# HG changeset patch
# User Alex Williamson <alex.williamson@xxxxxx>
# Date 1193780291 21600
# Node ID bac78e892e42cb075d98a52c2c9a81df82f12e10
# Parent 0edeca96e7d7821f8413589e2e1f90bf68b696bd
# Parent 71a5313e13f5d99b1b550b1b5f1c0198fc4c9a70
merge with linux-2.6.18-xen.hg (staging)
---
arch/i386/kernel/entry-xen.S | 23
arch/i386/kernel/smp-xen.c | 8
arch/i386/kernel/sysenter.c | 46 +
arch/i386/kernel/time-xen.c | 41 -
arch/i386/mm/fault-xen.c | 2
arch/i386/mm/hypervisor.c | 1
arch/um/kernel/physmem.c | 4
arch/x86_64/ia32/Makefile | 17
arch/x86_64/ia32/ia32entry-xen.S | 166 +-----
arch/x86_64/ia32/syscall32-xen.c | 49 +
arch/x86_64/ia32/syscall32_syscall-xen.S | 2
arch/x86_64/kernel/entry-xen.S | 7
arch/x86_64/kernel/smp-xen.c | 8
arch/x86_64/kernel/traps-xen.c | 2
arch/x86_64/mm/fault-xen.c | 2
arch/x86_64/mm/init-xen.c | 4
arch/x86_64/mm/pageattr-xen.c | 2
buildconfigs/linux-defconfig_xen0_ia64 | 1
buildconfigs/linux-defconfig_xen0_x86_32 | 1
buildconfigs/linux-defconfig_xen0_x86_64 | 1
buildconfigs/linux-defconfig_xenU_ia64 | 1
buildconfigs/linux-defconfig_xenU_x86_32 | 1
buildconfigs/linux-defconfig_xenU_x86_64 | 1
buildconfigs/linux-defconfig_xen_ia64 | 1
buildconfigs/linux-defconfig_xen_x86_32 | 1
buildconfigs/linux-defconfig_xen_x86_64 | 1
drivers/cpufreq/cpufreq_ondemand.c | 81 ++
drivers/xen/Kconfig | 6
drivers/xen/blktap/blktap.c | 8
drivers/xen/core/evtchn.c | 10
drivers/xen/core/smpboot.c | 1
drivers/xen/netfront/accel.c | 820 +++++++++++-------------------
drivers/xen/netfront/netfront.c | 20
drivers/xen/netfront/netfront.h | 82 ---
drivers/xen/xenbus/xenbus_probe.c | 8
include/asm-i386/mach-xen/asm/page.h | 8
include/asm-ia64/page.h | 9
include/asm-powerpc/page.h | 9
include/asm-um/page.h | 2
include/asm-x86_64/mach-xen/asm/page.h | 8
include/asm-x86_64/mach-xen/asm/pgtable.h | 2
include/linux/gfp.h | 6
include/xen/evtchn.h | 1
include/xen/interface/callback.h | 37 +
include/xen/interface/domctl.h | 56 +-
include/xen/interface/platform.h | 22
include/xen/interface/sysctl.h | 25
include/xen/interface/vcpu.h | 3
include/xen/interface/xen-compat.h | 2
mm/page_alloc.c | 14
50 files changed, 753 insertions(+), 880 deletions(-)
diff -r 0edeca96e7d7 -r bac78e892e42 arch/i386/kernel/entry-xen.S
--- a/arch/i386/kernel/entry-xen.S Tue Oct 30 13:42:54 2007 -0600
+++ b/arch/i386/kernel/entry-xen.S Tue Oct 30 15:38:11 2007 -0600
@@ -381,6 +381,29 @@ sysexit_ecrit: /**** END OF SYSEXIT CRIT
#endif /* !CONFIG_XEN */
CFI_ENDPROC
+ # pv sysenter call handler stub
+ENTRY(sysenter_entry_pv)
+ RING0_INT_FRAME
+ movl $__USER_DS,16(%esp)
+ movl %ebp,12(%esp)
+ movl $__USER_CS,4(%esp)
+ addl $4,%esp
+ /* +5*4 is SS:ESP,EFLAGS,CS:EIP. +8 is esp0 setting. */
+ pushl (TI_sysenter_return-THREAD_SIZE+8+4*4)(%esp)
+/*
+ * Load the potential sixth argument from user stack.
+ * Careful about security.
+ */
+ cmpl $__PAGE_OFFSET-3,%ebp
+ jae syscall_fault
+1: movl (%ebp),%ebp
+.section __ex_table,"a"
+ .align 4
+ .long 1b,syscall_fault
+.previous
+ /* fall through */
+ CFI_ENDPROC
+ENDPROC(sysenter_entry_pv)
# system call handler stub
ENTRY(system_call)
diff -r 0edeca96e7d7 -r bac78e892e42 arch/i386/kernel/smp-xen.c
--- a/arch/i386/kernel/smp-xen.c Tue Oct 30 13:42:54 2007 -0600
+++ b/arch/i386/kernel/smp-xen.c Tue Oct 30 15:38:11 2007 -0600
@@ -558,9 +558,7 @@ static void stop_this_cpu (void * dummy)
*/
cpu_clear(smp_processor_id(), cpu_online_map);
local_irq_disable();
-#if 0
- disable_local_APIC();
-#endif
+ disable_all_local_evtchn();
if (cpu_data[smp_processor_id()].hlt_works_ok)
for(;;) halt();
for (;;);
@@ -575,9 +573,7 @@ void smp_send_stop(void)
smp_call_function(stop_this_cpu, NULL, 1, 0);
local_irq_disable();
-#if 0
- disable_local_APIC();
-#endif
+ disable_all_local_evtchn();
local_irq_enable();
}
diff -r 0edeca96e7d7 -r bac78e892e42 arch/i386/kernel/sysenter.c
--- a/arch/i386/kernel/sysenter.c Tue Oct 30 13:42:54 2007 -0600
+++ b/arch/i386/kernel/sysenter.c Tue Oct 30 15:38:11 2007 -0600
@@ -48,7 +48,7 @@ extern asmlinkage void sysenter_entry(vo
void enable_sep_cpu(void)
{
-#ifndef CONFIG_X86_NO_TSS
+#ifndef CONFIG_XEN
int cpu = get_cpu();
struct tss_struct *tss = &per_cpu(init_tss, cpu);
@@ -62,8 +62,36 @@ void enable_sep_cpu(void)
wrmsr(MSR_IA32_SYSENTER_CS, __KERNEL_CS, 0);
wrmsr(MSR_IA32_SYSENTER_ESP, tss->esp1, 0);
wrmsr(MSR_IA32_SYSENTER_EIP, (unsigned long) sysenter_entry, 0);
- put_cpu();
-#endif
+#else
+ extern asmlinkage void sysenter_entry_pv(void);
+ static struct callback_register sysenter = {
+ .type = CALLBACKTYPE_sysenter,
+ .address = { __KERNEL_CS, (unsigned long)sysenter_entry_pv },
+ };
+
+ if (!boot_cpu_has(X86_FEATURE_SEP))
+ return;
+
+ get_cpu();
+
+ if (xen_feature(XENFEAT_supervisor_mode_kernel))
+ sysenter.address.eip = (unsigned long)sysenter_entry;
+
+ switch (HYPERVISOR_callback_op(CALLBACKOP_register, &sysenter)) {
+ case 0:
+ break;
+#if CONFIG_XEN_COMPAT < 0x030200
+ case -ENOSYS:
+ sysenter.type = CALLBACKTYPE_sysenter_deprecated;
+ if (HYPERVISOR_callback_op(CALLBACKOP_register, &sysenter) == 0)
+ break;
+#endif
+ default:
+ clear_bit(X86_FEATURE_SEP, boot_cpu_data.x86_capability);
+ break;
+ }
+#endif
+ put_cpu();
}
/*
@@ -77,18 +105,6 @@ int __init sysenter_setup(void)
int __init sysenter_setup(void)
{
syscall_page = (void *)get_zeroed_page(GFP_ATOMIC);
-
-#ifdef CONFIG_XEN
- if (boot_cpu_has(X86_FEATURE_SEP)) {
- static struct callback_register __initdata sysenter = {
- .type = CALLBACKTYPE_sysenter,
- .address = { __KERNEL_CS, (unsigned long)sysenter_entry
},
- };
-
- if (HYPERVISOR_callback_op(CALLBACKOP_register, &sysenter) < 0)
- clear_bit(X86_FEATURE_SEP,
boot_cpu_data.x86_capability);
- }
-#endif
#ifdef CONFIG_COMPAT_VDSO
__set_fixmap(FIX_VDSO, __pa(syscall_page), PAGE_READONLY);
diff -r 0edeca96e7d7 -r bac78e892e42 arch/i386/kernel/time-xen.c
--- a/arch/i386/kernel/time-xen.c Tue Oct 30 13:42:54 2007 -0600
+++ b/arch/i386/kernel/time-xen.c Tue Oct 30 15:38:11 2007 -0600
@@ -833,44 +833,15 @@ void notify_arch_cmos_timer(void)
mod_timer(&sync_xen_wallclock_timer, jiffies + 1);
}
-static long clock_cmos_diff, sleep_start;
-
-static int timer_suspend(struct sys_device *dev, pm_message_t state)
-{
- /*
- * Estimate time zone so that set_time can update the clock
- */
- clock_cmos_diff = -get_cmos_time();
- clock_cmos_diff += get_seconds();
- sleep_start = get_cmos_time();
+static int timer_resume(struct sys_device *dev)
+{
+ extern void time_resume(void);
+ time_resume();
return 0;
}
-static int timer_resume(struct sys_device *dev)
-{
- unsigned long flags;
- unsigned long sec;
- unsigned long sleep_length;
-
-#ifdef CONFIG_HPET_TIMER
- if (is_hpet_enabled())
- hpet_reenable();
-#endif
- sec = get_cmos_time() + clock_cmos_diff;
- sleep_length = (get_cmos_time() - sleep_start) * HZ;
- 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);
- touch_softlockup_watchdog();
- return 0;
-}
-
-void time_resume(void);
static struct sysdev_class timer_sysclass = {
- .resume = time_resume,
+ .resume = timer_resume,
set_kset_name("timer"),
};
@@ -1121,7 +1092,7 @@ void local_teardown_timer(unsigned int c
}
#endif
-#if CONFIG_CPU_FREQ
+#ifdef CONFIG_CPU_FREQ
static int time_cpufreq_notifier(struct notifier_block *nb, unsigned long val,
void *data)
{
diff -r 0edeca96e7d7 -r bac78e892e42 arch/i386/mm/fault-xen.c
--- a/arch/i386/mm/fault-xen.c Tue Oct 30 13:42:54 2007 -0600
+++ b/arch/i386/mm/fault-xen.c Tue Oct 30 15:38:11 2007 -0600
@@ -346,7 +346,7 @@ static int spurious_fault(struct pt_regs
if ((error_code & 0x02) && !pte_write(*pte))
return 0;
#ifdef CONFIG_X86_PAE
- if ((error_code & 0x10) && (pte_val(*pte) & _PAGE_NX))
+ if ((error_code & 0x10) && (__pte_val(*pte) & _PAGE_NX))
return 0;
#endif
diff -r 0edeca96e7d7 -r bac78e892e42 arch/i386/mm/hypervisor.c
--- a/arch/i386/mm/hypervisor.c Tue Oct 30 13:42:54 2007 -0600
+++ b/arch/i386/mm/hypervisor.c Tue Oct 30 15:38:11 2007 -0600
@@ -56,6 +56,7 @@ void xen_l1_entry_update(pte_t *ptr, pte
u.val = __pte_val(val);
BUG_ON(HYPERVISOR_mmu_update(&u, 1, NULL, DOMID_SELF) < 0);
}
+EXPORT_SYMBOL_GPL(xen_l1_entry_update);
void xen_l2_entry_update(pmd_t *ptr, pmd_t val)
{
diff -r 0edeca96e7d7 -r bac78e892e42 arch/um/kernel/physmem.c
--- a/arch/um/kernel/physmem.c Tue Oct 30 13:42:54 2007 -0600
+++ b/arch/um/kernel/physmem.c Tue Oct 30 15:38:11 2007 -0600
@@ -226,7 +226,7 @@ EXPORT_SYMBOL(physmem_remove_mapping);
EXPORT_SYMBOL(physmem_remove_mapping);
EXPORT_SYMBOL(physmem_subst_mapping);
-int arch_free_page(struct page *page, int order)
+void arch_free_page(struct page *page, int order)
{
void *virt;
int i;
@@ -235,8 +235,6 @@ int arch_free_page(struct page *page, in
virt = __va(page_to_phys(page + i));
physmem_remove_mapping(virt);
}
-
- return 0;
}
int is_remapped(void *virt)
diff -r 0edeca96e7d7 -r bac78e892e42 arch/x86_64/ia32/Makefile
--- a/arch/x86_64/ia32/Makefile Tue Oct 30 13:42:54 2007 -0600
+++ b/arch/x86_64/ia32/Makefile Tue Oct 30 15:38:11 2007 -0600
@@ -14,11 +14,14 @@ audit-class-$(CONFIG_AUDIT) := audit.o
audit-class-$(CONFIG_AUDIT) := audit.o
obj-$(CONFIG_IA32_EMULATION) += $(audit-class-y)
+syscall32-types-y := sysenter syscall
+syscall32-types-$(subst 1,$(CONFIG_XEN),$(shell expr $(CONFIG_XEN_COMPAT)0 '<'
0x0302000)) += int80
+
$(obj)/syscall32_syscall.o: \
- $(foreach F,sysenter syscall,$(obj)/vsyscall-$F.so)
+ $(foreach F,$(syscall32-types-y),$(obj)/vsyscall-$F.so)
# Teach kbuild about targets
-targets := $(foreach F,sysenter syscall,vsyscall-$F.o vsyscall-$F.so)
+targets := $(foreach F,$(syscall32-types-y),vsyscall-$F.o vsyscall-$F.so)
# The DSO images are built using a special linker script
quiet_cmd_syscall = SYSCALL $@
@@ -27,18 +30,10 @@ quiet_cmd_syscall = SYSCALL $@
-Wl,-soname=linux-gate.so.1 -o $@ \
-Wl,-T,$(filter-out FORCE,$^)
-$(obj)/vsyscall-int80.so \
-$(obj)/vsyscall-sysenter.so $(obj)/vsyscall-syscall.so: \
+$(foreach F,$(syscall32-types-y),$(obj)/vsyscall-$F.so): \
$(obj)/vsyscall-%.so: $(src)/vsyscall.lds $(obj)/vsyscall-%.o FORCE
$(call if_changed,syscall)
AFLAGS_vsyscall-sysenter.o = -m32 -Wa,-32 -Iarch/i386/kernel
AFLAGS_vsyscall-syscall.o = -m32 -Wa,-32 -Iarch/i386/kernel
-
-ifdef CONFIG_XEN
AFLAGS_vsyscall-int80.o = -m32 -Wa,-32 -Iarch/i386/kernel
-CFLAGS_syscall32.o += -DUSE_INT80
-AFLAGS_syscall32_syscall.o += -DUSE_INT80
-
-$(obj)/syscall32_syscall.o: $(obj)/vsyscall-int80.so
-endif
diff -r 0edeca96e7d7 -r bac78e892e42 arch/x86_64/ia32/ia32entry-xen.S
--- a/arch/x86_64/ia32/ia32entry-xen.S Tue Oct 30 13:42:54 2007 -0600
+++ b/arch/x86_64/ia32/ia32entry-xen.S Tue Oct 30 15:38:11 2007 -0600
@@ -92,7 +92,7 @@
* %ebp user stack
* 0(%ebp) Arg6
*
- * Interrupts off.
