Skip to contentSkip to conversation

qemu-img: Add salvaging mode to convert

PATCH Superseded Max Reitz 3 participants 15 messages Dec 3, 2018 – Feb 15, 2019 Block layer

v1 of 4 Superseded Latest is v4 →

lore
Share & tools
lore

Max Reitz

author
#0 lore raw

cover qemu-img: Add salvaging mode to convert

Hi,

This series adds a --salvage option to qemu-img convert. With this, qemu-img will not abort when it encounters an I/O error. Instead, it tries to narrow it down and will treat the affected sectors as being completely 0 (and print a warning).

Testing this is not so easy, because while real I/O errors during read operations should be treated as described above, errors encountered during bdrv_block_status() should just be ignored and the affected sectors should be considered allocated. But blkdebug does not yet have a way to intercept this, and:

(1) Just adding a new block-status event would be silly, because I don't want an event, I want it to fail on a certain kind of operation, on a certain sector range, independently of any events, so why can't we just do that? See patch 4.

(2) If we just make blkdebug intercept .bdrv_co_block_status() like all other kinds of operations, at least iotest 041 fails, which does exactly that silly thing: It uses the read_aio event to wait for any read. But it turns out that there may be a bdrv_*block_status() call in between, so suddenly the wrong operation yields an error. As I said, the real fault here is that it does not really make sense to pray that the operation you want to fail is the one that is immediately executed after some event that you hope will trigger that operation.
See patch 3.

So patch 3 allows blkdebug users to select which kind of I/O operation they actually want to make fail, and patch 4 allows them to not use any event, but to have a rule active all the time.

Together, we can then enable error injection for block-status in patch 5 and make use of event=none iotype=block-status in patch 6.

Max Reitz (6):
qemu-img: Move quiet into ImgConvertState
qemu-img: Add salvaging mode to convert
blkdebug: Add @iotype error option
blkdebug: Add "none" event
blkdebug: Inject errors on .bdrv_co_block_status()
iotests: Test qemu-img convert --salvage

 qapi/block-core.json       |  33 +++++++-
 block/blkdebug.c           |  60 ++++++++++++--
 qemu-img.c                 |  97 ++++++++++++++++------
 qemu-img-cmds.hx           |   4 +-
 qemu-img.texi              |   5 ++
 tests/qemu-iotests/236     | 164 +++++++++++++++++++++++++++++++++++++
 tests/qemu-iotests/236.out |  43 ++++++++++
 tests/qemu-iotests/group   |   1 +
 8 files changed, 370 insertions(+), 37 deletions(-)
 create mode 100755 tests/qemu-iotests/236
 create mode 100644 tests/qemu-iotests/236.out
Signature · 1 line
2.19.2
author
#1 lore raw

PATCH 1/6 qemu-img: Move quiet into ImgConvertState

Move img_convert()'s quiet flag into the ImgConvertState so it is accessible by nested functions. -q dictates that it suppresses anything but errors, so if those functions want to emit warnings, they need to query this flag first. (There currently are no such warnings, but there will be as of the next patch.)

---
 qemu-img.c | 13 +++++++------
 1 file changed, 7 insertions(+), 6 deletions(-)
qemu-img.c +7 -6
index ad04f59565..1582ef63f3 100644
@@ -1569,6 +1569,7 @@ typedef struct ImgConvertState {
     int64_t target_backing_sectors; /* negative if unknown */
     bool wr_in_order;
     bool copy_range;
+    bool quiet;
     int min_sparse;
     int alignment;
     size_t cluster_sectors;
@@ -2008,7 +2009,7 @@ static int img_convert(int argc, char **argv)
     QDict *open_opts = NULL;
     char *options = NULL;
     Error *local_err = NULL;
-    bool writethrough, src_writethrough, quiet = false, image_opts = false,
+    bool writethrough, src_writethrough, image_opts = false,
          skip_create = false, progress = false, tgt_image_opts = false;
     int64_t ret = -EINVAL;
     bool force_share = false;
@@ -2116,7 +2117,7 @@ static int img_convert(int argc, char **argv)
             src_cache = optarg;
             break;
         case 'q':
-            quiet = true;
+            s.quiet = true;
             break;
         case 'n':
             skip_create = true;
@@ -2205,7 +2206,7 @@ static int img_convert(int argc, char **argv)
     }
 
