* [Qemu-devel] [PATCH v2 1/9] qcow2: Fix leaks in dirty images
2014-08-15 15:16 [Qemu-devel] [PATCH v2 0/9] qcow2: Fix image repairing Max Reitz
@ 2014-08-15 15:16 ` Max Reitz
2014-08-22 16:09 ` Eric Blake
2014-08-15 15:16 ` [Qemu-devel] [PATCH v2 2/9] qcow2: Factor out refcount accounting for check Max Reitz
` (7 subsequent siblings)
8 siblings, 1 reply; 20+ messages in thread
From: Max Reitz @ 2014-08-15 15:16 UTC (permalink / raw)
To: qemu-devel; +Cc: Kevin Wolf, Benoît Canet, Stefan Hajnoczi, Max Reitz
When opening dirty images, qcow2's repair function should not only
repair errors but leaks as well.
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
block/qcow2.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/block/qcow2.c b/block/qcow2.c
index 435e0e1..8faa75d 100644
--- a/block/qcow2.c
+++ b/block/qcow2.c
@@ -790,7 +790,7 @@ static int qcow2_open(BlockDriverState *bs, QDict *options, int flags,
(s->incompatible_features & QCOW2_INCOMPAT_DIRTY)) {
BdrvCheckResult result = {0};
- ret = qcow2_check(bs, &result, BDRV_FIX_ERRORS);
+ ret = qcow2_check(bs, &result, BDRV_FIX_ERRORS | BDRV_FIX_LEAKS);
if (ret < 0) {
error_setg_errno(errp, -ret, "Could not repair dirty image");
goto fail;
--
2.0.3
^ permalink raw reply related [flat|nested] 20+ messages in thread* Re: [Qemu-devel] [PATCH v2 1/9] qcow2: Fix leaks in dirty images
2014-08-15 15:16 ` [Qemu-devel] [PATCH v2 1/9] qcow2: Fix leaks in dirty images Max Reitz
@ 2014-08-22 16:09 ` Eric Blake
0 siblings, 0 replies; 20+ messages in thread
From: Eric Blake @ 2014-08-22 16:09 UTC (permalink / raw)
To: Max Reitz, qemu-devel; +Cc: Kevin Wolf, Stefan Hajnoczi, Benoît Canet
[-- Attachment #1: Type: text/plain, Size: 1052 bytes --]
On 08/15/2014 09:16 AM, Max Reitz wrote:
> When opening dirty images, qcow2's repair function should not only
> repair errors but leaks as well.
>
> Signed-off-by: Max Reitz <mreitz@redhat.com>
> ---
> block/qcow2.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
Reviewed-by: Eric Blake <eblake@redhat.com>
>
> diff --git a/block/qcow2.c b/block/qcow2.c
> index 435e0e1..8faa75d 100644
> --- a/block/qcow2.c
> +++ b/block/qcow2.c
> @@ -790,7 +790,7 @@ static int qcow2_open(BlockDriverState *bs, QDict *options, int flags,
> (s->incompatible_features & QCOW2_INCOMPAT_DIRTY)) {
> BdrvCheckResult result = {0};
>
> - ret = qcow2_check(bs, &result, BDRV_FIX_ERRORS);
> + ret = qcow2_check(bs, &result, BDRV_FIX_ERRORS | BDRV_FIX_LEAKS);
> if (ret < 0) {
> error_setg_errno(errp, -ret, "Could not repair dirty image");
> goto fail;
>
--
Eric Blake eblake redhat com +1-919-301-3266
Libvirt virtualization library http://libvirt.org
[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 539 bytes --]
^ permalink raw reply [flat|nested] 20+ messages in thread
* [Qemu-devel] [PATCH v2 2/9] qcow2: Factor out refcount accounting for check
2014-08-15 15:16 [Qemu-devel] [PATCH v2 0/9] qcow2: Fix image repairing Max Reitz
2014-08-15 15:16 ` [Qemu-devel] [PATCH v2 1/9] qcow2: Fix leaks in dirty images Max Reitz
@ 2014-08-15 15:16 ` Max Reitz
2014-08-21 18:47 ` Benoît Canet
2014-08-21 18:57 ` Benoît Canet
2014-08-15 15:16 ` [Qemu-devel] [PATCH v2 3/9] qcow2: Factor out refcount comparison " Max Reitz
` (6 subsequent siblings)
8 siblings, 2 replies; 20+ messages in thread
From: Max Reitz @ 2014-08-15 15:16 UTC (permalink / raw)
To: qemu-devel; +Cc: Kevin Wolf, Benoît Canet, Stefan Hajnoczi, Max Reitz
Put the code for calculating the reference counts during qemu-img check
into an own function.
Also, do not use g_realloc() for increasing the size of the in-memory
refcount table, but rather g_try_realloc().
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
block/qcow2-refcount.c | 188 ++++++++++++++++++++++++++++++-------------------
1 file changed, 115 insertions(+), 73 deletions(-)
diff --git a/block/qcow2-refcount.c b/block/qcow2-refcount.c
index d60e2fe..b3c4edd 100644
--- a/block/qcow2-refcount.c
+++ b/block/qcow2-refcount.c
@@ -1496,69 +1496,15 @@ done:
}
/*
- * Checks an image for refcount consistency.
- *
- * Returns 0 if no errors are found, the number of errors in case the image is
- * detected as corrupted, and -errno when an internal error occurred.
+ * Checks consistency of refblocks and accounts for each refblock in
+ * *refcount_table.
*/
-int qcow2_check_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
- BdrvCheckMode fix)
+static int check_refblocks(BlockDriverState *bs, BdrvCheckResult *res,
+ BdrvCheckMode fix, uint16_t **refcount_table,
+ int64_t *nb_clusters)
{
BDRVQcowState *s = bs->opaque;
- int64_t size, i, highest_cluster, nb_clusters;
- int refcount1, refcount2;
- QCowSnapshot *sn;
- uint16_t *refcount_table;
- int ret;
-
- size = bdrv_getlength(bs->file);
- if (size < 0) {
- res->check_errors++;
- return size;
- }
-
- nb_clusters = size_to_clusters(s, size);
- if (nb_clusters > INT_MAX) {
- res->check_errors++;
- return -EFBIG;
- }
-
- refcount_table = g_try_malloc0(nb_clusters * sizeof(uint16_t));
- if (nb_clusters && refcount_table == NULL) {
- res->check_errors++;
- return -ENOMEM;
- }
-
- res->bfi.total_clusters =
- size_to_clusters(s, bs->total_sectors * BDRV_SECTOR_SIZE);
-
- /* header */
- inc_refcounts(bs, res, refcount_table, nb_clusters,
- 0, s->cluster_size);
-
- /* current L1 table */
- ret = check_refcounts_l1(bs, res, refcount_table, nb_clusters,
- s->l1_table_offset, s->l1_size, CHECK_FRAG_INFO);
- if (ret < 0) {
- goto fail;
- }
-
- /* snapshots */
- for(i = 0; i < s->nb_snapshots; i++) {
- sn = s->snapshots + i;
- ret = check_refcounts_l1(bs, res, refcount_table, nb_clusters,
- sn->l1_table_offset, sn->l1_size, 0);
- if (ret < 0) {
- goto fail;
- }
- }
- inc_refcounts(bs, res, refcount_table, nb_clusters,
- s->snapshots_offset, s->snapshots_size);
-
- /* refcount data */
- inc_refcounts(bs, res, refcount_table, nb_clusters,
- s->refcount_table_offset,
- s->refcount_table_size * sizeof(uint64_t));
+ int64_t i;
for(i = 0; i < s->refcount_table_size; i++) {
uint64_t offset, cluster;
@@ -1573,7 +1519,7 @@ int qcow2_check_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
continue;
}
- if (cluster >= nb_clusters) {
+ if (cluster >= *nb_clusters) {
fprintf(stderr, "ERROR refcount block %" PRId64
" is outside image\n", i);
res->corruptions++;
@@ -1581,14 +1527,14 @@ int qcow2_check_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
}
if (offset != 0) {
- inc_refcounts(bs, res, refcount_table, nb_clusters,
+ inc_refcounts(bs, res, *refcount_table, *nb_clusters,
offset, s->cluster_size);
- if (refcount_table[cluster] != 1) {
+ if ((*refcount_table)[cluster] != 1) {
fprintf(stderr, "%s refcount block %" PRId64
" refcount=%d\n",
fix & BDRV_FIX_ERRORS ? "Repairing" :
"ERROR",
- i, refcount_table[cluster]);
+ i, (*refcount_table)[cluster]);
if (fix & BDRV_FIX_ERRORS) {
int64_t new_offset;
@@ -1600,17 +1546,24 @@ int qcow2_check_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
}
/* update refcounts */
- if ((new_offset >> s->cluster_bits) >= nb_clusters) {
+ if ((new_offset >> s->cluster_bits) >= *nb_clusters) {
/* increase refcount_table size if necessary */
- int old_nb_clusters = nb_clusters;
- nb_clusters = (new_offset >> s->cluster_bits) + 1;
- refcount_table = g_realloc(refcount_table,
- nb_clusters * sizeof(uint16_t));
- memset(&refcount_table[old_nb_clusters], 0, (nb_clusters
- - old_nb_clusters) * sizeof(uint16_t));
+ int old_nb_clusters = *nb_clusters;
+ *nb_clusters = (new_offset >> s->cluster_bits) + 1;
+
+ *refcount_table = g_try_realloc(*refcount_table,
+ *nb_clusters * sizeof(uint16_t));
+ if (!*refcount_table) {
+ res->check_errors++;
+ return -ENOMEM;
+ }
+
+ memset(&(*refcount_table)[old_nb_clusters], 0,
+ (*nb_clusters - old_nb_clusters) *
+ sizeof(uint16_t));
}
- refcount_table[cluster]--;
- inc_refcounts(bs, res, refcount_table, nb_clusters,
+ (*refcount_table)[cluster]--;
+ inc_refcounts(bs, res, *refcount_table, *nb_clusters,
new_offset, s->cluster_size);
res->corruptions_fixed++;
@@ -1621,6 +1574,95 @@ int qcow2_check_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
}
}
+ return 0;
+}
+
+/*
+ * Calculates an in-memory refcount table.
+ */
+static int calculate_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
+ BdrvCheckMode fix, uint16_t **refcount_table,
+ int64_t *nb_clusters)
+{
+ BDRVQcowState *s = bs->opaque;
+ QCowSnapshot *sn;
+ int64_t i;
+ int ret;
+
+ if (!*refcount_table) {
+ *refcount_table = g_try_malloc0(*nb_clusters * sizeof(uint16_t));
+ if (*nb_clusters && !*refcount_table) {
+ res->check_errors++;
+ return -ENOMEM;
+ }
+ }
+
+ /* header */
+ inc_refcounts(bs, res, *refcount_table, *nb_clusters,
+ 0, s->cluster_size);
+
+ /* current L1 table */
+ ret = check_refcounts_l1(bs, res, *refcount_table, *nb_clusters,
+ s->l1_table_offset, s->l1_size, CHECK_FRAG_INFO);
+ if (ret < 0) {
+ return ret;
+ }
+
+ /* snapshots */
+ for (i = 0; i < s->nb_snapshots; i++) {
+ sn = s->snapshots + i;
+ ret = check_refcounts_l1(bs, res, *refcount_table, *nb_clusters,
+ sn->l1_table_offset, sn->l1_size, 0);
+ if (ret < 0) {
+ return ret;
+ }
+ }
+ inc_refcounts(bs, res, *refcount_table, *nb_clusters,
+ s->snapshots_offset, s->snapshots_size);
+
+ /* refcount data */
+ inc_refcounts(bs, res, *refcount_table, *nb_clusters,
+ s->refcount_table_offset,
+ s->refcount_table_size * sizeof(uint64_t));
+
+ return check_refblocks(bs, res, fix, refcount_table, nb_clusters);
+}
+
+/*
+ * Checks an image for refcount consistency.
+ *
+ * Returns 0 if no errors are found, the number of errors in case the image is
+ * detected as corrupted, and -errno when an internal error occurred.