+ * Interrupts on.
*
* This is purely a fast path. For anything complicated we use the int 0x80
* path below. Set up a complete hardware stack frame to share code
@@ -100,38 +100,26 @@
*/
ENTRY(ia32_sysenter_target)
CFI_STARTPROC32 simple
- CFI_DEF_CFA rsp,0
- CFI_REGISTER rsp,rbp
- __swapgs
- movq %gs:pda_kernelstack, %rsp
- addq $(PDA_STACKOFFSET),%rsp
- /*
- * No need to follow this irqs on/off section: the syscall
- * disabled irqs, here we enable it straight after entry:
- */
- XEN_UNBLOCK_EVENTS(%r11)
- __sti
+ CFI_DEF_CFA rsp,SS+8-RIP+16
+ /*CFI_REL_OFFSET ss,SS-RIP+16*/
+ CFI_REL_OFFSET rsp,RSP-RIP+16
+ /*CFI_REL_OFFSET rflags,EFLAGS-RIP+16*/
+ /*CFI_REL_OFFSET cs,CS-RIP+16*/
+ CFI_REL_OFFSET rip,RIP-RIP+16
+ CFI_REL_OFFSET r11,8
+ CFI_REL_OFFSET rcx,0
+ movq 8(%rsp),%r11
+ CFI_RESTORE r11
+ popq %rcx
+ CFI_ADJUST_CFA_OFFSET -8
+ CFI_RESTORE rcx
movl %ebp,%ebp /* zero extension */
- pushq $__USER32_DS
- CFI_ADJUST_CFA_OFFSET 8
- /*CFI_REL_OFFSET ss,0*/
- pushq %rbp
- CFI_ADJUST_CFA_OFFSET 8
- CFI_REL_OFFSET rsp,0
- pushfq
- CFI_ADJUST_CFA_OFFSET 8
- /*CFI_REL_OFFSET rflags,0*/
- movl $VSYSCALL32_SYSEXIT, %r10d
- CFI_REGISTER rip,r10
- pushq $__USER32_CS
- CFI_ADJUST_CFA_OFFSET 8
- /*CFI_REL_OFFSET cs,0*/
- movl %eax, %eax
- pushq %r10
- CFI_ADJUST_CFA_OFFSET 8
- CFI_REL_OFFSET rip,0
- pushq %rax
- CFI_ADJUST_CFA_OFFSET 8
+ movl %eax,%eax
+ movl $__USER32_DS,40(%rsp)
+ movq %rbp,32(%rsp)
+ movl $__USER32_CS,16(%rsp)
+ movl $VSYSCALL32_SYSEXIT,8(%rsp)
+ movq %rax,(%rsp)
cld
SAVE_ARGS 0,0,0
/* no need to do an access_ok check here because rbp has been
@@ -143,7 +131,6 @@ 1: movl (%rbp),%r9d
GET_THREAD_INFO(%r10)
orl $TS_COMPAT,threadinfo_status(%r10)
testl
$(_TIF_SYSCALL_TRACE|_TIF_SYSCALL_AUDIT|_TIF_SECCOMP),threadinfo_flags(%r10)
- CFI_REMEMBER_STATE
jnz sysenter_tracesys
sysenter_do_call:
cmpl $(IA32_NR_syscalls-1),%eax
@@ -151,33 +138,9 @@ sysenter_do_call:
IA32_ARG_FIXUP 1
call *ia32_sys_call_table(,%rax,8)
movq %rax,RAX-ARGOFFSET(%rsp)
- GET_THREAD_INFO(%r10)
- XEN_BLOCK_EVENTS(%r11)
- __cli
- TRACE_IRQS_OFF
- testl $_TIF_ALLWORK_MASK,threadinfo_flags(%r10)
- jnz int_ret_from_sys_call
- andl $~TS_COMPAT,threadinfo_status(%r10)
- /* clear IF, that popfq doesn't enable interrupts early */
- andl $~0x200,EFLAGS-R11(%rsp)
- RESTORE_ARGS 1,24,1,1,1,1
- popfq
- CFI_ADJUST_CFA_OFFSET -8
- /*CFI_RESTORE rflags*/
- popq %rcx /* User %esp */
- CFI_ADJUST_CFA_OFFSET -8
- CFI_REGISTER rsp,rcx
- movl $VSYSCALL32_SYSEXIT,%edx /* User %eip */
- CFI_REGISTER rip,rdx
- TRACE_IRQS_ON
- __swapgs
- XEN_UNBLOCK_EVENTS(%r11)
- __sti /* sti only takes effect after the next instruction */
- /* sysexit */
- .byte 0xf, 0x35 /* TBD */
+ jmp int_ret_from_sys_call
sysenter_tracesys:
- CFI_RESTORE_STATE
SAVE_REST
CLEAR_RREGS
movq $-ENOSYS,RAX(%rsp) /* really needed? */
@@ -210,7 +173,7 @@ ENDPROC(ia32_sysenter_target)
* %esp user stack
* 0(%esp) Arg6
*
- * Interrupts off.
+ * Interrupts on.
*
* This is purely a fast path. For anything complicated we use the int 0x80
* path below. Set up a complete hardware stack frame to share code
@@ -218,32 +181,20 @@ ENDPROC(ia32_sysenter_target)
*/
ENTRY(ia32_cstar_target)
CFI_STARTPROC32 simple
- CFI_DEF_CFA rsp,PDA_STACKOFFSET
- CFI_REGISTER rip,rcx
- /*CFI_REGISTER rflags,r11*/
- __swapgs
- movl %esp,%r8d
- CFI_REGISTER rsp,r8
- movq %gs:pda_kernelstack,%rsp
- /*
- * No need to follow this irqs on/off section: the syscall
- * disabled irqs and here we enable it straight after entry:
- */
- XEN_UNBLOCK_EVENTS(%r11)
- __sti
- SAVE_ARGS 8,1,1
+ CFI_DEF_CFA rsp,SS+8-RIP+16
+ /*CFI_REL_OFFSET ss,SS-RIP+16*/
+ CFI_REL_OFFSET rsp,RSP-RIP+16
+ /*CFI_REL_OFFSET rflags,EFLAGS-RIP+16*/
+ /*CFI_REL_OFFSET cs,CS-RIP+16*/
+ CFI_REL_OFFSET rip,RIP-RIP+16
movl %eax,%eax /* zero extension */
+ movl RSP-RIP+16(%rsp),%r8d
+ SAVE_ARGS -8,1,1
movq %rax,ORIG_RAX-ARGOFFSET(%rsp)
- movq %rcx,RIP-ARGOFFSET(%rsp)
- CFI_REL_OFFSET rip,RIP-ARGOFFSET
movq %rbp,RCX-ARGOFFSET(%rsp) /* this lies slightly to ptrace */
movl %ebp,%ecx
- movq $__USER32_CS,CS-ARGOFFSET(%rsp)
- movq $__USER32_DS,SS-ARGOFFSET(%rsp)
- movq %r11,EFLAGS-ARGOFFSET(%rsp)
- /*CFI_REL_OFFSET rflags,EFLAGS-ARGOFFSET*/
- movq %r8,RSP-ARGOFFSET(%rsp)
- CFI_REL_OFFSET rsp,RSP-ARGOFFSET
+ movl $__USER32_CS,CS-ARGOFFSET(%rsp)
+ movl $__USER32_DS,SS-ARGOFFSET(%rsp)
/* no need to do an access_ok check here because r8 has been
32bit zero extended */
/* hardware stack frame is complete now */
@@ -254,7 +205,6 @@ 1: movl (%r8),%r9d
GET_THREAD_INFO(%r10)
orl $TS_COMPAT,threadinfo_status(%r10)
testl
$(_TIF_SYSCALL_TRACE|_TIF_SYSCALL_AUDIT|_TIF_SECCOMP),threadinfo_flags(%r10)
- CFI_REMEMBER_STATE
jnz cstar_tracesys
cstar_do_call:
cmpl $IA32_NR_syscalls-1,%eax
@@ -262,26 +212,9 @@ cstar_do_call:
IA32_ARG_FIXUP 1
call *ia32_sys_call_table(,%rax,8)
movq %rax,RAX-ARGOFFSET(%rsp)
- GET_THREAD_INFO(%r10)
- XEN_BLOCK_EVENTS(%r11)
- __cli
- TRACE_IRQS_OFF
- testl $_TIF_ALLWORK_MASK,threadinfo_flags(%r10)
- jnz int_ret_from_sys_call
- andl $~TS_COMPAT,threadinfo_status(%r10)
- RESTORE_ARGS 1,-ARG_SKIP,1,1,1
- movl RIP-ARGOFFSET(%rsp),%ecx
- CFI_REGISTER rip,rcx
- movl EFLAGS-ARGOFFSET(%rsp),%r11d
- /*CFI_REGISTER rflags,r11*/
- TRACE_IRQS_ON
- movl RSP-ARGOFFSET(%rsp),%esp
- CFI_RESTORE rsp
- __swapgs
- sysretl /* TBD */
+ jmp int_ret_from_sys_call
cstar_tracesys:
- CFI_RESTORE_STATE
SAVE_REST
CLEAR_RREGS
movq $-ENOSYS,RAX(%rsp) /* really needed? */
@@ -322,32 +255,27 @@ ia32_badarg:
* Arguments are zero extended. For system calls that want sign extension and
* take long arguments a wrapper is needed. Most calls can just be called
* directly.
- * Assumes it is only called from user space and entered with interrupts off.
+ * Assumes it is only called from user space and entered with interrupts on.
*/
ENTRY(ia32_syscall)
CFI_STARTPROC simple
- CFI_DEF_CFA rsp,SS+8-RIP
- /*CFI_REL_OFFSET ss,SS-RIP*/
- CFI_REL_OFFSET rsp,RSP-RIP
- /*CFI_REL_OFFSET rflags,EFLAGS-RIP*/
- /*CFI_REL_OFFSET cs,CS-RIP*/
- CFI_REL_OFFSET rip,RIP-RIP
- __swapgs
- /*
- * No need to follow this irqs on/off section: the syscall
- * disabled irqs and here we enable it straight after entry:
- */
- XEN_UNBLOCK_EVENTS(%r11)
- __sti
- movq (%rsp),%rcx
+ CFI_DEF_CFA rsp,SS+8-RIP+16
+ /*CFI_REL_OFFSET ss,SS-RIP+16*/
+ CFI_REL_OFFSET rsp,RSP-RIP+16
+ /*CFI_REL_OFFSET rflags,EFLAGS-RIP+16*/
+ /*CFI_REL_OFFSET cs,CS-RIP+16*/
+ CFI_REL_OFFSET rip,RIP-RIP+16
+ CFI_REL_OFFSET r11,8
+ CFI_REL_OFFSET rcx,0
movq 8(%rsp),%r11
- addq $0x10,%rsp /* skip rcx and r11 */
+ CFI_RESTORE r11
+ popq %rcx
+ CFI_ADJUST_CFA_OFFSET -8
+ CFI_RESTORE rcx
movl %eax,%eax
- pushq %rax
- CFI_ADJUST_CFA_OFFSET 8
+ movq %rax,(%rsp)
cld
-/* 1: jmp 1b */
/* note the registers are not zero extended to the sf.
this could be a problem. */
SAVE_ARGS 0,0,1
diff -r 0edeca96e7d7 -r bac78e892e42 arch/x86_64/ia32/syscall32-xen.c
--- a/arch/x86_64/ia32/syscall32-xen.c Tue Oct 30 13:42:54 2007 -0600
+++ b/arch/x86_64/ia32/syscall32-xen.c Tue Oct 30 15:38:11 2007 -0600
@@ -13,17 +13,18 @@
#include <asm/proto.h>
#include <asm/tlbflush.h>
#include <asm/ia32_unistd.h>
+#include <xen/interface/callback.h>
-#ifdef USE_INT80
-extern unsigned char syscall32_int80[], syscall32_int80_end[];
-#endif
extern unsigned char syscall32_syscall[], syscall32_syscall_end[];
extern unsigned char syscall32_sysenter[], syscall32_sysenter_end[];
extern int sysctl_vsyscall32;
char *syscall32_page;
-#ifndef USE_INT80
static int use_sysenter = -1;
+
+#if CONFIG_XEN_COMPAT < 0x030200
+extern unsigned char syscall32_int80[], syscall32_int80_end[];
+static int use_int80 = 1;
#endif
static struct page *
@@ -86,13 +87,12 @@ static int __init init_syscall32(void)
if (!syscall32_page)
panic("Cannot allocate syscall32 page");
-#ifdef USE_INT80
- /*
- * At this point we use int 0x80.
