2004-04-06 00:05:18

by Jan Killius

[permalink] [raw]
Subject: Oops with cpufreq on 2.6.5-mm1

Hi,
cpufreq make an Oops on loading. I have attached the oops.

/proc/cpuinfo:
processor : 0
vendor_id : AuthenticAMD
cpu family : 15
model : 4
model name : AMD Athlon(tm) 64 Processor 3200+
stepping : 8
cpu MHz : 2002.580
cache size : 1024 KB
fpu : yes
fpu_exception : yes
cpuid level : 1
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge
mca cmov pat pse36 clflush mmx fxsr sse sse2 syscall nx mmxext lm
3dnowext 3dnow
bogomips : 3923.96
TLB size : 1088 4K pages
clflush size : 64
cache_alignment : 64
address sizes : 40 bits physical, 48 bits virtual
power management: ts fid vid ttp

/proc/modules:
nvidia 2564468 0 - Live 0xffffffffa017d000
ipv6 242688 8 - Live 0xffffffffa0140000
nfs 102496 1 - Live 0xffffffffa0125000
lockd 60880 2 nfs, Live 0xffffffffa0115000
sunrpc 136032 5 nfs,lockd, Live 0xffffffffa00f2000
ehci_hcd 27528 0 - Live 0xffffffffa00ea000
uhci_hcd 28456 0 - Live 0xffffffffa00e2000
sata_via 5892 0 - Live 0xffffffffa00df000
libata 35456 1 sata_via,[permanent], Live 0xffffffffa00d5000
snd_seq 52928 0 - Live 0xffffffffa00c7000
snd_via82xx 21568 0 - Live 0xffffffffa00c0000
snd_pcm 89996 1 snd_via82xx, Live 0xffffffffa00a9000
snd_timer 22152 2 snd_seq,snd_pcm, Live 0xffffffffa00a2000
snd_ac97_codec 65156 1 snd_via82xx, Live 0xffffffffa0091000
snd_page_alloc 9864 2 snd_via82xx,snd_pcm, Live 0xffffffffa008d000
snd_mpu401_uart 6912 1 snd_via82xx, Live 0xffffffffa008a000
snd_rawmidi 20608 1 snd_mpu401_uart, Live 0xffffffffa0083000
snd_seq_device 7308 2 snd_seq,snd_rawmidi, Live 0xffffffffa0080000
snd 41648 8
snd_seq,snd_via82xx,snd_pcm,snd_timer,snd_ac97_codec,snd_mpu401_uart,snd_rawmidi,snd_seq_device,
Live 0xffffffffa0074000
usbcore 99056 4 ehci_hcd,uhci_hcd, Live 0xffffffffa005a000
vfat 13824 5 - Live 0xffffffffa0055000
fat 43200 1 vfat, Live 0xffffffffa0049000
sk98lin 149068 1 - Live 0xffffffffa0023000
w83781d 34432 0 - Live 0xffffffffa0019000
i2c_sensor 2816 1 w83781d, Live 0xffffffffa0017000
i2c_isa 2560 0 - Live 0xffffffffa0015000
i2c_dev 10496 0 - Live 0xffffffffa0011000
i2c_core 21124 4 w83781d,i2c_sensor,i2c_isa,i2c_dev, Live
0xffffffffa000a000
cpufreq_powersave 1920 0 - Live 0xffffffffa0008000
cpufreq_userspace 5344 0 - Live 0xffffffffa0005000
powernow_k8 9643 2 - Loading 0xffffffffa0000000

--
Jan


Attachments:
(No filename) (2.33 kB)
oops.txt (1.69 kB)
Download all attachments

2004-04-06 08:45:39

by Jan Killius

[permalink] [raw]

2004-04-06 09:52:46

by Angelo Dell'Aera

[permalink] [raw]
Subject: Re: Oops with cpufreq on 2.6.5-mm1

On Tue, 6 Apr 2004 02:05:00 +0200
Jan Killius <[email protected]> wrote:

>Hi,
>cpufreq make an Oops on loading. I have attached the oops.

