This kernel requires the following features not present on the CPU... (on a VIA C7 CPU)

This kernel requires the following features not present on the CPU... (on a VIA C7 CPU)

[Claas Langbehn]

[-- Attachment #1: Type: text/plain, Size: 1554 bytes --]

Hello,

I have got a VIA EPIA EX15000G Mini-ITX mainboard with a C7 VIA Esther 
processor (1500MHz).
There are two options in the BIOS that affect CPU feature flags:

C7 CMPXCHG8 - enable/disable (Disable to install windows NT 4.0) and
C7 No Execute (NX) - enable/disable

I enabled both and compiled a new 2.6.22-rc1-mm1 kernel. So far so good.

The problem is that the system is very unstable and crashes regularly 
(1-3 times a day).
After a crash the bios resets the two processor flags back to disabled. 
But like this
the kernel complains and does not continiue booting:

 > This kernel requires the following features not present on the CPU:
 > 0:8

Each time this happens, I have to go to the BIOS and manually enable 
these flags again.
That is pretty annoying! :(
I attached a cpuinfo.txt. When the two flags are disabled, cx8 and nx 
are not listed there.


Now my question is:

Would it be possible to override the BIOS settings of cx8 and nx and 
activate it with linux anyway?
The CPU supports it and I don't see any reason to disable it.


I already wrote several emails to VIA months ago saying that this is 
probably a bug in their BIOS, but they
are neitherreleasing a new BIOS nor reacting in any way. Their so called 
support website viaarena.com
is not helping neither. Since VIA is not helping at all, the only 
soulution I see would be to override the
BIOS'es decision.

Concerning the crashes I already changed the memory module, but with no 
success. I'm not sure
wether this is a hardware or software bug.


Many regards,
claas

[-- Attachment #2: cpuinfo.txt --]
[-- Type: text/plain, Size: 481 bytes --]

processor	: 0
vendor_id	: CentaurHauls
cpu family	: 6
model		: 10
model name	: VIA Esther processor 1500MHz
stepping	: 9
cpu MHz		: 799.952
cache size	: 128 KB
fdiv_bug	: no
hlt_bug		: no
f00f_bug	: no
coma_bug	: no
fpu		: yes
fpu_exception	: yes
cpuid level	: 1
wp		: yes
flags		: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge cmov pat clflush acpi mmx fxsr sse sse2 nx pni est rng rng_en ace ace_en ace2 ace2_en phe phe_en pmm pmm_en
bogomips	: 1601.30
clflush size	: 64


[-- Attachment #3: dmesg.txt --]
[-- Type: text/plain, Size: 1678 bytes --]

000000 (gap: 3bf00000:a4100000)
Built 1 zonelists in Zone order, mobility grouping on.  Total pages: 243571
Kernel command line: root=/dev/sda3 ro quiet 
mapped APIC to ffffd000 (fee00000)
mapped IOAPIC to ffffc000 (fec00000)
Enabling fast FPU save and restore... done.
Enabling unmasked SIMD FPU exception support... done.
Initializing CPU#0
PID hash table entries: 4096 (order: 12, 16384 bytes)
Detected 1499.916 MHz processor.
Console: colour VGA+ 80x25
Dentry cache hash table entries: 131072 (order: 7, 524288 bytes)
Inode-cache hash table entries: 65536 (order: 6, 262144 bytes)
Memory: 969928k/981952k available (1931k kernel code, 11472k reserved, 659k data, 192k init, 64448k highmem)
virtual kernel memory layout:
    fixmap  : 0xfffaa000 - 0xfffff000   ( 340 kB)
    pkmap   : 0xff800000 - 0xffc00000   (4096 kB)
    vmalloc : 0xf8800000 - 0xff7fe000   ( 111 MB)
    lowmem  : 0xc0000000 - 0xf8000000   ( 896 MB)
      .init : 0xc038a000 - 0xc03ba000   ( 192 kB)
      .data : 0xc02e2c3b - 0xc03879d4   ( 659 kB)
      .text : 0xc0100000 - 0xc02e2c3b   (1931 kB)
Checking if this processor honours the WP bit even in supervisor mode... Ok.
Calibrating delay using timer specific routine.. 3002.19 BogoMIPS (lpj=5001917)
Mount-cache hash table entries: 512
CPU: After generic identify, caps: 87c9bbff 00100000 00000000 00000000 00000081 00000000 00000000
CPU: L1 I Cache: 64K (64 bytes/line), D cache 64K (64 bytes/line)
CPU: L2 Cache: 128K (64 bytes/line)
CPU: After all inits, caps: 07c9bbff 00100000 00000000 00000000 00000081 00003fcc 00000000
Compat vDSO mapped to ffffe000.
CPU: Centaur VIA Esther processor 1500MHz stepping 09
Checking 'hlt' instruction... OK.