- */
- memcpy(syscall32_page, syscall32_int80,
- syscall32_int80_end - syscall32_int80);
-#else
+#if CONFIG_XEN_COMPAT < 0x030200
+ if (use_int80) {
+ memcpy(syscall32_page, syscall32_int80,
+ syscall32_int80_end - syscall32_int80);
+ } else
+#endif
if (use_sysenter > 0) {
memcpy(syscall32_page, syscall32_sysenter,
syscall32_sysenter_end - syscall32_sysenter);
@@ -100,7 +100,6 @@ static int __init init_syscall32(void)
memcpy(syscall32_page, syscall32_syscall,
syscall32_syscall_end - syscall32_syscall);
}
-#endif
return 0;
}
@@ -113,16 +112,26 @@ core_initcall(init_syscall32);
/* May not be __init: called during resume */
void syscall32_cpu_init(void)
{
-#ifndef USE_INT80
- if (use_sysenter < 0)
- use_sysenter = (boot_cpu_data.x86_vendor == X86_VENDOR_INTEL);
+ static struct callback_register cstar = {
+ .type = CALLBACKTYPE_syscall32,
+ .address = (unsigned long)ia32_cstar_target
+ };
+ static struct callback_register sysenter = {
+ .type = CALLBACKTYPE_sysenter,
+ .address = (unsigned long)ia32_sysenter_target
+ };
/* Load these always in case some future AMD CPU supports
SYSENTER from compat mode too. */
- checking_wrmsrl(MSR_IA32_SYSENTER_CS, (u64)__KERNEL_CS);
- checking_wrmsrl(MSR_IA32_SYSENTER_ESP, 0ULL);
- checking_wrmsrl(MSR_IA32_SYSENTER_EIP, (u64)ia32_sysenter_target);
+ if ((HYPERVISOR_callback_op(CALLBACKOP_register, &sysenter) < 0) ||
+ (HYPERVISOR_callback_op(CALLBACKOP_register, &cstar) < 0))
+#if CONFIG_XEN_COMPAT < 0x030200
+ return;
+ use_int80 = 0;
+#else
+ BUG();
+#endif
- wrmsrl(MSR_CSTAR, ia32_cstar_target);
-#endif
+ if (use_sysenter < 0)
+ use_sysenter = (boot_cpu_data.x86_vendor == X86_VENDOR_INTEL);
}
diff -r 0edeca96e7d7 -r bac78e892e42 arch/x86_64/ia32/syscall32_syscall-xen.S
--- a/arch/x86_64/ia32/syscall32_syscall-xen.S Tue Oct 30 13:42:54 2007 -0600
+++ b/arch/x86_64/ia32/syscall32_syscall-xen.S Tue Oct 30 15:38:11 2007 -0600
@@ -2,7 +2,7 @@
.section ".init.data","aw"
-#ifdef USE_INT80
+#if CONFIG_XEN_COMPAT < 0x030200
.globl syscall32_int80
.globl syscall32_int80_end
diff -r 0edeca96e7d7 -r bac78e892e42 arch/x86_64/kernel/entry-xen.S
--- a/arch/x86_64/kernel/entry-xen.S Tue Oct 30 13:42:54 2007 -0600
+++ b/arch/x86_64/kernel/entry-xen.S Tue Oct 30 15:38:11 2007 -0600
@@ -237,7 +237,7 @@ END(ret_from_fork)
* r11 eflags for syscall/sysret, temporary for C
* r12-r15,rbp,rbx saved by C code, not touched.
*
- * Interrupts are off on entry.
+ * Interrupts are enabled on entry.
* Only called from user space.
*
* XXX if we had a free scratch register we could save the RSP into the stack
frame
@@ -252,11 +252,6 @@ ENTRY(system_call)
_frame (RIP-0x10)
SAVE_ARGS -8,0
movq %rax,ORIG_RAX-ARGOFFSET(%rsp)
- /*
- * No need to follow this irqs off/on section - it's straight
- * and short:
- */
- XEN_UNBLOCK_EVENTS(%r11)
GET_THREAD_INFO(%rcx)
testl
$(_TIF_SYSCALL_TRACE|_TIF_SYSCALL_AUDIT|_TIF_SECCOMP),threadinfo_flags(%rcx)
CFI_REMEMBER_STATE
diff -r 0edeca96e7d7 -r bac78e892e42 arch/x86_64/kernel/smp-xen.c
--- a/arch/x86_64/kernel/smp-xen.c Tue Oct 30 13:42:54 2007 -0600
+++ b/arch/x86_64/kernel/smp-xen.c Tue Oct 30 15:38:11 2007 -0600
@@ -482,9 +482,7 @@ void smp_stop_cpu(void)
*/
cpu_clear(smp_processor_id(), cpu_online_map);
local_irq_save(flags);
-#ifndef CONFIG_XEN
- disable_local_APIC();
-#endif
+ disable_all_local_evtchn();
local_irq_restore(flags);
}
@@ -512,9 +510,7 @@ void smp_send_stop(void)
spin_unlock(&call_lock);
local_irq_disable();
-#ifndef CONFIG_XEN
- disable_local_APIC();
-#endif
+ disable_all_local_evtchn();
local_irq_enable();
}
diff -r 0edeca96e7d7 -r bac78e892e42 arch/x86_64/kernel/traps-xen.c
--- a/arch/x86_64/kernel/traps-xen.c Tue Oct 30 13:42:54 2007 -0600
+++ b/arch/x86_64/kernel/traps-xen.c Tue Oct 30 15:38:11 2007 -0600
@@ -1110,7 +1110,7 @@ static trap_info_t trap_table[] = {
#endif
{ 19, 0|4, __KERNEL_CS, (unsigned long)simd_coprocessor_error },
#ifdef CONFIG_IA32_EMULATION
- { IA32_SYSCALL_VECTOR, 3|4, __KERNEL_CS, (unsigned long)ia32_syscall},
+ { IA32_SYSCALL_VECTOR, 3, __KERNEL_CS, (unsigned long)ia32_syscall},
#endif
{ 0, 0, 0, 0 }
};
diff -r 0edeca96e7d7 -r bac78e892e42 arch/x86_64/mm/fault-xen.c
--- a/arch/x86_64/mm/fault-xen.c Tue Oct 30 13:42:54 2007 -0600
+++ b/arch/x86_64/mm/fault-xen.c Tue Oct 30 15:38:11 2007 -0600
@@ -380,7 +380,7 @@ static int spurious_fault(struct pt_regs
return 0;
if ((error_code & PF_WRITE) && !pte_write(*pte))
return 0;
- if ((error_code & PF_INSTR) && (pte_val(*pte) & _PAGE_NX))
+ if ((error_code & PF_INSTR) && (__pte_val(*pte) & _PAGE_NX))
return 0;
return 1;
diff -r 0edeca96e7d7 -r bac78e892e42 arch/x86_64/mm/init-xen.c
--- a/arch/x86_64/mm/init-xen.c Tue Oct 30 13:42:54 2007 -0600
+++ b/arch/x86_64/mm/init-xen.c Tue Oct 30 15:38:11 2007 -0600
@@ -272,7 +272,7 @@ static __init void set_pte_phys(unsigned
pte = pte_offset_kernel(pmd, vaddr);
if (!pte_none(*pte) &&
- pte_val(*pte) != (pte_val(new_pte) & __supported_pte_mask))
+ __pte_val(*pte) != (__pte_val(new_pte) & __supported_pte_mask))
pte_ERROR(*pte);
set_pte(pte, new_pte);
@@ -868,7 +868,7 @@ void __init clear_kernel_mapping(unsigne
pmd = pmd_offset(pud, address);
if (!pmd || pmd_none(*pmd))
continue;
- if (0 == (pmd_val(*pmd) & _PAGE_PSE)) {
+ if (0 == (__pmd_val(*pmd) & _PAGE_PSE)) {
/* Could handle this, but it should not happen
currently. */
printk(KERN_ERR
"clear_kernel_mapping: mapping has been split. will leak
memory\n");
diff -r 0edeca96e7d7 -r bac78e892e42 arch/x86_64/mm/pageattr-xen.c
--- a/arch/x86_64/mm/pageattr-xen.c Tue Oct 30 13:42:54 2007 -0600
+++ b/arch/x86_64/mm/pageattr-xen.c Tue Oct 30 15:38:11 2007 -0600
@@ -370,7 +370,7 @@ static void revert_page(unsigned long ad
pud = pud_offset(pgd,address);
BUG_ON(pud_none(*pud));
pmd = pmd_offset(pud, address);
- BUG_ON(pmd_val(*pmd) & _PAGE_PSE);
+ BUG_ON(__pmd_val(*pmd) & _PAGE_PSE);
pgprot_val(ref_prot) |= _PAGE_PSE;
large_pte = mk_pte_phys(__pa(address) & LARGE_PAGE_MASK, ref_prot);
set_pte((pte_t *)pmd, large_pte);
diff -r 0edeca96e7d7 -r bac78e892e42 buildconfigs/linux-defconfig_xen0_ia64
--- a/buildconfigs/linux-defconfig_xen0_ia64 Tue Oct 30 13:42:54 2007 -0600
+++ b/buildconfigs/linux-defconfig_xen0_ia64 Tue Oct 30 15:38:11 2007 -0600
@@ -1683,6 +1683,7 @@ CONFIG_XEN_SYSFS=y
CONFIG_XEN_SYSFS=y
CONFIG_XEN_COMPAT_030002_AND_LATER=y
# CONFIG_XEN_COMPAT_030004_AND_LATER is not set
+# CONFIG_XEN_COMPAT_030100_AND_LATER is not set
# CONFIG_XEN_COMPAT_LATEST_ONLY is not set
CONFIG_XEN_COMPAT=0x030002
CONFIG_HAVE_IRQ_IGNORE_UNHANDLED=y
diff -r 0edeca96e7d7 -r bac78e892e42 buildconfigs/linux-defconfig_xen0_x86_32
--- a/buildconfigs/linux-defconfig_xen0_x86_32 Tue Oct 30 13:42:54 2007 -0600
+++ b/buildconfigs/linux-defconfig_xen0_x86_32 Tue Oct 30 15:38:11 2007 -0600
@@ -1421,6 +1421,7 @@ CONFIG_XEN_SYSFS=y
CONFIG_XEN_SYSFS=y
CONFIG_XEN_COMPAT_030002_AND_LATER=y
# CONFIG_XEN_COMPAT_030004_AND_LATER is not set
+# CONFIG_XEN_COMPAT_030100_AND_LATER is not set
# CONFIG_XEN_COMPAT_LATEST_ONLY is not set
CONFIG_XEN_COMPAT=0x030002
CONFIG_HAVE_IRQ_IGNORE_UNHANDLED=y
diff -r 0edeca96e7d7 -r bac78e892e42 buildconfigs/linux-defconfig_xen0_x86_64
--- a/buildconfigs/linux-defconfig_xen0_x86_64 Tue Oct 30 13:42:54 2007 -0600
+++ b/buildconfigs/linux-defconfig_xen0_x86_64 Tue Oct 30 15:38:11 2007 -0600
@@ -1369,6 +1369,7 @@ CONFIG_XEN_SYSFS=y
CONFIG_XEN_SYSFS=y
CONFIG_XEN_COMPAT_030002_AND_LATER=y
# CONFIG_XEN_COMPAT_030004_AND_LATER is not set
+# CONFIG_XEN_COMPAT_030100_AND_LATER is not set
# CONFIG_XEN_COMPAT_LATEST_ONLY is not set
CONFIG_XEN_COMPAT=0x030002
CONFIG_HAVE_IRQ_IGNORE_UNHANDLED=y
diff -r 0edeca96e7d7 -r bac78e892e42 buildconfigs/linux-defconfig_xenU_ia64
--- a/buildconfigs/linux-defconfig_xenU_ia64 Tue Oct 30 13:42:54 2007 -0600
+++ b/buildconfigs/linux-defconfig_xenU_ia64 Tue Oct 30 15:38:11 2007 -0600
@@ -1493,6 +1493,7 @@ CONFIG_XEN_SYSFS=y
CONFIG_XEN_SYSFS=y
CONFIG_XEN_COMPAT_030002_AND_LATER=y
# CONFIG_XEN_COMPAT_030004_AND_LATER is not set
+# CONFIG_XEN_COMPAT_030100_AND_LATER is not set
# CONFIG_XEN_COMPAT_LATEST_ONLY is not set
CONFIG_XEN_COMPAT=0x030002
CONFIG_HAVE_IRQ_IGNORE_UNHANDLED=y
diff -r 0edeca96e7d7 -r bac78e892e42 buildconfigs/linux-defconfig_xenU_x86_32
--- a/buildconfigs/linux-defconfig_xenU_x86_32 Tue Oct 30 13:42:54 2007 -0600
+++ b/buildconfigs/linux-defconfig_xenU_x86_32 Tue Oct 30 15:38:11 2007 -0600
@@ -920,6 +920,7 @@ CONFIG_XEN_SYSFS=y
CONFIG_XEN_SYSFS=y
CONFIG_XEN_COMPAT_030002_AND_LATER=y
# CONFIG_XEN_COMPAT_030004_AND_LATER is not set
+# CONFIG_XEN_COMPAT_030100_AND_LATER is not set
# CONFIG_XEN_COMPAT_LATEST_ONLY is not set
CONFIG_XEN_COMPAT=0x030002
CONFIG_HAVE_IRQ_IGNORE_UNHANDLED=y
diff -r 0edeca96e7d7 -r bac78e892e42 buildconfigs/linux-defconfig_xenU_x86_64
--- a/buildconfigs/linux-defconfig_xenU_x86_64 Tue Oct 30 13:42:54 2007 -0600
+++ b/buildconfigs/linux-defconfig_xenU_x86_64 Tue Oct 30 15:38:11 2007 -0600
@@ -1217,6 +1217,7 @@ CONFIG_XEN_SYSFS=y
CONFIG_XEN_SYSFS=y
CONFIG_XEN_COMPAT_030002_AND_LATER=y
# CONFIG_XEN_COMPAT_030004_AND_LATER is not set
+# CONFIG_XEN_COMPAT_030100_AND_LATER is not set
# CONFIG_XEN_COMPAT_LATEST_ONLY is not set
CONFIG_XEN_COMPAT=0x030002
CONFIG_HAVE_IRQ_IGNORE_UNHANDLED=y
diff -r 0edeca96e7d7 -r bac78e892e42 buildconfigs/linux-defconfig_xen_ia64
--- a/buildconfigs/linux-defconfig_xen_ia64 Tue Oct 30 13:42:54 2007 -0600
+++ b/buildconfigs/linux-defconfig_xen_ia64 Tue Oct 30 15:38:11 2007 -0600
@@ -1683,6 +1683,7 @@ CONFIG_XEN_SYSFS=y
CONFIG_XEN_SYSFS=y
CONFIG_XEN_COMPAT_030002_AND_LATER=y
# CONFIG_XEN_COMPAT_030004_AND_LATER is not set
+# CONFIG_XEN_COMPAT_030100_AND_LATER is not set
# CONFIG_XEN_COMPAT_LATEST_ONLY is not set
CONFIG_XEN_COMPAT=0x030002
CONFIG_HAVE_IRQ_IGNORE_UNHANDLED=y
diff -r 0edeca96e7d7 -r bac78e892e42 buildconfigs/linux-defconfig_xen_x86_32
--- a/buildconfigs/linux-defconfig_xen_x86_32 Tue Oct 30 13:42:54 2007 -0600
+++ b/buildconfigs/linux-defconfig_xen_x86_32 Tue Oct 30 15:38:11 2007 -0600
@@ -3280,6 +3280,7 @@ CONFIG_XEN_SYSFS=y
CONFIG_XEN_SYSFS=y
CONFIG_XEN_COMPAT_030002_AND_LATER=y
# CONFIG_XEN_COMPAT_030004_AND_LATER is not set
+# CONFIG_XEN_COMPAT_030100_AND_LATER is not set