     /* Initialize before goto out */
-    if (quiet) {
+    if (s.quiet) {
         progress = false;
     }
     qemu_progress_init(progress, 1.0);
@@ -2216,7 +2217,7 @@ static int img_convert(int argc, char **argv)
 
     for (bs_i = 0; bs_i < s.src_num; bs_i++) {
         s.src[bs_i] = img_open(image_opts, argv[optind + bs_i],
-                               fmt, src_flags, src_writethrough, quiet,
+                               fmt, src_flags, src_writethrough, s.quiet,
                                force_share);
         if (!s.src[bs_i]) {
             ret = -1;
@@ -2379,7 +2380,7 @@ static int img_convert(int argc, char **argv)
 
     if (skip_create) {
         s.target = img_open(tgt_image_opts, out_filename, out_fmt,
-                            flags, writethrough, quiet, false);
+                            flags, writethrough, s.quiet, false);
     } else {
         /* TODO ultimately we should allow --target-image-opts
          * to be used even when -n is not given.
@@ -2387,7 +2388,7 @@ static int img_convert(int argc, char **argv)
          * to allow filenames in option syntax
          */
         s.target = img_open_file(out_filename, open_opts, out_fmt,
-                                 flags, writethrough, quiet, false);
+                                 flags, writethrough, s.quiet, false);
         open_opts = NULL; /* blk_new_open will have freed it */
     }
     if (!s.target) {
Signature · 1 line
2.19.2
Reviewed-by on v1 1/6
#2 lore raw

Re: [Qemu-devel] [PATCH for-next 1/6] qemu-img: Move quiet into ImgConvertState

Quoted 10 lines · Max Reitz

Move img_convert()'s quiet flag into the ImgConvertState so it is accessible by nested functions. -q dictates that it suppresses anything but errors, so if those functions want to emit warnings, they need to query this flag first. (There currently are no such warnings, but there will be as of the next patch.)

---
 qemu-img.c | 13 +++++++------
 1 file changed, 7 insertions(+), 6 deletions(-)
Signature · 3 lines
Eric Blake, Principal Software Engineer
Red Hat, Inc.           +1-919-301-3226
Virtualization:  qemu.org | libvirt.org
author
#3 lore raw

PATCH 2/6 qemu-img: Add salvaging mode to convert

In reply to Max Reitz

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.

---
 qemu-img.c       | 84 +++++++++++++++++++++++++++++++++++++-----------
 qemu-img-cmds.hx |  4 +--
 qemu-img.texi    |  5 +++
 3 files changed, 72 insertions(+), 21 deletions(-)
qemu-img.c +65 -19
index 1582ef63f3..395c1973fe 100644
@@ -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;
@@ -1616,23 +1618,43 @@ 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("Failed to inquire block status of %s "
+                                        "at offset %#" PRIx64 ": %s",
+                                        blk_bs(s->src[src_cur])->filename,
+                                        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 {
+                    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) {
-            return ret;
-        }
         n = DIV_ROUND_UP(count, BDRV_SECTOR_SIZE);
 
         if (ret & BDRV_BLOCK_ZERO) {
@@ -1669,6 +1691,7 @@ 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;
     struct iovec iov;
@@ -1678,6 +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
@@ -1686,16 +1710,34 @@ 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;
+        }
+
         iov.iov_base = buf;
         iov.iov_len = n << BDRV_SECTOR_BITS;
         qemu_iovec_init_external(&qiov, &iov, 1);
 
-        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("Failed to read offset %#" PRIx64 " from "
+                                    "%s: %s", offset, blk_bs(blk)->filename,
+                                    strerror(-ret));
+                    }
+                    memset(buf, 0, BDRV_SECTOR_SIZE);
+                }
+            } else {
+                return ret;
+            }
         }
 
         sector_num += n;
@@ -2031,6 +2073,7 @@ 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",
@@ -2148,6 +2191,9 @@ 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;
qemu-img-cmds.hx +2 -2
index 1526f327a5..d6f444586c 100644
@@ -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,
qemu-img.texi +5 -0
index 3b6710a580..827ee9fe32 100644
@@ -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:
Signature · 1 line
2.19.2
author
#4 lore raw

PATCH 3/6 blkdebug: Add @iotype error option

In reply to Max Reitz

This new error option allows users of blkdebug to inject errors only on certain kinds of I/O operations. Users usually want to make a very specific operation fail, not just any; but right now they simply hope that the event that triggers the error injection is followed up with that very operation. That may not be true, however, because the block layer is changing (including blkdebug, which may increase the number of types of I/O operations on which to inject errors).