Re: This kernel requires the following features not present on the CPU... (on a VIA C7 CPU)

[Alan Cox]

> Concerning the crashes I already changed the memory module, but with no 
> success. I'm not sure
> wether this is a hardware or software bug.

Testing with memtest86 should help tell you if the problem is a dodgy
DIMM specifically rather than another hardware or software fault. 

Re: This kernel requires the following features not present on the CPU... (on a VIA C7 CPU)

[Claas Langbehn]

Alan Cox wrote:
>> Concerning the crashes I already changed the memory module, but with no 
>> success. I'm not sure
>> wether this is a hardware or software bug.
>>     
>
> Testing with memtest86 should help tell you if the problem is a dodgy
> DIMM specifically rather than another hardware or software fault.
>   

I already did this. I had a broken aeneon DIMM at first. I got a new 
working one.
But the crashes were still there. Then I got a samsung DIMM instead and 
it did
not get better. Therefore I think that the memory is not causing this.

Actually the system freezes when I click somewhere with my USB mouse, even
with no load at all. Unfortunately the system does not have a serial 
port for
a serial console. Netconsole was not working with the via-rhine. I wanted to
try serial console over ttyUSB, but I did not have time for it, yet.


Claas

Re: This kernel requires the following features not present on the CPU... (on a VIA C7 CPU)

[H. Peter Anvin]

Claas Langbehn wrote:
> 
> Now my question is:
> 
> Would it be possible to override the BIOS settings of cx8 and nx and
> activate it with linux anyway?
> The CPU supports it and I don't see any reason to disable it.
> 

Yes, that code is already in the git.newsetup tree.

	-hpa

Re: This kernel requires the following features not present on the CPU... (on a VIA C7 CPU)

[Claas Langbehn]

[-- Attachment #1: Type: text/plain, Size: 588 bytes --]

H. Peter Anvin schrieb:
> Claas Langbehn wrote:
>   
>> Now my question is:
>>
>> Would it be possible to override the BIOS settings of cx8 and nx and
>> activate it with linux anyway?
>> The CPU supports it and I don't see any reason to disable it.
>>
>>     
>
> Yes, that code is already in the git.newsetup tree.
>   

I just tried this code and it did not work. I am still getting the error 
message from the topic.
This crappy BIOS disables the CPU flags each time. I have to reactivate 
it before each boot time :(

I attached a "x86info -a -v", if it helps.


many regards,
claas

[-- Attachment #2: VIA_C7_Esther.txt --]
[-- Type: text/plain, Size: 3511 bytes --]

x86info v1.18.  Dave Jones 2001-2006
Feedback to <davej@redhat.com>.

Found 1 CPU
MP Table:
#	APIC ID	Version	State		Family	Model	Step	Flags
#	 0	 0x11	 BSP, usable	 6	 10	 9	 0xfbff

--------------------------------------------------------------------------
eax in: 0x00000000, eax = 00000001 ebx = 746e6543 ecx = 736c7561 edx = 48727561
eax in: 0x00000001, eax = 000006a9 ebx = 00010800 ecx = 00000081 edx = 87c9bbff

eax in: 0x80000000, eax = 80000006 ebx = 00000000 ecx = 00000000 edx = 00000000
eax in: 0x80000001, eax = 00000000 ebx = 00000000 ecx = 00000000 edx = 00100000
eax in: 0x80000002, eax = 20202020 ebx = 20202020 ecx = 20202020 edx = 20202020
eax in: 0x80000003, eax = 56202020 ebx = 45204149 ecx = 65687473 edx = 72702072
eax in: 0x80000004, eax = 7365636f ebx = 20726f73 ecx = 30303531 edx = 007a484d
eax in: 0x80000005, eax = 00000000 ebx = 08800880 ecx = 40040140 edx = 40040140
eax in: 0x80000006, eax = 00000000 ebx = 00000000 ecx = 0080a140 edx = 00000000

