Re: suspend/resume fails on second attempt in LNXVIDEO:00

Re: suspend/resume fails on second attempt in LNXVIDEO:00

[Brian J. Murrell]

On Mon, 15 Dec 2008 16:56:55 +0100, Pavel Machek wrote:
 
> Sounds like acpi video driver to me. So you may want to rmmod it before
> suspend at see what happens. (and cc linux-acpi).

Please forgive my ignorance.  Which driver would that be?  I currently
have:

$ lsmod
Module                  Size  Used by
af_packet              25088  2 
i915                   64004  2 
drm                    93736  3 i915
rfcomm                 41744  2 
l2cap                  29312  13 rfcomm
bluetooth              62436  4 rfcomm,l2cap
ppdev                  15620  0 
ipv6                  254836  24 
acpi_cpufreq           16268  1 
cpufreq_conservative    14472  0 
cpufreq_powersave       9856  0 
cpufreq_ondemand       15244  1 
cpufreq_userspace      11268  0 
cpufreq_stats          13188  0 
freq_table             12672  3 acpi_cpufreq,cpufreq_ondemand,cpufreq_stats
container              11648  0 
pci_slot               12680  0 
sbs                    18952  0 
sbshc                  13568  1 sbs
iptable_filter         11008  0 
ip_tables              19216  1 iptable_filter
x_tables               22916  1 ip_tables
parport_pc             40356  0 
lp                     17156  0 
parport                42348  3 ppdev,parport_pc,lp
joydev                 17856  0 
snd_hda_intel         416052  0 
snd_seq_dummy          11012  0 
snd_pcsp               17832  0 
snd_seq_oss            37248  0 
snd_pcm_oss            44928  0 
snd_mixer_oss          22272  1 snd_pcm_oss
psmouse                49168  0 
snd_seq_midi           14464  0 
snd_rawmidi            29184  1 snd_seq_midi
snd_seq_midi_event     15232  2 snd_seq_oss,snd_seq_midi
evdev                  17696  6 
snd_seq                55984  6 snd_seq_dummy,snd_seq_oss,snd_seq_midi,snd_seq_midi_event
snd_pcm                81156  3 snd_hda_intel,snd_pcsp,snd_pcm_oss
serio_raw              13444  0 
snd_timer              28808  2 snd_seq,snd_pcm
snd_seq_device         15244  5 snd_seq_dummy,snd_seq_oss,snd_seq_midi,snd_rawmidi,snd_seq
iTCO_wdt               19236  0 
iTCO_vendor_support    11908  1 iTCO_wdt
video                  25360  0 
output                 11136  1 video
snd                    61476  10 snd_hda_intel,snd_pcsp,snd_seq_oss,snd_pcm_oss,snd_mixer_oss,snd_rawmidi,snd_seq,snd_pcm,snd_timer,snd_seq_device
soundcore              15328  1 snd
ac                     12420  0 
snd_page_alloc         17160  2 snd_hda_intel,snd_pcm
wmi                    14888  0 
battery                18436  0 
shpchp                 39828  0 
pci_hotplug            34976  1 shpchp
intel_agp              33596  1 
agpgart                41928  3 drm,intel_agp
button                 14480  0 
ext3                  130568  1 
jbd                    53908  1 ext3
mbcache                16004  1 ext3
loop                   22540  2 
usb_storage            78656  0 
libusual               30356  1 usb_storage
sd_mod                 41240  2 
crc_t10dif             10112  1 sd_mod
sr_mod                 21956  0 
cdrom                  42272  1 sr_mod
sg                     36148  0 
ahci                   37260  1 
ata_piix               29700  0 
pata_acpi              12288  0 
ata_generic            13060  0 
8139too                31744  0 
8139cp                 27776  0 
mii                    13440  2 8139too,8139cp
libata                176160  4 ahci,ata_piix,pata_acpi,ata_generic
scsi_mod              156948  5 usb_storage,sd_mod,sr_mod,sg,libata
ehci_hcd               41996  0 
uhci_hcd               30352  0 
usbcore               149392  5 usb_storage,libusual,ehci_hcd,uhci_hcd
thermal                23708  0 
processor              49708  4 acpi_cpufreq,thermal
fan                    12676  0 
fuse                   58780  3 

Do you mean the i915 driver?  I don't think I will be able to rmmod that
one given the number of references on it.