+ */
+int qcow2_check_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
+ BdrvCheckMode fix)
+{
+ BDRVQcowState *s = bs->opaque;
+ int64_t size, i, highest_cluster, nb_clusters;
+ int refcount1, refcount2;
+ uint16_t *refcount_table = NULL;
+ int ret;
+
+ size = bdrv_getlength(bs->file);
+ if (size < 0) {
+ res->check_errors++;
+ return size;
+ }
+
+ nb_clusters = size_to_clusters(s, size);
+ if (nb_clusters > INT_MAX) {
+ res->check_errors++;
+ return -EFBIG;
+ }
+
+ res->bfi.total_clusters =
+ size_to_clusters(s, bs->total_sectors * BDRV_SECTOR_SIZE);
+
+ ret = calculate_refcounts(bs, res, fix, &refcount_table, &nb_clusters);
+ if (ret < 0) {
+ goto fail;
+ }
+
/* compare ref counts */
for (i = 0, highest_cluster = 0; i < nb_clusters; i++) {
refcount1 = get_refcount(bs, i);
--
2.0.3
^ permalink raw reply related [flat|nested] 20+ messages in thread* Re: [Qemu-devel] [PATCH v2 2/9] qcow2: Factor out refcount accounting for check
2014-08-15 15:16 ` [Qemu-devel] [PATCH v2 2/9] qcow2: Factor out refcount accounting for check Max Reitz
@ 2014-08-21 18:47 ` Benoît Canet
2014-08-21 18:57 ` Benoît Canet
1 sibling, 0 replies; 20+ messages in thread
From: Benoît Canet @ 2014-08-21 18:47 UTC (permalink / raw)
To: Max Reitz; +Cc: Kevin Wolf, qemu-devel, Stefan Hajnoczi, Benoît Canet
On Fri, Aug 15, 2014 at 05:16:19PM +0200, Max Reitz wrote:
> Put the code for calculating the reference counts during qemu-img check
> into an own function.
>
> Also, do not use g_realloc() for increasing the size of the in-memory
> refcount table, but rather g_try_realloc().
>
> Signed-off-by: Max Reitz <mreitz@redhat.com>
> ---
> block/qcow2-refcount.c | 188 ++++++++++++++++++++++++++++++-------------------
> 1 file changed, 115 insertions(+), 73 deletions(-)
>
> diff --git a/block/qcow2-refcount.c b/block/qcow2-refcount.c
> index d60e2fe..b3c4edd 100644
> --- a/block/qcow2-refcount.c
> +++ b/block/qcow2-refcount.c
> @@ -1496,69 +1496,15 @@ done:
> }
>
> /*
> - * Checks an image for refcount consistency.
> - *
> - * Returns 0 if no errors are found, the number of errors in case the image is
> - * detected as corrupted, and -errno when an internal error occurred.
> + * Checks consistency of refblocks and accounts for each refblock in
> + * *refcount_table.
> */
> -int qcow2_check_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
> - BdrvCheckMode fix)
> +static int check_refblocks(BlockDriverState *bs, BdrvCheckResult *res,
> + BdrvCheckMode fix, uint16_t **refcount_table,
> + int64_t *nb_clusters)
> {
> BDRVQcowState *s = bs->opaque;
> - int64_t size, i, highest_cluster, nb_clusters;
> - int refcount1, refcount2;
> - QCowSnapshot *sn;
> - uint16_t *refcount_table;
> - int ret;
> -
> - size = bdrv_getlength(bs->file);
> - if (size < 0) {
> - res->check_errors++;
> - return size;
> - }
> -
> - nb_clusters = size_to_clusters(s, size);
> - if (nb_clusters > INT_MAX) {
> - res->check_errors++;
> - return -EFBIG;
> - }
> -
> - refcount_table = g_try_malloc0(nb_clusters * sizeof(uint16_t));
> - if (nb_clusters && refcount_table == NULL) {
> - res->check_errors++;
> - return -ENOMEM;
> - }
> -
> - res->bfi.total_clusters =
> - size_to_clusters(s, bs->total_sectors * BDRV_SECTOR_SIZE);
> -
> - /* header */
> - inc_refcounts(bs, res, refcount_table, nb_clusters,
> - 0, s->cluster_size);
> -
> - /* current L1 table */
> - ret = check_refcounts_l1(bs, res, refcount_table, nb_clusters,
> - s->l1_table_offset, s->l1_size, CHECK_FRAG_INFO);
> - if (ret < 0) {
> - goto fail;
> - }
> -
> - /* snapshots */
> - for(i = 0; i < s->nb_snapshots; i++) {
> - sn = s->snapshots + i;
> - ret = check_refcounts_l1(bs, res, refcount_table, nb_clusters,
> - sn->l1_table_offset, sn->l1_size, 0);
> - if (ret < 0) {
> - goto fail;
> - }
> - }
> - inc_refcounts(bs, res, refcount_table, nb_clusters,
> - s->snapshots_offset, s->snapshots_size);
> -
> - /* refcount data */
> - inc_refcounts(bs, res, refcount_table, nb_clusters,
> - s->refcount_table_offset,
> - s->refcount_table_size * sizeof(uint64_t));
> + int64_t i;
>
> for(i = 0; i < s->refcount_table_size; i++) {
> uint64_t offset, cluster;
> @@ -1573,7 +1519,7 @@ int qcow2_check_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
> continue;
> }
>
> - if (cluster >= nb_clusters) {
> + if (cluster >= *nb_clusters) {
> fprintf(stderr, "ERROR refcount block %" PRId64
> " is outside image\n", i);
> res->corruptions++;
> @@ -1581,14 +1527,14 @@ int qcow2_check_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
> }
>
> if (offset != 0) {
> - inc_refcounts(bs, res, refcount_table, nb_clusters,
> + inc_refcounts(bs, res, *refcount_table, *nb_clusters,
> offset, s->cluster_size);
> - if (refcount_table[cluster] != 1) {
> + if ((*refcount_table)[cluster] != 1) {
> fprintf(stderr, "%s refcount block %" PRId64
> " refcount=%d\n",
> fix & BDRV_FIX_ERRORS ? "Repairing" :
> "ERROR",
> - i, refcount_table[cluster]);
> + i, (*refcount_table)[cluster]);
>
> if (fix & BDRV_FIX_ERRORS) {
> int64_t new_offset;
> @@ -1600,17 +1546,24 @@ int qcow2_check_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
> }
>
> /* update refcounts */
> - if ((new_offset >> s->cluster_bits) >= nb_clusters) {
> + if ((new_offset >> s->cluster_bits) >= *nb_clusters) {
> /* increase refcount_table size if necessary */
> - int old_nb_clusters = nb_clusters;
> - nb_clusters = (new_offset >> s->cluster_bits) + 1;
> - refcount_table = g_realloc(refcount_table,
> - nb_clusters * sizeof(uint16_t));
> - memset(&refcount_table[old_nb_clusters], 0, (nb_clusters
> - - old_nb_clusters) * sizeof(uint16_t));
> + int old_nb_clusters = *nb_clusters;
> + *nb_clusters = (new_offset >> s->cluster_bits) + 1;
> +
> + *refcount_table = g_try_realloc(*refcount_table,
> + *nb_clusters * sizeof(uint16_t));
> + if (!*refcount_table) {
> + res->check_errors++;
> + return -ENOMEM;
> + }
> +
> + memset(&(*refcount_table)[old_nb_clusters], 0,
> + (*nb_clusters - old_nb_clusters) *
> + sizeof(uint16_t));
> }
> - refcount_table[cluster]--;
> - inc_refcounts(bs, res, refcount_table, nb_clusters,
> + (*refcount_table)[cluster]--;
> + inc_refcounts(bs, res, *refcount_table, *nb_clusters,
> new_offset, s->cluster_size);
>
> res->corruptions_fixed++;
> @@ -1621,6 +1574,95 @@ int qcow2_check_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
> }
> }
>
> + return 0;
> +}
> +
> +/*
> + * Calculates an in-memory refcount table.
> + */
> +static int calculate_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
> + BdrvCheckMode fix, uint16_t **refcount_table,
> + int64_t *nb_clusters)
> +{
> + BDRVQcowState *s = bs->opaque;
> + QCowSnapshot *sn;
> + int64_t i;
> + int ret;
> +
> + if (!*refcount_table) {
> + *refcount_table = g_try_malloc0(*nb_clusters * sizeof(uint16_t));
> + if (*nb_clusters && !*refcount_table) {
> + res->check_errors++;
> + return -ENOMEM;
> + }
> + }
> +
> + /* header */
> + inc_refcounts(bs, res, *refcount_table, *nb_clusters,
> + 0, s->cluster_size);
> +
> + /* current L1 table */
> + ret = check_refcounts_l1(bs, res, *refcount_table, *nb_clusters,
> + s->l1_table_offset, s->l1_size, CHECK_FRAG_INFO);
> + if (ret < 0) {
> + return ret;
> + }
> +
> + /* snapshots */
> + for (i = 0; i < s->nb_snapshots; i++) {
> + sn = s->snapshots + i;
> + ret = check_refcounts_l1(bs, res, *refcount_table, *nb_clusters,
> + sn->l1_table_offset, sn->l1_size, 0);
> + if (ret < 0) {
> + return ret;
> + }
> + }
> + inc_refcounts(bs, res, *refcount_table, *nb_clusters,
> + s->snapshots_offset, s->snapshots_size);
> +
> + /* refcount data */
> + inc_refcounts(bs, res, *refcount_table, *nb_clusters,
> + s->refcount_table_offset,
> + s->refcount_table_size * sizeof(uint64_t));
> +
> + return check_refblocks(bs, res, fix, refcount_table, nb_clusters);
> +}
> +
> +/*
> + * Checks an image for refcount consistency.
> + *
> + * Returns 0 if no errors are found, the number of errors in case the image is
> + * detected as corrupted, and -errno when an internal error occurred.
> + */
> +int qcow2_check_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
> + BdrvCheckMode fix)
> +{
> + BDRVQcowState *s = bs->opaque;
> + int64_t size, i, highest_cluster, nb_clusters;
> + int refcount1, refcount2;
> + uint16_t *refcount_table = NULL;
> + int ret;
> +
> + size = bdrv_getlength(bs->file);
> + if (size < 0) {
> + res->check_errors++;
> + return size;
> + }
> +
> + nb_clusters = size_to_clusters(s, size);
> + if (nb_clusters > INT_MAX) {
> + res->check_errors++;
> + return -EFBIG;
> + }
> +
> + res->bfi.total_clusters =
> + size_to_clusters(s, bs->total_sectors * BDRV_SECTOR_SIZE);
> +
> + ret = calculate_refcounts(bs, res, fix, &refcount_table, &nb_clusters);
> + if (ret < 0) {
> + goto fail;
> + }
> +
> /* compare ref counts */
> for (i = 0, highest_cluster = 0; i < nb_clusters; i++) {
> refcount1 = get_refcount(bs, i);
> --
> 2.0.3
>
I tried during half an hour to find a git option (--patience etc) or a tool
(Kompare) showing me that these code motions are right.
The problem is that you extract subfunctions and move them up at the same time.
I will try again tomorow to find something making sense of this patch without
having to spell check every single character.
Best regards
Benoît
^ permalink raw reply [flat|nested] 20+ messages in thread* Re: [Qemu-devel] [PATCH v2 2/9] qcow2: Factor out refcount accounting for check
2014-08-15 15:16 ` [Qemu-devel] [PATCH v2 2/9] qcow2: Factor out refcount accounting for check Max Reitz
2014-08-21 18:47 ` Benoît Canet
@ 2014-08-21 18:57 ` Benoît Canet
2014-08-21 19:13 ` Eric Blake
1 sibling, 1 reply; 20+ messages in thread
From: Benoît Canet @ 2014-08-21 18:57 UTC (permalink / raw)
To: Max Reitz; +Cc: Kevin Wolf, qemu-devel, Stefan Hajnoczi, Benoît Canet
On Fri, Aug 15, 2014 at 05:16:19PM +0200, Max Reitz wrote:
> Put the code for calculating the reference counts during qemu-img check
> into an own function.
>
> Also, do not use g_realloc() for increasing the size of the in-memory
> refcount table, but rather g_try_realloc().
>
> Signed-off-by: Max Reitz <mreitz@redhat.com>
> ---
> block/qcow2-refcount.c | 188 ++++++++++++++++++++++++++++++-------------------
> 1 file changed, 115 insertions(+), 73 deletions(-)
>
> diff --git a/block/qcow2-refcount.c b/block/qcow2-refcount.c
> index d60e2fe..b3c4edd 100644
> --- a/block/qcow2-refcount.c
> +++ b/block/qcow2-refcount.c
> @@ -1496,69 +1496,15 @@ done:
> }
>
> /*
> - * Checks an image for refcount consistency.
