On Thu, May 23, 2024 at 07:05:29PM +0200, Christoph Hellwig wrote:

On Thu, May 23, 2024 at 12:38:24PM -0400, Mike Snitzer wrote:

quoted

Happy to see mptest get folded into blktests (its just bash code) --
but it doesn't reproduce for you so not a reliable safeguard.

It is a lot better than not running it.  And I'll look into why
it doesn't reproduce.  Right now the only thing I can think of is
different kernel configs, maybe related to schedulers.  Can you send
me your .config?

Will do off-list.

Is adding mptests something you want to do, or you'd prefer outhers
to take care of?

I won't have time in the near or mid-term.  So if someone else would
like to convert mptests over to blktests that'd be wonderful.

Fanning out to test the various supported test permutations would be
cool (meaning, test with both MULTIPATH_BACKEND_MODULE=scsidebug and
MULTIPATH_BACKEND_MODULE=tcmloop ... but tcmloop is brittle due to
sysfs changes vs targetcli's sysfs expectations -- but that's a
targetcli issue that might have been fixed, assuming that code is
supported still?  Not revisited in a few months)