Fwd: Re: [Suspend-devel] s2both on powerpc

Fwd: Re: [Suspend-devel] s2both on powerpc

[Rafael J. Wysocki]

Alan, Dave, could you please have a look at this?

Rafael


----------  Forwarded Message  ----------

Subject: Re: [Suspend-devel] s2both on powerpc
Date: Sunday, 1 April 2007 18:18
From: "Cédric Boutillier" <cedric.boutillier@gmail.com>
To: "Johannes Berg" <johannes@sipsolutions.net>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>, "Tim Dijkstra" <newsuser@famdijkstra.org>, suspend-devel@lists.sourceforge.net

Hi !
On 4/1/07, Johannes Berg <johannes@sipsolutions.net> wrote:
> Have you gotten it to resume?
> I'd test s2both but I haven't gotten s2disk to resume yet... Need to
> test that patch.

Yes it resumes, but I am stuck at some point because after the image
is reloaded from swap, the system freezes when (usb) modules are
reloaded. That is a problem I got with suspend-to-disk with recent
kernels : 2.6.18 leads to oops, and 2.6.20 / 2.6.21-rc5 lead to
lockup.

If I unload usb modules before suspending, the system resumes and I
get a prompt, but I get also some errors from usb (root hub lost power
or was reset). (see
http://marc.info/?l=linux-usb-devel&m=117528005511118&w=2)
I would appreciate If you had an idea on that .


Cedric

Re: Fwd: Re: [Suspend-devel] s2both on powerpc

[David Brownell]

On Sunday 01 April 2007 9:37 am, Rafael J. Wysocki wrote:
> Alan, Dave, could you please have a look at this?
> 
> Rafael
> 
> 
> ----------  Forwarded Message  ----------
> 
> Subject: Re: [Suspend-devel] s2both on powerpc
> Date: Sunday, 1 April 2007 18:18
> From: "Cédric Boutillier" <cedric.boutillier@gmail.com>
> To: "Johannes Berg" <johannes@sipsolutions.net>
> Cc: "Rafael J. Wysocki" <rjw@sisk.pl>, "Tim Dijkstra" <newsuser@famdijkstra.org>,
>	 suspend-devel@lists.sourceforge.net 
> 
> Hi !
> On 4/1/07, Johannes Berg <johannes@sipsolutions.net> wrote:
> > Have you gotten it to resume?
> > I'd test s2both but I haven't gotten s2disk to resume yet... Need to
> > test that patch.
> 
> Yes it resumes, but I am stuck at some point because after the image
> is reloaded from swap, the system freezes when (usb) modules are
> reloaded.

Which module?  I'm guessing just ehci-hcd, but you should try each
of the relevant HCDs by themselves.  (And if you have any external
USB devices connected, their drivers ... and also see whether the
connection of the external device is itself a factor.)

There's something platform-specific going on, and some clues will
be needed to sort it out.

All this used to work, as I recall...

- Dave




> That is a problem I got with suspend-to-disk with recent 
> kernels : 2.6.18 leads to oops, and 2.6.20 / 2.6.21-rc5 lead to
> lockup.
> 
> If I unload usb modules before suspending, the system resumes and I
> get a prompt, but I get also some errors from usb (root hub lost power
> or was reset). (see
> http://marc.info/?l=linux-usb-devel&m=117528005511118&w=2)
> I would appreciate If you had an idea on that .
> 
> 
> Cedric
> 

Re: Fwd: Re: [Suspend-devel] s2both on powerpc

[Rafael J. Wysocki]

On Thursday, 5 April 2007 23:38, David Brownell wrote:
> On Sunday 01 April 2007 9:37 am, Rafael J. Wysocki wrote:
> > Alan, Dave, could you please have a look at this?
> > 
> > Rafael
> > 
> > 
> > ----------  Forwarded Message  ----------
> > 
> > Subject: Re: [Suspend-devel] s2both on powerpc
> > Date: Sunday, 1 April 2007 18:18
> > From: "Cédric Boutillier" <cedric.boutillier@gmail.com>
> > To: "Johannes Berg" <johannes@sipsolutions.net>
> > Cc: "Rafael J. Wysocki" <rjw@sisk.pl>, "Tim Dijkstra" <newsuser@famdijkstra.org>,
> >	 suspend-devel@lists.sourceforge.net 
> > 
> > Hi !
> > On 4/1/07, Johannes Berg <johannes@sipsolutions.net> wrote:
> > > Have you gotten it to resume?
> > > I'd test s2both but I haven't gotten s2disk to resume yet... Need to
> > > test that patch.
> > 
> > Yes it resumes, but I am stuck at some point because after the image
> > is reloaded from swap, the system freezes when (usb) modules are
> > reloaded.
> 
> Which module?  I'm guessing just ehci-hcd, but you should try each
> of the relevant HCDs by themselves.  (And if you have any external
> USB devices connected, their drivers ... and also see whether the
> connection of the external device is itself a factor.)
> 
> There's something platform-specific going on, and some clues will
> be needed to sort it out.
> 
> All this used to work, as I recall...

This has already been handled by Alan (identified as a hardware bug, patch
is available).

Thanks,
Rafael

Re: Fwd: Re: [Suspend-devel] s2both on powerpc

[Alan Stern]

On Fri, 6 Apr 2007, Rafael J. Wysocki wrote:

> On Thursday, 5 April 2007 23:38, David Brownell wrote:
> > On Sunday 01 April 2007 9:37 am, Rafael J. Wysocki wrote:
> > > Alan, Dave, could you please have a look at this?
> > > 
> > > Rafael
> > > 
> > > 
> > > ----------  Forwarded Message  ----------
> > > 
> > > Subject: Re: [Suspend-devel] s2both on powerpc
> > > Date: Sunday, 1 April 2007 18:18
> > > From: "Cédric Boutillier" <cedric.boutillier@gmail.com>
> > > To: "Johannes Berg" <johannes@sipsolutions.net>
> > > Cc: "Rafael J. Wysocki" <rjw@sisk.pl>, "Tim Dijkstra" <newsuser@famdijkstra.org>,
> > >	 suspend-devel@lists.sourceforge.net 
> > > 
> > > Hi !
> > > On 4/1/07, Johannes Berg <johannes@sipsolutions.net> wrote:
> > > > Have you gotten it to resume?
> > > > I'd test s2both but I haven't gotten s2disk to resume yet... Need to
> > > > test that patch.
> > > 
> > > Yes it resumes, but I am stuck at some point because after the image
> > > is reloaded from swap, the system freezes when (usb) modules are
> > > reloaded.
> > 
> > Which module?  I'm guessing just ehci-hcd, but you should try each
> > of the relevant HCDs by themselves.  (And if you have any external
> > USB devices connected, their drivers ... and also see whether the
> > connection of the external device is itself a factor.)
> > 
> > There's something platform-specific going on, and some clues will
> > be needed to sort it out.
> > 
> > All this used to work, as I recall...
> 
> This has already been handled by Alan (identified as a hardware bug, patch
> is available).

In case you're curious, the problem and the patch are here:

http://marc.info/?l=linux-usb-devel&m=117545150508671&w=2

I'm just about to submit it to Greg.  AFAICS, removing that code will 
never cause any problems.  Hubs should automatically clear the suspend 
feature whenever a port is reset, which is the next thing to happen on 
that pathway.

Alan Stern