On 08/28/2017 01:56 PM, Sagi Grimberg wrote: >>>> Meh, I don't think that'll work out - the rdma refactoring from >>>> Sagi deeply depends on the unіnit_ctrl split.. >>> >>> I can do a for-4.14/block-postmerge branch, which is for-4.14/block >>> with -rc7 pulled in. Should apply on top of that. >>> >>> Did that branch, doesn't apply. But should be easier for you to >>> rebase on top of that branch. Pushed it out. >> >> I've pushed out a nvme-4.14-rebase branch with the rebase, but I didn't >> have time for non-trivial testing yet, that'll have to wait for >> tomorrow. >> >> Sagi: can you double check the resolutions around blk_mq_reinit_tagse? > > Thanks for doing it Christoph, > > It looks good to me, there is one glitch in patch: > "nvme-rdma: reuse configure/destroy_admin_queue" > > The "nvme_rdma_configure_admin_queue" block needs to > be squashed into "nvme-rdma: don't free tagset on resets> > I'll do that now. Let me know when it's ready. -- Jens Axboe