# CONFIG_XEN_COMPAT_LATEST_ONLY is not set
CONFIG_XEN_COMPAT=0x030002
CONFIG_HAVE_IRQ_IGNORE_UNHANDLED=y
diff -r 0edeca96e7d7 -r bac78e892e42 buildconfigs/linux-defconfig_xen_x86_64
--- a/buildconfigs/linux-defconfig_xen_x86_64 Tue Oct 30 13:42:54 2007 -0600
+++ b/buildconfigs/linux-defconfig_xen_x86_64 Tue Oct 30 15:38:11 2007 -0600
@@ -3111,6 +3111,7 @@ CONFIG_XEN_SYSFS=y
CONFIG_XEN_SYSFS=y
CONFIG_XEN_COMPAT_030002_AND_LATER=y
# CONFIG_XEN_COMPAT_030004_AND_LATER is not set
+# CONFIG_XEN_COMPAT_030100_AND_LATER is not set
# CONFIG_XEN_COMPAT_LATEST_ONLY is not set
CONFIG_XEN_COMPAT=0x030002
CONFIG_HAVE_IRQ_IGNORE_UNHANDLED=y
diff -r 0edeca96e7d7 -r bac78e892e42 drivers/cpufreq/cpufreq_ondemand.c
--- a/drivers/cpufreq/cpufreq_ondemand.c Tue Oct 30 13:42:54 2007 -0600
+++ b/drivers/cpufreq/cpufreq_ondemand.c Tue Oct 30 15:38:11 2007 -0600
@@ -222,17 +222,14 @@ static struct attribute_group dbs_attr_g
/************************** sysfs end ************************/
-static void dbs_check_cpu(struct cpu_dbs_info_s *this_dbs_info)
-{
- unsigned int idle_ticks, total_ticks;
- unsigned int load;
+#ifndef CONFIG_XEN
+static int dbs_calc_load(struct cpu_dbs_info_s *this_dbs_info)
+{
+ struct cpufreq_policy *policy;
cputime64_t cur_jiffies;
-
- struct cpufreq_policy *policy;
+ cputime64_t total_ticks, idle_ticks;
unsigned int j;
-
- if (!this_dbs_info->enable)
- return;
+ int load;
policy = this_dbs_info->cur_policy;
cur_jiffies = jiffies64_to_cputime64(get_jiffies_64());
@@ -240,7 +237,7 @@ static void dbs_check_cpu(struct cpu_dbs
this_dbs_info->prev_cpu_wall);
this_dbs_info->prev_cpu_wall = cur_jiffies;
if (!total_ticks)
- return;
+ return 200;
/*
* Every sampling_rate, we check, if current idle time is less
* than 20% (default), then we try to increase frequency
@@ -270,6 +267,70 @@ static void dbs_check_cpu(struct cpu_dbs
idle_ticks = tmp_idle_ticks;
}
load = (100 * (total_ticks - idle_ticks)) / total_ticks;
+ return load;
+}
+#else
+
+#include <xen/interface/platform.h>
+static int dbs_calc_load(struct cpu_dbs_info_s *this_dbs_info)
+{
+ int load = 0;
+ struct xen_platform_op op;
+ uint64_t idletime[NR_CPUS];
+ struct cpufreq_policy *policy;
+ unsigned int j;
+ cpumask_t cpumap;
+
+ op.cmd = XENPF_getidletime;
+ set_xen_guest_handle(op.u.getidletime.cpumap_bitmap, (uint8_t *)
cpus_addr(cpumap));
+ op.u.getidletime.cpumap_nr_cpus = NR_CPUS;// num_online_cpus();
+ set_xen_guest_handle(op.u.getidletime.idletime, idletime);
+ if (HYPERVISOR_platform_op(&op))
+ return 200;
+
+ policy = this_dbs_info->cur_policy;
+ cpumap = policy->cpus;
+ for_each_cpu_mask(j, cpumap) {
+ cputime64_t total_idle_nsecs, tmp_idle_nsecs;
+ cputime64_t total_wall_nsecs, tmp_wall_nsecs;
+ struct cpu_dbs_info_s *j_dbs_info;
+ unsigned long tmp_load;
+
+ j_dbs_info = &per_cpu(cpu_dbs_info, j);
+ total_idle_nsecs = idletime[j];
+ tmp_idle_nsecs = cputime64_sub(total_idle_nsecs,
+ j_dbs_info->prev_cpu_idle);
+ total_wall_nsecs = op.u.getidletime.now;
+ tmp_wall_nsecs = cputime64_sub(total_wall_nsecs,
+ j_dbs_info->prev_cpu_wall);
+
+ if (tmp_wall_nsecs == 0)
+ return 200;
+
+ j_dbs_info->prev_cpu_wall = total_wall_nsecs;
+ j_dbs_info->prev_cpu_idle = total_idle_nsecs;
+
+ tmp_load = (100 * (tmp_wall_nsecs - tmp_idle_nsecs)) /
+ tmp_wall_nsecs;
+ load = max(load, min(100, (int) tmp_load));
+ }
+ return load;
+}
+#endif
+
+static void dbs_check_cpu(struct cpu_dbs_info_s *this_dbs_info)
+{
+ int load;
+
+ struct cpufreq_policy *policy;
+
+ if (!this_dbs_info->enable)
+ return;
+
+ policy = this_dbs_info->cur_policy;
+ load = dbs_calc_load(this_dbs_info);
+ if (load > 100)
+ return;
/* Check for frequency increase */
if (load > dbs_tuners_ins.up_threshold) {
diff -r 0edeca96e7d7 -r bac78e892e42 drivers/xen/Kconfig
--- a/drivers/xen/Kconfig Tue Oct 30 13:42:54 2007 -0600
+++ b/drivers/xen/Kconfig Tue Oct 30 15:38:11 2007 -0600
@@ -13,7 +13,7 @@ if XEN
if XEN
config XEN_INTERFACE_VERSION
hex
- default 0x00030206
+ default 0x00030207
menu "XEN"
@@ -247,6 +247,9 @@ choice
config XEN_COMPAT_030004_AND_LATER
bool "3.0.4 and later"
+ config XEN_COMPAT_030100_AND_LATER
+ bool "3.1.0 and later"
+
config XEN_COMPAT_LATEST_ONLY
bool "no compatibility code"
@@ -255,6 +258,7 @@ config XEN_COMPAT
config XEN_COMPAT
hex
default 0xffffff if XEN_COMPAT_LATEST_ONLY
+ default 0x030100 if XEN_COMPAT_030100_AND_LATER
default 0x030004 if XEN_COMPAT_030004_AND_LATER
default 0x030002 if XEN_COMPAT_030002_AND_LATER
default 0
diff -r 0edeca96e7d7 -r bac78e892e42 drivers/xen/blktap/blktap.c
--- a/drivers/xen/blktap/blktap.c Tue Oct 30 13:42:54 2007 -0600
+++ b/drivers/xen/blktap/blktap.c Tue Oct 30 15:38:11 2007 -0600
@@ -115,13 +115,7 @@ static struct tap_blkif *tapfds[MAX_TAP_
static struct tap_blkif *tapfds[MAX_TAP_DEV];
static int blktap_next_minor;
-static int __init set_blkif_reqs(char *str)
-{
- get_option(&str, &blkif_reqs);
- return 1;
-}
-__setup("blkif_reqs=", set_blkif_reqs);
-
+module_param(blkif_reqs, int, 0);
/* Run-time switchable: /sys/module/blktap/parameters/ */
static unsigned int log_stats = 0;
static unsigned int debug_lvl = 0;
diff -r 0edeca96e7d7 -r bac78e892e42 drivers/xen/core/evtchn.c
--- a/drivers/xen/core/evtchn.c Tue Oct 30 13:42:54 2007 -0600
+++ b/drivers/xen/core/evtchn.c Tue Oct 30 15:38:11 2007 -0600
@@ -892,6 +892,16 @@ void unmask_evtchn(int port)
}
EXPORT_SYMBOL_GPL(unmask_evtchn);
+void disable_all_local_evtchn(void)
+{
+ unsigned i, cpu = smp_processor_id();
+ shared_info_t *s = HYPERVISOR_shared_info;
+
+ for (i = 0; i < NR_EVENT_CHANNELS; ++i)
+ if (cpu_from_evtchn(i) == cpu)
+ synch_set_bit(i, &s->evtchn_mask[0]);
+}
+
static void restore_cpu_virqs(int cpu)
{
struct evtchn_bind_virq bind_virq;
diff -r 0edeca96e7d7 -r bac78e892e42 drivers/xen/core/smpboot.c
--- a/drivers/xen/core/smpboot.c Tue Oct 30 13:42:54 2007 -0600
+++ b/drivers/xen/core/smpboot.c Tue Oct 30 15:38:11 2007 -0600
@@ -174,6 +174,7 @@ void cpu_bringup(void)
void cpu_bringup(void)
{
cpu_init();
+ identify_cpu(cpu_data + smp_processor_id());
touch_softlockup_watchdog();
preempt_disable();
local_irq_enable();
diff -r 0edeca96e7d7 -r bac78e892e42 drivers/xen/netfront/accel.c
--- a/drivers/xen/netfront/accel.c Tue Oct 30 13:42:54 2007 -0600
+++ b/drivers/xen/netfront/accel.c Tue Oct 30 15:38:11 2007 -0600
@@ -31,7 +31,7 @@
#include <linux/netdevice.h>
#include <linux/skbuff.h>
#include <linux/list.h>
-#include <linux/kref.h>
+#include <linux/mutex.h>
#include <xen/xenbus.h>
@@ -45,34 +45,40 @@
#define WPRINTK(fmt, args...) \
printk(KERN_WARNING "netfront/accel: " fmt, ##args)
-#if LINUX_VERSION_CODE <= KERNEL_VERSION(2, 6, 5)
-#define kref_init(x,y) kref_init(x,y)
-#define kref_put(x,y) kref_put(x)
-#else
-#define kref_init(x,y) kref_init(x)
-#define kref_put(x,y) kref_put(x,y)
-#endif
-
/*
* List of all netfront accelerator plugin modules available. Each
* list entry is of type struct netfront_accelerator.
*/
static struct list_head accelerators_list;
-/*
- * Lock to protect access to accelerators_list
- */
+/* Lock to protect access to accelerators_list */
static spinlock_t accelerators_lock;
-/* Forward declaration of kref cleanup functions */
-static void accel_kref_release(struct kref *ref);
-static void vif_kref_release(struct kref *ref);
-
+/* Mutex to prevent concurrent loads and suspends, etc. */
+DEFINE_MUTEX(accelerator_mutex);
void netif_init_accel(void)
{
INIT_LIST_HEAD(&accelerators_list);
spin_lock_init(&accelerators_lock);
+}
+
+void netif_exit_accel(void)
+{
+ struct netfront_accelerator *accelerator, *tmp;
+ unsigned flags;
+
+ spin_lock_irqsave(&accelerators_lock, flags);
+
+ list_for_each_entry_safe(accelerator, tmp, &accelerators_list, link) {
+ BUG_ON(!list_empty(&accelerator->vif_states));
+
+ list_del(&accelerator->link);
+ kfree(accelerator->frontend);
+ kfree(accelerator);
+ }
+
+ spin_unlock_irqrestore(&accelerators_lock, flags);
}
@@ -87,9 +93,6 @@ void init_accelerator_vif(struct netfron
/* It's assumed that these things don't change */
np->accel_vif_state.np = np;
np->accel_vif_state.dev = dev;
-
- np->accel_vif_state.ready_for_probe = 1;
- np->accel_vif_state.need_probe = NULL;
}
@@ -112,18 +115,25 @@ static void add_accelerator_vif(struct n
static void add_accelerator_vif(struct netfront_accelerator *accelerator,
struct netfront_info *np)
{
+ unsigned flags;
+
+ /* Need lock to write list */
+ spin_lock_irqsave(&accelerator->vif_states_lock, flags);
+
if (np->accelerator == NULL) {
np->accelerator = accelerator;
list_add(&np->accel_vif_state.link, &accelerator->vif_states);
} else {
/*
- * May get here legitimately if reconnecting to the
- * same accelerator, eg. after resume, so check that
- * is the case
+ * May get here legitimately if suspend_cancel is
+ * called, but in that case configuration should not
+ * have changed
*/
BUG_ON(np->accelerator != accelerator);
}
+
+ spin_unlock_irqrestore(&accelerator->vif_states_lock, flags);
}
@@ -131,7 +141,8 @@ static void add_accelerator_vif(struct n
* Initialise the state to track an accelerator plugin module.
*/
static int init_accelerator(const char *frontend,
- struct netfront_accelerator **result)
+ struct netfront_accelerator **result,
+ struct netfront_accel_hooks *hooks)
{
struct netfront_accelerator *accelerator =
kmalloc(sizeof(struct netfront_accelerator), GFP_KERNEL);
@@ -154,10 +165,7 @@ static int init_accelerator(const char *
INIT_LIST_HEAD(&accelerator->vif_states);
spin_lock_init(&accelerator->vif_states_lock);
- accelerator->hooks = NULL;
-
- accelerator->ready_for_probe = 1;
- accelerator->need_probe = NULL;
+ accelerator->hooks = hooks;
list_add(&accelerator->link, &accelerators_list);
@@ -174,17 +182,9 @@ static void
static void
accelerator_set_vif_state_hooks(struct netfront_accel_vif_state *vif_state)
{
- /* This function must be called with the vif_state_lock held */
+ /* This function must be called with the vif_states_lock held */
DPRINTK("%p\n",vif_state);
-
- /*
- * Take references to stop hooks disappearing.
- * This persists until vif_kref gets to zero.
