This adds a salvaging mode (--salvage) to qemu-img convert which ignores
read errors and treats the respective areas as containing only zeroes.
This can be used for instance to at least partially recover the data
from terminally corrupted qcow2 images.
Signed-off-by: Max Reitz <mreitz@redhat.com>
---
qemu-img.c | 85 ++++++++++++++++++++++++++++++++++++------------
qemu-img-cmds.hx | 4 +--
qemu-img.texi | 5 +++
3 files changed, 71 insertions(+), 23 deletions(-)
diff --git a/qemu-img.c b/qemu-img.c
index c53666aa41..c2216e67a6 100644
--- a/qemu-img.c
+++ b/qemu-img.c
@@ -66,6 +66,7 @@ enum {
OPTION_SIZE = 264,
OPTION_PREALLOCATION = 265,
OPTION_SHRINK = 266,
+ OPTION_SALVAGE = 267,
};
typedef enum OutputFormat {
@@ -1569,6 +1570,7 @@ typedef struct ImgConvertState {
int64_t target_backing_sectors; /* negative if unknown */
bool wr_in_order;
bool copy_range;
+ bool salvage;
bool quiet;
int min_sparse;
int alignment;
@@ -1616,25 +1618,44 @@ static int convert_iteration_sectors(ImgConvertState *s, int64_t sector_num)
}
if (s->sector_next_status <= sector_num) {
- int64_t count = n * BDRV_SECTOR_SIZE;
+ uint64_t offset = (sector_num - src_cur_offset) * BDRV_SECTOR_SIZE;
+ int64_t count;
- if (s->target_has_backing) {
+ do {
+ count = n * BDRV_SECTOR_SIZE;
+
+ if (s->target_has_backing) {
+ ret = bdrv_block_status(blk_bs(s->src[src_cur]), offset,
+ count, &count, NULL, NULL);
+ } else {
+ ret = bdrv_block_status_above(blk_bs(s->src[src_cur]), NULL,
+ offset, count, &count, NULL,
+ NULL);
+ }
+
+ if (ret < 0) {
+ if (s->salvage) {
+ if (n == 1) {
+ if (!s->quiet) {
+ warn_report("error while reading block status at "
+ "offset %" PRIu64 ": %s", offset,
+ strerror(-ret));
+ }
+ /* Just try to read the data, then */
+ ret = BDRV_BLOCK_DATA;
+ count = BDRV_SECTOR_SIZE;
+ } else {
+ /* Retry on a shorter range */
+ n = DIV_ROUND_UP(n, 4);
+ }
+ } else {
+ error_report("error while reading block status at offset "
+ "%" PRIu64 ": %s", offset, strerror(-ret));
+ return ret;
+ }
+ }
+ } while (ret < 0);
- ret = bdrv_block_status(blk_bs(s->src[src_cur]),
- (sector_num - src_cur_offset) *
- BDRV_SECTOR_SIZE,
- count, &count, NULL, NULL);
- } else {
- ret = bdrv_block_status_above(blk_bs(s->src[src_cur]), NULL,
- (sector_num - src_cur_offset) *
- BDRV_SECTOR_SIZE,
- count, &count, NULL, NULL);
- }
- if (ret < 0) {
- error_report("error while reading block status of sector %" PRId64
- ": %s", sector_num, strerror(-ret));
- return ret;
- }
n = DIV_ROUND_UP(count, BDRV_SECTOR_SIZE);
if (ret & BDRV_BLOCK_ZERO) {
@@ -1671,6 +1692,7 @@ static int convert_iteration_sectors(ImgConvertState *s, int64_t sector_num)
static int coroutine_fn convert_co_read(ImgConvertState *s, int64_t sector_num,
int nb_sectors, uint8_t *buf)
{
+ uint64_t single_read_until = 0;
int n, ret;
QEMUIOVector qiov;
@@ -1679,6 +1701,7 @@ static int coroutine_fn convert_co_read(ImgConvertState *s, int64_t sector_num,
BlockBackend *blk;
int src_cur;
int64_t bs_sectors, src_cur_offset;
+ uint64_t offset;
/* In the case of compression with multiple source files, we can get a
* nb_sectors that spreads into the next part. So we must be able to
@@ -1687,14 +1710,30 @@ static int coroutine_fn convert_co_read(ImgConvertState *s, int64_t sector_num,
blk = s->src[src_cur];
bs_sectors = s->src_sectors[src_cur];
