[lm-sensors] [PATCH] hwmon: (coretemp) Fix oops on driver load

[lm-sensors] [PATCH] hwmon: (coretemp) Fix oops on driver load

[Jean Delvare]

On Thu, 1 Dec 2011 01:26:39 +0100, Jean Delvare wrote:
> On Thu, 01 Dec 2011 07:25:15 +0800 (HKT), undersys wrote:
> > Hi All,
> > 
> > I can no longer load the core temp module, it gives me the following kernel Opps.
> > (...)
> > This used to work fine before I updated the kernel to 3.0.x. (...)
> 
> Ah, yes, I can reproduce that, and the bug is pretty obvious. I'll send
> a patch tomorrow.
> 
> Meanwhile you may be able to work around the problem by installing the
> microcode_ctl package and ensuring that the microcode is updated before
> loading the coretemp driver.

If the CPU microcode is too old, the coretemp driver won't work. But
instead of failing gracefully, it currently oops. Check for NULL
platform device data to avoid this.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: Durgadoss R <Durgadoss.r@intel.com>
Cc: Guenter Roeck <guenter.roeck@ericsson.com>
---
This is for stable kernel branch 3.0 only. Previous and later versions
have different code paths and are not affected by this bug.

 drivers/hwmon/coretemp.c |    2 ++
 1 file changed, 2 insertions(+)

--- linux-3.0.orig/drivers/hwmon/coretemp.c	2011-12-01 08:39:45.000000000 +0100
+++ linux-3.0/drivers/hwmon/coretemp.c	2011-12-01 08:49:32.000000000 +0100
@@ -539,6 +539,8 @@ static void coretemp_add_core(unsigned i
 		return;
 
 	pdata = platform_get_drvdata(pdev);
+	if (!pdata)
+		return;
 
 	err = create_core_data(pdata, pdev, cpu, pkg_flag);
 	if (err)

-- 
Jean Delvare

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] [PATCH] hwmon: (coretemp) Fix oops on driver load

[R, Durgadoss]

Thanks Jean for the fix.

> If the CPU microcode is too old, the coretemp driver won't work. But
> instead of failing gracefully, it currently oops. Check for NULL
> platform device data to avoid this.

Acked-by: Durgadoss R <durgadoss.r@intel.com>

Thanks,
Durga

> Signed-off-by: Jean Delvare <khali@linux-fr.org>
> Cc: Fenghua Yu <fenghua.yu@intel.com>
> Cc: Durgadoss R <Durgadoss.r@intel.com>
> Cc: Guenter Roeck <guenter.roeck@ericsson.com>
> ---
> This is for stable kernel branch 3.0 only. Previous and later versions
> have different code paths and are not affected by this bug.
> 
>  drivers/hwmon/coretemp.c |    2 ++
>  1 file changed, 2 insertions(+)
> 
> --- linux-3.0.orig/drivers/hwmon/coretemp.c	2011-12-01 08:39:45.000000000
> +0100
> +++ linux-3.0/drivers/hwmon/coretemp.c	2011-12-01 08:49:32.000000000 +0100
> @@ -539,6 +539,8 @@ static void coretemp_add_core(unsigned i
>  		return;
> 
>  	pdata = platform_get_drvdata(pdev);
> +	if (!pdata)
> +		return;
> 
>  	err = create_core_data(pdata, pdev, cpu, pkg_flag);
>  	if (err)
> 
> --
> Jean Delvare

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] [PATCH] hwmon: (coretemp) Fix oops on driver load

[Guenter Roeck]

On Thu, 2011-12-01 at 05:00 -0500, Jean Delvare wrote:
> On Thu, 1 Dec 2011 01:26:39 +0100, Jean Delvare wrote:
> > On Thu, 01 Dec 2011 07:25:15 +0800 (HKT), undersys wrote:
> > > Hi All,
> > > 
> > > I can no longer load the core temp module, it gives me the following kernel Opps.
> > > (...)
> > > This used to work fine before I updated the kernel to 3.0.x. (...)
> > 
> > Ah, yes, I can reproduce that, and the bug is pretty obvious. I'll send
> > a patch tomorrow.
> > 
> > Meanwhile you may be able to work around the problem by installing the
> > microcode_ctl package and ensuring that the microcode is updated before
> > loading the coretemp driver.
> 
> If the CPU microcode is too old, the coretemp driver won't work. But
> instead of failing gracefully, it currently oops. Check for NULL
> platform device data to avoid this.
> 
> Signed-off-by: Jean Delvare <khali@linux-fr.org>
> Cc: Fenghua Yu <fenghua.yu@intel.com>
> Cc: Durgadoss R <Durgadoss.r@intel.com>
> Cc: Guenter Roeck <guenter.roeck@ericsson.com>
Acked-by: Guenter Roeck <guenter.roeck@ericsson.com>