- */
- kref_get(&vif_state->np->accelerator->accel_kref);
- /* This persists until vif_state->hooks are cleared */
- kref_init(&vif_state->vif_kref, vif_kref_release);
/* Make sure there are no data path operations going on */
netif_poll_disable(vif_state->np->netdev);
@@ -202,47 +202,22 @@ static void accelerator_probe_new_vif(st
struct netfront_accelerator *accelerator)
{
struct netfront_accel_hooks *hooks;
- unsigned flags;
-
+
DPRINTK("\n");
- spin_lock_irqsave(&accelerator->vif_states_lock, flags);
-
- /*
- * Include this frontend device on the accelerator's list
- */
+ /* Include this frontend device on the accelerator's list */
add_accelerator_vif(accelerator, np);
hooks = accelerator->hooks;
if (hooks) {
- if (np->accel_vif_state.ready_for_probe) {
- np->accel_vif_state.ready_for_probe = 0;
-
- kref_get(&accelerator->accel_kref);
-
- spin_unlock_irqrestore(&accelerator->vif_states_lock,
- flags);
-
- hooks->new_device(np->netdev, dev);
-
- kref_put(&accelerator->accel_kref,
- accel_kref_release);
- /*
- * Hooks will get linked into vif_state by a
- * future call by the accelerator to
- * netfront_accelerator_ready()
- */
- return;
- } else {
- if (np->accel_vif_state.need_probe != NULL)
- DPRINTK("Probe request on vif awaiting
probe\n");
- np->accel_vif_state.need_probe = hooks;
- }
- }
-
- spin_unlock_irqrestore(&accelerator->vif_states_lock,
- flags);
+ hooks->new_device(np->netdev, dev);
+ /*
+ * Hooks will get linked into vif_state by a future
+ * call by the accelerator to netfront_accelerator_ready()
+ */
+ }
+
return;
}
@@ -256,10 +231,12 @@ int netfront_load_accelerator(struct net
const char *frontend)
{
struct netfront_accelerator *accelerator;
- int rc;
+ int rc = 0;
unsigned flags;
DPRINTK(" %s\n", frontend);
+
+ mutex_lock(&accelerator_mutex);
spin_lock_irqsave(&accelerators_lock, flags);
@@ -273,22 +250,24 @@ int netfront_load_accelerator(struct net
accelerator_probe_new_vif(np, dev, accelerator);
+ mutex_unlock(&accelerator_mutex);
return 0;
}
}
/* Couldn't find it, so create a new one and load the module */
- if ((rc = init_accelerator(frontend, &accelerator)) < 0) {
+ if ((rc = init_accelerator(frontend, &accelerator, NULL)) < 0) {
spin_unlock_irqrestore(&accelerators_lock, flags);
+ mutex_unlock(&accelerator_mutex);
return rc;
}
spin_unlock_irqrestore(&accelerators_lock, flags);
/* Include this frontend device on the accelerator's list */
- spin_lock_irqsave(&accelerator->vif_states_lock, flags);
add_accelerator_vif(accelerator, np);
- spin_unlock_irqrestore(&accelerator->vif_states_lock, flags);
+
+ mutex_unlock(&accelerator_mutex);
DPRINTK("requesting module %s\n", frontend);
@@ -300,7 +279,7 @@ int netfront_load_accelerator(struct net
* it's up and running, and we can continue from there
*/
- return 0;
+ return rc;
}
@@ -312,21 +291,11 @@ int netfront_load_accelerator(struct net
*/
static void
accelerator_probe_vifs(struct netfront_accelerator *accelerator,
- struct netfront_accel_hooks *hooks,
- unsigned lock_flags)
+ struct netfront_accel_hooks *hooks)
{
struct netfront_accel_vif_state *vif_state, *tmp;
- /* Calling function must have taken the vif_states_lock */
-
DPRINTK("%p\n", accelerator);
-
- /*
- * kref_init() takes a single reference to the hooks that will
- * persist until the accelerator hooks are removed (e.g. by
- * accelerator module unload)
- */
- kref_init(&accelerator->accel_kref, accel_kref_release);
/*
* Store the hooks for future calls to probe a new device, and
@@ -335,71 +304,23 @@ accelerator_probe_vifs(struct netfront_a
*/
BUG_ON(hooks == NULL);
accelerator->hooks = hooks;
-
+
+ /*
+ * currently hold accelerator_mutex, so don't need
+ * vif_states_lock to read the list
+ */
list_for_each_entry_safe(vif_state, tmp, &accelerator->vif_states,
link) {
struct netfront_info *np = vif_state->np;
-
- if (vif_state->ready_for_probe) {
- vif_state->ready_for_probe = 0;
- kref_get(&accelerator->accel_kref);
-
- /*
- * drop lock before calling hook. hooks are
- * protected by the kref
- */
- spin_unlock_irqrestore(&accelerator->vif_states_lock,
- lock_flags);
-
- hooks->new_device(np->netdev, vif_state->dev);
-
- kref_put(&accelerator->accel_kref, accel_kref_release);
-
- /* Retake lock for next go round the loop */
- spin_lock_irqsave(&accelerator->vif_states_lock,
lock_flags);
-
- /*
- * Hooks will get linked into vif_state by a call to
- * netfront_accelerator_ready() once accelerator
- * plugin is ready for action
- */
- } else {
- if (vif_state->need_probe != NULL)
- DPRINTK("Probe request on vif awaiting
probe\n");
- vif_state->need_probe = hooks;
- }
- }
-
- /* Return with vif_states_lock held, as on entry */
-}
-
-
-/*
- * Wrapper for accelerator_probe_vifs that checks now is a good time
- * to do the probe, and postpones till previous state cleared up if
- * necessary
- */
-static void
-accelerator_probe_vifs_on_load(struct netfront_accelerator *accelerator,
- struct netfront_accel_hooks *hooks)
-{
- unsigned flags;
-
- DPRINTK("\n");
-
- spin_lock_irqsave(&accelerator->vif_states_lock, flags);
-
- if (accelerator->ready_for_probe) {
- accelerator->ready_for_probe = 0;
- accelerator_probe_vifs(accelerator, hooks, flags);
- } else {
- if (accelerator->need_probe)
- DPRINTK("Probe request on accelerator awaiting
probe\n");
- accelerator->need_probe = hooks;
- }
-
- spin_unlock_irqrestore(&accelerator->vif_states_lock,
- flags);
+
+ hooks->new_device(np->netdev, vif_state->dev);
+
+ /*
+ * Hooks will get linked into vif_state by a call to
+ * netfront_accelerator_ready() once accelerator
+ * plugin is ready for action
+ */
+ }
}
@@ -427,6 +348,8 @@ int netfront_accelerator_loaded(int vers
}
}
+ mutex_lock(&accelerator_mutex);
+
spin_lock_irqsave(&accelerators_lock, flags);
/*
@@ -437,9 +360,9 @@ int netfront_accelerator_loaded(int vers
if (match_accelerator(frontend, accelerator)) {
spin_unlock_irqrestore(&accelerators_lock, flags);
- accelerator_probe_vifs_on_load(accelerator, hooks);
-
- return 0;
+ accelerator_probe_vifs(accelerator, hooks);
+
+ goto out;
}
}
@@ -450,10 +373,12 @@ int netfront_accelerator_loaded(int vers
DPRINTK("Couldn't find matching accelerator (%s)\n",
frontend);
- init_accelerator(frontend, &accelerator);
+ init_accelerator(frontend, &accelerator, hooks);
spin_unlock_irqrestore(&accelerators_lock, flags);
+ out:
+ mutex_unlock(&accelerator_mutex);
return 0;
}
EXPORT_SYMBOL_GPL(netfront_accelerator_loaded);
@@ -477,6 +402,10 @@ void netfront_accelerator_ready(const ch
list_for_each_entry(accelerator, &accelerators_list, link) {
if (match_accelerator(frontend, accelerator)) {
+ /*
+ * Mutex not held so need vif_states_lock for
+ * list
+ */
spin_lock_irqsave
(&accelerator->vif_states_lock, flags1);
@@ -489,11 +418,9 @@ void netfront_accelerator_ready(const ch
spin_unlock_irqrestore
(&accelerator->vif_states_lock, flags1);
- goto done;
+ break;
}
}
-
- done:
spin_unlock_irqrestore(&accelerators_lock, flags);
}
EXPORT_SYMBOL_GPL(netfront_accelerator_ready);
@@ -526,36 +453,30 @@ accelerator_remove_single_hook(struct ne
/*
* Safely remove the accelerator function hooks from a netfront state.
*/
-static void accelerator_remove_hooks(struct netfront_accelerator *accelerator,
- int remove_master)
-{
+static void accelerator_remove_hooks(struct netfront_accelerator *accelerator)
+{
+ struct netfront_accel_hooks *hooks;
struct netfront_accel_vif_state *vif_state, *tmp;
unsigned flags;
- spin_lock_irqsave(&accelerator->vif_states_lock, flags);
-
+ /* Mutex is held so don't need vif_states_lock to iterate list */
list_for_each_entry_safe(vif_state, tmp,
&accelerator->vif_states,
link) {
+ spin_lock_irqsave(&accelerator->vif_states_lock, flags);
+
+ BUG_ON(vif_state->hooks == NULL);
+ hooks = vif_state->hooks;
accelerator_remove_single_hook(accelerator, vif_state);
- /*
- * Remove the reference taken when the vif_state hooks
- * were set, must be called without lock held
- */
spin_unlock_irqrestore(&accelerator->vif_states_lock, flags);
- kref_put(&vif_state->vif_kref, vif_kref_release);
- spin_lock_irqsave(&accelerator->vif_states_lock, flags);
+
+ /* Last chance to get statistics from the accelerator */
+ hooks->get_stats(vif_state->np->netdev, &vif_state->np->stats);
+ hooks->remove(vif_state->dev);
}
- if(remove_master)
- accelerator->hooks = NULL;
-
- spin_unlock_irqrestore(&accelerator->vif_states_lock, flags);
-
- if(remove_master)
- /* Remove the reference taken when module loaded */
- kref_put(&accelerator->accel_kref, accel_kref_release);
+ accelerator->hooks = NULL;
}
@@ -564,369 +485,266 @@ static void accelerator_remove_hooks(str
* removes the hooks into the plugin and blocks until all devices have
* finished using it, so on return it is safe to unload.
*/
-void netfront_accelerator_stop(const char *frontend, int unloading)
-{
- struct netfront_accelerator *accelerator;
- unsigned flags;
-
+void netfront_accelerator_stop(const char *frontend)
+{
+ struct netfront_accelerator *accelerator;
+ unsigned flags;
+
+ mutex_lock(&accelerator_mutex);
spin_lock_irqsave(&accelerators_lock, flags);
list_for_each_entry(accelerator, &accelerators_list, link) {
if (match_accelerator(frontend, accelerator)) {
spin_unlock_irqrestore(&accelerators_lock, flags);
- /*
- * Use semaphore to ensure we know when all
- * uses of hooks are complete
- */
- sema_init(&accelerator->exit_semaphore, 0);
-
- accelerator_remove_hooks(accelerator, unloading);
-
- if (unloading)
- /* Wait for hooks to be unused, then return */
- down(&accelerator->exit_semaphore);
-
- return;
+ accelerator_remove_hooks(accelerator);
+
+ goto out;
}
}
spin_unlock_irqrestore(&accelerators_lock, flags);
+ out:
+ mutex_unlock(&accelerator_mutex);
}
EXPORT_SYMBOL_GPL(netfront_accelerator_stop);
-
-int netfront_check_accelerator_queue_busy(struct net_device *dev,
- struct netfront_info *np)
-{
- struct netfront_accel_hooks *hooks;
- int rc = 1;
- unsigned flags;
-
- /*
- * Call the check busy accelerator hook. The use count for the
- * accelerator's hooks is incremented for the duration of the
- * call to prevent the accelerator being able to modify the
- * hooks in the middle (by, for example, unloading)
- */
+/* Helper for call_remove and do_suspend */
+static int do_remove(struct netfront_info *np, struct xenbus_device *dev,
+ unsigned *lock_flags)
+{
+ struct netfront_accelerator *accelerator = np->accelerator;
+ struct netfront_accel_hooks *hooks;
+ int rc = 0;
+
if (np->accel_vif_state.hooks) {
- spin_lock_irqsave(&np->accelerator->vif_states_lock, flags);
hooks = np->accel_vif_state.hooks;
- if (hooks) {
- kref_get(&np->accel_vif_state.vif_kref);
- spin_unlock_irqrestore
- (&np->accelerator->vif_states_lock, flags);
-
- rc = np->accel_vif_state.hooks->check_busy(dev);
-
- kref_put(&np->accel_vif_state.vif_kref,
- vif_kref_release);
- } else {
- spin_unlock_irqrestore
- (&np->accelerator->vif_states_lock, flags);
- }
- }
-
- return rc;
-}
-
-
+
+ /*
+ * Try and do the opposite of accelerator_probe_new_vif
+ * to ensure there's no state pointing back at the
+ * netdev
+ */
+ accelerator_remove_single_hook(accelerator,
+ &np->accel_vif_state);
+
+ /* Last chance to get statistics from the accelerator */
+ hooks->get_stats(np->netdev, &np->stats);
+
+ spin_unlock_irqrestore(&accelerator->vif_states_lock,
+ *lock_flags);
+
+ rc = hooks->remove(dev);
+
+ spin_lock_irqsave(&accelerator->vif_states_lock, *lock_flags);
+ }
+
+
+ return rc;
+}
+
+
int netfront_accelerator_call_remove(struct netfront_info *np,
- struct xenbus_device *dev)
-{
- struct netfront_accelerator *accelerator = np->accelerator;
- struct netfront_accel_vif_state *tmp_vif_state;
- struct netfront_accel_hooks *hooks;
- unsigned flags;
- int rc = 0;
-
- /*
- * Call the remove accelerator hook. The use count for the
- * accelerator's hooks is incremented for the duration of the
- * call to prevent the accelerator being able to modify the
- * hooks in the middle (by, for example, unloading)
- */
- spin_lock_irqsave(&np->accelerator->vif_states_lock, flags);
- hooks = np->accel_vif_state.hooks;
-
- /*
- * Remove this vif_state from the accelerator's list
- */
- list_for_each_entry(tmp_vif_state, &accelerator->vif_states, link) {
+ struct xenbus_device *dev)
+{
+ struct netfront_accelerator *accelerator;
+ struct netfront_accel_vif_state *tmp_vif_state;
+ unsigned flags;
+ int rc = 0;
+
+ mutex_lock(&accelerator_mutex);
+
+ /* Check that we've got a device that was accelerated */
+ if (np->accelerator == NULL)
+ goto out;
+
+ accelerator = np->accelerator;
+
+ spin_lock_irqsave(&accelerator->vif_states_lock, flags);
+
+ list_for_each_entry(tmp_vif_state, &accelerator->vif_states,
+ link) {
if (tmp_vif_state == &np->accel_vif_state) {
list_del(&np->accel_vif_state.link);
break;
}
}
-
- if (hooks) {
- kref_get(&np->accel_vif_state.vif_kref);
- spin_unlock_irqrestore
- (&np->accelerator->vif_states_lock, flags);
-
- rc = np->accel_vif_state.hooks->remove(dev);
-
- kref_put(&np->accel_vif_state.vif_kref,
- vif_kref_release);
-
- spin_lock_irqsave(&np->accelerator->vif_states_lock,
- flags);
-
- /*
- * Try and do the opposite of accelerator_probe_new_vif
- * to ensure there's no state pointing back at the
- * netdev
- */
- accelerator_remove_single_hook(accelerator,
- &np->accel_vif_state);
-
- /*
- * Remove the reference taken when the vif_state hooks
- * were set, must be called without lock held
- */
- spin_unlock_irqrestore(&accelerator->vif_states_lock,
- flags);
- kref_put(&np->accel_vif_state.vif_kref,
- vif_kref_release);
- } else {
- spin_unlock_irqrestore(&np->accelerator->vif_states_lock,
- flags);
- }
-
+
+ rc = do_remove(np, dev, &flags);
+
+ np->accelerator = NULL;
+
+ spin_unlock_irqrestore(&accelerator->vif_states_lock, flags);
+ out:
+ mutex_unlock(&accelerator_mutex);
return rc;
}
-
-
-int netfront_accelerator_call_suspend(struct netfront_info *np,
- struct xenbus_device *dev)
-{
+
+
+int netfront_accelerator_suspend(struct netfront_info *np,
+ struct xenbus_device *dev)
+{
+ unsigned flags;
+ int rc = 0;
+
+ mutex_lock(&accelerator_mutex);
+
+ /* Check that we've got a device that was accelerated */
+ if (np->accelerator == NULL)
+ goto out;
+
+ /*
+ * Call the remove accelerator hook, but leave the vif_state
+ * on the accelerator's list in case there is a suspend_cancel.
