Re: Linux 2.6.12-rc2

Re: Linux 2.6.12-rc2

[Andrew Morton]

Moritz Muehlenhoff <jmm@inutil.org> wrote:
>
> Linus Torvalds wrote:
> > Benjamin Herrenschmidt:
> >   o radeonfb: Implement proper workarounds for PLL accesses
> >   o radeonfb: DDC i2c fix
> >   o radeonfb: Fix mode setting on CRT monitors
> >   o radeonfb: Preserve TMDS setting
> 
> One of these patches introduced two regressions on my Thinkpad X31 with
> "ATI Technologies Inc Radeon Mobility M6 LY (prog-if 00 [VGA])":
> 
> 1. When resuming from S3 suspend and having switched off the backlight
> with radeontool the backlight isn't switched back on any more.
> 
> 2. I'm using fbcon as my primary work environment, but tty switching has
> become _very_ sloppy, it's at least a second now, while with 2.6.11 it
> was as fast as a few ms. Is this caused by the "proper PLL accesses"?
> 

Moritz, can you tell us whether either of these problems remain in 2.6.12-rc5?

Thanks.


-------------------------------------------------------
SF.Net email is sponsored by: GoToMeeting - the easiest way to collaborate
online with coworkers and clients while avoiding the high cost of travel and
communications. There is no equipment to buy and you can meet as often as
you want. Try it free.http://ads.osdn.com/?ad_id=7402&alloc_id=16135&op=click

Re: Linux 2.6.12-rc2

[Benjamin Herrenschmidt]

On Wed, 2005-05-25 at 21:56 -0700, Andrew Morton wrote:
> Moritz Muehlenhoff <jmm@inutil.org> wrote:
> >
> > Linus Torvalds wrote:
> > > Benjamin Herrenschmidt:
> > >   o radeonfb: Implement proper workarounds for PLL accesses
> > >   o radeonfb: DDC i2c fix
> > >   o radeonfb: Fix mode setting on CRT monitors
> > >   o radeonfb: Preserve TMDS setting
> > 
> > One of these patches introduced two regressions on my Thinkpad X31 with
> > "ATI Technologies Inc Radeon Mobility M6 LY (prog-if 00 [VGA])":
> > 
> > 1. When resuming from S3 suspend and having switched off the backlight
> > with radeontool the backlight isn't switched back on any more.
> > 
> > 2. I'm using fbcon as my primary work environment, but tty switching has
> > become _very_ sloppy, it's at least a second now, while with 2.6.11 it
> > was as fast as a few ms. Is this caused by the "proper PLL accesses"?
> > 
> 
> Moritz, can you tell us whether either of these problems remain in 2.6.12-rc5?

The later will not be fixed. It's a side effect of some workarounds ATI
had me put in for the M6 which has a theorical hardware bug... unless we
decide the bug never happens and comment out the fix =P

The former, I'm not sure. x86 BIOSes are doing all sort of funny things,
I'm not sure what's going on specifically in this case. Also, the
backlight control code has known issues that I haven't find a way to fix
completely for everybody yet, it seems that different "races" of panels
need different solutions here, and by fixing one sort, I break another.
I'm still trying to get ATI to give me more infos about that, but
without luck so far.

However, why would you need to switch the backlight off with radeontool
when going to S3 ? You should let the firmware do it, or radeonfb do it,
but not manually. What if you just leave it "alone", what happens when
you go to/from S3 ? And if stays off, can you bring it back with
radeontool ?

Ben.




-------------------------------------------------------
SF.Net email is sponsored by: GoToMeeting - the easiest way to collaborate
online with coworkers and clients while avoiding the high cost of travel and
communications. There is no equipment to buy and you can meet as often as
you want. Try it free.http://ads.osdn.com/?ad_id=7402&alloc_id=16135&op=click

Re: Linux 2.6.12-rc2

[Moritz Muehlenhoff]

Benjamin Herrenschmidt wrote:
> On Wed, 2005-05-25 at 21:56 -0700, Andrew Morton wrote:
> > Moritz Muehlenhoff <jmm@inutil.org> wrote:
> > >
> > > Linus Torvalds wrote:
> > > > Benjamin Herrenschmidt:
> > > >   o radeonfb: Implement proper workarounds for PLL accesses
> > > >   o radeonfb: DDC i2c fix
> > > >   o radeonfb: Fix mode setting on CRT monitors
> > > >   o radeonfb: Preserve TMDS setting
> > > 
> > > One of these patches introduced two regressions on my Thinkpad X31 with
> > > "ATI Technologies Inc Radeon Mobility M6 LY (prog-if 00 [VGA])":
> > > 
> > > 1. When resuming from S3 suspend and having switched off the backlight
> > > with radeontool the backlight isn't switched back on any more.
> > > 
> > > 2. I'm using fbcon as my primary work environment, but tty switching has
> > > become _very_ sloppy, it's at least a second now, while with 2.6.11 it
> > > was as fast as a few ms. Is this caused by the "proper PLL accesses"?
> > 
> > Moritz, can you tell us whether either of these problems remain in 2.6.12-rc5?
> 
> The later will not be fixed. It's a side effect of some workarounds ATI
> had me put in for the M6 which has a theorical hardware bug... unless we
> decide the bug never happens and comment out the fix =P

It's "fixed" now. I think you've made a second change as well that reduced the
excessive delays in some function, which was called some hundred times. With
stock 2.6.12-rc5 it's usable again, as it's down to a tenth of a second or so.
Thanks a lot.

> The former, I'm not sure. x86 BIOSes are doing all sort of funny things,
> I'm not sure what's going on specifically in this case. Also, the
> backlight control code has known issues that I haven't find a way to fix
> completely for everybody yet, it seems that different "races" of panels
> need different solutions here, and by fixing one sort, I break another.
> I'm still trying to get ATI to give me more infos about that, but
> without luck so far.

Anyway, with stock 2.6.12-rc5 it's fixed as well. Suspend resumes properly
again.

> However, why would you need to switch the backlight off with radeontool
> when going to S3 ? You should let the firmware do it, or radeonfb do it,
> but not manually. What if you just leave it "alone", what happens when
> you go to/from S3 ? And if stays off, can you bring it back with
> radeontool ?

Well, maybe they should, but they don't do it. This may very well be a firmware
issue as it's documented on all the X31 on Linux websites.

If I do a simple echo mem > /sys/power/state my X31 Thinkpad goes into S3, but
the backlight remains one which consumes a lot of power and renders suspend
more or less useless. radeontool switches it off, and it's automatically turned
through the lid ACPI event. 

Cheers,
        Moritz


-------------------------------------------------------
SF.Net email is sponsored by: GoToMeeting - the easiest way to collaborate
online with coworkers and clients while avoiding the high cost of travel and
communications. There is no equipment to buy and you can meet as often as
you want. Try it free.http://ads.osdn.com/?ad_id=7402&alloc_id=16135&op=click