[0]);
+ return 1;
+ }
+ printf ("res=%d\n", res);
+ fputs (buf, stdout);
+
+ return 0;
+}
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/x86/acpi/Makefile
--- /dev/null Mon Mar 20 08:56:46 2006
+++ b/xen/arch/x86/acpi/Makefile Mon Mar 20 08:56:54 2006
@@ -0,0 +1,5 @@
+include $(BASEDIR)/Rules.mk
+
+obj-y += boot.o
+
+include $(BASEDIR)/Post.mk
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/x86/cpu/Makefile
--- /dev/null Mon Mar 20 08:56:46 2006
+++ b/xen/arch/x86/cpu/Makefile Mon Mar 20 08:56:54 2006
@@ -0,0 +1,16 @@
+include $(BASEDIR)/Rules.mk
+
+subdirs-y += mcheck
+subdirs-y += mtrr
+
+obj-y += amd.o
+obj-y += common.o
+obj-y += intel.o
+obj-y += intel_cacheinfo.o
+
+obj-$(x86_32) += centaur.o
+obj-$(x86_32) += cyrix.o
+obj-$(x86_32) += rise.o
+obj-$(x86_32) += transmeta.o
+
+include $(BASEDIR)/Post.mk
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/x86/cpu/mcheck/Makefile
--- /dev/null Mon Mar 20 08:56:46 2006
+++ b/xen/arch/x86/cpu/mcheck/Makefile Mon Mar 20 08:56:54 2006
@@ -0,0 +1,11 @@
+include $(BASEDIR)/Rules.mk
+
+obj-y += k7.o
+obj-y += mce.o
+obj-y += non-fatal.o
+obj-y += p4.o
+obj-y += p5.o
+obj-y += p6.o
+obj-y += winchip.o
+
+include $(BASEDIR)/Post.mk
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/x86/cpu/mtrr/Makefile
--- /dev/null Mon Mar 20 08:56:46 2006
+++ b/xen/arch/x86/cpu/mtrr/Makefile Mon Mar 20 08:56:54 2006
@@ -0,0 +1,10 @@
+include $(BASEDIR)/Rules.mk
+
+obj-y += amd.o
+obj-y += centaur.o
+obj-y += cyrix.o
+obj-y += generic.o
+obj-y += main.o
+obj-y += state.o
+
+include $(BASEDIR)/Post.mk
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/x86/genapic/Makefile
--- /dev/null Mon Mar 20 08:56:46 2006
+++ b/xen/arch/x86/genapic/Makefile Mon Mar 20 08:56:54 2006
@@ -0,0 +1,10 @@
+include $(BASEDIR)/Rules.mk
+
+obj-y += bigsmp.o
+obj-y += default.o
+obj-y += es7000.o
+obj-y += es7000plat.o
+obj-y += probe.o
+obj-y += summit.o
+
+include $(BASEDIR)/Post.mk
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/x86/hvm/Makefile
--- /dev/null Mon Mar 20 08:56:46 2006
+++ b/xen/arch/x86/hvm/Makefile Mon Mar 20 08:56:54 2006
@@ -0,0 +1,14 @@
+include $(BASEDIR)/Rules.mk
+
+subdirs-y += svm
+subdirs-y += vmx
+
+obj-y += hvm.o
+obj-y += i8259.o
+obj-y += intercept.o
+obj-y += io.o
+obj-y += platform.o
+obj-y += vioapic.o
+obj-y += vlapic.o
+
+include $(BASEDIR)/Post.mk
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/x86/hvm/svm/Makefile
--- /dev/null Mon Mar 20 08:56:46 2006
+++ b/xen/arch/x86/hvm/svm/Makefile Mon Mar 20 08:56:54 2006
@@ -0,0 +1,12 @@
+include $(BASEDIR)/Rules.mk
+
+subdirs-$(x86_32) += x86_32
+subdirs-$(x86_64) += x86_64
+
+obj-y += emulate.o
+obj-y += instrlen.o
+obj-y += intr.o
+obj-y += svm.o
+obj-y += vmcb.o
+
+include $(BASEDIR)/Post.mk
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/x86/hvm/svm/x86_32/Makefile
--- /dev/null Mon Mar 20 08:56:46 2006
+++ b/xen/arch/x86/hvm/svm/x86_32/Makefile Mon Mar 20 08:56:54 2006
@@ -0,0 +1,5 @@
+include $(BASEDIR)/Rules.mk
+
+obj-y += exits.o
+
+include $(BASEDIR)/Post.mk
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/x86/hvm/svm/x86_64/Makefile
--- /dev/null Mon Mar 20 08:56:46 2006
+++ b/xen/arch/x86/hvm/svm/x86_64/Makefile Mon Mar 20 08:56:54 2006
@@ -0,0 +1,5 @@
+include $(BASEDIR)/Rules.mk
+
+obj-y += exits.o
+
+include $(BASEDIR)/Post.mk
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/x86/hvm/vmx/Makefile
--- /dev/null Mon Mar 20 08:56:46 2006
+++ b/xen/arch/x86/hvm/vmx/Makefile Mon Mar 20 08:56:54 2006
@@ -0,0 +1,10 @@
+include $(BASEDIR)/Rules.mk
+
+subdirs-$(x86_32) += x86_32
+subdirs-$(x86_64) += x86_64
+
+obj-y += io.o
+obj-y += vmcs.o
+obj-y += vmx.o
+
+include $(BASEDIR)/Post.mk
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/x86/hvm/vmx/x86_32/Makefile
--- /dev/null Mon Mar 20 08:56:46 2006
+++ b/xen/arch/x86/hvm/vmx/x86_32/Makefile Mon Mar 20 08:56:54 2006
@@ -0,0 +1,5 @@
+include $(BASEDIR)/Rules.mk
+
+obj-y += exits.o
+
+include $(BASEDIR)/Post.mk
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/x86/hvm/vmx/x86_64/Makefile
--- /dev/null Mon Mar 20 08:56:46 2006
+++ b/xen/arch/x86/hvm/vmx/x86_64/Makefile Mon Mar 20 08:56:54 2006
@@ -0,0 +1,5 @@
+include $(BASEDIR)/Rules.mk
+
+obj-y += exits.o
+
+include $(BASEDIR)/Post.mk
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/x86/x86_32/Makefile
--- /dev/null Mon Mar 20 08:56:46 2006
+++ b/xen/arch/x86/x86_32/Makefile Mon Mar 20 08:56:54 2006
@@ -0,0 +1,11 @@
+include $(BASEDIR)/Rules.mk
+
+obj-y += domain_page.o
+obj-y += entry.o
+obj-y += mm.o
+obj-y += seg_fixup.o
+obj-y += traps.o
+
+obj-$(supervisor_mode_kernel) += supervisor_mode_kernel.o
+
+include $(BASEDIR)/Post.mk
diff -r 516cf6553011 -r cf89e8f0831b xen/arch/x86/x86_64/Makefile
--- /dev/null Mon Mar 20 08:56:46 2006
+++ b/xen/arch/x86/x86_64/Makefile Mon Mar 20 08:56:54 2006
@@ -0,0 +1,7 @@
+include $(BASEDIR)/Rules.mk
+
+obj-y += entry.o
+obj-y += mm.o
+obj-y += traps.o
+
+include $(BASEDIR)/Post.mk
diff -r 516cf6553011 -r cf89e8f0831b xen/include/asm-ia64/vlsapic.h
--- /dev/null Mon Mar 20 08:56:46 2006
+++ b/xen/include/asm-ia64/vlsapic.h Mon Mar 20 08:56:54 2006
@@ -0,0 +1,35 @@
+
+
+/* -*- Mode:C; c-basic-offset:4; tab-width:4; indent-tabs-mode:nil -*- */
+/*
+ * Copyright (c) 2004, Intel Corporation.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program; if not, write to the Free Software Foundation, Inc., 59 Temple
+ * Place - Suite 330, Boston, MA 02111-1307 USA.
+ *
+ *
+ */
+
+#ifndef _LSAPIC_H
+#define _LSAPIC_H
+#include <xen/sched.h>
+
+extern void vmx_virq_line_init(struct domain *d);
+extern void vtm_init(struct vcpu *vcpu);
+extern void vtm_set_itc(struct vcpu *vcpu, uint64_t new_itc);
+extern void vtm_set_itm(struct vcpu *vcpu, uint64_t val);
+extern void vtm_set_itv(struct vcpu *vcpu, uint64_t val);
+extern void vmx_vexirq(struct vcpu *vcpu);
+extern void vhpi_detection(struct vcpu *vcpu);
+
+#endif
diff -r 516cf6553011 -r cf89e8f0831b
patches/linux-2.6.16-rc5/i386-mach-io-check-nmi.patch
--- a/patches/linux-2.6.16-rc5/i386-mach-io-check-nmi.patch Mon Mar 20
08:56:46 2006
+++ /dev/null Mon Mar 20 08:56:54 2006
@@ -1,45 +0,0 @@
-diff -pruN ../pristine-linux-2.6.16-rc5/arch/i386/kernel/traps.c
./arch/i386/kernel/traps.c
---- ../pristine-linux-2.6.16-rc5/arch/i386/kernel/traps.c 2006-02-27
15:46:58.000000000 +0000
-+++ ./arch/i386/kernel/traps.c 2006-02-27 15:55:23.000000000 +0000
-@@ -567,18 +567,11 @@ static void mem_parity_error(unsigned ch
-
- static void io_check_error(unsigned char reason, struct pt_regs * regs)
- {
-- unsigned long i;
--
- printk(KERN_EMERG "NMI: IOCK error (debug interrupt?)\n");
- show_registers(regs);
-
- /* Re-enable the IOCK line, wait for a few seconds */
-- reason = (reason & 0xf) | 8;
-- outb(reason, 0x61);
-- i = 2000;
-- while (--i) udelay(1000);
-- reason &= ~8;
-- outb(reason, 0x61);
-+ clear_io_check_error(reason);
- }
-
- static void unknown_nmi_error(unsigned char reason, struct pt_regs * regs)
-diff -pruN
../pristine-linux-2.6.16-rc5/include/asm-i386/mach-default/mach_traps.h
./include/asm-i386/mach-default/mach_traps.h
---- ../pristine-linux-2.6.16-rc5/include/asm-i386/mach-default/mach_traps.h
2006-01-03 03:21:10.000000000 +0000
-+++ ./include/asm-i386/mach-default/mach_traps.h 2006-02-27
15:55:23.000000000 +0000
-@@ -15,6 +15,18 @@ static inline void clear_mem_error(unsig
