On Wed, Apr 19, 2017 at 06:55:30PM -0600, Jens Axboe wrote: > On 04/19/2017 06:44 PM, Ming Lei wrote: > > On Wed, Apr 19, 2017 at 02:17:45PM -0600, Jens Axboe wrote: > >> On 04/15/2017 06:38 AM, Ming Lei wrote: > >>> The 1st patch enhances BLK_MQ_F_NO_SCHED so that we can't change/ > >>> show available io schedulers on devices which don't support io > >>> scheduler. > >>> > >>> The 2nd patch passes BLK_MQ_F_NO_SCHED for avoiding one regression > >>> on mtip32xx, which is introduced by blk-mq io scheduler. > >> > >> I've added 1-2 for 4.11 for now, and then let's plan on getting > >> a real fix done for 4.12. > > > > The real fix has been posted out already: > > > > http://marc.info/?t=149256196700001&r=1&w=2 > > > > Could you consider that for 4.12? > > I didn't particularly like that approach. And we are really late > in the 4.11 cycle, and your first two patches fit the bill for > how we should fix it in the interim. If we don't need to reserve tag for internal command, I am happy to fix it in the interim. However, the mtip32xx driver has to ask blk-mq to reserve the tag zero for internal command. Then if we can't allow the driver to use that reserved tag, it looks quite awkward, doesn't it? > I'll take a closer look for 4.12, I want to get this fixed > for real, obviously. OK, thanks, let's discuss on solutions. Thanks, Ming