eax in: 0xc0000000, eax = c0000002 ebx = 00000000 ecx = 00000000 edx = 00000000
eax in: 0xc0000001, eax = 00000000 ebx = 00000000 ecx = 00000000 edx = 00003fcc
eax in: 0xc0000002, eax = 00000000 ebx = 08000810 ecx = 08100f13 edx = 42000000

Family: 6 Model: 10 Stepping: 9
CPU Model : VIA C3 (Esther) [C7-M]
Processor name string: VIA Esther processor 1500MHz

Feature flags:
	Onboard FPU
	Virtual Mode Extensions
	Debugging Extensions
	Page Size Extensions
	Time Stamp Counter
	Model-Specific Registers
	Physical Address Extensions
	Machine Check Architecture
	CMPXCHG8 instruction
	Onboard APIC
	SYSENTER/SYSEXIT
	Memory Type Range Registers
	Page Global Enable
	CMOV instruction
	Page Attribute Table
	CLFLUSH instruction
	ACPI via MSR
	MMX support
	FXSAVE and FXRESTORE instructions
	SSE support
	SSE2 support
	Pending Break Enable

Extended feature flags:

Instruction TLB: 8-way associative. 128 entries.
Data TLB: 8-way associative. 128 entries.
L1 Data cache:
	Size: 64Kb	4-way associative.
	lines per tag=1	line size=64 bytes.
L1 Instruction cache:
	Size: 64Kb	4-way associative.
	lines per tag=1	line size=64 bytes.
L2 (on CPU) cache:
	Size: 128Kb	10-way associative.
	lines per tag=1	line size=64 bytes.
FCR: MSR: 0x00001107=0x9f1f1ac6 : 10011111 00011111 00011010 11000110
Power management: 
MTRR registers:
MTRRcap (0xfe): 0x0000000000000508
MTRRphysBase0 (0x200): 0x0000000000000006
MTRRphysMask0 (0x201): 0x0000000fc0000800
MTRRphysBase1 (0x202): 0x000000003c000000
MTRRphysMask1 (0x203): 0x0000000ffc000800
MTRRphysBase2 (0x204): 0x00000000d8000001
MTRRphysMask2 (0x205): 0x0000000ffc000800
MTRRphysBase3 (0x206): 0x000000003bf00000
MTRRphysMask3 (0x207): 0x0000000ffff00800
MTRRphysBase4 (0x208): 0x00000000a0000001
MTRRphysMask4 (0x209): 0x00000000fc000800
MTRRphysBase5 (0x20a): 0x0000000000000000
MTRRphysMask5 (0x20b): 0x0000000000000000
MTRRphysBase6 (0x20c): 0x0000000000000000
MTRRphysMask6 (0x20d): 0x0000000000000000
MTRRphysBase7 (0x20e): 0x0000000000000000
MTRRphysMask7 (0x20f): 0x0000000000000000
MTRRfix64K_00000 (0x250): 0x0606060606060606
MTRRfix16K_80000 (0x258): 0x0606060606060606
MTRRfix16K_A0000 (0x259): 0x0000000000000000
MTRRfix4K_C8000 (0x269): 0x0505050505050505
MTRRfix4K_D0000 0x26a: 0x0000000000000000
MTRRfix4K_D8000 0x26b: 0x0000000000000000
MTRRfix4K_E0000 0x26c: 0x0000000000000000
MTRRfix4K_E8000 0x26d: 0x0000000000000000
MTRRfix4K_F0000 0x26e: 0x0404040404040404
MTRRfix4K_F8000 0x26f: 0x0404040404040400
MTRRdefType (0x2ff): 0x0000000000000c00


800MHz processor (estimate).

Re: This kernel requires the following features not present on the CPU... (on a VIA C7 CPU)

[Simon Arlott]

[-- Attachment #1: Type: text/plain, Size: 1948 bytes --]

On 17/05/07 13:27, Claas Langbehn wrote:
> I have got a VIA EPIA EX15000G Mini-ITX mainboard with a C7 VIA Esther 
> processor (1500MHz).
I have Jetway's J7F4K with the same CPU.

