# HG changeset patch
# User Alex Williamson <alex.williamson@xxxxxx>
# Date 1176147625 21600
# Node ID 1bde28f762a66276008da38e3fb24751f83dc128
# Parent cd3cfddcb055e65c38dcfd9199cefbf64e958dc8
# Parent 400a3dca237e856c53934f4b81137581d727fdb3
merge with xen-unstable.hg
---
linux-2.6-xen-sparse/arch/i386/mm/hypervisor.c | 2
linux-2.6-xen-sparse/include/asm-i386/mach-xen/asm/maddr.h | 36
linux-2.6-xen-sparse/include/asm-x86_64/mach-xen/asm/maddr.h | 8
tools/ioemu/target-i386-dm/exec-dm.c | 25
tools/ioemu/vl.c | 162 +-
tools/ioemu/vl.h | 15
tools/libxc/xc_domain_restore.c | 370 +++---
tools/libxc/xc_hvm_save.c | 63 -
tools/libxc/xc_linux_save.c | 580
+++++-----
tools/libxc/xc_private.c | 12
tools/libxc/xc_private.h | 12
tools/libxc/xenctrl.h | 15
unmodified_drivers/linux-2.6/compat-include/xen/platform-compat.h | 4
unmodified_drivers/linux-2.6/platform-pci/platform-pci.c | 4
xen/arch/x86/mm/shadow/common.c | 3
xen/arch/x86/traps.c | 15
xen/arch/x86/x86_32/entry.S | 1
xen/arch/x86/x86_64/compat/entry.S | 4
xen/arch/x86/x86_64/entry.S | 6
xen/arch/x86/x86_64/traps.c | 12
xen/arch/x86/x86_emulate.c | 2
xen/common/page_alloc.c | 14
xen/include/asm-x86/desc.h | 2
xen/include/asm-x86/processor.h | 1
xen/include/asm-x86/x86_32/asm_defns.h | 10
xen/include/asm-x86/x86_64/asm_defns.h | 10
xen/include/public/domctl.h | 18
27 files changed, 779 insertions(+), 627 deletions(-)
diff -r cd3cfddcb055 -r 1bde28f762a6
linux-2.6-xen-sparse/arch/i386/mm/hypervisor.c
--- a/linux-2.6-xen-sparse/arch/i386/mm/hypervisor.c Mon Apr 09 13:39:35
2007 -0600
+++ b/linux-2.6-xen-sparse/arch/i386/mm/hypervisor.c Mon Apr 09 13:40:25
2007 -0600
@@ -353,6 +353,7 @@ int xen_create_contiguous_region(
return success ? 0 : -ENOMEM;
}
+EXPORT_SYMBOL_GPL(xen_create_contiguous_region);
void xen_destroy_contiguous_region(unsigned long vstart, unsigned int order)
{
@@ -437,6 +438,7 @@ void xen_destroy_contiguous_region(unsig
balloon_unlock(flags);
}
+EXPORT_SYMBOL_GPL(xen_destroy_contiguous_region);
#ifdef __i386__
int write_ldt_entry(void *ldt, int entry, __u32 entry_a, __u32 entry_b)
diff -r cd3cfddcb055 -r 1bde28f762a6
linux-2.6-xen-sparse/include/asm-i386/mach-xen/asm/maddr.h
--- a/linux-2.6-xen-sparse/include/asm-i386/mach-xen/asm/maddr.h Mon Apr
09 13:39:35 2007 -0600
+++ b/linux-2.6-xen-sparse/include/asm-i386/mach-xen/asm/maddr.h Mon Apr
09 13:40:25 2007 -0600
@@ -153,23 +153,6 @@ static inline paddr_t pte_machine_to_phy
}
#endif
-#else /* !CONFIG_XEN */
-
-#define pfn_to_mfn(pfn) (pfn)
-#define mfn_to_pfn(mfn) (mfn)
-#define mfn_to_local_pfn(mfn) (mfn)
-#define set_phys_to_machine(pfn, mfn) ((void)0)
-#define phys_to_machine_mapping_valid(pfn) (1)
-#define phys_to_machine(phys) ((maddr_t)(phys))
-#define machine_to_phys(mach) ((paddr_t)(mach))
-
-#endif /* !CONFIG_XEN */
-
-/* VIRT <-> MACHINE conversion */
-#define virt_to_machine(v) (phys_to_machine(__pa(v)))
-#define virt_to_mfn(v) (pfn_to_mfn(__pa(v) >> PAGE_SHIFT))
-#define mfn_to_virt(m) (__va(mfn_to_pfn(m) << PAGE_SHIFT))
-
#ifdef CONFIG_X86_PAE
static inline pte_t pfn_pte_ma(unsigned long page_nr, pgprot_t pgprot)
{
@@ -188,4 +171,23 @@ static inline pte_t pfn_pte_ma(unsigned
#define __pte_ma(x) ((pte_t) { (x) } )
+#else /* !CONFIG_XEN */
+
+#define pfn_to_mfn(pfn) (pfn)
+#define mfn_to_pfn(mfn) (mfn)
+#define mfn_to_local_pfn(mfn) (mfn)
+#define set_phys_to_machine(pfn, mfn) ((void)0)
+#define phys_to_machine_mapping_valid(pfn) (1)
+#define phys_to_machine(phys) ((maddr_t)(phys))
+#define machine_to_phys(mach) ((paddr_t)(mach))
+#define pfn_pte_ma(pfn, prot) pfn_pte(pfn, prot)
+#define __pte_ma(x) __pte(x)
+
+#endif /* !CONFIG_XEN */
+
+/* VIRT <-> MACHINE conversion */
+#define virt_to_machine(v) (phys_to_machine(__pa(v)))
+#define virt_to_mfn(v) (pfn_to_mfn(__pa(v) >> PAGE_SHIFT))
+#define mfn_to_virt(m) (__va(mfn_to_pfn(m) << PAGE_SHIFT))
+
#endif /* _I386_MADDR_H */
diff -r cd3cfddcb055 -r 1bde28f762a6
linux-2.6-xen-sparse/include/asm-x86_64/mach-xen/asm/maddr.h
--- a/linux-2.6-xen-sparse/include/asm-x86_64/mach-xen/asm/maddr.h Mon Apr
09 13:39:35 2007 -0600
+++ b/linux-2.6-xen-sparse/include/asm-x86_64/mach-xen/asm/maddr.h Mon Apr
09 13:40:25 2007 -0600
@@ -135,6 +135,9 @@ static inline paddr_t pte_machine_to_phy
return phys;
}
+#define __pte_ma(x) ((pte_t) { (x) } )
+#define pfn_pte_ma(pfn, prot) __pte_ma((((pfn) << PAGE_SHIFT) |
pgprot_val(prot)) & __supported_pte_mask)
+
#else /* !CONFIG_XEN */
#define pfn_to_mfn(pfn) (pfn)
@@ -144,6 +147,8 @@ static inline paddr_t pte_machine_to_phy
#define phys_to_machine_mapping_valid(pfn) (1)
#define phys_to_machine(phys) ((maddr_t)(phys))
#define machine_to_phys(mach) ((paddr_t)(mach))
+#define pfn_pte_ma(pfn, prot) pfn_pte(pfn, prot)
+#define __pte_ma(x) __pte(x)
#endif /* !CONFIG_XEN */
@@ -152,8 +157,5 @@ static inline paddr_t pte_machine_to_phy
#define virt_to_mfn(v) (pfn_to_mfn(__pa(v) >> PAGE_SHIFT))
#define mfn_to_virt(m) (__va(mfn_to_pfn(m) << PAGE_SHIFT))
-#define __pte_ma(x) ((pte_t) { (x) } )
-#define pfn_pte_ma(pfn, prot) __pte_ma((((pfn) << PAGE_SHIFT) |
pgprot_val(prot)) & __supported_pte_mask)
-
#endif /* _X86_64_MADDR_H */
diff -r cd3cfddcb055 -r 1bde28f762a6 tools/ioemu/target-i386-dm/exec-dm.c
--- a/tools/ioemu/target-i386-dm/exec-dm.c Mon Apr 09 13:39:35 2007 -0600
+++ b/tools/ioemu/target-i386-dm/exec-dm.c Mon Apr 09 13:40:25 2007 -0600
@@ -128,11 +128,9 @@ FILE *logfile;
FILE *logfile;
int loglevel;
-
#ifdef MAPCACHE
pthread_mutex_t mapcache_mutex;
#endif
-
void cpu_exec_init(CPUState *env)
{
@@ -427,21 +425,10 @@ int iomem_index(target_phys_addr_t addr)
return 0;
}
-static inline int paddr_is_ram(target_phys_addr_t addr)
-{
- /* Is this guest physical address RAM-backed? */
-#if defined(CONFIG_DM) && (defined(__i386__) || defined(__x86_64__))
- return ((addr < HVM_BELOW_4G_MMIO_START) ||
- (addr >= HVM_BELOW_4G_MMIO_START + HVM_BELOW_4G_MMIO_LENGTH));
-#else
- return (addr < ram_size);
-#endif
-}
-
#if defined(__i386__) || defined(__x86_64__)
#define phys_ram_addr(x) (qemu_map_cache(x))
#elif defined(__ia64__)
-#define phys_ram_addr(x) (phys_ram_base + (x))
+#define phys_ram_addr(x) ((addr < ram_size) ? (phys_ram_base + (x)) : NULL)
#endif
extern unsigned long *logdirty_bitmap;
@@ -481,16 +468,15 @@ void cpu_physical_memory_rw(target_phys_
io_mem_write[io_index][0](io_mem_opaque[io_index], addr,
val);
l = 1;
}
- } else if (paddr_is_ram(addr)) {
+ } else if ((ptr = phys_ram_addr(addr)) != NULL) {
/* Writing to RAM */
- ptr = phys_ram_addr(addr);
memcpy(ptr, buf, l);
if (logdirty_bitmap != NULL) {
/* Record that we have dirtied this frame */
unsigned long pfn = addr >> TARGET_PAGE_BITS;
if (pfn / 8 >= logdirty_bitmap_size) {
- fprintf(logfile, "dirtying pfn %x >= bitmap size %x\n",
- pfn, logdirty_bitmap_size * 8);
+ fprintf(logfile, "dirtying pfn %lx >= bitmap "
+ "size %lx\n", pfn, logdirty_bitmap_size * 8);
} else {
logdirty_bitmap[pfn / HOST_LONG_BITS]
|= 1UL << pfn % HOST_LONG_BITS;
@@ -518,9 +504,8 @@ void cpu_physical_memory_rw(target_phys_
stb_raw(buf, val);
l = 1;
}
- } else if (paddr_is_ram(addr)) {
+ } else if ((ptr = phys_ram_addr(addr)) != NULL) {
/* Reading from RAM */
- ptr = phys_ram_addr(addr);
memcpy(buf, ptr, l);
} else {
/* Neither RAM nor known MMIO space */
diff -r cd3cfddcb055 -r 1bde28f762a6 tools/ioemu/vl.c
--- a/tools/ioemu/vl.c Mon Apr 09 13:39:35 2007 -0600
+++ b/tools/ioemu/vl.c Mon Apr 09 13:40:25 2007 -0600
@@ -5894,7 +5894,32 @@ void suspend(int sig)
suspend_requested = 1;
}
-#if defined(__i386__) || defined(__x86_64__)
+#if defined(MAPCACHE)
+
+#if defined(__i386__)
+#define MAX_MCACHE_SIZE 0x40000000 /* 1GB max for x86 */
+#define MCACHE_BUCKET_SHIFT 16
+#elif defined(__x86_64__)
+#define MAX_MCACHE_SIZE 0x1000000000 /* 64GB max for x86_64 */
+#define MCACHE_BUCKET_SHIFT 20
+#endif
+
+#define MCACHE_BUCKET_SIZE (1UL << MCACHE_BUCKET_SHIFT)
+
+#define BITS_PER_LONG (sizeof(long)*8)
+#define BITS_TO_LONGS(bits) \
+ (((bits)+BITS_PER_LONG-1)/BITS_PER_LONG)
+#define DECLARE_BITMAP(name,bits) \
+ unsigned long name[BITS_TO_LONGS(bits)]
+#define test_bit(bit,map) \
+ (!!((map)[(bit)/BITS_PER_LONG] & (1UL << ((bit)%BITS_PER_LONG))))
+
+struct map_cache {
+ unsigned long paddr_index;
+ uint8_t *vaddr_base;
+ DECLARE_BITMAP(valid_mapping, MCACHE_BUCKET_SIZE>>PAGE_SHIFT);
+};
+
static struct map_cache *mapcache_entry;
static unsigned long nr_buckets;
@@ -5928,69 +5953,85 @@ static int qemu_map_cache_init(void)
return 0;
}
-uint8_t *qemu_map_cache(target_phys_addr_t phys_addr)
-{
- struct map_cache *entry;
- unsigned long address_index = phys_addr >> MCACHE_BUCKET_SHIFT;
- unsigned long address_offset = phys_addr & (MCACHE_BUCKET_SIZE-1);
-
- if (address_index == last_address_index)
- return last_address_vaddr + address_offset;
-
- entry = &mapcache_entry[address_index % nr_buckets];
-
- if (entry->vaddr_base == NULL || entry->paddr_index != address_index) {
- /* We need to remap a bucket. */
- uint8_t *vaddr_base;
- unsigned long pfns[MCACHE_BUCKET_SIZE >> PAGE_SHIFT];
- unsigned int i;
-
- if (entry->vaddr_base != NULL) {
- errno = munmap(entry->vaddr_base, MCACHE_BUCKET_SIZE);
- if (errno) {
- fprintf(logfile, "unmap fails %d\n", errno);
- exit(-1);
- }
- }
-
- for (i = 0; i < MCACHE_BUCKET_SIZE >> PAGE_SHIFT; i++)
- pfns[i] = (address_index << (MCACHE_BUCKET_SHIFT-PAGE_SHIFT)) + i;
-
- vaddr_base = xc_map_foreign_batch(
- xc_handle, domid, PROT_READ|PROT_WRITE,
- pfns, MCACHE_BUCKET_SIZE >> PAGE_SHIFT);
- if (vaddr_base == NULL) {
- fprintf(logfile, "xc_map_foreign_batch error %d\n", errno);
- exit(-1);
- }
-
- entry->vaddr_base = vaddr_base;
- entry->paddr_index = address_index;;
- }
-
- last_address_index = address_index;
- last_address_vaddr = entry->vaddr_base;
-
- return last_address_vaddr + address_offset;
-}
-
-void qemu_invalidate_map_cache(void)
-{
- unsigned long i;
-
- mapcache_lock();
-
- for (i = 0; i < nr_buckets; i++) {
- struct map_cache *entry = &mapcache_entry[i];
-
- if (entry->vaddr_base == NULL)
- continue;
-
+static void qemu_remap_bucket(struct map_cache *entry,
+ unsigned long address_index)
+{
+ uint8_t *vaddr_base;
+ unsigned long pfns[MCACHE_BUCKET_SIZE >> PAGE_SHIFT];
+ unsigned int i, j;
+
+ if (entry->vaddr_base != NULL) {
errno = munmap(entry->vaddr_base, MCACHE_BUCKET_SIZE);
if (errno) {
fprintf(logfile, "unmap fails %d\n", errno);
exit(-1);
}
+ }
+
+ for (i = 0; i < MCACHE_BUCKET_SIZE >> PAGE_SHIFT; i++)
+ pfns[i] = (address_index << (MCACHE_BUCKET_SHIFT-PAGE_SHIFT)) + i;
+
+ vaddr_base = xc_map_foreign_batch(xc_handle, domid, PROT_READ|PROT_WRITE,
+ pfns, MCACHE_BUCKET_SIZE >> PAGE_SHIFT);
+ if (vaddr_base == NULL) {
+ fprintf(logfile, "xc_map_foreign_batch error %d\n", errno);
+ exit(-1);
+ }
+
+ entry->vaddr_base = vaddr_base;
+ entry->paddr_index = address_index;
+
+ for (i = 0; i < MCACHE_BUCKET_SIZE >> PAGE_SHIFT; i += BITS_PER_LONG) {
+ unsigned long word = 0;
+ j = ((i + BITS_PER_LONG) > (MCACHE_BUCKET_SIZE >> PAGE_SHIFT)) ?