Thanx!

b.

Re: suspend/resume fails on second attempt in LNXVIDEO:00

[Alexey Starikovskiy]

Brian J. Murrell wrote:
> On Mon, 15 Dec 2008 16:56:55 +0100, Pavel Machek wrote:
>  
>   
>> Sounds like acpi video driver to me. So you may want to rmmod it before
>> suspend at see what happens. (and cc linux-acpi).
>>     
>
> Please forgive my ignorance.  Which driver would that be?  I currently
> have:
>
> $ lsmod
> Module                  Size  Used by
> af_packet              25088  2 
> i915                   64004  2 
> drm                    93736  3 i915
> rfcomm                 41744  2 
> l2cap                  29312  13 rfcomm
> bluetooth              62436  4 rfcomm,l2cap
> ppdev                  15620  0 
> ipv6                  254836  24 
> acpi_cpufreq           16268  1 
> cpufreq_conservative    14472  0 
> cpufreq_powersave       9856  0 
> cpufreq_ondemand       15244  1 
> cpufreq_userspace      11268  0 
> cpufreq_stats          13188  0 
> freq_table             12672  3 acpi_cpufreq,cpufreq_ondemand,cpufreq_stats
> container              11648  0 
> pci_slot               12680  0 
> sbs                    18952  0 
> sbshc                  13568  1 sbs
> iptable_filter         11008  0 
> ip_tables              19216  1 iptable_filter
> x_tables               22916  1 ip_tables
> parport_pc             40356  0 
> lp                     17156  0 
> parport                42348  3 ppdev,parport_pc,lp
> joydev                 17856  0 
> snd_hda_intel         416052  0 
> snd_seq_dummy          11012  0 
> snd_pcsp               17832  0 
> snd_seq_oss            37248  0 
> snd_pcm_oss            44928  0 
> snd_mixer_oss          22272  1 snd_pcm_oss
> psmouse                49168  0 
> snd_seq_midi           14464  0 
> snd_rawmidi            29184  1 snd_seq_midi
> snd_seq_midi_event     15232  2 snd_seq_oss,snd_seq_midi
> evdev                  17696  6 
> snd_seq                55984  6 snd_seq_dummy,snd_seq_oss,snd_seq_midi,snd_seq_midi_event
> snd_pcm                81156  3 snd_hda_intel,snd_pcsp,snd_pcm_oss
> serio_raw              13444  0 
> snd_timer              28808  2 snd_seq,snd_pcm
> snd_seq_device         15244  5 snd_seq_dummy,snd_seq_oss,snd_seq_midi,snd_rawmidi,snd_seq
> iTCO_wdt               19236  0 
> iTCO_vendor_support    11908  1 iTCO_wdt
> video                  25360  0 
>   
this one ^^^^^
> output                 11136  1 video
> snd                    61476  10 snd_hda_intel,snd_pcsp,snd_seq_oss,snd_pcm_oss,snd_mixer_oss,snd_rawmidi,snd_seq,snd_pcm,snd_timer,snd_seq_device
> soundcore              15328  1 snd
> ac                     12420  0 
> snd_page_alloc         17160  2 snd_hda_intel,snd_pcm
> wmi                    14888  0 
> battery                18436  0 
> shpchp                 39828  0 
> pci_hotplug            34976  1 shpchp
> intel_agp              33596  1 
> agpgart                41928  3 drm,intel_agp
> button                 14480  0 
> ext3                  130568  1 
> jbd                    53908  1 ext3
> mbcache                16004  1 ext3
> loop                   22540  2 
> usb_storage            78656  0 
> libusual               30356  1 usb_storage
> sd_mod                 41240  2 
> crc_t10dif             10112  1 sd_mod
> sr_mod                 21956  0 
> cdrom                  42272  1 sr_mod
> sg                     36148  0 
> ahci                   37260  1 
> ata_piix               29700  0 
> pata_acpi              12288  0 
> ata_generic            13060  0 
> 8139too                31744  0 
> 8139cp                 27776  0 
> mii                    13440  2 8139too,8139cp
> libata                176160  4 ahci,ata_piix,pata_acpi,ata_generic
> scsi_mod              156948  5 usb_storage,sd_mod,sr_mod,sg,libata
> ehci_hcd               41996  0 
> uhci_hcd               30352  0 
> usbcore               149392  5 usb_storage,libusual,ehci_hcd,uhci_hcd
> thermal                23708  0 
> processor              49708  4 acpi_cpufreq,thermal
> fan                    12676  0 
> fuse                   58780  3 
>
> Do you mean the i915 driver?  I don't think I will be able to rmmod that
> one given the number of references on it.
>
> Thanx!
>
> b.
>
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>   

Re: suspend/resume fails on second attempt in LNXVIDEO:00

[Brian J. Murrell]

On Mon, 15 Dec 2008 20:33:44 +0300, Alexey Starikovskiy wrote:

>> video                  25360  0
>>   
> this one ^^^^^

Ahhh.  Got it.

Unfortunately that was of no help.  Still only a single suspend/resume 
cycle worked.  I removed it before the first cycle, (and made sure it was 
not loaded before the second) for the record.

b.

Re: suspend/resume fails on second attempt in LNXVIDEO:00

[Rafael J. Wysocki]

On Monday, 15 of December 2008, Brian J. Murrell wrote:
> On Mon, 15 Dec 2008 20:33:44 +0300, Alexey Starikovskiy wrote:
> 
> >> video                  25360  0
> >>   
> > this one ^^^^^
> 
> Ahhh.  Got it.
> 
> Unfortunately that was of no help.  Still only a single suspend/resume 
> cycle worked.  I removed it before the first cycle, (and made sure it was 
> not loaded before the second) for the record.

Does it also happen in the minimal configuration (ie. with init=/bin/bash)?

Rafael

Re: suspend/resume fails on second attempt in LNXVIDEO:00

[Brian J. Murrell]

On Mon, 15 Dec 2008 21:19:34 +0100, Rafael J. Wysocki wrote:
> Does it also happen in the minimal configuration (ie. with
> init=/bin/bash)?

Now, that's an interesting result.  Using the above, pm-suspend does 
suspend fine and hitting the flashing power button will resume fine with 
the exception of the screen is still black.  That is backlight problems 
isn't it?  I can see by disk light activity though that it's actually 
still alive though.

b.

Re: suspend/resume fails on second attempt in LNXVIDEO:00

[Rafael J. Wysocki]

On Monday, 15 of December 2008, Brian J. Murrell wrote:
> On Mon, 15 Dec 2008 21:19:34 +0100, Rafael J. Wysocki wrote:
> > Does it also happen in the minimal configuration (ie. with
> > init=/bin/bash)?
> 
> Now, that's an interesting result.  Using the above, pm-suspend does 
> suspend fine and hitting the flashing power button will resume fine with 
> the exception of the screen is still black.  That is backlight problems 
> isn't it?  I can see by disk light activity though that it's actually 
> still alive though.

No, most probably your graphics adapter is not handled correctly in this case,
because it needs some special user space quirks, which are done by hald that is
not running.

You can try to use the s2ram binary (http://en.opensuse.org/s2ram) to work
around this issue.

Anyway, if you suspend for the second time from this state, does it resume?

Rafael

Re: suspend/resume fails on second attempt in LNXVIDEO:00

[Brian J. Murrell]

On Tue, 16 Dec 2008 00:02:43 +0100, Rafael J. Wysocki wrote:
> 
> No, most probably your graphics adapter is not handled correctly in this
> case, because it needs some special user space quirks, which are done by
> hald that is not running.

Hrm.  OK.
 
> You can try to use the s2ram binary (http://en.opensuse.org/s2ram) to
> work around this issue.

That sounds like I am setting for something less than optimal.  Is that 
the case?  Why would I want to use what I've been trying to get working 
rather than s2ram?  Maybe put another way, why isn't everyone using s2ram?

Ultimately I'm going to want to be able to suspend as well as hibernate, 
for whatever that's worth.

> Anyway, if you suspend for the second time from this state, does it
> resume?

No.  It fails just as miserably.  :-(

b.

