block: Fix resize (extending) of short overlays
v4 cover block: Fix resize (extending) of short overlays
v4:
- Rewrote the series to move the actual zeroing to the block drivers so that error paths can work correctly and potentially long-running fallback to writing explicit zeroes is avoided.
- Fixed output filtering order in the test case [Max]
v3:
- Don't allow blocking the monitor for a zero write in block_resize (even though we can already blockfor other reasons there). This is mainly responsible for the increased complexity compared to v2. Personally, I think this is not an improvement over v2, but if this is what it takes to fix a corruption issue in 4.2... [Max]
- Don't use huge image files in the test case [Vladimir]
v2:
- Switched order of bs->total_sectors update and zero write [Vladimir]
- Fixed coding style [Vladimir]
- Changed the commit message to contain what was in the cover letter
- Test all preallocation modes
- Test allocation status with qemu-io 'map' [Vladimir]
Kevin Wolf (9):
block: Add flags to BlockDriver.bdrv_co_truncate()
block: Add flags to bdrv(_co)_truncate()
block-backend: Add flags to blk_truncate()
qcow2: Support BDRV_REQ_ZERO_WRITE for truncate
raw-format: Support BDRV_REQ_ZERO_WRITE for truncate
file-posix: Support BDRV_REQ_ZERO_WRITE for truncate
block: truncate: Don't make backing file data visible
iotests: Filter testfiles out in img_info_log()
iotests: Test committing to short backing file
include/block/block.h | 5 +- include/block/block_int.h | 10 +- include/sysemu/block-backend.h | 2 +- block.c | 3 +- block/block-backend.c | 4 +- block/commit.c | 4 +- block/crypto.c | 7 +- block/file-posix.c | 6 +- block/gluster.c | 1 + block/io.c | 32 +++++- block/iscsi.c | 2 +- block/mirror.c | 2 +- block/nfs.c | 3 +- block/parallels.c | 6 +- block/qcow.c | 4 +- block/qcow2-refcount.c | 2 +- block/qcow2.c | 28 +++-- block/qed.c | 3 +- block/raw-format.c | 5 +- block/rbd.c | 1 + block/sheepdog.c | 4 +- block/ssh.c | 2 +- block/vdi.c | 2 +- block/vhdx-log.c | 2 +- block/vhdx.c | 6 +- block/vmdk.c | 8 +- block/vpc.c | 2 +- blockdev.c | 2 +- qemu-img.c | 2 +- qemu-io-cmds.c | 2 +- tests/test-block-iothread.c | 9 +- tests/qemu-iotests/iotests.py | 5 +- tests/qemu-iotests/206.out | 12 +- tests/qemu-iotests/242.out | 12 +- tests/qemu-iotests/274 | 152 +++++++++++++++++++++++++ tests/qemu-iotests/274.out | 202 +++++++++++++++++++++++++++++++++ tests/qemu-iotests/group | 1 + 37 files changed, 483 insertions(+), 72 deletions(-) create mode 100755 tests/qemu-iotests/274 create mode 100644 tests/qemu-iotests/274.out
Signature · 1 line
2.20.1
PATCH v4 1/9 block: Add flags to BlockDriver.bdrv_co_truncate()
This adds a new BdrvRequestFlags parameter to the .bdrv_co_truncate() driver callbacks, and a supported_truncate_flags field in BlockDriverState that allows drivers to advertise support for request flags in the context of truncate.
For now, we always pass 0 and no drivers declare support for any flag.
--- include/block/block_int.h | 10 +++++++++- block/crypto.c | 3 ++- block/file-posix.c | 2 +- block/gluster.c | 1 + block/io.c | 8 +++++++- block/iscsi.c | 2 +- block/nfs.c | 3 ++- block/qcow2.c | 2 +- block/qed.c | 1 + block/raw-format.c | 2 +- block/rbd.c | 1 + block/sheepdog.c | 4 ++-- block/ssh.c | 2 +- tests/test-block-iothread.c | 3 ++- 14 files changed, 32 insertions(+), 12 deletions(-)
include/block/block_int.h +9 -1
@@ -355,7 +355,7 @@ struct BlockDriver {*/ int coroutine_fn (*bdrv_co_truncate)(BlockDriverState *bs, int64_t offset,bool exact,PreallocMode prealloc, -Error **errp); +BdrvRequestFlags flags,Error **errp);int64_t (*bdrv_getlength)(BlockDriverState *bs);bool has_variable_length;
@@ -847,6 +847,14 @@ struct BlockDriverState {/* Flags honored during pwrite_zeroes (so far: BDRV_REQ_FUA, * BDRV_REQ_MAY_UNMAP, BDRV_REQ_WRITE_UNCHANGED) */ unsigned int supported_zero_flags; +/* +* Flags honoured during truncate (so far: BDRV_REQ_ZERO_WRITE). +* +* If BDRV_REQ_ZERO_WRITE is given, the truncate operation must make sure +* that any added space reads as all zeros. If this can't be guaranteed, +* the operation must fail. +*/ +unsigned int supported_truncate_flags;/* the following member gives a name to every node on the bs graph. */ char node_name[32 ];
block/crypto.c +2 -1
@@ -299,7 +299,8 @@ static int block_crypto_co_create_generic(BlockDriverState *bs, static int coroutine_fnblock_crypto_co_truncate (BlockDriverState *bs,int64_t offset,bool exact, -PreallocMode prealloc,Error **errp) +PreallocMode prealloc,BdrvRequestFlags flags, +Error **errp) {BlockCrypto *crypto = bs->opaque; uint64_t payload_offset =
block/file-posix.c +1 -1
@@ -2080,7 +2080,7 @@ raw_regular_truncate(BlockDriverState *bs, int fd, int64_t offset, static int coroutine_fn raw_co_truncate(BlockDriverState *bs, int64_t offset,bool exact,PreallocMode prealloc, -Error **errp) +BdrvRequestFlags flags,Error **errp) {BDRVRawState *s = bs->opaque;struct stat st;
block/gluster.c +1 -0
@@ -1228,6 +1228,7 @@ static coroutine_fn int qemu_gluster_co_truncate(BlockDriverState *bs,int64_t offset, bool exact, PreallocMode prealloc, + BdrvRequestFlags flags, Error **errp) { BDRVGlusterState *s = bs->opaque;
block/io.c +7 -1
@@ -3344,6 +3344,7 @@ int coroutine_fn bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact,BlockDriverState *bs = child->bs;BlockDriver *drv = bs->drv;BdrvTrackedRequest req; +BdrvRequestFlags flags =0 ;int64_t old_size, new_bytes;int ret;
@@ -3394,7 +3395,12 @@ int coroutine_fn bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact, }if (drv->bdrv_co_truncate) { - ret = drv->bdrv_co_truncate (bs, offset, exact, prealloc, errp); +if (flags & ~bs->supported_truncate_flags) { +error_setg (errp,"Block driver does not support requested flags" ); + ret = -ENOTSUP ; + goto out; + } + ret = drv->bdrv_co_truncate (bs, offset, exact, prealloc, flags, errp); }else if (bs->file && drv->is_filter) { ret =bdrv_co_truncate (bs->file, offset, exact, prealloc, errp); }else {
block/iscsi.c +1 -1
@@ -2125,7 +2125,7 @@ static void iscsi_reopen_commit(BDRVReopenState *reopen_state) static int coroutine_fn iscsi_co_truncate(BlockDriverState *bs, int64_t offset,bool exact,PreallocMode prealloc, -Error **errp) +BdrvRequestFlags flags,Error **errp) {IscsiLun *iscsilun = bs->opaque;int64_t cur_length;
block/nfs.c +2 -1
@@ -755,7 +755,8 @@ static int64_t nfs_get_allocated_file_size(BlockDriverState *bs) static int coroutine_fnnfs_file_co_truncate (BlockDriverState *bs,int64_t offset,bool exact, -PreallocMode prealloc,Error **errp) +PreallocMode prealloc,BdrvRequestFlags flags, +Error **errp) {NFSClient *client = bs->opaque;int ret;
block/qcow2.c +1 -1
@@ -3964,7 +3964,7 @@ fail: static int coroutine_fn qcow2_co_truncate(BlockDriverState *bs, int64_t offset,bool exact,PreallocMode prealloc, -Error **errp) +BdrvRequestFlags flags,Error **errp) {BDRVQcow2State *s = bs->opaque;uint64_t old_length;
block/qed.c +1 -0
@@ -1467,6 +1467,7 @@ static int coroutine_fn bdrv_qed_co_truncate(BlockDriverState *bs,int64_t offset, bool exact, PreallocMode prealloc, + BdrvRequestFlags flags, Error **errp) { BDRVQEDState *s = bs->opaque;
block/raw-format.c +1 -1
@@ -371,7 +371,7 @@ static void raw_refresh_limits(BlockDriverState *bs, Error **errp) static int coroutine_fn raw_co_truncate(BlockDriverState *bs, int64_t offset,bool exact,PreallocMode prealloc, -Error **errp) +BdrvRequestFlags flags,Error **errp) {BDRVRawState *s = bs->opaque;
block/rbd.c +1 -0
@@ -1108,6 +1108,7 @@ static int coroutine_fn qemu_rbd_co_truncate(BlockDriverState *bs,int64_t offset, bool exact, PreallocMode prealloc, + BdrvRequestFlags flags, Error **errp) { int r;
block/sheepdog.c +2 -2
@@ -2288,7 +2288,7 @@ static int64_t sd_getlength(BlockDriverState *bs) static int coroutine_fn sd_co_truncate(BlockDriverState *bs, int64_t offset,bool exact,PreallocMode prealloc, -Error **errp) +BdrvRequestFlags flags,Error **errp) {BDRVSheepdogState *s = bs->opaque;int ret, fd;
@@ -2604,7 +2604,7 @@ static coroutine_fn int sd_co_writev(BlockDriverState *bs, int64_t sector_num,assert (!flags);if (offset > s->inode.vdi_size) { - ret =sd_co_truncate (bs, offset, false,PREALLOC_MODE_OFF ,NULL ); + ret =sd_co_truncate (bs, offset, false,PREALLOC_MODE_OFF ,0 ,NULL );if (ret <0 ) {return ret; }
block/ssh.c +1 -1
@@ -1298,7 +1298,7 @@ static int64_t ssh_getlength(BlockDriverState *bs) static int coroutine_fn ssh_co_truncate(BlockDriverState *bs, int64_t offset,bool exact,PreallocMode prealloc, -Error **errp) +BdrvRequestFlags flags,Error **errp) {BDRVSSHState *s = bs->opaque;
tests/test-block-iothread.c +2 -1
@@ -46,7 +46,8 @@ static int coroutine_fn bdrv_test_co_pdiscard(BlockDriverState *bs, static int coroutine_fnbdrv_test_co_truncate (BlockDriverState *bs,int64_t offset,bool exact, -PreallocMode prealloc,Error **errp) +PreallocMode prealloc,BdrvRequestFlags flags, +Error **errp) {return 0 ; }
Signature · 1 line
2.20.1
Re: [PATCH v4 1/9] block: Add flags to BlockDriver.bdrv_co_truncate()
Quoted 8 lines · Kevin Wolf
This adds a new BdrvRequestFlags parameter to the .bdrv_co_truncate() driver callbacks, and a
supported_truncate_flagsfield in BlockDriverState that allows drivers to advertise support for request flags in the context of truncate.For now, we always pass 0 and no drivers declare support for any flag.
