Skip to contentSkip to conversation

Block patches for 0.12.x

GIT PULL New Kevin Wolf 2 participants 8 messages May 28, 2010 – Jun 9, 2010

lore
Share & tools
lore

Kevin Wolf

author
#0 lore raw

The following changes since commit 0c459361a1117a6c434c7b2b008a4c6c035eb4bf:
Rabin Vincent (1):
arm_timer: fix oneshot mode

are available in the git repository at:

git://repo.or.cz/qemu/kevin.git for-stable-0.12

Kevin Wolf (6):
ide: Fix ide_dma_cancel
qcow2: Clear L2 table cache after write error
qcow2: Fix error handling in l2_allocate
block: Fix multiwrite with overlapping requests
qcow2: Fix corruption after refblock allocation
qcow2: Fix corruption after error in update_refcount

 block.c                |    2 +-
 block/qcow2-cluster.c  |   26 +++++++++++++++-----------
 block/qcow2-refcount.c |   15 +++++++++++++--
 hw/ide/core.c          |    8 ++++----
 4 files changed, 33 insertions(+), 18 deletions(-)
author
#1 lore raw

PATCH 1/6 ide: Fix ide_dma_cancel

When cancelling a request, bdrv_aio_cancel may decide that it waits for completion of a request rather than for cancellation. IDE therefore can't abandon its DMA status before calling bdrv_aio_cancel; otherwise the callback of a completed request would use invalid data.

(cherry picked from commit 38d8dfa193e9a45f0f08b06aab2ba2a94f40a041)
---
 hw/ide/core.c |    8 ++++----
 1 files changed, 4 insertions(+), 4 deletions(-)
hw/ide/core.c +4 -4
index 64aebc2..f9bb338 100644
@@ -2827,10 +2827,6 @@ static void ide_dma_restart(IDEState *s, int is_read)
 void ide_dma_cancel(BMDMAState *bm)
 {
     if (bm->status & BM_STATUS_DMAING) {
-        bm->status &= ~BM_STATUS_DMAING;
-        /* cancel DMA request */
-        bm->unit = -1;
-        bm->dma_cb = NULL;
         if (bm->aiocb) {
 #ifdef DEBUG_AIO
             printf("aio_cancel\n");
@@ -2838,6 +2834,10 @@ void ide_dma_cancel(BMDMAState *bm)
             bdrv_aio_cancel(bm->aiocb);
             bm->aiocb = NULL;
         }
+        bm->status &= ~BM_STATUS_DMAING;
+        /* cancel DMA request */
+        bm->unit = -1;
+        bm->dma_cb = NULL;
     }
 }
 
Signature · 1 line
1.6.6.1
author
#2 lore raw

PATCH 2/6 qcow2: Clear L2 table cache after write error

In reply to Kevin Wolf

If the L2 table was already updated in cache, but writing it to disk has failed, we must not continue using the changed version in the cache to stay consistent with what's on the disk.

(cherry picked from commit 1b7c801b40ce90795397bb566d019c9b76ef9c13)

Conflicts:

block/qcow2-cluster.c

---
 block/qcow2-cluster.c |    5 +++--
 1 files changed, 3 insertions(+), 2 deletions(-)
block/qcow2-cluster.c +3 -2
index c7057b1..0dc4f1d 100644
@@ -672,8 +672,9 @@ int qcow2_alloc_cluster_link_l2(BlockDriverState *bs, QCowL2Meta *m)
                     (i << s->cluster_bits)) | QCOW_OFLAG_COPIED);
      }
 
-    if (write_l2_entries(s, l2_table, l2_offset, l2_index, m->nb_clusters) < 0) {
-        ret = -1;
+    ret = write_l2_entries(s, l2_table, l2_offset, l2_index, m->nb_clusters);
+    if (ret < 0) {
+        qcow2_l2_cache_reset(bs);
         goto err;
     }
 
Signature · 1 line
1.6.6.1
author
#3 lore raw

PATCH 3/6 qcow2: Fix error handling in l2_allocate

In reply to Kevin Wolf

l2_allocate has some intermediate states in which the image is inconsistent. Change the order to write to the L1 table only after the new L2 table has successfully been initialized.

Also reset the L2 cache in failure case, it's very likely wrong.

(cherry picked from commit 175e11526e2613b3dc031c23fec3107aa4a80307)

Conflicts:

block/qcow2-cluster.c

---
 block/qcow2-cluster.c |   21 ++++++++++++---------
 1 files changed, 12 insertions(+), 9 deletions(-)
block/qcow2-cluster.c +12 -9
index 0dc4f1d..b7a5b35 100644
@@ -231,13 +231,6 @@ static uint64_t *l2_allocate(BlockDriverState *bs, int l1_index)
         return NULL;
     }
 
-    /* update the L1 entry */
-
-    s->l1_table[l1_index] = l2_offset | QCOW_OFLAG_COPIED;
-    if (write_l1_entry(s, l1_index) < 0) {
-        return NULL;
-    }
-
     /* allocate a new entry in the l2 cache */
 
     min_index = l2_cache_new_entry(bs);
@@ -251,13 +244,19 @@ static uint64_t *l2_allocate(BlockDriverState *bs, int l1_index)
         if (bdrv_pread(s->hd, old_l2_offset,
                        l2_table, s->l2_size * sizeof(uint64_t)) !=
             s->l2_size * sizeof(uint64_t))
-            return NULL;
+            goto fail;
     }
     /* write the l2 table to the file */
     if (bdrv_pwrite(s->hd, l2_offset,
                     l2_table, s->l2_size * sizeof(uint64_t)) !=
         s->l2_size * sizeof(uint64_t))
-        return NULL;
+        goto fail;
+
+    /* update the L1 entry */
+    s->l1_table[l1_index] = l2_offset | QCOW_OFLAG_COPIED;
+    if (write_l1_entry(s, l1_index) < 0) {
+        goto fail;
+    }
 
     /* update the l2 cache entry */
 
@@ -265,6 +264,10 @@ static uint64_t *l2_allocate(BlockDriverState *bs, int l1_index)
     s->l2_cache_counts[min_index] = 1;
 
     return l2_table;
+
+fail:
+    qcow2_l2_cache_reset(bs);
+    return NULL;
 }
 
 static int count_contiguous_clusters(uint64_t nb_clusters, int cluster_size,
Signature · 1 line
1.6.6.1
author
#4 lore raw

PATCH 4/6 block: Fix multiwrite with overlapping requests

In reply to Kevin Wolf

With overlapping requests, the total number of sectors is smaller than the sum of the nb_sectors of both requests.

(cherry picked from commit cbf1dff2f1033cadcb15c0ffc9c0a3d039d8ed42)
---
 block.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
block.c +1 -1
index 955eeaa..298414c 100644
@@ -1711,7 +1711,7 @@ static int multiwrite_merge(BlockDriverState *bs, BlockRequest *reqs,
             // Add the second request
             qemu_iovec_concat(qiov, reqs[i].qiov, reqs[i].qiov->size);
 
-            reqs[outidx].nb_sectors += reqs[i].nb_sectors;
+            reqs[outidx].nb_sectors = qiov->size >> 9;
             reqs[outidx].qiov = qiov;
 
             mcb->callbacks[i].free_qiov = reqs[outidx].qiov;
Signature · 1 line
1.6.6.1
author
#5 lore raw

PATCH 5/6 qcow2: Fix corruption after refblock allocation

In reply to Kevin Wolf

Refblock allocation code needs to take into consideration that update_refcount will load a different refcount block into the cache, so it must initialize the cache for a new refcount block only afterwards. Not doing this means that not only the refcount in the wrong block is updated, but also that the caller will work on the wrong block.

(cherry picked from commit 25408c09502be036e5575754fe54019ed4ed5dfa)
---
 block/qcow2-refcount.c |   11 +++++++++--
 1 files changed, 9 insertions(+), 2 deletions(-)
block/qcow2-refcount.c +9 -2
index 5ebbcb6..fa78e46 100644
@@ -215,8 +215,6 @@ static int64_t alloc_refcount_block(BlockDriverState *bs, int64_t cluster_index)
 
     /* Allocate the refcount block itself and mark it as used */
     uint64_t new_block = alloc_clusters_noref(bs, s->cluster_size);
-    memset(s->refcount_block_cache, 0, s->cluster_size);
-    s->refcount_block_cache_offset = new_block;
 
 #ifdef DEBUG_ALLOC2
     fprintf(stderr, "qcow2: Allocate refcount block %d for %" PRIx64
@@ -225,6 +223,10 @@ static int64_t alloc_refcount_block(BlockDriverState *bs, int64_t cluster_index)
 #endif
 
     if (in_same_refcount_block(s, new_block, cluster_index << s->cluster_bits)) {
+        /* Zero the new refcount block before updating it */
+        memset(s->refcount_block_cache, 0, s->cluster_size);
+        s->refcount_block_cache_offset = new_block;
+
         /* The block describes itself, need to update the cache */
         int block_index = (new_block >> s->cluster_bits) &
             ((1 << (s->cluster_bits - REFCOUNT_SHIFT)) - 1);
@@ -236,6 +238,11 @@ static int64_t alloc_refcount_block(BlockDriverState *bs, int64_t cluster_index)
         if (ret < 0) {
             goto fail_block;
         }
+
+        /* Initialize the new refcount block only after updating its refcount,
+         * update_refcount uses the refcount cache itself */
+        memset(s->refcount_block_cache, 0, s->cluster_size);
+        s->refcount_block_cache_offset = new_block;
     }
 
     /* Now the new refcount block needs to be written to disk */
Signature · 1 line
1.6.6.1
author
#6 lore raw

PATCH 6/6 qcow2: Fix corruption after error in update_refcount

In reply to Kevin Wolf

After it is done with updating refcounts in the cache, update_refcount writes all changed entries to disk. If a refcount block allocation fails, however, there was no change yet and therefore first_index = last_index = -1. Don't treat -1 as a normal sector index (resulting in a 512 byte write!) but return without updating anything in this case.

(cherry picked from commit 86fa8da83771238de55dc44819a1a27bafef5353)
---
 block/qcow2-refcount.c |    4 ++++
 1 files changed, 4 insertions(+), 0 deletions(-)
block/qcow2-refcount.c +4 -0
index fa78e46..465d5d3 100644
@@ -402,6 +402,10 @@ static int write_refcount_block_entries(BDRVQcowState *s,
         return 0;
     }
 
+    if (first_index < 0) {
+        return 0;
+    }
+
     first_index &= ~(REFCOUNTS_PER_SECTOR - 1);
     last_index = (last_index + REFCOUNTS_PER_SECTOR)
         & ~(REFCOUNTS_PER_SECTOR - 1);
Signature · 1 line
1.6.6.1
#7 lore raw
In reply to Kevin Wolf
Quoted 16 lines · Kevin Wolf

The following changes since commit 0c459361a1117a6c434c7b2b008a4c6c035eb4bf:
Rabin Vincent (1):
arm_timer: fix oneshot mode

are available in the git repository at:

git://repo.or.cz/qemu/kevin.git for-stable-0.12

Kevin Wolf (6):
ide: Fix ide_dma_cancel
qcow2: Clear L2 table cache after write error
qcow2: Fix error handling in l2_allocate
block: Fix multiwrite with overlapping requests
qcow2: Fix corruption after refblock allocation
qcow2: Fix corruption after error in update_refcount

Quoted 7 lines
 block.c                |    2 +-
 block/qcow2-cluster.c  |   26 +++++++++++++++-----------
 block/qcow2-refcount.c |   15 +++++++++++++--
 hw/ide/core.c          |    8 ++++----
 4 files changed, 33 insertions(+), 18 deletions(-)

Thanks, all applied.

Signature · 2 lines
Aurelien Jarno                          GPG: 1024D/F1BCDB73
[email protected]                 http://www.aurel32.net