From: "Darrick J. Wong" <darrick.wong@oracle.com>
To: Eric Sandeen <sandeen@sandeen.net>
Cc: linux-xfs <linux-xfs@vger.kernel.org>
Subject: Re: [PATCH] xfsprogs: fix geometry calls on older kernels for 5.2.1
Date: Tue, 20 Aug 2019 14:18:28 -0700 [thread overview]
Message-ID: <20190820211828.GC1037350@magnolia> (raw)
In-Reply-To: <7d83cd0d-8a15-201e-9ebf-e1f859270b92@sandeen.net>
On Tue, Aug 20, 2019 at 03:47:29PM -0500, Eric Sandeen wrote:
> I didn't think 5.2.0 through; the udpate of the geometry ioctl means
> that the tools won't work on older kernels that don't support the
> v5 ioctls, since I failed to merge Darrick's wrappers.
>
> As a very quick one-off I'd like to merge this to just revert every
> geometry call back to the original ioctl, so it keeps working on
> older kernels and I'll release 5.2.1. This hack can go away when
> Darrick's wrappers get merged.
>
> Signed-off-by: Eric Sandeen <sandeen@redhat.com>
For the four line code fix,
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
> ---
>
> I'm a little concerned that 3rd party existing code which worked fine
> before will now get the new XFS_IOC_FSGEOMETRY definition if they get
> rebuilt, and suddenly stop working on older kernels. Am I overreacting
> or misunderstanding our compatibility goals?
As for this question ^^^ ... <URRRK>.
I thought the overall strategy was to get everything in xfsprogs using
libfrog wrappers that would degrade gracefully on old kernels.
For xfsdump/restore, I think we should just merge it into xfsprogs and
then it can use our wrappers.
For everything else... I thought the story was that you shouldn't really
be using xfs ioctls unless you're keeping up with upstream.
<shrug> Feel free to differ, that's just a braindump of my shattered
mind. :P
--D
> diff --git a/libxfs/xfs_fs.h b/libxfs/xfs_fs.h
> index f1158a79..253b706c 100644
> --- a/libxfs/xfs_fs.h
> +++ b/libxfs/xfs_fs.h
> @@ -720,7 +720,10 @@ struct xfs_scrub_metadata {
> #define XFS_IOC_ATTRMULTI_BY_HANDLE _IOW ('X', 123, struct xfs_fsop_attrmulti_handlereq)
> #define XFS_IOC_FSGEOMETRY_V4 _IOR ('X', 124, struct xfs_fsop_geom_v4)
> #define XFS_IOC_GOINGDOWN _IOR ('X', 125, uint32_t)
> -#define XFS_IOC_FSGEOMETRY _IOR ('X', 126, struct xfs_fsop_geom)
> +/* For backwards compatibility in 5.2.1, just for now */
> +/* #define XFS_IOC_FSGEOMETRY _IOR ('X', 126, struct xfs_fsop_geom_v5) */
> +#define XFS_IOC_FSGEOMETRY XFS_IOC_FSGEOMETRY_V4
> +
> /* XFS_IOC_GETFSUUID ---------- deprecated 140 */
>
> /* reflink ioctls; these MUST match the btrfs ioctl definitions */
>
next prev parent reply other threads:[~2019-08-20 21:19 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-08-20 20:47 [PATCH] xfsprogs: fix geometry calls on older kernels for 5.2.1 Eric Sandeen
2019-08-20 21:18 ` Darrick J. Wong [this message]
2019-08-20 22:46 ` Dave Chinner
2019-08-21 15:39 ` Eric Sandeen
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20190820211828.GC1037350@magnolia \
--to=darrick.wong@oracle.com \
--cc=linux-xfs@vger.kernel.org \
--cc=sandeen@sandeen.net \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox