PMU_IOC_SLEEP failed

PMU_IOC_SLEEP failed

[jjDaNiMoTh]

Hi all,

what could be the problem here?

[root@myhost ~]# /usr/lib/hal/hal-system-power-pmu sleep
power-pmu : PMU_IOC_SLEEP failed

(same with pm-suspend, kernel version: any >= 2.6.33. I'm now on a .37-rc).

KMS is disabled (modprobe radeon modeset=0, and I can't switch over VT).
userspace thing has kms for radeon disabled (libdrm, mesa, xf86-video-ati).

There are some important parts of config should I attach?

Thanks

Re: PMU_IOC_SLEEP failed

[Michel Dänzer]

On Die, 2010-12-28 at 13:25 +0100, jjDaNiMoTh wrote:=20
>=20
> what could be the problem here?
>=20
> [root@myhost ~]# /usr/lib/hal/hal-system-power-pmu sleep
> power-pmu : PMU_IOC_SLEEP failed
>=20
> (same with pm-suspend, kernel version: any >=3D 2.6.33. I'm now on a .37-=
rc).
>=20
> KMS is disabled (modprobe radeon modeset=3D0, and I can't switch over VT)=
.

What exactly does "can't switch over VT" mean?

Sleep requires radeonfb, is that enabled?


--=20
Earthling Michel D=C3=A4nzer           |                http://www.vmware.c=
om
Libre software enthusiast         |          Debian, X and DRI developer

Re: PMU_IOC_SLEEP failed

[jjDaNiMoTh]

2010/12/28 Michel D=C3=A4nzer <michel@daenzer.net>:
[cut]
> What exactly does "can't switch over VT" mean?
After X is running, I can't switch over ctrl + alt + f{1,2,3,4,5,6}
consoles ( the screen flicker, then blank ). If I switch to f7, where
X live, all return to normal and I can see normal X window.
This doesn't happend with X active.

> Sleep requires radeonfb, is that enabled?

Of course, but I think there is something wrong:

(dmesg..)
[...]
radeonfb_pci_register BEGIN
radeonfb 0000:00:10.0: BAR 0: can't reserve [mem 0xb8000000-0xbfffffff pref=
]
radeonfb (0000:00:10.0): cannot request region 0.
radeonfb: probe of 0000:00:10.0 failed with error -16

Re: PMU_IOC_SLEEP failed

[Benjamin Herrenschmidt]

On Tue, 2010-12-28 at 17:39 +0100, jjDaNiMoTh wrote:
> 2010/12/28 Michel Dänzer <michel@daenzer.net>:
> [cut]
> > What exactly does "can't switch over VT" mean?
> After X is running, I can't switch over ctrl + alt + f{1,2,3,4,5,6}
> consoles ( the screen flicker, then blank ). If I switch to f7, where
> X live, all return to normal and I can see normal X window.
> This doesn't happend with X active.
> 
> > Sleep requires radeonfb, is that enabled?
> 
> Of course, but I think there is something wrong:
> 
> (dmesg..)
> [...]
> radeonfb_pci_register BEGIN
> radeonfb 0000:00:10.0: BAR 0: can't reserve [mem 0xb8000000-0xbfffffff pref]
> radeonfb (0000:00:10.0): cannot request region 0.
> radeonfb: probe of 0000:00:10.0 failed with error -16

That looks bad indeed. Send me the complete dmesg, a snapshot of the
device-tree (tar -c /proc/device-tree >foo.tar) and the output
of /proc/iomem please.

Cheers,
Ben.

Re: PMU_IOC_SLEEP failed

[jjDaNiMoTh]

2010/12/28 Benjamin Herrenschmidt <benh@kernel.crashing.org>:
> On Tue, 2010-12-28 at 17:39 +0100, jjDaNiMoTh wrote:
>> 2010/12/28 Michel D=C3=A4nzer <michel@daenzer.net>:
>> [cut]
>> > What exactly does "can't switch over VT" mean?
>> After X is running, I can't switch over ctrl + alt + f{1,2,3,4,5,6}
>> consoles ( the screen flicker, then blank ). If I switch to f7, where
>> X live, all return to normal and I can see normal X window.
>> This doesn't happend with X active.
>>
>> > Sleep requires radeonfb, is that enabled?
>>
>> Of course, but I think there is something wrong:
>>
>> (dmesg..)
>> [...]
>> radeonfb_pci_register BEGIN
>> radeonfb 0000:00:10.0: BAR 0: can't reserve [mem 0xb8000000-0xbfffffff p=
ref]
>> radeonfb (0000:00:10.0): cannot request region 0.
>> radeonfb: probe of 0000:00:10.0 failed with error -16
>
> That looks bad indeed. Send me the complete dmesg, a snapshot of the
> device-tree (tar -c /proc/device-tree >foo.tar) and the output
> of /proc/iomem please.

[1] http://nat.vacau.com/linux/iomem.txt.gz
[2] http://nat.vacau.com/linux/dmesg_37-rc.txt.gz
[3] http://nat.vacau.com/linux/device_tree.tar

Thank you

Re: PMU_IOC_SLEEP failed

[Benjamin Herrenschmidt]

On Wed, 2010-12-29 at 00:35 +0100, jjDaNiMoTh wrote:
> >> (dmesg..)
> >> [...]
> >> radeonfb_pci_register BEGIN
> >> radeonfb 0000:00:10.0: BAR 0: can't reserve [mem
> 0xb8000000-0xbfffffff pref]
> >> radeonfb (0000:00:10.0): cannot request region 0.
> >> radeonfb: probe of 0000:00:10.0 failed with error -16
> >
> > That looks bad indeed. Send me the complete dmesg, a snapshot of the
> > device-tree (tar -c /proc/device-tree >foo.tar) and the output
> > of /proc/iomem please.
> 
> [1] http://nat.vacau.com/linux/iomem.txt.gz
> [2] http://nat.vacau.com/linux/dmesg_37-rc.txt.gz
> [3] http://nat.vacau.com/linux/device_tree.tar 

Are you trying to use radeonfb as a module ? The dmesg seems to indicate
that you have offb taking control of the framebuffer, preventing
radeonfb from loading.

There's no offb->radeonfb handover (the handover mechanism is new and
only used by KMS for now). On those machines, radeonfb should be
built-in.

Cheers,
Ben.