Jean, how do we send this to Greg ? stable@vger.kernel.org ?

Guenter

> ---
> This is for stable kernel branch 3.0 only. Previous and later versions
> have different code paths and are not affected by this bug.
> 
>  drivers/hwmon/coretemp.c |    2 ++
>  1 file changed, 2 insertions(+)
> 
> --- linux-3.0.orig/drivers/hwmon/coretemp.c	2011-12-01 08:39:45.000000000 +0100
> +++ linux-3.0/drivers/hwmon/coretemp.c	2011-12-01 08:49:32.000000000 +0100
> @@ -539,6 +539,8 @@ static void coretemp_add_core(unsigned i
>  		return;
>  
>  	pdata = platform_get_drvdata(pdev);
> +	if (!pdata)
> +		return;
>  
>  	err = create_core_data(pdata, pdev, cpu, pkg_flag);
>  	if (err)
> 



_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] [PATCH] hwmon: (coretemp) Fix oops on driver load

[Jean Delvare]

On Thu, 1 Dec 2011 08:05:42 -0800, Guenter Roeck wrote:
> On Thu, 2011-12-01 at 05:00 -0500, Jean Delvare wrote:
> > On Thu, 1 Dec 2011 01:26:39 +0100, Jean Delvare wrote:
> > > On Thu, 01 Dec 2011 07:25:15 +0800 (HKT), undersys wrote:
> > > > Hi All,
> > > > 
> > > > I can no longer load the core temp module, it gives me the following kernel Opps.
> > > > (...)
> > > > This used to work fine before I updated the kernel to 3.0.x. (...)
> > > 
> > > Ah, yes, I can reproduce that, and the bug is pretty obvious. I'll send
> > > a patch tomorrow.
> > > 
> > > Meanwhile you may be able to work around the problem by installing the
> > > microcode_ctl package and ensuring that the microcode is updated before
> > > loading the coretemp driver.
> > 
> > If the CPU microcode is too old, the coretemp driver won't work. But
> > instead of failing gracefully, it currently oops. Check for NULL
> > platform device data to avoid this.
> > 
> > Signed-off-by: Jean Delvare <khali@linux-fr.org>
> > Cc: Fenghua Yu <fenghua.yu@intel.com>
> > Cc: Durgadoss R <Durgadoss.r@intel.com>
> > Cc: Guenter Roeck <guenter.roeck@ericsson.com>
> Acked-by: Guenter Roeck <guenter.roeck@ericsson.com>

Thanks.

> Jean, how do we send this to Greg ? stable@vger.kernel.org ?

Yes. I'll do so right now.

-- 
Jean Delvare

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] [PATCH] hwmon: (coretemp) Fix oops on driver load

[undersys]

----- Original Message -----
From: "Jean Delvare" <khali@linux-fr.org>
To: "guenter roeck" <guenter.roeck@ericsson.com>
Cc: undersys@basstech.net, lm-sensors@lm-sensors.org, "Durgadoss R" <Durgadoss.r@intel.com>
Sent: Friday, 2 December, 2011 12:19:25 AM
Subject: Re: [PATCH] hwmon: (coretemp) Fix oops on driver load