+ (MCACHE_BUCKET_SIZE >> PAGE_SHIFT) % BITS_PER_LONG : BITS_PER_LONG;
+ while (j > 0)
+ word = (word << 1) | !(pfns[i + --j] & 0xF0000000UL);
+ entry->valid_mapping[i / BITS_PER_LONG] = word;
+ }
+}
+
+uint8_t *qemu_map_cache(target_phys_addr_t phys_addr)
+{
+ struct map_cache *entry;
+ unsigned long address_index = phys_addr >> MCACHE_BUCKET_SHIFT;
+ unsigned long address_offset = phys_addr & (MCACHE_BUCKET_SIZE-1);
+
+ if (address_index == last_address_index)
+ return last_address_vaddr + address_offset;
+
+ entry = &mapcache_entry[address_index % nr_buckets];
+
+ if (entry->vaddr_base == NULL || entry->paddr_index != address_index ||
+ !test_bit(address_offset>>PAGE_SHIFT, entry->valid_mapping))
+ qemu_remap_bucket(entry, address_index);
+
+ if (!test_bit(address_offset>>PAGE_SHIFT, entry->valid_mapping))
+ return NULL;
+
+ last_address_index = address_index;
+ last_address_vaddr = entry->vaddr_base;
+
+ return last_address_vaddr + address_offset;
+}
+
+void qemu_invalidate_map_cache(void)
+{
+ unsigned long i;
+
+ mapcache_lock();
+
+ for (i = 0; i < nr_buckets; i++) {
+ struct map_cache *entry = &mapcache_entry[i];
+
+ if (entry->vaddr_base == NULL)
+ continue;
+
+ errno = munmap(entry->vaddr_base, MCACHE_BUCKET_SIZE);
+ if (errno) {
+ fprintf(logfile, "unmap fails %d\n", errno);
+ exit(-1);
+ }
entry->paddr_index = 0;
entry->vaddr_base = NULL;
@@ -6001,7 +6042,8 @@ void qemu_invalidate_map_cache(void)
mapcache_unlock();
}
-#endif
+
+#endif /* defined(MAPCACHE) */
int main(int argc, char **argv)
{
diff -r cd3cfddcb055 -r 1bde28f762a6 tools/ioemu/vl.h
--- a/tools/ioemu/vl.h Mon Apr 09 13:39:35 2007 -0600
+++ b/tools/ioemu/vl.h Mon Apr 09 13:40:25 2007 -0600
@@ -161,21 +161,6 @@ extern FILE *logfile;
#define MAPCACHE
-#if defined(__i386__)
-#define MAX_MCACHE_SIZE 0x40000000 /* 1GB max for x86 */
-#define MCACHE_BUCKET_SHIFT 16
-#elif defined(__x86_64__)
-#define MAX_MCACHE_SIZE 0x1000000000 /* 64GB max for x86_64 */
-#define MCACHE_BUCKET_SHIFT 20
-#endif
-
-#define MCACHE_BUCKET_SIZE (1UL << MCACHE_BUCKET_SHIFT)
-
-struct map_cache {
- unsigned long paddr_index;
- uint8_t *vaddr_base;
-};
-
uint8_t *qemu_map_cache(target_phys_addr_t phys_addr);
void qemu_invalidate_map_cache(void);
diff -r cd3cfddcb055 -r 1bde28f762a6 tools/libxc/xc_domain_restore.c
--- a/tools/libxc/xc_domain_restore.c Mon Apr 09 13:39:35 2007 -0600
+++ b/tools/libxc/xc_domain_restore.c Mon Apr 09 13:40:25 2007 -0600
@@ -62,17 +62,17 @@ read_exact(int fd, void *buf, size_t cou
int r = 0, s;
unsigned char *b = buf;
- while (r < count) {
+ while ( r < count )
+ {
s = read(fd, &b[r], count - r);
- if ((s == -1) && (errno == EINTR))
+ if ( (s == -1) && (errno == EINTR) )
continue;
- if (s <= 0) {
+ if ( s <= 0 )
break;
- }
r += s;
}
- return (r == count) ? 1 : 0;
+ return (r == count);
}
/*
@@ -93,20 +93,21 @@ static int uncanonicalize_pagetable(int
pte_last = PAGE_SIZE / ((pt_levels == 2)? 4 : 8);
/* First pass: work out how many (if any) MFNs we need to alloc */
- for(i = 0; i < pte_last; i++) {
-
- if(pt_levels == 2)
+ for ( i = 0; i < pte_last; i++ )
+ {
+ if ( pt_levels == 2 )
pte = ((uint32_t *)page)[i];
else
pte = ((uint64_t *)page)[i];
-
+
/* XXX SMH: below needs fixing for PROT_NONE etc */
- if(!(pte & _PAGE_PRESENT))
- continue;
+ if ( !(pte & _PAGE_PRESENT) )
+ continue;
pfn = (pte >> PAGE_SHIFT) & MFN_MASK_X86;
- if(pfn >= p2m_size) {
+ if ( pfn >= p2m_size )
+ {
/* This "page table page" is probably not one; bail. */
ERROR("Frame number in type %lu page table is out of range: "
"i=%d pfn=0x%lx p2m_size=%lu",
@@ -114,16 +115,18 @@ static int uncanonicalize_pagetable(int
return 0;
}
- if(p2m[pfn] == INVALID_P2M_ENTRY) {
+ if ( p2m[pfn] == INVALID_P2M_ENTRY )
+ {
/* Have a 'valid' PFN without a matching MFN - need to alloc */
p2m_batch[nr_mfns++] = pfn;
}
}
-
-
- /* Allocate the requistite number of mfns */
- if (nr_mfns && xc_domain_memory_populate_physmap(
- xc_handle, dom, nr_mfns, 0, 0, p2m_batch) != 0) {
+
+ /* Allocate the requisite number of mfns. */
+ if ( nr_mfns &&
+ (xc_domain_memory_populate_physmap(xc_handle, dom, nr_mfns, 0, 0,
+ p2m_batch) != 0) )
+ {
ERROR("Failed to allocate memory for batch.!\n");
errno = ENOMEM;
return 0;
@@ -131,26 +134,26 @@ static int uncanonicalize_pagetable(int
/* Second pass: uncanonicalize each present PTE */
nr_mfns = 0;
- for(i = 0; i < pte_last; i++) {
-
- if(pt_levels == 2)
+ for ( i = 0; i < pte_last; i++ )
+ {
+ if ( pt_levels == 2 )
pte = ((uint32_t *)page)[i];
else
pte = ((uint64_t *)page)[i];
/* XXX SMH: below needs fixing for PROT_NONE etc */
- if(!(pte & _PAGE_PRESENT))
+ if ( !(pte & _PAGE_PRESENT) )
continue;
pfn = (pte >> PAGE_SHIFT) & MFN_MASK_X86;
- if(p2m[pfn] == INVALID_P2M_ENTRY)
+ if ( p2m[pfn] == INVALID_P2M_ENTRY )
p2m[pfn] = p2m_batch[nr_mfns++];
pte &= ~MADDR_MASK_X86;
pte |= (uint64_t)p2m[pfn] << PAGE_SHIFT;
- if(pt_levels == 2)
+ if ( pt_levels == 2 )
((uint32_t *)page)[i] = (uint32_t)pte;
else
((uint64_t *)page)[i] = (uint64_t)pte;
@@ -161,62 +164,72 @@ static int uncanonicalize_pagetable(int
/* Load the p2m frame list, plus potential extended info chunk */
-static xen_pfn_t * load_p2m_frame_list(int io_fd, int *pae_extended_cr3)
+static xen_pfn_t *load_p2m_frame_list(int io_fd, int *pae_extended_cr3)
{
xen_pfn_t *p2m_frame_list;
vcpu_guest_context_t ctxt;
- if (!(p2m_frame_list = malloc(P2M_FL_SIZE))) {
+ if ( (p2m_frame_list = malloc(P2M_FL_SIZE)) == NULL )
+ {
ERROR("Couldn't allocate p2m_frame_list array");
return NULL;
}
/* Read first entry of P2M list, or extended-info signature (~0UL). */
- if (!read_exact(io_fd, p2m_frame_list, sizeof(long))) {
- ERROR("read extended-info signature failed");
- return NULL;
- }
+ if ( !read_exact(io_fd, p2m_frame_list, sizeof(long)) )
+ {
+ ERROR("read extended-info signature failed");
+ return NULL;
+ }
- if (p2m_frame_list[0] == ~0UL) {
+ if ( p2m_frame_list[0] == ~0UL )
+ {
uint32_t tot_bytes;
/* Next 4 bytes: total size of following extended info. */
- if (!read_exact(io_fd, &tot_bytes, sizeof(tot_bytes))) {
+ if ( !read_exact(io_fd, &tot_bytes, sizeof(tot_bytes)) )
+ {
ERROR("read extended-info size failed");
return NULL;
}
- while (tot_bytes) {
+ while ( tot_bytes )
+ {
uint32_t chunk_bytes;
char chunk_sig[4];
/* 4-character chunk signature + 4-byte remaining chunk size. */
- if (!read_exact(io_fd, chunk_sig, sizeof(chunk_sig)) ||
- !read_exact(io_fd, &chunk_bytes, sizeof(chunk_bytes))) {
+ if ( !read_exact(io_fd, chunk_sig, sizeof(chunk_sig)) ||
+ !read_exact(io_fd, &chunk_bytes, sizeof(chunk_bytes)) )
+ {
ERROR("read extended-info chunk signature failed");
return NULL;
}
tot_bytes -= 8;
-
+
/* VCPU context structure? */
- if (!strncmp(chunk_sig, "vcpu", 4)) {
- if (!read_exact(io_fd, &ctxt, sizeof(ctxt))) {
+ if ( !strncmp(chunk_sig, "vcpu", 4) )
+ {
+ if ( !read_exact(io_fd, &ctxt, sizeof(ctxt)) )
+ {
ERROR("read extended-info vcpu context failed");
return NULL;
}
tot_bytes -= sizeof(struct vcpu_guest_context);
chunk_bytes -= sizeof(struct vcpu_guest_context);
- if (ctxt.vm_assist & (1UL << VMASST_TYPE_pae_extended_cr3))
+ if ( ctxt.vm_assist & (1UL << VMASST_TYPE_pae_extended_cr3) )
*pae_extended_cr3 = 1;
}
/* Any remaining bytes of this chunk: read and discard. */
- while (chunk_bytes) {
+ while ( chunk_bytes )
+ {
unsigned long sz = chunk_bytes;
if ( sz > P2M_FL_SIZE )
sz = P2M_FL_SIZE;
- if (!read_exact(io_fd, p2m_frame_list, sz)) {
+ if ( !read_exact(io_fd, p2m_frame_list, sz) )
+ {
ERROR("read-and-discard extended-info chunk bytes failed");
return NULL;
}
@@ -224,24 +237,24 @@ static xen_pfn_t * load_p2m_frame_list(i
tot_bytes -= sz;
}
}
-
+
/* Now read the real first entry of P2M list. */
- if (!read_exact(io_fd, p2m_frame_list, sizeof(long))) {
+ if ( !read_exact(io_fd, p2m_frame_list, sizeof(long)) )
+ {
ERROR("read first entry of p2m_frame_list failed");
return NULL;
}
}
-
+
/* First entry is already read into the p2m array. */
- if (!read_exact(io_fd, &p2m_frame_list[1], P2M_FL_SIZE - sizeof(long))) {
- ERROR("read p2m_frame_list failed");
- return NULL;
+ if ( !read_exact(io_fd, &p2m_frame_list[1], P2M_FL_SIZE - sizeof(long)) )
+ {
+ ERROR("read p2m_frame_list failed");
+ return NULL;
}
return p2m_frame_list;
}
-
-
int xc_domain_restore(int xc_handle, int io_fd, uint32_t dom,
unsigned int store_evtchn, unsigned long *store_mfn,
@@ -284,7 +297,7 @@ int xc_domain_restore(int xc_handle, int
/* Our mapping of the current region (batch) */
char *region_base;
- xc_mmu_t *mmu = NULL;
+ struct xc_mmu *mmu = NULL;
/* used by debug verify code */
unsigned long buf[PAGE_SIZE/sizeof(unsigned long)];
@@ -323,20 +336,23 @@ int xc_domain_restore(int xc_handle, int
domctl.cmd = XEN_DOMCTL_set_address_size;
domctl.u.address_size.size = sizeof(unsigned long) * 8;
rc = do_domctl(xc_handle, &domctl);
- if ( rc != 0 ) {
+ if ( rc != 0 )
+ {
ERROR("Unable to set guest address size.");
goto out;
}
rc = 1;
}
- if(!get_platform_info(xc_handle, dom,
- &max_mfn, &hvirt_start, &pt_levels)) {
+ if ( !get_platform_info(xc_handle, dom,
+ &max_mfn, &hvirt_start, &pt_levels) )
+ {
ERROR("Unable to get platform info.");
return 1;
}
- if (lock_pages(&ctxt, sizeof(ctxt))) {
+ if ( lock_pages(&ctxt, sizeof(ctxt)) )
+ {
/* needed for build domctl, but might as well do early */
ERROR("Unable to lock ctxt");
return 1;
@@ -356,19 +372,22 @@ int xc_domain_restore(int xc_handle, int
region_mfn = calloc(MAX_BATCH_SIZE, sizeof(xen_pfn_t));
p2m_batch = calloc(MAX_BATCH_SIZE, sizeof(xen_pfn_t));
- if ((p2m == NULL) || (pfn_type == NULL) ||
- (region_mfn == NULL) || (p2m_batch == NULL)) {
+ if ( (p2m == NULL) || (pfn_type == NULL) ||
+ (region_mfn == NULL) || (p2m_batch == NULL) )
+ {
ERROR("memory alloc failed");
errno = ENOMEM;
goto out;
}
- if (lock_pages(region_mfn, sizeof(xen_pfn_t) * MAX_BATCH_SIZE)) {
+ if ( lock_pages(region_mfn, sizeof(xen_pfn_t) * MAX_BATCH_SIZE) )
+ {
ERROR("Could not lock region_mfn");
goto out;
}
- if (lock_pages(p2m_batch, sizeof(xen_pfn_t) * MAX_BATCH_SIZE)) {
+ if ( lock_pages(p2m_batch, sizeof(xen_pfn_t) * MAX_BATCH_SIZE) )
+ {
ERROR("Could not lock p2m_batch");
goto out;
}
@@ -376,7 +395,8 @@ int xc_domain_restore(int xc_handle, int