Signed-off-by Kevin Wolf <[email protected]>
Berto
Re: [PATCH v4 1/9] block: Add flags to BlockDriver.bdrv_co_truncate()
In reply to Kevin WolfQuoted 9 lines · Kevin Wolf
This adds a new BdrvRequestFlags parameter to the .bdrv_co_truncate() driver callbacks, and a
supported_truncate_flagsfield in BlockDriverState that allows drivers to advertise support for request flags in the context of truncate.For now, we always pass 0 and no drivers declare support for any flag.
Signed-off-by Kevin Wolf<[email protected]>---
[..]
Quoted 13 lines
block/io.c +1
-0@@ -3344,6 +3344,7 @@ int coroutine_fn bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact,BlockDriverState *bs = child->bs;BlockDriver *drv = bs->drv;BdrvTrackedRequest req; +BdrvRequestFlags flags =0 ;int64_t old_size, new_bytes;int ret;@@ -3394,7 +3395,12 @@ int coroutine_fn bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact, }
block/io.c int coroutine_fn bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact, · line 3405
Quoted 10 lines
block/io.c +6
-1if (drv->bdrv_co_truncate) { - ret = drv->bdrv_co_truncate (bs, offset, exact, prealloc, errp); +if (flags & ~bs->supported_truncate_flags) { +error_setg (errp,"Block driver does not support requested flags" ); + ret = -ENOTSUP ; + goto out; + } + ret = drv->bdrv_co_truncate (bs, offset, exact, prealloc, flags, errp); }else if (bs->file && drv->is_filter) { ret =bdrv_co_truncate (bs->file, offset, exact, prealloc, errp);
Hmm, I think it's incorrect to silently omit flags here.. But in current patch flags are always 0, and you update this place in the following patch. OK
With updated block/file-win32.c:
Signature · 2 lines
Best regards, Vladimir
PATCH v4 2/9 block: Add flags to bdrv(_co)_truncate()
In reply to Kevin WolfNow that block drivers can support flags for .bdrv_co_truncate, expose the parameter in the node level interfaces bdrv_co_truncate() and bdrv_truncate().
--- include/block/block.h | 5 +++-- block/block-backend.c | 2 +- block/crypto.c | 2 +- block/io.c | 12 +++++++----- block/parallels.c | 6 +++--- block/qcow.c | 4 ++-- block/qcow2-refcount.c | 2 +- block/qcow2.c | 13 +++++++------ block/raw-format.c | 2 +- block/vhdx-log.c | 2 +- block/vhdx.c | 2 +- block/vmdk.c | 2 +- tests/test-block-iothread.c | 6 +++--- 13 files changed, 32 insertions(+), 28 deletions(-)
include/block/block.h +3 -2
@@ -339,9 +339,10 @@ BlockDriverState *bdrv_find_backing_image(BlockDriverState *bs,void bdrv_refresh_filename (BlockDriverState *bs); int coroutine_fn bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact, -PreallocMode prealloc,Error **errp); +PreallocMode prealloc,BdrvRequestFlags flags, +Error **errp);int bdrv_truncate (BdrvChild *child,int64_t offset,bool exact, -PreallocMode prealloc,Error **errp); +PreallocMode prealloc,BdrvRequestFlags flags,Error **errp);int64_t bdrv_nb_sectors (BlockDriverState *bs);int64_t bdrv_getlength (BlockDriverState *bs);
block/block-backend.c +1 -1
@@ -2144,7 +2144,7 @@ int blk_truncate(BlockBackend *blk, int64_t offset, bool exact,return -ENOMEDIUM ; } -return bdrv_truncate (blk->root, offset, exact, prealloc, errp); +return bdrv_truncate (blk->root, offset, exact, prealloc,0 , errp); } int blk_save_vmstate(BlockBackend *blk, const uint8_t *buf,
block/crypto.c +1 -1
@@ -313,7 +313,7 @@ block_crypto_co_truncate(BlockDriverState *bs, int64_t offset, bool exact, offset += payload_offset; -return bdrv_co_truncate (bs->file, offset, exact, prealloc, errp); +return bdrv_co_truncate (bs->file, offset, exact, prealloc,0 , errp); }static void block_crypto_close (BlockDriverState *bs)
block/io.c +7 -5
@@ -3339,12 +3339,12 @@ static void bdrv_parent_cb_resize(BlockDriverState *bs)* 'offset' bytes in length. */ int coroutine_fn bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact, -PreallocMode prealloc,Error **errp) +PreallocMode prealloc,BdrvRequestFlags flags, +Error **errp) {BlockDriverState *bs = child->bs;BlockDriver *drv = bs->drv;BdrvTrackedRequest req; -BdrvRequestFlags flags =0 ;int64_t old_size, new_bytes;int ret;
@@ -3402,7 +3402,7 @@ int coroutine_fn bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact, } ret = drv->bdrv_co_truncate (bs, offset, exact, prealloc, flags, errp); }else if (bs->file && drv->is_filter) { - ret =bdrv_co_truncate (bs->file, offset, exact, prealloc, errp); + ret =bdrv_co_truncate (bs->file, offset, exact, prealloc, flags, errp); }else {error_setg (errp,"Image format driver does not support resize" ); ret = -ENOTSUP ;
@@ -3435,6 +3435,7 @@ typedef struct TruncateCo {int64_t offset;bool exact;PreallocMode prealloc; +BdrvRequestFlags flags;Error **errp;int ret; } TruncateCo;
@@ -3443,12 +3444,12 @@ static void coroutine_fn bdrv_truncate_co_entry(void *opaque) {TruncateCo *tco = opaque; tco->ret =bdrv_co_truncate (tco->child, tco->offset, tco->exact, - tco->prealloc, tco->errp); + tco->prealloc, tco->flags, tco->errp);aio_wait_kick (); }int bdrv_truncate (BdrvChild *child,int64_t offset,bool exact, -PreallocMode prealloc,Error **errp) +PreallocMode prealloc,BdrvRequestFlags flags,Error **errp) {Coroutine *co;TruncateCo tco = {
@@ -3456,6 +3457,7 @@ int bdrv_truncate(BdrvChild *child, int64_t offset, bool exact, .offset = offset, .exact = exact, .prealloc = prealloc, + .flags = flags, .errp = errp, .ret = NOT_DONE, };
block/parallels.c +3 -3
@@ -203,7 +203,7 @@ static int64_t allocate_clusters(BlockDriverState *bs, int64_t sector_num, }else { ret =bdrv_truncate (bs->file, (s->data_end + space) <<BDRV_SECTOR_BITS , - false,PREALLOC_MODE_OFF ,NULL ); + false,PREALLOC_MODE_OFF ,0 ,NULL ); }if (ret <0 ) {return ret;
@@ -493,7 +493,7 @@ static int coroutine_fn parallels_co_check(BlockDriverState *bs,* That means we have to pass exact=true. */ ret =bdrv_truncate (bs->file, res->image_end_offset, true, -PREALLOC_MODE_OFF , &local_err); +PREALLOC_MODE_OFF ,0 , &local_err);if (ret <0 ) {error_report_err (local_err); res->check_errors++;
@@ -889,7 +889,7 @@ static void parallels_close(BlockDriverState *bs)/* errors are ignored, so we might as well pass exact=true */ bdrv_truncate (bs->file, s->data_end <<BDRV_SECTOR_BITS , true, -PREALLOC_MODE_OFF ,NULL ); +PREALLOC_MODE_OFF ,0 ,NULL ); }g_free (s->bat_dirty_bmap);
block/qcow.c +2 -2
@@ -480,7 +480,7 @@ static int get_cluster_offset(BlockDriverState *bs,return -E2BIG ; } ret =bdrv_truncate (bs->file, cluster_offset + s->cluster_size, - false,PREALLOC_MODE_OFF ,NULL ); + false,PREALLOC_MODE_OFF ,0 ,NULL );if (ret <0 ) {return ret; }
@@ -1035,7 +1035,7 @@ static int qcow_make_empty(BlockDriverState *bs) l1_length) < 0)return -1 ; ret =bdrv_truncate (bs->file, s->l1_table_offset + l1_length, false, -PREALLOC_MODE_OFF ,NULL ); +PREALLOC_MODE_OFF ,0 ,NULL );if (ret <0 )return ret;
block/qcow2-refcount.c +1 -1
@@ -2018,7 +2018,7 @@ static int check_refblocks(BlockDriverState *bs, BdrvCheckResult *res, } ret =bdrv_truncate (bs->file, offset + s->cluster_size, false, -PREALLOC_MODE_OFF , &local_err); +PREALLOC_MODE_OFF ,0 , &local_err);if (ret <0 ) {error_report_err (local_err); goto resize_fail;
block/qcow2.c +7 -6
@@ -3095,7 +3095,7 @@ static int coroutine_fn preallocate_co(BlockDriverState *bs, uint64_t offset, mode =PREALLOC_MODE_OFF ; } ret =bdrv_co_truncate (s->data_file, host_offset + cur_bytes, false, - mode, errp); + mode,0 , errp);if (ret <0 ) {return ret; }
@@ -4061,7 +4061,7 @@ static int coroutine_fn qcow2_co_truncate(BlockDriverState *bs, int64_t offset,* always fulfilled, so there is no need to pass it on.) */ bdrv_co_truncate (bs->file, (last_cluster +1 ) * s->cluster_size, - false,PREALLOC_MODE_OFF , &local_err); + false,PREALLOC_MODE_OFF ,0 , &local_err);if (local_err) {warn_reportf_err (local_err,"Failed to truncate the tail of the image: " );
@@ -4083,7 +4083,8 @@ static int coroutine_fn qcow2_co_truncate(BlockDriverState *bs, int64_t offset,* file should be resized to the exact target size, too, * so we pass @exact here. */ - ret =bdrv_co_truncate (s->data_file, offset, exact, prealloc, errp); + ret =bdrv_co_truncate (s->data_file, offset, exact, prealloc,0 , + errp);if (ret <0 ) { goto fail; }
@@ -4169,7 +4170,7 @@ static int coroutine_fn qcow2_co_truncate(BlockDriverState *bs, int64_t offset, new_file_size = allocation_start + nb_new_data_clusters * s->cluster_size;/* Image file grows, so @exact does not matter */ - ret =bdrv_co_truncate (bs->file, new_file_size, false, prealloc, errp); + ret =bdrv_co_truncate (bs->file, new_file_size, false, prealloc,0 , errp);if (ret <0 ) {error_prepend (errp,"Failed to resize underlying file: " ); qcow2_free_clusters(bs, allocation_start,
@@ -4348,7 +4349,7 @@ qcow2_co_pwritev_compressed_part(BlockDriverState *bs,if (len <0 ) {return len; } -return bdrv_co_truncate (bs->file, len, false,PREALLOC_MODE_OFF ,NULL ); +return bdrv_co_truncate (bs->file, len, false,PREALLOC_MODE_OFF ,0 ,NULL ); }if (offset_into_cluster (s, offset)) {
@@ -4563,7 +4564,7 @@ static int make_completely_empty(BlockDriverState *bs) } ret =bdrv_truncate (bs->file, (3 + l1_clusters) * s->cluster_size, false, -PREALLOC_MODE_OFF , &local_err); +PREALLOC_MODE_OFF ,0 , &local_err);if (ret <0 ) {error_report_err (local_err); goto fail;
block/raw-format.c +1 -1
@@ -387,7 +387,7 @@ static int coroutine_fn raw_co_truncate(BlockDriverState *bs, int64_t offset, s->size = offset; offset += s->offset; -return bdrv_co_truncate (bs->file, offset, exact, prealloc, errp); +return bdrv_co_truncate (bs->file, offset, exact, prealloc,0 , errp); }static void raw_eject (BlockDriverState *bs,bool eject_flag)
block/vhdx-log.c +1 -1
@@ -558,7 +558,7 @@ static int vhdx_log_flush(BlockDriverState *bs, BDRVVHDXState *s, goto exit; } ret =bdrv_truncate (bs->file, new_file_size, false, -PREALLOC_MODE_OFF ,NULL ); +PREALLOC_MODE_OFF ,0 ,NULL );if (ret <0 ) { goto exit; }
block/vhdx.c +1 -1
@@ -1264,7 +1264,7 @@ static int vhdx_allocate_block(BlockDriverState *bs, BDRVVHDXState *s, }return bdrv_truncate (bs->file, *new_offset + s->block_size, false, -PREALLOC_MODE_OFF ,NULL ); +PREALLOC_MODE_OFF ,0 ,NULL ); }/*
block/vmdk.c +1 -1
@@ -2077,7 +2077,7 @@ vmdk_co_pwritev_compressed(BlockDriverState *bs, uint64_t offset, } length =QEMU_ALIGN_UP (length,BDRV_SECTOR_SIZE ); ret =bdrv_truncate (s->extents[i].file, length, false, -PREALLOC_MODE_OFF ,NULL ); +PREALLOC_MODE_OFF ,0 ,NULL );if (ret <0 ) {return ret; }
tests/test-block-iothread.c +3 -3
@@ -186,18 +186,18 @@ static void test_sync_op_truncate(BdrvChild *c)int ret;/* Normal success path */ - ret =bdrv_truncate (c,65536 , false,PREALLOC_MODE_OFF ,NULL ); + ret =bdrv_truncate (c,65536 , false,PREALLOC_MODE_OFF ,0 ,NULL ); g_assert_cmpint(ret, ==, 0);/* Early error: Negative offset */ - ret =bdrv_truncate (c,-2 , false,PREALLOC_MODE_OFF ,NULL ); + ret =bdrv_truncate (c,-2 , false,PREALLOC_MODE_OFF ,0 ,NULL ); g_assert_cmpint(ret, ==, -EINVAL);/* Error: Read-only image */ c->bs->read_only = true; c->bs->open_flags &= ~BDRV_O_RDWR ; - ret =bdrv_truncate (c,65536 , false,PREALLOC_MODE_OFF ,NULL ); + ret =bdrv_truncate (c,65536 , false,PREALLOC_MODE_OFF ,0 ,NULL ); g_assert_cmpint(ret, ==, -EACCES); c->bs->read_only = false;
Signature · 1 line
2.20.1
Re: [PATCH v4 2/9] block: Add flags to bdrv(_co)_truncate()
Quoted 5 lines · Kevin Wolf
Now that block drivers can support flags for .bdrv_co_truncate, expose the parameter in the node level interfaces
bdrv_co_truncate()andbdrv_truncate().Signed-off-by Kevin Wolf <[email protected]>
Berto
Re: [PATCH v4 2/9] block: Add flags to bdrv(_co)_truncate()
In reply to Kevin WolfQuoted 5 lines · Kevin Wolf
Now that block drivers can support flags for .bdrv_co_truncate, expose the parameter in the node level interfaces
bdrv_co_truncate()andbdrv_truncate().Signed-off-by Kevin Wolf <[email protected]>
Quoted 1 line
---
[..]
block/qcow2.c static int coroutine_fn qcow2_co_truncate(BlockDriverState *bs, int64_t offset, · line 4173
Quoted 6 lines
block/qcow2.c +1
-1@@ -4169,7 +4170,7 @@ static int coroutine_fn qcow2_co_truncate(BlockDriverState *bs, int64_t offset, new_file_size = allocation_start + nb_new_data_clusters * s->cluster_size;/* Image file grows, so @exact does not matter */ - ret =bdrv_co_truncate (bs->file, new_file_size, false, prealloc, errp); + ret =bdrv_co_truncate (bs->file, new_file_size, false, prealloc,0 , errp);
over-80 line
block/qcow2.c qcow2_co_pwritev_compressed_part(BlockDriverState *bs, · line 4352
Quoted 9 lines
block/qcow2.c +1
-1if (ret <0 ) {error_prepend (errp,"Failed to resize underlying file: " ); qcow2_free_clusters(bs, allocation_start,@@ -4348,7 +4349,7 @@ qcow2_co_pwritev_compressed_part(BlockDriverState *bs,if (len <0 ) {return len; } -return bdrv_co_truncate (bs->file, len, false,PREALLOC_MODE_OFF ,NULL ); +return bdrv_co_truncate (bs->file, len, false,PREALLOC_MODE_OFF ,0 ,NULL );
and this one
Signature · 2 lines
Best regards, Vladimir
PATCH v4 3/9 block-backend: Add flags to blk_truncate()
In reply to Kevin WolfNow that node level interface bdrv_truncate() supports passing request flags to the block driver, expose this on the BlockBackend level, too.
