Re: [PATCH] generic: concurrent IO test with mixed IO types

[Dave Chinner <david@fromorbit.com>]

On Mon, Jun 08, 2015 at 01:02:48PM +0200, Lukáš Czerner wrote:
> On Mon, 8 Jun 2015, Eryu Guan wrote:
> 
> > Date: Mon,  8 Jun 2015 18:41:41 +0800 From: Eryu Guan
> > <eguan@redhat.com> To: fstests@vger.kernel.org Cc:
> > lczerner@redhat.com, Eryu Guan <eguan@redhat.com> Subject:
> > [PATCH] generic: concurrent IO test with mixed IO types
> > 
> > Test concurrent buffered I/O, DIO, AIO, mmap I/O and splice I/O
> > on the same files.
> 
> Hi Eryu,
> 
> Great! Thanks for turning it into a test. Couple of comments
> below.
....
> > + +seq=`basename $0` +seqres=$RESULT_DIR/$seq +echo "QA output
> > created by $seq" + +here=`pwd` +tmp=/tmp/$$
> 
> Not a big deal, but using mktemp woud probably be better.

Not better at all - this reduces the easy of debugging problems.
e.g.  if a test is hung, I know where it's temporary output files
are just by looking at the process listing and grabbing the pid of
the test process being run. IOWs, we *want* deterministic temporary
filenames here - obscuring the temporary file names that tests use
doesn't gain us anything.

Yes, I see that some btrfs tests have used mktemp  and a couple of
generic tests use $TEST_DIR/$$ - they should be reverted to /tmp/$$
so they are consistent with the majority of tests and the new test
template....

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com