Re: suspend/resume fails on second attempt in LNXVIDEO:00

[Rafael J. Wysocki]

On Tuesday, 16 of December 2008, Brian J. Murrell wrote:
> On Tue, 16 Dec 2008 00:02:43 +0100, Rafael J. Wysocki wrote:
> > 
> > No, most probably your graphics adapter is not handled correctly in this
> > case, because it needs some special user space quirks, which are done by
> > hald that is not running.
> 
> Hrm.  OK.
>  
> > You can try to use the s2ram binary (http://en.opensuse.org/s2ram) to
> > work around this issue.
> 
> That sounds like I am setting for something less than optimal.  Is that 
> the case?  Why would I want to use what I've been trying to get working 
> rather than s2ram?  Maybe put another way, why isn't everyone using s2ram?

Because the hal-driven whitelist allows us to match systems in a more detailed
way.  If your system is in the s2ram whitelist already and works with it, there
shouldn't be any difference.

Still, I was only considering that as a debugging aid in your case, because
s2ram works in the minimal configuration, while hald doesn't.

> Ultimately I'm going to want to be able to suspend as well as hibernate, 
> for whatever that's worth.
> 
> > Anyway, if you suspend for the second time from this state, does it
> > resume?
> 
> No.  It fails just as miserably.  :-(

That's a bit of new information.

What's the list of modules loaded in the minimal configuration?

Rafael

Re: suspend/resume fails on second attempt in LNXVIDEO:00

[Brian J. Murrell]

On Tue, 16 Dec 2008 16:25:37 +0100, Rafael J. Wysocki wrote:

> Because the hal-driven whitelist allows us to match systems in a more
> detailed way.  If your system is in the s2ram whitelist already and
> works with it, there shouldn't be any difference.

Unfortunately, Ubuntu in their "wisdom" removed s2ram and claim "s2both" 
supersedes it in that it does both suspend to disk and suspend to ram and 
then suspends, but if your battery dies while suspended you can still 
resume from the suspend to disk.
 
> Still, I was only considering that as a debugging aid in your case,
> because s2ram works in the minimal configuration, while hald doesn't.

Right.  I will give s2both a try and see what happens... no joy.  It 
fails to even run without the swap device mounted, and my swap device is 
not big enough to hibernate to anyway.  Damn you Ubuntu.

> That's a bit of new information.
> 
> What's the list of modules loaded in the minimal configuration?

Module                  Size  Used by
ext3                  130568  1 
jbd                    53908  1 ext3
mbcache                16004  1 ext3
loop                   22540  2 
usb_storage            78656  0 
libusual               30356  1 usb_storage
sd_mod                 41240  2 
crc_t10dif             10112  1 sd_mod
sr_mod                 21956  0 
cdrom                  42272  1 sr_mod
sg                     36148  0 
ahci                   37260  1 
ata_piix               29700  0 
pata_acpi              12288  0 
ata_generic            13060  0 
libata                176160  4 ahci,ata_piix,pata_acpi,ata_generic
scsi_mod              156948  5 usb_storage,sd_mod,sr_mod,sg,libata
8139too                31744  0 
8139cp                 27776  0 
mii                    13440  2 8139too,8139cp
ehci_hcd               41996  0 
uhci_hcd               30352  0 
usbcore               149392  5 usb_storage,libusual,ehci_hcd,uhci_hcd
thermal                23708  0 
processor              49708  3 thermal
fan                    12676  0 
fuse                   58780  3 

Cheers,
b.

Re: suspend/resume fails on second attempt in LNXVIDEO:00

[Rafael J. Wysocki]

On Tuesday, 16 of December 2008, Brian J. Murrell wrote:
> On Tue, 16 Dec 2008 16:25:37 +0100, Rafael J. Wysocki wrote:
> 
> > Because the hal-driven whitelist allows us to match systems in a more
> > detailed way.  If your system is in the s2ram whitelist already and
> > works with it, there shouldn't be any difference.
> 
> Unfortunately, Ubuntu in their "wisdom" removed s2ram and claim "s2both" 
> supersedes it in that it does both suspend to disk and suspend to ram and 
> then suspends, but if your battery dies while suspended you can still 
> resume from the suspend to disk.