> C7 CMPXCHG8 - enable/disable (Disable to install windows NT 4.0) and
> C7 No Execute (NX) - enable/disable
Without those options in the BIOS setup.

>  > This kernel requires the following features not present on the CPU:
>  > 0:8
> I attached a cpuinfo.txt. When the two flags are disabled, cx8 and nx 
> are not listed there.
I don't have cx8, but I do have nx (which is unused because I have PAE 
disabled...).


On 17/05/07 19:33, H. Peter Anvin wrote:
>> Claas Langbehn wrote:
>> Would it be possible to override the BIOS settings of cx8 and nx and
>> > activate it with linux anyway?
>> > The CPU supports it and I don't see any reason to disable it.
>> > 
>  
> Yes, that code is already in the git.newsetup tree.

Is it automatic? I have CONFIG_X86_CMPXCHG=y without cx8 showing in 
cpuinfo, and it appears to work fine.

Will your changes needlessly prevent the kernel running? Would I be 
right in thinking that the kernel is successfully using cmpxchg even 
though it's considered disabled? I realise people compile kernels 
for the wrong CPU but preventing them working when it's been chosen 
correctly seems wrong.

processor       : 0
vendor_id       : CentaurHauls
cpu family      : 6
model           : 10
model name      : VIA Esther processor 1500MHz
stepping        : 9
cpu MHz         : 800.000 (cpufreq)
cache size      : 128 KB
fdiv_bug        : no
hlt_bug         : no
f00f_bug        : no
coma_bug        : no
fpu             : yes
fpu_exception   : yes
cpuid level     : 1
wp              : yes
flags           : fpu vme de pse tsc msr pae mce apic sep mtrr pge cmov pat clflush acpi mmx fxsr sse sse2 tm nx pni est tm2 rng rng_en ace ace_en ace2 ace2_en phe phe_en pmm pmm_en
bogomips        : 1601.18
clflush size    : 64

-- 
Simon Arlott

[-- Attachment #2: config.gz --]
[-- Type: application/x-gzip, Size: 10555 bytes --]

Re: This kernel requires the following features not present on the CPU... (on a VIA C7 CPU)

[H. Peter Anvin]

Simon Arlott wrote:
> 
> Is it automatic? I have CONFIG_X86_CMPXCHG=y without cx8 showing in
> cpuinfo, and it appears to work fine.
> 
> Will your changes needlessly prevent the kernel running? Would I be
> right in thinking that the kernel is successfully using cmpxchg even
> though it's considered disabled? I realise people compile kernels for
> the wrong CPU but preventing them working when it's been chosen
> correctly seems wrong.
> 

CX8 isn't cmpxchg; it's cmpxchg8b.

	-hpa

Re: This kernel requires the following features not present on the CPU... (on a VIA C7 CPU)

[Simon Arlott]

On 17/05/07 21:15, H. Peter Anvin wrote:
> Simon Arlott wrote:
>> Is it automatic? I have CONFIG_X86_CMPXCHG=y without cx8 showing in
>> cpuinfo, and it appears to work fine.
>>
>> Will your changes needlessly prevent the kernel running? Would I be
>> right in thinking that the kernel is successfully using cmpxchg even
>> though it's considered disabled? I realise people compile kernels for
>> the wrong CPU but preventing them working when it's been chosen
>> correctly seems wrong.
>>
> 
> CX8 isn't cmpxchg; it's cmpxchg8b.

Ok, but I have CONFIG_X86_CMPXCHG64=y too.

-- 
Simon Arlott

Re: This kernel requires the following features not present on the CPU... (on a VIA C7 CPU)

[H. Peter Anvin]

Simon Arlott wrote:
> On 17/05/07 21:15, H. Peter Anvin wrote:
>> Simon Arlott wrote:
>>> Is it automatic? I have CONFIG_X86_CMPXCHG=y without cx8 showing in
>>> cpuinfo, and it appears to work fine.
>>>
>>> Will your changes needlessly prevent the kernel running? Would I be
>>> right in thinking that the kernel is successfully using cmpxchg even
>>> though it's considered disabled? I realise people compile kernels for
>>> the wrong CPU but preventing them working when it's been chosen
>>> correctly seems wrong.
>>>
>>
>> CX8 isn't cmpxchg; it's cmpxchg8b.
> 
> Ok, but I have CONFIG_X86_CMPXCHG64=y too.
> 

OK then, then VIA just masks the CPUID bit as opposed to the
instruction.  Point still holds.