> - *
> - * Returns 0 if no errors are found, the number of errors in case the image is
> - * detected as corrupted, and -errno when an internal error occurred.
> + * Checks consistency of refblocks and accounts for each refblock in
> + * *refcount_table.
> */
> -int qcow2_check_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
> - BdrvCheckMode fix)
> +static int check_refblocks(BlockDriverState *bs, BdrvCheckResult *res,
> + BdrvCheckMode fix, uint16_t **refcount_table,
> + int64_t *nb_clusters)
> {
> BDRVQcowState *s = bs->opaque;
> - int64_t size, i, highest_cluster, nb_clusters;
> - int refcount1, refcount2;
> - QCowSnapshot *sn;
> - uint16_t *refcount_table;
> - int ret;
> -
> - size = bdrv_getlength(bs->file);
> - if (size < 0) {
> - res->check_errors++;
> - return size;
> - }
> -
> - nb_clusters = size_to_clusters(s, size);
> - if (nb_clusters > INT_MAX) {
> - res->check_errors++;
> - return -EFBIG;
> - }
> -
> - refcount_table = g_try_malloc0(nb_clusters * sizeof(uint16_t));
> - if (nb_clusters && refcount_table == NULL) {
> - res->check_errors++;
> - return -ENOMEM;
> - }
> -
> - res->bfi.total_clusters =
> - size_to_clusters(s, bs->total_sectors * BDRV_SECTOR_SIZE);
> -
> - /* header */
> - inc_refcounts(bs, res, refcount_table, nb_clusters,
> - 0, s->cluster_size);
> -
> - /* current L1 table */
> - ret = check_refcounts_l1(bs, res, refcount_table, nb_clusters,
> - s->l1_table_offset, s->l1_size, CHECK_FRAG_INFO);
> - if (ret < 0) {
> - goto fail;
> - }
> -
> - /* snapshots */
> - for(i = 0; i < s->nb_snapshots; i++) {
> - sn = s->snapshots + i;
> - ret = check_refcounts_l1(bs, res, refcount_table, nb_clusters,
> - sn->l1_table_offset, sn->l1_size, 0);
> - if (ret < 0) {
> - goto fail;
> - }
> - }
> - inc_refcounts(bs, res, refcount_table, nb_clusters,
> - s->snapshots_offset, s->snapshots_size);
> -
> - /* refcount data */
> - inc_refcounts(bs, res, refcount_table, nb_clusters,
> - s->refcount_table_offset,
> - s->refcount_table_size * sizeof(uint64_t));
> + int64_t i;
>
> for(i = 0; i < s->refcount_table_size; i++) {
> uint64_t offset, cluster;
> @@ -1573,7 +1519,7 @@ int qcow2_check_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
> continue;
> }
>
> - if (cluster >= nb_clusters) {
> + if (cluster >= *nb_clusters) {
> fprintf(stderr, "ERROR refcount block %" PRId64
> " is outside image\n", i);
> res->corruptions++;
> @@ -1581,14 +1527,14 @@ int qcow2_check_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
> }
>
> if (offset != 0) {
> - inc_refcounts(bs, res, refcount_table, nb_clusters,
> + inc_refcounts(bs, res, *refcount_table, *nb_clusters,
> offset, s->cluster_size);
> - if (refcount_table[cluster] != 1) {
> + if ((*refcount_table)[cluster] != 1) {
> fprintf(stderr, "%s refcount block %" PRId64
> " refcount=%d\n",
> fix & BDRV_FIX_ERRORS ? "Repairing" :
> "ERROR",
> - i, refcount_table[cluster]);
> + i, (*refcount_table)[cluster]);
>
> if (fix & BDRV_FIX_ERRORS) {
> int64_t new_offset;
> @@ -1600,17 +1546,24 @@ int qcow2_check_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
> }
>
> /* update refcounts */
> - if ((new_offset >> s->cluster_bits) >= nb_clusters) {
> + if ((new_offset >> s->cluster_bits) >= *nb_clusters) {
> /* increase refcount_table size if necessary */
> - int old_nb_clusters = nb_clusters;
> - nb_clusters = (new_offset >> s->cluster_bits) + 1;
> - refcount_table = g_realloc(refcount_table,
> - nb_clusters * sizeof(uint16_t));
> - memset(&refcount_table[old_nb_clusters], 0, (nb_clusters
> - - old_nb_clusters) * sizeof(uint16_t));
> + int old_nb_clusters = *nb_clusters;
> + *nb_clusters = (new_offset >> s->cluster_bits) + 1;
> +
> + *refcount_table = g_try_realloc(*refcount_table,
> + *nb_clusters * sizeof(uint16_t));
> + if (!*refcount_table) {
> + res->check_errors++;
> + return -ENOMEM;
> + }
> +
> + memset(&(*refcount_table)[old_nb_clusters], 0,
> + (*nb_clusters - old_nb_clusters) *
> + sizeof(uint16_t));
> }
> - refcount_table[cluster]--;
> - inc_refcounts(bs, res, refcount_table, nb_clusters,
> + (*refcount_table)[cluster]--;
> + inc_refcounts(bs, res, *refcount_table, *nb_clusters,
> new_offset, s->cluster_size);
>
> res->corruptions_fixed++;
> @@ -1621,6 +1574,95 @@ int qcow2_check_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
> }
> }
>
> + return 0;
> +}
> +
> +/*
> + * Calculates an in-memory refcount table.
> + */
> +static int calculate_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
> + BdrvCheckMode fix, uint16_t **refcount_table,
> + int64_t *nb_clusters)
> +{
> + BDRVQcowState *s = bs->opaque;
> + QCowSnapshot *sn;
> + int64_t i;
> + int ret;
> +
> + if (!*refcount_table) {
> + *refcount_table = g_try_malloc0(*nb_clusters * sizeof(uint16_t));
> + if (*nb_clusters && !*refcount_table) {
> + res->check_errors++;
> + return -ENOMEM;
> + }
> + }
> +
> + /* header */
> + inc_refcounts(bs, res, *refcount_table, *nb_clusters,
> + 0, s->cluster_size);
> +
> + /* current L1 table */
> + ret = check_refcounts_l1(bs, res, *refcount_table, *nb_clusters,
> + s->l1_table_offset, s->l1_size, CHECK_FRAG_INFO);
> + if (ret < 0) {
> + return ret;
> + }
> +
> + /* snapshots */
> + for (i = 0; i < s->nb_snapshots; i++) {
> + sn = s->snapshots + i;
> + ret = check_refcounts_l1(bs, res, *refcount_table, *nb_clusters,
> + sn->l1_table_offset, sn->l1_size, 0);
> + if (ret < 0) {
> + return ret;
> + }
> + }
> + inc_refcounts(bs, res, *refcount_table, *nb_clusters,
> + s->snapshots_offset, s->snapshots_size);
> +
> + /* refcount data */
> + inc_refcounts(bs, res, *refcount_table, *nb_clusters,
> + s->refcount_table_offset,
> + s->refcount_table_size * sizeof(uint64_t));
> +
> + return check_refblocks(bs, res, fix, refcount_table, nb_clusters);
> +}
> +
> +/*
> + * Checks an image for refcount consistency.
> + *
> + * Returns 0 if no errors are found, the number of errors in case the image is
> + * detected as corrupted, and -errno when an internal error occurred.
> + */
> +int qcow2_check_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
> + BdrvCheckMode fix)
> +{
> + BDRVQcowState *s = bs->opaque;
> + int64_t size, i, highest_cluster, nb_clusters;
> + int refcount1, refcount2;
> + uint16_t *refcount_table = NULL;
> + int ret;
> +
> + size = bdrv_getlength(bs->file);
> + if (size < 0) {
> + res->check_errors++;
> + return size;
> + }
> +
> + nb_clusters = size_to_clusters(s, size);
> + if (nb_clusters > INT_MAX) {
> + res->check_errors++;
> + return -EFBIG;
> + }
> +
> + res->bfi.total_clusters =
> + size_to_clusters(s, bs->total_sectors * BDRV_SECTOR_SIZE);
> +
> + ret = calculate_refcounts(bs, res, fix, &refcount_table, &nb_clusters);
> + if (ret < 0) {
> + goto fail;
> + }
> +
> /* compare ref counts */
> for (i = 0, highest_cluster = 0; i < nb_clusters; i++) {
> refcount1 = get_refcount(bs, i);
> --
> 2.0.3
>
Another point is that I think these two extractions patches will totaly confuse
git blame the day we will need it.
Best regards
Benoît
^ permalink raw reply [flat|nested] 20+ messages in thread* Re: [Qemu-devel] [PATCH v2 2/9] qcow2: Factor out refcount accounting for check
2014-08-21 18:57 ` Benoît Canet
@ 2014-08-21 19:13 ` Eric Blake
2014-08-21 21:16 ` Benoît Canet
0 siblings, 1 reply; 20+ messages in thread
From: Eric Blake @ 2014-08-21 19:13 UTC (permalink / raw)
To: Benoît Canet, Max Reitz; +Cc: Kevin Wolf, qemu-devel, Stefan Hajnoczi
[-- Attachment #1: Type: text/plain, Size: 1451 bytes --]
On 08/21/2014 12:57 PM, Benoît Canet wrote:
> On Fri, Aug 15, 2014 at 05:16:19PM +0200, Max Reitz wrote:
>> Put the code for calculating the reference counts during qemu-img check
>> into an own function.
>>
>> Also, do not use g_realloc() for increasing the size of the in-memory
>> refcount table, but rather g_try_realloc().
The "Also," may be a sign of doing two things in one patch that could
have been two.
>>
>> Signed-off-by: Max Reitz <mreitz@redhat.com>
>> ---
>> block/qcow2-refcount.c | 188 ++++++++++++++++++++++++++++++-------------------
>> 1 file changed, 115 insertions(+), 73 deletions(-)
But overall, this patch seems like a reasonable refactor to me,
splitting out a reusable piece.
> Another point is that I think these two extractions patches will totaly confuse
> git blame the day we will need it.
I disagree. When refactoring to split a large function into multiple
smaller functions, where the original function calls out to a helper
function for what it used to do inline, git blame tracks things
perfectly. Someone following blame may end up on this commit as the
source of a given line in its new location, but the blame viewer will
also show this commit is a refactor, and it should be fairly easy to
find where the line was moved from, and resume blaming even further.
--
Eric Blake eblake redhat com +1-919-301-3266
Libvirt virtualization library http://libvirt.org
[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 539 bytes --]
^ permalink raw reply [flat|nested] 20+ messages in thread
* Re: [Qemu-devel] [PATCH v2 2/9] qcow2: Factor out refcount accounting for check
2014-08-21 19:13 ` Eric Blake
@ 2014-08-21 21:16 ` Benoît Canet
2014-08-22 15:26 ` Max Reitz
0 siblings, 1 reply; 20+ messages in thread
From: Benoît Canet @ 2014-08-21 21:16 UTC (permalink / raw)
To: Eric Blake
Cc: Kevin Wolf, Stefan Hajnoczi, qemu-devel, Benoît Canet,
Max Reitz
On Thu, Aug 21, 2014 at 01:13:20PM -0600, Eric Blake wrote:
> On 08/21/2014 12:57 PM, Benoît Canet wrote:
> > On Fri, Aug 15, 2014 at 05:16:19PM +0200, Max Reitz wrote:
> >> Put the code for calculating the reference counts during qemu-img check
> >> into an own function.
> >>
> >> Also, do not use g_realloc() for increasing the size of the in-memory
> >> refcount table, but rather g_try_realloc().
>
> The "Also," may be a sign of doing two things in one patch that could
> have been two.
>
> >>
> >> Signed-off-by: Max Reitz <mreitz@redhat.com>
> >> ---
> >> block/qcow2-refcount.c | 188 ++++++++++++++++++++++++++++++-------------------
> >> 1 file changed, 115 insertions(+), 73 deletions(-)
>
> But overall, this patch seems like a reasonable refactor to me,
> splitting out a reusable piece.
>
> > Another point is that I think these two extractions patches will totaly confuse
> > git blame the day we will need it.