Hm, this really isn't reasonable.  s2both saves the image, which takes quite
a lot of time, while s2ram doesn't.  Also, s2both is not really ACPI-compliant.

> > Still, I was only considering that as a debugging aid in your case,
> > because s2ram works in the minimal configuration, while hald doesn't.
> 
> Right.  I will give s2both a try and see what happens... no joy.  It 
> fails to even run without the swap device mounted, and my swap device is 
> not big enough to hibernate to anyway.  Damn you Ubuntu.

You can compile s2ram from sources, it's not too difficult.

> > That's a bit of new information.
> > 
> > What's the list of modules loaded in the minimal configuration?
> 
> Module                  Size  Used by
> ext3                  130568  1 
> jbd                    53908  1 ext3
> mbcache                16004  1 ext3
> loop                   22540  2 
> usb_storage            78656  0 
> libusual               30356  1 usb_storage
> sd_mod                 41240  2 
> crc_t10dif             10112  1 sd_mod
> sr_mod                 21956  0 
> cdrom                  42272  1 sr_mod
> sg                     36148  0 
> ahci                   37260  1 
> ata_piix               29700  0 
> pata_acpi              12288  0 
> ata_generic            13060  0 
> libata                176160  4 ahci,ata_piix,pata_acpi,ata_generic
> scsi_mod              156948  5 usb_storage,sd_mod,sr_mod,sg,libata
> 8139too                31744  0 
> 8139cp                 27776  0 
> mii                    13440  2 8139too,8139cp
> ehci_hcd               41996  0 
> uhci_hcd               30352  0 
> usbcore               149392  5 usb_storage,libusual,ehci_hcd,uhci_hcd
> thermal                23708  0 
> processor              49708  3 thermal
> fan                    12676  0 
> fuse                   58780  3 

Hm.  Is that all loaded when you boot with init=/bin/bash (by which I mean the
minimal config)?

Rafael

Re: suspend/resume fails on second attempt in LNXVIDEO:00

[Brian J. Murrell]

On Tue, 16 Dec 2008 17:13:08 +0100, Rafael J. Wysocki wrote:
> 
> Hm, this really isn't reasonable.

Many people agree with you.  :-)  There is a long bug open with ubuntu to 
restore this s2ram.  They seem to be ignoring it though.

> s2both saves the image, which takes
> quite a lot of time, while s2ram doesn't.

Indeed.
 
> You can compile s2ram from sources, it's not too difficult.

Yeah.  I will probably just grab the debian package and dpkg-buildpackage 
it on my Ubuntu system.
 
> Hm.  Is that all loaded when you boot with init=/bin/bash (by which I
> mean the minimal config)?

It sure is.  Most assuredly by the initrd.  Debian/Ubuntu systems seem to 
put everything but the kitchen sink into their initrds.  Which on the one 
hand makes them pretty portable and resilient, but on the other, yes, 
bloated.

b.

Re: suspend/resume fails on second attempt in LNXVIDEO:00

[Rafael J. Wysocki]

On Tuesday, 16 of December 2008, Brian J. Murrell wrote:
> On Tue, 16 Dec 2008 17:13:08 +0100, Rafael J. Wysocki wrote:
> > 
> > Hm, this really isn't reasonable.
> 
> Many people agree with you.  :-)  There is a long bug open with ubuntu to 
> restore this s2ram.  They seem to be ignoring it though.
> 
> > s2both saves the image, which takes
> > quite a lot of time, while s2ram doesn't.
> 
> Indeed.
>  
> > You can compile s2ram from sources, it's not too difficult.
> 
> Yeah.  I will probably just grab the debian package and dpkg-buildpackage 
> it on my Ubuntu system.
>  
> > Hm.  Is that all loaded when you boot with init=/bin/bash (by which I
> > mean the minimal config)?
> 
> It sure is.  Most assuredly by the initrd.  Debian/Ubuntu systems seem to 
> put everything but the kitchen sink into their initrds.  Which on the one 
> hand makes them pretty portable and resilient, but on the other, yes, 
> bloated.

Well, please remove as many modules as you can (using rmmod) and see if your
second resume still fails.