- outb(reason, 0x61);
- }
-
-+static inline void clear_io_check_error(unsigned char reason)
-+{
-+ unsigned long i;
-+
-+ reason = (reason & 0xf) | 8;
-+ outb(reason, 0x61);
-+ i = 2000;
-+ while (--i) udelay(1000);
-+ reason &= ~8;
-+ outb(reason, 0x61);
-+}
-+
- static inline unsigned char get_nmi_reason(void)
- {
- return inb(0x61);
diff -r 516cf6553011 -r cf89e8f0831b patches/linux-2.6.16-rc5/net-csum.patch
--- a/patches/linux-2.6.16-rc5/net-csum.patch Mon Mar 20 08:56:46 2006
+++ /dev/null Mon Mar 20 08:56:54 2006
@@ -1,41 +0,0 @@
-diff -pruN ../pristine-linux-2.6.16-rc5/net/ipv4/netfilter/ip_nat_proto_tcp.c
./net/ipv4/netfilter/ip_nat_proto_tcp.c
---- ../pristine-linux-2.6.16-rc5/net/ipv4/netfilter/ip_nat_proto_tcp.c
2006-02-27 15:47:38.000000000 +0000
-+++ ./net/ipv4/netfilter/ip_nat_proto_tcp.c 2006-02-27 15:55:25.000000000
+0000
-@@ -129,10 +129,14 @@ tcp_manip_pkt(struct sk_buff **pskb,
- if (hdrsize < sizeof(*hdr))
- return 1;
-
-- hdr->check = ip_nat_cheat_check(~oldip, newip,
-+ if ((*pskb)->proto_csum_blank) {
-+ hdr->check = ip_nat_cheat_check(oldip, ~newip, hdr->check);
-+ } else {
-+ hdr->check = ip_nat_cheat_check(~oldip, newip,
- ip_nat_cheat_check(oldport ^ 0xFFFF,
- newport,
- hdr->check));
-+ }
- return 1;
- }
-
-diff -pruN ../pristine-linux-2.6.16-rc5/net/ipv4/netfilter/ip_nat_proto_udp.c
./net/ipv4/netfilter/ip_nat_proto_udp.c
---- ../pristine-linux-2.6.16-rc5/net/ipv4/netfilter/ip_nat_proto_udp.c
2006-02-27 15:47:38.000000000 +0000
-+++ ./net/ipv4/netfilter/ip_nat_proto_udp.c 2006-02-27 15:55:25.000000000
+0000
-@@ -113,11 +113,16 @@ udp_manip_pkt(struct sk_buff **pskb,
- newport = tuple->dst.u.udp.port;
- portptr = &hdr->dest;
- }
-- if (hdr->check) /* 0 is a special case meaning no checksum */
-- hdr->check = ip_nat_cheat_check(~oldip, newip,
-+ if (hdr->check) { /* 0 is a special case meaning no checksum */
-+ if ((*pskb)->proto_csum_blank) {
-+ hdr->check = ip_nat_cheat_check(oldip, ~newip,
hdr->check);
-+ } else {
-+ hdr->check = ip_nat_cheat_check(~oldip, newip,
- ip_nat_cheat_check(*portptr ^ 0xFFFF,
- newport,
- hdr->check));
-+ }
-+ }
- *portptr = newport;
- return 1;
- }
diff -r 516cf6553011 -r cf89e8f0831b patches/linux-2.6.16-rc5/pmd-shared.patch
--- a/patches/linux-2.6.16-rc5/pmd-shared.patch Mon Mar 20 08:56:46 2006
+++ /dev/null Mon Mar 20 08:56:54 2006
@@ -1,111 +0,0 @@
-diff -pruN ../pristine-linux-2.6.16-rc5/arch/i386/mm/pageattr.c
./arch/i386/mm/pageattr.c
---- ../pristine-linux-2.6.16-rc5/arch/i386/mm/pageattr.c 2006-02-27
15:46:58.000000000 +0000
-+++ ./arch/i386/mm/pageattr.c 2006-02-27 15:55:31.000000000 +0000
-@@ -78,7 +78,7 @@ static void set_pmd_pte(pte_t *kpte, uns
- unsigned long flags;
-
- set_pte_atomic(kpte, pte); /* change init_mm */
-- if (PTRS_PER_PMD > 1)
-+ if (HAVE_SHARED_KERNEL_PMD)
- return;
-
- spin_lock_irqsave(&pgd_lock, flags);
-diff -pruN ../pristine-linux-2.6.16-rc5/arch/i386/mm/pgtable.c
./arch/i386/mm/pgtable.c
---- ../pristine-linux-2.6.16-rc5/arch/i386/mm/pgtable.c 2006-01-03
03:21:10.000000000 +0000
-+++ ./arch/i386/mm/pgtable.c 2006-02-27 15:55:31.000000000 +0000
-@@ -215,9 +215,10 @@ void pgd_ctor(void *pgd, kmem_cache_t *c
- spin_lock_irqsave(&pgd_lock, flags);
- }
-
-- clone_pgd_range((pgd_t *)pgd + USER_PTRS_PER_PGD,
-- swapper_pg_dir + USER_PTRS_PER_PGD,
-- KERNEL_PGD_PTRS);
-+ if (PTRS_PER_PMD == 1 || HAVE_SHARED_KERNEL_PMD)
-+ clone_pgd_range((pgd_t *)pgd + USER_PTRS_PER_PGD,
-+ swapper_pg_dir + USER_PTRS_PER_PGD,
-+ KERNEL_PGD_PTRS);
- if (PTRS_PER_PMD > 1)
- return;
-
-@@ -249,6 +250,30 @@ pgd_t *pgd_alloc(struct mm_struct *mm)
- goto out_oom;
- set_pgd(&pgd[i], __pgd(1 + __pa(pmd)));
- }
-+
-+ if (!HAVE_SHARED_KERNEL_PMD) {
-+ unsigned long flags;
-+
-+ for (i = USER_PTRS_PER_PGD; i < PTRS_PER_PGD; i++) {
-+ pmd_t *pmd = kmem_cache_alloc(pmd_cache, GFP_KERNEL);
-+ if (!pmd)
-+ goto out_oom;
-+ set_pgd(&pgd[USER_PTRS_PER_PGD], __pgd(1 + __pa(pmd)));
-+ }
-+
-+ spin_lock_irqsave(&pgd_lock, flags);
-+ for (i = USER_PTRS_PER_PGD; i < PTRS_PER_PGD; i++) {
-+ unsigned long v = (unsigned long)i << PGDIR_SHIFT;
-+ pgd_t *kpgd = pgd_offset_k(v);
-+ pud_t *kpud = pud_offset(kpgd, v);
-+ pmd_t *kpmd = pmd_offset(kpud, v);
-+ pmd_t *pmd = (void *)__va(pgd_val(pgd[i])-1);
-+ memcpy(pmd, kpmd, PAGE_SIZE);
-+ }
-+ pgd_list_add(pgd);
-+ spin_unlock_irqrestore(&pgd_lock, flags);
-+ }
-+
- return pgd;
-
- out_oom:
-@@ -263,9 +288,23 @@ void pgd_free(pgd_t *pgd)
- int i;
-
- /* in the PAE case user pgd entries are overwritten before usage */
-- if (PTRS_PER_PMD > 1)
-- for (i = 0; i < USER_PTRS_PER_PGD; ++i)
-- kmem_cache_free(pmd_cache, (void
*)__va(pgd_val(pgd[i])-1));
-+ if (PTRS_PER_PMD > 1) {
-+ for (i = 0; i < USER_PTRS_PER_PGD; ++i) {
-+ pmd_t *pmd = (void *)__va(pgd_val(pgd[i])-1);
-+ kmem_cache_free(pmd_cache, pmd);
-+ }
-+ if (!HAVE_SHARED_KERNEL_PMD) {
-+ unsigned long flags;
-+ spin_lock_irqsave(&pgd_lock, flags);
-+ pgd_list_del(pgd);
-+ spin_unlock_irqrestore(&pgd_lock, flags);
-+ for (i = USER_PTRS_PER_PGD; i < PTRS_PER_PGD; i++) {
-+ pmd_t *pmd = (void *)__va(pgd_val(pgd[i])-1);
-+ memset(pmd, 0, PTRS_PER_PMD*sizeof(pmd_t));
-+ kmem_cache_free(pmd_cache, pmd);
-+ }
-+ }
-+ }
- /* in the non-PAE case, free_pgtables() clears user pgd entries */
- kmem_cache_free(pgd_cache, pgd);
- }
-diff -pruN ../pristine-linux-2.6.16-rc5/include/asm-i386/pgtable-2level-defs.h
./include/asm-i386/pgtable-2level-defs.h
---- ../pristine-linux-2.6.16-rc5/include/asm-i386/pgtable-2level-defs.h
2006-01-03 03:21:10.000000000 +0000
-+++ ./include/asm-i386/pgtable-2level-defs.h 2006-02-27 15:55:31.000000000
+0000
-@@ -1,6 +1,8 @@
- #ifndef _I386_PGTABLE_2LEVEL_DEFS_H
- #define _I386_PGTABLE_2LEVEL_DEFS_H
-
-+#define HAVE_SHARED_KERNEL_PMD 0
-+
- /*
- * traditional i386 two-level paging structure:
- */
-diff -pruN ../pristine-linux-2.6.16-rc5/include/asm-i386/pgtable-3level-defs.h
./include/asm-i386/pgtable-3level-defs.h
---- ../pristine-linux-2.6.16-rc5/include/asm-i386/pgtable-3level-defs.h
2006-01-03 03:21:10.000000000 +0000
-+++ ./include/asm-i386/pgtable-3level-defs.h 2006-02-27 15:55:31.000000000
+0000
-@@ -1,6 +1,8 @@
- #ifndef _I386_PGTABLE_3LEVEL_DEFS_H
- #define _I386_PGTABLE_3LEVEL_DEFS_H
-
-+#define HAVE_SHARED_KERNEL_PMD 1
-+
- /*
- * PGDIR_SHIFT determines what a top-level page table entry can map
- */
diff -r 516cf6553011 -r cf89e8f0831b patches/linux-2.6.16-rc5/smp-alts.patch
--- a/patches/linux-2.6.16-rc5/smp-alts.patch Mon Mar 20 08:56:46 2006
+++ /dev/null Mon Mar 20 08:56:54 2006
@@ -1,591 +0,0 @@
-diff -pruN ../pristine-linux-2.6.16-rc5/arch/i386/Kconfig ./arch/i386/Kconfig
---- ../pristine-linux-2.6.16-rc5/arch/i386/Kconfig 2006-02-27
15:46:58.000000000 +0000
-+++ ./arch/i386/Kconfig 2006-02-27 15:55:34.000000000 +0000
-@@ -202,6 +202,19 @@ config SMP
-
- If you don't know what to do here, say N.
-
-+config SMP_ALTERNATIVES
-+ bool "SMP alternatives support (EXPERIMENTAL)"
-+ depends on SMP && EXPERIMENTAL
-+ help
-+ Try to reduce the overhead of running an SMP kernel on a uniprocessor
-+ host slightly by replacing certain key instruction sequences
-+ according to whether we currently have more than one CPU available.
-+ This should provide a noticeable boost to performance when
-+ running SMP kernels on UP machines, and have negligible impact
-+ when running on an true SMP host.
-+
-+ If unsure, say N.
