xen-mapcache: Avoid entry->lock overflow
PATCH 1/1 xen-mapcache: Avoid entry->lock overflow
In some cases, a particular mapcache entry may be mapped 256 times causing the lock field to wrap to 0. For example, this may happen when using emulated NVME and the guest submits a large scatter-gather write. At this point, the entry map be remapped causing QEMU to write the wrong data or crash (since remap is not atomic).
Avoid this overflow by increasing the lock field to a uint16_t and also detect it and abort rather than continuing regardless.
--- hw/i386/xen/xen-mapcache.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-)
hw/i386/xen/xen-mapcache.c +7 -1
@@ -52,7 +52,7 @@ typedef struct MapCacheEntry {hwaddr paddr_index;uint8_t *vaddr_base;unsigned long *valid_mapping; -uint8_t lock; +uint16_t lock;#define XEN_MAPCACHE_ENTRY_DUMMY (1 << 0)uint8_t flags;hwaddr size;
@@ -355,6 +355,12 @@ tryagain:if (lock) {MapCacheRev *reventry =g_malloc0 (sizeof (MapCacheRev)); entry->lock++; +if (entry->lock ==0 ) { +fprintf (stderr, +"mapcache entry lock overflow: " TARGET_FMT_plx" -> %p\n" , + entry->paddr_index, entry->vaddr_base); +abort (); + } reventry->dma = dma; reventry->vaddr_req = mapcache->last_entry->vaddr_base + address_offset; reventry->paddr_index = mapcache->last_entry->paddr_index;
Signature · 1 line
2.27.0
Quoted 8 lines · Ross Lagerwall
In some cases, a particular mapcache entry may be mapped 256 times causing the lock field to wrap to 0. For example, this may happen when using emulated NVME and the guest submits a large scatter-gather write. At this point, the entry map be remapped causing QEMU to write the wrong data or crash (since remap is not atomic).
Avoid this overflow by increasing the lock field to a
uint16_tand also detect it and abort rather than continuing regardless.
Nice catch!
Quoted 10 lines
Signed-off-by Ross Lagerwall <[email protected]>--- hw/i386/xen/xen-mapcache.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-)hw/i386/xen/xen-mapcache.c +0
-0
hw/i386/xen/xen-mapcache.c typedef struct MapCacheEntry { · line 55
Quoted 5 lines
hw/i386/xen/xen-mapcache.c +1
-1@@ -52,7 +52,7 @@ typedef struct MapCacheEntry {hwaddr paddr_index;uint8_t *vaddr_base;unsigned long *valid_mapping; -uint8_t lock; +uint16_t lock;
As struct MapCacheEntry is not packed, we might as well switch to uint32_t as it doesn't make a difference.
hw/i386/xen/xen-mapcache.c tryagain: · line 362
Quoted 12 lines
hw/i386/xen/xen-mapcache.c +5
-0#define XEN_MAPCACHE_ENTRY_DUMMY (1 << 0)uint8_t flags;hwaddr size;@@ -355,6 +355,12 @@ tryagain:if (lock) {MapCacheRev *reventry =g_malloc0 (sizeof (MapCacheRev)); entry->lock++; +if (entry->lock ==0 ) { +fprintf (stderr, +"mapcache entry lock overflow: " TARGET_FMT_plx" -> %p\n" , + entry->paddr_index, entry->vaddr_base); +abort ();
I was going to suggest that it might be better to return NULL and let the caller handle the failure but it looks like the caller *cannot* handle the failure.
Quoted 7 lines
hw/i386/xen/xen-mapcache.c +1
-0+ } reventry->dma = dma; reventry->vaddr_req = mapcache->last_entry->vaddr_base + address_offset; reventry->paddr_index = mapcache->last_entry->paddr_index;Signature · 1 line
2.27.0