Thanks,
Rafael

Re: suspend/resume fails on second attempt in LNXVIDEO:00

[Brian J. Murrell]

On Tue, 16 Dec 2008 17:22:40 +0100, Rafael J. Wysocki wrote:
> 
> Well, please remove as many modules as you can (using rmmod) and see if
> your second resume still fails.

OK.  I got the list of modules down to:

Module                  Size  Used by
ipv6                  254836  14 
ext3                  130568  1 
jbd                    53908  1 ext3
mbcache                16004  1 ext3
loop                   22540  2 
sd_mod                 41240  2 
crc_t10dif             10112  1 sd_mod
ahci                   37260  1 
libata                176160  1 ahci
scsi_mod              156948  2 sd_mod,libata
processor              49708  2 
fuse                   58780  3 

and that still failed the second resume.  Just to remind you, the first 
suspend/resume (echo mem >/sys/power/state) works, but the screen is 
black on returning from the resume.  I can then (blindly typing) do a
second suspend but resume still hangs the machine.

b.

Re: suspend/resume fails on second attempt in LNXVIDEO:00

[Matthew Garrett]

On Tue, Dec 16, 2008 at 05:13:08PM +0100, Rafael J. Wysocki wrote:
> On Tuesday, 16 of December 2008, Brian J. Murrell wrote:
> > On Tue, 16 Dec 2008 16:25:37 +0100, Rafael J. Wysocki wrote:
> > 
> > > Because the hal-driven whitelist allows us to match systems in a more
> > > detailed way.  If your system is in the s2ram whitelist already and
> > > works with it, there shouldn't be any difference.
> > 
> > Unfortunately, Ubuntu in their "wisdom" removed s2ram and claim "s2both" 
> > supersedes it in that it does both suspend to disk and suspend to ram and 
> > then suspends, but if your battery dies while suspended you can still 
> > resume from the suspend to disk.

I seem to be missing the original message, but: no, s2ram doesn't exist 
in Ubuntu because pm-utils handles that role.

-- 
Matthew Garrett | mjg59@srcf.ucam.org

Re: suspend/resume fails on second attempt in LNXVIDEO:00

[Brian J. Murrell]

On Tue, 16 Dec 2008 16:37:46 +0000, Matthew Garrett wrote:
> 
> I seem to be missing the original message, but: no, s2ram doesn't exist
> in Ubuntu because pm-utils handles that role.

Well, that's *their* argument but lots of people think otherwise.  There 
appears to be no shortage of people for whom pm-utils does NOT work yet 
s2ram does.  https://bugs.launchpad.net/ubuntu/+source/uswsusp/+bug/134238

There is also indication earlier in this thread that it may work for me 
where pm-utils is not working.

b.

Re: suspend/resume fails on second attempt in LNXVIDEO:00

[Matthew Garrett]

On Tue, Dec 16, 2008 at 06:07:15PM +0000, Brian J. Murrell wrote:
> On Tue, 16 Dec 2008 16:37:46 +0000, Matthew Garrett wrote:
> > 
> > I seem to be missing the original message, but: no, s2ram doesn't exist
> > in Ubuntu because pm-utils handles that role.
> 
> Well, that's *their* argument but lots of people think otherwise.  There 
> appears to be no shortage of people for whom pm-utils does NOT work yet 
> s2ram does.  https://bugs.launchpad.net/ubuntu/+source/uswsusp/+bug/134238

The right answer to "This piece of software contains bugs" is not 
"Provide two pieces of software with the same features but different 
bugs". The only real functional difference between the two is that s2ram 
has its own internal whitelist rather than just using the hal entries. 
Suse have normally been good at making sure that they're synced across, 
but if that's not happening you probably want to poke them.

-- 
Matthew Garrett | mjg59@srcf.ucam.org

Re: suspend/resume fails on second attempt in LNXVIDEO:00

[Pavel Machek]

On Tue 2008-12-16 20:53:35, Matthew Garrett wrote:
> On Tue, Dec 16, 2008 at 06:07:15PM +0000, Brian J. Murrell wrote:
> > On Tue, 16 Dec 2008 16:37:46 +0000, Matthew Garrett wrote:
> > > 
> > > I seem to be missing the original message, but: no, s2ram doesn't exist
> > > in Ubuntu because pm-utils handles that role.
> > 
> > Well, that's *their* argument but lots of people think otherwise.  There 
> > appears to be no shortage of people for whom pm-utils does NOT work yet 
> > s2ram does.  https://bugs.launchpad.net/ubuntu/+source/uswsusp/+bug/134238
> 
> The right answer to "This piece of software contains bugs" is not 
> "Provide two pieces of software with the same features but different 
> bugs". The only real functional difference between the two is that

Yes, and that's why pm-utils should die: they have design problems
(depend on hal, can't be pagelocked, unusable for suspend debugging).
									Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

Re: suspend/resume fails on second attempt in LNXVIDEO:00

[Justin P. Mattock]

Pavel Machek wrote:
> On Tue 2008-12-16 20:53:35, Matthew Garrett wrote:
>   
>> On Tue, Dec 16, 2008 at 06:07:15PM +0000, Brian J. Murrell wrote:
>>     
>>> On Tue, 16 Dec 2008 16:37:46 +0000, Matthew Garrett wrote:
>>>       
>>>> I seem to be missing the original message, but: no, s2ram doesn't exist
>>>> in Ubuntu because pm-utils handles that role.
>>>>         
>>> Well, that's *their* argument but lots of people think otherwise.  There 
>>> appears to be no shortage of people for whom pm-utils does NOT work yet 
>>> s2ram does.  https://bugs.launchpad.net/ubuntu/+source/uswsusp/+bug/134238
>>>       
>> The right answer to "This piece of software contains bugs" is not 
>> "Provide two pieces of software with the same features but different 
>> bugs". The only real functional difference between the two is that
>>     
>
> Yes, and that's why pm-utils should die: they have design problems
> (depend on hal, can't be pagelocked, unusable for suspend debugging).
> 									Pavel
>   
When using ubuntu intrepid, s2ram
did work, but instead of grabbing
the package from ubuntu, I used Debian
SID instead. (I have a tendency of mixing
packages);

regards;

Justin P. Mattock

Re: suspend/resume fails on second attempt in LNXVIDEO:00

[Brian J. Murrell]

On Tue, 16 Dec 2008 00:02:43 +0100, Rafael J. Wysocki wrote:
> 
> You can try to use the s2ram binary (http://en.opensuse.org/s2ram) to
> work around this issue.

OK.  I've done my s2ram experiments and go no better than with pm-utils/
kernel mode suspend (where first suspend/resume works, second suspend 
works, but second resume hangs the machine).  Here's the full array of 
suggestions on http://en.opensuse.org/S2ram and their results:

s2ram -f -a 3
 - hung on first resume

s2ram -f -a 2
 - restored first resume but no backlight
s2ram -f -a 1
 - hung on first resume

s2ram -f -p -m 
 - resumed first time but console font changed
 - hung on second resume

s2ram -f -p -s 
 - resumed first time but console font changed
 - hung on second resume

s2ram -f -m
 - restored first resume but no backlight

s2ram -f -s
 - resumed first time but console font changed
    - message "Function not supported?\nSwitching back to vt1"
 - hung on second resume

s2ram -f -p
 - resumed first time but console font changed
 - hung on second resume

s2ram -f -a 2 -m
 - restored first resume but no backlight

s2ram -f -a 2 -s
 - resumed first time but console font changed
    - message "Function not supported?\nSwitching back to vt1"

These were always booting with init=/bin/bash and removing all removable 
(i.e. unreferenced) modules.

Where I mention "console font changed" that means it reverted back to the 
font that it was in when it was booting the kernel.  You see, Ubuntu, at 
some point through boot actually change the font to be a "leaner" (i.e. 
not so "fat") font.  This is not to say that I am running on the console 
in a framebuffer (to the best of my knowledge).  I believe this font 
setting is being done with a "/usr/bin/consolechars" tool.

Anyway, thanx much for continuing to look into this.

Cheers,
b.

