Can we get the PowerPC config changed?

Tim Gardner tim.gardner at canonical.com
Mon Apr 23 16:15:42 UTC 2012


On 04/18/2012 06:47 AM, o jordan wrote:
> Hi list,
> 
> Is it too late to get some config changes into the 12.04 PowerPC kernels
> please?  It would vastly improve the ease of use of
> Ubuntu/Lubuntu/Kubuntu/etc if the ATI and radeon framebuffer modules
> were built back into the kernel:
> 
> CONFIG_FB_RADEON=y
> CONFIG_FB_ATY128=y
> CONFIG_FB_ATY=y
> 
> I've had this logged as a bug since early March
> (https://bugs.launchpad.net/ubuntu/+source/linux/+bug/949288 ).  The
> request for the radeon framebuffer
> (https://bugs.launchpad.net/ubuntu/+source/linux/+bug/779110 ) has been
> in since May last year!!!  The bug has been logged on the QA tracker
> (http://iso.qa.ubuntu.com/qatracker/reports/bugs/949288 ) and is the
> reason testcases have failed.  I've also already mentioned the problem
> to the list when the PowerPC kernels were discussed last month
> (https://lists.ubuntu.com/archives/kernel-team/2012-March/019402.html ).
> 
> I've tracked down when the framebuffers were removed and the reason
> given
> (https://lists.ubuntu.com/archives/kernel-team/2011-April/015249.html
> ).  It doesn't really make sense to remove FB_ATY and FB_ATY128 since
> KMS is not available on pre-radeon ATI cards.  The lack of the FB_ATY128
> framebuffer does, however, stop the r128 driver working on
> PowerPC.  Unlike for other architectures, with PowerPC the r128 card
> expects there to be a framebuffer in place (see "UseFBDev" option of
> manual http://manpages.ubuntu.com/manpages/precise/en/man4/r128.4.html
> ).  It doesn't work with the openfirmware framebuffer.  This causes
> computers with a rage 128 card to always boot into a low graphics mode
> or worse a terminal.  Building back in the framebuffer would solve this
> problem.
> 
> The radeonfb module does stop KMS working and I can see why it was
> removed for this reason (although it is very easy to turn
> off).  However, at the same time KMS was defaulted to off
> (https://lists.ubuntu.com/archives/kernel-team/2011-April/015248.html
> ).  So currently the kernel is not setup for UMS or KMS under radeon.
> 
> I've emailed the PowerPC kernel list to check the current status of KMS
> under PowerPC radeon.  They confirm it is largely currently broken
> (http://lists.ozlabs.org/pipermail/linuxppc-dev/2012-April/097503.html
> ).  It crashes with AGP and there is no suspend.
> 
> Currently, the only way to get suspend working with a radeon card is to
> use the radeonfb framebuffer.  It is possible to use the radeonfb
> module, but to do so requires the editing of several files.  This is
> beyond the ability of many people.  It would be much easier if this
> framebuffer is built in as suspend would just work out-of-the-box.  I'm
> already fed up of explaining to people on the UbuntuForums how to solve
> their suspend problems in 12.04.  People are walking away from Ubuntu
> and its derivatives because of the perceived lack of suspend.
> 
> It is also worth noting that some people seem to loose their tty
> consoles when the above framebuffers are not loaded.
> 
> In the bug 949288 report  I also asked for CONFIG_DRM_RADEON_KMS=y.  It
> would have made the setup almost the same as 10.04 (and other
> distributions) and I couldn't see it would cause any major problems
> (because of the presence of radeonfb).   However, I cannot justify
> lifting the kernel freeze on CONFIG_DRM_RADEON_KMS.  There is not enough
> time now to test this particular change.
> 
> If some KMS patch comes out of the linuxppc-dev KMS discussion would it
> be possible to get this eventually in 12.04 too?
> 
> Thanks
> 
> Adam
> 
> 

See updates to http://bugs.launchpad.net/bugs/949288

-- 
Tim Gardner tim.gardner at canonical.com




More information about the kernel-team mailing list