/* Get the domain's shared-info frame. */
domctl.cmd = XEN_DOMCTL_getdomaininfo;
domctl.domain = (domid_t)dom;
- if (xc_domctl(xc_handle, &domctl) < 0) {
+ if ( xc_domctl(xc_handle, &domctl) < 0 )
+ {
ERROR("Could not get information on new domain");
goto out;
}
@@ -386,7 +406,9 @@ int xc_domain_restore(int xc_handle, int
for ( pfn = 0; pfn < p2m_size; pfn++ )
p2m[pfn] = INVALID_P2M_ENTRY;
- if(!(mmu = xc_init_mmu_updates(xc_handle, dom))) {
+ mmu = xc_alloc_mmu_updates(xc_handle, dom);
+ if ( mmu == NULL )
+ {
ERROR("Could not initialise for MMU updates");
goto out;
}
@@ -400,8 +422,8 @@ int xc_domain_restore(int xc_handle, int
prev_pc = 0;
n = m = 0;
- while (1) {
-
+ for ( ; ; )
+ {
int j, nr_mfns = 0;
this_pc = (n * 100) / p2m_size;
@@ -411,39 +433,45 @@ int xc_domain_restore(int xc_handle, int
prev_pc = this_pc;
}
- if (!read_exact(io_fd, &j, sizeof(int))) {
+ if ( !read_exact(io_fd, &j, sizeof(int)) )
+ {
ERROR("Error when reading batch size");
goto out;
}
PPRINTF("batch %d\n",j);
- if (j == -1) {
+ if ( j == -1 )
+ {
verify = 1;
DPRINTF("Entering page verify mode\n");
continue;
}
- if (j == -2) {
+ if ( j == -2 )
+ {
new_ctxt_format = 1;
- if (!read_exact(io_fd, &max_vcpu_id, sizeof(int)) ||
- (max_vcpu_id >= 64) ||
- !read_exact(io_fd, &vcpumap, sizeof(uint64_t))) {
+ if ( !read_exact(io_fd, &max_vcpu_id, sizeof(int)) ||
+ (max_vcpu_id >= 64) ||
+ !read_exact(io_fd, &vcpumap, sizeof(uint64_t)) )
+ {
ERROR("Error when reading max_vcpu_id");
goto out;
}
continue;
}
- if (j == 0)
+ if ( j == 0 )
break; /* our work here is done */
- if (j > MAX_BATCH_SIZE) {
+ if ( j > MAX_BATCH_SIZE )
+ {
ERROR("Max batch size exceeded. Giving up.");
goto out;
}
- if (!read_exact(io_fd, region_pfn_type, j*sizeof(unsigned long))) {
+ if ( !read_exact(io_fd, region_pfn_type, j*sizeof(unsigned long)) )
+ {
ERROR("Error when reading region pfn types");
goto out;
}
@@ -464,10 +492,11 @@ int xc_domain_restore(int xc_handle, int
}
}
-
/* Now allocate a bunch of mfns for this batch */
- if (nr_mfns && xc_domain_memory_populate_physmap(
- xc_handle, dom, nr_mfns, 0, 0, p2m_batch) != 0) {
+ if ( nr_mfns &&
+ (xc_domain_memory_populate_physmap(xc_handle, dom, nr_mfns, 0,
+ 0, p2m_batch) != 0) )
+ {
ERROR("Failed to allocate memory for batch.!\n");
errno = ENOMEM;
goto out;
@@ -481,11 +510,12 @@ int xc_domain_restore(int xc_handle, int
pfn = region_pfn_type[i] & ~XEN_DOMCTL_PFINFO_LTAB_MASK;
pagetype = region_pfn_type[i] & XEN_DOMCTL_PFINFO_LTAB_MASK;
- if ( pagetype == XEN_DOMCTL_PFINFO_XTAB)
+ if ( pagetype == XEN_DOMCTL_PFINFO_XTAB )
region_mfn[i] = ~0UL; /* map will fail but we don't care */
else
{
- if (p2m[pfn] == INVALID_P2M_ENTRY) {
+ if ( p2m[pfn] == INVALID_P2M_ENTRY )
+ {
/* We just allocated a new mfn above; update p2m */
p2m[pfn] = p2m_batch[nr_mfns++];
nr_pfns++;
@@ -532,7 +562,8 @@ int xc_domain_restore(int xc_handle, int
/* In verify mode, we use a copy; otherwise we work in place */
page = verify ? (void *)buf : (region_base + i*PAGE_SIZE);
- if (!read_exact(io_fd, page, PAGE_SIZE)) {
+ if ( !read_exact(io_fd, page, PAGE_SIZE) )
+ {
ERROR("Error when reading page (type was %lx)", pagetype);
goto out;
}
@@ -577,13 +608,11 @@ int xc_domain_restore(int xc_handle, int
}
-
- if (verify) {
-
+ if ( verify )
+ {
int res = memcmp(buf, (region_base + i*PAGE_SIZE), PAGE_SIZE);
-
- if (res) {
-
+ if ( res )
+ {
int v;
DPRINTF("************** pfn=%lx type=%lx gotcs=%08lx "
@@ -591,20 +620,21 @@ int xc_domain_restore(int xc_handle, int
csum_page(region_base + i*PAGE_SIZE),
csum_page(buf));
- for (v = 0; v < 4; v++) {
-
+ for ( v = 0; v < 4; v++ )
+ {
unsigned long *p = (unsigned long *)
(region_base + i*PAGE_SIZE);
- if (buf[v] != p[v])
+ if ( buf[v] != p[v] )
DPRINTF(" %d: %08lx %08lx\n", v, buf[v], p[v]);
}
}
}
- if (!hvm
- && xc_add_mmu_update(xc_handle, mmu,
- (((unsigned long long)mfn) << PAGE_SHIFT)
- | MMU_MACHPHYS_UPDATE, pfn)) {
+ if ( !hvm &&
+ xc_add_mmu_update(xc_handle, mmu,
+ (((unsigned long long)mfn) << PAGE_SHIFT)
+ | MMU_MACHPHYS_UPDATE, pfn) )
+ {
ERROR("failed machpys update mfn=%lx pfn=%lx", mfn, pfn);
goto out;
}
@@ -629,8 +659,9 @@ int xc_domain_restore(int xc_handle, int
* Ensure we flush all machphys updates before potential PAE-specific
* reallocations below.
*/
- if (!hvm && xc_finish_mmu_updates(xc_handle, mmu)) {
- ERROR("Error doing finish_mmu_updates()");
+ if ( !hvm && xc_flush_mmu_updates(xc_handle, mmu) )
+ {
+ ERROR("Error doing flush_mmu_updates()");
goto out;
}
@@ -664,9 +695,9 @@ int xc_domain_restore(int xc_handle, int
*store_mfn = magic_pfns[2];
/* Read vcpu contexts */
- for (i = 0; i <= max_vcpu_id; i++)
- {
- if (!(vcpumap & (1ULL << i)))
+ for ( i = 0; i <= max_vcpu_id; i++ )
+ {
+ if ( !(vcpumap & (1ULL << i)) )
continue;
if ( !read_exact(io_fd, &(ctxt), sizeof(ctxt)) )
@@ -713,8 +744,8 @@ int xc_domain_restore(int xc_handle, int
/* Non-HVM guests only from here on */
- if ((pt_levels == 3) && !pae_extended_cr3) {
-
+ if ( (pt_levels == 3) && !pae_extended_cr3 )
+ {
/*
** XXX SMH on PAE we need to ensure PGDs are in MFNs < 4G. This
** is a little awkward and involves (a) finding all such PGDs and
@@ -744,21 +775,24 @@ int xc_domain_restore(int xc_handle, int
xc_map_foreign_range(xc_handle, dom, PAGE_SIZE,
PROT_READ, p2m[i]);
- for(j = 0; j < 4; j++)
+ for ( j = 0; j < 4; j++ )
l3ptes[j] = l3tab[j];
munmap(l3tab, PAGE_SIZE);
- if (!(new_mfn=xc_make_page_below_4G(xc_handle, dom, p2m[i]))) {
+ new_mfn = xc_make_page_below_4G(xc_handle, dom, p2m[i]);
+ if ( !new_mfn )
+ {
ERROR("Couldn't get a page below 4GB :-(");
goto out;
}
p2m[i] = new_mfn;
- if (xc_add_mmu_update(xc_handle, mmu,
- (((unsigned long long)new_mfn)
- << PAGE_SHIFT) |
- MMU_MACHPHYS_UPDATE, i)) {
+ if ( xc_add_mmu_update(xc_handle, mmu,
+ (((unsigned long long)new_mfn)
+ << PAGE_SHIFT) |
+ MMU_MACHPHYS_UPDATE, i) )
+ {
ERROR("Couldn't m2p on PAE root pgdir");
goto out;
}
@@ -767,11 +801,10 @@ int xc_domain_restore(int xc_handle, int
xc_map_foreign_range(xc_handle, dom, PAGE_SIZE,
PROT_READ | PROT_WRITE, p2m[i]);
- for(j = 0; j < 4; j++)
+ for ( j = 0; j < 4; j++ )
l3tab[j] = l3ptes[j];
munmap(l3tab, PAGE_SIZE);
-
}
}
@@ -787,19 +820,22 @@ int xc_domain_restore(int xc_handle, int
j++;
}
- if(i == (p2m_size-1) || j == MAX_BATCH_SIZE) {
-
- if (!(region_base = xc_map_foreign_batch(
- xc_handle, dom, PROT_READ | PROT_WRITE,
- region_mfn, j))) {
+ if ( (i == (p2m_size-1)) || (j == MAX_BATCH_SIZE) )
+ {
+ region_base = xc_map_foreign_batch(
+ xc_handle, dom, PROT_READ | PROT_WRITE, region_mfn, j);
+ if ( region_base == NULL )
+ {
ERROR("map batch failed");
goto out;
}
- for(k = 0; k < j; k++) {
- if(!uncanonicalize_pagetable(xc_handle, dom,
- XEN_DOMCTL_PFINFO_L1TAB,
- region_base + k*PAGE_SIZE)) {
+ for ( k = 0; k < j; k++ )
+ {
+ if ( !uncanonicalize_pagetable(
+ xc_handle, dom, XEN_DOMCTL_PFINFO_L1TAB,
+ region_base + k*PAGE_SIZE) )
+ {
ERROR("failed uncanonicalize pt!");
goto out;
}
@@ -810,8 +846,9 @@ int xc_domain_restore(int xc_handle, int
}
}
- if (xc_finish_mmu_updates(xc_handle, mmu)) {
- ERROR("Error doing finish_mmu_updates()");
+ if ( xc_flush_mmu_updates(xc_handle, mmu) )
+ {
+ ERROR("Error doing xc_flush_mmu_updates()");
goto out;
}
}
@@ -852,8 +889,10 @@ int xc_domain_restore(int xc_handle, int
nr_pins++;
/* Batch full? Then flush. */
- if (nr_pins == MAX_PIN_BATCH) {
- if (xc_mmuext_op(xc_handle, pin, nr_pins, dom) < 0) {
+ if ( nr_pins == MAX_PIN_BATCH )
+ {
+ if ( xc_mmuext_op(xc_handle, pin, nr_pins, dom) < 0 )
+ {
ERROR("Failed to pin batch of %d page tables", nr_pins);
goto out;
}
@@ -862,7 +901,8 @@ int xc_domain_restore(int xc_handle, int
}
/* Flush final partial batch. */
- if ((nr_pins != 0) && (xc_mmuext_op(xc_handle, pin, nr_pins, dom) < 0)) {
+ if ( (nr_pins != 0) && (xc_mmuext_op(xc_handle, pin, nr_pins, dom) < 0) )
+ {
ERROR("Failed to pin batch of %d page tables", nr_pins);
goto out;
}
@@ -876,36 +916,40 @@ int xc_domain_restore(int xc_handle, int
unsigned long *pfntab;
int nr_frees, rc;
- if (!read_exact(io_fd, &count, sizeof(count))) {
+ if ( !read_exact(io_fd, &count, sizeof(count)) )
+ {
ERROR("Error when reading pfn count");
goto out;
}
- if(!(pfntab = malloc(sizeof(unsigned long) * count))) {
+ if ( !(pfntab = malloc(sizeof(unsigned long) * count)) )
+ {
ERROR("Out of memory");
goto out;
}
- if (!read_exact(io_fd, pfntab, sizeof(unsigned long)*count)) {
+ if ( !read_exact(io_fd, pfntab, sizeof(unsigned long)*count) )
+ {
ERROR("Error when reading pfntab");
goto out;
}
nr_frees = 0;
- for (i = 0; i < count; i++) {
-
+ for ( i = 0; i < count; i++ )
+ {
unsigned long pfn = pfntab[i];
- if(p2m[pfn] != INVALID_P2M_ENTRY) {
+ if ( p2m[pfn] != INVALID_P2M_ENTRY )
+ {
/* pfn is not in physmap now, but was at some point during
the save/migration process - need to free it */
pfntab[nr_frees++] = p2m[pfn];
- p2m[pfn] = INVALID_P2M_ENTRY; // not in pseudo-physical map
- }
- }
-
- if (nr_frees > 0) {
-
+ p2m[pfn] = INVALID_P2M_ENTRY; /* not in pseudo-physical map */
+ }
+ }
+
+ if ( nr_frees > 0 )
+ {
struct xen_memory_reservation reservation = {
.nr_extents = nr_frees,
.extent_order = 0,
@@ -913,20 +957,24 @@ int xc_domain_restore(int xc_handle, int
};
set_xen_guest_handle(reservation.extent_start, pfntab);
- if ((rc = xc_memory_op(xc_handle, XENMEM_decrease_reservation,
- &reservation)) != nr_frees) {
+ if ( (rc = xc_memory_op(xc_handle, XENMEM_decrease_reservation,
+ &reservation)) != nr_frees )
+ {
ERROR("Could not decrease reservation : %d", rc);
goto out;
- } else
+ }
+ else
DPRINTF("Decreased reservation by %d pages\n", count);
}
}
- for (i = 0; i <= max_vcpu_id; i++) {
- if (!(vcpumap & (1ULL << i)))
+ for ( i = 0; i <= max_vcpu_id; i++ )
+ {
+ if ( !(vcpumap & (1ULL << i)) )
continue;
- if (!read_exact(io_fd, &ctxt, sizeof(ctxt))) {
+ if ( !read_exact(io_fd, &ctxt, sizeof(ctxt)) )
+ {
ERROR("Error when reading ctxt %d", i);
goto out;
}
@@ -934,14 +982,16 @@ int xc_domain_restore(int xc_handle, int
if ( !new_ctxt_format )
ctxt.flags |= VGCF_online;
- if (i == 0) {
+ if ( i == 0 )
+ {
/*
* Uncanonicalise the suspend-record frame number and poke
* resume record.
*/
pfn = ctxt.user_regs.edx;
- if ((pfn >= p2m_size) ||
- (pfn_type[pfn] != XEN_DOMCTL_PFINFO_NOTAB)) {
+ if ( (pfn >= p2m_size) ||
+ (pfn_type[pfn] != XEN_DOMCTL_PFINFO_NOTAB) )
+ {
ERROR("Suspend record frame number is bad");
goto out;
}
@@ -960,15 +1010,18 @@ int xc_domain_restore(int xc_handle, int
}
/* Uncanonicalise each GDT frame number. */
- if (ctxt.gdt_ents > 8192) {
+ if ( ctxt.gdt_ents > 8192 )
+ {
ERROR("GDT entry count out of range");
goto out;
}
- for (j = 0; (512*j) < ctxt.gdt_ents; j++) {
+ for ( j = 0; (512*j) < ctxt.gdt_ents; j++ )
+ {
pfn = ctxt.gdt_frames[j];
- if ((pfn >= p2m_size) ||
- (pfn_type[pfn] != XEN_DOMCTL_PFINFO_NOTAB)) {
+ if ( (pfn >= p2m_size) ||
+ (pfn_type[pfn] != XEN_DOMCTL_PFINFO_NOTAB) )
+ {
ERROR("GDT frame number is bad");
goto out;
}
@@ -978,14 +1031,16 @@ int xc_domain_restore(int xc_handle, int
/* Uncanonicalise the page table base pointer. */
pfn = xen_cr3_to_pfn(ctxt.ctrlreg[3]);
- if (pfn >= p2m_size) {
+ if ( pfn >= p2m_size )
+ {
ERROR("PT base is bad: pfn=%lu p2m_size=%lu type=%08lx",
pfn, p2m_size, pfn_type[pfn]);
goto out;
}
if ( (pfn_type[pfn] & XEN_DOMCTL_PFINFO_LTABTYPE_MASK) !=
- ((unsigned long)pt_levels<<XEN_DOMCTL_PFINFO_LTAB_SHIFT) ) {
+ ((unsigned long)pt_levels<<XEN_DOMCTL_PFINFO_LTAB_SHIFT) )
+ {
ERROR("PT base is bad. pfn=%lu nr=%lu type=%08lx %08lx",
pfn, p2m_size, pfn_type[pfn],
(unsigned long)pt_levels<<XEN_DOMCTL_PFINFO_LTAB_SHIFT);
@@ -999,14 +1054,16 @@ int xc_domain_restore(int xc_handle, int
{
pfn = xen_cr3_to_pfn(ctxt.ctrlreg[1]);
- if (pfn >= p2m_size) {
+ if ( pfn >= p2m_size )
+ {
ERROR("User PT base is bad: pfn=%lu p2m_size=%lu type=%08lx",
pfn, p2m_size, pfn_type[pfn]);
goto out;
}
if ( (pfn_type[pfn] & XEN_DOMCTL_PFINFO_LTABTYPE_MASK) !=
- ((unsigned long)pt_levels<<XEN_DOMCTL_PFINFO_LTAB_SHIFT) ) {
+ ((unsigned long)pt_levels<<XEN_DOMCTL_PFINFO_LTAB_SHIFT) )
+ {
ERROR("User PT base is bad. pfn=%lu nr=%lu type=%08lx %08lx",
pfn, p2m_size, pfn_type[pfn],
(unsigned long)pt_levels<<XEN_DOMCTL_PFINFO_LTAB_SHIFT);
@@ -1021,14 +1078,16 @@ int xc_domain_restore(int xc_handle, int
domctl.u.vcpucontext.vcpu = i;
set_xen_guest_handle(domctl.u.vcpucontext.ctxt, &ctxt);
rc = xc_domctl(xc_handle, &domctl);
- if (rc != 0) {
+ if ( rc != 0 )
+ {
ERROR("Couldn't build vcpu%d", i);
goto out;
}
rc = 1;
}
- if (!read_exact(io_fd, shared_info_page, PAGE_SIZE)) {
+ if ( !read_exact(io_fd, shared_info_page, PAGE_SIZE) )
+ {
ERROR("Error when reading shared info page");
goto out;
}
@@ -1046,9 +1105,11 @@ int xc_domain_restore(int xc_handle, int
munmap(page, PAGE_SIZE);
/* Uncanonicalise the pfn-to-mfn table frame-number list. */
- for (i = 0; i < P2M_FL_ENTRIES; i++) {
+ for ( i = 0; i < P2M_FL_ENTRIES; i++ )
+ {
pfn = p2m_frame_list[i];
- if ((pfn >= p2m_size) || (pfn_type[pfn] != XEN_DOMCTL_PFINFO_NOTAB)) {
+ if ( (pfn >= p2m_size) || (pfn_type[pfn] != XEN_DOMCTL_PFINFO_NOTAB) )
+ {
ERROR("PFN-to-MFN frame number is bad");
goto out;
}
@@ -1057,8 +1118,9 @@ int xc_domain_restore(int xc_handle, int
}
/* Copy the P2M we've constructed to the 'live' P2M */
- if (!(live_p2m = xc_map_foreign_batch(xc_handle, dom, PROT_WRITE,
- p2m_frame_list, P2M_FL_ENTRIES))) {
+ if ( !(live_p2m = xc_map_foreign_batch(xc_handle, dom, PROT_WRITE,
+ p2m_frame_list, P2M_FL_ENTRIES)) )
+ {
ERROR("Couldn't map p2m table");
goto out;
}
diff -r cd3cfddcb055 -r 1bde28f762a6 tools/libxc/xc_hvm_save.c
--- a/tools/libxc/xc_hvm_save.c Mon Apr 09 13:39:35 2007 -0600
+++ b/tools/libxc/xc_hvm_save.c Mon Apr 09 13:40:25 2007 -0600
@@ -119,8 +119,8 @@ static uint64_t llgettimeofday(void)
static uint64_t tv_delta(struct timeval *new, struct timeval *old)
{
- return ((new->tv_sec - old->tv_sec)*1000000 ) +
- (new->tv_usec - old->tv_usec);
+ return (((new->tv_sec - old->tv_sec)*1000000) +
+ (new->tv_usec - old->tv_usec));
}
@@ -130,9 +130,7 @@ static uint64_t tv_delta(struct timeval
static inline ssize_t write_exact(int fd, void *buf, size_t count)
{
- if(write(fd, buf, count) != count)
- return 0;
- return 1;
+ return (write(fd, buf, count) == count);
}
static int print_stats(int xc_handle, uint32_t domid, int pages_sent,
@@ -156,15 +154,14 @@ static int print_stats(int xc_handle, ui
DPRINTF("ARRHHH!!\n");
wall_delta = tv_delta(&wall_now,&wall_last)/1000;
-
- if (wall_delta == 0) wall_delta = 1;
+ if ( wall_delta == 0 )
+ wall_delta = 1;
d0_cpu_delta = (d0_cpu_now - d0_cpu_last)/1000;
d1_cpu_delta = (d1_cpu_now - d1_cpu_last)/1000;
- if (print)
- DPRINTF(
- "delta %lldms, dom0 %d%%, target %d%%, sent %dMb/s, "
+ if ( print )
+ DPRINTF("delta %lldms, dom0 %d%%, target %d%%, sent %dMb/s, "
"dirtied %dMb/s %" PRId32 " pages\n",
wall_delta,
(int)((d0_cpu_delta*100)/wall_delta),
@@ -189,18 +186,19 @@ static int analysis_phase(int xc_handle,
start = llgettimeofday();
- for (j = 0; j < runs; j++) {
+ for ( j = 0; j < runs; j++ )
+ {
int i;
xc_shadow_control(xc_handle, domid, XEN_DOMCTL_SHADOW_OP_CLEAN,
arr, pfn_array_size, NULL, 0, NULL);
DPRINTF("#Flush\n");
- for ( i = 0; i < 40; i++ ) {
+ for ( i = 0; i < 40; i++ )
+ {
usleep(50000);
now = llgettimeofday();
xc_shadow_control(xc_handle, domid, XEN_DOMCTL_SHADOW_OP_PEEK,
NULL, 0, NULL, 0, &stats);
-
DPRINTF("now= %lld faults= %"PRId32" dirty= %"PRId32"\n",
((now-start)+500)/1000,
stats.fault_count, stats.dirty_count);
@@ -216,39 +214,39 @@ static int suspend_and_state(int (*suspe
{
int i = 0;
- if (!(*suspend)(dom)) {
+ if ( !(*suspend)(dom) )
+ {
ERROR("Suspend request failed");
return -1;
}
retry:
- if (xc_domain_getinfo(xc_handle, dom, 1, info) != 1) {
+ if ( xc_domain_getinfo(xc_handle, dom, 1, info) != 1 )
+ {
ERROR("Could not get domain info");
return -1;
}
- if ( xc_vcpu_getcontext(xc_handle, dom, 0 /* XXX */, ctxt))
+ if ( xc_vcpu_getcontext(xc_handle, dom, 0, ctxt) )
ERROR("Could not get vcpu context");
-
- if (info->shutdown && info->shutdown_reason == SHUTDOWN_suspend)
- return 0; // success
-
- if (info->paused) {
- // try unpausing domain, wait, and retest
+ if ( info->shutdown && (info->shutdown_reason == SHUTDOWN_suspend) )
+ return 0; /* success */
+
+ if ( info->paused )
+ {
+ /* Try unpausing domain, wait, and retest. */
xc_domain_unpause( xc_handle, dom );
-
ERROR("Domain was paused. Wait and re-test.");
- usleep(10000); // 10ms
-
+ usleep(10000); /* 10ms */
goto retry;
}
-
- if( ++i < 100 ) {
+ if ( ++i < 100 )
+ {
ERROR("Retry suspend domain.");
- usleep(10000); // 10ms
+ usleep(10000); /* 10ms */
goto retry;
}
@@ -350,8 +348,6 @@ int xc_hvm_save(int xc_handle, int io_fd
ERROR("Couldn't enable shadow mode");
goto out;
}
-
- DPRINTF("hvm domain live migration debug start: logdirty enable.\n");
}
else
{
@@ -378,7 +374,6 @@ int xc_hvm_save(int xc_handle, int io_fd
ERROR("Error when writing to state file (1)");
goto out;
}
-
/* pretend we sent all the pages last iteration */
sent_last_iter = pfn_array_size;
@@ -452,7 +447,7 @@ int xc_hvm_save(int xc_handle, int io_fd
prev_pc = 0;
N=0;
- DPRINTF("Saving HVM domain memory pages: iter %d 0%%", iter);
+ DPRINTF("Saving memory pages: iter %d 0%%", iter);
while ( N < pfn_array_size )
{
@@ -474,7 +469,7 @@ int xc_hvm_save(int xc_handle, int io_fd
pfn_array_size, NULL, 0, NULL);
if ( rc != pfn_array_size )
{
- ERROR("Error peeking HVM shadow bitmap");
+ ERROR("Error peeking shadow bitmap");
goto out;
}
}
@@ -703,7 +698,7 @@ int xc_hvm_save(int xc_handle, int io_fd
/* save vcpu/vmcs contexts */
for ( i = 0; i < nr_vcpus; i++ )
{
- if (!(vcpumap & (1ULL << i)))
+ if ( !(vcpumap & (1ULL << i)) )
continue;
if ( xc_vcpu_getcontext(xc_handle, dom, i, &ctxt) )
diff -r cd3cfddcb055 -r 1bde28f762a6 tools/libxc/xc_linux_save.c
--- a/tools/libxc/xc_linux_save.c Mon Apr 09 13:39:35 2007 -0600
+++ b/tools/libxc/xc_linux_save.c Mon Apr 09 13:40:25 2007 -0600
@@ -27,7 +27,6 @@
#define DEF_MAX_ITERS 29 /* limit us to 30 times round loop */
#define DEF_MAX_FACTOR 3 /* never send more than 3x p2m_size */
-
/* max mfn of the whole machine */
static unsigned long max_mfn;
@@ -56,10 +55,9 @@ static unsigned long m2p_mfn0;
* in the guest's pseudophysical map.
*/
#define MFN_IS_IN_PSEUDOPHYS_MAP(_mfn) \
-(((_mfn) < (max_mfn)) && \
- ((mfn_to_pfn(_mfn) < (p2m_size)) && \
- (live_p2m[mfn_to_pfn(_mfn)] == (_mfn))))
-
+ (((_mfn) < (max_mfn)) && \
+ ((mfn_to_pfn(_mfn) < (p2m_size)) && \
+ (live_p2m[mfn_to_pfn(_mfn)] == (_mfn))))
/* Returns TRUE if MFN is successfully converted to a PFN. */
#define translate_mfn_to_pfn(_pmfn) \
@@ -116,7 +114,7 @@ static inline int count_bits ( int nr, v
int i, count = 0;
volatile unsigned long *p = (volatile unsigned long *)addr;
/* We know that the array is padded to unsigned long. */
- for( i = 0; i < (nr / (sizeof(unsigned long)*8)); i++, p++ )
+ for ( i = 0; i < (nr / (sizeof(unsigned long)*8)); i++, p++ )
count += hweight32(*p);
return count;
}
@@ -151,9 +149,6 @@ static inline int permute( int i, int nr
return i;
}
-
-
-
static uint64_t tv_to_us(struct timeval *new)
{
return (new->tv_sec * 1000000) + new->tv_usec;
@@ -168,8 +163,8 @@ static uint64_t llgettimeofday(void)
static uint64_t tv_delta(struct timeval *new, struct timeval *old)
{
- return ((new->tv_sec - old->tv_sec)*1000000 ) +
- (new->tv_usec - old->tv_usec);
+ return (((new->tv_sec - old->tv_sec)*1000000) +
+ (new->tv_usec - old->tv_usec));
}
static int noncached_write(int fd, int live, void *buffer, int len)
@@ -179,8 +174,8 @@ static int noncached_write(int fd, int l
int rc = write(fd,buffer,len);
write_count += len;
-
- if (write_count >= MAX_PAGECACHE_USAGE*PAGE_SIZE) {
+ if ( write_count >= (MAX_PAGECACHE_USAGE * PAGE_SIZE) )
+ {
/* Time to discard cache - dont care if this fails */
discard_file_cache(fd, 0 /* no flush */);
write_count = 0;
@@ -190,7 +185,6 @@ static int noncached_write(int fd, int l
}
#ifdef ADAPTIVE_SAVE
-
/*
** We control the rate at which we transmit (or save) to minimize impact
@@ -200,26 +194,22 @@ static int noncached_write(int fd, int l
#define MAX_MBIT_RATE 500 /* maximum transmit rate for migrate */
#define START_MBIT_RATE 100 /* initial transmit rate for migrate */
-
/* Scaling factor to convert between a rate (in Mb/s) and time (in usecs) */
#define RATE_TO_BTU 781250
/* Amount in bytes we allow ourselves to send in a burst */
#define BURST_BUDGET (100*1024)
-
/* We keep track of the current and previous transmission rate */
static int mbit_rate, ombit_rate = 0;
/* Have we reached the maximum transmission rate? */
#define RATE_IS_MAX() (mbit_rate == MAX_MBIT_RATE)
-
static inline void initialize_mbit_rate()
{
mbit_rate = START_MBIT_RATE;
}
-
static int ratewrite(int io_fd, int live, void *buf, int n)
{
@@ -230,39 +220,47 @@ static int ratewrite(int io_fd, int live
struct timespec delay;
long long delta;
- if (START_MBIT_RATE == 0)
+ if ( START_MBIT_RATE == 0 )
return noncached_write(io_fd, live, buf, n);
budget -= n;
- if (budget < 0) {
- if (mbit_rate != ombit_rate) {
+ if ( budget < 0 )
+ {
+ if ( mbit_rate != ombit_rate )
+ {
burst_time_us = RATE_TO_BTU / mbit_rate;
ombit_rate = mbit_rate;
DPRINTF("rate limit: %d mbit/s burst budget %d slot time %d\n",
mbit_rate, BURST_BUDGET, burst_time_us);
}
- if (last_put.tv_sec == 0) {
+ if ( last_put.tv_sec == 0 )
+ {
budget += BURST_BUDGET;
gettimeofday(&last_put, NULL);
- } else {
- while (budget < 0) {
+ }
+ else
+ {
+ while ( budget < 0 )
+ {
gettimeofday(&now, NULL);
delta = tv_delta(&now, &last_put);
- while (delta > burst_time_us) {
+ while ( delta > burst_time_us )
+ {
budget += BURST_BUDGET;
last_put.tv_usec += burst_time_us;
- if (last_put.tv_usec > 1000000) {
+ if ( last_put.tv_usec > 1000000
+ {
last_put.tv_usec -= 1000000;
last_put.tv_sec++;
}
delta -= burst_time_us;
}
- if (budget > 0)
+ if ( budget > 0 )
break;
delay.tv_sec = 0;
delay.tv_nsec = 1000 * (burst_time_us - delta);
- while (delay.tv_nsec > 0)
- if (nanosleep(&delay, &delay) == 0)
+ while ( delay.tv_nsec > 0 )
+ if ( nanosleep(&delay, &delay) == 0 )
break;
}
}
@@ -278,15 +276,10 @@ static int ratewrite(int io_fd, int live
#endif
-
static inline ssize_t write_exact(int fd, void *buf, size_t count)
{
- if(write(fd, buf, count) != count)
- return 0;
- return 1;
-}
-
-
+ return (write(fd, buf, count) == count);
+}
static int print_stats(int xc_handle, uint32_t domid, int pages_sent,
xc_shadow_op_stats_t *stats, int print)
@@ -309,15 +302,14 @@ static int print_stats(int xc_handle, ui
DPRINTF("ARRHHH!!\n");
wall_delta = tv_delta(&wall_now,&wall_last)/1000;
-
- if (wall_delta == 0) wall_delta = 1;
+ if ( wall_delta == 0 )
+ wall_delta = 1;
d0_cpu_delta = (d0_cpu_now - d0_cpu_last)/1000;
d1_cpu_delta = (d1_cpu_now - d1_cpu_last)/1000;
- if (print)
- DPRINTF(
- "delta %lldms, dom0 %d%%, target %d%%, sent %dMb/s, "
+ if ( print )
+ DPRINTF("delta %lldms, dom0 %d%%, target %d%%, sent %dMb/s, "
"dirtied %dMb/s %" PRId32 " pages\n",
wall_delta,
(int)((d0_cpu_delta*100)/wall_delta),
@@ -327,10 +319,11 @@ static int print_stats(int xc_handle, ui
stats->dirty_count);
#ifdef ADAPTIVE_SAVE
- if (((stats->dirty_count*PAGE_SIZE)/(wall_delta*(1000/8))) > mbit_rate) {
+ if ( ((stats->dirty_count*PAGE_SIZE)/(wall_delta*(1000/8))) > mbit_rate )
+ {
mbit_rate = (int)((stats->dirty_count*PAGE_SIZE)/(wall_delta*(1000/8)))
+ 50;
- if (mbit_rate > MAX_MBIT_RATE)
+ if ( mbit_rate > MAX_MBIT_RATE )
mbit_rate = MAX_MBIT_RATE;
}
#endif
@@ -352,18 +345,19 @@ static int analysis_phase(int xc_handle,
start = llgettimeofday();
- for (j = 0; j < runs; j++) {
+ for ( j = 0; j < runs; j++ )
+ {
int i;
xc_shadow_control(xc_handle, domid, XEN_DOMCTL_SHADOW_OP_CLEAN,
arr, p2m_size, NULL, 0, NULL);
DPRINTF("#Flush\n");
- for ( i = 0; i < 40; i++ ) {
+ for ( i = 0; i < 40; i++ )
+ {
usleep(50000);
now = llgettimeofday();
xc_shadow_control(xc_handle, domid, XEN_DOMCTL_SHADOW_OP_PEEK,
NULL, 0, NULL, 0, &stats);
-
DPRINTF("now= %lld faults= %"PRId32" dirty= %"PRId32"\n",
((now-start)+500)/1000,
stats.fault_count, stats.dirty_count);
@@ -380,34 +374,40 @@ static int suspend_and_state(int (*suspe
{
int i = 0;
- if (!(*suspend)(dom)) {
+ if ( !(*suspend)(dom) )
+ {
ERROR("Suspend request failed");
return -1;
}
retry:
- if (xc_domain_getinfo(xc_handle, dom, 1, info) != 1) {
+ if ( xc_domain_getinfo(xc_handle, dom, 1, info) != 1 )
+ {
ERROR("Could not get domain info");
return -1;
}
- if ( xc_vcpu_getcontext(xc_handle, dom, 0 /* XXX */, ctxt))
+ if ( xc_vcpu_getcontext(xc_handle, dom, 0, ctxt) )
ERROR("Could not get vcpu context");
- if (info->dying) {
+ if ( info->dying )
+ {
ERROR("domain is dying");
return -1;
}
- if (info->crashed) {
+ if ( info->crashed )
+ {
ERROR("domain has crashed");
return -1;
}
- if (info->shutdown) {
- switch (info->shutdown_reason) {
+ if ( info->shutdown )
+ {
+ switch ( info->shutdown_reason )
+ {
case SHUTDOWN_poweroff:
case SHUTDOWN_reboot:
ERROR("domain has shut down");
@@ -420,20 +420,19 @@ static int suspend_and_state(int (*suspe
}
}
- if (info->paused) {
- // try unpausing domain, wait, and retest
+ if ( info->paused )
+ {
+ /* Try unpausing domain, wait, and retest. */
xc_domain_unpause( xc_handle, dom );
-
ERROR("Domain was paused. Wait and re-test.");
- usleep(10000); // 10ms
-
+ usleep(10000); /* 10ms */
goto retry;
}
-
- if( ++i < 100 ) {
+ if ( ++i < 100 )
+ {
ERROR("Retry suspend domain");
- usleep(10000); // 10ms
+ usleep(10000); /* 10ms */
goto retry;
}
@@ -453,18 +452,18 @@ static void *map_frame_list_list(int xc_
int count = 100;
void *p;
- while (count-- && shinfo->arch.pfn_to_mfn_frame_list_list == 0)
+ while ( count-- && (shinfo->arch.pfn_to_mfn_frame_list_list == 0) )
usleep(10000);
- if (shinfo->arch.pfn_to_mfn_frame_list_list == 0) {
+ if ( shinfo->arch.pfn_to_mfn_frame_list_list == 0 )
+ {
ERROR("Timed out waiting for frame list updated.");
return NULL;
}
p = xc_map_foreign_range(xc_handle, dom, PAGE_SIZE, PROT_READ,
shinfo->arch.pfn_to_mfn_frame_list_list);
-
- if (p == NULL)
+ if ( p == NULL )
ERROR("Couldn't map p2m_frame_list_list (errno %d)", errno);
return p;
@@ -493,10 +492,10 @@ static int canonicalize_pagetable(unsign
*/
xen_start = xen_end = pte_last = PAGE_SIZE / ((pt_levels == 2) ? 4 : 8);
- if (pt_levels == 2 && type == XEN_DOMCTL_PFINFO_L2TAB)
+ if ( (pt_levels == 2) && (type == XEN_DOMCTL_PFINFO_L2TAB) )
xen_start = (hvirt_start >> L2_PAGETABLE_SHIFT);
- if (pt_levels == 3 && type == XEN_DOMCTL_PFINFO_L3TAB)
+ if ( (pt_levels == 3) && (type == XEN_DOMCTL_PFINFO_L3TAB) )
xen_start = L3_PAGETABLE_ENTRIES_PAE;
/*
@@ -505,27 +504,31 @@ static int canonicalize_pagetable(unsign
** Xen always ensures is present in that L2. Guests must ensure
** that this check will fail for other L2s.
*/
- if (pt_levels == 3 && type == XEN_DOMCTL_PFINFO_L2TAB) {
+ if ( (pt_levels == 3) && (type == XEN_DOMCTL_PFINFO_L2TAB) )
+ {
int hstart;
uint64_t he;
hstart = (hvirt_start >> L2_PAGETABLE_SHIFT_PAE) & 0x1ff;
he = ((const uint64_t *) spage)[hstart];
- if ( ((he >> PAGE_SHIFT) & MFN_MASK_X86) == m2p_mfn0 ) {
+ if ( ((he >> PAGE_SHIFT) & MFN_MASK_X86) == m2p_mfn0 )
+ {
/* hvirt starts with xen stuff... */
xen_start = hstart;
- } else if ( hvirt_start != 0xf5800000 ) {
+ }
+ else if ( hvirt_start != 0xf5800000 )
+ {
/* old L2s from before hole was shrunk... */
hstart = (0xf5800000 >> L2_PAGETABLE_SHIFT_PAE) & 0x1ff;
he = ((const uint64_t *) spage)[hstart];
-
- if( ((he >> PAGE_SHIFT) & MFN_MASK_X86) == m2p_mfn0 )
+ if ( ((he >> PAGE_SHIFT) & MFN_MASK_X86) == m2p_mfn0 )
xen_start = hstart;
}
}
- if (pt_levels == 4 && type == XEN_DOMCTL_PFINFO_L4TAB) {
+ if ( (pt_levels == 4) && (type == XEN_DOMCTL_PFINFO_L4TAB) )
+ {
/*
** XXX SMH: should compute these from hvirt_start (which we have)
** and hvirt_end (which we don't)
@@ -535,27 +538,29 @@ static int canonicalize_pagetable(unsign
}
/* Now iterate through the page table, canonicalizing each PTE */
- for (i = 0; i < pte_last; i++ ) {
-
+ for (i = 0; i < pte_last; i++ )
+ {
unsigned long pfn, mfn;
- if (pt_levels == 2)
+ if ( pt_levels == 2 )
pte = ((const uint32_t*)spage)[i];
else
pte = ((const uint64_t*)spage)[i];
- if (i >= xen_start && i < xen_end)
+ if ( (i >= xen_start) && (i < xen_end) )
pte = 0;
- if (pte & _PAGE_PRESENT) {
-
+ if ( pte & _PAGE_PRESENT )
+ {
mfn = (pte >> PAGE_SHIFT) & MFN_MASK_X86;
- if (!MFN_IS_IN_PSEUDOPHYS_MAP(mfn)) {
+ if ( !MFN_IS_IN_PSEUDOPHYS_MAP(mfn) )
+ {
/* This will happen if the type info is stale which
is quite feasible under live migration */
pfn = 0; /* zap it - we'll retransmit this page later */
race = 1; /* inform the caller of race; fatal if !live */
- } else
+ }
+ else
pfn = mfn_to_pfn(mfn);
pte &= ~MADDR_MASK_X86;
@@ -566,25 +571,20 @@ static int canonicalize_pagetable(unsign
* a 64bit hypervisor. We zap these here to avoid any
* surprise at restore time...
*/
- if ( pt_levels == 3 &&
- type == XEN_DOMCTL_PFINFO_L3TAB &&
- pte & (_PAGE_USER|_PAGE_RW|_PAGE_ACCESSED) )
- {
+ if ( (pt_levels == 3) &&
+ (type == XEN_DOMCTL_PFINFO_L3TAB) &&
+ (pte & (_PAGE_USER|_PAGE_RW|_PAGE_ACCESSED)) )
pte &= ~(_PAGE_USER|_PAGE_RW|_PAGE_ACCESSED);
- }
- }
-
- if (pt_levels == 2)
+ }
+
+ if ( pt_levels == 2 )
((uint32_t*)dpage)[i] = pte;
else
((uint64_t*)dpage)[i] = pte;
-
- }
-
- return race;
-}
-
-
+ }
+
+ return race;
+}
static xen_pfn_t *xc_map_m2p(int xc_handle,
unsigned long max_mfn,
@@ -601,37 +601,43 @@ static xen_pfn_t *xc_map_m2p(int xc_hand
m2p_chunks = M2P_CHUNKS(max_mfn);
xmml.max_extents = m2p_chunks;
- if (!(extent_start = malloc(m2p_chunks * sizeof(xen_pfn_t)))) {
+ if ( !(extent_start = malloc(m2p_chunks * sizeof(xen_pfn_t))) )
+ {
ERROR("failed to allocate space for m2p mfns");
return NULL;
}
set_xen_guest_handle(xmml.extent_start, extent_start);
- if (xc_memory_op(xc_handle, XENMEM_machphys_mfn_list, &xmml) ||
- (xmml.nr_extents != m2p_chunks)) {
+ if ( xc_memory_op(xc_handle, XENMEM_machphys_mfn_list, &xmml) ||
+ (xmml.nr_extents != m2p_chunks) )
+ {
ERROR("xc_get_m2p_mfns");
return NULL;
}
- if ((m2p = mmap(NULL, m2p_size, prot,
- MAP_SHARED, xc_handle, 0)) == MAP_FAILED) {
+ if ( (m2p = mmap(NULL, m2p_size, prot,
+ MAP_SHARED, xc_handle, 0)) == MAP_FAILED )
+ {
ERROR("failed to mmap m2p");
return NULL;
}
- if (!(entries = malloc(m2p_chunks * sizeof(privcmd_mmap_entry_t)))) {
+ if ( !(entries = malloc(m2p_chunks * sizeof(privcmd_mmap_entry_t))) )
+ {
ERROR("failed to allocate space for mmap entries");
return NULL;
}
- for (i=0; i < m2p_chunks; i++) {
+ for ( i = 0; i < m2p_chunks; i++ )
+ {
entries[i].va = (unsigned long)(((void *)m2p) + (i * M2P_CHUNK_SIZE));
entries[i].mfn = extent_start[i];
entries[i].npages = M2P_CHUNK_SIZE >> PAGE_SHIFT;
}
- if ((rc = xc_map_foreign_ranges(xc_handle, DOMID_XEN,
- entries, m2p_chunks)) < 0) {
+ if ( (rc = xc_map_foreign_ranges(xc_handle, DOMID_XEN,
+ entries, m2p_chunks)) < 0 )
+ {
ERROR("xc_mmap_foreign_ranges failed (rc = %d)", rc);
return NULL;
}
@@ -643,8 +649,6 @@ static xen_pfn_t *xc_map_m2p(int xc_hand
return m2p;
}
-
-
int xc_linux_save(int xc_handle, int io_fd, uint32_t dom, uint32_t max_iters,
uint32_t max_factor, uint32_t flags, int (*suspend)(int))
@@ -699,33 +703,35 @@ int xc_linux_save(int xc_handle, int io_
uint64_t vcpumap = 1ULL;
/* If no explicit control parameters given, use defaults */
- if(!max_iters)
- max_iters = DEF_MAX_ITERS;
- if(!max_factor)
- max_factor = DEF_MAX_FACTOR;
+ max_iters = max_iters ? : DEF_MAX_ITERS;
+ max_factor = max_factor ? : DEF_MAX_FACTOR;
initialize_mbit_rate();
- if(!get_platform_info(xc_handle, dom,
- &max_mfn, &hvirt_start, &pt_levels)) {
+ if ( !get_platform_info(xc_handle, dom,
+ &max_mfn, &hvirt_start, &pt_levels) )
+ {
ERROR("Unable to get platform info.");
return 1;
}
- if (xc_domain_getinfo(xc_handle, dom, 1, &info) != 1) {
+ if ( xc_domain_getinfo(xc_handle, dom, 1, &info) != 1 )
+ {
ERROR("Could not get domain info");
return 1;
}
- if (xc_vcpu_getcontext(xc_handle, dom, 0, &ctxt)) {
+ if ( xc_vcpu_getcontext(xc_handle, dom, 0, &ctxt) )
+ {
ERROR("Could not get vcpu context");
goto out;
}
shared_info_frame = info.shared_info_frame;
/* Map the shared info frame */
- if(!(live_shinfo = xc_map_foreign_range(xc_handle, dom, PAGE_SIZE,
- PROT_READ, shared_info_frame))) {
+ if ( !(live_shinfo = xc_map_foreign_range(xc_handle, dom, PAGE_SIZE,
+ PROT_READ, shared_info_frame)) )
+ {
ERROR("Couldn't map live_shinfo");
goto out;
}
@@ -734,16 +740,15 @@ int xc_linux_save(int xc_handle, int io_
live_p2m_frame_list_list = map_frame_list_list(xc_handle, dom,
live_shinfo);
-
- if (!live_p2m_frame_list_list)
+ if ( !live_p2m_frame_list_list )
goto out;
live_p2m_frame_list =
xc_map_foreign_batch(xc_handle, dom, PROT_READ,
live_p2m_frame_list_list,
P2M_FLL_ENTRIES);
-
- if (!live_p2m_frame_list) {
+ if ( !live_p2m_frame_list )
+ {
ERROR("Couldn't map p2m_frame_list");
goto out;
}
@@ -756,69 +761,70 @@ int xc_linux_save(int xc_handle, int io_
live_p2m = xc_map_foreign_batch(xc_handle, dom, PROT_READ,
live_p2m_frame_list,
P2M_FL_ENTRIES);
-
- if (!live_p2m) {
+ if ( !live_p2m )
+ {
ERROR("Couldn't map p2m table");
goto out;
}
/* Setup the mfn_to_pfn table mapping */
- if(!(live_m2p = xc_map_m2p(xc_handle, max_mfn, PROT_READ))) {
+ if ( !(live_m2p = xc_map_m2p(xc_handle, max_mfn, PROT_READ)) )
+ {
ERROR("Failed to map live M2P table");
goto out;
}
/* Get a local copy of the live_P2M_frame_list */
- if(!(p2m_frame_list = malloc(P2M_FL_SIZE))) {
+ if ( !(p2m_frame_list = malloc(P2M_FL_SIZE)) )
+ {
ERROR("Couldn't allocate p2m_frame_list array");
goto out;
}
memcpy(p2m_frame_list, live_p2m_frame_list, P2M_FL_SIZE);
/* Canonicalise the pfn-to-mfn table frame-number list. */
- for (i = 0; i < p2m_size; i += fpp) {
- if (!translate_mfn_to_pfn(&p2m_frame_list[i/fpp])) {
+ for ( i = 0; i < p2m_size; i += fpp )
+ {
+ if ( !translate_mfn_to_pfn(&p2m_frame_list[i/fpp]) )
+ {
ERROR("Frame# in pfn-to-mfn frame list is not in pseudophys");
ERROR("entry %d: p2m_frame_list[%ld] is 0x%"PRIx64, i, i/fpp,
- (uint64_t)p2m_frame_list[i/fpp]);
+ (uint64_t)p2m_frame_list[i/fpp]);
goto out;
}
}
/* Domain is still running at this point */
- if (live) {
-
- if (xc_shadow_control(xc_handle, dom,
- XEN_DOMCTL_SHADOW_OP_ENABLE_LOGDIRTY,
- NULL, 0, NULL, 0, NULL) < 0) {
+ if ( live )
+ {
+ /* Live suspend. Enable log-dirty mode. */
+ if ( xc_shadow_control(xc_handle, dom,
+ XEN_DOMCTL_SHADOW_OP_ENABLE_LOGDIRTY,
+ NULL, 0, NULL, 0, NULL) < 0 )
+ {
ERROR("Couldn't enable shadow mode");
goto out;
}
-
- last_iter = 0;
-
- } else {
-
- /* This is a non-live suspend. Issue the call back to get the
- domain suspended */
-
- last_iter = 1;
-
- if (suspend_and_state(suspend, xc_handle, io_fd, dom, &info, &ctxt)) {
+ }
+ else
+ {
+ /* This is a non-live suspend. Suspend the domain .*/
+ if ( suspend_and_state(suspend, xc_handle, io_fd, dom, &info, &ctxt) )
+ {
ERROR("Domain appears not to have suspended");
goto out;
}
-
- }
+ }
+
+ last_iter = !live;
/* pretend we sent all the pages last iteration */
sent_last_iter = p2m_size;
-
/* calculate the power of 2 order of p2m_size, e.g.
15->4 16->4 17->5 */
- for (i = p2m_size-1, order_nr = 0; i ; i >>= 1, order_nr++)
+ for ( i = p2m_size-1, order_nr = 0; i ; i >>= 1, order_nr++ )
continue;
/* Setup to_send / to_fix and to_skip bitmaps */
@@ -826,20 +832,23 @@ int xc_linux_save(int xc_handle, int io_
to_fix = calloc(1, BITMAP_SIZE);
to_skip = malloc(BITMAP_SIZE);
- if (!to_send || !to_fix || !to_skip) {
+ if ( !to_send || !to_fix || !to_skip )
+ {
ERROR("Couldn't allocate to_send array");
goto out;
}
memset(to_send, 0xff, BITMAP_SIZE);
- if (lock_pages(to_send, BITMAP_SIZE)) {
+ if ( lock_pages(to_send, BITMAP_SIZE) )
+ {
ERROR("Unable to lock to_send");
return 1;
}
/* (to fix is local only) */
- if (lock_pages(to_skip, BITMAP_SIZE)) {
+ if ( lock_pages(to_skip, BITMAP_SIZE) )
+ {
ERROR("Unable to lock to_skip");
return 1;
}
@@ -849,14 +858,15 @@ int xc_linux_save(int xc_handle, int io_
/* We want zeroed memory so use calloc rather than malloc. */
pfn_type = calloc(MAX_BATCH_SIZE, sizeof(*pfn_type));
pfn_batch = calloc(MAX_BATCH_SIZE, sizeof(*pfn_batch));
-
- if ((pfn_type == NULL) || (pfn_batch == NULL)) {
+ if ( (pfn_type == NULL) || (pfn_batch == NULL) )
+ {
ERROR("failed to alloc memory for pfn_type and/or pfn_batch arrays");
errno = ENOMEM;
goto out;
}
- if (lock_pages(pfn_type, MAX_BATCH_SIZE * sizeof(*pfn_type))) {
+ if ( lock_pages(pfn_type, MAX_BATCH_SIZE * sizeof(*pfn_type)) )
+ {
ERROR("Unable to lock");
goto out;
}
@@ -867,10 +877,11 @@ int xc_linux_save(int xc_handle, int io_
{
int err=0;
unsigned long mfn;
- for (i = 0; i < p2m_size; i++) {
-
+ for ( i = 0; i < p2m_size; i++ )
+ {
mfn = live_p2m[i];
- if((mfn != INVALID_P2M_ENTRY) && (mfn_to_pfn(mfn) != i)) {
+ if( (mfn != INVALID_P2M_ENTRY) && (mfn_to_pfn(mfn) != i) )
+ {
DPRINTF("i=0x%x mfn=%lx live_m2p=%lx\n", i,
mfn, mfn_to_pfn(mfn));
err++;
@@ -879,10 +890,9 @@ int xc_linux_save(int xc_handle, int io_
DPRINTF("Had %d unexplained entries in p2m table\n", err);
}
-
/* Start writing out the saved-domain record. */
-
- if (!write_exact(io_fd, &p2m_size, sizeof(unsigned long))) {
+ if ( !write_exact(io_fd, &p2m_size, sizeof(unsigned long)) )
+ {
ERROR("write: p2m_size");
goto out;
}
@@ -892,23 +902,26 @@ int xc_linux_save(int xc_handle, int io_
* a PAE guest understands extended CR3 (PDPTs above 4GB). Turns off
* slow paths in the restore code.
*/
- if ((pt_levels == 3) &&
- (ctxt.vm_assist & (1UL << VMASST_TYPE_pae_extended_cr3))) {
+ if ( (pt_levels == 3) &&
+ (ctxt.vm_assist & (1UL << VMASST_TYPE_pae_extended_cr3)) )
+ {
unsigned long signature = ~0UL;
uint32_t tot_sz = sizeof(struct vcpu_guest_context) + 8;
uint32_t chunk_sz = sizeof(struct vcpu_guest_context);
char chunk_sig[] = "vcpu";
- if (!write_exact(io_fd, &signature, sizeof(signature)) ||
- !write_exact(io_fd, &tot_sz, sizeof(tot_sz)) ||
- !write_exact(io_fd, &chunk_sig, 4) ||
- !write_exact(io_fd, &chunk_sz, sizeof(chunk_sz)) ||
- !write_exact(io_fd, &ctxt, sizeof(ctxt))) {
+ if ( !write_exact(io_fd, &signature, sizeof(signature)) ||
+ !write_exact(io_fd, &tot_sz, sizeof(tot_sz)) ||
+ !write_exact(io_fd, &chunk_sig, 4) ||
+ !write_exact(io_fd, &chunk_sz, sizeof(chunk_sz)) ||
+ !write_exact(io_fd, &ctxt, sizeof(ctxt)) )
+ {
ERROR("write: extended info");
goto out;
}
}
- if (!write_exact(io_fd, p2m_frame_list, P2M_FL_SIZE)) {
+ if ( !write_exact(io_fd, p2m_frame_list, P2M_FL_SIZE) )
+ {
ERROR("write: p2m_frame_list");
goto out;
}
@@ -916,57 +929,65 @@ int xc_linux_save(int xc_handle, int io_
print_stats(xc_handle, dom, 0, &stats, 0);
/* Now write out each data page, canonicalising page tables as we go... */
-
- while(1) {
-
+ for ( ; ; )
+ {
unsigned int prev_pc, sent_this_iter, N, batch;
iter++;
sent_this_iter = 0;
skip_this_iter = 0;
prev_pc = 0;
- N=0;
+ N = 0;
DPRINTF("Saving memory pages: iter %d 0%%", iter);
- while( N < p2m_size ){
-
+ while ( N < p2m_size )
+ {
unsigned int this_pc = (N * 100) / p2m_size;
-
- if ((this_pc - prev_pc) >= 5) {
+ int rc;
+
+ if ( (this_pc - prev_pc) >= 5 )
+ {
DPRINTF("\b\b\b\b%3d%%", this_pc);
prev_pc = this_pc;
}
- /* slightly wasteful to peek the whole array evey time,
- but this is fast enough for the moment. */
- if (!last_iter && xc_shadow_control(
- xc_handle, dom, XEN_DOMCTL_SHADOW_OP_PEEK,
- to_skip, p2m_size, NULL, 0, NULL) != p2m_size) {
- ERROR("Error peeking shadow bitmap");
- goto out;
- }
-
+ if ( !last_iter )
+ {
+ /* Slightly wasteful to peek the whole array evey time,
+ but this is fast enough for the moment. */
+ rc = xc_shadow_control(
+ xc_handle, dom, XEN_DOMCTL_SHADOW_OP_PEEK, to_skip,
+ p2m_size, NULL, 0, NULL);
+ if ( rc != p2m_size )
+ {
+ ERROR("Error peeking shadow bitmap");
+ goto out;
+ }
+ }
/* load pfn_type[] with the mfn of all the pages we're doing in
this batch. */
- for (batch = 0; batch < MAX_BATCH_SIZE && N < p2m_size ; N++) {
-
+ for ( batch = 0;
+ (batch < MAX_BATCH_SIZE) && (N < p2m_size);
+ N++ )
+ {
int n = permute(N, p2m_size, order_nr);
- if (debug) {
+ if ( debug )
DPRINTF("%d pfn= %08lx mfn= %08lx %d [mfn]= %08lx\n",
iter, (unsigned long)n, live_p2m[n],
test_bit(n, to_send),
mfn_to_pfn(live_p2m[n]&0xFFFFF));
- }
-
- if (!last_iter && test_bit(n, to_send)&& test_bit(n, to_skip))
+
+ if ( !last_iter &&
+ test_bit(n, to_send) &&
+ test_bit(n, to_skip) )
skip_this_iter++; /* stats keeping */
- if (!((test_bit(n, to_send) && !test_bit(n, to_skip)) ||
- (test_bit(n, to_send) && last_iter) ||
- (test_bit(n, to_fix) && last_iter)))
+ if ( !((test_bit(n, to_send) && !test_bit(n, to_skip)) ||
+ (test_bit(n, to_send) && last_iter) ||
+ (test_bit(n, to_fix) && last_iter)) )
continue;
/*
@@ -979,20 +1000,22 @@ int xc_linux_save(int xc_handle, int io_
pfn_batch[batch] = n;
pfn_type[batch] = live_p2m[n];
- if(!is_mapped(pfn_type[batch])) {
-
+ if ( !is_mapped(pfn_type[batch]) )
+ {
/*
** not currently in psuedo-physical map -- set bit
** in to_fix since we must send this page in last_iter
** unless its sent sooner anyhow, or it never enters
** pseudo-physical map (e.g. for ballooned down domains)
*/
-
set_bit(n, to_fix);
continue;
}
- if(last_iter && test_bit(n, to_fix) && !test_bit(n, to_send)) {
+ if ( last_iter &&
+ test_bit(n, to_fix) &&
+ !test_bit(n, to_send) )
+ {
needed_to_fix++;
DPRINTF("Fix! iter %d, pfn %x. mfn %lx\n",
iter, n, pfn_type[batch]);
@@ -1003,11 +1026,13 @@ int xc_linux_save(int xc_handle, int io_
batch++;
}
- if (batch == 0)
+ if ( batch == 0 )
goto skip; /* vanishingly unlikely... */
- if ((region_base = xc_map_foreign_batch(
- xc_handle, dom, PROT_READ, pfn_type, batch)) == 0) {
+ region_base = xc_map_foreign_batch(
+ xc_handle, dom, PROT_READ, pfn_type, batch);
+ if ( region_base == NULL )
+ {
ERROR("map batch failed");
goto out;
}
@@ -1033,7 +1058,7 @@ int xc_linux_save(int xc_handle, int io_
continue;
}
- if (debug)
+ if ( debug )
DPRINTF("%d pfn= %08lx mfn= %08lx [mfn]= %08lx"
" sum= %08lx\n",
iter,
@@ -1049,13 +1074,15 @@ int xc_linux_save(int xc_handle, int io_
pfn_batch[j];
}
- if(!write_exact(io_fd, &batch, sizeof(unsigned int))) {
+ if ( !write_exact(io_fd, &batch, sizeof(unsigned int)) )
+ {
ERROR("Error when writing to state file (2) (errno %d)",
errno);
goto out;
}
- if(!write_exact(io_fd, pfn_type, sizeof(unsigned long)*j)) {
+ if ( !write_exact(io_fd, pfn_type, sizeof(unsigned long)*j) )
+ {
ERROR("Error when writing to state file (3) (errno %d)",
errno);
goto out;
@@ -1083,22 +1110,26 @@ int xc_linux_save(int xc_handle, int io_
race =
canonicalize_pagetable(pagetype, pfn, spage, page);
- if(race && !live) {
+ if ( race && !live )
+ {
ERROR("Fatal PT race (pfn %lx, type %08lx)", pfn,
pagetype);
goto out;
}
- if (ratewrite(io_fd, live, page, PAGE_SIZE) != PAGE_SIZE) {
+ if ( ratewrite(io_fd, live, page, PAGE_SIZE) != PAGE_SIZE )
+ {
ERROR("Error when writing to state file (4)"
" (errno %d)", errno);
goto out;
}
-
- } else {
-
+ }
+ else
+ {
/* We have a normal page: just write it directly. */
- if (ratewrite(io_fd, live, spage, PAGE_SIZE) != PAGE_SIZE)
{
+ if ( ratewrite(io_fd, live, spage, PAGE_SIZE) !=
+ PAGE_SIZE )
+ {
ERROR("Error when writing to state file (5)"
" (errno %d)", errno);
goto out;
@@ -1119,7 +1150,8 @@ int xc_linux_save(int xc_handle, int io_
DPRINTF("\r %d: sent %d, skipped %d, ",
iter, sent_this_iter, skip_this_iter );
- if (last_iter) {
+ if ( last_iter )
+ {
print_stats( xc_handle, dom, sent_this_iter, &stats, 1);
DPRINTF("Total pages sent= %ld (%.2fx)\n",
@@ -1127,14 +1159,16 @@ int xc_linux_save(int xc_handle, int io_
DPRINTF("(of which %ld were fixups)\n", needed_to_fix );
}
- if (last_iter && debug) {
+ if ( last_iter && debug )
+ {
int minusone = -1;
memset(to_send, 0xff, BITMAP_SIZE);
debug = 0;
DPRINTF("Entering debug resend-all mode\n");
/* send "-1" to put receiver into debug mode */
- if(!write_exact(io_fd, &minusone, sizeof(int))) {
+ if ( !write_exact(io_fd, &minusone, sizeof(int)) )
+ {
ERROR("Error when writing to state file (6) (errno %d)",
errno);
goto out;
@@ -1143,19 +1177,22 @@ int xc_linux_save(int xc_handle, int io_
continue;
}
- if (last_iter)
+ if ( last_iter )
break;
- if (live) {
- if (((sent_this_iter > sent_last_iter) && RATE_IS_MAX()) ||
- (iter >= max_iters) ||
- (sent_this_iter+skip_this_iter < 50) ||
- (total_sent > p2m_size*max_factor)) {
+ if ( live )
+ {
+ if ( ((sent_this_iter > sent_last_iter) && RATE_IS_MAX()) ||
+ (iter >= max_iters) ||
+ (sent_this_iter+skip_this_iter < 50) ||
+ (total_sent > p2m_size*max_factor) )
+ {
DPRINTF("Start last iteration\n");
last_iter = 1;
- if (suspend_and_state(suspend, xc_handle, io_fd, dom, &info,
- &ctxt)) {
+ if ( suspend_and_state(suspend, xc_handle, io_fd, dom, &info,
+ &ctxt) )
+ {
ERROR("Domain appears not to have suspended");
goto out;
}
@@ -1166,9 +1203,10 @@ int xc_linux_save(int xc_handle, int io_
(unsigned long)ctxt.user_regs.edx);
}
- if (xc_shadow_control(xc_handle, dom,
- XEN_DOMCTL_SHADOW_OP_CLEAN, to_send,
- p2m_size, NULL, 0, &stats) != p2m_size) {
+ if ( xc_shadow_control(xc_handle, dom,
+ XEN_DOMCTL_SHADOW_OP_CLEAN, to_send,
+ p2m_size, NULL, 0, &stats) != p2m_size )
+ {
ERROR("Error flushing shadow PT");
goto out;
}
@@ -1178,7 +1216,7 @@ int xc_linux_save(int xc_handle, int io_
print_stats(xc_handle, dom, sent_this_iter, &stats, 1);
}
- } /* end of while 1 */
+ } /* end of infinite for loop */
DPRINTF("All memory is saved\n");
@@ -1189,20 +1227,23 @@ int xc_linux_save(int xc_handle, int io_
uint64_t vcpumap;
} chunk = { -2, info.max_vcpu_id };
- if (info.max_vcpu_id >= 64) {
+ if ( info.max_vcpu_id >= 64 )
+ {
ERROR("Too many VCPUS in guest!");
goto out;
}
- for (i = 1; i <= info.max_vcpu_id; i++) {
+ for ( i = 1; i <= info.max_vcpu_id; i++ )
+ {
xc_vcpuinfo_t vinfo;
- if ((xc_vcpu_getinfo(xc_handle, dom, i, &vinfo) == 0) &&
- vinfo.online)
+ if ( (xc_vcpu_getinfo(xc_handle, dom, i, &vinfo) == 0) &&
+ vinfo.online )
vcpumap |= 1ULL << i;
}
chunk.vcpumap = vcpumap;
- if(!write_exact(io_fd, &chunk, sizeof(chunk))) {
+ if ( !write_exact(io_fd, &chunk, sizeof(chunk)) )
+ {
ERROR("Error when writing to state file (errno %d)", errno);
goto out;
}
@@ -1210,7 +1251,8 @@ int xc_linux_save(int xc_handle, int io_
/* Zero terminate */
i = 0;
- if (!write_exact(io_fd, &i, sizeof(int))) {
+ if ( !write_exact(io_fd, &i, sizeof(int)) )
+ {
ERROR("Error when writing to state file (6') (errno %d)", errno);
goto out;
}
@@ -1220,24 +1262,28 @@ int xc_linux_save(int xc_handle, int io_
unsigned int i,j;
unsigned long pfntab[1024];
- for (i = 0, j = 0; i < p2m_size; i++) {
- if (!is_mapped(live_p2m[i]))
+ for ( i = 0, j = 0; i < p2m_size; i++ )
+ {
+ if ( !is_mapped(live_p2m[i]) )
j++;
}
- if(!write_exact(io_fd, &j, sizeof(unsigned int))) {
+ if ( !write_exact(io_fd, &j, sizeof(unsigned int)) )
+ {
ERROR("Error when writing to state file (6a) (errno %d)", errno);
goto out;
}
- for (i = 0, j = 0; i < p2m_size; ) {
-
- if (!is_mapped(live_p2m[i]))
+ for ( i = 0, j = 0; i < p2m_size; )
+ {
+ if ( !is_mapped(live_p2m[i]) )
pfntab[j++] = i;
i++;
- if (j == 1024 || i == p2m_size) {
- if(!write_exact(io_fd, &pfntab, sizeof(unsigned long)*j)) {
+ if ( (j == 1024) || (i == p2m_size) )
+ {
+ if ( !write_exact(io_fd, &pfntab, sizeof(unsigned long)*j) )
+ {
ERROR("Error when writing to state file (6b) (errno %d)",
errno);
goto out;
@@ -1245,34 +1291,39 @@ int xc_linux_save(int xc_handle, int io_
j = 0;
}
}
-
}
/* Canonicalise the suspend-record frame number. */
- if ( !translate_mfn_to_pfn(&ctxt.user_regs.edx) ){
+ if ( !translate_mfn_to_pfn(&ctxt.user_regs.edx) )
+ {
ERROR("Suspend record is not in range of pseudophys map");
goto out;
}
- for (i = 0; i <= info.max_vcpu_id; i++) {
- if (!(vcpumap & (1ULL << i)))
+ for ( i = 0; i <= info.max_vcpu_id; i++ )
+ {
+ if ( !(vcpumap & (1ULL << i)) )
continue;
- if ((i != 0) && xc_vcpu_getcontext(xc_handle, dom, i, &ctxt)) {
+ if ( (i != 0) && xc_vcpu_getcontext(xc_handle, dom, i, &ctxt) )
+ {
ERROR("No context for VCPU%d", i);
goto out;
}
/* Canonicalise each GDT frame number. */
- for ( j = 0; (512*j) < ctxt.gdt_ents; j++ ) {
- if ( !translate_mfn_to_pfn(&ctxt.gdt_frames[j]) ) {
+ for ( j = 0; (512*j) < ctxt.gdt_ents; j++ )
+ {
+ if ( !translate_mfn_to_pfn(&ctxt.gdt_frames[j]) )
+ {
ERROR("GDT frame is not in range of pseudophys map");
goto out;
}
}
/* Canonicalise the page table base pointer. */
- if ( !MFN_IS_IN_PSEUDOPHYS_MAP(xen_cr3_to_pfn(ctxt.ctrlreg[3])) ) {
+ if ( !MFN_IS_IN_PSEUDOPHYS_MAP(xen_cr3_to_pfn(ctxt.ctrlreg[3])) )
+ {
ERROR("PT base is not in range of pseudophys map");
goto out;
}
@@ -1282,7 +1333,8 @@ int xc_linux_save(int xc_handle, int io_
/* Guest pagetable (x86/64) stored in otherwise-unused CR1. */
if ( (pt_levels == 4) && ctxt.ctrlreg[1] )
{
- if ( !MFN_IS_IN_PSEUDOPHYS_MAP(xen_cr3_to_pfn(ctxt.ctrlreg[1])) ) {
+ if ( !MFN_IS_IN_PSEUDOPHYS_MAP(xen_cr3_to_pfn(ctxt.ctrlreg[1])) )
+ {
ERROR("PT base is not in range of pseudophys map");
goto out;
}
@@ -1291,7 +1343,8 @@ int xc_linux_save(int xc_handle, int io_
xen_pfn_to_cr3(mfn_to_pfn(xen_cr3_to_pfn(ctxt.ctrlreg[1])));
}
- if (!write_exact(io_fd, &ctxt, sizeof(ctxt))) {
+ if ( !write_exact(io_fd, &ctxt, sizeof(ctxt)) )
+ {
ERROR("Error when writing to state file (1) (errno %d)", errno);
goto out;
}
@@ -1302,7 +1355,8 @@ int xc_linux_save(int xc_handle, int io_
*/
memcpy(page, live_shinfo, PAGE_SIZE);
((shared_info_t *)page)->arch.pfn_to_mfn_frame_list_list = 0;
- if (!write_exact(io_fd, page, PAGE_SIZE)) {
+ if ( !write_exact(io_fd, page, PAGE_SIZE) )
+ {
ERROR("Error when writing to state file (1) (errno %d)", errno);
goto out;
}
@@ -1312,30 +1366,30 @@ int xc_linux_save(int xc_handle, int io_
out:
- if (live) {
- if(xc_shadow_control(xc_handle, dom,
- XEN_DOMCTL_SHADOW_OP_OFF,
- NULL, 0, NULL, 0, NULL) < 0) {
+ if ( live )
+ {
+ if ( xc_shadow_control(xc_handle, dom,
+ XEN_DOMCTL_SHADOW_OP_OFF,
+ NULL, 0, NULL, 0, NULL) < 0 )
DPRINTF("Warning - couldn't disable shadow mode");
- }
- }
-
- // flush last write and discard cache for file
+ }
+
+ /* Flush last write and discard cache for file. */
discard_file_cache(io_fd, 1 /* flush */);
- if (live_shinfo)
+ if ( live_shinfo )
munmap(live_shinfo, PAGE_SIZE);
- if (live_p2m_frame_list_list)
+ if ( live_p2m_frame_list_list )
munmap(live_p2m_frame_list_list, PAGE_SIZE);
- if (live_p2m_frame_list)
+ if ( live_p2m_frame_list )
munmap(live_p2m_frame_list, P2M_FLL_ENTRIES * PAGE_SIZE);
- if (live_p2m)
+ if ( live_p2m )
munmap(live_p2m, ROUNDUP(p2m_size * sizeof(xen_pfn_t), PAGE_SHIFT));
- if (live_m2p)
+ if ( live_m2p )
munmap(live_m2p, M2P_SIZE(max_mfn));
free(pfn_type);
diff -r cd3cfddcb055 -r 1bde28f762a6 tools/libxc/xc_private.c
--- a/tools/libxc/xc_private.c Mon Apr 09 13:39:35 2007 -0600
+++ b/tools/libxc/xc_private.c Mon Apr 09 13:40:25 2007 -0600
@@ -145,7 +145,7 @@ int xc_mmuext_op(
return ret;
}
-static int flush_mmu_updates(int xc_handle, xc_mmu_t *mmu)
+static int flush_mmu_updates(int xc_handle, struct xc_mmu *mmu)
{
int err = 0;
DECLARE_HYPERCALL;
@@ -180,9 +180,9 @@ static int flush_mmu_updates(int xc_hand
return err;
}
-xc_mmu_t *xc_init_mmu_updates(int xc_handle, domid_t dom)
-{
- xc_mmu_t *mmu = malloc(sizeof(xc_mmu_t));
+struct xc_mmu *xc_alloc_mmu_updates(int xc_handle, domid_t dom)
+{
+ struct xc_mmu *mmu = malloc(sizeof(*mmu));
if ( mmu == NULL )
return mmu;
mmu->idx = 0;
@@ -190,7 +190,7 @@ xc_mmu_t *xc_init_mmu_updates(int xc_han
return mmu;
}
-int xc_add_mmu_update(int xc_handle, xc_mmu_t *mmu,
+int xc_add_mmu_update(int xc_handle, struct xc_mmu *mmu,
unsigned long long ptr, unsigned long long val)
{
mmu->updates[mmu->idx].ptr = ptr;
@@ -202,7 +202,7 @@ int xc_add_mmu_update(int xc_handle, xc_
return 0;
}
-int xc_finish_mmu_updates(int xc_handle, xc_mmu_t *mmu)
+int xc_flush_mmu_updates(int xc_handle, struct xc_mmu *mmu)
{
return flush_mmu_updates(xc_handle, mmu);
}
diff -r cd3cfddcb055 -r 1bde28f762a6 tools/libxc/xc_private.h
--- a/tools/libxc/xc_private.h Mon Apr 09 13:39:35 2007 -0600
+++ b/tools/libxc/xc_private.h Mon Apr 09 13:40:25 2007 -0600
@@ -168,4 +168,16 @@ void bitmap_byte_to_64(uint64_t *lp, con
/* Optionally flush file to disk and discard page cache */
void discard_file_cache(int fd, int flush);
+#define MAX_MMU_UPDATES 1024
+struct xc_mmu {
+ mmu_update_t updates[MAX_MMU_UPDATES];
+ int idx;
+ domid_t subject;
+};
+/* Structure returned by xc_alloc_mmu_updates must be free()'ed by caller. */
+struct xc_mmu *xc_alloc_mmu_updates(int xc_handle, domid_t dom);
+int xc_add_mmu_update(int xc_handle, struct xc_mmu *mmu,
+ unsigned long long ptr, unsigned long long val);
+int xc_flush_mmu_updates(int xc_handle, struct xc_mmu *mmu);
+
#endif /* __XC_PRIVATE_H__ */
diff -r cd3cfddcb055 -r 1bde28f762a6 tools/libxc/xenctrl.h
--- a/tools/libxc/xenctrl.h Mon Apr 09 13:39:35 2007 -0600
+++ b/tools/libxc/xenctrl.h Mon Apr 09 13:40:25 2007 -0600
@@ -666,21 +666,6 @@ int xc_sysctl(int xc_handle, struct xen_
int xc_version(int xc_handle, int cmd, void *arg);
-/*
- * MMU updates.
- */
-#define MAX_MMU_UPDATES 1024
-struct xc_mmu {
- mmu_update_t updates[MAX_MMU_UPDATES];
- int idx;
- domid_t subject;
-};
-typedef struct xc_mmu xc_mmu_t;
-xc_mmu_t *xc_init_mmu_updates(int xc_handle, domid_t dom);
-int xc_add_mmu_update(int xc_handle, xc_mmu_t *mmu,
- unsigned long long ptr, unsigned long long val);
-int xc_finish_mmu_updates(int xc_handle, xc_mmu_t *mmu);
-
int xc_acm_op(int xc_handle, int cmd, void *arg, unsigned long arg_size);
/*
diff -r cd3cfddcb055 -r 1bde28f762a6
unmodified_drivers/linux-2.6/compat-include/xen/platform-compat.h
--- a/unmodified_drivers/linux-2.6/compat-include/xen/platform-compat.h Mon Apr
09 13:39:35 2007 -0600
+++ b/unmodified_drivers/linux-2.6/compat-include/xen/platform-compat.h Mon Apr
09 13:40:25 2007 -0600
@@ -96,6 +96,10 @@ extern char *kasprintf(gfp_t gfp, const
#define handle_sysrq(x,y,z) handle_sysrq(x,y)
#endif
+#if defined(_PAGE_PRESENT) && !defined(_PAGE_NX)
+#define _PAGE_NX 0
+#endif
+
/*
* This variable at present is referenced by netfront, but only in code that
* is dead when running in hvm guests. To detect potential active uses of it
diff -r cd3cfddcb055 -r 1bde28f762a6
unmodified_drivers/linux-2.6/platform-pci/platform-pci.c
--- a/unmodified_drivers/linux-2.6/platform-pci/platform-pci.c Mon Apr 09
13:39:35 2007 -0600
+++ b/unmodified_drivers/linux-2.6/platform-pci/platform-pci.c Mon Apr 09
13:40:25 2007 -0600
@@ -1,6 +1,6 @@
/******************************************************************************
- * evtchn-pci.c
- * xen event channel fake PCI device driver
+ * platform-pci.c
+ * Xen platform PCI device driver
* Copyright (C) 2005, Intel Corporation.
*
* This program is free software; you can redistribute it and/or modify it
diff -r cd3cfddcb055 -r 1bde28f762a6 xen/arch/x86/mm/shadow/common.c
--- a/xen/arch/x86/mm/shadow/common.c Mon Apr 09 13:39:35 2007 -0600
+++ b/xen/arch/x86/mm/shadow/common.c Mon Apr 09 13:40:25 2007 -0600
@@ -1245,9 +1245,6 @@ static unsigned int sh_set_allocation(st
list_del(&sp->list);
d->arch.paging.shadow.free_pages -= 1<<SHADOW_MAX_ORDER;
d->arch.paging.shadow.total_pages -= 1<<SHADOW_MAX_ORDER;
- for ( j = 0; j < 1<<SHADOW_MAX_ORDER; j++ )
- /* Keep the page allocator happy */
- ((struct page_info *)sp)[j].count_info = 0;
free_domheap_pages((struct page_info *)sp, SHADOW_MAX_ORDER);
}
diff -r cd3cfddcb055 -r 1bde28f762a6 xen/arch/x86/traps.c
--- a/xen/arch/x86/traps.c Mon Apr 09 13:39:35 2007 -0600
+++ b/xen/arch/x86/traps.c Mon Apr 09 13:40:25 2007 -0600
@@ -268,21 +268,6 @@ void show_stack(struct cpu_user_regs *re
printk("\n");
show_trace(regs);
-}
-
-void show_xen_trace()
-{
- struct cpu_user_regs regs;
-#ifdef __x86_64
- __asm__("movq %%rsp,%0" : "=m" (regs.rsp));
- __asm__("movq %%rbp,%0" : "=m" (regs.rbp));
- __asm__("leaq 0(%%rip),%0" : "=a" (regs.rip));
-#else
- __asm__("movl %%esp,%0" : "=m" (regs.esp));
- __asm__("movl %%ebp,%0" : "=m" (regs.ebp));
- __asm__("call 1f; 1: popl %0" : "=a" (regs.eip));
-#endif
- show_trace(®s);
}
void show_stack_overflow(unsigned int cpu, unsigned long esp)
diff -r cd3cfddcb055 -r 1bde28f762a6 xen/arch/x86/x86_32/entry.S
--- a/xen/arch/x86/x86_32/entry.S Mon Apr 09 13:39:35 2007 -0600
+++ b/xen/arch/x86/x86_32/entry.S Mon Apr 09 13:40:25 2007 -0600
@@ -283,6 +283,7 @@ bad_hypercall:
/* %edx == trap_bounce, %ebx == struct vcpu */
/* %eax,%ecx are clobbered. %gs:%esi contain new UREGS_ss/UREGS_esp. */
create_bounce_frame:
+ ASSERT_INTERRUPTS_ENABLED
movl UREGS_eflags+4(%esp),%ecx
movb UREGS_cs+4(%esp),%cl
testl $(2|X86_EFLAGS_VM),%ecx
diff -r cd3cfddcb055 -r 1bde28f762a6 xen/arch/x86/x86_64/compat/entry.S
--- a/xen/arch/x86/x86_64/compat/entry.S Mon Apr 09 13:39:35 2007 -0600
+++ b/xen/arch/x86/x86_64/compat/entry.S Mon Apr 09 13:40:25 2007 -0600
@@ -137,6 +137,7 @@ compat_bad_hypercall:
/* %rbx: struct vcpu, interrupts disabled */
compat_restore_all_guest:
+ ASSERT_INTERRUPTS_DISABLED
RESTORE_ALL
addq $8,%rsp
.Lft0: iretq
@@ -188,13 +189,14 @@ ENTRY(compat_post_handle_exception)
ENTRY(compat_int80_direct_trap)
call compat_create_bounce_frame
- jmp compat_restore_all_guest
+ jmp compat_test_all_events
/* CREATE A BASIC EXCEPTION FRAME ON GUEST OS (RING-1) STACK: */
/* {[ERRCODE,] EIP, CS, EFLAGS, [ESP, SS]} */
/* %rdx: trap_bounce, %rbx: struct vcpu */
/* On return only %rbx is guaranteed non-clobbered. */
compat_create_bounce_frame:
+ ASSERT_INTERRUPTS_ENABLED
mov %fs,%edi
testb $2,UREGS_cs+8(%rsp)
jz 1f
diff -r cd3cfddcb055 -r 1bde28f762a6 xen/arch/x86/x86_64/entry.S
--- a/xen/arch/x86/x86_64/entry.S Mon Apr 09 13:39:35 2007 -0600
+++ b/xen/arch/x86/x86_64/entry.S Mon Apr 09 13:40:25 2007 -0600
@@ -38,6 +38,7 @@ 1: call create_bounce_frame
/* %rbx: struct vcpu, interrupts disabled */
restore_all_guest:
+ ASSERT_INTERRUPTS_DISABLED
RESTORE_ALL
testw $TRAP_syscall,4(%rsp)
jz iret_exit_to_guest
@@ -230,7 +231,7 @@ ENTRY(int80_direct_trap)
/* Check that the callback is non-null. */
leaq VCPU_int80_bounce(%rbx),%rdx
- cmp $0, TRAPBOUNCE_flags(%rdx)
+ cmp $0,TRAPBOUNCE_flags(%rdx)
jz int80_slow_path
movq VCPU_domain(%rbx),%rax
@@ -238,7 +239,7 @@ ENTRY(int80_direct_trap)
jnz compat_int80_direct_trap
call create_bounce_frame
- jmp restore_all_guest
+ jmp test_all_events
int80_slow_path:
/*
@@ -256,6 +257,7 @@ int80_slow_path:
/* %rdx: trap_bounce, %rbx: struct vcpu */
/* On return only %rbx is guaranteed non-clobbered. */
create_bounce_frame:
+ ASSERT_INTERRUPTS_ENABLED
testb $TF_kernel_mode,VCPU_thread_flags(%rbx)
jnz 1f
/* Push new frame at registered guest-OS stack base. */
diff -r cd3cfddcb055 -r 1bde28f762a6 xen/arch/x86/x86_64/traps.c
--- a/xen/arch/x86/x86_64/traps.c Mon Apr 09 13:39:35 2007 -0600
+++ b/xen/arch/x86/x86_64/traps.c Mon Apr 09 13:40:25 2007 -0600
@@ -170,7 +170,8 @@ asmlinkage void do_double_fault(struct c
regs->r9, regs->r10, regs->r11);
printk("r12: %016lx r13: %016lx r14: %016lx\n",
regs->r12, regs->r13, regs->r14);
- printk("r15: %016lx\n", regs->r15);
+ printk("r15: %016lx cs: %016lx ss: %016lx\n",
+ regs->r15, (long)regs->cs, (long)regs->ss);
show_stack_overflow(cpu, regs->rsp);
panic("DOUBLE FAULT -- system shutdown\n");
@@ -260,11 +261,14 @@ void __init percpu_traps_init(void)
idt_table[TRAP_double_fault].a |= 1UL << 32; /* IST1 */
idt_table[TRAP_nmi].a |= 2UL << 32; /* IST2 */
-#ifdef CONFIG_COMPAT
- /* The hypercall entry vector is only accessible from ring 1. */
+ /*
+ * The 32-on-64 hypercall entry vector is only accessible from ring 1.
+ * Also note that this is a trap gate, not an interrupt gate.
+ */
_set_gate(idt_table+HYPERCALL_VECTOR, 15, 1, &compat_hypercall);
+
+ /* Fast trap for int80 (faster than taking the #GP-fixup path). */
_set_gate(idt_table+0x80, 15, 3, &int80_direct_trap);
-#endif
}
stack_bottom = (char *)get_stack_bottom();
diff -r cd3cfddcb055 -r 1bde28f762a6 xen/arch/x86/x86_emulate.c
--- a/xen/arch/x86/x86_emulate.c Mon Apr 09 13:39:35 2007 -0600
+++ b/xen/arch/x86/x86_emulate.c Mon Apr 09 13:40:25 2007 -0600
@@ -2413,7 +2413,7 @@ x86_emulate(
goto writeback;
cannot_emulate:
-#ifdef __XEN__
+#if 0
gdprintk(XENLOG_DEBUG, "Instr:");
for ( ea.mem.off = ctxt->regs->eip; ea.mem.off < _regs.eip; ea.mem.off++ )
{
diff -r cd3cfddcb055 -r 1bde28f762a6 xen/common/page_alloc.c
--- a/xen/common/page_alloc.c Mon Apr 09 13:39:35 2007 -0600
+++ b/xen/common/page_alloc.c Mon Apr 09 13:40:25 2007 -0600
@@ -445,7 +445,19 @@ static void free_heap_pages(
for ( i = 0; i < (1 << order); i++ )
{
- BUG_ON(pg[i].count_info != 0);
+ /*
+ * Cannot assume that count_info == 0, as there are some corner cases
+ * where it isn't the case and yet it isn't a bug:
+ * 1. page_get_owner() is NULL
+ * 2. page_get_owner() is a domain that was never accessible by
+ * its domid (e.g., failed to fully construct the domain).
+ * 3. page was never addressable by the guest (e.g., it's an
+ * auto-translate-physmap guest and the page was never included
+ * in its pseudophysical address space).
+ * In all the above cases there can be no guest mappings of this page.
+ */
+ pg[i].count_info = 0;
+
if ( (d = page_get_owner(&pg[i])) != NULL )
{
pg[i].tlbflush_timestamp = tlbflush_current_time();
diff -r cd3cfddcb055 -r 1bde28f762a6 xen/include/asm-x86/desc.h
--- a/xen/include/asm-x86/desc.h Mon Apr 09 13:39:35 2007 -0600
+++ b/xen/include/asm-x86/desc.h Mon Apr 09 13:40:25 2007 -0600
@@ -106,7 +106,7 @@
((((sel)>>3) < FIRST_RESERVED_GDT_ENTRY) || /* Guest seg? */ \
((sel) == (!IS_COMPAT(d) ? \
FLAT_KERNEL_CS : /* Xen default seg? */ \
- FLAT_COMPAT_KERNEL_CS)) || /* Xen default compat seg? */
\
+ FLAT_COMPAT_KERNEL_CS)) || \
((sel) & 4)) /* LDT seg? */
#endif /* __ASSEMBLY__ */
diff -r cd3cfddcb055 -r 1bde28f762a6 xen/include/asm-x86/processor.h
--- a/xen/include/asm-x86/processor.h Mon Apr 09 13:39:35 2007 -0600
+++ b/xen/include/asm-x86/processor.h Mon Apr 09 13:40:25 2007 -0600
@@ -552,7 +552,6 @@ extern always_inline void prefetchw(cons
#endif
void show_stack(struct cpu_user_regs *regs);
-void show_xen_trace(void);
void show_stack_overflow(unsigned int cpu, unsigned long esp);
void show_registers(struct cpu_user_regs *regs);
void show_execution_state(struct cpu_user_regs *regs);
diff -r cd3cfddcb055 -r 1bde28f762a6 xen/include/asm-x86/x86_32/asm_defns.h
--- a/xen/include/asm-x86/x86_32/asm_defns.h Mon Apr 09 13:39:35 2007 -0600
+++ b/xen/include/asm-x86/x86_32/asm_defns.h Mon Apr 09 13:40:25 2007 -0600
@@ -8,9 +8,19 @@
#define SETUP_EXCEPTION_FRAME_POINTER \
movl %esp,%ebp; \
notl %ebp
+#define ASSERT_INTERRUPT_STATUS(x) \
+ pushf; \
+ testb $X86_EFLAGS_IF>>8,1(%esp); \
+ j##x 1f; \
+ ud2a; \
+1: addl $4,%esp;
#else
#define SETUP_EXCEPTION_FRAME_POINTER
+#define ASSERT_INTERRUPT_STATUS(x)
#endif
+
+#define ASSERT_INTERRUPTS_ENABLED ASSERT_INTERRUPT_STATUS(nz)
+#define ASSERT_INTERRUPTS_DISABLED ASSERT_INTERRUPT_STATUS(z)
#define __SAVE_ALL_PRE \
cld; \
diff -r cd3cfddcb055 -r 1bde28f762a6 xen/include/asm-x86/x86_64/asm_defns.h
--- a/xen/include/asm-x86/x86_64/asm_defns.h Mon Apr 09 13:39:35 2007 -0600
+++ b/xen/include/asm-x86/x86_64/asm_defns.h Mon Apr 09 13:40:25 2007 -0600
@@ -8,9 +8,19 @@
#define SETUP_EXCEPTION_FRAME_POINTER \
movq %rsp,%rbp; \
notq %rbp
+#define ASSERT_INTERRUPT_STATUS(x) \
+ pushf; \
+ testb $X86_EFLAGS_IF>>8,1(%rsp); \
+ j##x 1f; \
+ ud2a; \
+1: addq $8,%rsp;
#else
#define SETUP_EXCEPTION_FRAME_POINTER
+#define ASSERT_INTERRUPT_STATUS(x)
#endif
+
+#define ASSERT_INTERRUPTS_ENABLED ASSERT_INTERRUPT_STATUS(nz)
+#define ASSERT_INTERRUPTS_DISABLED ASSERT_INTERRUPT_STATUS(z)
#define SAVE_ALL \
cld; \
diff -r cd3cfddcb055 -r 1bde28f762a6 xen/include/public/domctl.h
--- a/xen/include/public/domctl.h Mon Apr 09 13:39:35 2007 -0600
+++ b/xen/include/public/domctl.h Mon Apr 09 13:40:25 2007 -0600
@@ -123,15 +123,15 @@ DEFINE_XEN_GUEST_HANDLE(xen_domctl_getme
#define XEN_DOMCTL_getpageframeinfo 7
#define XEN_DOMCTL_PFINFO_LTAB_SHIFT 28
-#define XEN_DOMCTL_PFINFO_NOTAB (0x0<<28)
-#define XEN_DOMCTL_PFINFO_L1TAB (0x1<<28)
-#define XEN_DOMCTL_PFINFO_L2TAB (0x2<<28)
-#define XEN_DOMCTL_PFINFO_L3TAB (0x3<<28)
-#define XEN_DOMCTL_PFINFO_L4TAB (0x4<<28)
-#define XEN_DOMCTL_PFINFO_LTABTYPE_MASK (0x7<<28)
-#define XEN_DOMCTL_PFINFO_LPINTAB (0x1<<31)
-#define XEN_DOMCTL_PFINFO_XTAB (0xf<<28) /* invalid page */
-#define XEN_DOMCTL_PFINFO_LTAB_MASK (0xf<<28)
+#define XEN_DOMCTL_PFINFO_NOTAB (0x0U<<28)
+#define XEN_DOMCTL_PFINFO_L1TAB (0x1U<<28)
+#define XEN_DOMCTL_PFINFO_L2TAB (0x2U<<28)
+#define XEN_DOMCTL_PFINFO_L3TAB (0x3U<<28)
+#define XEN_DOMCTL_PFINFO_L4TAB (0x4U<<28)
+#define XEN_DOMCTL_PFINFO_LTABTYPE_MASK (0x7U<<28)
+#define XEN_DOMCTL_PFINFO_LPINTAB (0x1U<<31)
+#define XEN_DOMCTL_PFINFO_XTAB (0xfU<<28) /* invalid page */
+#define XEN_DOMCTL_PFINFO_LTAB_MASK (0xfU<<28)
struct xen_domctl_getpageframeinfo {
/* IN variables. */
_______________________________________________
Xen-changelog mailing list
Xen-changelog@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-changelog
|
Home