Comment 35 for bug 1776887

That's not enough information. At this point you should report a new bug, please.

Notice the linked bug #1760450 includes a crash trace. (Not saying for sure you'd be able to do the same). The fatal signal is SIGBUS, as opposed to the usual SIGSEGV or abort. If you've got a SIGSEGV or an abort instead, you don't have this bug.

If you do, you can link to this bug for the historical reminder. (The analysis of these crash traces is a bit unusual).

I'd be pretty surprised if there was a regression on this. Axboe demanded a regression test for blktests, so it can't be exactly the same. And the cause of the bug was something pretty silly during the transition to blk-mq... I don't know why it would get introduced elsewhere.