On 10/16/21 1:29 AM, Dave Chinner wrote:

On Fri, Oct 15, 2021 at 09:10:17PM -0700, Darrick J. Wong wrote:

quoted

On Fri, Oct 15, 2021 at 03:32:41PM -0500, Eric Sandeen wrote:

quoted

On 10/15/21 2:42 PM, Eric Sandeen wrote:

quoted

Hi folks,

The for-next branch of the xfsprogs repository at:

      git://git.kernel.org/pub/scm/fs/xfs/xfsprogs-dev.git

has just been updated.

Patches often get missed, so please check if your outstanding
patches were in this update. If they have not been in this update,
please resubmit them to linux-xfs@vger.kernel.org so they can be
picked up in the next update.

This is really just the libxfs-5.14 sync (finally!).  Big thanks
to chandan, djwong, dchinner who all helped significantly with what
was a much more challenging libxfs sync this time.

Odds are this will be the bulk of the final 5.14 release. I will just
add Darrick's deprecation warning, and anything else I get reminded
of in the next week.  :)

I missed Derrick's "libxfs: fix crash on second attempt to initialize library"
because my old userspace rcu library did not exhibit the problem. :/

Rather than leave a few dozen commits with regressed behavior as a bisect bomb,
I have force-pushed and anybody who pulled in the last hour will need to rebase.
Sorry about that!

Er... /me notices the following discrepancy between the kernel and
xfsprogs in libxfs/xfs_shared.h:

--- a/libxfs/xfs_shared.h
+++ b/libxfs/xfs_shared.h

@@ -174,24 +174,4 @@ struct xfs_ino_geometry {
  
  };
  
-/* Faked up kernel bits */
-struct rb_root {
-};
-
-#define RB_ROOT                (struct rb_root) { }
-
-typedef struct wait_queue_head {
-} wait_queue_head_t;
-
-#define init_waitqueue_head(wqh)       do { } while(0)
-
-struct rhashtable {
-};
-
-struct delayed_work {
-};
-
-#define INIT_DELAYED_WORK(work, func)  do { } while(0)
-#define cancel_delayed_work_sync(work) do { } while(0)
-
  #endif /* __XFS_SHARED_H__ */

Shouldn't those be libxfs/libxfs_priv.h along with all the other
faked up kernel bits?

IIRC that doesn't work but I don't remember why - has to do w/ the ag code move,
and use in userspace IIRC. And TBH I forgot that "xfs_shared" was uhhhh actually
shared with the kernel and failed to check.

I'll figure something out.  Tired of pushing this stuff from file to file,
subdir to subdir ... was trying to avoid libfrog.

But thanks for noticing, Darrick - and sorry for missing it.

-Eric