-+
- config NR_CPUS
- int "Maximum number of CPUs (2-255)"
- range 2 255
-diff -pruN ../pristine-linux-2.6.16-rc5/arch/i386/kernel/Makefile
./arch/i386/kernel/Makefile
---- ../pristine-linux-2.6.16-rc5/arch/i386/kernel/Makefile 2006-02-27
15:46:58.000000000 +0000
-+++ ./arch/i386/kernel/Makefile 2006-02-27 15:55:34.000000000 +0000
-@@ -37,6 +37,7 @@ obj-$(CONFIG_EFI) += efi.o efi_stub.o
- obj-$(CONFIG_DOUBLEFAULT) += doublefault.o
- obj-$(CONFIG_VM86) += vm86.o
- obj-$(CONFIG_EARLY_PRINTK) += early_printk.o
-+obj-$(CONFIG_SMP_ALTERNATIVES) += smpalts.o
-
- EXTRA_AFLAGS := -traditional
-
-diff -pruN ../pristine-linux-2.6.16-rc5/arch/i386/kernel/smpalts.c
./arch/i386/kernel/smpalts.c
---- ../pristine-linux-2.6.16-rc5/arch/i386/kernel/smpalts.c 1970-01-01
01:00:00.000000000 +0100
-+++ ./arch/i386/kernel/smpalts.c 2006-02-27 15:55:34.000000000 +0000
-@@ -0,0 +1,85 @@
-+#include <linux/kernel.h>
-+#include <asm/system.h>
-+#include <asm/smp_alt.h>
-+#include <asm/processor.h>
-+#include <asm/string.h>
-+
-+struct smp_replacement_record {
-+ unsigned char targ_size;
-+ unsigned char smp1_size;
-+ unsigned char smp2_size;
-+ unsigned char up_size;
-+ unsigned char feature;
-+ unsigned char data[0];
-+};
-+
-+struct smp_alternative_record {
-+ void *targ_start;
-+ struct smp_replacement_record *repl;
-+};
-+
-+extern struct smp_alternative_record __start_smp_alternatives_table,
-+ __stop_smp_alternatives_table;
-+extern unsigned long __init_begin, __init_end;
-+
-+void prepare_for_smp(void)
-+{
-+ struct smp_alternative_record *r;
-+ printk(KERN_INFO "Enabling SMP...\n");
-+ for (r = &__start_smp_alternatives_table;
-+ r != &__stop_smp_alternatives_table;
-+ r++) {
-+ BUG_ON(r->repl->targ_size < r->repl->smp1_size);
-+ BUG_ON(r->repl->targ_size < r->repl->smp2_size);
-+ BUG_ON(r->repl->targ_size < r->repl->up_size);
-+ if (system_state == SYSTEM_RUNNING &&
-+ r->targ_start >= (void *)&__init_begin &&
-+ r->targ_start < (void *)&__init_end)
-+ continue;
-+ if (r->repl->feature != (unsigned char)-1 &&
-+ boot_cpu_has(r->repl->feature)) {
-+ memcpy(r->targ_start,
-+ r->repl->data + r->repl->smp1_size,
-+ r->repl->smp2_size);
-+ memset(r->targ_start + r->repl->smp2_size,
-+ 0x90,
-+ r->repl->targ_size - r->repl->smp2_size);
-+ } else {
-+ memcpy(r->targ_start,
-+ r->repl->data,
-+ r->repl->smp1_size);
-+ memset(r->targ_start + r->repl->smp1_size,
-+ 0x90,
-+ r->repl->targ_size - r->repl->smp1_size);
-+ }
-+ }
-+ /* Paranoia */
-+ asm volatile ("jmp 1f\n1:");
-+ mb();
-+}
-+
-+void unprepare_for_smp(void)
-+{
-+ struct smp_alternative_record *r;
-+ printk(KERN_INFO "Disabling SMP...\n");
-+ for (r = &__start_smp_alternatives_table;
-+ r != &__stop_smp_alternatives_table;
-+ r++) {
-+ BUG_ON(r->repl->targ_size < r->repl->smp1_size);
-+ BUG_ON(r->repl->targ_size < r->repl->smp2_size);
-+ BUG_ON(r->repl->targ_size < r->repl->up_size);
-+ if (system_state == SYSTEM_RUNNING &&
-+ r->targ_start >= (void *)&__init_begin &&
-+ r->targ_start < (void *)&__init_end)
-+ continue;
-+ memcpy(r->targ_start,
-+ r->repl->data + r->repl->smp1_size + r->repl->smp2_size,
-+ r->repl->up_size);
-+ memset(r->targ_start + r->repl->up_size,
-+ 0x90,
-+ r->repl->targ_size - r->repl->up_size);
-+ }
-+ /* Paranoia */
-+ asm volatile ("jmp 1f\n1:");
-+ mb();
-+}
-diff -pruN ../pristine-linux-2.6.16-rc5/arch/i386/kernel/smpboot.c
./arch/i386/kernel/smpboot.c
---- ../pristine-linux-2.6.16-rc5/arch/i386/kernel/smpboot.c 2006-02-27
15:46:58.000000000 +0000
-+++ ./arch/i386/kernel/smpboot.c 2006-02-27 15:55:34.000000000 +0000
-@@ -1208,6 +1208,11 @@ static void __init smp_boot_cpus(unsigne
- if (max_cpus <= cpucount+1)
- continue;
-
-+#ifdef CONFIG_SMP_ALTERNATIVES
-+ if (kicked == 1)
-+ prepare_for_smp();
-+#endif
-+
- if (((cpu = alloc_cpu_id()) <= 0) || do_boot_cpu(apicid, cpu))
- printk("CPU #%d not responding - cannot use it.\n",
- apicid);
-@@ -1386,6 +1391,11 @@ int __devinit __cpu_up(unsigned int cpu)
- return -EIO;
- }
-
-+#ifdef CONFIG_SMP_ALTERNATIVES
-+ if (num_online_cpus() == 1)
-+ prepare_for_smp();
-+#endif
-+
- local_irq_enable();
- per_cpu(cpu_state, cpu) = CPU_UP_PREPARE;
- /* Unleash the CPU! */
-diff -pruN ../pristine-linux-2.6.16-rc5/arch/i386/kernel/vmlinux.lds.S
./arch/i386/kernel/vmlinux.lds.S
---- ../pristine-linux-2.6.16-rc5/arch/i386/kernel/vmlinux.lds.S
2006-01-03 03:21:10.000000000 +0000
-+++ ./arch/i386/kernel/vmlinux.lds.S 2006-02-27 15:55:34.000000000 +0000
-@@ -34,6 +34,13 @@ SECTIONS
- __ex_table : AT(ADDR(__ex_table) - LOAD_OFFSET) { *(__ex_table) }
- __stop___ex_table = .;
-
-+ . = ALIGN(16);
-+ __start_smp_alternatives_table = .;
-+ __smp_alternatives : { *(__smp_alternatives) }
-+ __stop_smp_alternatives_table = .;
-+
-+ __smp_replacements : { *(__smp_replacements) }
-+
- RODATA
-
- /* writeable */
-diff -pruN ../pristine-linux-2.6.16-rc5/include/asm-i386/atomic.h
./include/asm-i386/atomic.h
---- ../pristine-linux-2.6.16-rc5/include/asm-i386/atomic.h 2006-02-27
15:47:25.000000000 +0000
-+++ ./include/asm-i386/atomic.h 2006-02-27 15:55:34.000000000 +0000
-@@ -4,18 +4,13 @@
- #include <linux/config.h>
- #include <linux/compiler.h>
- #include <asm/processor.h>
-+#include <asm/smp_alt.h>
-
- /*
- * Atomic operations that C can't guarantee us. Useful for
- * resource counting etc..
- */
-
--#ifdef CONFIG_SMP
--#define LOCK "lock ; "
--#else
--#define LOCK ""
--#endif
--
- /*
- * Make sure gcc doesn't try to be clever and move things around
- * on us. We need to use _exactly_ the address the user gave us,
-diff -pruN ../pristine-linux-2.6.16-rc5/include/asm-i386/bitops.h
./include/asm-i386/bitops.h
---- ../pristine-linux-2.6.16-rc5/include/asm-i386/bitops.h 2006-02-27
15:47:25.000000000 +0000
-+++ ./include/asm-i386/bitops.h 2006-02-27 15:55:34.000000000 +0000
-@@ -7,6 +7,7 @@
-
- #include <linux/config.h>
- #include <linux/compiler.h>
-+#include <asm/smp_alt.h>
-
- /*
- * These have to be done with inline assembly: that way the bit-setting
-@@ -16,12 +17,6 @@
- * bit 0 is the LSB of addr; bit 32 is the LSB of (addr+1).
- */
-
--#ifdef CONFIG_SMP
--#define LOCK_PREFIX "lock ; "
--#else
--#define LOCK_PREFIX ""
--#endif
--
- #define ADDR (*(volatile long *) addr)
-
- /**
-@@ -41,7 +36,7 @@
- */
- static inline void set_bit(int nr, volatile unsigned long * addr)
- {
-- __asm__ __volatile__( LOCK_PREFIX
-+ __asm__ __volatile__( LOCK
- "btsl %1,%0"
- :"+m" (ADDR)
- :"Ir" (nr));
-@@ -76,7 +71,7 @@ static inline void __set_bit(int nr, vol
- */
- static inline void clear_bit(int nr, volatile unsigned long * addr)
- {
-- __asm__ __volatile__( LOCK_PREFIX
-+ __asm__ __volatile__( LOCK
- "btrl %1,%0"
- :"+m" (ADDR)
- :"Ir" (nr));
-@@ -121,7 +116,7 @@ static inline void __change_bit(int nr,
- */
- static inline void change_bit(int nr, volatile unsigned long * addr)
- {
-- __asm__ __volatile__( LOCK_PREFIX
-+ __asm__ __volatile__( LOCK
- "btcl %1,%0"
- :"+m" (ADDR)
- :"Ir" (nr));
-@@ -140,7 +135,7 @@ static inline int test_and_set_bit(int n
- {
- int oldbit;
-
-- __asm__ __volatile__( LOCK_PREFIX
-+ __asm__ __volatile__( LOCK
- "btsl %2,%1\n\tsbbl %0,%0"
- :"=r" (oldbit),"+m" (ADDR)
- :"Ir" (nr) : "memory");
-@@ -180,7 +175,7 @@ static inline int test_and_clear_bit(int
- {
- int oldbit;
-
-- __asm__ __volatile__( LOCK_PREFIX
-+ __asm__ __volatile__( LOCK
- "btrl %2,%1\n\tsbbl %0,%0"
- :"=r" (oldbit),"+m" (ADDR)
- :"Ir" (nr) : "memory");
-@@ -231,7 +226,7 @@ static inline int test_and_change_bit(in
- {
- int oldbit;
-
-- __asm__ __volatile__( LOCK_PREFIX
-+ __asm__ __volatile__( LOCK
- "btcl %2,%1\n\tsbbl %0,%0"
- :"=r" (oldbit),"+m" (ADDR)
- :"Ir" (nr) : "memory");
-diff -pruN ../pristine-linux-2.6.16-rc5/include/asm-i386/futex.h
./include/asm-i386/futex.h
---- ../pristine-linux-2.6.16-rc5/include/asm-i386/futex.h 2006-02-27
15:47:25.000000000 +0000
-+++ ./include/asm-i386/futex.h 2006-02-27 15:55:34.000000000 +0000
-@@ -28,7 +28,7 @@
- "1: movl %2, %0\n\
- movl %0, %3\n" \
- insn "\n" \
--"2: " LOCK_PREFIX "cmpxchgl %3, %2\n\
-+"2: " LOCK "cmpxchgl %3, %2\n\
- jnz 1b\n\
- 3: .section .fixup,\"ax\"\n\
- 4: mov %5, %1\n\
-@@ -68,7 +68,7 @@ futex_atomic_op_inuser (int encoded_op,
- #endif
- switch (op) {
- case FUTEX_OP_ADD:
-- __futex_atomic_op1(LOCK_PREFIX "xaddl %0, %2", ret,
-+ __futex_atomic_op1(LOCK "xaddl %0, %2", ret,
- oldval, uaddr, oparg);
- break;
- case FUTEX_OP_OR:
-diff -pruN ../pristine-linux-2.6.16-rc5/include/asm-i386/rwsem.h
./include/asm-i386/rwsem.h
---- ../pristine-linux-2.6.16-rc5/include/asm-i386/rwsem.h 2006-01-03
03:21:10.000000000 +0000
-+++ ./include/asm-i386/rwsem.h 2006-02-27 15:55:34.000000000 +0000
-@@ -40,6 +40,7 @@
-
- #include <linux/list.h>
- #include <linux/spinlock.h>
-+#include <asm/smp_alt.h>
-
- struct rwsem_waiter;
-
-@@ -99,7 +100,7 @@ static inline void __down_read(struct rw
- {
- __asm__ __volatile__(
- "# beginning down_read\n\t"
--LOCK_PREFIX " incl (%%eax)\n\t" /* adds 0x00000001, returns the old
value */
-+LOCK " incl (%%eax)\n\t" /* adds 0x00000001, returns the old
value */
- " js 2f\n\t" /* jump if we weren't granted the lock */
- "1:\n\t"
- LOCK_SECTION_START("")
-@@ -130,7 +131,7 @@ static inline int __down_read_trylock(st
- " movl %1,%2\n\t"
- " addl %3,%2\n\t"
- " jle 2f\n\t"
--LOCK_PREFIX " cmpxchgl %2,%0\n\t"
-+LOCK " cmpxchgl %2,%0\n\t"
- " jnz 1b\n\t"
- "2:\n\t"
- "# ending __down_read_trylock\n\t"
-@@ -150,7 +151,7 @@ static inline void __down_write(struct r
- tmp = RWSEM_ACTIVE_WRITE_BIAS;
- __asm__ __volatile__(
- "# beginning down_write\n\t"
--LOCK_PREFIX " xadd %%edx,(%%eax)\n\t" /* subtract 0x0000ffff, returns
the old value */
-+LOCK " xadd %%edx,(%%eax)\n\t" /* subtract 0x0000ffff, returns
the old value */
- " testl %%edx,%%edx\n\t" /* was the count 0 before? */
- " jnz 2f\n\t" /* jump if we weren't granted the lock */
- "1:\n\t"
-@@ -188,7 +189,7 @@ static inline void __up_read(struct rw_s
- __s32 tmp = -RWSEM_ACTIVE_READ_BIAS;
- __asm__ __volatile__(
- "# beginning __up_read\n\t"
--LOCK_PREFIX " xadd %%edx,(%%eax)\n\t" /* subtracts 1, returns the old
value */
-+LOCK " xadd %%edx,(%%eax)\n\t" /* subtracts 1, returns the old
value */
- " js 2f\n\t" /* jump if the lock is being waited upon */
- "1:\n\t"
- LOCK_SECTION_START("")
-@@ -214,7 +215,7 @@ static inline void __up_write(struct rw_
- __asm__ __volatile__(
- "# beginning __up_write\n\t"
- " movl %2,%%edx\n\t"
--LOCK_PREFIX " xaddl %%edx,(%%eax)\n\t" /* tries to transition
0xffff0001 -> 0x00000000 */
-+LOCK " xaddl %%edx,(%%eax)\n\t" /* tries to transition
0xffff0001 -> 0x00000000 */
- " jnz 2f\n\t" /* jump if the lock is being waited upon */
- "1:\n\t"
- LOCK_SECTION_START("")
-@@ -239,7 +240,7 @@ static inline void __downgrade_write(str
- {
- __asm__ __volatile__(
- "# beginning __downgrade_write\n\t"
--LOCK_PREFIX " addl %2,(%%eax)\n\t" /* transitions 0xZZZZ0001 ->
0xYYYY0001 */
-+LOCK " addl %2,(%%eax)\n\t" /* transitions 0xZZZZ0001 ->
0xYYYY0001 */
- " js 2f\n\t" /* jump if the lock is being waited upon */
- "1:\n\t"
- LOCK_SECTION_START("")
-@@ -263,7 +264,7 @@ LOCK_PREFIX " addl %2,(%%eax)\n\t"
- static inline void rwsem_atomic_add(int delta, struct rw_semaphore *sem)
- {
- __asm__ __volatile__(
--LOCK_PREFIX "addl %1,%0"
-+LOCK "addl %1,%0"
- : "=m"(sem->count)
- : "ir"(delta), "m"(sem->count));
- }
-@@ -276,7 +277,7 @@ static inline int rwsem_atomic_update(in
- int tmp = delta;
-
- __asm__ __volatile__(
--LOCK_PREFIX "xadd %0,(%2)"
-+LOCK "xadd %0,(%2)"
- : "+r"(tmp), "=m"(sem->count)
- : "r"(sem), "m"(sem->count)
- : "memory");
-diff -pruN ../pristine-linux-2.6.16-rc5/include/asm-i386/smp_alt.h
./include/asm-i386/smp_alt.h
---- ../pristine-linux-2.6.16-rc5/include/asm-i386/smp_alt.h 1970-01-01
01:00:00.000000000 +0100
-+++ ./include/asm-i386/smp_alt.h 2006-02-27 15:55:34.000000000 +0000
-@@ -0,0 +1,32 @@
-+#ifndef __ASM_SMP_ALT_H__
-+#define __ASM_SMP_ALT_H__
-+
-+#include <linux/config.h>
-+
-+#ifdef CONFIG_SMP
-+#if defined(CONFIG_SMP_ALTERNATIVES) && !defined(MODULE)
-+#define LOCK \
-+ "6677: nop\n" \
-+ ".section __smp_alternatives,\"a\"\n" \
-+ ".long 6677b\n" \
-+ ".long 6678f\n" \
-+ ".previous\n" \
-+ ".section __smp_replacements,\"a\"\n" \
-+ "6678: .byte 1\n" \
-+ ".byte 1\n" \
-+ ".byte 0\n" \
-+ ".byte 1\n" \
-+ ".byte -1\n" \
-+ "lock\n" \
-+ "nop\n" \
-+ ".previous\n"
-+void prepare_for_smp(void);
-+void unprepare_for_smp(void);
-+#else
-+#define LOCK "lock ; "
-+#endif
-+#else
-+#define LOCK ""
-+#endif
-+
-+#endif /* __ASM_SMP_ALT_H__ */
-diff -pruN ../pristine-linux-2.6.16-rc5/include/asm-i386/spinlock.h
./include/asm-i386/spinlock.h
---- ../pristine-linux-2.6.16-rc5/include/asm-i386/spinlock.h 2006-01-03
03:21:10.000000000 +0000
-+++ ./include/asm-i386/spinlock.h 2006-02-27 15:55:34.000000000 +0000
-@@ -6,6 +6,7 @@
- #include <asm/page.h>
- #include <linux/config.h>
- #include <linux/compiler.h>
-+#include <asm/smp_alt.h>
-
- /*
- * Your basic SMP spinlocks, allowing only a single CPU anywhere
-@@ -23,7 +24,8 @@
-
- #define __raw_spin_lock_string \
- "\n1:\t" \
-- "lock ; decb %0\n\t" \
-+ LOCK \
-+ "decb %0\n\t" \
- "jns 3f\n" \
- "2:\t" \
- "rep;nop\n\t" \
-@@ -34,7 +36,8 @@
-
- #define __raw_spin_lock_string_flags \
- "\n1:\t" \
-- "lock ; decb %0\n\t" \
-+ LOCK \
-+ "decb %0\n\t" \
- "jns 4f\n\t" \
- "2:\t" \
- "testl $0x200, %1\n\t" \
-@@ -65,10 +68,34 @@ static inline void __raw_spin_lock_flags
- static inline int __raw_spin_trylock(raw_spinlock_t *lock)
- {
- char oldval;
-+#ifdef CONFIG_SMP_ALTERNATIVES
- __asm__ __volatile__(
-- "xchgb %b0,%1"
-+ "1:movb %1,%b0\n"
-+ "movb $0,%1\n"
-+ "2:"
-+ ".section __smp_alternatives,\"a\"\n"
-+ ".long 1b\n"
-+ ".long 3f\n"
-+ ".previous\n"
-+ ".section __smp_replacements,\"a\"\n"
-+ "3: .byte 2b - 1b\n"
-+ ".byte 5f-4f\n"
-+ ".byte 0\n"
-+ ".byte 6f-5f\n"
-+ ".byte -1\n"
-+ "4: xchgb %b0,%1\n"
-+ "5: movb %1,%b0\n"
-+ "movb $0,%1\n"
-+ "6:\n"
-+ ".previous\n"
- :"=q" (oldval), "=m" (lock->slock)
- :"0" (0) : "memory");
-+#else
-+ __asm__ __volatile__(
-+ "xchgb %b0,%1\n"
-+ :"=q" (oldval), "=m" (lock->slock)
-+ :"0" (0) : "memory");
-+#endif
- return oldval > 0;
- }
-
-@@ -178,12 +205,12 @@ static inline int __raw_write_trylock(ra
-
- static inline void __raw_read_unlock(raw_rwlock_t *rw)
- {
-- asm volatile("lock ; incl %0" :"=m" (rw->lock) : : "memory");
-+ asm volatile(LOCK "incl %0" :"=m" (rw->lock) : : "memory");
- }
-
- static inline void __raw_write_unlock(raw_rwlock_t *rw)
- {
-- asm volatile("lock ; addl $" RW_LOCK_BIAS_STR ", %0"
-+ asm volatile(LOCK "addl $" RW_LOCK_BIAS_STR ", %0"
- : "=m" (rw->lock) : : "memory");
- }
-
-diff -pruN ../pristine-linux-2.6.16-rc5/include/asm-i386/system.h
./include/asm-i386/system.h
---- ../pristine-linux-2.6.16-rc5/include/asm-i386/system.h 2006-02-27
15:47:25.000000000 +0000
-+++ ./include/asm-i386/system.h 2006-02-27 15:55:34.000000000 +0000
-@@ -5,7 +5,7 @@
- #include <linux/kernel.h>
- #include <asm/segment.h>
- #include <asm/cpufeature.h>
--#include <linux/bitops.h> /* for LOCK_PREFIX */
-+#include <asm/smp_alt.h>
-
- #ifdef __KERNEL__
-
-@@ -271,19 +271,19 @@ static inline unsigned long __cmpxchg(vo
- unsigned long prev;
- switch (size) {
- case 1:
-- __asm__ __volatile__(LOCK_PREFIX "cmpxchgb %b1,%2"
-+ __asm__ __volatile__(LOCK "cmpxchgb %b1,%2"
- : "=a"(prev)
- : "q"(new), "m"(*__xg(ptr)), "0"(old)
- : "memory");
- return prev;
- case 2:
-- __asm__ __volatile__(LOCK_PREFIX "cmpxchgw %w1,%2"
-+ __asm__ __volatile__(LOCK "cmpxchgw %w1,%2"
- : "=a"(prev)
- : "r"(new), "m"(*__xg(ptr)), "0"(old)
- : "memory");
- return prev;
- case 4:
-- __asm__ __volatile__(LOCK_PREFIX "cmpxchgl %1,%2"
-+ __asm__ __volatile__(LOCK "cmpxchgl %1,%2"
- : "=a"(prev)
- : "r"(new), "m"(*__xg(ptr)), "0"(old)
- : "memory");
-@@ -336,7 +336,7 @@ static inline unsigned long long __cmpxc
- unsigned long long new)
- {
- unsigned long long prev;
-- __asm__ __volatile__(LOCK_PREFIX "cmpxchg8b %3"
-+ __asm__ __volatile__(LOCK "cmpxchg8b %3"
- : "=A"(prev)
- : "b"((unsigned long)new),
- "c"((unsigned long)(new >> 32)),
-@@ -503,11 +503,55 @@ struct alt_instr {
- #endif
-
- #ifdef CONFIG_SMP
-+#if defined(CONFIG_SMP_ALTERNATIVES) && !defined(MODULE)
-+#define smp_alt_mb(instr) \
-+__asm__ __volatile__("6667:\nnop\nnop\nnop\nnop\nnop\nnop\n6668:\n" \
-+ ".section __smp_alternatives,\"a\"\n" \
-+ ".long 6667b\n" \
-+ ".long 6673f\n" \
-+ ".previous\n" \
-+ ".section __smp_replacements,\"a\"\n" \
-+ "6673:.byte 6668b-6667b\n" \
-+ ".byte 6670f-6669f\n" \
-+ ".byte 6671f-6670f\n" \
-+ ".byte 0\n" \
-+ ".byte %c0\n" \
-+ "6669:lock;addl $0,0(%%esp)\n" \
-+ "6670:" instr "\n" \
-+ "6671:\n" \
-+ ".previous\n" \
-+ : \
-+ : "i" (X86_FEATURE_XMM2) \
-+ : "memory")
-+#define smp_rmb() smp_alt_mb("lfence")
-+#define smp_mb() smp_alt_mb("mfence")
-+#define set_mb(var, value) do { \
-+unsigned long __set_mb_temp; \
-+__asm__ __volatile__("6667:movl %1, %0\n6668:\n" \
-+ ".section __smp_alternatives,\"a\"\n" \
-+ ".long 6667b\n" \
-+ ".long 6673f\n" \
-+ ".previous\n" \
-+ ".section __smp_replacements,\"a\"\n" \
-+ "6673: .byte 6668b-6667b\n" \
-+ ".byte 6670f-6669f\n" \
-+ ".byte 0\n" \
-+ ".byte 6671f-6670f\n" \
-+ ".byte -1\n" \
-+ "6669: xchg %1, %0\n" \
-+ "6670:movl %1, %0\n" \
-+ "6671:\n" \
-+ ".previous\n" \
-+ : "=m" (var), "=r" (__set_mb_temp) \
-+ : "1" (value) \
-+ : "memory"); } while (0)
-+#else
- #define smp_mb() mb()
- #define smp_rmb() rmb()
-+#define set_mb(var, value) do { (void) xchg(&var, value); } while (0)
-+#endif
- #define smp_wmb() wmb()
- #define smp_read_barrier_depends() read_barrier_depends()
--#define set_mb(var, value) do { (void) xchg(&var, value); } while (0)
- #else
- #define smp_mb() barrier()
- #define smp_rmb() barrier()
diff -r 516cf6553011 -r cf89e8f0831b
xen/include/asm-ia64/linux/asm-generic/pci-dma-compat.h
--- a/xen/include/asm-ia64/linux/asm-generic/pci-dma-compat.h Mon Mar 20
08:56:46 2006
+++ /dev/null Mon Mar 20 08:56:54 2006
@@ -1,107 +0,0 @@
-/* include this file if the platform implements the dma_ DMA Mapping API
- * and wants to provide the pci_ DMA Mapping API in terms of it */
-
-#ifndef _ASM_GENERIC_PCI_DMA_COMPAT_H
-#define _ASM_GENERIC_PCI_DMA_COMPAT_H
-
-#include <linux/dma-mapping.h>