The problem seems to be located in find_psb_table(). In fact,
while powernow_table is correctly initialized, it's never assigned
to data->powernow_table thus leading to a NULL pointer dereferencing
in cpufreq_frequency_table_cpuinfo(). Try this patch please.

Regards.

--

Angelo Dell'Aera 'buffer'
Antifork Research, Inc. http://buffer.antifork.org


--- linux-2.6.5-mm1/arch/i386/kernel/cpu/cpufreq/powernow-k8.c.old 2004-04-06 11:46:41.000000000 +0200
+++ linux-2.6.5-mm1/arch/i386/kernel/cpu/cpufreq/powernow-k8.c 2004-04-06 11:48:29.000000000 +0200
@@ -634,6 +634,8 @@
return -EIO;
}

+ data->powernow_table = powernow_table;
+
printk(KERN_INFO PFX "currfid 0x%x (%d MHz), currvid 0x%x\n",
data->currfid, find_freq_from_fid(data->currfid), data->currvid);




2004-04-06 10:16:31

by Jan Killius

[permalink] [raw]
Subject: Re: Oops with cpufreq on 2.6.5-mm1

The patch have fixed the problem. thx
--
Jan

2004-04-06 10:52:24

by Dave Jones

[permalink] [raw]
Subject: Re: Oops with cpufreq on 2.6.5-mm1

On Tue, Apr 06, 2004 at 11:51:49AM +0200, Angelo Dell'Aera wrote:
> On Tue, 6 Apr 2004 02:05:00 +0200
> Jan Killius <[email protected]> wrote:
>
> >Hi,
> >cpufreq make an Oops on loading. I have attached the oops.
>
> The problem seems to be located in find_psb_table(). In fact,
> while powernow_table is correctly initialized, it's never assigned
> to data->powernow_table thus leading to a NULL pointer dereferencing
> in cpufreq_frequency_table_cpuinfo(). Try this patch please.

-mm has an incomplete merge of the large set of changes in this
driver. I'm working on it, and should get the merge complete today.

Dave

2004-04-06 13:58:13

by Dave Jones

[permalink] [raw]
Subject: Re: Oops with cpufreq on 2.6.5-mm1

On Tue, Apr 06, 2004 at 12:16:09PM +0200, Jan Killius wrote:
> The patch have fixed the problem. thx

Can you try out the fully merged patch at
http://www.codemonkey.org.uk/projects/bitkeeper/cpufreq
too please ?

Dave

2004-04-06 14:53:21

by Jan Killius

[permalink] [raw]
Subject: Re: Oops with cpufreq on 2.6.5-mm1

Hi,
the patch don't apply clean to 2.6.5-mm1.
On Tuesday 06 April 2004 15:54, Dave Jones wrote:
> On Tue, Apr 06, 2004 at 12:16:09PM +0200, Jan Killius wrote:
> > The patch have fixed the problem. thx
>
> Can you try out the fully merged patch at
> http://www.codemonkey.org.uk/projects/bitkeeper/cpufreq
> too please ?
>
> Dave

2004-04-06 15:07:11

by Dave Jones

[permalink] [raw]
Subject: Re: Oops with cpufreq on 2.6.5-mm1

On Tue, Apr 06, 2004 at 04:53:14PM +0200, Jan Killius wrote:

> > > The patch have fixed the problem. thx
> >
> > Can you try out the fully merged patch at
> > http://www.codemonkey.org.uk/projects/bitkeeper/cpufreq
> > too please ?
>
> the patch don't apply clean to 2.6.5-mm1.

No, it's against 2.6.5 vanilla. You'll get it to apply to -mm
if you backout (patch -R) the bk-cpufreq.patch first from
http://www.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.5/2.6.5-mm1/broken-out/

Dave

2004-04-07 00:26:56

by Jan Killius

[permalink] [raw]
Subject: Re: Oops with cpufreq on 2.6.5-mm1

The patch applies cleanly and seems to be working.
--
Jan