>
> I disagree. When refactoring to split a large function into multiple
> smaller functions, where the original function calls out to a helper
> function for what it used to do inline, git blame tracks things
> perfectly. Someone following blame may end up on this commit as the
> source of a given line in its new location, but the blame viewer will
> also show this commit is a refactor, and it should be fairly easy to
> find where the line was moved from, and resume blaming even further.
I think we could separate the extractions and the respectives moves into their own
patches.
This way the extraction patch would be cleaner (no code disapearing and appearing
elsewere with another author) and the operations could be reviewed more easily
with the various code review tools.
>
> --
> Eric Blake eblake redhat com +1-919-301-3266
> Libvirt virtualization library http://libvirt.org
>
^ permalink raw reply [flat|nested] 20+ messages in thread
* Re: [Qemu-devel] [PATCH v2 2/9] qcow2: Factor out refcount accounting for check
2014-08-21 21:16 ` Benoît Canet
@ 2014-08-22 15:26 ` Max Reitz
2014-08-22 15:37 ` Benoît Canet
0 siblings, 1 reply; 20+ messages in thread
From: Max Reitz @ 2014-08-22 15:26 UTC (permalink / raw)
To: Benoît Canet, Eric Blake; +Cc: Kevin Wolf, qemu-devel, Stefan Hajnoczi
On 21.08.2014 23:16, Benoît Canet wrote:
> On Thu, Aug 21, 2014 at 01:13:20PM -0600, Eric Blake wrote:
>> On 08/21/2014 12:57 PM, Benoît Canet wrote:
>>> On Fri, Aug 15, 2014 at 05:16:19PM +0200, Max Reitz wrote:
>>>> Put the code for calculating the reference counts during qemu-img check
>>>> into an own function.
>>>>
>>>> Also, do not use g_realloc() for increasing the size of the in-memory
>>>> refcount table, but rather g_try_realloc().
>> The "Also," may be a sign of doing two things in one patch that could
>> have been two.
The second change fit really well into this patch, so I left it in. On
the other hand, I just noticed I don't need to do that at all, because
that call will be dropped in patch 5 anyway.
>>>> Signed-off-by: Max Reitz <mreitz@redhat.com>
>>>> ---
>>>> block/qcow2-refcount.c | 188 ++++++++++++++++++++++++++++++-------------------
>>>> 1 file changed, 115 insertions(+), 73 deletions(-)
>> But overall, this patch seems like a reasonable refactor to me,
>> splitting out a reusable piece.
>>
>>> Another point is that I think these two extractions patches will totaly confuse
>>> git blame the day we will need it.
>> I disagree. When refactoring to split a large function into multiple
>> smaller functions, where the original function calls out to a helper
>> function for what it used to do inline, git blame tracks things
>> perfectly. Someone following blame may end up on this commit as the
>> source of a given line in its new location, but the blame viewer will
>> also show this commit is a refactor, and it should be fairly easy to
>> find where the line was moved from, and resume blaming even further.
> I think we could separate the extractions and the respectives moves into their own
> patches.
> This way the extraction patch would be cleaner (no code disapearing and appearing
> elsewere with another author) and the operations could be reviewed more easily
> with the various code review tools.
The main cause for the diff cluttering is making nb_clusters and
refcount_table pointers in the function. Unfortunately (or maybe
fortunately?) this is not C++, so I cannot make them references, but
they need to be passed by reference. The function absolutely must be
able to change their values and I don't see any way of avoiding this
even in the first extraction patch.
I'll do my best on the rest, but the diff not being trivial is mainly
simply due to the movement not being absolutely trivial; all accesses to
nb_clusters and refcount_table have to be modified.
Max
^ permalink raw reply [flat|nested] 20+ messages in thread
* Re: [Qemu-devel] [PATCH v2 2/9] qcow2: Factor out refcount accounting for check
2014-08-22 15:26 ` Max Reitz
@ 2014-08-22 15:37 ` Benoît Canet
2014-08-22 15:44 ` Max Reitz
0 siblings, 1 reply; 20+ messages in thread
From: Benoît Canet @ 2014-08-22 15:37 UTC (permalink / raw)
To: Max Reitz; +Cc: Kevin Wolf, Stefan Hajnoczi, qemu-devel, Benoît Canet
On Fri, Aug 22, 2014 at 05:26:45PM +0200, Max Reitz wrote:
> On 21.08.2014 23:16, Benoît Canet wrote:
> >On Thu, Aug 21, 2014 at 01:13:20PM -0600, Eric Blake wrote:
> >>On 08/21/2014 12:57 PM, Benoît Canet wrote:
> >>>On Fri, Aug 15, 2014 at 05:16:19PM +0200, Max Reitz wrote:
> >>>>Put the code for calculating the reference counts during qemu-img check
> >>>>into an own function.
> >>>>
> >>>>Also, do not use g_realloc() for increasing the size of the in-memory
> >>>>refcount table, but rather g_try_realloc().
> >>The "Also," may be a sign of doing two things in one patch that could
> >>have been two.
>
> The second change fit really well into this patch, so I left it in.
> On the other hand, I just noticed I don't need to do that at all,
> because that call will be dropped in patch 5 anyway.
>
> >>>>Signed-off-by: Max Reitz <mreitz@redhat.com>
> >>>>---
> >>>> block/qcow2-refcount.c | 188 ++++++++++++++++++++++++++++++-------------------
> >>>> 1 file changed, 115 insertions(+), 73 deletions(-)
> >>But overall, this patch seems like a reasonable refactor to me,
> >>splitting out a reusable piece.
> >>
> >>>Another point is that I think these two extractions patches will totaly confuse
> >>>git blame the day we will need it.
> >>I disagree. When refactoring to split a large function into multiple
> >>smaller functions, where the original function calls out to a helper
> >>function for what it used to do inline, git blame tracks things
> >>perfectly. Someone following blame may end up on this commit as the
> >>source of a given line in its new location, but the blame viewer will
> >>also show this commit is a refactor, and it should be fairly easy to
> >>find where the line was moved from, and resume blaming even further.
> >I think we could separate the extractions and the respectives moves into their own
> >patches.
> >This way the extraction patch would be cleaner (no code disapearing and appearing
> >elsewere with another author) and the operations could be reviewed more easily
> >with the various code review tools.
>
> The main cause for the diff cluttering is making nb_clusters and
> refcount_table pointers in the function. Unfortunately (or maybe
> fortunately?) this is not C++, so I cannot make them references, but
> they need to be passed by reference. The function absolutely must be
> able to change their values and I don't see any way of avoiding this
> even in the first extraction patch.
Do a: git show f75c744 | kompare -
You will see a whole block of code being morphed to int64_t i and the same whole
block reapear below.
I think this is a consequence of the extracted functions moving up at the same time
they are extracted.
Best regards
Benoît
>
> I'll do my best on the rest, but the diff not being trivial is
> mainly simply due to the movement not being absolutely trivial; all
> accesses to nb_clusters and refcount_table have to be modified.
>
> Max
^ permalink raw reply [flat|nested] 20+ messages in thread
* Re: [Qemu-devel] [PATCH v2 2/9] qcow2: Factor out refcount accounting for check
2014-08-22 15:37 ` Benoît Canet
@ 2014-08-22 15:44 ` Max Reitz
2014-08-22 15:46 ` Max Reitz
2014-08-22 15:48 ` Eric Blake
0 siblings, 2 replies; 20+ messages in thread
From: Max Reitz @ 2014-08-22 15:44 UTC (permalink / raw)
To: Benoît Canet; +Cc: Kevin Wolf, qemu-devel, Stefan Hajnoczi
On 22.08.2014 17:37, Benoît Canet wrote:
> On Fri, Aug 22, 2014 at 05:26:45PM +0200, Max Reitz wrote:
>> On 21.08.2014 23:16, Benoît Canet wrote:
>>> On Thu, Aug 21, 2014 at 01:13:20PM -0600, Eric Blake wrote:
>>>> On 08/21/2014 12:57 PM, Benoît Canet wrote:
>>>>> On Fri, Aug 15, 2014 at 05:16:19PM +0200, Max Reitz wrote:
>>>>>> Put the code for calculating the reference counts during qemu-img check
>>>>>> into an own function.
>>>>>>
>>>>>> Also, do not use g_realloc() for increasing the size of the in-memory
>>>>>> refcount table, but rather g_try_realloc().
>>>> The "Also," may be a sign of doing two things in one patch that could
>>>> have been two.
>> The second change fit really well into this patch, so I left it in.
>> On the other hand, I just noticed I don't need to do that at all,
>> because that call will be dropped in patch 5 anyway.
>>
>>>>>> Signed-off-by: Max Reitz <mreitz@redhat.com>
>>>>>> ---
>>>>>> block/qcow2-refcount.c | 188 ++++++++++++++++++++++++++++++-------------------
>>>>>> 1 file changed, 115 insertions(+), 73 deletions(-)
>>>> But overall, this patch seems like a reasonable refactor to me,
>>>> splitting out a reusable piece.
>>>>
>>>>> Another point is that I think these two extractions patches will totaly confuse
>>>>> git blame the day we will need it.
>>>> I disagree. When refactoring to split a large function into multiple
>>>> smaller functions, where the original function calls out to a helper
>>>> function for what it used to do inline, git blame tracks things
>>>> perfectly. Someone following blame may end up on this commit as the
>>>> source of a given line in its new location, but the blame viewer will
>>>> also show this commit is a refactor, and it should be fairly easy to
>>>> find where the line was moved from, and resume blaming even further.
>>> I think we could separate the extractions and the respectives moves into their own
>>> patches.
>>> This way the extraction patch would be cleaner (no code disapearing and appearing
>>> elsewere with another author) and the operations could be reviewed more easily
>>> with the various code review tools.
>> The main cause for the diff cluttering is making nb_clusters and
>> refcount_table pointers in the function. Unfortunately (or maybe
>> fortunately?) this is not C++, so I cannot make them references, but
>> they need to be passed by reference. The function absolutely must be
>> able to change their values and I don't see any way of avoiding this
>> even in the first extraction patch.
> Do a: git show f75c744 | kompare -
>
> You will see a whole block of code being morphed to int64_t i and the same whole
> block reapear below.
> I think this is a consequence of the extracted functions moving up at the same time
> they are extracted.
This is not a consequence of the extracted functions moving up, but
because of check_refblocks() preceding calculate_refcounts(). Splitting
the check won't help; and I know that Eric dislikes
(forward-)declarations of static functions (which I have to do if I pull
calculate_refcounts() before check_refblocks()).
Max
^ permalink raw reply [flat|nested] 20+ messages in thread
* Re: [Qemu-devel] [PATCH v2 2/9] qcow2: Factor out refcount accounting for check
2014-08-22 15:44 ` Max Reitz
@ 2014-08-22 15:46 ` Max Reitz
2014-08-22 15:48 ` Eric Blake
1 sibling, 0 replies; 20+ messages in thread
From: Max Reitz @ 2014-08-22 15:46 UTC (permalink / raw)
To: Benoît Canet; +Cc: Kevin Wolf, qemu-devel, Stefan Hajnoczi
On 22.08.2014 17:44, Max Reitz wrote:
> [snip]
>
> […] Splitting the check won't help; […]
Sorry, I meant “patch” instead of “check”.
Max
^ permalink raw reply [flat|nested] 20+ messages in thread
* Re: [Qemu-devel] [PATCH v2 2/9] qcow2: Factor out refcount accounting for check
2014-08-22 15:44 ` Max Reitz
2014-08-22 15:46 ` Max Reitz
@ 2014-08-22 15:48 ` Eric Blake
1 sibling, 0 replies; 20+ messages in thread
From: Eric Blake @ 2014-08-22 15:48 UTC (permalink / raw)
To: Max Reitz, Benoît Canet; +Cc: Kevin Wolf, qemu-devel, Stefan Hajnoczi
[-- Attachment #1: Type: text/plain, Size: 859 bytes --]
On 08/22/2014 09:44 AM, Max Reitz wrote:
> This is not a consequence of the extracted functions moving up, but
> because of check_refblocks() preceding calculate_refcounts(). Splitting
> the check won't help; and I know that Eric dislikes
> (forward-)declarations of static functions (which I have to do if I pull
> calculate_refcounts() before check_refblocks()).
I may not like them in the end result, but they are perfectly fine for
an intermediate patch that is refactoring things, when coupled with a
followup patch that moves the refactored code into topological order.