The new option's default has been chosen to keep backwards compatibility.

Note that similar to the internal representation, we could choose to expose this option as a list of I/O types. But there is no practical use for this, because as described above, users usually know exactly which kind of operation they want to make fail, so there is no need to specify multiple I/O types at once. In addition, exposing this option as a list would require non-trivial changes to qemu_opts_absorb_qdict().

---
If you want @iotype be a list, I have patches.

---
 qapi/block-core.json | 26 +++++++++++++++++++++++
 block/blkdebug.c     | 50 ++++++++++++++++++++++++++++++++++++--------
 2 files changed, 67 insertions(+), 9 deletions(-)
qapi/block-core.json +26 -0
index d4fe710836..469d889753 100644
@@ -3024,6 +3024,26 @@
             'l1_shrink_write_table', 'l1_shrink_free_l2_clusters',
             'cor_write'] }
 
+##
+# @BlkdebugIOType:
+#
+# Kinds of I/O that blkdebug can inject errors in.
+#
+# @read: .bdrv_co_preadv()
+#
+# @write: .bdrv_co_pwritev()
+#
+# @write-zeroes: .bdrv_co_pwrite_zeroes()
+#
+# @discard: .bdrv_co_pdiscard()
+#
+# @flush: .bdrv_co_flush_to_disk()
+#
+# Since: 4.0
+##
+{ 'enum': 'BlkdebugIOType',
+  'data': [ 'read', 'write', 'write-zeroes', 'discard', 'flush' ] }
+
 ##
 # @BlkdebugInjectErrorOptions:
 #
@@ -3034,6 +3054,11 @@
 # @state:       the state identifier blkdebug needs to be in to
 #               actually trigger the event; defaults to "any"
 #
+# @iotype:      the type of I/O operations on which this error should
+#               be injected; defaults to "all read, write,
+#               write-zeroes, discard, and flush operations"
+#               (since: 4.0)
+#
 # @errno:       error identifier (errno) to be returned; defaults to
 #               EIO
 #
@@ -3051,6 +3076,7 @@
 { 'struct': 'BlkdebugInjectErrorOptions',
   'data': { 'event': 'BlkdebugEvent',
             '*state': 'int',
+            '*iotype': 'BlkdebugIOType',
             '*errno': 'int',
             '*sector': 'int',
             '*once': 'bool',
block/blkdebug.c +41 -9

122 lines not on this page · raw message

Signature · 1 line
2.19.2
Reviewed-by on v1 3/6
#5 lore raw

Re: [Qemu-devel] [PATCH for-next 3/6] blkdebug: Add @iotype error option

Quoted 18 lines · Max Reitz

This new error option allows users of blkdebug to inject errors only on certain kinds of I/O operations. Users usually want to make a very specific operation fail, not just any; but right now they simply hope that the event that triggers the error injection is followed up with that very operation. That may not be true, however, because the block layer is changing (including blkdebug, which may increase the number of types of I/O operations on which to inject errors).

The new option's default has been chosen to keep backwards compatibility.

Note that similar to the internal representation, we could choose to expose this option as a list of I/O types. But there is no practical use for this, because as described above, users usually know exactly which kind of operation they want to make fail, so there is no need to specify multiple I/O types at once. In addition, exposing this option as a list would require non-trivial changes to qemu_opts_absorb_qdict().

Quoted 3 lines

---
If you want @iotype be a list, I have patches.

Might be interesting to see, whether or not we take them.

Quoted 5 lines
---
 qapi/block-core.json | 26 +++++++++++++++++++++++
 block/blkdebug.c     | 50 ++++++++++++++++++++++++++++++++++++--------
 2 files changed, 67 insertions(+), 9 deletions(-)
Signature · 3 lines
Eric Blake, Principal Software Engineer
Red Hat, Inc.           +1-919-301-3226
Virtualization:  qemu.org | libvirt.org
author
#6 lore raw

Re: [Qemu-devel] [PATCH for-next 3/6] blkdebug: Add @iotype error option

Quoted 19 lines · Eric Blake
Quoted 18 lines · Max Reitz

This new error option allows users of blkdebug to inject errors only on certain kinds of I/O operations. Users usually want to make a very specific operation fail, not just any; but right now they simply hope that the event that triggers the error injection is followed up with that very operation. That may not be true, however, because the block layer is changing (including blkdebug, which may increase the number of types of I/O operations on which to inject errors).

