* Re: [PATCHv2 0/7] direct-io: even more flexible io vectors
[not found] ` <aJU0scj_dR8_37S8@kbusch-mbp>
@ 2025-08-11 10:32 ` Christoph Hellwig
0 siblings, 0 replies; only message in thread
From: Christoph Hellwig @ 2025-08-11 10:32 UTC (permalink / raw)
To: Keith Busch
Cc: Christoph Hellwig, Keith Busch, linux-block, linux-fsdevel,
linux-kernel, snitzer, axboe, dw, brauner, linux-xfs
On Thu, Aug 07, 2025 at 05:20:17PM -0600, Keith Busch wrote:
> Sure. I wrote up some for blktest, and the same test works as-is for
> filesystems too. Potential question: where do such programs go
> (xfstests, blktests, both, or some common place)?
We currently have no good way to share tests between xfstest and
blktests, so I think it would require duplicating the helper programs.
> I tested on loop, nvme, and virtio-blk, both raw block (blktests) and
> xfs (fstests). Seems fine.
Cool. I'd like the hear from the other XFS folks if the possibility
of easily introducing preallocated space (that's what it looks like on
disk) for writes failed because of wrong alignment is fine. Given that
the same user could introduce them using fallocate it's definitively not
a security issue, but also rather unusual and unexpected. And from the
other file system maintainers if they have similar issues.
^ permalink raw reply [flat|nested] only message in thread