--- include/sysemu/block-backend.h | 2 +- block.c | 3 ++- block/block-backend.c | 4 ++-- block/commit.c | 4 ++-- block/crypto.c | 2 +- block/mirror.c | 2 +- block/qcow2.c | 4 ++-- block/qed.c | 2 +- block/vdi.c | 2 +- block/vhdx.c | 4 ++-- block/vmdk.c | 6 +++--- block/vpc.c | 2 +- blockdev.c | 2 +- qemu-img.c | 2 +- qemu-io-cmds.c | 2 +- 15 files changed, 22 insertions(+), 21 deletions(-)
include/sysemu/block-backend.h +1 -1
@@ -237,7 +237,7 @@ int coroutine_fn blk_co_pwrite_zeroes(BlockBackend *blk, int64_t offset,int blk_pwrite_compressed (BlockBackend *blk,int64_t offset,const void *buf,int bytes);int blk_truncate (BlockBackend *blk,int64_t offset,bool exact, -PreallocMode prealloc,Error **errp); +PreallocMode prealloc,BdrvRequestFlags flags,Error **errp);int blk_pdiscard (BlockBackend *blk,int64_t offset,int bytes);int blk_save_vmstate (BlockBackend *blk,const uint8_t *buf,int64_t pos,int size);
block.c +2 -1
@@ -548,7 +548,8 @@ static int64_t create_file_fallback_truncate(BlockBackend *blk,int64_t size;int ret; - ret =blk_truncate (blk, minimum_size, false,PREALLOC_MODE_OFF , &local_err); + ret =blk_truncate (blk, minimum_size, false,PREALLOC_MODE_OFF ,0 , + &local_err);if (ret <0 && ret != -ENOTSUP ) {error_propagate (errp, local_err);return ret;
block/block-backend.c +2 -2
@@ -2137,14 +2137,14 @@ int blk_pwrite_compressed(BlockBackend *blk, int64_t offset, const void *buf, }int blk_truncate (BlockBackend *blk,int64_t offset,bool exact, -PreallocMode prealloc,Error **errp) +PreallocMode prealloc,BdrvRequestFlags flags,Error **errp) {if (!blk_is_available (blk)) {error_setg (errp,"No medium inserted" );return -ENOMEDIUM ; } -return bdrv_truncate (blk->root, offset, exact, prealloc,0 , errp); +return bdrv_truncate (blk->root, offset, exact, prealloc, flags, errp); } int blk_save_vmstate(BlockBackend *blk, const uint8_t *buf,
block/commit.c +2 -2
@@ -133,7 +133,7 @@ static int coroutine_fn commit_run(Job *job, Error **errp) }if (base_len < len) { - ret =blk_truncate (s->base, len, false,PREALLOC_MODE_OFF ,NULL ); + ret =blk_truncate (s->base, len, false,PREALLOC_MODE_OFF ,0 ,NULL );if (ret) { goto out; }
@@ -458,7 +458,7 @@ int bdrv_commit(BlockDriverState *bs)* grow the backing file image if possible. If not possible, * we must return an error */ if (length > backing_length) { - ret =blk_truncate (backing, length, false,PREALLOC_MODE_OFF , + ret =blk_truncate (backing, length, false,PREALLOC_MODE_OFF ,0 , &local_err);if (ret <0 ) {error_report_err (local_err);
block/crypto.c +1 -1
@@ -115,7 +115,7 @@ static ssize_t block_crypto_init_func(QCryptoBlock *block,* which will be used by the crypto header */ return blk_truncate (data->blk, data->size + headerlen, false, - data->prealloc, errp); + data->prealloc,0 , errp); }
block/mirror.c +1 -1
@@ -900,7 +900,7 @@ static int coroutine_fn mirror_run(Job *job, Error **errp)if (s->bdev_length > base_length) { ret =blk_truncate (s->target, s->bdev_length, false, -PREALLOC_MODE_OFF ,NULL ); +PREALLOC_MODE_OFF ,0 ,NULL );if (ret <0 ) { goto immediate_exit; }
block/qcow2.c +2 -2
@@ -3511,7 +3511,7 @@ qcow2_co_create(BlockdevCreateOptions *create_options, Error **errp)/* Okay, now that we have a valid image, let's give it the right size */ ret =blk_truncate (blk, qcow2_opts->size, false, qcow2_opts->preallocation, - errp); +0 , errp);if (ret <0 ) {error_prepend (errp,"Could not resize image: " ); goto out;
@@ -5372,7 +5372,7 @@ static int qcow2_amend_options(BlockDriverState *bs, QemuOpts *opts,* Amending image options should ensure that the image has * exactly the given new values, so pass exact=true here. */ - ret =blk_truncate (blk, new_size, true,PREALLOC_MODE_OFF , errp); + ret =blk_truncate (blk, new_size, true,PREALLOC_MODE_OFF ,0 , errp);blk_unref (blk);if (ret <0 ) {return ret;
block/qed.c +1 -1
@@ -677,7 +677,7 @@ static int coroutine_fn bdrv_qed_co_create(BlockdevCreateOptions *opts,* The QED format associates file length with allocation status, * so a new file (which is empty) must have a length of 0. */ - ret =blk_truncate (blk,0 , true,PREALLOC_MODE_OFF , errp); + ret =blk_truncate (blk,0 , true,PREALLOC_MODE_OFF ,0 , errp);if (ret <0 ) { goto out; }
block/vdi.c +1 -1
@@ -875,7 +875,7 @@ static int coroutine_fn vdi_co_do_create(BlockdevCreateOptions *create_options,if (image_type ==VDI_TYPE_STATIC ) { ret =blk_truncate (blk, offset + blocks * block_size, false, -PREALLOC_MODE_OFF , errp); +PREALLOC_MODE_OFF ,0 , errp);if (ret <0 ) {error_prepend (errp,"Failed to statically allocate file" ); goto exit;
block/vhdx.c +2 -2
@@ -1703,13 +1703,13 @@ static int vhdx_create_bat(BlockBackend *blk, BDRVVHDXState *s,/* All zeroes, so we can just extend the file - the end of the BAT * is the furthest thing we have written yet */ ret =blk_truncate (blk, data_file_offset, false,PREALLOC_MODE_OFF , - errp); +0 , errp);if (ret <0 ) { goto exit; } }else if (type ==VHDX_TYPE_FIXED ) { ret =blk_truncate (blk, data_file_offset + image_size, false, -PREALLOC_MODE_OFF , errp); +PREALLOC_MODE_OFF ,0 , errp);if (ret <0 ) { goto exit; }
block/vmdk.c +3 -3
@@ -2118,7 +2118,7 @@ static int vmdk_init_extent(BlockBackend *blk,int gd_buf_size;if (flat) { - ret =blk_truncate (blk, filesize, false,PREALLOC_MODE_OFF , errp); + ret =blk_truncate (blk, filesize, false,PREALLOC_MODE_OFF ,0 , errp); goto exit; } magic =cpu_to_be32 (VMDK4_MAGIC );
@@ -2182,7 +2182,7 @@ static int vmdk_init_extent(BlockBackend *blk, } ret =blk_truncate (blk,le64_to_cpu (header.grain_offset) <<9 , false, -PREALLOC_MODE_OFF , errp); +PREALLOC_MODE_OFF ,0 , errp);if (ret <0 ) { goto exit; }
@@ -2523,7 +2523,7 @@ static int coroutine_fn vmdk_co_do_create(int64_t size,/* bdrv_pwrite write padding zeros to align to sector, we don't need that * for description file */ if (desc_offset ==0 ) { - ret =blk_truncate (blk, desc_len, false,PREALLOC_MODE_OFF , errp); + ret =blk_truncate (blk, desc_len, false,PREALLOC_MODE_OFF ,0 , errp);if (ret <0 ) { goto exit; }
block/vpc.c +1 -1
@@ -898,7 +898,7 @@ static int create_fixed_disk(BlockBackend *blk, uint8_t *buf,/* Add footer to total size */ total_size +=HEADER_SIZE ; - ret =blk_truncate (blk, total_size, false,PREALLOC_MODE_OFF , errp); + ret =blk_truncate (blk, total_size, false,PREALLOC_MODE_OFF ,0 , errp);if (ret <0 ) {return ret; }
blockdev.c +1 -1
@@ -2741,7 +2741,7 @@ void qmp_block_resize(bool has_device, const char *device, }bdrv_drained_begin (bs); - ret =blk_truncate (blk, size, false,PREALLOC_MODE_OFF , errp); + ret =blk_truncate (blk, size, false,PREALLOC_MODE_OFF ,0 , errp);bdrv_drained_end (bs); out:
qemu-img.c +1 -1
@@ -3866,7 +3866,7 @@ static int img_resize(int argc, char **argv)* resizing, so pass @exact=true. It is of no use to report * success when the image has not actually been resized. */ - ret =blk_truncate (blk, total_size, true, prealloc, &err); + ret =blk_truncate (blk, total_size, true, prealloc,0 , &err);if (!ret) {qprintf (quiet,"Image resized.\n" ); }else {
qemu-io-cmds.c +1 -1
@@ -1715,7 +1715,7 @@ static int truncate_f(BlockBackend *blk, int argc, char **argv)* exact=true. It is better to err on the "emit more errors" side * than to be overly permissive. */ - ret =blk_truncate (blk, offset, true,PREALLOC_MODE_OFF , &local_err); + ret =blk_truncate (blk, offset, true,PREALLOC_MODE_OFF ,0 , &local_err);if (ret <0 ) {error_report_err (local_err);return ret;
Signature · 1 line
2.20.1
Re: [PATCH v4 3/9] block-backend: Add flags to blk_truncate()
Quoted 4 lines · Kevin Wolf
Now that node level interface
bdrv_truncate()supports passing request flags to the block driver, expose this on the BlockBackend level, too.Signed-off-by Kevin Wolf <[email protected]>
Berto
Re: [PATCH v4 3/9] block-backend: Add flags to blk_truncate()
In reply to Kevin WolfQuoted 4 lines · Kevin Wolf
Now that node level interface
bdrv_truncate()supports passing request flags to the block driver, expose this on the BlockBackend level, too.Signed-off-by Kevin Wolf<[email protected]>
Signature · 2 lines
Best regards, Vladimir
PATCH v4 4/9 qcow2: Support BDRV_REQ_ZERO_WRITE for truncate
In reply to Kevin WolfIf BDRV_REQ_ZERO_WRITE is set and we're extending the image, calling qcow2_cluster_zeroize() with flags=0 does the right thing: It doesn't undo any previous preallocation, but just adds the zero flag to all relevant L2 entries. If an external data file is in use, a write_zeroes request to the data file is made instead.