-
-/* note pci_set_dma_mask isn't here, since it's a public function
- * exported from drivers/pci, use dma_supported instead */
-
-static inline int
-pci_dma_supported(struct pci_dev *hwdev, u64 mask)
-{
- return dma_supported(hwdev == NULL ? NULL : &hwdev->dev, mask);
-}
-
-static inline void *
-pci_alloc_consistent(struct pci_dev *hwdev, size_t size,
- dma_addr_t *dma_handle)
-{
- return dma_alloc_coherent(hwdev == NULL ? NULL : &hwdev->dev, size,
dma_handle, GFP_ATOMIC);
-}
-
-static inline void
-pci_free_consistent(struct pci_dev *hwdev, size_t size,
- void *vaddr, dma_addr_t dma_handle)
-{
- dma_free_coherent(hwdev == NULL ? NULL : &hwdev->dev, size, vaddr,
dma_handle);
-}
-
-static inline dma_addr_t
-pci_map_single(struct pci_dev *hwdev, void *ptr, size_t size, int direction)
-{
- return dma_map_single(hwdev == NULL ? NULL : &hwdev->dev, ptr, size,
(enum dma_data_direction)direction);
-}
-
-static inline void
-pci_unmap_single(struct pci_dev *hwdev, dma_addr_t dma_addr,
- size_t size, int direction)
-{
- dma_unmap_single(hwdev == NULL ? NULL : &hwdev->dev, dma_addr, size,
(enum dma_data_direction)direction);
-}
-
-static inline dma_addr_t
-pci_map_page(struct pci_dev *hwdev, struct page *page,
- unsigned long offset, size_t size, int direction)
-{
- return dma_map_page(hwdev == NULL ? NULL : &hwdev->dev, page, offset,
size, (enum dma_data_direction)direction);
-}
-
-static inline void
-pci_unmap_page(struct pci_dev *hwdev, dma_addr_t dma_address,
- size_t size, int direction)
-{
- dma_unmap_page(hwdev == NULL ? NULL : &hwdev->dev, dma_address, size,
(enum dma_data_direction)direction);
-}
-
-static inline int
-pci_map_sg(struct pci_dev *hwdev, struct scatterlist *sg,
- int nents, int direction)
-{
- return dma_map_sg(hwdev == NULL ? NULL : &hwdev->dev, sg, nents, (enum
dma_data_direction)direction);
-}
-
-static inline void
-pci_unmap_sg(struct pci_dev *hwdev, struct scatterlist *sg,
- int nents, int direction)
-{
- dma_unmap_sg(hwdev == NULL ? NULL : &hwdev->dev, sg, nents, (enum
dma_data_direction)direction);
-}
-
-static inline void
-pci_dma_sync_single_for_cpu(struct pci_dev *hwdev, dma_addr_t dma_handle,
- size_t size, int direction)
-{
- dma_sync_single_for_cpu(hwdev == NULL ? NULL : &hwdev->dev, dma_handle,
size, (enum dma_data_direction)direction);
-}
-
-static inline void
-pci_dma_sync_single_for_device(struct pci_dev *hwdev, dma_addr_t dma_handle,
- size_t size, int direction)
-{
- dma_sync_single_for_device(hwdev == NULL ? NULL : &hwdev->dev,
dma_handle, size, (enum dma_data_direction)direction);
-}
-
-static inline void
-pci_dma_sync_sg_for_cpu(struct pci_dev *hwdev, struct scatterlist *sg,
- int nelems, int direction)
-{
- dma_sync_sg_for_cpu(hwdev == NULL ? NULL : &hwdev->dev, sg, nelems,
(enum dma_data_direction)direction);
-}
-
-static inline void
-pci_dma_sync_sg_for_device(struct pci_dev *hwdev, struct scatterlist *sg,
- int nelems, int direction)
-{
- dma_sync_sg_for_device(hwdev == NULL ? NULL : &hwdev->dev, sg, nelems,
(enum dma_data_direction)direction);
-}
-
-static inline int
-pci_dma_mapping_error(dma_addr_t dma_addr)
-{
- return dma_mapping_error(dma_addr);
-}
-
-#endif
diff -r 516cf6553011 -r cf89e8f0831b
xen/include/asm-ia64/linux/asm/scatterlist.h
--- a/xen/include/asm-ia64/linux/asm/scatterlist.h Mon Mar 20 08:56:46 2006
+++ /dev/null Mon Mar 20 08:56:54 2006
@@ -1,28 +0,0 @@
-#ifndef _ASM_IA64_SCATTERLIST_H
-#define _ASM_IA64_SCATTERLIST_H
-
-/*
- * Modified 1998-1999, 2001-2002, 2004
- * David Mosberger-Tang <davidm@xxxxxxxxxx>, Hewlett-Packard Co
- */
-
-struct scatterlist {
- struct page *page;
- unsigned int offset;
- unsigned int length; /* buffer length */
-
- dma_addr_t dma_address;
- unsigned int dma_length;
-};
-
-/*
- * It used to be that ISA_DMA_THRESHOLD had something to do with the
- * DMA-limits of ISA-devices. Nowadays, its only remaining use (apart
- * from the aha1542.c driver, which isn't 64-bit clean anyhow) is to
- * tell the block-layer (via BLK_BOUNCE_ISA) what the max. physical
- * address of a page is that is allocated with GFP_DMA. On IA-64,
- * that's 4GB - 1.
- */
-#define ISA_DMA_THRESHOLD 0xffffffff
-
-#endif /* _ASM_IA64_SCATTERLIST_H */
diff -r 516cf6553011 -r cf89e8f0831b xen/include/asm-ia64/linux/mmzone.h
--- a/xen/include/asm-ia64/linux/mmzone.h Mon Mar 20 08:56:46 2006
+++ /dev/null Mon Mar 20 08:56:54 2006
@@ -1,592 +0,0 @@
-#ifndef _LINUX_MMZONE_H
-#define _LINUX_MMZONE_H
-
-#ifdef __KERNEL__
-#ifndef __ASSEMBLY__
-
-#include <linux/config.h>
-#include <linux/spinlock.h>
-#include <linux/list.h>
-#include <linux/wait.h>
-#include <linux/cache.h>
-#include <linux/threads.h>
-#include <linux/numa.h>
-#include <linux/init.h>
-#include <asm/atomic.h>
-
-/* Free memory management - zoned buddy allocator. */
-#ifndef CONFIG_FORCE_MAX_ZONEORDER
-#define MAX_ORDER 11
-#else
-#define MAX_ORDER CONFIG_FORCE_MAX_ZONEORDER
-#endif
-
-struct free_area {
- struct list_head free_list;
- unsigned long nr_free;
-};
-
-struct pglist_data;
-
-/*
- * zone->lock and zone->lru_lock are two of the hottest locks in the kernel.
- * So add a wild amount of padding here to ensure that they fall into separate
- * cachelines. There are very few zone structures in the machine, so space
- * consumption is not a concern here.
- */
-#if defined(CONFIG_SMP)
-struct zone_padding {
- char x[0];
-} ____cacheline_maxaligned_in_smp;
-#define ZONE_PADDING(name) struct zone_padding name;
-#else
-#define ZONE_PADDING(name)
-#endif
-
-struct per_cpu_pages {
- int count; /* number of pages in the list */
- int low; /* low watermark, refill needed */
- int high; /* high watermark, emptying needed */
- int batch; /* chunk size for buddy add/remove */
- struct list_head list; /* the list of pages */
-};
-
-struct per_cpu_pageset {
- struct per_cpu_pages pcp[2]; /* 0: hot. 1: cold */
-#ifdef CONFIG_NUMA
- unsigned long numa_hit; /* allocated in intended node */
- unsigned long numa_miss; /* allocated in non intended node */
- unsigned long numa_foreign; /* was intended here, hit elsewhere */
- unsigned long interleave_hit; /* interleaver prefered this zone */
- unsigned long local_node; /* allocation from local node */
- unsigned long other_node; /* allocation from other node */
-#endif
-} ____cacheline_aligned_in_smp;
-
-#ifdef CONFIG_NUMA
-#define zone_pcp(__z, __cpu) ((__z)->pageset[(__cpu)])
-#else
-#define zone_pcp(__z, __cpu) (&(__z)->pageset[(__cpu)])
-#endif
-
-#define ZONE_DMA 0
-#define ZONE_NORMAL 1
-#define ZONE_HIGHMEM 2
-
-#define MAX_NR_ZONES 3 /* Sync this with ZONES_SHIFT */
-#define ZONES_SHIFT 2 /* ceil(log2(MAX_NR_ZONES)) */
-
-
-/*
- * When a memory allocation must conform to specific limitations (such
- * as being suitable for DMA) the caller will pass in hints to the
- * allocator in the gfp_mask, in the zone modifier bits. These bits
- * are used to select a priority ordered list of memory zones which
- * match the requested limits. GFP_ZONEMASK defines which bits within
- * the gfp_mask should be considered as zone modifiers. Each valid
- * combination of the zone modifier bits has a corresponding list
- * of zones (in node_zonelists). Thus for two zone modifiers there
- * will be a maximum of 4 (2 ** 2) zonelists, for 3 modifiers there will
- * be 8 (2 ** 3) zonelists. GFP_ZONETYPES defines the number of possible
- * combinations of zone modifiers in "zone modifier space".
- */
-#define GFP_ZONEMASK 0x03
-/*
- * As an optimisation any zone modifier bits which are only valid when
- * no other zone modifier bits are set (loners) should be placed in
- * the highest order bits of this field. This allows us to reduce the
- * extent of the zonelists thus saving space. For example in the case
- * of three zone modifier bits, we could require up to eight zonelists.
- * If the left most zone modifier is a "loner" then the highest valid
- * zonelist would be four allowing us to allocate only five zonelists.
- * Use the first form when the left most bit is not a "loner", otherwise
- * use the second.
- */
-/* #define GFP_ZONETYPES (GFP_ZONEMASK + 1) */ /* Non-loner */
-#define GFP_ZONETYPES ((GFP_ZONEMASK + 1) / 2 + 1) /* Loner */
-
-/*
- * On machines where it is needed (eg PCs) we divide physical memory
- * into multiple physical zones. On a PC we have 3 zones:
- *
- * ZONE_DMA < 16 MB ISA DMA capable memory
- * ZONE_NORMAL 16-896 MB direct mapped by the kernel
- * ZONE_HIGHMEM > 896 MB only page cache and user processes
- */
-
-struct zone {
- /* Fields commonly accessed by the page allocator */
- unsigned long free_pages;
- unsigned long pages_min, pages_low, pages_high;
- /*
- * We don't know if the memory that we're going to allocate will be
freeable
- * or/and it will be released eventually, so to avoid totally wasting
several
- * GB of ram we must reserve some of the lower zone memory (otherwise
we risk
- * to run OOM on the lower zones despite there's tons of freeable ram
- * on the higher zones). This array is recalculated at runtime if the
- * sysctl_lowmem_reserve_ratio sysctl changes.