The new option's default has been chosen to keep backwards compatibility.

Note that similar to the internal representation, we could choose to expose this option as a list of I/O types. But there is no practical use for this, because as described above, users usually know exactly which kind of operation they want to make fail, so there is no need to specify multiple I/O types at once. In addition, exposing this option as a list would require non-trivial changes to qemu_opts_absorb_qdict().

Quoted 5 lines
Quoted 3 lines

---
If you want @iotype be a list, I have patches.

Might be interesting to see, whether or not we take them.

Here you go:

https://git.xanclic.moe/XanClic/qemu/commit/f4fcf47cb24b24040d3f0b2306403b0a17b51439
https://git.xanclic.moe/XanClic/qemu/commit/bffaabee7e1995df2adfb6183dec5a2cf6a8e8c8

Quoted 7 lines
Quoted 5 lines
---
 qapi/block-core.json | 26 +++++++++++++++++++++++
 block/blkdebug.c     | 50 ++++++++++++++++++++++++++++++++++++--------
 2 files changed, 67 insertions(+), 9 deletions(-)

Once more, thanks!

Max

author
#7 lore raw

PATCH 4/6 blkdebug: Add "none" event

In reply to Max Reitz

Together with @iotypes and @sector, this can be used to trap e.g. the first read or write access to a certain sector without having to know what happens internally in the block layer, i.e. which "real" events happen right before such an access.

---
 qapi/block-core.json | 4 +++-
 block/blkdebug.c     | 2 ++
 2 files changed, 5 insertions(+), 1 deletion(-)
qapi/block-core.json +3 -1
index 469d889753..045206196b 100644
@@ -3004,6 +3004,8 @@
 #
 # @cor_write: a write due to copy-on-read (since 2.11)
 #
+# @none: triggers once at creation of the blkdebug node (since 4.0)
+#
 # Since: 2.9
 ##
 { 'enum': 'BlkdebugEvent', 'prefix': 'BLKDBG',
@@ -3022,7 +3024,7 @@
             'pwritev_rmw_tail', 'pwritev_rmw_after_tail', 'pwritev',
             'pwritev_zero', 'pwritev_done', 'empty_image_prepare',
             'l1_shrink_write_table', 'l1_shrink_free_l2_clusters',
-            'cor_write'] }
+            'cor_write', 'none' ] }
 
 ##
 # @BlkdebugIOType:
block/blkdebug.c +2 -0
index d15b6c5fef..51f07648ba 100644
@@ -491,6 +491,8 @@ static int blkdebug_open(BlockDriverState *bs, QDict *options, int flags,
         goto out;
     }
 
+    bdrv_debug_event(bs, BLKDBG_NONE);
+
     ret = 0;
 out:
     if (ret < 0) {
Signature · 1 line
2.19.2
Reviewed-by on v1 4/6
#8 lore raw

Re: [Qemu-devel] [PATCH for-next 4/6] blkdebug: Add "none" event

Quoted 11 lines · Max Reitz

Together with @iotypes and @sector, this can be used to trap e.g. the first read or write access to a certain sector without having to know what happens internally in the block layer, i.e. which "real" events happen right before such an access.

---
 qapi/block-core.json | 4 +++-
 block/blkdebug.c     | 2 ++
 2 files changed, 5 insertions(+), 1 deletion(-)
Signature · 3 lines
Eric Blake, Principal Software Engineer
Red Hat, Inc.           +1-919-301-3226
Virtualization:  qemu.org | libvirt.org
author
#9 lore raw

PATCH 5/6 blkdebug: Inject errors on .bdrv_co_block_status()

