Problems with 2.4.7-ac6 + SMP + FastTrak100

Problems with 2.4.7-ac6 + SMP + FastTrak100

[Arnvid Karstad]

Hiya,

I rebuilt my kernel 2.4.7 after applied Alan Cox patch 6 to get FastTrak
100 support in my kernel. The kernel boots fine up until after it's
freeing memory used by kernel with the following printout..


Freeing unused kernel memory: 236kbytes
Invalid operand: 0000
CPU:0
EIP: 0010:[<c010ca24>]
EFALGS: 00010206
... then all the registers

then
..
Process swapper
...
code: 0f ae 82 90 03 00 00 db c2 eb 08 90 dd b2 90 03 00 00 9b 0f


And then it dies in "idle task something" and freezes.. ;>
(it's a thunder storm here atm, so I will boot to get more info later)

My system is a 

Dual P2 300 Mhz with 192MB ram
Promise FastTrak 100 + 4x IBM ATA 100 75GB disks
Intel EtherPro/100 and an old Mach64CT gfx adapter

Regards

Arnvid

Re: Problems with 2.4.7-ac6 + SMP + FastTrak100

[Alan Cox]

> I rebuilt my kernel 2.4.7 after applied Alan Cox patch 6 to get FastTrak
> 100 support in my kernel. The kernel boots fine up until after it's
> freeing memory used by kernel with the following printout..
> 
> 
> Freeing unused kernel memory: 236kbytes
> Invalid operand: 0000
> CPU:0
> EIP: 0010:[<c010ca24>]
> EFALGS: 00010206
> ... then all the registers

If yo can grab the EIP and the calltrace after the thunderstorm, then look
them up in System.map that would be great. Also see if disabling pnpbios
support helps

Re: Problems with 2.4.7-ac6 + SMP + FastTrak100

[Arnvid Karstad]

On Sun, 5 Aug 2001 20:59:47 +0100 (BST)
Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:

> > I rebuilt my kernel 2.4.7 after applied Alan Cox patch 6 to get FastTrak
> > 100 support in my kernel. The kernel boots fine up until after it's
> > freeing memory used by kernel with the following printout..
> > 
> > 
> > Freeing unused kernel memory: 236kbytes
> > Invalid operand: 0000
> > CPU:0
> > EIP: 0010:[<c010ca24>]
> > EFALGS: 00010206
> > ... then all the registers
> 
> If yo can grab the EIP and the calltrace after the thunderstorm, then look
> them up in System.map that would be great. Also see if disabling pnpbios
> support helps


I tried to both disable the pnpbios in the config but it didn't help

EIP 0010:[<c010c764>]
Process swapper : (pid: 0, stackspace c0279000)
CALLTRACE
[<c01057c2>][<c0112d1e>][<c01051b0>][<c010501b0>][<c010524e>][<c0105000>][<c0105043>]

Can't really seem to find any of this in system.map .. but I'm prolly
looking wrong =)

Re: Problems with 2.4.7-ac6 + SMP + FastTrak100

[Alan Cox]

> I tried to both disable the pnpbios in the config but it didn't help
> 
> EIP 0010:[<c010c764>]
> Process swapper : (pid: 0, stackspace c0279000)
> CALLTRACE
> [<c01057c2>][<c0112d1e>][<c01051b0>][<c010501b0>][<c010524e>][<c0105000>][<c0105043>]
> 
> Can't really seem to find any of this in system.map .. but I'm prolly
> looking wrong =)

For each of the eip and call trace values find the System.map entry that is
the nearest address _before_ the one given

Re: Problems with 2.4.7-ac6 + SMP + FastTrak100

[Alan Cox]

> Even the 2.4.7 vanilla did die.. I booted a 2.4.5 perfectly before.. and
> 2.2.19 SMP
> Gonna try them both again now

Does 2.4.6 fail ?

> CPU:    0
> EIP:    0010:[<c010bef8>]
> EFLAGS: 00010206

Crashed doing FPU setup

Your boot CPU ends "cmov mmx" (ie Pentium II)

> bogomips        : 614.40

> flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca =
> cmov pat pse36 mmx fxsr
> bogomips        : 616.03

And the second processor has fxsr (ie SSE - preventium III)

It looks like that is tripping a bug that got re-introduced in 2.4.6 or
2.4.7 (hence I asked which failed). That would explain why there haven't
been millions of reports

Alan

Re: Problems with 2.4.7-ac6 + SMP + FastTrak100

[Arnvid Karstad]

I've tested more now

It spits one of these msgs during the init script... but continues on
2.4.5 SMP

But on 2.4.7UP + ac6 + the promise driver enabled it works like a charm
if I go back to 2.2.19 it still works like a charm.. Time to  test more
tho.. I still cannot make the kernel see my raid drive .. since I've
raided the 4 disks in bios to 1 large drive.. it stil shows me 4 drives
all being 75GB.. wierd ?