- */
- unsigned long lowmem_reserve[MAX_NR_ZONES];
-
-#ifdef CONFIG_NUMA
- struct per_cpu_pageset *pageset[NR_CPUS];
-#else
- struct per_cpu_pageset pageset[NR_CPUS];
-#endif
- /*
- * free areas of different sizes
- */
- spinlock_t lock;
- struct free_area free_area[MAX_ORDER];
-
-
- ZONE_PADDING(_pad1_)
-
- /* Fields commonly accessed by the page reclaim scanner */
- spinlock_t lru_lock;
- struct list_head active_list;
- struct list_head inactive_list;
- unsigned long nr_scan_active;
- unsigned long nr_scan_inactive;
- unsigned long nr_active;
- unsigned long nr_inactive;
- unsigned long pages_scanned; /* since last reclaim */
- int all_unreclaimable; /* All pages pinned */
-
- /*
- * Does the allocator try to reclaim pages from the zone as soon
- * as it fails a watermark_ok() in __alloc_pages?
- */
- int reclaim_pages;
- /* A count of how many reclaimers are scanning this zone */
- atomic_t reclaim_in_progress;
-
- /*
- * prev_priority holds the scanning priority for this zone. It is
- * defined as the scanning priority at which we achieved our reclaim
- * target at the previous try_to_free_pages() or balance_pgdat()
- * invokation.
- *
- * We use prev_priority as a measure of how much stress page reclaim is
- * under - it drives the swappiness decision: whether to unmap mapped
- * pages.
- *
- * temp_priority is used to remember the scanning priority at which
- * this zone was successfully refilled to free_pages == pages_high.
- *
- * Access to both these fields is quite racy even on uniprocessor. But
- * it is expected to average out OK.
- */
- int temp_priority;
- int prev_priority;
-
-
- ZONE_PADDING(_pad2_)
- /* Rarely used or read-mostly fields */
-
- /*
- * wait_table -- the array holding the hash table
- * wait_table_size -- the size of the hash table array
- * wait_table_bits -- wait_table_size == (1 << wait_table_bits)
- *
- * The purpose of all these is to keep track of the people
- * waiting for a page to become available and make them
- * runnable again when possible. The trouble is that this
- * consumes a lot of space, especially when so few things
- * wait on pages at a given time. So instead of using
- * per-page waitqueues, we use a waitqueue hash table.
- *
- * The bucket discipline is to sleep on the same queue when
- * colliding and wake all in that wait queue when removing.
- * When something wakes, it must check to be sure its page is
- * truly available, a la thundering herd. The cost of a
- * collision is great, but given the expected load of the
- * table, they should be so rare as to be outweighed by the
- * benefits from the saved space.
- *
- * __wait_on_page_locked() and unlock_page() in mm/filemap.c, are the
- * primary users of these fields, and in mm/page_alloc.c
- * free_area_init_core() performs the initialization of them.
- */
- wait_queue_head_t * wait_table;
- unsigned long wait_table_size;
- unsigned long wait_table_bits;
-
- /*
- * Discontig memory support fields.
- */
- struct pglist_data *zone_pgdat;
- struct page *zone_mem_map;
- /* zone_start_pfn == zone_start_paddr >> PAGE_SHIFT */
- unsigned long zone_start_pfn;
-
- unsigned long spanned_pages; /* total size, including holes
*/
- unsigned long present_pages; /* amount of memory (excluding
holes) */
-
- /*
- * rarely used fields:
- */
- char *name;
-} ____cacheline_maxaligned_in_smp;
-
-
-/*
- * The "priority" of VM scanning is how much of the queues we will scan in one
- * go. A value of 12 for DEF_PRIORITY implies that we will scan 1/4096th of the
- * queues ("queue_length >> 12") during an aging round.
- */
-#define DEF_PRIORITY 12
-
-/*
- * One allocation request operates on a zonelist. A zonelist
- * is a list of zones, the first one is the 'goal' of the
- * allocation, the other zones are fallback zones, in decreasing
- * priority.
- *
- * Right now a zonelist takes up less than a cacheline. We never
- * modify it apart from boot-up, and only a few indices are used,
- * so despite the zonelist table being relatively big, the cache
- * footprint of this construct is very small.
- */
-struct zonelist {
- struct zone *zones[MAX_NUMNODES * MAX_NR_ZONES + 1]; // NULL delimited
-};
-
-
-/*
- * The pg_data_t structure is used in machines with CONFIG_DISCONTIGMEM
- * (mostly NUMA machines?) to denote a higher-level memory zone than the
- * zone denotes.
- *
- * On NUMA machines, each NUMA node would have a pg_data_t to describe
- * it's memory layout.
- *
- * Memory statistics and page replacement data structures are maintained on a
- * per-zone basis.
- */
-struct bootmem_data;
-typedef struct pglist_data {
- struct zone node_zones[MAX_NR_ZONES];
- struct zonelist node_zonelists[GFP_ZONETYPES];
- int nr_zones;
-#ifdef CONFIG_FLAT_NODE_MEM_MAP
- struct page *node_mem_map;
-#endif
- struct bootmem_data *bdata;
- unsigned long node_start_pfn;
- unsigned long node_present_pages; /* total number of physical pages */
- unsigned long node_spanned_pages; /* total size of physical page
- range, including holes */
- int node_id;
- struct pglist_data *pgdat_next;
- wait_queue_head_t kswapd_wait;
- struct task_struct *kswapd;
- int kswapd_max_order;
-} pg_data_t;
-
-#define node_present_pages(nid) (NODE_DATA(nid)->node_present_pages)
-#define node_spanned_pages(nid) (NODE_DATA(nid)->node_spanned_pages)
-#ifdef CONFIG_FLAT_NODE_MEM_MAP
-#define pgdat_page_nr(pgdat, pagenr) ((pgdat)->node_mem_map + (pagenr))
-#else
-#define pgdat_page_nr(pgdat, pagenr) mfn_to_page((pgdat)->node_start_pfn +
(pagenr))
-#endif
-#define nid_page_nr(nid, pagenr) pgdat_page_nr(NODE_DATA(nid),(pagenr))
-
-extern struct pglist_data *pgdat_list;
-
-void __get_zone_counts(unsigned long *active, unsigned long *inactive,
- unsigned long *free, struct pglist_data *pgdat);
-void get_zone_counts(unsigned long *active, unsigned long *inactive,
- unsigned long *free);
-void build_all_zonelists(void);
-void wakeup_kswapd(struct zone *zone, int order);
-int zone_watermark_ok(struct zone *z, int order, unsigned long mark,
- int alloc_type, int can_try_harder, int gfp_high);
-
-#ifdef CONFIG_HAVE_MEMORY_PRESENT
-void memory_present(int nid, unsigned long start, unsigned long end);
-#else
-static inline void memory_present(int nid, unsigned long start, unsigned long
end) {}
-#endif
-
-#ifdef CONFIG_NEED_NODE_MEMMAP_SIZE
-unsigned long __init node_memmap_size_bytes(int, unsigned long, unsigned long);
-#endif
-
-/*
- * zone_idx() returns 0 for the ZONE_DMA zone, 1 for the ZONE_NORMAL zone, etc.
- */
-#define zone_idx(zone) ((zone) - (zone)->zone_pgdat->node_zones)
-
-/**
- * for_each_pgdat - helper macro to iterate over all nodes
- * @pgdat - pointer to a pg_data_t variable
- *
- * Meant to help with common loops of the form
- * pgdat = pgdat_list;
- * while(pgdat) {
- * ...
- * pgdat = pgdat->pgdat_next;
- * }
- */
-#define for_each_pgdat(pgdat) \
- for (pgdat = pgdat_list; pgdat; pgdat = pgdat->pgdat_next)
-
-/*
- * next_zone - helper magic for for_each_zone()
- * Thanks to William Lee Irwin III for this piece of ingenuity.
- */
-static inline struct zone *next_zone(struct zone *zone)
-{
- pg_data_t *pgdat = zone->zone_pgdat;
-
- if (zone < pgdat->node_zones + MAX_NR_ZONES - 1)
- zone++;
- else if (pgdat->pgdat_next) {
- pgdat = pgdat->pgdat_next;
- zone = pgdat->node_zones;
- } else
- zone = NULL;
-
- return zone;
-}
-
-/**
- * for_each_zone - helper macro to iterate over all memory zones
- * @zone - pointer to struct zone variable
- *
- * The user only needs to declare the zone variable, for_each_zone
- * fills it in. This basically means for_each_zone() is an
- * easier to read version of this piece of code:
- *
- * for (pgdat = pgdat_list; pgdat; pgdat = pgdat->node_next)
- * for (i = 0; i < MAX_NR_ZONES; ++i) {
- * struct zone * z = pgdat->node_zones + i;
- * ...
- * }
- * }
- */
-#define for_each_zone(zone) \
- for (zone = pgdat_list->node_zones; zone; zone = next_zone(zone))
-
-static inline int is_highmem_idx(int idx)
-{
- return (idx == ZONE_HIGHMEM);
-}
-
-static inline int is_normal_idx(int idx)
-{
- return (idx == ZONE_NORMAL);
-}
-/**
- * is_highmem - helper function to quickly check if a struct zone is a
- * highmem zone or not. This is an attempt to keep references
- * to ZONE_{DMA/NORMAL/HIGHMEM/etc} in general code to a minimum.
- * @zone - pointer to struct zone variable
- */
-static inline int is_highmem(struct zone *zone)
-{
- return zone == zone->zone_pgdat->node_zones + ZONE_HIGHMEM;
-}
-
-static inline int is_normal(struct zone *zone)
-{
- return zone == zone->zone_pgdat->node_zones + ZONE_NORMAL;
-}
-
-/* These two functions are used to setup the per zone pages min values */
-struct ctl_table;
-struct file;
-int min_free_kbytes_sysctl_handler(struct ctl_table *, int, struct file *,
- void __user *, size_t *, loff_t *);
-extern int sysctl_lowmem_reserve_ratio[MAX_NR_ZONES-1];
-int lowmem_reserve_ratio_sysctl_handler(struct ctl_table *, int, struct file *,
- void __user *, size_t *, loff_t *);
-
-#include <linux/topology.h>
-/* Returns the number of the current Node. */
-#define numa_node_id() (cpu_to_node(raw_smp_processor_id()))
-
-#ifndef CONFIG_NEED_MULTIPLE_NODES
-
-extern struct pglist_data contig_page_data;
-#define NODE_DATA(nid) (&contig_page_data)
-#define NODE_MEM_MAP(nid) mem_map
-#define MAX_NODES_SHIFT 1
-#define pfn_to_nid(pfn) (0)
-
-#else /* CONFIG_NEED_MULTIPLE_NODES */
-
-#include <asm/mmzone.h>
-
-#endif /* !CONFIG_NEED_MULTIPLE_NODES */
-
-#ifdef CONFIG_SPARSEMEM
-#include <asm/sparsemem.h>
-#endif
-
-#if BITS_PER_LONG == 32 || defined(ARCH_HAS_ATOMIC_UNSIGNED)
-/*
- * with 32 bit page->flags field, we reserve 8 bits for node/zone info.
- * there are 3 zones (2 bits) and this leaves 8-2=6 bits for nodes.
- */
-#define FLAGS_RESERVED 8
-
-#elif BITS_PER_LONG == 64
-/*
- * with 64 bit flags field, there's plenty of room.
- */
-#define FLAGS_RESERVED 32
-
-#else
-
-#error BITS_PER_LONG not defined
-
-#endif
-
-#ifndef CONFIG_HAVE_ARCH_EARLY_PFN_TO_NID
-#define early_pfn_to_nid(nid) (0UL)
-#endif
-
-#define pfn_to_section_nr(pfn) ((pfn) >> PFN_SECTION_SHIFT)
-#define section_nr_to_pfn(sec) ((sec) << PFN_SECTION_SHIFT)
-
-#ifdef CONFIG_SPARSEMEM
-
-/*
- * SECTION_SHIFT #bits space required to store a section #
- *
- * PA_SECTION_SHIFT physical address to/from section number
- * PFN_SECTION_SHIFT pfn to/from section number
- */
-#define SECTIONS_SHIFT (MAX_PHYSMEM_BITS - SECTION_SIZE_BITS)
-
-#define PA_SECTION_SHIFT (SECTION_SIZE_BITS)
-#define PFN_SECTION_SHIFT (SECTION_SIZE_BITS - PAGE_SHIFT)
-
-#define NR_MEM_SECTIONS (1UL << SECTIONS_SHIFT)
-
-#define PAGES_PER_SECTION (1UL << PFN_SECTION_SHIFT)
-#define PAGE_SECTION_MASK (~(PAGES_PER_SECTION-1))
-
-#if (MAX_ORDER - 1 + PAGE_SHIFT) > SECTION_SIZE_BITS
-#error Allocator MAX_ORDER exceeds SECTION_SIZE
-#endif
-
-struct page;
-struct mem_section {
- /*
- * This is, logically, a pointer to an array of struct
- * pages. However, it is stored with some other magic.
