On Sun, 2024-07-21 at 02:49 +0200, Bruno Haible wrote:
> Hi Jeff,
>
> You wrote on 2024-06-28:
> > There are some proposed changes [1] to track finer-grained timestamps in
> > the Linux kernel that will break the assumptions that nap() uses to
> > gauge the delay. In particular, writing to a file will almost always
> > show a change in the timestamp now, so usually this method will settle
> > on a delay of 1ns.
> > ...
> > test-stat-time still fails without this change.
>
> The paperwork being complete now, I wanted to verify that the patch
> indeed fixes the problem. But I can't reproduce the problem in the first
> place.
>
> Here's what I did:
> - Install Fedora Rawhide (Linux
> 6.11.0-0.rc0.20240719git720261cfc732.7.fc41.x86_64).
> - Create a testdir
> $ ./gnulib-tool --create-testdir --dir=../testdir1 --single-configure
> chown fchownat fdutimensat futimens lchown stat-time utime utimens utimensat
> - Build it, both on an ext4 file system and a btrfs file system.
> => No test failure is seen.
>
> What am I missing? What do I need to change in my setup, in order to
> reproduce the problem?
>
> Bruno
>
>
>
Unfortunately, these patches aren't in rawhide kernels yet. Christian
will pull these into his tree once the dust settles from the merge
window. Once that happens, it should go into rawhide soon after.
If you want to test before that, the patches are in the "mgtime" branch
of my public tree:
https://git.kernel.org/pub/scm/linux/kernel/git/jlayton/linux.git/
...but you'll have to build your own kernel for that.
--
Jeff Layton <jlay...@kernel.org>