+ */
+ spin_lock_irqsave(&np->accelerator->vif_states_lock, flags);
+
+ rc = do_remove(np, dev, &flags);
+
+ spin_unlock_irqrestore(&np->accelerator->vif_states_lock, flags);
+ out:
+ mutex_unlock(&accelerator_mutex);
+ return rc;
+}
+
+
+int netfront_accelerator_suspend_cancel(struct netfront_info *np,
+ struct xenbus_device *dev)
+{
+ struct netfront_accel_vif_state *accel_vif_state = NULL;
+
+ mutex_lock(&accelerator_mutex);
+
+ /* Check that we've got a device that was accelerated */
+ if (np->accelerator == NULL)
+ goto out;
+
+ /* Find the vif_state from the accelerator's list */
+ list_for_each_entry(accel_vif_state, &np->accelerator->vif_states,
+ link) {
+ if (accel_vif_state->dev == dev) {
+ BUG_ON(accel_vif_state != &np->accel_vif_state);
+
+ /*
+ * Kick things off again to restore
+ * acceleration as it was before suspend
+ */
+ accelerator_probe_new_vif(np, dev, np->accelerator);
+
+ break;
+ }
+ }
+
+ out:
+ mutex_unlock(&accelerator_mutex);
+ return 0;
+}
+
+
+void netfront_accelerator_resume(struct netfront_info *np,
+ struct xenbus_device *dev)
+{
+ struct netfront_accel_vif_state *accel_vif_state = NULL;
+ spinlock_t *vif_states_lock;
+ unsigned flags;
+
+ mutex_lock(&accelerator_mutex);
+
+ /* Check that we've got a device that was accelerated */
+ if(np->accelerator == NULL)
+ goto out;
+
+ /* Find the vif_state from the accelerator's list */
+ list_for_each_entry(accel_vif_state, &np->accelerator->vif_states,
+ link) {
+ if (accel_vif_state->dev == dev) {
+ BUG_ON(accel_vif_state != &np->accel_vif_state);
+
+ vif_states_lock = &np->accelerator->vif_states_lock;
+ spin_lock_irqsave(vif_states_lock, flags);
+
+ /*
+ * Remove it from the accelerator's list so
+ * state is consistent for probing new vifs
+ * when they get connected
+ */
+ list_del(&accel_vif_state->link);
+ np->accelerator = NULL;
+
+ spin_unlock_irqrestore(vif_states_lock, flags);
+
+ break;
+ }
+ }
+
+ out:
+ mutex_unlock(&accelerator_mutex);
+ return;
+}
+
+
+int netfront_check_accelerator_queue_ready(struct net_device *dev,
+ struct netfront_info *np)
+{
+ struct netfront_accelerator *accelerator;
struct netfront_accel_hooks *hooks;
- unsigned flags;
- int rc = 0;
-
- IPRINTK("netfront_accelerator_call_suspend\n");
-
- /*
- * Call the suspend accelerator hook. The use count for the
- * accelerator's hooks is incremented for the duration of
- * the call to prevent the accelerator being able to modify
- * the hooks in the middle (by, for example, unloading)
- */
- if (np->accel_vif_state.hooks) {
- spin_lock_irqsave(&np->accelerator->vif_states_lock, flags);
+ int rc = 1;
+ unsigned flags;
+
+ accelerator = np->accelerator;
+
+ /* Call the check_ready accelerator hook. */
+ if (np->accel_vif_state.hooks && accelerator) {
+ spin_lock_irqsave(&accelerator->vif_states_lock, flags);
hooks = np->accel_vif_state.hooks;
- if (hooks) {
- kref_get(&np->accel_vif_state.vif_kref);
- spin_unlock_irqrestore
- (&np->accelerator->vif_states_lock, flags);
-
- rc = np->accel_vif_state.hooks->suspend(dev);
-
- kref_put(&np->accel_vif_state.vif_kref,
- vif_kref_release);
- } else {
- spin_unlock_irqrestore
- (&np->accelerator->vif_states_lock, flags);
- }
- }
+ if (hooks && np->accelerator == accelerator)
+ rc = np->accel_vif_state.hooks->check_ready(dev);
+ spin_unlock_irqrestore(&accelerator->vif_states_lock, flags);
+ }
+
return rc;
-}
-
-
-int netfront_accelerator_call_suspend_cancel(struct netfront_info *np,
- struct xenbus_device *dev)
-{
- struct netfront_accel_hooks *hooks;
- unsigned flags;
- int rc = 0;
-
- IPRINTK(" netfront_accelerator_call_suspend_cancel\n");
-
- /*
- * Call the suspend_cancel accelerator hook. The use count
- * for the accelerator's hooks is incremented for the
- * duration of the call to prevent the accelerator being able
- * to modify the hooks in the middle (by, for example,
- * unloading)
- */
- if (np->accel_vif_state.hooks) {
- spin_lock_irqsave(&np->accelerator->vif_states_lock, flags);
- hooks = np->accel_vif_state.hooks;
- if (hooks) {
- kref_get(&np->accel_vif_state.vif_kref);
- spin_unlock_irqrestore
- (&np->accelerator->vif_states_lock, flags);
-
- rc = np->accel_vif_state.hooks->suspend_cancel(dev);
-
- kref_put(&np->accel_vif_state.vif_kref,
- vif_kref_release);
- } else {
- spin_unlock_irqrestore
- (&np->accelerator->vif_states_lock, flags);
- }
- }
- return rc;
-}
-
-
-int netfront_accelerator_call_resume(struct netfront_info *np,
- struct xenbus_device *dev)
-{
- struct netfront_accel_hooks *hooks;
- unsigned flags;
- int rc = 0;
-
- /*
- * Call the resume accelerator hook. The use count for the
- * accelerator's hooks is incremented for the duration of
- * the call to prevent the accelerator being able to modify
- * the hooks in the middle (by, for example, unloading)
- */
- if (np->accel_vif_state.hooks) {
- spin_lock_irqsave(&np->accelerator->vif_states_lock, flags);
- hooks = np->accel_vif_state.hooks;
- if (hooks) {
- kref_get(&np->accel_vif_state.vif_kref);
- spin_unlock_irqrestore
- (&np->accelerator->vif_states_lock, flags);
-
- rc = np->accel_vif_state.hooks->resume(dev);
-
- kref_put(&np->accel_vif_state.vif_kref,
- vif_kref_release);
- } else {
- spin_unlock_irqrestore
- (&np->accelerator->vif_states_lock, flags);
- }
- }
- return rc;
-}
-
-
-void netfront_accelerator_call_backend_changed(struct netfront_info *np,
- struct xenbus_device *dev,
- enum xenbus_state backend_state)
-{
- struct netfront_accel_hooks *hooks;
- unsigned flags;
-
- /*
- * Call the backend_changed accelerator hook. The use count
- * for the accelerator's hooks is incremented for the duration
- * of the call to prevent the accelerator being able to modify
- * the hooks in the middle (by, for example, unloading)
- */
- if (np->accel_vif_state.hooks) {
- spin_lock_irqsave(&np->accelerator->vif_states_lock, flags);
- hooks = np->accel_vif_state.hooks;
- if (hooks) {
- kref_get(&np->accel_vif_state.vif_kref);
- spin_unlock_irqrestore
- (&np->accelerator->vif_states_lock, flags);
-
- np->accel_vif_state.hooks->backend_changed
- (dev, backend_state);
-
- kref_put(&np->accel_vif_state.vif_kref,
- vif_kref_release);
- } else {
- spin_unlock_irqrestore
- (&np->accelerator->vif_states_lock, flags);
- }
- }
}
void netfront_accelerator_call_stop_napi_irq(struct netfront_info *np,
struct net_device *dev)
{
+ struct netfront_accelerator *accelerator;
struct netfront_accel_hooks *hooks;
unsigned flags;
- /*
- * Call the stop_napi_interrupts accelerator hook. The use
- * count for the accelerator's hooks is incremented for the
- * duration of the call to prevent the accelerator being able
- * to modify the hooks in the middle (by, for example,
- * unloading)
- */
-
- if (np->accel_vif_state.hooks) {
- spin_lock_irqsave(&np->accelerator->vif_states_lock, flags);
+ accelerator = np->accelerator;
+
+ /* Call the stop_napi_interrupts accelerator hook. */
+ if (np->accel_vif_state.hooks && accelerator != NULL) {
+ spin_lock_irqsave(&accelerator->vif_states_lock, flags);
hooks = np->accel_vif_state.hooks;
- if (hooks) {
- kref_get(&np->accel_vif_state.vif_kref);
- spin_unlock_irqrestore
- (&np->accelerator->vif_states_lock, flags);
-
+ if (hooks && np->accelerator == accelerator)
np->accel_vif_state.hooks->stop_napi_irq(dev);
-
- kref_put(&np->accel_vif_state.vif_kref,
- vif_kref_release);
- } else {
- spin_unlock_irqrestore
- (&np->accelerator->vif_states_lock, flags);
- }
- }
-}
-
-
-/*
- * Once all users of hooks have kref_put()'d we can signal that it's
- * safe to unload
- */
-static void accel_kref_release(struct kref *ref)
-{
- struct netfront_accelerator *accelerator =
- container_of(ref, struct netfront_accelerator, accel_kref);
+ spin_unlock_irqrestore(&accelerator->vif_states_lock, flags);
+ }
+}
+
+
+int netfront_accelerator_call_get_stats(struct netfront_info *np,
+ struct net_device *dev)
+{
+ struct netfront_accelerator *accelerator;
struct netfront_accel_hooks *hooks;
unsigned flags;
-
- DPRINTK("%p\n", accelerator);
-
- /* Signal that all users of hooks are done */
- up(&accelerator->exit_semaphore);
-
- spin_lock_irqsave(&accelerator->vif_states_lock, flags);
- if (accelerator->need_probe) {
- hooks = accelerator->need_probe;
- accelerator->need_probe = NULL;
- accelerator_probe_vifs(accelerator, hooks, flags);
- }
- else
- accelerator->ready_for_probe = 1;
-
- spin_unlock_irqrestore(&accelerator->vif_states_lock, flags);
-}
-
-
-static void vif_kref_release(struct kref *ref)
-{
- struct netfront_accel_vif_state *vif_state =
- container_of(ref, struct netfront_accel_vif_state, vif_kref);
- struct netfront_accel_hooks *hooks;
- unsigned flags;
-
- DPRINTK("%p\n", vif_state);
-
- /*
- * Now that this vif has finished using the hooks, it can
- * decrement the accelerator's global copy ref count
- */
- kref_put(&vif_state->np->accelerator->accel_kref, accel_kref_release);
-
- spin_lock_irqsave(&vif_state->np->accelerator->vif_states_lock, flags);
- if (vif_state->need_probe) {
- hooks = vif_state->need_probe;
- vif_state->need_probe = NULL;
- spin_unlock_irqrestore
- (&vif_state->np->accelerator->vif_states_lock, flags);
- hooks->new_device(vif_state->np->netdev, vif_state->dev);
- } else {
- vif_state->ready_for_probe = 1;
- spin_unlock_irqrestore
- (&vif_state->np->accelerator->vif_states_lock, flags);
- }
-}
-
+ int rc = 0;
+
+ accelerator = np->accelerator;
+
+ /* Call the get_stats accelerator hook. */
+ if (np->accel_vif_state.hooks && accelerator != NULL) {
+ spin_lock_irqsave(&accelerator->vif_states_lock, flags);
+ hooks = np->accel_vif_state.hooks;
+ if (hooks && np->accelerator == accelerator)
+ rc = np->accel_vif_state.hooks->get_stats(dev,
+ &np->stats);
+ spin_unlock_irqrestore(&accelerator->vif_states_lock, flags);
+ }
+ return rc;
+}
+
diff -r 0edeca96e7d7 -r bac78e892e42 drivers/xen/netfront/netfront.c
--- a/drivers/xen/netfront/netfront.c Tue Oct 30 13:42:54 2007 -0600
+++ b/drivers/xen/netfront/netfront.c Tue Oct 30 15:38:11 2007 -0600
@@ -308,14 +308,14 @@ static int netfront_suspend(struct xenbu
static int netfront_suspend(struct xenbus_device *dev)
{
struct netfront_info *info = dev->dev.driver_data;
- return netfront_accelerator_call_suspend(info, dev);
+ return netfront_accelerator_suspend(info, dev);
}
static int netfront_suspend_cancel(struct xenbus_device *dev)
{
struct netfront_info *info = dev->dev.driver_data;
- return netfront_accelerator_call_suspend_cancel(info, dev);
+ return netfront_accelerator_suspend_cancel(info, dev);
}
@@ -331,7 +331,7 @@ static int netfront_resume(struct xenbus
DPRINTK("%s\n", dev->nodename);
- netfront_accelerator_call_resume(info, dev);
+ netfront_accelerator_resume(info, dev);
netif_disconnect_backend(info);
return 0;
@@ -549,8 +549,6 @@ static void backend_changed(struct xenbu
xenbus_frontend_closed(dev);
break;
}
-
- netfront_accelerator_call_backend_changed(np, dev, backend_state);
}
/** Send a packet on a net device to encourage switches to learn the
@@ -595,12 +593,12 @@ static inline void network_maybe_wake_tx
if (unlikely(netif_queue_stopped(dev)) &&
netfront_tx_slot_available(np) &&
likely(netif_running(dev)) &&
- netfront_check_accelerator_queue_busy(dev, np))
+ netfront_check_accelerator_queue_ready(dev, np))
netif_wake_queue(dev);
}
-int netfront_check_queue_busy(struct net_device *dev)
+int netfront_check_queue_ready(struct net_device *dev)
{
struct netfront_info *np = netdev_priv(dev);
@@ -608,7 +606,7 @@ int netfront_check_queue_busy(struct net
netfront_tx_slot_available(np) &&
likely(netif_running(dev));
}
-EXPORT_SYMBOL(netfront_check_queue_busy);
+EXPORT_SYMBOL(netfront_check_queue_ready);
static int network_open(struct net_device *dev)
@@ -1507,7 +1505,7 @@ err:
* fast path is likewise
*/
accel_more_to_do =
- np->accel_vif_state.hooks->start_napi_irq(dev);
+ np->accel_vif_state.hooks->start_napi_irq(dev);
}
if (!more_to_do && !accel_more_to_do)
@@ -1674,6 +1672,8 @@ static struct net_device_stats *network_
static struct net_device_stats *network_get_stats(struct net_device *dev)
{
struct netfront_info *np = netdev_priv(dev);
+
+ netfront_accelerator_call_get_stats(np, dev);
return &np->stats;
}
@@ -2210,6 +2210,8 @@ static void __exit netif_exit(void)
unregister_inetaddr_notifier(¬ifier_inetdev);
+ netif_exit_accel();
+
return xenbus_unregister_driver(&netfront);
}
module_exit(netif_exit);
diff -r 0edeca96e7d7 -r bac78e892e42 drivers/xen/netfront/netfront.h
--- a/drivers/xen/netfront/netfront.h Tue Oct 30 13:42:54 2007 -0600
+++ b/drivers/xen/netfront/netfront.h Tue Oct 30 15:38:11 2007 -0600
@@ -37,7 +37,6 @@
#include <linux/netdevice.h>
#include <linux/skbuff.h>
#include <linux/list.h>
-#include <linux/kref.h>
#define NET_TX_RING_SIZE __RING_SIZE((struct netif_tx_sring *)0, PAGE_SIZE)
#define NET_RX_RING_SIZE __RING_SIZE((struct netif_rx_sring *)0, PAGE_SIZE)
@@ -60,19 +59,9 @@ struct netfront_accel_hooks {
*/
int (*new_device)(struct net_device *net_dev, struct xenbus_device
*dev);
/*
- * suspend, suspend_cancel, resume, remove: Equivalent to the
- * normal xenbus_* callbacks
- */
- int (*suspend)(struct xenbus_device *dev);
- int (*suspend_cancel)(struct xenbus_device *dev);
- int (*resume)(struct xenbus_device *dev);
+ * remove: Opposite of new_device
+ */
int (*remove)(struct xenbus_device *dev);
- /*
- * backend_changed: Callback from watch based on backend's
- * xenbus state changing
- */
- void (*backend_changed)(struct xenbus_device *dev,
- enum xenbus_state backend_state);
/*
* The net_device is being polled, check the accelerated
* hardware for any pending packets
@@ -95,13 +84,18 @@ struct netfront_accel_hooks {
* Called before re-enabling the TX queue to check the fast
* path has slots too
*/
- int (*check_busy)(struct net_device *dev);
+ int (*check_ready)(struct net_device *dev);
+ /*
+ * Get the fastpath network statistics
+ */
+ int (*get_stats)(struct net_device *dev,
+ struct net_device_stats *stats);
};
/* Version of API/protocol for communication between netfront and
acceleration plugin supported */
-#define NETFRONT_ACCEL_VERSION 0x00010000
+#define NETFRONT_ACCEL_VERSION 0x00010002
/*
* Per-netfront device state for the accelerator. This is used to
@@ -114,14 +108,6 @@ struct netfront_accel_vif_state {
struct xenbus_device *dev;
struct netfront_info *np;
struct netfront_accel_hooks *hooks;
-
- /*
- * Protect against removal of hooks while in use.
- */
- struct kref vif_kref;
-
- unsigned ready_for_probe;
- struct netfront_accel_hooks *need_probe;
};
/*
@@ -142,10 +128,7 @@ struct netfront_accelerator {
char *frontend;
/* The hooks into the accelerator plugin module */
struct netfront_accel_hooks *hooks;
- /*
- * Protect against removal of hooks while in use.
- */
- struct kref accel_kref;
+
/*
* List of per-netfront device state (struct
* netfront_accel_vif_state) for each netfront device that is
@@ -153,14 +136,6 @@ struct netfront_accelerator {
*/
struct list_head vif_states;
spinlock_t vif_states_lock;
- /*
- * Semaphore to signal that all users of this accelerator have
- * finished using it before module is unloaded
- */
- struct semaphore exit_semaphore;
-
- unsigned ready_for_probe;
- struct netfront_accel_hooks *need_probe;
};
struct netfront_info {
@@ -251,18 +226,15 @@ extern void netfront_accelerator_ready(c
*
* frontend: the string describing the accelerator. Must match the
* one passed to netfront_accelerator_loaded()
- *
- * wait: 1 => wait for all users of module to complete before
- * returning, thus making it safe to unload on return
*/
-extern void netfront_accelerator_stop(const char *frontend, int wait);
+extern void netfront_accelerator_stop(const char *frontend);
/*
* Called by an accelerator before waking the net device's TX queue to
* ensure the slow path has available slots. Returns true if OK to
* wake, false if still busy
*/
-extern int netfront_check_queue_busy(struct net_device *net_dev);
+extern int netfront_check_queue_ready(struct net_device *net_dev);
@@ -274,28 +246,26 @@ extern int netfront_check_queue_busy(str
* if still busy
*/
extern
-int netfront_check_accelerator_queue_busy(struct net_device *dev,
- struct netfront_info *np);
+int netfront_check_accelerator_queue_ready(struct net_device *dev,
+ struct netfront_info *np);
extern
int netfront_accelerator_call_remove(struct netfront_info *np,
struct xenbus_device *dev);
extern
-int netfront_accelerator_call_suspend(struct netfront_info *np,
- struct xenbus_device *dev);
-extern
-int netfront_accelerator_call_suspend_cancel(struct netfront_info *np,
- struct xenbus_device *dev);
-extern
-int netfront_accelerator_call_resume(struct netfront_info *np,
- struct xenbus_device *dev);
-extern
-void netfront_accelerator_call_backend_changed(struct netfront_info *np,
- struct xenbus_device *dev,
- enum xenbus_state backend_state);
+int netfront_accelerator_suspend(struct netfront_info *np,
+ struct xenbus_device *dev);
+extern
+int netfront_accelerator_suspend_cancel(struct netfront_info *np,
+ struct xenbus_device *dev);
+extern
+void netfront_accelerator_resume(struct netfront_info *np,
+ struct xenbus_device *dev);
extern
void netfront_accelerator_call_stop_napi_irq(struct netfront_info *np,
struct net_device *dev);
-
+extern
+int netfront_accelerator_call_get_stats(struct netfront_info *np,
+ struct net_device *dev);
extern
int netfront_load_accelerator(struct netfront_info *np,
struct xenbus_device *dev,
@@ -303,6 +273,8 @@ int netfront_load_accelerator(struct net
extern
void netif_init_accel(void);
+extern
+void netif_exit_accel(void);
extern
void init_accelerator_vif(struct netfront_info *np,
diff -r 0edeca96e7d7 -r bac78e892e42 drivers/xen/xenbus/xenbus_probe.c
--- a/drivers/xen/xenbus/xenbus_probe.c Tue Oct 30 13:42:54 2007 -0600
+++ b/drivers/xen/xenbus/xenbus_probe.c Tue Oct 30 15:38:11 2007 -0600
@@ -69,7 +69,7 @@ static unsigned long xen_store_mfn;
extern struct mutex xenwatch_mutex;
-static ATOMIC_NOTIFIER_HEAD(xenstore_chain);
+static BLOCKING_NOTIFIER_HEAD(xenstore_chain);
static void wait_for_devices(struct xenbus_driver *xendrv);
@@ -825,7 +825,7 @@ int register_xenstore_notifier(struct no
if (xenstored_ready > 0)
ret = nb->notifier_call(nb, 0, NULL);
else
- atomic_notifier_chain_register(&xenstore_chain, nb);
+ blocking_notifier_chain_register(&xenstore_chain, nb);
return ret;
}
@@ -833,7 +833,7 @@ EXPORT_SYMBOL_GPL(register_xenstore_noti
void unregister_xenstore_notifier(struct notifier_block *nb)
{
- atomic_notifier_chain_unregister(&xenstore_chain, nb);
+ blocking_notifier_chain_unregister(&xenstore_chain, nb);
}
EXPORT_SYMBOL_GPL(unregister_xenstore_notifier);
@@ -848,7 +848,7 @@ void xenbus_probe(void *unused)
xenbus_backend_probe_and_watch();
/* Notify others that xenstore is up */
- atomic_notifier_call_chain(&xenstore_chain, 0, NULL);
+ blocking_notifier_call_chain(&xenstore_chain, 0, NULL);
}
diff -r 0edeca96e7d7 -r bac78e892e42 include/asm-i386/mach-xen/asm/page.h
--- a/include/asm-i386/mach-xen/asm/page.h Tue Oct 30 13:42:54 2007 -0600
+++ b/include/asm-i386/mach-xen/asm/page.h Tue Oct 30 15:38:11 2007 -0600
@@ -36,14 +36,6 @@
#include <asm/bug.h>
#include <xen/interface/xen.h>
#include <xen/features.h>
-
-#define arch_free_page(_page,_order) \
-({ int foreign = PageForeign(_page); \
- if (foreign) \
- PageForeignDestructor(_page); \
- foreign; \
-})
-#define HAVE_ARCH_FREE_PAGE
#ifdef CONFIG_X86_USE_3DNOW
diff -r 0edeca96e7d7 -r bac78e892e42 include/asm-ia64/page.h
--- a/include/asm-ia64/page.h Tue Oct 30 13:42:54 2007 -0600
+++ b/include/asm-ia64/page.h Tue Oct 30 15:38:11 2007 -0600
@@ -239,15 +239,6 @@ get_order (unsigned long size)
#include <asm/maddr.h>
-#define arch_free_page(_page, _order) \
-({ \
- int foreign = PageForeign(_page); \
- if (foreign) \
- PageForeignDestructor(_page); \
- foreign; \
-})
-#define HAVE_ARCH_FREE_PAGE
-
#endif /* CONFIG_XEN */
#endif /* __ASSEMBLY__ */
diff -r 0edeca96e7d7 -r bac78e892e42 include/asm-powerpc/page.h
--- a/include/asm-powerpc/page.h Tue Oct 30 13:42:54 2007 -0600
+++ b/include/asm-powerpc/page.h Tue Oct 30 15:38:11 2007 -0600
@@ -194,15 +194,6 @@ struct vm_area_struct;
struct vm_area_struct;
extern const char *arch_vma_name(struct vm_area_struct *vma);
-#define arch_free_page(_page, _order) \
-({ \
- int foreign = PageForeign(_page); \
- if (foreign) \
- PageForeignDestructor(_page); \
- foreign; \
-})
-#define HAVE_ARCH_FREE_PAGE
-
#include <asm-generic/memory_model.h>
#endif /* __ASSEMBLY__ */
diff -r 0edeca96e7d7 -r bac78e892e42 include/asm-um/page.h
--- a/include/asm-um/page.h Tue Oct 30 13:42:54 2007 -0600
+++ b/include/asm-um/page.h Tue Oct 30 15:38:11 2007 -0600
@@ -114,7 +114,7 @@ extern struct page *arch_validate(struct
extern struct page *arch_validate(struct page *page, gfp_t mask, int order);
#define HAVE_ARCH_VALIDATE
-extern int arch_free_page(struct page *page, int order);
+extern void arch_free_page(struct page *page, int order);
#define HAVE_ARCH_FREE_PAGE
#include <asm-generic/memory_model.h>
diff -r 0edeca96e7d7 -r bac78e892e42 include/asm-x86_64/mach-xen/asm/page.h
--- a/include/asm-x86_64/mach-xen/asm/page.h Tue Oct 30 13:42:54 2007 -0600
+++ b/include/asm-x86_64/mach-xen/asm/page.h Tue Oct 30 15:38:11 2007 -0600
@@ -15,14 +15,6 @@
* below. The preprocessor will warn if the two definitions aren't identical.
*/
#define _PAGE_PRESENT 0x001
-
-#define arch_free_page(_page,_order) \
-({ int foreign = PageForeign(_page); \
- if (foreign) \
- PageForeignDestructor(_page); \
- foreign; \
-})
-#define HAVE_ARCH_FREE_PAGE
/* PAGE_SHIFT determines the page size */
#define PAGE_SHIFT 12
diff -r 0edeca96e7d7 -r bac78e892e42 include/asm-x86_64/mach-xen/asm/pgtable.h
--- a/include/asm-x86_64/mach-xen/asm/pgtable.h Tue Oct 30 13:42:54 2007 -0600
+++ b/include/asm-x86_64/mach-xen/asm/pgtable.h Tue Oct 30 15:38:11 2007 -0600
@@ -505,7 +505,7 @@ static inline pte_t pte_modify(pte_t pte
#define __swp_type(x) (((x).val >> 1) & 0x3f)
#define __swp_offset(x) ((x).val >> 8)
#define __swp_entry(type, offset) ((swp_entry_t) { ((type) << 1) |
((offset) << 8) })
-#define __pte_to_swp_entry(pte) ((swp_entry_t) { pte_val(pte) })
+#define __pte_to_swp_entry(pte) ((swp_entry_t) { __pte_val(pte)
})
#define __swp_entry_to_pte(x) ((pte_t) { (x).val })
extern spinlock_t pgd_lock;
diff -r 0edeca96e7d7 -r bac78e892e42 include/linux/gfp.h
--- a/include/linux/gfp.h Tue Oct 30 13:42:54 2007 -0600
+++ b/include/linux/gfp.h Tue Oct 30 15:38:11 2007 -0600
@@ -99,11 +99,7 @@ static inline int gfp_zone(gfp_t gfp)
*/
#ifndef HAVE_ARCH_FREE_PAGE
-/*
- * If arch_free_page returns non-zero then the generic free_page code can
- * immediately bail: the arch-specific function has done all the work.