+ offset = (sector_num - src_cur_offset) << BDRV_SECTOR_BITS;
+
n = MIN(nb_sectors, bs_sectors - (sector_num - src_cur_offset));
+ if (single_read_until > offset) {
+ n = 1;
+ }
qemu_iovec_init_buf(&qiov, buf, n << BDRV_SECTOR_BITS);
- ret = blk_co_preadv(
- blk, (sector_num - src_cur_offset) << BDRV_SECTOR_BITS,
- n << BDRV_SECTOR_BITS, &qiov, 0);
+ ret = blk_co_preadv(blk, offset, n << BDRV_SECTOR_BITS, &qiov, 0);
if (ret < 0) {
- return ret;
+ if (s->salvage) {
+ if (n > 1) {
+ single_read_until = offset + (n << BDRV_SECTOR_BITS);
+ continue;
+ } else {
+ if (!s->quiet) {
+ warn_report("error while reading offset %" PRIu64
+ ": %s", offset, strerror(-ret));
+ }
+ memset(buf, 0, BDRV_SECTOR_SIZE);
+ }
+ } else {
+ return ret;
+ }
}
sector_num += n;
@@ -2028,6 +2067,7 @@ static int img_convert(int argc, char **argv)
{"image-opts", no_argument, 0, OPTION_IMAGE_OPTS},
{"force-share", no_argument, 0, 'U'},
{"target-image-opts", no_argument, 0, OPTION_TARGET_IMAGE_OPTS},
+ {"salvage", no_argument, 0, OPTION_SALVAGE},
{0, 0, 0, 0}
};
c = getopt_long(argc, argv, ":hf:O:B:Cco:l:S:pt:T:qnm:WU",
@@ -2145,6 +2185,9 @@ static int img_convert(int argc, char **argv)
case OPTION_IMAGE_OPTS:
image_opts = true;
break;
+ case OPTION_SALVAGE:
+ s.salvage = true;
+ break;
case OPTION_TARGET_IMAGE_OPTS:
tgt_image_opts = true;
break;
diff --git a/qemu-img-cmds.hx b/qemu-img-cmds.hx
index 1526f327a5..d6f444586c 100644
--- a/qemu-img-cmds.hx
+++ b/qemu-img-cmds.hx
@@ -44,9 +44,9 @@ STEXI
ETEXI
DEF("convert", img_convert,
- "convert [--object objectdef] [--image-opts] [--target-image-opts] [-U] [-C] [-c] [-p] [-q] [-n] [-f fmt] [-t cache] [-T src_cache] [-O output_fmt] [-B backing_file] [-o options] [-l snapshot_param] [-S sparse_size] [-m num_coroutines] [-W] filename [filename2 [...]] output_filename")
+ "convert [--object objectdef] [--image-opts] [--target-image-opts] [-U] [-C] [-c] [-p] [-q] [-n] [-f fmt] [-t cache] [-T src_cache] [-O output_fmt] [-B backing_file] [-o options] [-l snapshot_param] [-S sparse_size] [-m num_coroutines] [-W] [--salvage] filename [filename2 [...]] output_filename")
STEXI
-@item convert [--object @var{objectdef}] [--image-opts] [--target-image-opts] [-U] [-c] [-p] [-q] [-n] [-f @var{fmt}] [-t @var{cache}] [-T @var{src_cache}] [-O @var{output_fmt}] [-B @var{backing_file}] [-o @var{options}] [-l @var{snapshot_param}] [-S @var{sparse_size}] [-m @var{num_coroutines}] [-W] @var{filename} [@var{filename2} [...]] @var{output_filename}
+@item convert [--object @var{objectdef}] [--image-opts] [--target-image-opts] [-U] [-c] [-p] [-q] [-n] [-f @var{fmt}] [-t @var{cache}] [-T @var{src_cache}] [-O @var{output_fmt}] [-B @var{backing_file}] [-o @var{options}] [-l @var{snapshot_param}] [-S @var{sparse_size}] [-m @var{num_coroutines}] [-W] [--salvage] @var{filename} [@var{filename2} [...]] @var{output_filename}
ETEXI
DEF("create", img_create,
diff --git a/qemu-img.texi b/qemu-img.texi
index 3b6710a580..827ee9fe32 100644
--- a/qemu-img.texi
+++ b/qemu-img.texi
@@ -175,6 +175,11 @@ improve performance if the data is remote, such as with NFS or iSCSI backends,
but will not automatically sparsify zero sectors, and may result in a fully
allocated target image depending on the host support for getting allocation
information.