- * (see sparse.c::sparse_init_one_section())
- *
- * Making it a UL at least makes someone do a cast
- * before using it wrong.
- */
- unsigned long section_mem_map;
-};
-
-extern struct mem_section mem_section[NR_MEM_SECTIONS];
-
-static inline struct mem_section *__nr_to_section(unsigned long nr)
-{
- return &mem_section[nr];
-}
-
-/*
- * We use the lower bits of the mem_map pointer to store
- * a little bit of information. There should be at least
- * 3 bits here due to 32-bit alignment.
- */
-#define SECTION_MARKED_PRESENT (1UL<<0)
-#define SECTION_HAS_MEM_MAP (1UL<<1)
-#define SECTION_MAP_LAST_BIT (1UL<<2)
-#define SECTION_MAP_MASK (~(SECTION_MAP_LAST_BIT-1))
-
-static inline struct page *__section_mem_map_addr(struct mem_section *section)
-{
- unsigned long map = section->section_mem_map;
- map &= SECTION_MAP_MASK;
- return (struct page *)map;
-}
-
-static inline int valid_section(struct mem_section *section)
-{
- return (section->section_mem_map & SECTION_MARKED_PRESENT);
-}
-
-static inline int section_has_mem_map(struct mem_section *section)
-{
- return (section->section_mem_map & SECTION_HAS_MEM_MAP);
-}
-
-static inline int valid_section_nr(unsigned long nr)
-{
- return valid_section(__nr_to_section(nr));
-}
-
-/*
- * Given a kernel address, find the home node of the underlying memory.
- */
-#define kvaddr_to_nid(kaddr) pfn_to_nid(__pa(kaddr) >> PAGE_SHIFT)
-
-static inline struct mem_section *__pfn_to_section(unsigned long pfn)
-{
- return __nr_to_section(pfn_to_section_nr(pfn));
-}
-
-#define mfn_to_page(pfn) \
-({ \
- unsigned long __pfn = (pfn); \
- __section_mem_map_addr(__pfn_to_section(__pfn)) + __pfn; \
-})
-#define page_to_mfn(page) \
-({ \
- page - __section_mem_map_addr(__nr_to_section( \
- page_to_section(page))); \
-})
-
-static inline int mfn_valid(unsigned long pfn)
-{
- if (pfn_to_section_nr(pfn) >= NR_MEM_SECTIONS)
- return 0;
- return valid_section(__nr_to_section(pfn_to_section_nr(pfn)));
-}
-
-/*
- * These are _only_ used during initialisation, therefore they
- * can use __initdata ... They could have names to indicate
- * this restriction.
- */
-#ifdef CONFIG_NUMA
-#define pfn_to_nid early_pfn_to_nid
-#endif
-
-#define pfn_to_pgdat(pfn) \
-({ \
- NODE_DATA(pfn_to_nid(pfn)); \
-})
-
-#define early_mfn_valid(pfn) mfn_valid(pfn)
-void sparse_init(void);
-#else
-#define sparse_init() do {} while (0)
-#endif /* CONFIG_SPARSEMEM */
-
-#ifdef CONFIG_NODES_SPAN_OTHER_NODES
-#define early_pfn_in_nid(pfn, nid) (early_pfn_to_nid(pfn) == (nid))
-#else
-#define early_pfn_in_nid(pfn, nid) (1)
-#endif
-
-#ifndef early_mfn_valid
-#define early_mfn_valid(pfn) (1)
-#endif
-
-void memory_present(int nid, unsigned long start, unsigned long end);
-unsigned long __init node_memmap_size_bytes(int, unsigned long, unsigned long);
-
-#endif /* !__ASSEMBLY__ */
-#endif /* __KERNEL__ */
-#endif /* _LINUX_MMZONE_H */
diff -r 516cf6553011 -r cf89e8f0831b xen/include/asm-ia64/linux/page-flags.h
--- a/xen/include/asm-ia64/linux/page-flags.h Mon Mar 20 08:56:46 2006
+++ /dev/null Mon Mar 20 08:56:54 2006
@@ -1,324 +0,0 @@
-/*
- * Macros for manipulating and testing page->flags
- */
-
-#ifndef PAGE_FLAGS_H
-#define PAGE_FLAGS_H
-
-#include <linux/percpu.h>
-#include <linux/cache.h>
-#include <asm/pgtable.h>
-
-/*
- * Various page->flags bits:
- *
- * PG_reserved is set for special pages, which can never be swapped out. Some
- * of them might not even exist (eg empty_bad_page)...
- *
- * The PG_private bitflag is set if page->private contains a valid value.
- *
- * During disk I/O, PG_locked is used. This bit is set before I/O and
- * reset when I/O completes. page_waitqueue(page) is a wait queue of all tasks
- * waiting for the I/O on this page to complete.
- *
- * PG_uptodate tells whether the page's contents is valid. When a read
- * completes, the page becomes uptodate, unless a disk I/O error happened.
- *
- * For choosing which pages to swap out, inode pages carry a PG_referenced bit,
- * which is set any time the system accesses that page through the (mapping,
- * index) hash table. This referenced bit, together with the referenced bit
- * in the page tables, is used to manipulate page->age and move the page across
- * the active, inactive_dirty and inactive_clean lists.
- *
- * Note that the referenced bit, the page->lru list_head and the active,
- * inactive_dirty and inactive_clean lists are protected by the
- * zone->lru_lock, and *NOT* by the usual PG_locked bit!
- *
- * PG_error is set to indicate that an I/O error occurred on this page.
- *
- * PG_arch_1 is an architecture specific page state bit. The generic code
- * guarantees that this bit is cleared for a page when it first is entered into
- * the page cache.
- *
- * PG_highmem pages are not permanently mapped into the kernel virtual address
- * space, they need to be kmapped separately for doing IO on the pages. The
- * struct page (these bits with information) are always mapped into kernel
- * address space...
- */
-
-/*
- * Don't use the *_dontuse flags. Use the macros. Otherwise you'll break
- * locked- and dirty-page accounting. The top eight bits of page->flags are
- * used for page->zone, so putting flag bits there doesn't work.
- */
-#define PG_locked 0 /* Page is locked. Don't touch. */
-#define PG_error 1
-#define PG_referenced 2
-#define PG_uptodate 3
-
-#define PG_dirty 4
-#define PG_lru 5
-#define PG_active 6
-#define PG_slab 7 /* slab debug (Suparna wants
this) */
-
-#define PG_checked 8 /* kill me in 2.5.<early>. */
-#define PG_arch_1 9
-#define PG_reserved 10
-#define PG_private 11 /* Has something at ->private */
-
-#define PG_writeback 12 /* Page is under writeback */
-#define PG_nosave 13 /* Used for system suspend/resume */
-#define PG_compound 14 /* Part of a compound page */
-#define PG_swapcache 15 /* Swap page: swp_entry_t in private */
-
-#define PG_mappedtodisk 16 /* Has blocks allocated on-disk
*/
-#define PG_reclaim 17 /* To be reclaimed asap */
-#define PG_nosave_free 18 /* Free, should not be written */
-#define PG_uncached 19 /* Page has been mapped as uncached */
-
-/*
- * Global page accounting. One instance per CPU. Only unsigned longs are
- * allowed.
- */
-struct page_state {
- unsigned long nr_dirty; /* Dirty writeable pages */
- unsigned long nr_writeback; /* Pages under writeback */
- unsigned long nr_unstable; /* NFS unstable pages */
- unsigned long nr_page_table_pages;/* Pages used for pagetables */
- unsigned long nr_mapped; /* mapped into pagetables */
- unsigned long nr_slab; /* In slab */
-#define GET_PAGE_STATE_LAST nr_slab
-
- /*
- * The below are zeroed by get_page_state(). Use get_full_page_state()
- * to add up all these.
- */
- unsigned long pgpgin; /* Disk reads */
- unsigned long pgpgout; /* Disk writes */
- unsigned long pswpin; /* swap reads */
- unsigned long pswpout; /* swap writes */
- unsigned long pgalloc_high; /* page allocations */
-
- unsigned long pgalloc_normal;
- unsigned long pgalloc_dma;
- unsigned long pgfree; /* page freeings */
- unsigned long pgactivate; /* pages moved inactive->active */
- unsigned long pgdeactivate; /* pages moved active->inactive */
-
- unsigned long pgfault; /* faults (major+minor) */
- unsigned long pgmajfault; /* faults (major only) */
- unsigned long pgrefill_high; /* inspected in refill_inactive_zone */
- unsigned long pgrefill_normal;
- unsigned long pgrefill_dma;
-
- unsigned long pgsteal_high; /* total highmem pages reclaimed */
- unsigned long pgsteal_normal;
- unsigned long pgsteal_dma;
- unsigned long pgscan_kswapd_high;/* total highmem pages scanned */
- unsigned long pgscan_kswapd_normal;
-
- unsigned long pgscan_kswapd_dma;
- unsigned long pgscan_direct_high;/* total highmem pages scanned */
- unsigned long pgscan_direct_normal;
- unsigned long pgscan_direct_dma;
- unsigned long pginodesteal; /* pages reclaimed via inode freeing */
-
- unsigned long slabs_scanned; /* slab objects scanned */
- unsigned long kswapd_steal; /* pages reclaimed by kswapd */
- unsigned long kswapd_inodesteal;/* reclaimed via kswapd inode freeing */
- unsigned long pageoutrun; /* kswapd's calls to page reclaim */
- unsigned long allocstall; /* direct reclaim calls */
-
- unsigned long pgrotated; /* pages rotated to tail of the LRU */
- unsigned long nr_bounce; /* pages for bounce buffers */
-};
-
-extern void get_page_state(struct page_state *ret);
-extern void get_full_page_state(struct page_state *ret);
-extern unsigned long __read_page_state(unsigned long offset);
-extern void __mod_page_state(unsigned long offset, unsigned long delta);
-
-#define read_page_state(member) \
- __read_page_state(offsetof(struct page_state, member))
-
-#define mod_page_state(member, delta) \
- __mod_page_state(offsetof(struct page_state, member), (delta))
-
-#define inc_page_state(member) mod_page_state(member, 1UL)
-#define dec_page_state(member) mod_page_state(member, 0UL - 1)
-#define add_page_state(member,delta) mod_page_state(member, (delta))
-#define sub_page_state(member,delta) mod_page_state(member, 0UL - (delta))
-
-#define mod_page_state_zone(zone, member, delta)
\
- do {
\
- unsigned offset;
\
- if (is_highmem(zone))
\
- offset = offsetof(struct page_state, member##_high);
\
- else if (is_normal(zone))
\
- offset = offsetof(struct page_state, member##_normal);
\
- else
\
- offset = offsetof(struct page_state, member##_dma);
\
- __mod_page_state(offset, (delta));
\
- } while (0)
-
-/*
- * Manipulation of page state flags
- */
-#define PageLocked(page) \
- test_bit(PG_locked, &(page)->flags)
-#define SetPageLocked(page) \
- set_bit(PG_locked, &(page)->flags)
-#define TestSetPageLocked(page) \
- test_and_set_bit(PG_locked, &(page)->flags)
-#define ClearPageLocked(page) \
- clear_bit(PG_locked, &(page)->flags)
-#define TestClearPageLocked(page) \
- test_and_clear_bit(PG_locked, &(page)->flags)
-
-#define PageError(page) test_bit(PG_error, &(page)->flags)
-#define SetPageError(page) set_bit(PG_error, &(page)->flags)
-#define ClearPageError(page) clear_bit(PG_error, &(page)->flags)
-
-#define PageReferenced(page) test_bit(PG_referenced, &(page)->flags)
-#define SetPageReferenced(page) set_bit(PG_referenced, &(page)->flags)