	-hpa

Re: This kernel requires the following features not present on the CPU... (on a VIA C7 CPU)

[Jan Engelhardt]

Hi,


since VIA seems to be one of the quirkier CPUs...

On May 17 2007 21:12, Simon Arlott wrote:
> On 17/05/07 13:27, Claas Langbehn wrote:
>>> Claas Langbehn wrote:

>>>> Would it be possible to override the BIOS settings of cx8 and nx and
>>>> activate it with linux anyway?
>>>> The CPU supports it and I don't see any reason to disable it.
>> 
>> Yes, that code is already in the git.newsetup tree.
>
> Is it automatic? I have CONFIG_X86_CMPXCHG=y without cx8 showing in cpuinfo,
> and it appears to work fine.
>
> Will your changes needlessly prevent the kernel running? Would I be right in
> thinking that the kernel is successfully using cmpxchg even though it's
> considered disabled? I realise people compile kernels for the wrong CPU but
> preventing them working when it's been chosen correctly seems wrong.
>
> processor       : 0
> vendor_id       : CentaurHauls
> cpu family      : 6
> model           : 10
> model name      : VIA Esther processor 1500MHz
> stepping        : 9
> cpu MHz         : 800.000 (cpufreq)
> cache size      : 128 KB
> fdiv_bug        : no
> hlt_bug         : no
> f00f_bug        : no
> coma_bug        : no
> fpu             : yes
> fpu_exception   : yes
> cpuid level     : 1
> wp              : yes
> flags           : fpu vme de pse tsc msr pae mce apic sep mtrr pge cmov pat
> clflush acpi mmx fxsr sse sse2 tm nx pni est tm2 rng rng_en ace ace_en ace2
> ace2_en phe phe_en pmm pmm_en
> bogomips        : 1601.18
> clflush size    : 64

http://lkml.org/lkml/2007/5/5/134 , shows a Family-6 Model-*9* which
*does* have CX8 (with CONFIG_M586=y, CONFIG_X86_CMPXCHG=y,
CONFIG_X86_CMPXCHG64=y). Just a hint that you do not inadvertently
disable CX there ;-)


	Jan
-- 

Re: This kernel requires the following features not present on the CPU... (on a VIA C7 CPU)

[Simon Arlott]

On 17/05/07 13:27, Claas Langbehn wrote:
> Concerning the crashes I already changed the memory module, but with no 
> success. I'm not sure
> wether this is a hardware or software bug.

Get a replacement motherboard+cpu. I've had one board with rx-only 
ethernet, and one which would only POST if I pressed reset (a full 
size power supply made it just turn off). The third one has been 
working fine since I got it.

-- 
Simon Arlott

Re: This kernel requires the following features not present on the CPU... (on a VIA C7 CPU)

[Artur Kedzierski]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1


 	I purchased C7 VIA Esther and I experience similar problem with
random crashes. There is also a thread on Via Arena about "J7F4K - Hard 
Lockups" where people describe the same issue:
http://forums.viaarena.com/messageview.aspx?catid=28&threadid=77032
 	I've tried the following kernels:
2.6.21.1 - crashes pretty quick
2.6.21.1 with disabled CPU scaling - better mileage but still crashes
2.6.22-rc1 - hangs on boot
2.6.22-rc1-mm1 - doesn't compile
2.6.22-rc2 - hangs on boot
 	I've ran memtest86 and no memory problems were found.
 	Any ideas or suggestions?
 	Thank You.

=================================================================
Artur Kedzierski                    || Artur@Kedzierski.org


-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (GNU/Linux)

iD8DBQFGT6SdJzjfxpUNGawRAqPzAJ921Rk67oA+H28hilXkqqUTIcF60QCghI69
LLLVhIpwQjN5nTsriGfC2M0=
=rG40
-----END PGP SIGNATURE-----

Re: This kernel requires the following features not present on the CPU... (on a VIA C7 CPU)

[Claas Langbehn]

Hello Artur,
>     I purchased C7 VIA Esther and I experience similar problem with
> random crashes. There is also a thread on Via Arena about "J7F4K - 
> Hard Lockups" where people describe the same issue:
> http://forums.viaarena.com/messageview.aspx?catid=28&threadid=77032
>     I've tried the following kernels:
> 2.6.21.1 - crashes pretty quick
> 2.6.21.1 with disabled CPU scaling - better mileage but still crashes
> 2.6.22-rc1 - hangs on boot
> 2.6.22-rc1-mm1 - doesn't compile
> 2.6.22-rc2 - hangs on boot
here, too: crashes over crashes...
>     I've ran memtest86 and no memory problems were found.
>     Any ideas or suggestions?
>     Thank You.
The same here.


I applied the patch mentioned by Christian Volkmann in the thread "Re: 
2.6.22-rc1 does not boot on VIA C3_2 cause of X86_CMPXCHG64 II"  on May 
19th 2007. Since then I did not have a crash any more. Touch wood!

Could anyone explain to me what CMPXCHG64 / cx8 is and what happens if 
the kernel has been compiled to use it but the CPU does not have it? 
Especially if the CPU like the C7 supports it, but it's disabled? Would 
random crashes be a plausible effect?


Regards,
claas

Re: This kernel requires the following features not present on the CPU... (on a VIA C7 CPU)

[Christian Volkmann]

Claas Langbehn wrote:
> Could anyone explain to me what CMPXCHG64 / cx8 is and what happens if
> the kernel has been compiled to use it but the CPU does not have it?
> 
> Regards,
> claas
> 


Hi Claas,

the bug is at another place. But it's hidden if X86_CMPXCHG64 is
switched of.

There are 3 errors.

First: Via is capable of CMPXCHG64, but switch the related CPU bit off
to work with old Win NT versions. ( So the NT error is the fourth... but who cares
about that error. ;-) )

Second:  verify_cpu.S
This checks during the boot the capability of the CPU. It expected
the related bit, but... see first.
This error is switched of if X86_CMPXCHG64 is disabled.

Third: setup.S "PANIC: CPU too old for this kernel." is not printed
cause of a missing set register in setup.S.


As I have learned the NX-bit of the C7 is a BIOS issue.

The only real required patch is below. It enables the right bits for the C3/C7.

The first error is out of scope, and the fix for the "CPU too old.." is for
me just a cosmetic print out in case of boot errors.

I suppose .rc3 should contain a fix for this bug.


Best regards,

Christian







--- linux.orig/arch/i386/kernel/verify_cpu.S
+++ linux/arch/i386/kernel/verify_cpu.S
@@ -2,6 +2,7 @@
    This runs in 16bit mode so that the caller can still use the BIOS
    to output errors on the screen */
 #include <asm/cpufeature.h>
+#include <asm/msr.h>

 verify_cpu:
        pushfl                          # Save caller passed flags
@@ -45,6 +46,28 @@
        cmpl    $0x1,%eax
        jb      bad                     # no cpuid 1

+#if REQUIRED_MASK1 & NEED_CMPXCHG64
+       /* Some VIA C3s need magic MSRs to enable CX64. Do this here */
+       cmpl    $0x746e6543,%ebx        # Cent
+       jne     1f
+       cmpl    $0x48727561,%edx        # aurH
+       jne     1f
+       cmpl    $0x736c7561,%ecx        # auls
+       jne     1f
+       movl    $1,%eax                 # check model
+       cpuid
+       shr     $4,%eax
+       andl    $0xf,%eax               # get model
+       cmpl    $6,%eax
+       jb      1f
+       cmpl    $10,%eax                 # newer vias hopefully don't require
+       ja      1f                      # this anymore
+       movl    $MSR_VIA_FCR,%ecx
+       rdmsr
+       orl     $((1<<1)|(1<<7)),%eax   # enable CMPXCHG64 and PGE
+       wrmsr
+1:
+#endif
        movl    $0x1,%eax               # Does the cpu have what it takes
        cpuid