- */
-static inline int arch_free_page(struct page *page, int order) { return 0; }
+static inline void arch_free_page(struct page *page, int order) { }
#endif
extern struct page *
diff -r 0edeca96e7d7 -r bac78e892e42 include/xen/evtchn.h
--- a/include/xen/evtchn.h Tue Oct 30 13:42:54 2007 -0600
+++ b/include/xen/evtchn.h Tue Oct 30 15:38:11 2007 -0600
@@ -102,6 +102,7 @@ void evtchn_device_upcall(int port);
void evtchn_device_upcall(int port);
void mask_evtchn(int port);
+void disable_all_local_evtchn(void);
void unmask_evtchn(int port);
#ifdef CONFIG_SMP
diff -r 0edeca96e7d7 -r bac78e892e42 include/xen/interface/callback.h
--- a/include/xen/interface/callback.h Tue Oct 30 13:42:54 2007 -0600
+++ b/include/xen/interface/callback.h Tue Oct 30 15:38:11 2007 -0600
@@ -36,15 +36,39 @@
* @extra_args == Operation-specific extra arguments (NULL if none).
*/
+/* ia64, x86: Callback for event delivery. */
#define CALLBACKTYPE_event 0
+
+/* x86: Failsafe callback when guest state cannot be restored by Xen. */
#define CALLBACKTYPE_failsafe 1
-#define CALLBACKTYPE_syscall 2 /* x86_64 only */
+
+/* x86/64 hypervisor: Syscall by 64-bit guest app ('64-on-64-on-64'). */
+#define CALLBACKTYPE_syscall 2
+
/*
- * sysenter is only available on x86_32 with the
- * supervisor_mode_kernel option enabled.
+ * x86/32 hypervisor: Only available on x86/32 when supervisor_mode_kernel
+ * feature is enabled. Do not use this callback type in new code.
*/
-#define CALLBACKTYPE_sysenter 3
+#define CALLBACKTYPE_sysenter_deprecated 3
+
+/* x86: Callback for NMI delivery. */
#define CALLBACKTYPE_nmi 4
+
+/*
+ * x86: sysenter is only available as follows:
+ * - 32-bit hypervisor: with the supervisor_mode_kernel feature enabled
+ * - 64-bit hypervisor: 32-bit guest applications on Intel CPUs
+ * ('32-on-32-on-64', '32-on-64-on-64')
+ * [nb. also 64-bit guest applications on Intel CPUs
+ * ('64-on-64-on-64'), but syscall is preferred]
+ */
+#define CALLBACKTYPE_sysenter 5
+
+/*
+ * x86/64 hypervisor: Syscall by 32-bit guest app on AMD CPUs
+ * ('32-on-32-on-64', '32-on-64-on-64')
+ */
+#define CALLBACKTYPE_syscall32 7
/*
* Disable event deliver during callback? This flag is ignored for event and
@@ -79,6 +103,11 @@ typedef struct callback_unregister callb
typedef struct callback_unregister callback_unregister_t;
DEFINE_XEN_GUEST_HANDLE(callback_unregister_t);
+#if __XEN_INTERFACE_VERSION__ < 0x00030207
+#undef CALLBACKTYPE_sysenter
+#define CALLBACKTYPE_sysenter CALLBACKTYPE_sysenter_deprecated
+#endif
+
#endif /* __XEN_PUBLIC_CALLBACK_H__ */
/*
diff -r 0edeca96e7d7 -r bac78e892e42 include/xen/interface/domctl.h
--- a/include/xen/interface/domctl.h Tue Oct 30 13:42:54 2007 -0600
+++ b/include/xen/interface/domctl.h Tue Oct 30 15:38:11 2007 -0600
@@ -474,11 +474,11 @@ DEFINE_XEN_GUEST_HANDLE(xen_domctl_bind_
/* Bind machine I/O address range -> HVM address range. */
#define XEN_DOMCTL_memory_mapping 39
struct xen_domctl_memory_mapping {
- uint64_t first_gfn; /* first page (hvm guest phys page) in range */
- uint64_t first_mfn; /* first page (machine page) in range */
- uint64_t nr_mfns; /* number of pages in range (>0) */
- uint32_t add_mapping; /* add or remove mapping */
- uint32_t padding; /* padding for 64-bit aligned structure */
+ uint64_aligned_t first_gfn; /* first page (hvm guest phys page) in range */
+ uint64_aligned_t first_mfn; /* first page (machine page) in range */
+ uint64_aligned_t nr_mfns; /* number of pages in range (>0) */
+ uint32_t add_mapping; /* add or remove mapping */
+ uint32_t padding; /* padding for 64-bit aligned structure */
};
typedef struct xen_domctl_memory_mapping xen_domctl_memory_mapping_t;
DEFINE_XEN_GUEST_HANDLE(xen_domctl_memory_mapping_t);
@@ -494,6 +494,50 @@ struct xen_domctl_ioport_mapping {
};
typedef struct xen_domctl_ioport_mapping xen_domctl_ioport_mapping_t;
DEFINE_XEN_GUEST_HANDLE(xen_domctl_ioport_mapping_t);
+
+
+/*
+ * Pin caching type of RAM space for x86 HVM domU.
+ */
+#define XEN_DOMCTL_pin_mem_cacheattr 41
+/* Caching types: these happen to be the same as x86 MTRR/PAT type codes. */
+#define XEN_DOMCTL_MEM_CACHEATTR_UC 0
+#define XEN_DOMCTL_MEM_CACHEATTR_WC 1
+#define XEN_DOMCTL_MEM_CACHEATTR_WT 4
+#define XEN_DOMCTL_MEM_CACHEATTR_WP 5
+#define XEN_DOMCTL_MEM_CACHEATTR_WB 6
+#define XEN_DOMCTL_MEM_CACHEATTR_UCM 7
+struct xen_domctl_pin_mem_cacheattr {
+ uint64_aligned_t start, end;
+ unsigned int type; /* XEN_DOMCTL_MEM_CACHEATTR_* */
+};
+typedef struct xen_domctl_pin_mem_cacheattr xen_domctl_pin_mem_cacheattr_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_pin_mem_cacheattr_t);
+
+
+#define XEN_DOMCTL_set_ext_vcpucontext 42
+#define XEN_DOMCTL_get_ext_vcpucontext 43
+struct xen_domctl_ext_vcpucontext {
+ /* IN: VCPU that this call applies to. */
+ uint32_t vcpu;
+ /*
+ * SET: Size of struct (IN)
+ * GET: Size of struct (OUT)
+ */
+ uint32_t size;
+#if defined(__i386__) || defined(__x86_64__)
+ /* SYSCALL from 32-bit mode and SYSENTER callback information. */
+ /* NB. SYSCALL from 64-bit mode is contained in vcpu_guest_context_t */
+ uint64_aligned_t syscall32_callback_eip;
+ uint64_aligned_t sysenter_callback_eip;
+ uint16_t syscall32_callback_cs;
+ uint16_t sysenter_callback_cs;
+ uint8_t syscall32_disables_events;
+ uint8_t sysenter_disables_events;
+#endif
+};
+typedef struct xen_domctl_ext_vcpucontext xen_domctl_ext_vcpucontext_t;
+DEFINE_XEN_GUEST_HANDLE(xen_domctl_ext_vcpucontext_t);
struct xen_domctl {
@@ -529,6 +573,8 @@ struct xen_domctl {
struct xen_domctl_bind_pt_irq bind_pt_irq;
struct xen_domctl_memory_mapping memory_mapping;
struct xen_domctl_ioport_mapping ioport_mapping;
+ struct xen_domctl_pin_mem_cacheattr pin_mem_cacheattr;
+ struct xen_domctl_ext_vcpucontext ext_vcpucontext;
uint8_t pad[128];
} u;
};
diff -r 0edeca96e7d7 -r bac78e892e42 include/xen/interface/platform.h
--- a/include/xen/interface/platform.h Tue Oct 30 13:42:54 2007 -0600
+++ b/include/xen/interface/platform.h Tue Oct 30 15:38:11 2007 -0600
@@ -174,13 +174,27 @@ typedef struct xenpf_change_freq xenpf_c
typedef struct xenpf_change_freq xenpf_change_freq_t;
DEFINE_XEN_GUEST_HANDLE(xenpf_change_freq_t);
+/*
+ * Get idle times (nanoseconds since boot) for physical CPUs specified in the
+ * @cpumap_bitmap with range [0..@cpumap_nr_cpus-1]. The @idletime array is
+ * indexed by CPU number; only entries with the corresponding @cpumap_bitmap
+ * bit set are written to. On return, @cpumap_bitmap is modified so that any
+ * non-existent CPUs are cleared. Such CPUs have their @idletime array entry
+ * cleared.
+ */
#define XENPF_getidletime 53
struct xenpf_getidletime {
- /* IN variables. */
- uint32_t max_cpus;
+ /* IN/OUT variables */
+ /* IN: CPUs to interrogate; OUT: subset of IN which are present */
+ XEN_GUEST_HANDLE(uint8_t) cpumap_bitmap;
+ /* IN variables */
+ /* Size of cpumap bitmap. */
+ uint32_t cpumap_nr_cpus;
+ /* Must be indexable for every cpu in cpumap_bitmap. */
XEN_GUEST_HANDLE(uint64_t) idletime;
- /* OUT variables. */
- uint32_t nr_cpus;
+ /* OUT variables */
+ /* System time when the idletime snapshots were taken. */
+ uint64_t now;
};
typedef struct xenpf_getidletime xenpf_getidletime_t;
DEFINE_XEN_GUEST_HANDLE(xenpf_getidletime_t);
diff -r 0edeca96e7d7 -r bac78e892e42 include/xen/interface/sysctl.h
--- a/include/xen/interface/sysctl.h Tue Oct 30 13:42:54 2007 -0600
+++ b/include/xen/interface/sysctl.h Tue Oct 30 15:38:11 2007 -0600
@@ -34,18 +34,27 @@
#include "xen.h"
#include "domctl.h"
-#define XEN_SYSCTL_INTERFACE_VERSION 0x00000005
+#define XEN_SYSCTL_INTERFACE_VERSION 0x00000006
/*
* Read console content from Xen buffer ring.
*/
#define XEN_SYSCTL_readconsole 1
struct xen_sysctl_readconsole {
- /* IN variables. */
- uint32_t clear; /* Non-zero -> clear after reading. */
- XEN_GUEST_HANDLE_64(char) buffer; /* Buffer start */
- /* IN/OUT variables. */
- uint32_t count; /* In: Buffer size; Out: Used buffer size */
+ /* IN: Non-zero -> clear after reading. */
+ uint8_t clear;
+ /* IN: Non-zero -> start index specified by @index field. */
+ uint8_t incremental;
+ uint8_t pad0, pad1;
+ /*
+ * IN: Start index for consuming from ring buffer (if @incremental);
+ * OUT: End index after consuming from ring buffer.
+ */
+ uint32_t index;
+ /* IN: Virtual address to write console data. */
+ XEN_GUEST_HANDLE_64(char) buffer;
+ /* IN: Size of buffer; OUT: Bytes written to buffer. */
+ uint32_t count;
};
typedef struct xen_sysctl_readconsole xen_sysctl_readconsole_t;
DEFINE_XEN_GUEST_HANDLE(xen_sysctl_readconsole_t);
@@ -171,7 +180,7 @@ DEFINE_XEN_GUEST_HANDLE(xen_sysctl_debug
/* Get physical CPU information. */
#define XEN_SYSCTL_getcpuinfo 8
struct xen_sysctl_cpuinfo {
- uint64_t idletime;
+ uint64_aligned_t idletime;
};
typedef struct xen_sysctl_cpuinfo xen_sysctl_cpuinfo_t;
DEFINE_XEN_GUEST_HANDLE(xen_sysctl_cpuinfo_t);
@@ -192,7 +201,7 @@ struct xen_sysctl_availheap {
uint32_t max_bitwidth; /* Largest address width (zero if don't care). */
int32_t node; /* NUMA node of interest (-1 for all nodes). */
/* OUT variables. */
- uint64_t avail_bytes; /* Bytes available in the specified region. */
+ uint64_aligned_t avail_bytes;/* Bytes available in the specified region. */
};
typedef struct xen_sysctl_availheap xen_sysctl_availheap_t;
DEFINE_XEN_GUEST_HANDLE(xen_sysctl_availheap_t);
diff -r 0edeca96e7d7 -r bac78e892e42 include/xen/interface/vcpu.h
--- a/include/xen/interface/vcpu.h Tue Oct 30 13:42:54 2007 -0600
+++ b/include/xen/interface/vcpu.h Tue Oct 30 15:38:11 2007 -0600
@@ -179,6 +179,9 @@ typedef struct vcpu_register_vcpu_info v
typedef struct vcpu_register_vcpu_info vcpu_register_vcpu_info_t;
DEFINE_XEN_GUEST_HANDLE(vcpu_register_vcpu_info_t);
+/* Send an NMI to the specified VCPU. @extra_arg == NULL. */
+#define VCPUOP_send_nmi 11
+
#endif /* __XEN_PUBLIC_VCPU_H__ */
/*
diff -r 0edeca96e7d7 -r bac78e892e42 include/xen/interface/xen-compat.h
--- a/include/xen/interface/xen-compat.h Tue Oct 30 13:42:54 2007 -0600
+++ b/include/xen/interface/xen-compat.h Tue Oct 30 15:38:11 2007 -0600
@@ -27,7 +27,7 @@
#ifndef __XEN_PUBLIC_XEN_COMPAT_H__
#define __XEN_PUBLIC_XEN_COMPAT_H__
-#define __XEN_LATEST_INTERFACE_VERSION__ 0x00030206
+#define __XEN_LATEST_INTERFACE_VERSION__ 0x00030207
#if defined(__XEN__) || defined(__XEN_TOOLS__)
/* Xen is built with matching headers and implements the latest interface. */
diff -r 0edeca96e7d7 -r bac78e892e42 mm/page_alloc.c
--- a/mm/page_alloc.c Tue Oct 30 13:42:54 2007 -0600
+++ b/mm/page_alloc.c Tue Oct 30 15:38:11 2007 -0600
@@ -451,8 +451,13 @@ static void __free_pages_ok(struct page
int i;
int reserved = 0;
- if (arch_free_page(page, order))
+#ifdef CONFIG_XEN
+ if (PageForeign(page)) {
+ PageForeignDestructor(page);
return;
+ }
+#endif
+ arch_free_page(page, order);
if (!PageHighMem(page))
debug_check_no_locks_freed(page_address(page),
PAGE_SIZE<<order);
@@ -730,8 +735,13 @@ static void fastcall free_hot_cold_page(
struct per_cpu_pages *pcp;
unsigned long flags;
- if (arch_free_page(page, 0))
+#ifdef CONFIG_XEN
+ if (PageForeign(page)) {
+ PageForeignDestructor(page);
return;
+ }
+#endif
+ arch_free_page(page, 0);
if (PageAnon(page))
page->mapping = NULL;
_______________________________________________
Xen-changelog mailing list
Xen-changelog@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-changelog
|
Home