-#define ClearPageReferenced(page) clear_bit(PG_referenced, &(page)->flags)
-#define TestClearPageReferenced(page) test_and_clear_bit(PG_referenced,
&(page)->flags)
-
-#define PageUptodate(page) test_bit(PG_uptodate, &(page)->flags)
-#ifndef SetPageUptodate
-#define SetPageUptodate(page) set_bit(PG_uptodate, &(page)->flags)
-#endif
-#define ClearPageUptodate(page) clear_bit(PG_uptodate, &(page)->flags)
-
-#define PageDirty(page) test_bit(PG_dirty, &(page)->flags)
-#define SetPageDirty(page) set_bit(PG_dirty, &(page)->flags)
-#define TestSetPageDirty(page) test_and_set_bit(PG_dirty, &(page)->flags)
-#define ClearPageDirty(page) clear_bit(PG_dirty, &(page)->flags)
-#define TestClearPageDirty(page) test_and_clear_bit(PG_dirty, &(page)->flags)
-
-#define SetPageLRU(page) set_bit(PG_lru, &(page)->flags)
-#define PageLRU(page) test_bit(PG_lru, &(page)->flags)
-#define TestSetPageLRU(page) test_and_set_bit(PG_lru, &(page)->flags)
-#define TestClearPageLRU(page) test_and_clear_bit(PG_lru, &(page)->flags)
-
-#define PageActive(page) test_bit(PG_active, &(page)->flags)
-#define SetPageActive(page) set_bit(PG_active, &(page)->flags)
-#define ClearPageActive(page) clear_bit(PG_active, &(page)->flags)
-#define TestClearPageActive(page) test_and_clear_bit(PG_active, &(page)->flags)
-#define TestSetPageActive(page) test_and_set_bit(PG_active, &(page)->flags)
-
-#define PageSlab(page) test_bit(PG_slab, &(page)->flags)
-#define SetPageSlab(page) set_bit(PG_slab, &(page)->flags)
-#define ClearPageSlab(page) clear_bit(PG_slab, &(page)->flags)
-#define TestClearPageSlab(page) test_and_clear_bit(PG_slab,
&(page)->flags)
-#define TestSetPageSlab(page) test_and_set_bit(PG_slab, &(page)->flags)
-
-#ifdef CONFIG_HIGHMEM
-#define PageHighMem(page) is_highmem(page_zone(page))
-#else
-#define PageHighMem(page) 0 /* needed to optimize away at compile time */
-#endif
-
-#define PageChecked(page) test_bit(PG_checked, &(page)->flags)
-#define SetPageChecked(page) set_bit(PG_checked, &(page)->flags)
-#define ClearPageChecked(page) clear_bit(PG_checked, &(page)->flags)
-
-#define PageReserved(page) test_bit(PG_reserved, &(page)->flags)
-#define SetPageReserved(page) set_bit(PG_reserved, &(page)->flags)
-#define ClearPageReserved(page) clear_bit(PG_reserved, &(page)->flags)
-#define __ClearPageReserved(page) __clear_bit(PG_reserved, &(page)->flags)
-
-#define SetPagePrivate(page) set_bit(PG_private, &(page)->flags)
-#define ClearPagePrivate(page) clear_bit(PG_private, &(page)->flags)
-#define PagePrivate(page) test_bit(PG_private, &(page)->flags)
-#define __SetPagePrivate(page) __set_bit(PG_private, &(page)->flags)
-#define __ClearPagePrivate(page) __clear_bit(PG_private, &(page)->flags)
-
-#define PageWriteback(page) test_bit(PG_writeback, &(page)->flags)
-#define SetPageWriteback(page) \
- do { \
- if (!test_and_set_bit(PG_writeback, \
- &(page)->flags)) \
- inc_page_state(nr_writeback); \
- } while (0)
-#define TestSetPageWriteback(page) \
- ({ \
- int ret; \
- ret = test_and_set_bit(PG_writeback, \
- &(page)->flags); \
- if (!ret) \
- inc_page_state(nr_writeback); \
- ret; \
- })
-#define ClearPageWriteback(page) \
- do { \
- if (test_and_clear_bit(PG_writeback, \
- &(page)->flags)) \
- dec_page_state(nr_writeback); \
- } while (0)
-#define TestClearPageWriteback(page) \
- ({ \
- int ret; \
- ret = test_and_clear_bit(PG_writeback, \
- &(page)->flags); \
- if (ret) \
- dec_page_state(nr_writeback); \
- ret; \
- })
-
-#define PageNosave(page) test_bit(PG_nosave, &(page)->flags)
-#define SetPageNosave(page) set_bit(PG_nosave, &(page)->flags)
-#define TestSetPageNosave(page) test_and_set_bit(PG_nosave,
&(page)->flags)
-#define ClearPageNosave(page) clear_bit(PG_nosave, &(page)->flags)
-#define TestClearPageNosave(page) test_and_clear_bit(PG_nosave,
&(page)->flags)
-
-#define PageNosaveFree(page) test_bit(PG_nosave_free, &(page)->flags)
-#define SetPageNosaveFree(page) set_bit(PG_nosave_free, &(page)->flags)
-#define ClearPageNosaveFree(page) clear_bit(PG_nosave_free,
&(page)->flags)
-
-#define PageMappedToDisk(page) test_bit(PG_mappedtodisk, &(page)->flags)
-#define SetPageMappedToDisk(page) set_bit(PG_mappedtodisk, &(page)->flags)
-#define ClearPageMappedToDisk(page) clear_bit(PG_mappedtodisk, &(page)->flags)
-
-#define PageReclaim(page) test_bit(PG_reclaim, &(page)->flags)
-#define SetPageReclaim(page) set_bit(PG_reclaim, &(page)->flags)
-#define ClearPageReclaim(page) clear_bit(PG_reclaim, &(page)->flags)
-#define TestClearPageReclaim(page) test_and_clear_bit(PG_reclaim,
&(page)->flags)
-
-#ifdef CONFIG_HUGETLB_PAGE
-#define PageCompound(page) test_bit(PG_compound, &(page)->flags)
-#else
-#define PageCompound(page) 0
-#endif
-#define SetPageCompound(page) set_bit(PG_compound, &(page)->flags)
-#define ClearPageCompound(page) clear_bit(PG_compound, &(page)->flags)
-
-#ifdef CONFIG_SWAP
-#define PageSwapCache(page) test_bit(PG_swapcache, &(page)->flags)
-#define SetPageSwapCache(page) set_bit(PG_swapcache, &(page)->flags)
-#define ClearPageSwapCache(page) clear_bit(PG_swapcache, &(page)->flags)
-#else
-#define PageSwapCache(page) 0
-#endif
-
-#define PageUncached(page) test_bit(PG_uncached, &(page)->flags)
-#define SetPageUncached(page) set_bit(PG_uncached, &(page)->flags)
-#define ClearPageUncached(page) clear_bit(PG_uncached, &(page)->flags)
-
-struct page; /* forward declaration */
-
-int test_clear_page_dirty(struct page *page);
-int test_clear_page_writeback(struct page *page);
-int test_set_page_writeback(struct page *page);
-
-static inline void clear_page_dirty(struct page *page)
-{
- test_clear_page_dirty(page);
-}
-
-static inline void set_page_writeback(struct page *page)
-{
- test_set_page_writeback(page);
-}
-
-#endif /* PAGE_FLAGS_H */
diff -r 516cf6553011 -r cf89e8f0831b xen/include/asm-ia64/linux/rbtree.h
--- a/xen/include/asm-ia64/linux/rbtree.h Mon Mar 20 08:56:46 2006
+++ /dev/null Mon Mar 20 08:56:54 2006
@@ -1,141 +0,0 @@
-/*
- Red Black Trees
- (C) 1999 Andrea Arcangeli <andrea@xxxxxxx>
-
- This program is free software; you can redistribute it and/or modify
- it under the terms of the GNU General Public License as published by
- the Free Software Foundation; either version 2 of the License, or
- (at your option) any later version.
-
- This program is distributed in the hope that it will be useful,
- but WITHOUT ANY WARRANTY; without even the implied warranty of
- MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
- GNU General Public License for more details.
-
- You should have received a copy of the GNU General Public License
- along with this program; if not, write to the Free Software
- Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
-
- linux/include/linux/rbtree.h
-
- To use rbtrees you'll have to implement your own insert and search cores.
- This will avoid us to use callbacks and to drop drammatically performances.
- I know it's not the cleaner way, but in C (not in C++) to get
- performances and genericity...
-
- Some example of insert and search follows here. The search is a plain
- normal search over an ordered tree. The insert instead must be implemented
- int two steps: as first thing the code must insert the element in
- order as a red leaf in the tree, then the support library function
- rb_insert_color() must be called. Such function will do the
- not trivial work to rebalance the rbtree if necessary.
-
------------------------------------------------------------------------
-static inline struct page * rb_search_page_cache(struct inode * inode,
- unsigned long offset)
-{
- struct rb_node * n = inode->i_rb_page_cache.rb_node;
- struct page * page;
-
- while (n)
- {
- page = rb_entry(n, struct page, rb_page_cache);
-
- if (offset < page->offset)
- n = n->rb_left;
- else if (offset > page->offset)
- n = n->rb_right;
- else
- return page;
- }
- return NULL;
-}
-
-static inline struct page * __rb_insert_page_cache(struct inode * inode,
- unsigned long offset,
- struct rb_node * node)
-{
- struct rb_node ** p = &inode->i_rb_page_cache.rb_node;
- struct rb_node * parent = NULL;
- struct page * page;
-
- while (*p)
- {
- parent = *p;
- page = rb_entry(parent, struct page, rb_page_cache);
-
- if (offset < page->offset)
- p = &(*p)->rb_left;
- else if (offset > page->offset)
- p = &(*p)->rb_right;
- else
- return page;
- }
-
- rb_link_node(node, parent, p);
-
- return NULL;
-}
-
-static inline struct page * rb_insert_page_cache(struct inode * inode,
- unsigned long offset,
- struct rb_node * node)
-{
- struct page * ret;
- if ((ret = __rb_insert_page_cache(inode, offset, node)))
- goto out;
- rb_insert_color(node, &inode->i_rb_page_cache);
- out:
- return ret;
-}
------------------------------------------------------------------------
-*/
-
-#ifndef _LINUX_RBTREE_H
-#define _LINUX_RBTREE_H
-
-#include <linux/kernel.h>
-#include <linux/stddef.h>
-
-struct rb_node
-{
- struct rb_node *rb_parent;
- int rb_color;
-#define RB_RED 0
-#define RB_BLACK 1
- struct rb_node *rb_right;
- struct rb_node *rb_left;
-};
-
-struct rb_root
-{
- struct rb_node *rb_node;
-};
-
-#define RB_ROOT (struct rb_root) { NULL, }
-#define rb_entry(ptr, type, member) container_of(ptr, type, member)
-
-extern void rb_insert_color(struct rb_node *, struct rb_root *);
-extern void rb_erase(struct rb_node *, struct rb_root *);
-
-/* Find logical next and previous nodes in a tree */
-extern struct rb_node *rb_next(struct rb_node *);
-extern struct rb_node *rb_prev(struct rb_node *);
-extern struct rb_node *rb_first(struct rb_root *);
-extern struct rb_node *rb_last(struct rb_root *);
-
-/* Fast replacement of a single node without remove/rebalance/add/rebalance */
-extern void rb_replace_node(struct rb_node *victim, struct rb_node *new,
- struct rb_root *root);
-
-static inline void rb_link_node(struct rb_node * node, struct rb_node * parent,
- struct rb_node ** rb_link)
-{
- node->rb_parent = parent;
- node->rb_color = RB_RED;
- node->rb_left = node->rb_right = NULL;
-
- *rb_link = node;
-}
-
-#endif /* _LINUX_RBTREE_H */
_______________________________________________
Xen-changelog mailing list
Xen-changelog@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-changelog
|
Home