+@item --salvage
+Try to ignore I/O errors when reading. Unless in quiet mode (@code{-q}), errors
+will still be printed. Areas that cannot be read from the source will be
+treated as containing only zeroes. This option has no effect in copy offloading
+mode (@code{-C}).
@end table
Parameters to dd subcommand:
--
2.20.1
10.04.2019 23:57, Max Reitz wrote:
> This adds a salvaging mode (--salvage) to qemu-img convert which ignores
> read errors and treats the respective areas as containing only zeroes.
> This can be used for instance to at least partially recover the data
> from terminally corrupted qcow2 images.
>
> Signed-off-by: Max Reitz <mreitz@redhat.com>
> ---
[..]
> --- a/qemu-img.texi
> +++ b/qemu-img.texi
> @@ -175,6 +175,11 @@ improve performance if the data is remote, such as with NFS or iSCSI backends,
> but will not automatically sparsify zero sectors, and may result in a fully
> allocated target image depending on the host support for getting allocation
> information.
> +@item --salvage
> +Try to ignore I/O errors when reading. Unless in quiet mode (@code{-q}), errors
> +will still be printed. Areas that cannot be read from the source will be
> +treated as containing only zeroes. This option has no effect in copy offloading
> +mode (@code{-C}).
Isn't it better to restrict such it, instead of creating meaningless combination?
And drop the restriction if we support it in future, instead of silent move from
noop to the feature?
Otherwise, looks OK for me. I dislike mixing byte-based and sector-based logic, but it
is not about the series, we should move to byte-based like in other places, and it may
be done later.
--
Best regards,
Vladimir
On 11.04.19 16:33, Vladimir Sementsov-Ogievskiy wrote:
> 10.04.2019 23:57, Max Reitz wrote:
>> This adds a salvaging mode (--salvage) to qemu-img convert which ignores
>> read errors and treats the respective areas as containing only zeroes.
>> This can be used for instance to at least partially recover the data
>> from terminally corrupted qcow2 images.
>>
>> Signed-off-by: Max Reitz <mreitz@redhat.com>
>> ---
>
> [..]
>
>> --- a/qemu-img.texi
>> +++ b/qemu-img.texi
>> @@ -175,6 +175,11 @@ improve performance if the data is remote, such as with NFS or iSCSI backends,
>> but will not automatically sparsify zero sectors, and may result in a fully
>> allocated target image depending on the host support for getting allocation
>> information.
>> +@item --salvage
>> +Try to ignore I/O errors when reading. Unless in quiet mode (@code{-q}), errors
>> +will still be printed. Areas that cannot be read from the source will be
>> +treated as containing only zeroes. This option has no effect in copy offloading
>> +mode (@code{-C}).
>
> Isn't it better to restrict such it, instead of creating meaningless combination?
> And drop the restriction if we support it in future, instead of silent move from
> noop to the feature?
You’re right, I should do that.
> Otherwise, looks OK for me. I dislike mixing byte-based and sector-based logic, but it
> is not about the series, we should move to byte-based like in other places, and it may
> be done later.
Thanks for having a look!
Max
© 2016 - 2026 Red Hat, Inc.