Anything we can do to make reviewing easier, even if it requires going
through more commits and more lines of churn to get there, pays off in
the long run.
--
Eric Blake eblake redhat com +1-919-301-3266
Libvirt virtualization library http://libvirt.org
[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 539 bytes --]
^ permalink raw reply [flat|nested] 20+ messages in thread
* [Qemu-devel] [PATCH v2 3/9] qcow2: Factor out refcount comparison for check
2014-08-15 15:16 [Qemu-devel] [PATCH v2 0/9] qcow2: Fix image repairing Max Reitz
2014-08-15 15:16 ` [Qemu-devel] [PATCH v2 1/9] qcow2: Fix leaks in dirty images Max Reitz
2014-08-15 15:16 ` [Qemu-devel] [PATCH v2 2/9] qcow2: Factor out refcount accounting for check Max Reitz
@ 2014-08-15 15:16 ` Max Reitz
2014-08-15 15:16 ` [Qemu-devel] [PATCH v2 4/9] qcow2: Fix refcount blocks beyond image end Max Reitz
` (5 subsequent siblings)
8 siblings, 0 replies; 20+ messages in thread
From: Max Reitz @ 2014-08-15 15:16 UTC (permalink / raw)
To: qemu-devel; +Cc: Kevin Wolf, Benoît Canet, Stefan Hajnoczi, Max Reitz
Put the code for comparing the calculated refcounts against the on-disk
refcounts during qemu-img check into an own function.
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
block/qcow2-refcount.c | 81 +++++++++++++++++++++++++++++---------------------
1 file changed, 47 insertions(+), 34 deletions(-)
diff --git a/block/qcow2-refcount.c b/block/qcow2-refcount.c
index b3c4edd..c477f04 100644
--- a/block/qcow2-refcount.c
+++ b/block/qcow2-refcount.c
@@ -1629,42 +1629,18 @@ static int calculate_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
}
/*
- * Checks an image for refcount consistency.
- *
- * Returns 0 if no errors are found, the number of errors in case the image is
- * detected as corrupted, and -errno when an internal error occurred.
+ * Compares the actual reference count for each cluster in the image against the
+ * refcount as reported by the refcount structures on-disk.
*/
-int qcow2_check_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
- BdrvCheckMode fix)
+static void compare_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
+ BdrvCheckMode fix, int64_t *highest_cluster,
+ uint16_t *refcount_table, int64_t nb_clusters)
{
BDRVQcowState *s = bs->opaque;
- int64_t size, i, highest_cluster, nb_clusters;
- int refcount1, refcount2;
- uint16_t *refcount_table = NULL;
- int ret;
-
- size = bdrv_getlength(bs->file);
- if (size < 0) {
- res->check_errors++;
- return size;
- }
-
- nb_clusters = size_to_clusters(s, size);
- if (nb_clusters > INT_MAX) {
- res->check_errors++;
- return -EFBIG;
- }
-
- res->bfi.total_clusters =
- size_to_clusters(s, bs->total_sectors * BDRV_SECTOR_SIZE);
-
- ret = calculate_refcounts(bs, res, fix, &refcount_table, &nb_clusters);
- if (ret < 0) {
- goto fail;
- }
+ int64_t i;
+ int refcount1, refcount2, ret;
- /* compare ref counts */
- for (i = 0, highest_cluster = 0; i < nb_clusters; i++) {
+ for (i = 0, *highest_cluster = 0; i < nb_clusters; i++) {
refcount1 = get_refcount(bs, i);
if (refcount1 < 0) {
fprintf(stderr, "Can't get refcount for cluster %" PRId64 ": %s\n",
@@ -1676,11 +1652,10 @@ int qcow2_check_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
refcount2 = refcount_table[i];
if (refcount1 > 0 || refcount2 > 0) {
- highest_cluster = i;
+ *highest_cluster = i;
}
if (refcount1 != refcount2) {
-
/* Check if we're allowed to fix the mismatch */
int *num_fixed = NULL;
if (refcount1 > refcount2 && (fix & BDRV_FIX_LEAKS)) {
@@ -1713,6 +1688,44 @@ int qcow2_check_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
}
}
}
+}
+
+/*
+ * Checks an image for refcount consistency.
+ *
+ * Returns 0 if no errors are found, the number of errors in case the image is
+ * detected as corrupted, and -errno when an internal error occurred.
+ */
+int qcow2_check_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
+ BdrvCheckMode fix)
+{
+ BDRVQcowState *s = bs->opaque;
+ int64_t size, highest_cluster, nb_clusters;
+ uint16_t *refcount_table = NULL;
+ int ret;
+
+ size = bdrv_getlength(bs->file);
+ if (size < 0) {
+ res->check_errors++;
+ return size;
+ }
+
+ nb_clusters = size_to_clusters(s, size);
+ if (nb_clusters > INT_MAX) {
+ res->check_errors++;
+ return -EFBIG;
+ }
+
+ res->bfi.total_clusters =
+ size_to_clusters(s, bs->total_sectors * BDRV_SECTOR_SIZE);
+
+ ret = calculate_refcounts(bs, res, fix, &refcount_table, &nb_clusters);
+ if (ret < 0) {
+ goto fail;
+ }
+
+ compare_refcounts(bs, res, fix, &highest_cluster, refcount_table,
+ nb_clusters);
/* check OFLAG_COPIED */
ret = check_oflag_copied(bs, res, fix);
--
2.0.3
^ permalink raw reply related [flat|nested] 20+ messages in thread* [Qemu-devel] [PATCH v2 4/9] qcow2: Fix refcount blocks beyond image end
2014-08-15 15:16 [Qemu-devel] [PATCH v2 0/9] qcow2: Fix image repairing Max Reitz
` (2 preceding siblings ...)
2014-08-15 15:16 ` [Qemu-devel] [PATCH v2 3/9] qcow2: Factor out refcount comparison " Max Reitz
@ 2014-08-15 15:16 ` Max Reitz
2014-08-15 15:16 ` [Qemu-devel] [PATCH v2 5/9] qcow2: Do not perform potentially damaging repairs Max Reitz
` (4 subsequent siblings)
8 siblings, 0 replies; 20+ messages in thread
From: Max Reitz @ 2014-08-15 15:16 UTC (permalink / raw)
To: qemu-devel; +Cc: Kevin Wolf, Benoît Canet, Stefan Hajnoczi, Max Reitz
If the qcow2 check function detects a refcount block located beyond the
image end, grow the image appropriately. This cannot break anything and
is the logical fix for such a case.
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
block/qcow2-refcount.c | 50 ++++++++++++++++++++++++++++++++++++++++++++++----
1 file changed, 46 insertions(+), 4 deletions(-)
diff --git a/block/qcow2-refcount.c b/block/qcow2-refcount.c
index c477f04..1783720 100644
--- a/block/qcow2-refcount.c
+++ b/block/qcow2-refcount.c
@@ -1504,7 +1504,8 @@ static int check_refblocks(BlockDriverState *bs, BdrvCheckResult *res,
int64_t *nb_clusters)
{
BDRVQcowState *s = bs->opaque;
- int64_t i;
+ int64_t i, size;
+ int ret;
for(i = 0; i < s->refcount_table_size; i++) {
uint64_t offset, cluster;
@@ -1520,9 +1521,50 @@ static int check_refblocks(BlockDriverState *bs, BdrvCheckResult *res,
}
if (cluster >= *nb_clusters) {
- fprintf(stderr, "ERROR refcount block %" PRId64
- " is outside image\n", i);
- res->corruptions++;
+ fprintf(stderr, "%s refcount block %" PRId64 " is outside image\n",
+ fix & BDRV_FIX_ERRORS ? "Repairing" : "ERROR", i);
+
+ if (fix & BDRV_FIX_ERRORS) {
+ int64_t old_nb_clusters = *nb_clusters;
+
+ ret = bdrv_truncate(bs->file, offset + s->cluster_size);
+ if (ret < 0) {
+ goto resize_fail;
+ }
+ size = bdrv_getlength(bs->file);
+ if (size < 0) {
+ ret = size;
+ goto resize_fail;
+ }
+
+ *nb_clusters = size_to_clusters(s, size);
+ assert(*nb_clusters >= old_nb_clusters);
+
+ *refcount_table = g_try_realloc(*refcount_table,
+ *nb_clusters * sizeof(uint16_t));
+ if (!*refcount_table) {
+ res->check_errors++;
+ return -ENOMEM;
+ }
+
+ memset(*refcount_table + old_nb_clusters, 0,
+ (*nb_clusters - old_nb_clusters) * sizeof(uint16_t));
+
+ if (cluster >= *nb_clusters) {
+ ret = -EINVAL;
+ goto resize_fail;
+ }
+
+ res->corruptions_fixed++;
+ continue;
+
+resize_fail:
+ res->corruptions++;
+ fprintf(stderr, "ERROR could not resize image: %s\n",
+ strerror(-ret));
+ } else {
+ res->corruptions++;
+ }
continue;
}
--
2.0.3
^ permalink raw reply related [flat|nested] 20+ messages in thread* [Qemu-devel] [PATCH v2 5/9] qcow2: Do not perform potentially damaging repairs
2014-08-15 15:16 [Qemu-devel] [PATCH v2 0/9] qcow2: Fix image repairing Max Reitz
` (3 preceding siblings ...)
2014-08-15 15:16 ` [Qemu-devel] [PATCH v2 4/9] qcow2: Fix refcount blocks beyond image end Max Reitz
@ 2014-08-15 15:16 ` Max Reitz
2014-08-15 15:16 ` [Qemu-devel] [PATCH v2 6/9] qcow2: Rebuild refcount structure during check Max Reitz
` (3 subsequent siblings)
8 siblings, 0 replies; 20+ messages in thread
From: Max Reitz @ 2014-08-15 15:16 UTC (permalink / raw)
To: qemu-devel; +Cc: Kevin Wolf, Benoît Canet, Stefan Hajnoczi, Max Reitz
If a referenced cluster has a refcount of 0, increasing its refcount may
result in clusters being allocated for the refcount structures. This may
overwrite the referenced cluster, therefore we cannot simply increase
the refcount then.
In such cases, we can either try to replicate all the refcount
operations solely for the check operation, basing the allocations on the
in-memory refcount table; or we can simply rebuild the whole refcount
structure based on the in-memory refcount table. Since the latter will
be much easier, do that.
To prepare for this, introduce a "rebuild" boolean which should be set
to true whenever a fix is rather dangerous or too complicated using the
current refcount structures. Another example for this is refcount blocks
being referenced more than once.
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
block/qcow2-refcount.c | 192 ++++++++-----------------------------------------
1 file changed, 28 insertions(+), 164 deletions(-)
diff --git a/block/qcow2-refcount.c b/block/qcow2-refcount.c
index 1783720..adfca45 100644
--- a/block/qcow2-refcount.c
+++ b/block/qcow2-refcount.c
@@ -1381,127 +1381,12 @@ fail:
}
/*
- * Writes one sector of the refcount table to the disk
- */
-#define RT_ENTRIES_PER_SECTOR (512 / sizeof(uint64_t))
-static int write_reftable_entry(BlockDriverState *bs, int rt_index)
-{
- BDRVQcowState *s = bs->opaque;
- uint64_t buf[RT_ENTRIES_PER_SECTOR];
- int rt_start_index;
- int i, ret;
-
- rt_start_index = rt_index & ~(RT_ENTRIES_PER_SECTOR - 1);
- for (i = 0; i < RT_ENTRIES_PER_SECTOR; i++) {
- buf[i] = cpu_to_be64(s->refcount_table[rt_start_index + i]);
- }
-
- ret = qcow2_pre_write_overlap_check(bs, QCOW2_OL_REFCOUNT_TABLE,
- s->refcount_table_offset + rt_start_index * sizeof(uint64_t),
- sizeof(buf));
- if (ret < 0) {
- return ret;
- }
-
- BLKDBG_EVENT(bs->file, BLKDBG_REFTABLE_UPDATE);
- ret = bdrv_pwrite_sync(bs->file, s->refcount_table_offset +
- rt_start_index * sizeof(uint64_t), buf, sizeof(buf));
- if (ret < 0) {
- return ret;
- }
-
- return 0;
-}
-
-/*
- * Allocates a new cluster for the given refcount block (represented by its
- * offset in the image file) and copies the current content there. This function
- * does _not_ decrement the reference count for the currently occupied cluster.
