Re: [PATCH] selftests/lkdtm: Use "comm" instead of "diff" for dmesg

[Kees Cook <keescook@chromium.org>]

On Wed, Sep 09, 2020 at 04:29:50PM -0400, Joe Lawrence wrote:
> On 9/9/20 3:49 PM, Kees Cook wrote:
> > 
> > On Fri, Jun 26, 2020 at 01:59:43PM -0700, Kees Cook wrote:
> > > Instead of full GNU diff (which smaller boot environments may not have),
> > > use "comm" which is more available.
> > > 
> > > Reported-by: Naresh Kamboju <naresh.kamboju@linaro.org>
> > > Link: https://lore.kernel.org/lkml/CA+G9fYtHP+Gg+BrR_GkBMxu2oOi-_e9pATtpb6TVRswv1G1r1Q@mail.gmail.com
> > > Fixes: f131d9edc29d ("selftests/lkdtm: Don't clear dmesg when running tests")
> > > Signed-off-by: Kees Cook <keescook@chromium.org>
> > 
> > Shuah, this really needs to land to fix lkdtm tests on busybox. Can
> > you add this to -next? (Or is it better to direct this to Greg for the
> > lkdtm tree?)
> > 
> > Thanks!
> > 
> > -Kees
> > 
> > > ---
> > >   tools/testing/selftests/lkdtm/run.sh | 2 +-
> > >   1 file changed, 1 insertion(+), 1 deletion(-)
> > > 
> > > diff --git a/tools/testing/selftests/lkdtm/run.sh b/tools/testing/selftests/lkdtm/run.sh
> > > index 8383eb89d88a..5fe23009ae13 100755
> > > --- a/tools/testing/selftests/lkdtm/run.sh
> > > +++ b/tools/testing/selftests/lkdtm/run.sh
> > > @@ -82,7 +82,7 @@ dmesg > "$DMESG"
> > >   ($SHELL -c 'cat <(echo '"$test"') >'"$TRIGGER" 2>/dev/null) || true
> > >   # Record and dump the results
> > > -dmesg | diff --changed-group-format='%>' --unchanged-group-format='' "$DMESG" - > "$LOG" || true
> > > +dmesg | comm -13 "$DMESG" - > "$LOG" || true
> > >   cat "$LOG"
> > >   # Check for expected output
> > > -- 
> > > 2.25.1
> > > 
> > > 
> > > -- 
> > > Kees Cook
> > 
> 
> Hi Kees,
> 
> You may want to consider a similar follow up to the one Miroslav made to the
> livepatching equivalent:
> 
> https://lore.kernel.org/live-patching/nycvar.YFH.7.76.2008271528000.27422@cbobk.fhfr.pm/T/#m1c17812d2c005dd57e9a299a4a492026a156619e
> 
> basically 'comm' will complain if two lines from dmesg have the same
> timestamp prefix and their text portions are not sorted.

Ah-ha! Thank you. I will send a v2. :)

-- 
Kees Cook