Re: suspend/resume fails on second attempt in LNXVIDEO:00

[Brian J. Murrell]

On Thu, 18 Dec 2008 04:24:44 +0000, Brian J. Murrell wrote:

> 
> OK.  I've done my s2ram experiments and go no better than with pm-utils/
> kernel mode suspend (where first suspend/resume works, second suspend
> works, but second resume hangs the machine).  Here's the full array of
> suggestions on http://en.opensuse.org/S2ram and their results:

Just wondering if any of this experimentation is likely to help me or is 
this just a loser of a laptop that will be more trouble that it's worth 
trying to get to suspend?  I can take it back to the store still.  Maybe 
I should and try a different one?

Or does it seem like there still might be a relatively easy fix for this 
problem, which would be nice given that it was a pretty good deal (or was 
it?).  :-)

b.

Re: suspend/resume fails on second attempt in LNXVIDEO:00

[Rafael J. Wysocki]

On Friday, 19 of December 2008, Brian J. Murrell wrote:
> On Thu, 18 Dec 2008 04:24:44 +0000, Brian J. Murrell wrote:
> 
> > 
> > OK.  I've done my s2ram experiments and go no better than with pm-utils/
> > kernel mode suspend (where first suspend/resume works, second suspend
> > works, but second resume hangs the machine).  Here's the full array of
> > suggestions on http://en.opensuse.org/S2ram and their results:
> 
> Just wondering if any of this experimentation is likely to help me or is 
> this just a loser of a laptop that will be more trouble that it's worth 
> trying to get to suspend?  I can take it back to the store still.  Maybe 
> I should and try a different one?
> 
> Or does it seem like there still might be a relatively easy fix for this 
> problem, which would be nice given that it was a pretty good deal (or was 
> it?).  :-)

You're not the only person having a problem like this and unfortunately it's
not easy to debug.  I don't know of any fixes available.

Thanks,
Rafael

Re: suspend/resume fails on second attempt in LNXVIDEO:00

[Brian J. Murrell]

On Fri, 19 Dec 2008 18:33:10 +0100, Rafael J. Wysocki wrote:
> 
> You're not the only person having a problem like this

Indeed!

> and unfortunately
> it's not easy to debug.

It seems so, yes.  :-(

> I don't know of any fixes available.

So I should probably stop holding my breath on this laptop and just go 
get a different one I guess.  :-(  Too bad.  This is definitely a better 
laptop than I can buy otherwise for that money.

Maybe my wife will be happy to give up suspending for the better 
machine.  Suspend is very nice to have though.

If there is anything at all I can do to help progress this along, I'd be 
more than happy.  I've already built a kernel from source with the power 
management verbose debugging.  I can apply patches and test easily enough 
if that helps in any way.

I certainly appreciate the effort you and everyone has already put into 
this and if goes no further than this, I certainly cannot complain that 
nobody tried to help.

Thanx!

b.

Re: suspend/resume fails on second attempt in LNXVIDEO:00

[Sergio Monteiro Basto]

[-- Attachment #1: Type: text/plain, Size: 444 bytes --]

On Mon, 2008-12-15 at 17:00 +0000, Brian J. Murrell wrote:
> i915                   64004  2 
> drm                    93736  3 i915

Well , I am thinking if you have a intel-video , and yes ! , I am
getting in a same kind of trouble, I bet that is a problem with i915
drm , mesa 7.2 and xorg-x11-drv-intel (or i810) .
Maybe you can try resume/suspend using vesa video drive, or without
X.org running. 

Regards,
-- 
Sérgio M. B.

[-- Attachment #2: smime.p7s --]
[-- Type: application/x-pkcs7-signature, Size: 2192 bytes --]

Re: suspend/resume fails on second attempt in LNXVIDEO:00

[Brian J. Murrell]

On Sat, 20 Dec 2008 17:43:44 +0000, Sergio Monteiro Basto wrote:
> 
> Well , I am thinking if you have a intel-video , and yes ! , I am
> getting in a same kind of trouble, I bet that is a problem with i915 drm
> , mesa 7.2 and xorg-x11-drv-intel (or i810) .

Not likely.  I unload all of that before I do my tests.  My tests are 
nothing more running that a shell (literally, that's the only process) 
and I have my module stack down to about half a dozen modules.

> Maybe you can try
> resume/suspend using vesa video drive, or without X.org running.

Not a solution.  See above.

b.