Re: [PATCH 3/3] xfs_repair: add post-phase error injection points

["Darrick J. Wong" <djwong@kernel.org>]

On Tue, Feb 16, 2021 at 06:58:39AM -0500, Brian Foster wrote:
> On Fri, Feb 12, 2021 at 09:47:01PM -0800, Darrick J. Wong wrote:
> > From: Darrick J. Wong <djwong@kernel.org>
> > 
> > Create an error injection point so that we can simulate repair failing
> > after a certain phase.
> > 
> > Signed-off-by: Darrick J. Wong <djwong@kernel.org>
> > ---
> >  repair/globals.c    |    3 +++
> >  repair/globals.h    |    3 +++
> >  repair/progress.c   |    3 +++
> >  repair/xfs_repair.c |    4 ++++
> >  4 files changed, 13 insertions(+)
> > 
> > 
> ...
> > diff --git a/repair/progress.c b/repair/progress.c
> > index e5a9c1ef..5bbe58ec 100644
> > --- a/repair/progress.c
> > +++ b/repair/progress.c
> > @@ -410,6 +410,9 @@ timestamp(int end, int phase, char *buf)
> >  		current_phase = phase;
> >  	}
> >  
> > +	if (fail_after_phase && phase == fail_after_phase)
> > +		kill(getpid(), SIGKILL);
> > +
> 
> It seems a little hacky to bury this in timestamp(). Perhaps we should
> at least check for end == PHASE_END (even though PHASE_START is
> currently only used in one place). Otherwise seems reasonable..

Yeah, I don't know of a better place to put it -- adding another call
after every timestamp() just seems like clutter.

Will fix it to check for PHASE_END, though.

Thanks for reading. :)

--D

> Brian
> 
> >  	if (buf) {
> >  		tmp = localtime((const time_t *)&now);
> >  		sprintf(buf, _("%02d:%02d:%02d"), tmp->tm_hour, tmp->tm_min, tmp->tm_sec);
> > diff --git a/repair/xfs_repair.c b/repair/xfs_repair.c
> > index 12e319ae..6b60b8f4 100644
> > --- a/repair/xfs_repair.c
> > +++ b/repair/xfs_repair.c
> > @@ -362,6 +362,10 @@ process_args(int argc, char **argv)
> >  
> >  	if (report_corrected && no_modify)
> >  		usage();
> > +
> > +	p = getenv("XFS_REPAIR_FAIL_AFTER_PHASE");
> > +	if (p)
> > +		fail_after_phase = (int)strtol(p, NULL, 0);
> >  }
> >  
> >  void __attribute__((noreturn))
> > 
>