Bug#321678: xine-ui: and a kernel oops by the way

[Paul Brossier]

On Sat, Sep 10, 2005 at 11:02:44AM +0200, Siggi Langauf wrote:
> On Sat, 10 Sep 2005, Paul Brossier wrote:
> 
> >Sep 10 02:05:43 localhost kernel: Oops: kernel access of bad area, sig: 11 
> >[#5]
> >Sep 10 02:05:43 localhost kernel: SMP NR_CPUS=32
> >Sep 10 02:05:43 localhost kernel: NIP: C0005460 LR: C0004734 SP: D0539F20 
> >REGS: d0539e70 TRAP: 0300    Not tainted
> >Sep 10 02:05:43 localhost kernel: MSR: 00009032 EE: 1 PR: 0 FP: 0 ME: 1 
> >IR/DR: 11
> >Sep 10 02:05:43 localhost kernel: DAR: 00000088, DSISR: 40000000
> >Sep 10 02:05:43 localhost kernel: TASK = d4e7e810[16988] 'xine' THREAD: 
> >d0538000
> >Sep 10 02:05:43 localhost kernel: Last syscall: 174  CPU: 1
> >Sep 10 02:05:43 localhost kernel: GPR00: C0004734 D0539F20 D4E7E810 
> >00000004 00000004 00030001 00000000 0F70787C
> >Sep 10 02:05:43 localhost kernel: GPR08: 0F706F98 C0004734 00009032 
> >C0320000 00009032 100F117C 00000000 100F28E0
> >Sep 10 02:05:43 localhost kernel: GPR16: 00000015 00000000 00000000 
> >0000000F 00000001 0000000A 00000001 00000002
> >Sep 10 02:05:43 localhost kernel: GPR24: 00000000 00000000 100F0000 
> >100F28E0 0FE639DC 00000000 0FE6819C 00000004
> >Sep 10 02:05:43 localhost kernel: NIP [c0005460] 
> >AltivecUnavailException+0x24/0x98
> >Sep 10 02:05:43 localhost kernel: LR [c0004734] ret_from_except+0x0/0x14
> >Sep 10 02:05:43 localhost kernel: Call trace:
> >Sep 10 02:05:43 localhost kernel:  [c0004734] ret_from_except+0x0/0x14
> 
> Well, xine does work on G3 CPUs (without AltiVec) as well as on G4 (with 
> AltiVec). I'd strongly recommend you use an AltiVec enabled kernel if 
> running on an AltiVec CPU, because otherwise any application using AltiVec 
> extensions will lead to a kernel oops (or might even have unrecognized 
> side effects on other applications).

ok, good to know these 'without altivec' bugs are not present on G3.

> I'm not sure if there's a way to check for AltiVec support in the kernel 
> (in addition to the checks on CPU capability that xine already does), but 
> I'll see if there's a way to work around this crash (other than disabling 
> AltiVec support altogether)...

i would also like to know about detecting the presence of altivec. there
was a talk about it at last debconf, but i could not get any more info. 
there was a discussion in http://bugs.debian.org/225959

> Just out of curiosity: What kind of machine is that? Does it really have 
> 32 processors?!

it's a G5, or PowerMac 7,2 to be precise. these bugs are faced on debian
kernels -power4{,-smp} << 2.6.12 , which were compiled without altivec
support. so essentially xine and totem have these bugs in sarge.

cheers, piem