Weird suspend/resume problem on HP nx6325 w/ openSUSE 11.2 w/ -rc8

Weird suspend/resume problem on HP nx6325 w/ openSUSE 11.2 w/ -rc8

[Rafael J. Wysocki]

Hi,

I'm not really sure where this should be reported, but here it goes.

After installing openSUSE 11.2 on my testbed nx6325 I noticed that resume
(from suspend to RAM) stopped working on it.  Apparently, it hanged while
switching from the console back to X and the box remained completely
unresponsive after that (it didn't even react to the magic sysrq).  However,
it used to work just fine with -rc7 and openSUSE 11.1, so I figured the problem
was related to the new X driver, which is radeon module version 6.12.4 (xorg
server 1.6.5).

So, I started to investigate and found that after a suspend from a framebuffer
console, the resume had always worked.  However, when I switched from X to the
framebuffer console, suspended, resumed and switched back to X, the box hanged
immediately just like during resume after a suspend from under X.  [There also
were some problems when switching from X to the framebuffer console and back
which seemd to be independent of suspend/resume (in short, the screen stays
black after a switch from the framebuffer console to X and you need to kill X
to fix this).]

Later, however, I discovered that if the box was booted into runlevel 3, then
suspended to RAM and resumed, and then switched to runlevel 5, I could log
into X and suspend/resume without any problems for many times.  OTOH, after
booting directly into X, the first attempt to resume always resulted in a black
screen and a hard hang.

I wonder what's up in there.

[Hibernation/resume works 100% of the time AFAICS.]

Thanks,
Rafael

Re: Weird suspend/resume problem on HP nx6325 w/ openSUSE 11.2 w/ -rc8

[Justin P. Mattock]

[-- Attachment #1.1: Type: text/plain, Size: 2089 bytes --]

On 11/21/09 15:15, Rafael J. Wysocki wrote:
> Hi,
>
> I'm not really sure where this should be reported, but here it goes.
>
> After installing openSUSE 11.2 on my testbed nx6325 I noticed that resume
> (from suspend to RAM) stopped working on it.  Apparently, it hanged while
> switching from the console back to X and the box remained completely
> unresponsive after that (it didn't even react to the magic sysrq).  However,
> it used to work just fine with -rc7 and openSUSE 11.1, so I figured the problem
> was related to the new X driver, which is radeon module version 6.12.4 (xorg
> server 1.6.5).
>
> So, I started to investigate and found that after a suspend from a framebuffer
> console, the resume had always worked.  However, when I switched from X to the
> framebuffer console, suspended, resumed and switched back to X, the box hanged
> immediately just like during resume after a suspend from under X.  [There also
> were some problems when switching from X to the framebuffer console and back
> which seemd to be independent of suspend/resume (in short, the screen stays
> black after a switch from the framebuffer console to X and you need to kill X
> to fix this).]
>
> Later, however, I discovered that if the box was booted into runlevel 3, then
> suspended to RAM and resumed, and then switched to runlevel 5, I could log
> into X and suspend/resume without any problems for many times.  OTOH, after
> booting directly into X, the first attempt to resume always resulted in a black
> screen and a hard hang.
>
> I wonder what's up in there.
>
> [Hibernation/resume works 100% of the time AFAICS.]
>
> Thanks,
> Rafael
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
>
>    
I just noticed today that on my macbook(radeon) using
s2ram works, but I get a long pause(black screen) before
the screen shows up.(latest stable release works fine).

Justin P. Mattock

[-- Attachment #1.2: Type: text/html, Size: 2607 bytes --]

[-- Attachment #2: Type: text/plain, Size: 0 bytes --]

Re: Weird suspend/resume problem on HP nx6325 w/ openSUSE 11.2 w/ -rc8

[Dave Airlie]

On Sun, Nov 22, 2009 at 9:15 AM, Rafael J. Wysocki <rjw@sisk.pl> wrote:
> Hi,
>
> I'm not really sure where this should be reported, but here it goes.
>
> After installing openSUSE 11.2 on my testbed nx6325 I noticed that resume
> (from suspend to RAM) stopped working on it.  Apparently, it hanged while
> switching from the console back to X and the box remained completely
> unresponsive after that (it didn't even react to the magic sysrq).  However,
> it used to work just fine with -rc7 and openSUSE 11.1, so I figured the problem
> was related to the new X driver, which is radeon module version 6.12.4 (xorg
> server 1.6.5).
>
> So, I started to investigate and found that after a suspend from a framebuffer
> console, the resume had always worked.  However, when I switched from X to the
> framebuffer console, suspended, resumed and switched back to X, the box hanged
> immediately just like during resume after a suspend from under X.  [There also
> were some problems when switching from X to the framebuffer console and back
> which seemd to be independent of suspend/resume (in short, the screen stays
> black after a switch from the framebuffer console to X and you need to kill X
> to fix this).]

You are using radeonfb on it? or vesafb?

first test is to try without either of those installed.

>From memory though that machines an Xpress 200M which is the the worst
graphics card for use with open drivers and has always caused no-end of pain.

Its also mostly undocumented when it comes to errata and workarounds.

Dave.

Re: Weird suspend/resume problem on HP nx6325 w/ openSUSE 11.2 w/ -rc8

[Rafael J. Wysocki]

On Sunday 22 November 2009, Dave Airlie wrote:
> On Sun, Nov 22, 2009 at 9:15 AM, Rafael J. Wysocki <rjw@sisk.pl> wrote:
> > Hi,
> >
> > I'm not really sure where this should be reported, but here it goes.
> >
> > After installing openSUSE 11.2 on my testbed nx6325 I noticed that resume
> > (from suspend to RAM) stopped working on it.  Apparently, it hanged while
> > switching from the console back to X and the box remained completely
> > unresponsive after that (it didn't even react to the magic sysrq).  However,
> > it used to work just fine with -rc7 and openSUSE 11.1, so I figured the problem
> > was related to the new X driver, which is radeon module version 6.12.4 (xorg
> > server 1.6.5).
> >
> > So, I started to investigate and found that after a suspend from a framebuffer
> > console, the resume had always worked.  However, when I switched from X to the
> > framebuffer console, suspended, resumed and switched back to X, the box hanged
> > immediately just like during resume after a suspend from under X.  [There also
> > were some problems when switching from X to the framebuffer console and back
> > which seemd to be independent of suspend/resume (in short, the screen stays
> > black after a switch from the framebuffer console to X and you need to kill X
> > to fix this).]
> 
> You are using radeonfb on it? or vesafb?

vesafb

> first test is to try without either of those installed.

That doesn't change the behavior AFAICS.

> From memory though that machines an Xpress 200M

Yes, it has this one.

> which is the the worst graphics card for use with open drivers and has
> always caused no-end of pain.

Too bad. :-(

> Its also mostly undocumented when it comes to errata and workarounds.

Well, so I guess this is a hardware errata.

Still, it worked with the older X, so I wonder if I can make the new X work
like the old one?

Best,
Rafael

Re: Weird suspend/resume problem on HP nx6325 w/ openSUSE 11.2 w/ -rc8

[Justin Mattock]

On Thu, Nov 26, 2009 at 5:04 PM, Rafael J. Wysocki <rjw@sisk.pl> wrote:
> On Sunday 22 November 2009, Dave Airlie wrote:
>> On Sun, Nov 22, 2009 at 9:15 AM, Rafael J. Wysocki <rjw@sisk.pl> wrote:
>> > Hi,
>> >
>> > I'm not really sure where this should be reported, but here it goes.
>> >
>> > After installing openSUSE 11.2 on my testbed nx6325 I noticed that resume
>> > (from suspend to RAM) stopped working on it.  Apparently, it hanged while
>> > switching from the console back to X and the box remained completely
>> > unresponsive after that (it didn't even react to the magic sysrq).  However,
>> > it used to work just fine with -rc7 and openSUSE 11.1, so I figured the problem
>> > was related to the new X driver, which is radeon module version 6.12.4 (xorg
>> > server 1.6.5).
>> >
>> > So, I started to investigate and found that after a suspend from a framebuffer
>> > console, the resume had always worked.  However, when I switched from X to the
>> > framebuffer console, suspended, resumed and switched back to X, the box hanged
>> > immediately just like during resume after a suspend from under X.  [There also
>> > were some problems when switching from X to the framebuffer console and back
>> > which seemd to be independent of suspend/resume (in short, the screen stays
>> > black after a switch from the framebuffer console to X and you need to kill X
>> > to fix this).]
>>
>> You are using radeonfb on it? or vesafb?
>
> vesafb
>
>> first test is to try without either of those installed.
>
> That doesn't change the behavior AFAICS.
>
>> From memory though that machines an Xpress 200M
>
> Yes, it has this one.
>
>> which is the the worst graphics card for use with open drivers and has
>> always caused no-end of pain.
>
> Too bad. :-(
>
>> Its also mostly undocumented when it comes to errata and workarounds.
>
> Well, so I guess this is a hardware errata.
>
> Still, it worked with the older X, so I wonder if I can make the new X work
> like the old one?
>
> Best,
> Rafael
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
>

>From over here I was running the stable kernel
on a freshly built clfs system, with the latest xorg from git
suspend worked as is. Then loading the latest git kernel
noticed the lagging of the wake up. When I get a chance I'll
do a bisect and see If I come up with anything.


-- 
Justin P. Mattock