In reply to Max Reitz
---
 qapi/block-core.json | 5 ++++-
 block/blkdebug.c     | 8 ++++++++
 2 files changed, 12 insertions(+), 1 deletion(-)
qapi/block-core.json +4 -1
index 045206196b..bfe48f71f4 100644
@@ -3041,10 +3041,13 @@
 #
 # @flush: .bdrv_co_flush_to_disk()
 #
+# @block-status: .bdrv_co_block_status()
+#
 # Since: 4.0
 ##
 { 'enum': 'BlkdebugIOType',
-  'data': [ 'read', 'write', 'write-zeroes', 'discard', 'flush' ] }
+  'data': [ 'read', 'write', 'write-zeroes', 'discard', 'flush',
+            'block-status' ] }
 
 ##
 # @BlkdebugInjectErrorOptions:
block/blkdebug.c +8 -0
index 51f07648ba..b86dfe58b5 100644
@@ -670,7 +670,15 @@ static int coroutine_fn blkdebug_co_block_status(BlockDriverState *bs,
                                                  int64_t *map,
                                                  BlockDriverState **file)
 {
+    int err;
+
     assert(QEMU_IS_ALIGNED(offset | bytes, bs->bl.request_alignment));
+
+    err = rule_check(bs, offset, bytes, BLKDEBUGIO_TYPE_BLOCK_STATUS);
+    if (err) {
+        return err;
+    }
+
     return bdrv_co_block_status_from_file(bs, want_zero, offset, bytes,
                                           pnum, map, file);
 }
Signature · 1 line
2.19.2
Reviewed-by on v1 5/6
#10 lore raw

Re: [Qemu-devel] [PATCH for-next 5/6] blkdebug: Inject errors on .bdrv_co_block_status()

Quoted 6 lines · Max Reitz
---
 qapi/block-core.json | 5 ++++-
 block/blkdebug.c     | 8 ++++++++
 2 files changed, 12 insertions(+), 1 deletion(-)
Signature · 3 lines
Eric Blake, Principal Software Engineer
Red Hat, Inc.           +1-919-301-3226
Virtualization:  qemu.org | libvirt.org
author
#11 lore raw

PATCH 6/6 iotests: Test qemu-img convert --salvage

In reply to Max Reitz

This test converts a simple image to another, but blkdebug injects block_status and read faults at some offsets. The resulting image should be the same as the input image, except that sectors that could not be read have to be 0.

