From: Fiona Ebner <[email protected]> Commit 5634622bcb ("file-posix: allow BLKZEROOUT with -t writeback") enables the BLKZEROOUT ioctl when using 'writeback' cache, regressing certain 'qemu-img convert' invocations, because of a pre-existing issue. Namely, the BLKZEROOUT ioctl might fail with errno EINVAL when the request is shorter than the block size of the block device. Fallback to the bounce buffer, similar to when the ioctl is not supported at all, rather than treating such an error as fatal.
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/3257 Resolves: https://bugzilla.proxmox.com/show_bug.cgi?id=7197 Cc: [email protected] Signed-off-by: Fiona Ebner <[email protected]> Message-ID: <[email protected]> [Added TODO comment describing a larger fix that could be implemented in the future. --Stefan] Signed-off-by: Stefan Hajnoczi <[email protected]> --- block/io.c | 13 ++++++++++++- 1 file changed, 12 insertions(+), 1 deletion(-) diff --git a/block/io.c b/block/io.c index e8fb4ede4d..6c0bbdcf1e 100644 --- a/block/io.c +++ b/block/io.c @@ -1918,7 +1918,18 @@ bdrv_co_do_pwrite_zeroes(BlockDriverState *bs, int64_t offset, int64_t bytes, assert(!bs->supported_zero_flags); } - if (ret == -ENOTSUP && !(flags & BDRV_REQ_NO_FALLBACK)) { + /* + * TODO The ret == -EINVAL && num < alignment case is a workaround for + * when request_alignment is 1 on files with cache=writeback. The Linux + * ioctl(BLKZEROOUT) requires block alignment and will fail with + * EINVAL. The block layer should align the request to + * write_zeroes_alignment instead of trying the syscall, failing, and + * falling back to a bounce buffer. Doing that is not easy so for now + * we use a bounce buffer: + * https://lore.kernel.org/qemu-devel/[email protected]/ + */ + if ((ret == -ENOTSUP || (ret == -EINVAL && num < alignment)) && + !(flags & BDRV_REQ_NO_FALLBACK)) { /* Fall back to bounce buffer if write zeroes is unsupported */ BdrvRequestFlags write_flags = flags & ~BDRV_REQ_ZERO_WRITE; -- 2.54.0
