[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[PULL 08/20] block/file-posix: Fix problem with fallocate(PUNCH_HOLE) on
From: |
Kevin Wolf |
Subject: |
[PULL 08/20] block/file-posix: Fix problem with fallocate(PUNCH_HOLE) on GPFS |
Date: |
Wed, 2 Jun 2021 15:45:17 +0200 |
From: Thomas Huth <thuth@redhat.com>
A customer reported that running
qemu-img convert -t none -O qcow2 -f qcow2 input.qcow2 output.qcow2
fails for them with the following error message when the images are
stored on a GPFS file system :
qemu-img: error while writing sector 0: Invalid argument
After analyzing the strace output, it seems like the problem is in
handle_aiocb_write_zeroes(): The call to fallocate(FALLOC_FL_PUNCH_HOLE)
returns EINVAL, which can apparently happen if the file system has
a different idea of the granularity of the operation. It's arguably
a bug in GPFS, since the PUNCH_HOLE mode should not result in EINVAL
according to the man-page of fallocate(), but the file system is out
there in production and so we have to deal with it. In commit 294682cc3a
("block: workaround for unaligned byte range in fallocate()") we also
already applied the a work-around for the same problem to the earlier
fallocate(FALLOC_FL_ZERO_RANGE) call, so do it now similar with the
PUNCH_HOLE call. But instead of silently catching and returning
-ENOTSUP (which causes the caller to fall back to writing zeroes),
let's rather inform the user once about the buggy file system and
try the other fallback instead.
Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20210527172020.847617-2-thuth@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
---
block/file-posix.c | 11 +++++++++++
1 file changed, 11 insertions(+)
diff --git a/block/file-posix.c b/block/file-posix.c
index 10b71d9a13..6e24083f3f 100644
--- a/block/file-posix.c
+++ b/block/file-posix.c
@@ -1650,6 +1650,17 @@ static int handle_aiocb_write_zeroes(void *opaque)
return ret;
}
s->has_fallocate = false;
+ } else if (ret == -EINVAL) {
+ /*
+ * Some file systems like older versions of GPFS do not like un-
+ * aligned byte ranges, and return EINVAL in such a case, though
+ * they should not do it according to the man-page of fallocate().
+ * Warn about the bad filesystem and try the final fallback
instead.
+ */
+ warn_report_once("Your file system is misbehaving: "
+ "fallocate(FALLOC_FL_PUNCH_HOLE) returned EINVAL.
"
+ "Please report this bug to your file sytem "
+ "vendor.");
} else if (ret != -ENOTSUP) {
return ret;
} else {
--
2.30.2
- [PULL 00/20] Block layer patches, Kevin Wolf, 2021/06/02
- [PULL 02/20] qemu-io-cmds: assert that we don't have .perm requested in no-blk case, Kevin Wolf, 2021/06/02
- [PULL 01/20] block/quorum: Provide .bdrv_co_flush instead of .bdrv_co_flush_to_disk, Kevin Wolf, 2021/06/02
- [PULL 03/20] block/vvfat: child_vvfat_qcow: add .get_parent_aio_context, fix crash, Kevin Wolf, 2021/06/02
- [PULL 04/20] block/vvfat: fix vvfat_child_perm crash, Kevin Wolf, 2021/06/02
- [PULL 05/20] block: consistently use bdrv_is_read_only(), Kevin Wolf, 2021/06/02
- [PULL 06/20] block: drop BlockDriverState::read_only, Kevin Wolf, 2021/06/02
- [PULL 08/20] block/file-posix: Fix problem with fallocate(PUNCH_HOLE) on GPFS,
Kevin Wolf <=
- [PULL 07/20] block: drop BlockBackendRootState::read_only, Kevin Wolf, 2021/06/02
- [PULL 10/20] block: document child argument of bdrv_attach_child_common(), Kevin Wolf, 2021/06/02
- [PULL 09/20] block/file-posix: Try other fallbacks after invalid FALLOC_FL_ZERO_RANGE, Kevin Wolf, 2021/06/02
- [PULL 11/20] block-backend: improve blk_root_get_parent_desc(), Kevin Wolf, 2021/06/02
- [PULL 12/20] block: improve bdrv_child_get_parent_desc(), Kevin Wolf, 2021/06/02
- [PULL 13/20] block/vvfat: inherit child_vvfat_qcow from child_of_bds, Kevin Wolf, 2021/06/02
- [PULL 14/20] block: simplify bdrv_child_user_desc(), Kevin Wolf, 2021/06/02
- [PULL 15/20] block: improve permission conflict error message, Kevin Wolf, 2021/06/02
- [PULL 17/20] nbd/server: Use drained block ops to quiesce the server, Kevin Wolf, 2021/06/02
- [PULL 16/20] block-backend: add drained_poll, Kevin Wolf, 2021/06/02