---
 tests/qemu-iotests/236     | 164 +++++++++++++++++++++++++++++++++++++
 tests/qemu-iotests/236.out |  43 ++++++++++
 tests/qemu-iotests/group   |   1 +
 3 files changed, 208 insertions(+)
 create mode 100755 tests/qemu-iotests/236
 create mode 100644 tests/qemu-iotests/236.out
tests/qemu-iotests/236 new file +164 -0
new file mode 100755
index 0000000000..02b032ebb5
@@ -0,0 +1,164 @@
+#!/bin/bash
+#
+# Test qemu-img convert --salvage
+#
+# Copyright (C) 2018 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
[email protected]
+
+seq=$(basename $0)
+echo "QA output created by $seq"
+
+here=$PWD
+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
+. ./common.qemu
+
+_supported_fmt generic
+_supported_proto file
+_supported_os Linux
+
+
+TEST_IMG="$TEST_IMG.orig" _make_test_img 64M
+
+$QEMU_IO -c 'write -P 42 0 64M' "$TEST_IMG.orig" | _filter_qemu_io
+
+
+sector_size=512
+
+# Offsets on which to fail block-status.  Keep in ascending order so
+# the indexing done by _filter_offsets will appear in ascending order
+# in the output as well.
+status_fail_offsets="$((16 * 1024 * 1024 + 8192))
+                     $((33 * 1024 * 1024 + 512))"
+
+# Offsets on which to fail reads.  Keep in ascending order for the
+# same reason.
+# Element 1 is shared with $status_fail_offsets on purpose.
+# Elements 2 and later test what happens when a continuous range of
+# sectors is inaccessible.
+read_fail_offsets="$((32 * 1024 * 1024 - 65536))
+                   $((33 * 1024 * 1024 + 512))
+                   $(seq $((34 * 1024 * 1024)) $sector_size \
+                         $((34 * 1024 * 1024 + 4096 - $sector_size)))"
+
+
+# blkdebug must be above the format layer so it can intercept all
+# block-status events
+source_img="json:{'driver': 'blkdebug',
+                  'image': {
+                      'driver': '$IMGFMT',
+                      'file': {
+                          'driver': 'file',
+                          'filename': '$TEST_IMG.orig'
+                      }
+                  },
+                  'inject-error': ["
+
+for ofs in $status_fail_offsets
+do
+    source_img+="{ 'event': 'none',
+                   'iotype': 'block-status',
+                   'errno': 5,
+                   'sector': $((ofs / sector_size)) },"
+done
+
+for ofs in $read_fail_offsets
+do
+    source_img+="{ 'event': 'none',
+                   'iotype': 'read',
+                   'errno': 5,
+                   'sector': $((ofs / sector_size)) },"
+done
+
+# Remove the trailing comma and terminate @inject-error and json:{}
+source_img="${source_img%,} ] }"
+
+
+echo
+
+
+_filter_offsets() {
+    filters=
+
+    index=0
+    for ofs in $2
+    do
+        filters+=" -e s/$(printf "$1" $ofs)/status_fail_offset_$index/"
+        index=$((index + 1))
+    done
+
+    index=0
+    for ofs in $3
+    do
+        filters+=" -e s/$(printf "$1" $ofs)/read_fail_offset_$index/"
+        index=$((index + 1))
+    done
+
+    sed $filters
+}
+
+# While determining the number of allocated sectors in the input
+# image, we should see one block status warning per element of
+# $status_fail_offsets.
+#
+# Then, the image is read.  Since the block status is queried in
+# basically the same way, the same warnings as in the previous step
+# should reappear.  Interleaved with those we should see a read
+# warning per element of $read_fail_offsets.
+# Note that $read_fail_offsets and $status_fail_offsets share an
+# element (read_fail_offset_1 == status_fail_offset_1), so
+# "status_fail_offset_1" in the output is the same as
+# "read_fail_offset_1".
+$QEMU_IMG convert --salvage "$source_img" "$TEST_IMG" 2>&1 \
+    | sed -e 's/json:{.*}/SOURCE_IMG/' \
+    | _filter_offsets '%#x' "$status_fail_offsets" "$read_fail_offsets"
+
+echo
+
+# The offsets where the block status could not be determined should
+# have been treated as containing data and thus should be correct in
+# the output image.
+# The offsets where reading failed altogether should be 0.  Make them
+# 0 in the input image, too, so we can compare both images.
+for ofs in $read_fail_offsets
+do
+    $QEMU_IO -c "write -z $ofs $sector_size" "$TEST_IMG.orig" \
+        | _filter_qemu_io \
+        | _filter_offsets '%i' '' "$read_fail_offsets"
+done
+
+echo
+
+# These should be equal now.
+$QEMU_IMG compare "$TEST_IMG.orig" "$TEST_IMG"
+
+
+# success, all done
+echo "*** done"
+rm -f $seq.full
+status=0
tests/qemu-iotests/236.out new file +43 -0

43 lines not on this page · raw message

tests/qemu-iotests/group +1 -0
index 61a6d98ebd..f6b245917a 100644
@@ -233,3 +233,4 @@
 233 auto quick
 234 auto quick migration
 235 auto quick
+236 auto quick
Signature · 1 line
2.19.2
#12 lore raw
In reply to Max Reitz
Patchew URL: https://patchew.org/QEMU/[email protected]/

Patchew URL: https://patchew.org/QEMU/[email protected]/

Hi,

This series seems to have some coding style problems. See output below for more information:

Subject: [Qemu-devel] [PATCH for-next 0/6] qemu-img: Add salvaging mode to convert
Message-id: [email protected]
Type: series

=== TEST SCRIPT BEGIN ===
#!/bin/bash

BASE=base
n=1
total=$(git log --oneline $BASE.. | wc -l)
failed=0

git config --local diff.renamelimit 0
git config --local diff.renames True
git config --local diff.algorithm histogram

commits="$(git log --format=%H --reverse $BASE..)" for c in $commits; do

    echo "Checking PATCH $n/$total: $(git log -n 1 --format=%s $c)..."
    if ! git show $c --format=email | ./scripts/checkpatch.pl --mailback -; then
        failed=1
        echo
    fi
    n=$((n+1))

done

exit $failed
=== TEST SCRIPT END ===

Updating 3c8cf5a9c21ff8782164d1def7f44bd888713384
fatal: unable to access 'https://github.com/patchew-project/qemu/': Operation timed out after 300008 milliseconds with 0 out of 0 bytes received
error: Could not fetch 3c8cf5a9c21ff8782164d1def7f44bd888713384
Traceback (most recent call last):
File "patchew-tester/src/patchew-cli", line 521, in test_one

    git_clone_repo(clone, r["repo"], r["head"], logf, True)

