On Fri, 5 Jun 2020 13:11:34 -0700 Matthew Wilcox <willy@xxxxxxxxxxxxx> wrote:
> On Fri, Jun 05, 2020 at 03:58:48PM -0400, Qian Cai wrote:
> > This will trigger,
> >
> > [ 8853.759549] LTP: starting semget05
> > [ 8867.257088] BUG: sleeping function called from invalid context at mm/slab.h:567
> > [ 8867.270259] in_atomic(): 1, irqs_disabled(): 0, non_block: 0, pid: 22556, name: semget05
> > [ 8867.270309] 2 locks held by semget05/22556:
> > [ 8867.270345] #0: 00000000512de7e0 (&ids->rwsem){++++}-{3:3}, at: ipcget+0x4e/0x230
> > [ 8867.270426] #1: 00000000552b9018 (&new->lock){+.+.}-{2:2}, at: ipc_addid+0xf4/0xf50
>
> Did the fix for this not make it into -next?
I don't think I've seen this fix before. And I'm unclear on how it
pertains to the current patch(es) in -mm. And it has no changelog!
Perhaps it would be best to do a formal send of the latest version?