- *
- * This function prints an informative message to stderr on error (and returns
- * -errno); on success, the offset of the newly allocated cluster is returned.
- */
-static int64_t realloc_refcount_block(BlockDriverState *bs, int reftable_index,
- uint64_t offset)
-{
- BDRVQcowState *s = bs->opaque;
- int64_t new_offset = 0;
- void *refcount_block = NULL;
- int ret;
-
- /* allocate new refcount block */
- new_offset = qcow2_alloc_clusters(bs, s->cluster_size);
- if (new_offset < 0) {
- fprintf(stderr, "Could not allocate new cluster: %s\n",
- strerror(-new_offset));
- ret = new_offset;
- goto done;
- }
-
- /* fetch current refcount block content */
- ret = qcow2_cache_get(bs, s->refcount_block_cache, offset, &refcount_block);
- if (ret < 0) {
- fprintf(stderr, "Could not fetch refcount block: %s\n", strerror(-ret));
- goto fail_free_cluster;
- }
-
- /* new block has not yet been entered into refcount table, therefore it is
- * no refcount block yet (regarding this check) */
- ret = qcow2_pre_write_overlap_check(bs, 0, new_offset, s->cluster_size);
- if (ret < 0) {
- fprintf(stderr, "Could not write refcount block; metadata overlap "
- "check failed: %s\n", strerror(-ret));
- /* the image will be marked corrupt, so don't even attempt on freeing
- * the cluster */
- goto done;
- }
-
- /* write to new block */
- ret = bdrv_write(bs->file, new_offset / BDRV_SECTOR_SIZE, refcount_block,
- s->cluster_sectors);
- if (ret < 0) {
- fprintf(stderr, "Could not write refcount block: %s\n", strerror(-ret));
- goto fail_free_cluster;
- }
-
- /* update refcount table */
- assert(!offset_into_cluster(s, new_offset));
- s->refcount_table[reftable_index] = new_offset;
- ret = write_reftable_entry(bs, reftable_index);
- if (ret < 0) {
- fprintf(stderr, "Could not update refcount table: %s\n",
- strerror(-ret));
- goto fail_free_cluster;
- }
-
- goto done;
-
-fail_free_cluster:
- qcow2_free_clusters(bs, new_offset, s->cluster_size, QCOW2_DISCARD_OTHER);
-
-done:
- if (refcount_block) {
- /* This should never fail, as it would only do so if the given refcount
- * block cannot be found in the cache. As this is impossible as long as
- * there are no bugs, assert the success. */
- int tmp = qcow2_cache_put(bs, s->refcount_block_cache, &refcount_block);
- assert(tmp == 0);
- }
-
- if (ret < 0) {
- return ret;
- }
-
- return new_offset;
-}
-
-/*
* Checks consistency of refblocks and accounts for each refblock in
* *refcount_table.
*/
static int check_refblocks(BlockDriverState *bs, BdrvCheckResult *res,
- BdrvCheckMode fix, uint16_t **refcount_table,
- int64_t *nb_clusters)
+ BdrvCheckMode fix, bool *rebuild,
+ uint16_t **refcount_table, int64_t *nb_clusters)
{
BDRVQcowState *s = bs->opaque;
int64_t i, size;
@@ -1517,6 +1402,7 @@ static int check_refblocks(BlockDriverState *bs, BdrvCheckResult *res,
fprintf(stderr, "ERROR refcount block %" PRId64 " is not "
"cluster aligned; refcount table entry corrupted\n", i);
res->corruptions++;
+ *rebuild = true;
continue;
}
@@ -1560,6 +1446,7 @@ static int check_refblocks(BlockDriverState *bs, BdrvCheckResult *res,
resize_fail:
res->corruptions++;
+ *rebuild = true;
fprintf(stderr, "ERROR could not resize image: %s\n",
strerror(-ret));
} else {
@@ -1571,47 +1458,12 @@ resize_fail:
if (offset != 0) {
inc_refcounts(bs, res, *refcount_table, *nb_clusters,
offset, s->cluster_size);
- if ((*refcount_table)[cluster] != 1) {
- fprintf(stderr, "%s refcount block %" PRId64
- " refcount=%d\n",
- fix & BDRV_FIX_ERRORS ? "Repairing" :
- "ERROR",
- i, (*refcount_table)[cluster]);
-
- if (fix & BDRV_FIX_ERRORS) {
- int64_t new_offset;
- new_offset = realloc_refcount_block(bs, i, offset);
- if (new_offset < 0) {
- res->corruptions++;
- continue;
- }
-
- /* update refcounts */
- if ((new_offset >> s->cluster_bits) >= *nb_clusters) {
- /* increase refcount_table size if necessary */
- int old_nb_clusters = *nb_clusters;
- *nb_clusters = (new_offset >> s->cluster_bits) + 1;
-
- *refcount_table = g_try_realloc(*refcount_table,
- *nb_clusters * sizeof(uint16_t));
- if (!*refcount_table) {
- res->check_errors++;
- return -ENOMEM;
- }
-
- memset(&(*refcount_table)[old_nb_clusters], 0,
- (*nb_clusters - old_nb_clusters) *
- sizeof(uint16_t));
- }
- (*refcount_table)[cluster]--;
- inc_refcounts(bs, res, *refcount_table, *nb_clusters,
- new_offset, s->cluster_size);
-
- res->corruptions_fixed++;
- } else {
- res->corruptions++;
- }
+ if ((*refcount_table)[cluster] != 1) {
+ fprintf(stderr, "ERROR refcount block %" PRId64
+ " refcount=%d\n", i, (*refcount_table)[cluster]);
+ res->corruptions++;
+ *rebuild = true;
}
}
}
@@ -1623,8 +1475,8 @@ resize_fail:
* Calculates an in-memory refcount table.
*/
static int calculate_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
- BdrvCheckMode fix, uint16_t **refcount_table,
- int64_t *nb_clusters)
+ BdrvCheckMode fix, bool *rebuild,
+ uint16_t **refcount_table, int64_t *nb_clusters)
{
BDRVQcowState *s = bs->opaque;
QCowSnapshot *sn;
@@ -1667,7 +1519,7 @@ static int calculate_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
s->refcount_table_offset,
s->refcount_table_size * sizeof(uint64_t));
- return check_refblocks(bs, res, fix, refcount_table, nb_clusters);
+ return check_refblocks(bs, res, fix, rebuild, refcount_table, nb_clusters);
}
/*
@@ -1675,7 +1527,8 @@ static int calculate_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
* refcount as reported by the refcount structures on-disk.
*/
static void compare_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
- BdrvCheckMode fix, int64_t *highest_cluster,
+ BdrvCheckMode fix, bool *rebuild,
+ int64_t *highest_cluster,
uint16_t *refcount_table, int64_t nb_clusters)
{
BDRVQcowState *s = bs->opaque;
@@ -1702,10 +1555,15 @@ static void compare_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
int *num_fixed = NULL;
if (refcount1 > refcount2 && (fix & BDRV_FIX_LEAKS)) {
num_fixed = &res->leaks_fixed;
- } else if (refcount1 < refcount2 && (fix & BDRV_FIX_ERRORS)) {
+ } else if (refcount1 > 0 && refcount1 < refcount2 &&
+ (fix & BDRV_FIX_ERRORS)) {
num_fixed = &res->corruptions_fixed;
}
+ if (refcount1 == 0) {
+ *rebuild = true;
+ }
+
fprintf(stderr, "%s cluster %" PRId64 " refcount=%d reference=%d\n",
num_fixed != NULL ? "Repairing" :
refcount1 < refcount2 ? "ERROR" :
@@ -1744,6 +1602,7 @@ int qcow2_check_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
BDRVQcowState *s = bs->opaque;
int64_t size, highest_cluster, nb_clusters;
uint16_t *refcount_table = NULL;
+ bool rebuild = false;
int ret;
size = bdrv_getlength(bs->file);
@@ -1761,14 +1620,19 @@ int qcow2_check_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
res->bfi.total_clusters =
size_to_clusters(s, bs->total_sectors * BDRV_SECTOR_SIZE);
- ret = calculate_refcounts(bs, res, fix, &refcount_table, &nb_clusters);
+ ret = calculate_refcounts(bs, res, fix, &rebuild, &refcount_table,
+ &nb_clusters);
if (ret < 0) {
goto fail;
}
- compare_refcounts(bs, res, fix, &highest_cluster, refcount_table,
+ compare_refcounts(bs, res, fix, &rebuild, &highest_cluster, refcount_table,
nb_clusters);
+ if (rebuild) {
+ fprintf(stderr, "ERROR need to rebuild refcount structures\n");
+ }
+
/* check OFLAG_COPIED */
ret = check_oflag_copied(bs, res, fix);
if (ret < 0) {
--
2.0.3
^ permalink raw reply related [flat|nested] 20+ messages in thread* [Qemu-devel] [PATCH v2 6/9] qcow2: Rebuild refcount structure during check
2014-08-15 15:16 [Qemu-devel] [PATCH v2 0/9] qcow2: Fix image repairing Max Reitz
` (4 preceding siblings ...)
2014-08-15 15:16 ` [Qemu-devel] [PATCH v2 5/9] qcow2: Do not perform potentially damaging repairs Max Reitz
@ 2014-08-15 15:16 ` Max Reitz
2014-08-15 15:16 ` [Qemu-devel] [PATCH v2 7/9] qcow2: Clean up after refcount rebuild Max Reitz
` (2 subsequent siblings)
8 siblings, 0 replies; 20+ messages in thread
From: Max Reitz @ 2014-08-15 15:16 UTC (permalink / raw)
To: qemu-devel; +Cc: Kevin Wolf, Benoît Canet, Stefan Hajnoczi, Max Reitz
The previous commit introduced the "rebuild" variable to qcow2's
implementation of the image consistency check. Now make use of this by
adding a function which creates a completely new refcount structure
based solely on the in-memory information gathered before.
The old refcount structure will be leaked, however.
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
block/qcow2-refcount.c | 265 ++++++++++++++++++++++++++++++++++++++++++++++++-
1 file changed, 262 insertions(+), 3 deletions(-)
diff --git a/block/qcow2-refcount.c b/block/qcow2-refcount.c
index adfca45..a442b8d 100644
--- a/block/qcow2-refcount.c
+++ b/block/qcow2-refcount.c
@@ -1591,6 +1591,245 @@ static void compare_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
}
/*
+ * Allocates a cluster using an in-memory refcount table (IMRT) in contrast to
+ * the on-disk refcount structures.
+ *
+ * *first_free_cluster does not necessarily point to the first free cluster, but
+ * may point to one cluster as close as possible before it. The offset returned
+ * will never be before that cluster.
+ *
+ * Note that *first_free_cluster is a cluster index whereas the return value is
+ * an offset.
+ */
+static int64_t alloc_clusters_imrt(BlockDriverState *bs,
+ int cluster_count,
+ uint16_t **refcount_table,
+ int64_t *nb_clusters,
+ int64_t *first_free_cluster)
+{
+ BDRVQcowState *s = bs->opaque;
+ int64_t cluster = *first_free_cluster, i;
+ bool first_gap = true;
+ int contiguous_clusters;
+
+ for (contiguous_clusters = 0;
+ cluster < *nb_clusters && contiguous_clusters < cluster_count;
+ cluster++)
+ {
+ if (!(*refcount_table)[cluster]) {
+ contiguous_clusters++;
+ if (first_gap) {
+ *first_free_cluster = cluster;
+ first_gap = false;
+ }
+ } else if (contiguous_clusters) {
+ contiguous_clusters = 0;
+ }
+ }
+
+ if (contiguous_clusters < cluster_count) {
+ int64_t old_nb_clusters = *nb_clusters;
+
+ *nb_clusters = cluster + cluster_count - contiguous_clusters;
+ *refcount_table = g_try_realloc(*refcount_table,
+ *nb_clusters * sizeof(uint16_t));
+ if (!*refcount_table) {
+ return -ENOMEM;
+ }
+
+ memset(*refcount_table + old_nb_clusters, 0,
+ (*nb_clusters - old_nb_clusters) * sizeof(uint16_t));
+ }
+
+ cluster -= contiguous_clusters;
+ for (i = 0; i < cluster_count; i++) {
+ (*refcount_table)[cluster + i] = 1;
+ }
+
+ return cluster << s->cluster_bits;
+}
+
+/*
+ * Creates a new refcount structure based solely on the in-memory information
+ * given through *refcount_table. All necessary allocations will be reflected
+ * in that array.