On Thu, 1 Dec 2011 08:05:42 -0800, Guenter Roeck wrote:
> On Thu, 2011-12-01 at 05:00 -0500, Jean Delvare wrote:
> > On Thu, 1 Dec 2011 01:26:39 +0100, Jean Delvare wrote:
> > > On Thu, 01 Dec 2011 07:25:15 +0800 (HKT), undersys wrote:
> > > > Hi All,
> > > > 
> > > > I can no longer load the core temp module, it gives me the following kernel Opps.
> > > > (...)
> > > > This used to work fine before I updated the kernel to 3.0.x. (...)
> > > 
> > > Ah, yes, I can reproduce that, and the bug is pretty obvious. I'll send
> > > a patch tomorrow.
> > > 
> > > Meanwhile you may be able to work around the problem by installing the
> > > microcode_ctl package and ensuring that the microcode is updated before
> > > loading the coretemp driver.
> > 
> > If the CPU microcode is too old, the coretemp driver won't work. But
> > instead of failing gracefully, it currently oops. Check for NULL
> > platform device data to avoid this.
> > 
> > Signed-off-by: Jean Delvare <khali@linux-fr.org>
> > Cc: Fenghua Yu <fenghua.yu@intel.com>
> > Cc: Durgadoss R <Durgadoss.r@intel.com>
> > Cc: Guenter Roeck <guenter.roeck@ericsson.com>
> Acked-by: Guenter Roeck <guenter.roeck@ericsson.com>

Thanks.

> Jean, how do we send this to Greg ? stable@vger.kernel.org ?

Yes. I'll do so right now.

-- 
Jean Delvare

I've now got the latest microcode from Intel loading fine.
The coretemp module still will case an "OPPS" but the new code will solve that
from what i understand but it also still means that coretemp no longer pics up my CPU temp..

Thanks

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] [PATCH] hwmon: (coretemp) Fix oops on driver load

[Jean Delvare]

On Sat, 03 Dec 2011 09:32:24 +0800 (HKT), undersys wrote:
> I've now got the latest microcode from Intel loading fine.
> The coretemp module still will case an "OPPS" but the new code will solve that
> from what i understand but it also still means that coretemp no longer pics up my CPU temp..

Odd. Which microcode package version are you running?
What's your exact CPU family/model/stepping?
Did you get the coretemp driver to work for this CPU with any kernel
version in the 2.6.22 to 2.6.39 range?

-- 
Jean Delvare

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] [PATCH] hwmon: (coretemp) Fix oops on driver load

[undersys]

----- Original Message -----
From: "Jean Delvare" <khali@linux-fr.org>
To: undersys@basstech.net
Cc: lm-sensors@lm-sensors.org, "Durgadoss R" <Durgadoss.r@intel.com>, "guenter roeck" <guenter.roeck@ericsson.com>
Sent: Tuesday, 6 December, 2011 1:02:54 AM
Subject: Re: [PATCH] hwmon: (coretemp) Fix oops on driver load

On Sat, 03 Dec 2011 09:32:24 +0800 (HKT), undersys wrote:
> I've now got the latest microcode from Intel loading fine.
> The coretemp module still will case an "OPPS" but the new code will solve that
> from what i understand but it also still means that coretemp no longer pics up my CPU temp..

Odd. Which microcode package version are you running?
What's your exact CPU family/model/stepping?
Did you get the coretemp driver to work for this CPU with any kernel
version in the 2.6.22 to 2.6.39 range?

-- 
Jean Delvare

The microcode utility I am using is :-
http://urbanmyth.org/microcode

The version of micro code I am using is "microcode-20110915.tgz" :-
http://downloadcenter.intel.com/Detail_Desc.aspx?agr=Y&DwnldID 429&lang=eng

CPU is :-
Intel Xeon sossaman, model\x14, stepping=4. 
I don't have the SPEC number on hand let me know if you need more information.

Micro code was working on 2.6.29 and 2.6.39 , then I jumped to 3.0.6.

Thanks for your help

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] [PATCH] hwmon: (coretemp) Fix oops on driver load

[Jean Delvare]

On Tue, 06 Dec 2011 07:26:59 +0800 (HKT), undersys wrote:
> The microcode utility I am using is :-
> http://urbanmyth.org/microcode
> 
> The version of micro code I am using is "microcode-20110915.tgz" :-
> http://downloadcenter.intel.com/Detail_Desc.aspx?agr=Y&DwnldID 429&lang=eng

Did you have to do that with previous kernels too?

> CPU is :-
> Intel Xeon sossaman, model\x14, stepping=4. 
> I don't have the SPEC number on hand let me know if you need more information.

Is this a single or multi-cpu system?

I would like to see the contents of /proc/cpuinfo.

> Micro code was working on 2.6.29 and 2.6.39 , then I jumped to 3.0.6.