On Mon, 06 Aug 2001 01:33:38 +0200
Arnvid Karstad <arnvid@karstad.org> wrote:

> invalid operand: 0000
> CPU:    0
> EIP:    0010:[<c010bef8>]
> EFLAGS: 00010206
> eax: 0183fbff   ebx: cb648000   ecx: 40238cc0   edx: cb648000
> esi: bffff3a0   edi: 0808dce8   ebp: 00000014   esp: cb649fb8
> ds: 0018   es: 0018   ss: 0018
> Process sshd (pid: 77, stackpage=cb649000)
> Stack: c0107c69 cb648000 c0106d0c 400fb9d4 40238cc0 0808de50 bffff3a0 0808dce8 
>        00000014 00000000 0000002b 0000002b ffffffff 400978a8 00000023 00010246 
>        bffff374 0000002b 
> Call Trace: [<c0107c69>] [<c0106d0c>] 
> 
> Code: 0f ae 8a 90 03 00 00 c3 dd a2 90 03 00 00 c3 90 8b 54 24 04 
> 
> 
> 00000000c010bee8 T restore_fpu
> 00000000c0107c50 T math_state_restore
> 00000000c0106d0c t ret_from_exception
> 
> That's with 2.4.5 but it continues..
> 
> Gonna try 2.4.4 and 2.4.6 now
> 
> 2.2.19 still works
> 
> On Mon, 6 Aug 2001 00:16:25 +0100 (BST)
> Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:
> 
> > > Even the 2.4.7 vanilla did die.. I booted a 2.4.5 perfectly before.. and
> > > 2.2.19 SMP
> > > Gonna try them both again now
> > 
> > Does 2.4.6 fail ?
> > 
> > > CPU:    0
> > > EIP:    0010:[<c010bef8>]
> > > EFLAGS: 00010206
> > 
> > Crashed doing FPU setup
> > 
> > Your boot CPU ends "cmov mmx" (ie Pentium II)
> > 
> > > bogomips        : 614.40
> > 
> > > flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca =
> > > cmov pat pse36 mmx fxsr
> > > bogomips        : 616.03
> > 
> > And the second processor has fxsr (ie SSE - preventium III)
> > 
> > It looks like that is tripping a bug that got re-introduced in 2.4.6 or
> > 2.4.7 (hence I asked which failed). That would explain why there haven't
> > been millions of reports
> > 
> > Alan
> > -
> > To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> > the body of a message to majordomo@vger.kernel.org
> > More majordomo info at  http://vger.kernel.org/majordomo-info.html
> > Please read the FAQ at  http://www.tux.org/lkml/
> 
> Mvh, 
> 
> Arnvid Karstad,
> 
> 
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/

Re: Problems with 2.4.7-ac6 + SMP + FastTrak100

[Arnvid Karstad]

invalid operand: 0000
CPU:    0
EIP:    0010:[<c010bef8>]
EFLAGS: 00010206
eax: 0183fbff   ebx: cb648000   ecx: 40238cc0   edx: cb648000
esi: bffff3a0   edi: 0808dce8   ebp: 00000014   esp: cb649fb8
ds: 0018   es: 0018   ss: 0018
Process sshd (pid: 77, stackpage=cb649000)
Stack: c0107c69 cb648000 c0106d0c 400fb9d4 40238cc0 0808de50 bffff3a0 0808dce8 
       00000014 00000000 0000002b 0000002b ffffffff 400978a8 00000023 00010246 
       bffff374 0000002b 
Call Trace: [<c0107c69>] [<c0106d0c>] 

Code: 0f ae 8a 90 03 00 00 c3 dd a2 90 03 00 00 c3 90 8b 54 24 04 


00000000c010bee8 T restore_fpu
00000000c0107c50 T math_state_restore
00000000c0106d0c t ret_from_exception

That's with 2.4.5 but it continues..

Gonna try 2.4.4 and 2.4.6 now

2.2.19 still works

On Mon, 6 Aug 2001 00:16:25 +0100 (BST)
Alan Cox <alan@lxorguk.ukuu.org.uk> wrote:

> > Even the 2.4.7 vanilla did die.. I booted a 2.4.5 perfectly before.. and
> > 2.2.19 SMP
> > Gonna try them both again now
> 
> Does 2.4.6 fail ?
> 
> > CPU:    0
> > EIP:    0010:[<c010bef8>]
> > EFLAGS: 00010206
> 
> Crashed doing FPU setup
> 
> Your boot CPU ends "cmov mmx" (ie Pentium II)
> 
> > bogomips        : 614.40
> 
> > flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca =
> > cmov pat pse36 mmx fxsr
> > bogomips        : 616.03
> 
> And the second processor has fxsr (ie SSE - preventium III)
> 
> It looks like that is tripping a bug that got re-introduced in 2.4.6 or
> 2.4.7 (hence I asked which failed). That would explain why there haven't
> been millions of reports
> 
> Alan
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/

Mvh, 

Arnvid Karstad,