On Fri, Jan 04, 2019 at 11:54:11AM +0200, Nikolay Borisov wrote:
>
>
> On 4.01.19 г. 11:43 ч., Eryu Guan wrote:
> > On Fri, Jan 04, 2019 at 08:56:09AM +1100, Dave Chinner wrote:
> >> On Thu, Jan 03, 2019 at 02:29:08PM +0200, Nikolay Borisov wrote:
> >>>
> >>>
> >>> On 3.01.19 г. 12:01 ч., Nikolay Borisov wrote:
> >>>> Commit 7fc034868d5d ("common/config: create $RESULT_BASE before dumping kmemleak leaks")
> >>>> inadvertently broke $RESULT_BASE dir creation since it changed the logic
> >>>> to only create the directory only if this variable is not explicitly set
> >>>> by the user. Fix this by ensuring RESULT_BASE is always created.
> >>>>
> >>>> Signed-off-by: Nikolay Borisov <nborisov@xxxxxxxx>
> >>>
> >>> Eryu,
> >>>
> >>> I think Johannes' commit should actually be reverted. Currently
> >>> get_next_config is called at the beginning of the section code _AFTER_
> >>> _init_kmemleak so it's not really fixing the problem that Johannes
> >>> described. So care to revert his commit ?
> >>
> >> Yes, the original commit should be reverted. Using RESULT_BASE in
> >> the way that init_kmemleak is using it is fundamentally broken.
> >> Using RESULTS_BASE is incorrect - harness run state is supposed to
> >> be stored in the section-aware RESULTS_DIR (e.g. see
> >> _require_scratch()) which is constant and always exists for each
> >> section that is run.
> >>
> >> The code was architectected this way because section definitions can
> >> change RESULT_BASE, and that means the RESULT_BASE defined when
> >> init_kmemleak() is called may not point to the same location as when
> >> check_memleak() is called. Hence check_memleak will never run if a
> >> section definition changes RESULT_BASE.
> >>
> >> IOWs, the init_kmemleak() call needs to be done inside the
> >> section iteration loop, after the section config has been parsed and
> >> we've determined if the the section will be run. This is where the
> >> original code created RESULT_BASE if it didn't exist. Further, all
> >> the section run state is then stored in RESULTS_DIR, which is
> >> created from the current RESULT_BASE (e.g. see _require_scratch,
> >> check_dmesg, etc).
> >
> > This makes more sense, thanks for the suggestion!
> >
> >>
> >> Hence init_kmemleak() and check_kmemleak() should have a scope
> >> inside a valid RESULT_DIR path. And once this is done, you can then
> >> add a USE_KMEMLEAK section variable to turn kmemleak detection on
> >> and off via the config file on a per-section basis.....
> >>
> >> So, IMO, the correct thing to do here is revert the original broken
> >> change and fix the kmemleak infrastructure to be properly aware of
> >> config sections.
> >
> > I think we can take Nikolay's patch for now so we don't leave
> > user-specified RESULT_BASE not created for any longer, and revert it and
> > 7fc034868d5d when we rework kmemleak infrastructure.
>
> Actually no, for the following RESULT_BASE pattern:
> RESULT_BASE="$PWD/results/$HOST/$(uname -r)"/$(date +%Y-%m-%d-%H-%M-%N)/
>
>
> Running a test I can see the following mkdirs happening:
>
> created RESULT_BASE: /root/xfstests-dev/results/linux/4.12.14-nikbor/2019-01-04-09-48-027249346/
> created RESULT_BASE: /root/xfstests-dev/results/linux/4.12.14-nikbor/2019-01-04-09-48-093731409/
> created RESULT_BASE: /root/xfstests-dev/results/linux/4.12.14-nikbor/2019-01-04-09-48-093731409/
>
> In my case for every test run it potentially creates 2 directories, where only one is really
OK, I see it now, it only happens when defining RESULT_BASE as above in
section based config file, non-section based config is not affected and
it worked if RESULT_BASE is defined from the command line.
> populated with data. This is no good. I will strongly suggest that you revert the original patch and
> let kmemleak be broken rather than piling hacks (which have funky side effects as I have shown above)
> or leaving dir creation broken altogether.
Would you like to send a revert patch?
>
> While at it you might also consider reverting 074740a32c6a36c5ba7d4be66dd4ee63e9f744f3 until Darick
> (or anyone else interested) properly integrated kmemleak support.
It's been broken for almost one year, so I think it's fine to just
revert 7fc034868d5d. I'll look into rework kmemleak next week.
Thanks,
Eryu
