Bug#342053: DirectFrameBuffer crashes on PPC systems if HW accelerated drivers are used
On Fri, Sep 22, 2006 at 02:52:24PM +0200, Attilio Fiandrotti wrote:
> Sven Luther wrote:
> >>Last times the g-i was tested on PPC, it turned tout that
> >> Known good : radeonfb, vesafb (or whatever x86 uses).
> >vesafb is not built on powerpc, and the laptop we had in extremadura used a
> >radeon chipset, altough of the R300 variety.
> isn't "fbonly" the PPC equivalent to "vesafb" on i386?
That is offb, and no, it is totally different :) Most drivers are builtin the
kernel though, so therei s really no need for offb, except for real old
> I know claudio ciccani worked a lot on DFB's Radeon driver recently, so
> things may have been fixed in DFB 0.9.25.
Currently in unstable ?
> >> Known bad : atyfb, nvidiafb
> >>when acceleration was enabled.
> >>But that was with GTKDFB 2.0.9 and DFB 0.9.22, while now we have DFB
> >>0.9.25 and GTK 2.8.20 and many bugs got fixed.
> >And it will be better with gtk 2.10.x even, right ?
> On i386 this proved to be true, but i cannot speak for other archs as i
> never experimented anything personally.
Ok. Let me think to get you an efika board once the developper program is
> >>This is the most serious bug affecting g-i on PPC, and i wasn't able to
> >>fix it because i have no PPC HW.
> >Yes, i will. Need to build a netboot g-i image and will test that on my
> >pegasos with the radon 9250 board.
> If you experience crashes, you may want to run the d-i in a chroot cage,
> like explained in this  wiki page.
> >>Sven, do you think can give the PPC g-i a try? i will help you on
> >>debugging it as much as i can.
> >Indeed, that is my intentions, a first test today, and more over the
> thanks! Having the g-i avalable for PPCs too would be really nice!
Indeed. As said, i would have done this earlier, but well, the context was not
favourable to this kind of things, let's say.
>  http://wiki.debian.org/DebianInstaller/GUIBuild