File "patchew-tester/src/patchew-cli", line 48, in git_clone_repo
stdout=logf, stderr=logf)
File "/usr/lib64/python3.4/subprocess.py", line 558, in check_call
raise CalledProcessError(retcode, cmd)
subprocess.CalledProcessError: Command '['git', 'remote', 'add', '-f', '--mirror=fetch', '3c8cf5a9c21ff8782164d1def7f44bd888713384', 'https://github.com/patchew-project/qemu']' returned non-zero exit status 1

The full log is available at
http://patchew.org/logs/[email protected]/testing.checkpatch/?type=message.
---
Email generated automatically by Patchew [http://patchew.org/].
Please send your feedback to [email protected]

#13 lore raw
Quoted 6 lines · On 12/3/18 9:39 PM, [email protected] wrote:

Patchew URL: https://patchew.org/QEMU/[email protected]/

Hi,

Quoted 7 lines

This series seems to have some coding style problems. See output below for more information:

Updating 3c8cf5a9c21ff8782164d1def7f44bd888713384
fatal: unable to access 'https://github.com/patchew-project/qemu/': Operation timed out after 300008 milliseconds with 0 out of 0 bytes received

False negative. It would be nice to figure out why patchew had a connection problem, but it would also be nice if patchew didn't send out failure notices when the failure is due to a git clone issue unrelated to the patch itself.

Quoted 4 lines

---
Email generated automatically by Patchew [http://patchew.org/].
Please send your feedback to [email protected]

Signature · 3 lines
Eric Blake, Principal Software Engineer
Red Hat, Inc.           +1-919-301-3266
Virtualization:  qemu.org | libvirt.org
author
#14 lore raw
In reply to Max Reitz

Any opinions?

Quoted 57 lines · Max Reitz

Hi,

This series adds a --salvage option to qemu-img convert. With this, qemu-img will not abort when it encounters an I/O error. Instead, it tries to narrow it down and will treat the affected sectors as being completely 0 (and print a warning).

Testing this is not so easy, because while real I/O errors during read operations should be treated as described above, errors encountered during bdrv_block_status() should just be ignored and the affected sectors should be considered allocated. But blkdebug does not yet have a way to intercept this, and:

(1) Just adding a new block-status event would be silly, because I don't want an event, I want it to fail on a certain kind of operation, on a certain sector range, independently of any events, so why can't we just do that? See patch 4.

(2) If we just make blkdebug intercept .bdrv_co_block_status() like all other kinds of operations, at least iotest 041 fails, which does exactly that silly thing: It uses the read_aio event to wait for any read. But it turns out that there may be a bdrv_*block_status() call in between, so suddenly the wrong operation yields an error. As I said, the real fault here is that it does not really make sense to pray that the operation you want to fail is the one that is immediately executed after some event that you hope will trigger that operation.
See patch 3.

So patch 3 allows blkdebug users to select which kind of I/O operation they actually want to make fail, and patch 4 allows them to not use any event, but to have a rule active all the time.

Together, we can then enable error injection for block-status in patch 5 and make use of event=none iotype=block-status in patch 6.

Max Reitz (6):
qemu-img: Move quiet into ImgConvertState
qemu-img: Add salvaging mode to convert
blkdebug: Add @iotype error option
blkdebug: Add "none" event
blkdebug: Inject errors on .bdrv_co_block_status()
iotests: Test qemu-img convert --salvage

 qapi/block-core.json       |  33 +++++++-
 block/blkdebug.c           |  60 ++++++++++++--
 qemu-img.c                 |  97 ++++++++++++++++------
 qemu-img-cmds.hx           |   4 +-
 qemu-img.texi              |   5 ++
 tests/qemu-iotests/236     | 164 +++++++++++++++++++++++++++++++++++++
 tests/qemu-iotests/236.out |  43 ++++++++++
 tests/qemu-iotests/group   |   1 +
 8 files changed, 370 insertions(+), 37 deletions(-)
 create mode 100755 tests/qemu-iotests/236
 create mode 100644 tests/qemu-iotests/236.out