Hi. So, current summary:
1) first patch + debug patch: I can reproduce the issue in wild, but not with pm_test set; 2) first patch + debug patch + second patch: I cannot reproduce issue at all, neither with "none", nor with "mq-deadline".
Thus, "blk-mq: align to legacy path for implementing blk_execute_rq" + "blk-mq: add requests in the tail of hctx->dispatch" look like a proper fix. Thanks for that!
I setup dm-crypt on /dev/md0 directly, could you show me how to setup lvm on raid10?
If still relevant, here is a layout: === sda 8:0 0 8G 0 disk |-sda1 8:1 0 128M 0 part /boot/EFI `-sda2 8:2 0 7.9G 0 part `-md0 9:0 0 7.9G 0 raid10 `-system-root 253:0 0 7.9G 0 lvm / sdb 8:16 0 8G 0 disk |-sdb1 8:17 0 128M 0 part `-sdb2 8:18 0 7.9G 0 part `-md0 9:0 0 7.9G 0 raid10 `-system-root 253:0 0 7.9G 0 lvm / === Anything else you'd like /me to test? 30.08.2017 10:06, Ming Lei написав:
…SNIP… Please try the following patch and previous patch together and see if they make a difference: …SNIP…