+ *
+ * On success, the old refcount structure is leaked (it will be covered by the
+ * new refcount structure).
+ */
+static int rebuild_refcount_structure(BlockDriverState *bs,
+ BdrvCheckResult *res,
+ uint16_t **refcount_table,
+ int64_t *nb_clusters)
+{
+ BDRVQcowState *s = bs->opaque;
+ int64_t first_free_cluster = 0, rt_ofs = -1, cluster = 0;
+ int64_t rb_ofs, rb_start, rb_index;
+ uint32_t reftable_size = 0;
+ uint64_t *reftable = NULL;
+ uint16_t *on_disk_rb;
+ uint8_t rt_offset_and_clusters[sizeof(uint64_t) + sizeof(uint32_t)];
+ int i, ret = 0;
+
+ qcow2_cache_empty(bs, s->refcount_block_cache);
+
+write_refblocks:
+ for (; cluster < *nb_clusters; cluster++) {
+ if (!(*refcount_table)[cluster]) {
+ continue;
+ }
+
+ rb_index = cluster >> (s->cluster_bits - 1);
+ rb_start = rb_index << (s->cluster_bits - 1);
+
+ /* Don't allocate a cluster in a refblock already written to disk */
+ if (first_free_cluster < rb_start) {
+ first_free_cluster = rb_start;
+ }
+ rb_ofs = alloc_clusters_imrt(bs, 1, refcount_table, nb_clusters,
+ &first_free_cluster);
+ if (rb_ofs < 0) {
+ fprintf(stderr, "ERROR allocating refblock: %s\n", strerror(-ret));
+ res->check_errors++;
+ ret = rb_ofs;
+ goto fail;
+ }
+
+ if (reftable_size <= rb_index) {
+ uint32_t old_rt_size = reftable_size;
+ reftable_size = ROUND_UP((rb_index + 1) * sizeof(uint64_t),
+ s->cluster_size) / sizeof(uint64_t);
+ reftable = g_try_realloc(reftable,
+ reftable_size * sizeof(uint64_t));
+ if (!reftable) {
+ res->check_errors++;
+ ret = -ENOMEM;
+ goto fail;
+ }
+
+ memset(reftable + old_rt_size, 0,
+ (reftable_size - old_rt_size) * sizeof(uint64_t));
+
+ /* The offset we have for the reftable is now no longer valid;
+ * this will leak that range, but we can easily fix that by running
+ * a leak-fixing check after this rebuild operation */
+ rt_ofs = -1;
+ }
+ reftable[rb_index] = rb_ofs;
+
+ /* If this is apparently the last refblock (for now), try to squeeze the
+ * reftable in */
+ if (rb_index == (*nb_clusters - 1) >> (s->cluster_bits - 1) &&
+ rt_ofs < 0)
+ {
+ rt_ofs = alloc_clusters_imrt(bs, size_to_clusters(s, reftable_size *
+ sizeof(uint64_t)),
+ refcount_table, nb_clusters,
+ &first_free_cluster);
+ if (rt_ofs < 0) {
+ fprintf(stderr, "ERROR allocating reftable: %s\n",
+ strerror(-ret));
+ res->check_errors++;
+ ret = rt_ofs;
+ goto fail;
+ }
+ }
+
+ ret = qcow2_pre_write_overlap_check(bs, 0, rb_ofs, s->cluster_size);
+ if (ret < 0) {
+ fprintf(stderr, "ERROR writing refblock: %s\n", strerror(-ret));
+ goto fail;
+ }
+
+ on_disk_rb = g_malloc0(s->cluster_size);
+ for (i = 0; i < s->cluster_size / sizeof(uint16_t) &&
+ rb_start + i < *nb_clusters; i++)
+ {
+ on_disk_rb[i] = cpu_to_be16((*refcount_table)[rb_start + i]);
+ }
+
+ ret = bdrv_write(bs->file, rb_ofs / BDRV_SECTOR_SIZE,
+ (void *)on_disk_rb, s->cluster_sectors);
+ g_free(on_disk_rb);
+ if (ret < 0) {
+ fprintf(stderr, "ERROR writing refblock: %s\n", strerror(-ret));
+ goto fail;
+ }
+
+ /* Go to the end of this refblock */
+ cluster = rb_start + s->cluster_size / sizeof(uint16_t) - 1;
+ }
+
+ if (rt_ofs < 0) {
+ int64_t post_rb_start = ROUND_UP(*nb_clusters,
+ s->cluster_size / sizeof(uint16_t));
+
+ /* Not pretty but simple */
+ if (first_free_cluster < post_rb_start) {
+ first_free_cluster = post_rb_start;
+ }
+ rt_ofs = alloc_clusters_imrt(bs, size_to_clusters(s, reftable_size *
+ sizeof(uint64_t)),
+ refcount_table, nb_clusters,
+ &first_free_cluster);
+ if (rt_ofs < 0) {
+ fprintf(stderr, "ERROR allocating reftable: %s\n", strerror(-ret));
+ res->check_errors++;
+ ret = rt_ofs;
+ goto fail;
+ }
+
+ goto write_refblocks;
+ }
+
+ assert(reftable);
+
+ for (rb_index = 0; rb_index < reftable_size; rb_index++) {
+ cpu_to_be64s(&reftable[rb_index]);
+ }
+
+ ret = qcow2_pre_write_overlap_check(bs, 0, rt_ofs,
+ reftable_size * sizeof(uint64_t));
+ if (ret < 0) {
+ fprintf(stderr, "ERROR writing reftable: %s\n", strerror(-ret));
+ goto fail;
+ }
+
+ ret = bdrv_write(bs->file, rt_ofs / BDRV_SECTOR_SIZE, (void *)reftable,
+ reftable_size * sizeof(uint64_t) / BDRV_SECTOR_SIZE);
+ if (ret < 0) {
+ fprintf(stderr, "ERROR writing reftable: %s\n", strerror(-ret));
+ goto fail;
+ }
+
+ /* Enter new reftable into the image header */
+ cpu_to_be64w((uint64_t *)&rt_offset_and_clusters[0], rt_ofs);
+ cpu_to_be32w((uint32_t *)&rt_offset_and_clusters[sizeof(uint64_t)],
+ size_to_clusters(s, reftable_size * sizeof(uint64_t)));
+ ret = bdrv_pwrite_sync(bs->file, offsetof(QCowHeader,
+ refcount_table_offset),
+ rt_offset_and_clusters,
+ sizeof(rt_offset_and_clusters));
+ if (ret < 0) {
+ fprintf(stderr, "ERROR setting reftable: %s\n", strerror(-ret));
+ goto fail;
+ }
+
+ for (rb_index = 0; rb_index < reftable_size; rb_index++) {
+ be64_to_cpus(&reftable[rb_index]);
+ }
+ s->refcount_table = reftable;
+ s->refcount_table_offset = rt_ofs;
+ s->refcount_table_size = reftable_size;
+
+ return 0;
+
+fail:
+ g_free(reftable);
+ return ret;
+}
+
+/*
* Checks an image for refcount consistency.
*
* Returns 0 if no errors are found, the number of errors in case the image is
@@ -1600,6 +1839,7 @@ int qcow2_check_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
BdrvCheckMode fix)
{
BDRVQcowState *s = bs->opaque;
+ BdrvCheckResult pre_compare_res;
int64_t size, highest_cluster, nb_clusters;
uint16_t *refcount_table = NULL;
bool rebuild = false;
@@ -1626,11 +1866,30 @@ int qcow2_check_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
goto fail;
}
- compare_refcounts(bs, res, fix, &rebuild, &highest_cluster, refcount_table,
+ /* In case we don't need to rebuild the refcount structure (but want to fix
+ * something), this function is immediately called again, in which case the
+ * result should be ignored */
+ pre_compare_res = *res;
+ compare_refcounts(bs, res, 0, &rebuild, &highest_cluster, refcount_table,
nb_clusters);
- if (rebuild) {
- fprintf(stderr, "ERROR need to rebuild refcount structures\n");
+ if (rebuild && (fix & BDRV_FIX_ERRORS)) {
+ fprintf(stderr, "Rebuilding refcount structure\n");
+ ret = rebuild_refcount_structure(bs, res, &refcount_table,
+ &nb_clusters);
+ if (ret < 0) {
+ goto fail;
+ }
+ } else if (fix) {
+ if (rebuild) {
+ fprintf(stderr, "ERROR need to rebuild refcount structures\n");
+ }
+
+ if (res->leaks || res->corruptions) {
+ *res = pre_compare_res;
+ compare_refcounts(bs, res, fix, &rebuild, &highest_cluster,
+ refcount_table, nb_clusters);
+ }
}
/* check OFLAG_COPIED */
--
2.0.3
^ permalink raw reply related [flat|nested] 20+ messages in thread* [Qemu-devel] [PATCH v2 7/9] qcow2: Clean up after refcount rebuild
2014-08-15 15:16 [Qemu-devel] [PATCH v2 0/9] qcow2: Fix image repairing Max Reitz
` (5 preceding siblings ...)
2014-08-15 15:16 ` [Qemu-devel] [PATCH v2 6/9] qcow2: Rebuild refcount structure during check Max Reitz
@ 2014-08-15 15:16 ` Max Reitz
2014-08-15 15:16 ` [Qemu-devel] [PATCH v2 8/9] iotests: Fix test outputs Max Reitz
2014-08-15 15:16 ` [Qemu-devel] [PATCH v2 9/9] iotests: Add test for potentially damaging repairs Max Reitz
8 siblings, 0 replies; 20+ messages in thread
From: Max Reitz @ 2014-08-15 15:16 UTC (permalink / raw)
To: qemu-devel; +Cc: Kevin Wolf, Benoît Canet, Stefan Hajnoczi, Max Reitz
Because the old refcount structure will be leaked after having rebuilt
it, we need to recalculate the refcounts and run a leak-fixing operation
afterwards (if leaks should be fixed at all).
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
block/qcow2-refcount.c | 35 +++++++++++++++++++++++++++++++++++
1 file changed, 35 insertions(+)
diff --git a/block/qcow2-refcount.c b/block/qcow2-refcount.c
index a442b8d..cc2a304 100644
--- a/block/qcow2-refcount.c
+++ b/block/qcow2-refcount.c
@@ -1874,12 +1874,47 @@ int qcow2_check_refcounts(BlockDriverState *bs, BdrvCheckResult *res,
nb_clusters);
if (rebuild && (fix & BDRV_FIX_ERRORS)) {
+ BdrvCheckResult old_res = *res;
+
fprintf(stderr, "Rebuilding refcount structure\n");
ret = rebuild_refcount_structure(bs, res, &refcount_table,
&nb_clusters);
if (ret < 0) {
goto fail;
}
+
+ res->corruptions = 0;
+ res->leaks = 0;
+
+ /* Because the old reftable has been exchanged for a new one the
+ * references have to be recalculated */
+ rebuild = false;
+ memset(refcount_table, 0, nb_clusters * sizeof(uint16_t));
+ ret = calculate_refcounts(bs, res, 0, &rebuild, &refcount_table,
+ &nb_clusters);
+ if (ret < 0) {
+ goto fail;
+ }
+
+ if (fix & BDRV_FIX_LEAKS) {
+ /* The old refcount structures are now leaked, fix it; the result
+ * can be ignored */
+ pre_compare_res = *res;
+ compare_refcounts(bs, res, BDRV_FIX_LEAKS, &rebuild,
+ &highest_cluster, refcount_table, nb_clusters);
+ if (rebuild) {
+ fprintf(stderr, "ERROR rebuilt refcount structure is still "
+ "broken\n");
+ }
+ *res = pre_compare_res;
+ }
+
+ if (res->corruptions < old_res.corruptions) {
+ res->corruptions_fixed += old_res.corruptions - res->corruptions;
+ }
+ if (res->leaks < old_res.leaks) {
+ res->leaks_fixed += old_res.leaks - res->leaks;
+ }
} else if (fix) {
if (rebuild) {
fprintf(stderr, "ERROR need to rebuild refcount structures\n");
--
2.0.3
^ permalink raw reply related [flat|nested] 20+ messages in thread* [Qemu-devel] [PATCH v2 8/9] iotests: Fix test outputs
2014-08-15 15:16 [Qemu-devel] [PATCH v2 0/9] qcow2: Fix image repairing Max Reitz
` (6 preceding siblings ...)