I admit I don't quite get it for now. While the device probing logic
was modified a lot in the coretemp driver of kernel 3.0, the microcode
check did not change, so I just don't get why the new driver would
behave differently from the previous ones in this respect.

For testing purposes, I made the 2.6.39 version of the coretemp driver
available as a standalone driver:
http://khali.linux-fr.org/devel/misc/coretemp_old/
(Instructions at http://khali.linux-fr.org/devel/misc/INSTALL )

I would like you to test it on your 3.0 kernel and report whether it is
working or not. You'll have to rmmod the original coretemp driver
before loading coretemp_old, you can't have both loaded at once.

If coretemp_old works, then please rmmod it and modprobe coretemp
again, and see if it then works.

-- 
Jean Delvare
http://khali.linux-fr.org/wishlist.html

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] [PATCH] hwmon: (coretemp) Fix oops on driver load

[undersys]

[-- Attachment #1: Type: text/plain, Size: 2608 bytes --]

----- Original Message -----
From: "Jean Delvare" <khali@linux-fr.org>
To: undersys@basstech.net
Cc: lm-sensors@lm-sensors.org, "Durgadoss R" <Durgadoss.r@intel.com>, "guenter roeck" <guenter.roeck@ericsson.com>
Sent: Tuesday, 6 December, 2011 5:53:41 PM
Subject: Re: [PATCH] hwmon: (coretemp) Fix oops on driver load

On Tue, 06 Dec 2011 07:26:59 +0800 (HKT), undersys wrote:
> The microcode utility I am using is :-
> http://urbanmyth.org/microcode
> 
> The version of micro code I am using is "microcode-20110915.tgz" :-
> http://downloadcenter.intel.com/Detail_Desc.aspx?agr=Y&DwnldID=20429&lang=eng

Did you have to do that with previous kernels too?

undersys ->  I Did not ever need to use the microcode utility.
             Maybe i should in future...

> CPU is :-
> Intel Xeon sossaman, model=14, stepping=4. 
> I don't have the SPEC number on hand let me know if you need more information.

Is this a single or multi-cpu system?

undersys -> multi-cpu (dual socket)

I would like to see the contents of /proc/cpuinfo.

undersys -> attached

> Micro code was working on 2.6.29 and 2.6.39 , then I jumped to 3.0.6.

I admit I don't quite get it for now. While the device probing logic
was modified a lot in the coretemp driver of kernel 3.0, the microcode
check did not change, so I just don't get why the new driver would
behave differently from the previous ones in this respect.

For testing purposes, I made the 2.6.39 version of the coretemp driver
available as a standalone driver:
http://khali.linux-fr.org/devel/misc/coretemp_old/
(Instructions at http://khali.linux-fr.org/devel/misc/INSTALL )

I would like you to test it on your 3.0 kernel and report whether it is
working or not. You'll have to rmmod the original coretemp driver
before loading coretemp_old, you can't have both loaded at once.

If coretemp_old works, then please rmmod it and modprobe coretemp
again, and see if it then works.

undersys ->

I built your module(coretemp_old), works fine no kernel opps.
The only lines printed to "dmesg" are :-
[  133.603204] coretemp coretemp.0: Errata AE18 not fixed, update BIOS or microcode of the CPU!
[  133.603277] coretemp coretemp.1: Errata AE18 not fixed, update BIOS or microcode of the CPU!
[  133.603329] coretemp coretemp.2: Errata AE18 not fixed, update BIOS or microcode of the CPU!
[  133.603379] coretemp coretemp.3: Errata AE18 not fixed, update BIOS or microcode of the CPU!

If i remove it and then attempt to re-probe the new one, the new one fails.


-- 
Jean Delvare
http://khali.linux-fr.org/wishlist.html


Hi Jean,

Answers are inline.

Thanks for your time!

