Messages in this thread | | | Date | Tue, 7 May 2013 17:18:05 -0400 | Subject | Re: Immediate wakeup after suspend | From | Parag Warudkar <> |
| |
On Mon, May 6, 2013 at 8:57 PM, Parag Warudkar <parag.lkml@gmail.com> wrote: > On Mon, May 6, 2013 at 8:45 PM, Rafael J. Wysocki <rjw@sisk.pl> wrote: >>> I have tried disabling EHC1 and EHC2 without any difference. >> >> That may mean one of two things: (1) After you'd tried that they appeared as >> disabled, but the immediate wakeup still happened or (2) you'd tried to >> disable them, but they still appeared as enabled. >> >> Which one of them is it? >> > > #1 - that is they appeared disabled but the wakeup was still > immediate. I tried that at least twice before sending the initial > email and got the immediate wakeups both times but when I tried to > recreate this right now with the same kernel and now I get a bunch of > WARN_ON()s but no immediate wakeups. (I can still wake it up by > pressing power button.) > I will test more and also bisect to see what is going on.
So on a hunch I removed radeon UVD firmware to disable UVD from initializing and the problem goes away. Turns out, earlier when the problem did not reproduce I was running a kernel with a local patch that disables UVD.
I have tested this 3 times each - with UVD loaded the machine resumes instantly. Without it, it works as expected. I let the radeon/dri people know about it.
Thanks, Parag
| |