[LTP] [PATCH] Disable failure hints before we actually run the test

[Martin Doucha]

On 17. 09. 24 17:50, Cyril Hrubis wrote:
> Hi!
>> does it make sense to print failure hints after setup() failure? I can
>> think of a few cases where bug effects can be delayed until cleanup()
>> but it seems a bit early to me to cover setup() as well. I think the
>> flag could be moved to shared memory, enabled immediately before calling
>> run_tests() and disabled immediately after handling child's exit status
>> in fork_testrun().
> 
> My idea here was that we should always show the hints once we executed a
> test function at least once. The rationale is that if the run_test()
> corrupted kernel memory it may manifest at any point e.g. when we
> attempt to setup a next loop device for the next filesytem (for
> .all_filesystems).
> 
> So the idea is not to print them during the test library setup i.e.
> do_setup() function but enable them right after that and keeping them
> on.

Yes, that makes sense. Though the enablement is still quite deep in 
library setup. You could at least move it immediately before the fork() 
in fork_testrun(), if you don't want to use shared memory.

-- 
Martin Doucha   mdoucha@suse.cz
SW Quality Engineer
SUSE LINUX, s.r.o.
CORSO IIa
Krizikova 148/34
186 00 Prague 8
Czech Republic