[-- Attachment #2: cpuinfo.txt --]
[-- Type: text/plain, Size: 2761 bytes --]

# cat /proc/cpuinfo 
processor	: 0
vendor_id	: GenuineIntel
cpu family	: 6
model		: 14
model name	: Intel(R) Dual Pentium(R) M CPU   000  @ 1.83GHz
stepping	: 4
cpu MHz		: 996.000
cache size	: 2048 KB
physical id	: 0
siblings	: 2
core id		: 0
cpu cores	: 2
apicid		: 0
initial apicid	: 0
fdiv_bug	: no
hlt_bug		: no
f00f_bug	: no
coma_bug	: no
fpu		: yes
fpu_exception	: yes
cpuid level	: 6
wp		: yes
flags		: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe nx constant_tsc bts pni monitor vmx est tm2 xtpr dts
bogomips	: 3667.09
clflush size	: 64
cache_alignment	: 64
address sizes	: 36 bits physical, 32 bits virtual
power management:

processor	: 1
vendor_id	: GenuineIntel
cpu family	: 6
model		: 14
model name	: Intel(R) Dual Pentium(R) M CPU   000  @ 1.83GHz
stepping	: 4
cpu MHz		: 996.000
cache size	: 2048 KB
physical id	: 3
siblings	: 2
core id		: 0
cpu cores	: 2
apicid		: 6
initial apicid	: 6
fdiv_bug	: no
hlt_bug		: no
f00f_bug	: no
coma_bug	: no
fpu		: yes
fpu_exception	: yes
cpuid level	: 6
wp		: yes
flags		: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe nx constant_tsc bts pni monitor vmx est tm2 xtpr dts
bogomips	: 3666.16
clflush size	: 64
cache_alignment	: 64
address sizes	: 36 bits physical, 32 bits virtual
power management:

processor	: 2
vendor_id	: GenuineIntel
cpu family	: 6
model		: 14
model name	: Intel(R) Dual Pentium(R) M CPU   000  @ 1.83GHz
stepping	: 4
cpu MHz		: 996.000
cache size	: 2048 KB
physical id	: 0
siblings	: 2
core id		: 1
cpu cores	: 2
apicid		: 1
initial apicid	: 1
fdiv_bug	: no
hlt_bug		: no
f00f_bug	: no
coma_bug	: no
fpu		: yes
fpu_exception	: yes
cpuid level	: 6
wp		: yes
flags		: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe nx constant_tsc bts pni monitor vmx est tm2 xtpr dts
bogomips	: 3666.16
clflush size	: 64
cache_alignment	: 64
address sizes	: 36 bits physical, 32 bits virtual
power management:

processor	: 3
vendor_id	: GenuineIntel
cpu family	: 6
model		: 14
model name	: Intel(R) Dual Pentium(R) M CPU   000  @ 1.83GHz
stepping	: 4
cpu MHz		: 996.000
cache size	: 2048 KB
physical id	: 3
siblings	: 2
core id		: 1
cpu cores	: 2
apicid		: 7
initial apicid	: 7
fdiv_bug	: no
hlt_bug		: no
f00f_bug	: no
coma_bug	: no
fpu		: yes
fpu_exception	: yes
cpuid level	: 6
wp		: yes
flags		: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe nx constant_tsc bts pni monitor vmx est tm2 xtpr dts
bogomips	: 3666.16
clflush size	: 64
cache_alignment	: 64
address sizes	: 36 bits physical, 32 bits virtual
power management:


[-- Attachment #3: Type: text/plain, Size: 153 bytes --]

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] [PATCH] hwmon: (coretemp) Fix oops on driver load

[Jean Delvare]

On Tue, 06 Dec 2011 19:50:04 +0800 (HKT), undersys wrote:
> I would like to see the contents of /proc/cpuinfo.
> 
> undersys -> attached

I can only confirm that this is one of the series where the microcode
needs to be recent enough. No surprise here.