2014-08-15 15:16 ` [Qemu-devel] [PATCH v2 7/9] qcow2: Clean up after refcount rebuild Max Reitz
@ 2014-08-15 15:16 ` Max Reitz
2014-08-15 15:16 ` [Qemu-devel] [PATCH v2 9/9] iotests: Add test for potentially damaging repairs Max Reitz
8 siblings, 0 replies; 20+ messages in thread
From: Max Reitz @ 2014-08-15 15:16 UTC (permalink / raw)
To: qemu-devel; +Cc: Kevin Wolf, Benoît Canet, Stefan Hajnoczi, Max Reitz
039, 060 and 061 all create images with referenced clusters having a
refcount of 0. Because previous commits changed handling of such errors,
these tests now have a different output. Fix it.
Furthermore, 060 created a refblock with a refcount greater than one
which now results in having to rebuild the refcount structure as well.
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
tests/qemu-iotests/039.out | 10 ++++++++--
tests/qemu-iotests/060.out | 10 ++++++++--
tests/qemu-iotests/061.out | 18 ++++++++++++------
3 files changed, 28 insertions(+), 10 deletions(-)
diff --git a/tests/qemu-iotests/039.out b/tests/qemu-iotests/039.out
index 67e7744..0adf153 100644
--- a/tests/qemu-iotests/039.out
+++ b/tests/qemu-iotests/039.out
@@ -25,7 +25,10 @@ read 512/512 bytes at offset 0
incompatible_features 0x1
== Repairing the image file must succeed ==
-Repairing cluster 5 refcount=0 reference=1
+ERROR cluster 5 refcount=0 reference=1
+Rebuilding refcount structure
+Repairing cluster 1 refcount=1 reference=0
+Repairing cluster 2 refcount=1 reference=0
The following inconsistencies were found and repaired:
0 leaked clusters
@@ -45,7 +48,10 @@ wrote 512/512 bytes at offset 0
512 bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
./039: Aborted ( ulimit -c 0; exec "$@" )
incompatible_features 0x1
-Repairing cluster 5 refcount=0 reference=1
+ERROR cluster 5 refcount=0 reference=1
+Rebuilding refcount structure
+Repairing cluster 1 refcount=1 reference=0
+Repairing cluster 2 refcount=1 reference=0
wrote 512/512 bytes at offset 0
512 bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
incompatible_features 0x0
diff --git a/tests/qemu-iotests/060.out b/tests/qemu-iotests/060.out
index c27c952..41ecbc7 100644
--- a/tests/qemu-iotests/060.out
+++ b/tests/qemu-iotests/060.out
@@ -27,11 +27,15 @@ incompatible_features 0x0
qcow2: Preventing invalid write on metadata (overlaps with refcount block); image marked as corrupt.
write failed: Input/output error
incompatible_features 0x2
-Repairing refcount block 0 refcount=2
+ERROR refcount block 0 refcount=2
+ERROR cluster 2 refcount=1 reference=2
+Rebuilding refcount structure
+Repairing cluster 1 refcount=1 reference=0
+Repairing cluster 2 refcount=2 reference=1
The following inconsistencies were found and repaired:
0 leaked clusters
- 1 corruptions
+ 2 corruptions
Double checking the fixed image now...
No errors were found on the image.
@@ -59,6 +63,8 @@ incompatible_features 0x0
qcow2: Preventing invalid write on metadata (overlaps with inactive L2 table); image marked as corrupt.
write failed: Input/output error
incompatible_features 0x2
+ERROR cluster 4 refcount=1 reference=2
+Leaked cluster 9 refcount=1 reference=0
Repairing cluster 4 refcount=1 reference=2
Repairing cluster 9 refcount=1 reference=0
Repairing OFLAG_COPIED data cluster: l2_entry=8000000000040000 refcount=2
diff --git a/tests/qemu-iotests/061.out b/tests/qemu-iotests/061.out
index e372470..4ae6472 100644
--- a/tests/qemu-iotests/061.out
+++ b/tests/qemu-iotests/061.out
@@ -76,8 +76,11 @@ autoclear_features 0x0
refcount_order 4
header_length 104
-Repairing cluster 5 refcount=0 reference=1
-Repairing cluster 6 refcount=0 reference=1
+ERROR cluster 5 refcount=0 reference=1
+ERROR cluster 6 refcount=0 reference=1
+Rebuilding refcount structure
+Repairing cluster 1 refcount=1 reference=0
+Repairing cluster 2 refcount=1 reference=0
magic 0x514649fb
version 2
backing_file_offset 0x0
@@ -87,7 +90,7 @@ size 67108864
crypt_method 0
l1_size 1
l1_table_offset 0x30000
-refcount_table_offset 0x10000
+refcount_table_offset 0x80000
refcount_table_clusters 1
nb_snapshots 0
snapshot_offset 0x0
@@ -230,8 +233,11 @@ autoclear_features 0x0
refcount_order 4
header_length 104
-Repairing cluster 5 refcount=0 reference=1
-Repairing cluster 6 refcount=0 reference=1
+ERROR cluster 5 refcount=0 reference=1
+ERROR cluster 6 refcount=0 reference=1
+Rebuilding refcount structure
+Repairing cluster 1 refcount=1 reference=0
+Repairing cluster 2 refcount=1 reference=0
magic 0x514649fb
version 3
backing_file_offset 0x0
@@ -241,7 +247,7 @@ size 67108864
crypt_method 0
l1_size 1
l1_table_offset 0x30000
-refcount_table_offset 0x10000
+refcount_table_offset 0x80000
refcount_table_clusters 1
nb_snapshots 0
snapshot_offset 0x0
--
2.0.3
^ permalink raw reply related [flat|nested] 20+ messages in thread* [Qemu-devel] [PATCH v2 9/9] iotests: Add test for potentially damaging repairs
2014-08-15 15:16 [Qemu-devel] [PATCH v2 0/9] qcow2: Fix image repairing Max Reitz
` (7 preceding siblings ...)
2014-08-15 15:16 ` [Qemu-devel] [PATCH v2 8/9] iotests: Fix test outputs Max Reitz
@ 2014-08-15 15:16 ` Max Reitz
8 siblings, 0 replies; 20+ messages in thread
From: Max Reitz @ 2014-08-15 15:16 UTC (permalink / raw)
To: qemu-devel; +Cc: Kevin Wolf, Benoît Canet, Stefan Hajnoczi, Max Reitz
There are certain cases where repairing a qcow2 image might actually
damage it further (or rather, where repairing it has in fact damaged it
further with the old qcow2 check implementation). This should not
happen, so add a test for these cases.
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
tests/qemu-iotests/101 | 98 ++++++++++++++++++++++++++++++++++++++++++++++
tests/qemu-iotests/101.out | 46 ++++++++++++++++++++++
tests/qemu-iotests/group | 1 +
3 files changed, 145 insertions(+)
create mode 100755 tests/qemu-iotests/101
create mode 100644 tests/qemu-iotests/101.out
diff --git a/tests/qemu-iotests/101 b/tests/qemu-iotests/101
new file mode 100755
index 0000000..cc6d0b3
--- /dev/null
+++ b/tests/qemu-iotests/101
@@ -0,0 +1,98 @@
+#!/bin/bash
+#
+# Test case for repairing qcow2 images which cannot be repaired using
+# the on-disk refcount structures
+#
+# Copyright (C) 2013 Red Hat, Inc.
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program. If not, see <http://www.gnu.org/licenses/>.
+#
+
+# creator
+owner=mreitz@redhat.com
+
+seq="$(basename $0)"
+echo "QA output created by $seq"
+
+here="$PWD"
+tmp=/tmp/$$
+status=1 # failure is the default!
+
+_cleanup()
+{
+ _cleanup_test_img
+}
+trap "_cleanup; exit \$status" 0 1 2 3 15
+
+# get standard environment, filters and checks
+. ./common.rc
+. ./common.filter
+
+# This tests qocw2-specific low-level functionality
+_supported_fmt qcow2
+_supported_proto file
+_supported_os Linux
+
+echo
+echo '=== Repairing an image without any refcount table ==='
+echo
+
+_make_test_img 64M
+# just write some data
+$QEMU_IO -c 'write -P 42 0 64k' "$TEST_IMG" | _filter_qemu_io
+
+# refcount_table_offset
+poke_file "$TEST_IMG" 48 "\x00\x00\x00\x00\x00\x00\x00\x00"
+# refcount_table_clusters
+poke_file "$TEST_IMG" 56 "\x00\x00\x00\x00"
+
+_check_test_img -r all
+
+$QEMU_IO -c 'read -P 42 0 64k' "$TEST_IMG" | _filter_qemu_io
+
+echo
+echo '=== Repairing unreferenced data cluster in new refblock area ==='
+echo
+
+IMGOPTS='cluster_size=512' _make_test_img 64M
+# Allocate the first 128 kB in the image (first refblock)
+$QEMU_IO -c 'write 0 111104' "$TEST_IMG" | _filter_qemu_io
+# should be 131072
+stat -c '%s' "$TEST_IMG"
+
+# Enter a cluster at 128 kB (0x20000)
+# XXX: This (0x1ccc8) should be the first free entry in the last L2 table, but
+# we cannot be sure
+poke_file "$TEST_IMG" 117960 "\x80\x00\x00\x00\x00\x02\x00\x00"
+
+# Fill the cluster
+truncate -s 131584 "$TEST_IMG"
+$QEMU_IO -c "open -o driver=raw $TEST_IMG" -c 'write -P 42 128k 512' \
+ | _filter_qemu_io
+
+# The data should now appear at this guest offset
+$QEMU_IO -c 'read -P 42 111104 512' "$TEST_IMG" | _filter_qemu_io
+
+# This cluster is unallocated; fix it
+_check_test_img -r all
+
+# This repair operation must have allocated a new refblock; and that refblock
+# should not overlap with the unallocated data cluster. If it does, the data
+# will be damaged, so check it.
+$QEMU_IO -c 'read -P 42 111104 512' "$TEST_IMG" | _filter_qemu_io
+
+# success, all done
+echo '*** done'
+rm -f $seq.full
+status=0
diff --git a/tests/qemu-iotests/101.out b/tests/qemu-iotests/101.out
new file mode 100644
index 0000000..d0e2149
--- /dev/null
+++ b/tests/qemu-iotests/101.out
@@ -0,0 +1,46 @@
+QA output created by 101
+
+=== Repairing an image without any refcount table ===
+
+Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864
+wrote 65536/65536 bytes at offset 0
+64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+ERROR cluster 0 refcount=0 reference=1
+ERROR cluster 3 refcount=0 reference=1
+ERROR cluster 4 refcount=0 reference=1
+ERROR cluster 5 refcount=0 reference=1
+Rebuilding refcount structure
+The following inconsistencies were found and repaired:
+
+ 0 leaked clusters
+ 4 corruptions
+
+Double checking the fixed image now...
+No errors were found on the image.
+read 65536/65536 bytes at offset 0
+64 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+
+=== Repairing unreferenced data cluster in new refblock area ===
+
+Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=67108864
+wrote 111104/111104 bytes at offset 0
+108.500 KiB, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+131072
+wrote 512/512 bytes at offset 131072
+512 bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+read 512/512 bytes at offset 111104
+512 bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+ERROR cluster 256 refcount=0 reference=1
+Rebuilding refcount structure
+Repairing cluster 1 refcount=1 reference=0
+Repairing cluster 2 refcount=1 reference=0
+The following inconsistencies were found and repaired:
+
+ 0 leaked clusters
+ 1 corruptions
+
+Double checking the fixed image now...
+No errors were found on the image.
+read 512/512 bytes at offset 111104
+512 bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
+*** done
diff --git a/tests/qemu-iotests/group b/tests/qemu-iotests/group
index 6e67f61..e25e992 100644
--- a/tests/qemu-iotests/group
+++ b/tests/qemu-iotests/group
@@ -100,3 +100,4 @@
091 rw auto quick
092 rw auto quick
095 rw auto quick
+101 rw auto quick
--
2.0.3
^ permalink raw reply related [flat|nested] 20+ messages in thread