--- block/qcow2.c | 9 +++++++++ 1 file changed, 9 insertions(+)
block/qcow2.c +9 -0
@@ -1726,6 +1726,7 @@ static int coroutine_fn qcow2_do_open(BlockDriverState *bs, QDict *options, bs->supported_zero_flags = header.version >=3 ?BDRV_REQ_MAY_UNMAP |BDRV_REQ_NO_FALLBACK :0 ; + bs->supported_truncate_flags =BDRV_REQ_ZERO_WRITE ;/* Repair image if dirty */ if (!(flags & (BDRV_O_CHECK | BDRV_O_INACTIVE)) && !bs->read_only &&
@@ -4213,6 +4214,14 @@ static int coroutine_fn qcow2_co_truncate(BlockDriverState *bs, int64_t offset,g_assert_not_reached (); } +if ((flags &BDRV_REQ_ZERO_WRITE ) && offset > old_length) { + ret =qcow2_cluster_zeroize (bs, old_length, offset - old_length,0 ); +if (ret <0 ) { +error_setg_errno (errp, -ret,"Failed to zero out the new area" ); + goto fail; + } + } +if (prealloc !=PREALLOC_MODE_OFF ) {/* Flush metadata before actually changing the image size */ ret =qcow2_write_caches (bs);
Signature · 1 line
2.20.1
Re: [PATCH v4 4/9] qcow2: Support BDRV_REQ_ZERO_WRITE for truncate
Quoted 7 lines · Kevin Wolf
If
BDRV_REQ_ZERO_WRITEis set and we're extending the image, callingqcow2_cluster_zeroize()with flags=0 does the right thing: It doesn't undo any previous preallocation, but just adds the zero flag to all relevant L2 entries. If an external data file is in use, awrite_zeroesrequest to the data file is made instead.Signed-off-by Kevin Wolf <[email protected]>
Berto
Re: [PATCH v4 4/9] qcow2: Support BDRV_REQ_ZERO_WRITE for truncate
In reply to Kevin WolfQuoted 24 lines · Kevin Wolf
If
BDRV_REQ_ZERO_WRITEis set and we're extending the image, callingqcow2_cluster_zeroize()with flags=0 does the right thing: It doesn't undo any previous preallocation, but just adds the zero flag to all relevant L2 entries. If an external data file is in use, awrite_zeroesrequest to the data file is made instead.Signed-off-by Kevin Wolf <[email protected]>--- block/qcow2.c | 9 +++++++++ 1 file changed, 9 insertions(+)block/qcow2.c +1
-0@@ -1726,6 +1726,7 @@ static int coroutine_fn qcow2_do_open(BlockDriverState *bs, QDict *options, bs->supported_zero_flags = header.version >=3 ?BDRV_REQ_MAY_UNMAP |BDRV_REQ_NO_FALLBACK :0 ; + bs->supported_truncate_flags =BDRV_REQ_ZERO_WRITE ;/* Repair image if dirty */ if (!(flags & (BDRV_O_CHECK | BDRV_O_INACTIVE)) && !bs->read_only &&
block/qcow2.c static int coroutine_fn qcow2_co_truncate(BlockDriverState *bs, int64_t offset, · line 4218
Quoted 5 lines
block/qcow2.c +2
-0@@ -4213,6 +4214,14 @@ static int coroutine_fn qcow2_co_truncate(BlockDriverState *bs, int64_t offset,g_assert_not_reached (); } +if ((flags &BDRV_REQ_ZERO_WRITE ) && offset > old_length) { + ret =qcow2_cluster_zeroize (bs, old_length, offset - old_length,0 );
Hmm. As I understand, qcow2_cluster_zeroize is unprepared to cluster-unaligned offset/size. I think we should handle it somehow.
Quoted 10 lines
block/qcow2.c +6
-0+if (ret <0 ) { +error_setg_errno (errp, -ret,"Failed to zero out the new area" ); + goto fail; + } + } +if (prealloc !=PREALLOC_MODE_OFF ) {/* Flush metadata before actually changing the image size */ ret =qcow2_write_caches (bs);
Signature · 2 lines
Best regards, Vladimir
Re: [PATCH v4 4/9] qcow2: Support BDRV_REQ_ZERO_WRITE for truncate
Quoted 5 lines · Vladimir Sementsov-Ogievskiy
block/qcow2.c static int coroutine_fn qcow2_co_truncate(BlockDriverState *bs, int64_t offset, · line 4218
Quoted 2 lines
block/qcow2.c +2
-0+if ((flags &BDRV_REQ_ZERO_WRITE ) && offset > old_length) { + ret =qcow2_cluster_zeroize (bs, old_length, offset - old_length,0 );Hmm. As I understand,
qcow2_cluster_zeroizeis unprepared to cluster-unaligned offset/size. I think we should handle it somehow.
You're right, it actually hits an assertion :-/
I suppose you can simply round the size up to the next cluster.
Berto
Re: [PATCH v4 4/9] qcow2: Support BDRV_REQ_ZERO_WRITE for truncate
Quoted 11 lines · Alberto Garcia
Quoted 5 lines · Vladimir Sementsov-Ogievskiy
block/qcow2.c static int coroutine_fn qcow2_co_truncate(BlockDriverState *bs, int64_t offset, · line 4218
Quoted 2 lines
block/qcow2.c +2
-0+if ((flags &BDRV_REQ_ZERO_WRITE ) && offset > old_length) { + ret =qcow2_cluster_zeroize (bs, old_length, offset - old_length,0 );Hmm. As I understand,
qcow2_cluster_zeroizeis unprepared to cluster-unaligned offset/size. I think we should handle it somehow.You're right, it actually hits an assertion :-/
I suppose you can simply round the size up to the next cluster.
But we also need to handle unaligned old_length, we can't just align it down, we should write zero to existing cluster if it is NORMAL cluster..
Signature · 2 lines
Best regards, Vladimir
PATCH v4 5/9 raw-format: Support BDRV_REQ_ZERO_WRITE for truncate
In reply to Kevin WolfThe raw format driver can simply forward the flag and let its bs->file child take care of actually providing the zeros.
--- block/raw-format.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
block/raw-format.c +2 -1
@@ -387,7 +387,7 @@ static int coroutine_fn raw_co_truncate(BlockDriverState *bs, int64_t offset, s->size = offset; offset += s->offset; -return bdrv_co_truncate (bs->file, offset, exact, prealloc,0 , errp); +return bdrv_co_truncate (bs->file, offset, exact, prealloc, flags, errp); }static void raw_eject (BlockDriverState *bs,bool eject_flag)
@@ -445,6 +445,7 @@ static int raw_open(BlockDriverState *bs, QDict *options, int flags, bs->supported_zero_flags =BDRV_REQ_WRITE_UNCHANGED | ((BDRV_REQ_FUA |BDRV_REQ_MAY_UNMAP |BDRV_REQ_NO_FALLBACK ) & bs->file->bs->supported_zero_flags); + bs->supported_truncate_flags =BDRV_REQ_ZERO_WRITE ;if (bs->probed && !bdrv_is_read_only (bs)) {bdrv_refresh_filename (bs->file->bs);
Signature · 1 line
2.20.1
Re: [PATCH v4 5/9] raw-format: Support BDRV_REQ_ZERO_WRITE for truncate
Quoted 4 lines · Kevin Wolf
The raw format driver can simply forward the flag and let its
bs->filechild take care of actually providing the zeros.Signed-off-by Kevin Wolf <[email protected]>
Berto
Re: [PATCH v4 5/9] raw-format: Support BDRV_REQ_ZERO_WRITE for truncate
In reply to Kevin WolfQuoted 22 lines · Kevin Wolf
The raw format driver can simply forward the flag and let its
bs->filechild take care of actually providing the zeros.Signed-off-by Kevin Wolf <[email protected]>--- block/raw-format.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)block/raw-format.c +1
-1@@ -387,7 +387,7 @@ static int coroutine_fn raw_co_truncate(BlockDriverState *bs, int64_t offset, s->size = offset; offset += s->offset; -return bdrv_co_truncate (bs->file, offset, exact, prealloc,0 , errp); +return bdrv_co_truncate (bs->file, offset, exact, prealloc, flags, errp); }static void raw_eject (BlockDriverState *bs,bool eject_flag)
block/raw-format.c static int raw_open(BlockDriverState *bs, QDict *options, int flags, · line 448
Quoted 4 lines
block/raw-format.c +1
-0@@ -445,6 +445,7 @@ static int raw_open(BlockDriverState *bs, QDict *options, int flags, bs->supported_zero_flags =BDRV_REQ_WRITE_UNCHANGED | ((BDRV_REQ_FUA |BDRV_REQ_MAY_UNMAP |BDRV_REQ_NO_FALLBACK ) & bs->file->bs->supported_zero_flags); + bs->supported_truncate_flags =BDRV_REQ_ZERO_WRITE ;
Shouldn't this be:
bs->supported_truncate_flags = (bs->file->bs->supported_truncate_flags &
BDRV_REQ_ZERO_WRITE);
rather than unconditionally advertising something that the underlying layer may lack?
Signature · 3 lines
Eric Blake, Principal Software Engineer Red Hat, Inc. +1-919-301-3226 Virtualization: qemu.org | libvirt.org
Re: [PATCH v4 5/9] raw-format: Support BDRV_REQ_ZERO_WRITE for truncate
Am 20.04.2020 um 17:14 hat Eric Blake geschrieben:
Quoted 28 lines · Kevin Wolf
Quoted 20 lines
The raw format driver can simply forward the flag and let its
bs->filechild take care of actually providing the zeros.Signed-off-by Kevin Wolf <[email protected]>--- block/raw-format.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)block/raw-format.c +1
-1@@ -387,7 +387,7 @@ static int coroutine_fn raw_co_truncate(BlockDriverState *bs, int64_t offset, s->size = offset; offset += s->offset; -return bdrv_co_truncate (bs->file, offset, exact, prealloc,0 , errp); +return bdrv_co_truncate (bs->file, offset, exact, prealloc, flags, errp); }static void raw_eject (BlockDriverState *bs,bool eject_flag)block/raw-format.c static int raw_open(BlockDriverState *bs, QDict *options, int flags, · line 448
Quoted 4 lines
block/raw-format.c +1
-0@@ -445,6 +445,7 @@ static int raw_open(BlockDriverState *bs, QDict *options, int flags, bs->supported_zero_flags =BDRV_REQ_WRITE_UNCHANGED | ((BDRV_REQ_FUA |BDRV_REQ_MAY_UNMAP |BDRV_REQ_NO_FALLBACK ) & bs->file->bs->supported_zero_flags); + bs->supported_truncate_flags =BDRV_REQ_ZERO_WRITE ;Shouldn't this be:
Quoted 5 lines
bs->supported_truncate_flags= (bs->file->bs->supported_truncate_flags&BDRV_REQ_ZERO_WRITE);rather than unconditionally advertising something that the underlying layer may lack?
Maybe that makes more sense, yes.
I think in practice it wouldn't make a difference because the nested bdrv_co_truncate() would still fail rather than silently ignoring the flag. It would behave the same as filter drivers, which also recursively call bdrv_co_truncate() without checking the flag first (which is, of course, because I don't want to modify each filter driver).
Kevin
Re: [PATCH v4 5/9] raw-format: Support BDRV_REQ_ZERO_WRITE for truncate
Quoted 8 lines · Kevin Wolf
Quoted 8 lines
block/raw-format.c static int raw_open(BlockDriverState *bs, QDict *options, int flags, · line 448
Quoted 5 lines
block/raw-format.c +1
-0@@ -445,6 +445,7 @@ static int raw_open(BlockDriverState *bs, QDict *options, int flags, bs->supported_zero_flags =BDRV_REQ_WRITE_UNCHANGED | ((BDRV_REQ_FUA |BDRV_REQ_MAY_UNMAP |BDRV_REQ_NO_FALLBACK ) & bs->file->bs->supported_zero_flags); + bs->supported_truncate_flags =BDRV_REQ_ZERO_WRITE ;Shouldn't this be:
Quoted 7 lines
Quoted 5 lines
bs->supported_truncate_flags= (bs->file->bs->supported_truncate_flags&BDRV_REQ_ZERO_WRITE);rather than unconditionally advertising something that the underlying layer may lack?
Maybe that makes more sense, yes.
If nothing else, it is more consistent with what we are doing for supported_zero_flags. I also argue that having a reference to the passthrough is easier to grep for, if we ever add new flags in the future. That is, while keeping passthrough as opt-in rather than blind copying or blind assignment is slightly more code, it is easier to maintain.
Quoted 6 lines
I think in practice it wouldn't make a difference because the nested
bdrv_co_truncate()would still fail rather than silently ignoring the flag. It would behave the same as filter drivers, which also recursively callbdrv_co_truncate()without checking the flag first (which is, of course, because I don't want to modify each filter driver).
Probably true, but consistency and ease of maintenance are better than proving action at a distance :)
Signature · 3 lines
Eric Blake, Principal Software Engineer Red Hat, Inc. +1-919-301-3226 Virtualization: qemu.org | libvirt.org
PATCH v4 6/9 file-posix: Support BDRV_REQ_ZERO_WRITE for truncate
In reply to Kevin WolfFor regular files, we always get BDRV_REQ_ZERO_WRITE behaviour from the OS, so we can advertise the flag and just ignore it.
--- block/file-posix.c | 4 ++++ 1 file changed, 4 insertions(+)
block/file-posix.c +4 -0
@@ -702,6 +702,10 @@ static int raw_open_common(BlockDriverState *bs, QDict *options,#endif bs->supported_zero_flags =BDRV_REQ_MAY_UNMAP |BDRV_REQ_NO_FALLBACK ; +if (S_ISREG (st.st_mode)) { +/* When extending regular files, we get zeros from the OS */ + bs->supported_truncate_flags =BDRV_REQ_ZERO_WRITE ; + } ret =0 ; fail:if (filename && (bdrv_flags &BDRV_O_TEMPORARY )) {
Signature · 1 line
2.20.1
Re: [PATCH v4 6/9] file-posix: Support BDRV_REQ_ZERO_WRITE for truncate
Quoted 4 lines · Kevin Wolf
For regular files, we always get
BDRV_REQ_ZERO_WRITEbehaviour from the OS, so we can advertise the flag and just ignore it.Signed-off-by Kevin Wolf <[email protected]>
Berto
Re: [PATCH v4 6/9] file-posix: Support BDRV_REQ_ZERO_WRITE for truncate
In reply to Kevin WolfQuoted 4 lines · Kevin Wolf
For regular files, we always get
BDRV_REQ_ZERO_WRITEbehaviour from the OS, so we can advertise the flag and just ignore it.Signed-off-by Kevin Wolf<[email protected]>
Signature · 2 lines
Best regards, Vladimir
PATCH v4 7/9 block: truncate: Don't make backing file data visible
In reply to Kevin WolfWhen extending the size of an image that has a backing file larger than its old size, make sure that the backing file data doesn't become visible in the guest, but the added area is properly zeroed out.
Consider the following scenario where the overlay is shorter than its backing file:
base.qcow2: AAAAAAAA
overlay.qcow2: BBBB
When resizing (extending) overlay.qcow2, the new blocks should not stay unallocated and make the additional As from base.qcow2 visible like before this patch, but zeros should be read.
A similar case happens with the various variants of a commit job when an intermediate file is short (- for unallocated):
base.qcow2: A-A-AAAA
mid.qcow2: BB-B
top.qcow2: C--C--C-After commit top.qcow2 to mid.qcow2, the following happens:
mid.qcow2: CB-C00C0 (correct result)
mid.qcow2: CB-C--C- (before this fix)
Without the fix, blocks that previously read as zeros on top.qcow2 suddenly turn into A.
--- block/io.c | 14 ++++++++++++++ 1 file changed, 14 insertions(+)
block/io.c +14 -0
@@ -3394,6 +3394,20 @@ int coroutine_fn bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact, goto out; } +/* +* If the image has a backing file that is large enough that it would +* provide data for the new area, we cannot leave it unallocated because +* then the backing file content would become visible. Instead, zero-fill +* the area where backing file and new area overlap. +* +* Note that if the image has a backing file, but was opened without the +* backing file, taking care of keeping things consistent with that backing +* file is the user's responsibility. +*/ +if (new_bytes && bs->backing) { + flags |=BDRV_REQ_ZERO_WRITE ; + } +if (drv->bdrv_co_truncate) {if (flags & ~bs->supported_truncate_flags) {error_setg (errp,"Block driver does not support requested flags" );
Signature · 1 line
2.20.1
Re: [PATCH v4 7/9] block: truncate: Don't make backing file data visible
Quoted 23 lines · Kevin Wolf
When extending the size of an image that has a backing file larger than its old size, make sure that the backing file data doesn't become visible in the guest, but the added area is properly zeroed out.
Consider the following scenario where the overlay is shorter than its backing file:
base.qcow2: AAAAAAAA
overlay.qcow2: BBBBWhen resizing (extending) overlay.qcow2, the new blocks should not stay unallocated and make the additional As from base.qcow2 visible like before this patch, but zeros should be read.
A similar case happens with the various variants of a commit job when an intermediate file is short (- for unallocated):
base.qcow2: A-A-AAAA mid.qcow2: BB-B top.qcow2: C--C--C-After commit top.qcow2 to mid.qcow2, the following happens:
Quoted 7 lines
mid.qcow2: CB-C00C0 (correct result)
mid.qcow2: CB-C--C- (before this fix)Without the fix, blocks that previously read as zeros on top.qcow2 suddenly turn into A.
Signed-off-by Kevin Wolf <[email protected]>
Berto
Re: [PATCH v4 7/9] block: truncate: Don't make backing file data visible
In reply to Kevin WolfQuoted 39 lines · Kevin Wolf
When extending the size of an image that has a backing file larger than its old size, make sure that the backing file data doesn't become visible in the guest, but the added area is properly zeroed out.
Consider the following scenario where the overlay is shorter than its backing file:
base.qcow2: AAAAAAAA
overlay.qcow2: BBBBWhen resizing (extending) overlay.qcow2, the new blocks should not stay unallocated and make the additional As from base.qcow2 visible like before this patch, but zeros should be read.
A similar case happens with the various variants of a commit job when an intermediate file is short (- for unallocated):
base.qcow2: A-A-AAAA mid.qcow2: BB-B top.qcow2: C--C--C-After commit top.qcow2 to mid.qcow2, the following happens:
mid.qcow2: CB-C00C0 (correct result)
mid.qcow2: CB-C--C- (before this fix)Without the fix, blocks that previously read as zeros on top.qcow2 suddenly turn into A.
Signed-off-by Kevin Wolf <[email protected]>--- block/io.c | 14 ++++++++++++++ 1 file changed, 14 insertions(+)block/io.c +0
-0
block/io.c int coroutine_fn bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact, · line 3401
Quoted 8 lines
block/io.c +5
-0@@ -3394,6 +3394,20 @@ int coroutine_fn bdrv_co_truncate(BdrvChild *child, int64_t offset, bool exact, goto out; } +/* +* If the image has a backing file that is large enough that it would +* provide data for the new area, we cannot leave it unallocated because +* then the backing file content would become visible. Instead, zero-fill +* the area where backing file and new area overlap.
hmm, actually, you zero-fill the whole new area, not the overlap.
Quoted 13 lines
block/io.c +9
-0+* +* Note that if the image has a backing file, but was opened without the +* backing file, taking care of keeping things consistent with that backing +* file is the user's responsibility. +*/ +if (new_bytes && bs->backing) { + flags |=BDRV_REQ_ZERO_WRITE ; + } +if (drv->bdrv_co_truncate) {if (flags & ~bs->supported_truncate_flags) {error_setg (errp,"Block driver does not support requested flags" );
Signature · 2 lines
Best regards, Vladimir
PATCH v4 8/9 iotests: Filter testfiles out in img_info_log()
In reply to Kevin WolfThis changes the output of some files where instead of filter_img_info() now filter_testfiles() takes precedence, so update the reference output.
--- tests/qemu-iotests/iotests.py | 5 ++++- tests/qemu-iotests/206.out | 12 ++++++------ tests/qemu-iotests/242.out | 12 ++++++------ 3 files changed, 16 insertions(+), 13 deletions(-)
tests/qemu-iotests/iotests.py +4 -1
@@ -152,7 +152,10 @@ def img_info_log(filename, filter_path=None, imgopts=False, extra_args=[]): output = qemu_img_pipe(*args) if not filter_path: filter_path = filename - log(filter_img_info(output, filter_path)) + + output = filter_testfiles(output) + output = filter_img_info(output, filter_path) + log(output) def qemu_io(*args): '''Run qemu-io and return the stdout data'''
tests/qemu-iotests/206.out +6 -6
@@ -12,7 +12,7 @@ {"execute": "job-dismiss", "arguments": {"id": "job0"}} {"return": {}} -image: TEST_IMG +image: TEST_DIR/PID-t.IMGFMT file format: IMGFMT virtual size: 128 MiB (134217728 bytes) cluster_size: 65536
@@ -34,7 +34,7 @@ Format specific information: {"execute": "job-dismiss", "arguments": {"id": "job0"}} {"return": {}} -image: TEST_IMG +image: TEST_DIR/PID-t.IMGFMT file format: IMGFMT virtual size: 64 MiB (67108864 bytes) cluster_size: 65536
@@ -56,7 +56,7 @@ Format specific information: {"execute": "job-dismiss", "arguments": {"id": "job0"}} {"return": {}} -image: TEST_IMG +image: TEST_DIR/PID-t.IMGFMT file format: IMGFMT virtual size: 32 MiB (33554432 bytes) cluster_size: 2097152
@@ -78,11 +78,11 @@ Format specific information: {"execute": "job-dismiss", "arguments": {"id": "job0"}} {"return": {}} -image: TEST_IMG +image: TEST_DIR/PID-t.IMGFMT file format: IMGFMT virtual size: 32 MiB (33554432 bytes) cluster_size: 512 -backing file: TEST_IMG.base +backing file: TEST_DIR/PID-t.IMGFMT.base backing file format: IMGFMT Format specific information: compat: 0.10
@@ -95,7 +95,7 @@ Format specific information: {"execute": "job-dismiss", "arguments": {"id": "job0"}} {"return": {}} -image: TEST_IMG +image: TEST_DIR/PID-t.IMGFMT file format: IMGFMT virtual size: 32 MiB (33554432 bytes) encrypted: yes
tests/qemu-iotests/242.out +6 -6
@@ -6,7 +6,7 @@ wrote 262144/262144 bytes at offset 0 qemu-img info dump: -image: TEST_IMG +image: TEST_DIR/PID-disk file format: IMGFMT virtual size: 1 MiB (1048576 bytes) cluster_size: 65536
@@ -26,7 +26,7 @@ wrote 262144/262144 bytes at offset 262144 qemu-img info dump: -image: TEST_IMG +image: TEST_DIR/PID-disk file format: IMGFMT virtual size: 1 MiB (1048576 bytes) cluster_size: 65536
@@ -58,7 +58,7 @@ wrote 262144/262144 bytes at offset 524288 qemu-img info dump: -image: TEST_IMG +image: TEST_DIR/PID-disk file format: IMGFMT virtual size: 1 MiB (1048576 bytes) cluster_size: 65536
@@ -98,7 +98,7 @@ Test 4 Checking "in-use" flag... qemu-img info dump: -image: TEST_IMG +image: TEST_DIR/PID-disk file format: IMGFMT virtual size: 1 MiB (1048576 bytes) cluster_size: 65536
@@ -143,11 +143,11 @@ Test 5 {"execute": "block-dirty-bitmap-add", "arguments": {"disabled": false, "granularity": 16384, "name": "bitmap-0", "node": "drive0", "persistent": true}} {"return": {}} Write an unknown bitmap flag '0x4' into a new QCOW2 image at offset 327695 -qemu-img: Could not open 'TEST_IMG': Bitmap 'bitmap-0' doesn't satisfy the constraints +qemu-img: Could not open 'TEST_DIR/PID-disk': Bitmap 'bitmap-0' doesn't satisfy the constraints Unset the unknown bitmap flag '0x4' in the bitmap directory entry: -image: TEST_IMG +image: TEST_DIR/PID-disk file format: IMGFMT virtual size: 1 MiB (1048576 bytes) cluster_size: 65536
Signature · 1 line
2.20.1
Re: [PATCH v4 8/9] iotests: Filter testfiles out in img_info_log()
Quoted 15 lines · Kevin Wolf
This changes the output of some files where instead of
filter_img_info()nowfilter_testfiles()takes precedence, so update the reference output.Signed-off-by Kevin Wolf <[email protected]>--- tests/qemu-iotests/iotests.py | 5 ++++- tests/qemu-iotests/206.out | 12 ++++++------ tests/qemu-iotests/242.out | 12 ++++++------ 3 files changed, 16 insertions(+), 13 deletions(-)tests/qemu-iotests/iotests.py +0
-0
tests/qemu-iotests/iotests.py def img_info_log(filename, filter_path=None, imgopts=False, extra_args=[]): · line 158
Quoted 8 lines
tests/qemu-iotests/iotests.py +4
-1@@ -152,7 +152,10 @@ def img_info_log(filename, filter_path=None, imgopts=False, extra_args=[]): output = qemu_img_pipe(*args) if not filter_path: filter_path = filename - log(filter_img_info(output, filter_path)) + + output = filter_testfiles(output) + output = filter_img_info(output, filter_path) + log(output)
Looks weird for me, because filter_img_info is used only in img_info_log, and seems to be the function, which specifically prepare img_info output for printing. So, isn't it better to improve filter_img_info itself?
Hmm, qemu_io_log do the same thing, it uses filter_testfiles, and filter_qemu_io.. So, at least you are not the first.
Still, for applying several filters, seems better to use filters=[] argument of log, like in qemu_io_log.
tests/qemu-iotests/206.out line 15
Quoted 13 lines
tests/qemu-iotests/iotests.py +0
-0def qemu_io(*args): '''Run qemu-io and return the stdout data'''tests/qemu-iotests/206.out +1
-1@@ -12,7 +12,7 @@ {"execute": "job-dismiss", "arguments": {"id": "job0"}} {"return": {}} -image: TEST_IMG +image: TEST_DIR/PID-t.IMGFMT
Honestly, doesn't look as improvement.. But not a problem of course.
Quoted 99 lines
tests/qemu-iotests/206.out +5
-5file format: IMGFMT virtual size: 128 MiB (134217728 bytes) cluster_size: 65536@@ -34,7 +34,7 @@ Format specific information: {"execute": "job-dismiss", "arguments": {"id": "job0"}} {"return": {}} -image: TEST_IMG +image: TEST_DIR/PID-t.IMGFMT file format: IMGFMT virtual size: 64 MiB (67108864 bytes) cluster_size: 65536@@ -56,7 +56,7 @@ Format specific information: {"execute": "job-dismiss", "arguments": {"id": "job0"}} {"return": {}} -image: TEST_IMG +image: TEST_DIR/PID-t.IMGFMT file format: IMGFMT virtual size: 32 MiB (33554432 bytes) cluster_size: 2097152@@ -78,11 +78,11 @@ Format specific information: {"execute": "job-dismiss", "arguments": {"id": "job0"}} {"return": {}} -image: TEST_IMG +image: TEST_DIR/PID-t.IMGFMT file format: IMGFMT virtual size: 32 MiB (33554432 bytes) cluster_size: 512 -backing file: TEST_IMG.base +backing file: TEST_DIR/PID-t.IMGFMT.base backing file format: IMGFMT Format specific information: compat: 0.10@@ -95,7 +95,7 @@ Format specific information: {"execute": "job-dismiss", "arguments": {"id": "job0"}} {"return": {}} -image: TEST_IMG +image: TEST_DIR/PID-t.IMGFMT file format: IMGFMT virtual size: 32 MiB (33554432 bytes) encrypted: yestests/qemu-iotests/242.out +6
-6@@ -6,7 +6,7 @@ wrote 262144/262144 bytes at offset 0 qemu-img info dump: -image: TEST_IMG +image: TEST_DIR/PID-disk file format: IMGFMT virtual size: 1 MiB (1048576 bytes) cluster_size: 65536@@ -26,7 +26,7 @@ wrote 262144/262144 bytes at offset 262144 qemu-img info dump: -image: TEST_IMG +image: TEST_DIR/PID-disk file format: IMGFMT virtual size: 1 MiB (1048576 bytes) cluster_size: 65536@@ -58,7 +58,7 @@ wrote 262144/262144 bytes at offset 524288 qemu-img info dump: -image: TEST_IMG +image: TEST_DIR/PID-disk file format: IMGFMT virtual size: 1 MiB (1048576 bytes) cluster_size: 65536@@ -98,7 +98,7 @@ Test 4 Checking "in-use" flag... qemu-img info dump: -image: TEST_IMG +image: TEST_DIR/PID-disk file format: IMGFMT virtual size: 1 MiB (1048576 bytes) cluster_size: 65536@@ -143,11 +143,11 @@ Test 5 {"execute": "block-dirty-bitmap-add", "arguments": {"disabled": false, "granularity": 16384, "name": "bitmap-0", "node": "drive0", "persistent": true}} {"return": {}} Write an unknown bitmap flag '0x4' into a new QCOW2 image at offset 327695 -qemu-img: Could not open 'TEST_IMG': Bitmap 'bitmap-0' doesn't satisfy the constraints +qemu-img: Could not open 'TEST_DIR/PID-disk': Bitmap 'bitmap-0' doesn't satisfy the constraints Unset the unknown bitmap flag '0x4' in the bitmap directory entry: -image: TEST_IMG +image: TEST_DIR/PID-disk file format: IMGFMT virtual size: 1 MiB (1048576 bytes) cluster_size: 65536
Signature · 2 lines
Best regards, Vladimir
PATCH v4 9/9 iotests: Test committing to short backing file
In reply to Kevin Wolf--- tests/qemu-iotests/274 | 152 ++++++++++++++++++++++++++++ tests/qemu-iotests/274.out | 202 +++++++++++++++++++++++++++++++++++++ tests/qemu-iotests/group | 1 + 3 files changed, 355 insertions(+) create mode 100755 tests/qemu-iotests/274 create mode 100644 tests/qemu-iotests/274.out
tests/qemu-iotests/274 new file +152 -0
@@ -0,0 +1,152 @@ +#!/usr/bin/env python3 +# +# Copyright (C) 2019 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: Kevin Wolf <[email protected]> +# +# Some tests for short backing files and short overlays + +import iotests +import os + +iotests.verify_image_format(supported_fmts=['qcow2']) +iotests.verify_platform(['linux']) + +size_short = 1 * 1024 * 1024 +size_long = 2 * 1024 * 1024 +size_diff = size_long - size_short + +def create_chain(): + iotests.qemu_img_log('create', '-f', iotests.imgfmt, base, + str(size_long)) + iotests.qemu_img_log('create', '-f', iotests.imgfmt, '-b', base, mid, + str(size_short)) + iotests.qemu_img_log('create', '-f', iotests.imgfmt, '-b', mid, top, + str(size_long)) + + iotests.qemu_io_log('-c', 'write -P 1 0 %d' % size_long, base) + +def create_vm(): + vm = iotests.VM() + vm.add_blockdev('file,filename=%s,node-name=base-file' % (base)) + vm.add_blockdev('%s,file=base-file,node-name=base' % (iotests.imgfmt)) + vm.add_blockdev('file,filename=%s,node-name=mid-file' % (mid)) + vm.add_blockdev('%s,file=mid-file,node-name=mid,backing=base' % (iotests.imgfmt)) + vm.add_drive(top, 'backing=mid,node-name=top') + return vm + +with iotests.FilePath('base') as base, \ + iotests.FilePath('mid') as mid, \ + iotests.FilePath('top') as top: + + iotests.log('== Commit tests ==') + + create_chain() + + iotests.log('=== Check visible data ===') + + iotests.qemu_io_log('-c', 'read -P 1 0 %d' % size_short, top) + iotests.qemu_io_log('-c', 'read -P 0 %d %d' % (size_short, size_diff), top) + + iotests.log('=== Checking allocation status ===') + + iotests.qemu_io_log('-c', 'alloc 0 %d' % size_short, + '-c', 'alloc %d %d' % (size_short, size_diff), + base) + + iotests.qemu_io_log('-c', 'alloc 0 %d' % size_short, + '-c', 'alloc %d %d' % (size_short, size_diff), + mid) + + iotests.qemu_io_log('-c', 'alloc 0 %d' % size_short, + '-c', 'alloc %d %d' % (size_short, size_diff), + top) + + iotests.log('=== Checking map ===') + + iotests.qemu_img_log('map', '--output=json', base) + iotests.qemu_img_log('map', '--output=human', base) + iotests.qemu_img_log('map', '--output=json', mid) + iotests.qemu_img_log('map', '--output=human', mid) + iotests.qemu_img_log('map', '--output=json', top) + iotests.qemu_img_log('map', '--output=human', top) + + iotests.log('=== Testing qemu-img commit (top -> mid) ===') + + iotests.qemu_img_log('commit', top) + iotests.img_info_log(mid) + iotests.qemu_io_log('-c', 'read -P 1 0 %d' % size_short, mid) + iotests.qemu_io_log('-c', 'read -P 0 %d %d' % (size_short, size_diff), mid) + + iotests.log('=== Testing HMP commit (top -> mid) ===') + + create_chain() + with create_vm() as vm: + vm.launch() + vm.qmp_log('human-monitor-command', command_line='commit drive0') + + iotests.img_info_log(mid) + iotests.qemu_io_log('-c', 'read -P 1 0 %d' % size_short, mid) + iotests.qemu_io_log('-c', 'read -P 0 %d %d' % (size_short, size_diff), mid) + + iotests.log('=== Testing QMP active commit (top -> mid) ===') + + create_chain() + with create_vm() as vm: + vm.launch() + vm.qmp_log('block-commit', device='top', base_node='mid', + job_id='job0', auto_dismiss=False) + vm.run_job('job0', wait=5) + + iotests.img_info_log(mid) + iotests.qemu_io_log('-c', 'read -P 1 0 %d' % size_short, mid) + iotests.qemu_io_log('-c', 'read -P 0 %d %d' % (size_short, size_diff), mid) + + + iotests.log('== Resize tests ==') + + # Use different sizes for different allocation modes: + # + # We want to have at least one test where 32 bit truncation in the size of + # the overlapping area becomes visible. This is covered by the + # prealloc='off' case (1G to 6G is an overlap of 5G). + # + # However, we can only do this for modes that don't preallocate data + # because otherwise we might run out of space on the test host. + for (prealloc, base_size, top_size_old, top_size_new, off) in [ + ('off', '6G', '1G', '8G', '5G'), + ('metadata', '32G', '30G', '33G', '31G'), + ('falloc', '10M', '5M', '15M', '9M'), + ('full', '16M', '8M', '12M', '11M')]: + + iotests.log('=== preallocation=%s ===' % prealloc) + iotests.qemu_img_log('create', '-f', iotests.imgfmt, base, base_size) + iotests.qemu_img_log('create', '-f', iotests.imgfmt, '-b', base, top, + top_size_old) + iotests.qemu_io_log('-c', 'write -P 1 %s 64k' % off, base) + + # After this, 0 to base_size should be allocated/zeroed + # base_size to top_size_new should be unallocated with + # preallocation=off and allocated with preallocation enabled + iotests.qemu_img_log('resize', '-f', iotests.imgfmt, + '--preallocation', prealloc, top, top_size_new) + iotests.qemu_io_log('-c', 'read -P 0 %s 64k' % off, top) + + # Metadata preallocation doesn't have a defined result on the file + # system level with respect to holes, so skip it here + iotests.qemu_io_log('-c', 'map', top) + if prealloc != 'metadata': + iotests.qemu_img_log('map', '--output=json', top)
tests/qemu-iotests/274.out new file +202 -0
202 lines not on this page · raw message
tests/qemu-iotests/group +1 -0
@@ -286,6 +286,7 @@ 270 rw backing quick 272 rw 273 backing quick +274 rw backing 277 rw quick 279 rw backing quick 280 rw migration quick
Signature · 1 line
2.20.1
Re: [PATCH v4 9/9] iotests: Test committing to short backing file
Quoted 38 lines · Kevin Wolf
Signed-off-by Kevin Wolf <[email protected]>Reviewed-by Eric Blake <[email protected]>Reviewed-by Vladimir Sementsov-Ogievskiy <[email protected]>--- tests/qemu-iotests/274 | 152 ++++++++++++++++++++++++++++ tests/qemu-iotests/274.out | 202 +++++++++++++++++++++++++++++++++++++ tests/qemu-iotests/group | 1 + 3 files changed, 355 insertions(+)create mode 100755 tests/qemu-iotests/274
create mode 100644 tests/qemu-iotests/274.outtests/qemu-iotests/274 new file +20
-0@@ -0,0 +1,152 @@ +#!/usr/bin/env python3 +# +# Copyright (C) 2019 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: Kevin Wolf <[email protected]> +# +# Some tests for short backing files and short overlays
tests/qemu-iotests/274 line 30
Quoted 10 lines
tests/qemu-iotests/274 new file +10
-0+ +import iotests +import os + +iotests.verify_image_format(supported_fmts=['qcow2']) +iotests.verify_platform(['linux']) + +size_short = 1 * 1024 * 1024 +size_long = 2 * 1024 * 1024 +size_diff = size_long - size_short
Would be good to add unaligned-to-cluster testcase.
Signature · 2 lines
Best regards, Vladimir
Patchew URL: https://patchew.org/QEMU/[email protected]/
Patchew URL: https://patchew.org/QEMU/[email protected]/
Hi,
This series failed the docker-mingw@fedora build test. Please find the testing commands and their output below. If you have Docker installed, you can probably reproduce it locally.
=== TEST SCRIPT BEGIN ===
#! /bin/bash
export ARCH=x86_64
make docker-image-fedora V=1 NETWORK=1
time make docker-test-mingw@fedora J=14 NETWORK=1
=== TEST SCRIPT END ===
CC block/dirty-bitmap.o CC block/write-threshold.o CC block/backup.o
/tmp/qemu-test/src/block/file-win32.c:647:27: error: initialization of 'int (*)(BlockDriverState *, int64_t, _Bool, PreallocMode, BdrvRequestFlags, Error **)' {aka 'int (*)(struct BlockDriverState *, long long int, _Bool, enum PreallocMode, enum <anonymous>, struct Error **)'} from incompatible pointer type 'int (*)(BlockDriverState *, int64_t, _Bool, PreallocMode, Error **)' {aka 'int (*)(struct BlockDriverState *, long long int, _Bool, enum PreallocMode, struct Error **)'} [-Werror=incompatible-pointer-types]
.bdrv_co_truncate = raw_co_truncate,
^~~~~~~~~~~~~~~/tmp/qemu-test/src/block/file-win32.c:647:27: note: (near initialization for 'bdrv_file.bdrv_co_truncate')
cc1: all warnings being treated as errors
make: *** [/tmp/qemu-test/src/rules.mak:69: block/file-win32.o] Error 1
make: *** Waiting for unfinished jobs....
Traceback (most recent call last):
File "./tests/docker/docker.py", line 664, in <module>
---
raise CalledProcessError(retcode, cmd)
subprocess.CalledProcessError: Command '['sudo', '-n', 'docker', 'run', '--label', 'com.qemu.instance.uuid=67092de850d740e099c72a2716592e10', '-u', '1001', '--security-opt', 'seccomp=unconfined', '--rm', '-e', 'TARGET_LIST=', '-e', 'EXTRA_CONFIGURE_OPTS=', '-e', 'V=', '-e', 'J=14', '-e', 'DEBUG=', '-e', 'SHOW_ENV=', '-e', 'CCACHE_DIR=/var/tmp/ccache', '-v', '/home/patchew/.cache/qemu-docker-ccache:/var/tmp/ccache:z', '-v', '/var/tmp/patchew-tester-tmp-dnai7bqe/src/docker-src.2020-04-20-19.46.23.5204:/var/tmp/qemu:z,ro', 'qemu:fedora', '/var/tmp/qemu/run', 'test-mingw']' returned non-zero exit status 2. filter=--filter=label=com.qemu.instance.uuid=67092de850d740e099c72a2716592e10
make[1]: *** [docker-run] Error 1
make[1]: Leaving directory `/var/tmp/patchew-tester-tmp-dnai7bqe/src'
make: *** [docker-run-test-mingw@fedora] Error 2
real 2m42.450s
user 0m9.080s
The full log is available at
http://patchew.org/logs/[email protected]/testing.docker-mingw@fedora/?type=message.
---
Email generated automatically by Patchew [https://patchew.org/].
Please send your feedback to [email protected]