> > Micro code was working on 2.6.29 and 2.6.39 , then I jumped to 3.0.6.
> 
> I admit I don't quite get it for now. While the device probing logic
> was modified a lot in the coretemp driver of kernel 3.0, the microcode
> check did not change, so I just don't get why the new driver would
> behave differently from the previous ones in this respect.
> 
> For testing purposes, I made the 2.6.39 version of the coretemp driver
> available as a standalone driver:
> http://khali.linux-fr.org/devel/misc/coretemp_old/
> (Instructions at http://khali.linux-fr.org/devel/misc/INSTALL )
> 
> I would like you to test it on your 3.0 kernel and report whether it is
> working or not. You'll have to rmmod the original coretemp driver
> before loading coretemp_old, you can't have both loaded at once.
> 
> If coretemp_old works, then please rmmod it and modprobe coretemp
> again, and see if it then works.
> 
> undersys ->
> 
> I built your module(coretemp_old), works fine no kernel opps.
> The only lines printed to "dmesg" are :-
> [  133.603204] coretemp coretemp.0: Errata AE18 not fixed, update BIOS or microcode of the CPU!
> [  133.603277] coretemp coretemp.1: Errata AE18 not fixed, update BIOS or microcode of the CPU!
> [  133.603329] coretemp coretemp.2: Errata AE18 not fixed, update BIOS or microcode of the CPU!
> [  133.603379] coretemp coretemp.3: Errata AE18 not fixed, update BIOS or microcode of the CPU!
> 
> If i remove it and then attempt to re-probe the new one, the new one fails.

I'm giving up on this one. I tried reproducing on one of my systems by
simulating a wrong microcode version. I obtain the same behavior for
the 2.6.32, 2.6.39 and (fixed) 3.0 versions of the coretemp driver:
error message in the log, driver loads without crashing, and "sensors"
show no coretemp device. So this is all as expected and I don't see any
regression here, thus there is nothing for me to fix.

-- 
Jean Delvare

_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

Re: [lm-sensors] [PATCH] hwmon: (coretemp) Fix oops on driver load

[undersys]

----- Original Message -----
From: "Jean Delvare" <khali@linux-fr.org>
To: undersys@basstech.net
Cc: lm-sensors@lm-sensors.org, "Durgadoss R" <Durgadoss.r@intel.com>, "guenter roeck" <guenter.roeck@ericsson.com>
Sent: Sunday, 11 December, 2011 8:49:50 PM
Subject: Re: [PATCH] hwmon: (coretemp) Fix oops on driver load

On Tue, 06 Dec 2011 19:50:04 +0800 (HKT), undersys wrote:
> I would like to see the contents of /proc/cpuinfo.
> 
> undersys -> attached

I can only confirm that this is one of the series where the microcode
needs to be recent enough. No surprise here.

> > Micro code was working on 2.6.29 and 2.6.39 , then I jumped to 3.0.6.
> 
> I admit I don't quite get it for now. While the device probing logic
> was modified a lot in the coretemp driver of kernel 3.0, the microcode
> check did not change, so I just don't get why the new driver would
> behave differently from the previous ones in this respect.
> 
> For testing purposes, I made the 2.6.39 version of the coretemp driver
> available as a standalone driver:
> http://khali.linux-fr.org/devel/misc/coretemp_old/
> (Instructions at http://khali.linux-fr.org/devel/misc/INSTALL )
> 
> I would like you to test it on your 3.0 kernel and report whether it is
> working or not. You'll have to rmmod the original coretemp driver
> before loading coretemp_old, you can't have both loaded at once.
> 
> If coretemp_old works, then please rmmod it and modprobe coretemp
> again, and see if it then works.
> 
> undersys ->
> 
> I built your module(coretemp_old), works fine no kernel opps.
> The only lines printed to "dmesg" are :-
> [  133.603204] coretemp coretemp.0: Errata AE18 not fixed, update BIOS or microcode of the CPU!
> [  133.603277] coretemp coretemp.1: Errata AE18 not fixed, update BIOS or microcode of the CPU!
> [  133.603329] coretemp coretemp.2: Errata AE18 not fixed, update BIOS or microcode of the CPU!
> [  133.603379] coretemp coretemp.3: Errata AE18 not fixed, update BIOS or microcode of the CPU!
> 
> If i remove it and then attempt to re-probe the new one, the new one fails.

I'm giving up on this one. I tried reproducing on one of my systems by
simulating a wrong microcode version. I obtain the same behavior for
the 2.6.32, 2.6.39 and (fixed) 3.0 versions of the coretemp driver:
error message in the log, driver loads without crashing, and "sensors"
show no coretemp device. So this is all as expected and I don't see any
regression here, thus there is nothing for me to fix.

-- 
Jean Delvare

Hi Jean,

No worries. Thanks for your time spent ! For now I will go back to getting the data from the IPMI card.
It would be nice if this Xeon was oneday supported but hey :)

Thanks


_______________________________________________
lm-sensors mailing list
lm-sensors@lm-sensors.org
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors