qcow2: Fix cluster allocation
PATCH 1/1 qcow2: Fix cluster allocation
When allocating multiple clusters at once, the qcow2 implementation tries to find as many physically contiguous clusters as possible to allow larger writes. This search includes allocated clusters which are in the right place and still free clusters. If the range to allocate spans clusters in patterns like "10 allocated, then 10 free, then again 10 allocated" it is only checked that the chunks of allocated clusters are contiguous for themselves.
However, what is actually needed is to have _all_ allocated clusters contiguous, starting at the first cluster of the allocation and spanning multiple such chunks. This patch changes the check so that each offset is not compared to the offset of the first cluster in its own chunk but to the first cluster in the whole allocation.
I haven't seen it happen, but without this fix data corruption on qcow2 images is possible.
This patch is meant to be applied on top of Gleb's compressed image fix.
block-qcow2.c +6 -6
@@ -615,7 +615,7 @@ static int size_to_clusters(BDRVQcowStat }static int count_contiguous_clusters (uint64_t nb_clusters,int cluster_size, -uint64_t *l2_table,uint64_t mask) +uint64_t *l2_table,uint64_t start,uint64_t mask) {int i;uint64_t offset =be64_to_cpu (l2_table[0 ]) & ~mask;
@@ -623,11 +623,11 @@ static int count_contiguous_clusters(uinif (!offset)return 0 ; -for (i =0 ; i < nb_clusters; i++) +for (i = start; i < start + nb_clusters; i++)if (offset + i * cluster_size != (be64_to_cpu (l2_table[i]) & ~mask))break ; -return i; +return (i - start); }static int count_contiguous_free_clusters (uint64_t nb_clusters,uint64_t *l2_table)
@@ -714,7 +714,7 @@ static uint64_t get_cluster_offset(Block }else {/* how many allocated clusters ? */ c =count_contiguous_clusters (nb_clusters, s->cluster_size, - &l2_table[l2_index],QCOW_OFLAG_COPIED ); + &l2_table[l2_index],0 ,QCOW_OFLAG_COPIED ); } nb_available = (c * s->cluster_sectors);
@@ -968,7 +968,7 @@ static uint64_t alloc_cluster_offset(Bloif (cluster_offset &QCOW_OFLAG_COPIED ) { nb_clusters =count_contiguous_clusters (nb_clusters, s->cluster_size, - &l2_table[l2_index],0 ); + &l2_table[l2_index],0 ,0 ); cluster_offset &= ~QCOW_OFLAG_COPIED ; m->nb_clusters =0 ;
@@ -985,7 +985,7 @@ static uint64_t alloc_cluster_offset(Blowhile (i < nb_clusters) { i +=count_contiguous_clusters (nb_clusters - i, s->cluster_size, - &l2_table[l2_index + i],0 ); + &l2_table[l2_index], i,0 );if (be64_to_cpu (l2_table[l2_index + i]))break ;
Quoted 14 lines · Kevin Wolf
When allocating multiple clusters at once, the qcow2 implementation tries to find as many physically contiguous clusters as possible to allow larger writes. This search includes allocated clusters which are in the right place and still free clusters. If the range to allocate spans clusters in patterns like "10 allocated, then 10 free, then again 10 allocated" it is only checked that the chunks of allocated clusters are contiguous for themselves.
However, what is actually needed is to have _all_ allocated clusters contiguous, starting at the first cluster of the allocation and spanning multiple such chunks. This patch changes the check so that each offset is not compared to the offset of the first cluster in its own chunk but to the first cluster in the whole allocation.
Quoted 6 lines
I haven't seen it happen, but without this fix data corruption on qcow2 images is possible.
This patch is meant to be applied on top of Gleb's compressed image fix.
Signed-off-by Kevin Wolf <[email protected]>
Quoted 46 lines
Index:
qemu-svn/block-qcow2.c
===================================================================block-qcow2.c +5
-5@@ -615,7 +615,7 @@ static int size_to_clusters(BDRVQcowStat }static int count_contiguous_clusters (uint64_t nb_clusters,int cluster_size, -uint64_t *l2_table,uint64_t mask) +uint64_t *l2_table,uint64_t start,uint64_t mask) {int i;uint64_t offset =be64_to_cpu (l2_table[0 ]) & ~mask;@@ -623,11 +623,11 @@ static int count_contiguous_clusters(uinif (!offset)return 0 ; -for (i =0 ; i < nb_clusters; i++) +for (i = start; i < start + nb_clusters; i++)if (offset + i * cluster_size != (be64_to_cpu (l2_table[i]) & ~mask))break ; -return i; +return (i - start); }static int count_contiguous_free_clusters (uint64_t nb_clusters,uint64_t *l2_table)@@ -714,7 +714,7 @@ static uint64_t get_cluster_offset(Block }else {/* how many allocated clusters ? */ c =count_contiguous_clusters (nb_clusters, s->cluster_size, - &l2_table[l2_index],QCOW_OFLAG_COPIED ); + &l2_table[l2_index],0 ,QCOW_OFLAG_COPIED ); } nb_available = (c * s->cluster_sectors);@@ -968,7 +968,7 @@ static uint64_t alloc_cluster_offset(Bloif (cluster_offset &QCOW_OFLAG_COPIED ) { nb_clusters =count_contiguous_clusters (nb_clusters, s->cluster_size, - &l2_table[l2_index],0 ); + &l2_table[l2_index],0 ,0 ); cluster_offset &= ~QCOW_OFLAG_COPIED ; m->nb_clusters =0 ;
block-qcow2.c static uint64_t alloc_cluster_offset(Blo · line 991
Quoted 8 lines
block-qcow2.c +1
-1@@ -985,7 +985,7 @@ static uint64_t alloc_cluster_offset(Blowhile (i < nb_clusters) { i +=count_contiguous_clusters (nb_clusters - i, s->cluster_size, - &l2_table[l2_index + i],0 ); + &l2_table[l2_index], i,0 );if (be64_to_cpu (l2_table[l2_index + i]))break ;
--
Gleb.