Re: [fix] Re: 2.3.31 (and 2.3.32pre2) breaks cpp with segmentation fault (ok in 2.3.30), reproducabl

Thorsten Kranzkowski (th@Marvin.DL8BCU.ampr.org.borg.net)
Thu, 16 Dec 1999 16:46:39 +0000

On Thu, Dec 16, 1999 at 01:06:56PM +0000, Thorsten Kranzkowski wrote:
> On Wed, Dec 15, 1999 at 05:09:23PM -0500, Benjamin C.R. LaHaise wrote:
> > On Wed, 15 Dec 1999, Thorsten Kranzkowski wrote:
> >
> > > Hmm - it seems I have a different problem - your patch makes no difference
> > > (applied to 2.3.31)
> >
> > Please upgrade to 2.3.33 and check if the problem persists.
> >
> > -ben
>
> I tried but the resulting kernel doesn't boot at all. At the time I expect
> the first kernel boot message, MILO gets restarted instead.
>
> alpha AXPpci33 ('Noname', ev4), gcc 2.95.2
>

Following up myself - I've commented out the mcpu_done stuff in
arch/alpha/Makefile and now I get this:


Bootstrap complete, kernel @ 0xfffffc0000310000, sp @ 0xfffffc0000304000
...turning on virtual addressing and jumping to the Linux Kernel

------------------------------------------------------------

Linux version 2.3.33 (root@Marvin) (gcc version 2.95.2 19991024 (release)) #2 Thu Dec 16 13:37:21 UTC 1999
Booting on Noname using machine vector Noname from MILO
Command line: bootdevice=sda1 bootfile=alpha-2.3.33test2.gz root=/dev/sda1 ro ip=::::::off hisax=27,AVM console=tty0 console=ttyS0
memcluster 0, usage 2, start 0, end 1
memcluster 1, usage 0, start 1, end 52
memcluster 2, usage 2, start 52, end 54
memcluster 3, usage 0, start 54, end 256
memcluster 4, usage 2, start 256, end 328
memcluster 5, usage 0, start 328, end 384
memcluster 6, usage 2, start 384, end 869
memcluster 7, usage 0, start 869, end 16384
max_low_pfn 16384
_end fffffc00006c98c8, start_pfn 869
bootmap size: 1 pages
allocating bootmap in area 869:870
freeing pages 869:16384
reserving bootmap 869:870
On node 0 totalpages: 00004000
zone(0): 16384 pages.
zone(1): 0 pages.
zone(2): 0 pages.
HWRPB cycle frequency bogus. Estimated 167019819 Hz
Console: colour VGA+ 80x25
Calibrating delay loop... 162.00 BogoMIPS
Memory: 121928k/131072k available (2273k kernel code, 4664k reserved, 537k data, 216k init)
swapper(0): Instruction fault 4
pc = [<fffffc000053f944>] ra = [<fffffc000038c8b8>] ps = 0000
[...]

ksymoops 0.7c on alpha 2.3.31. Options used
-v /usr/src/linux-2.3.33-work/vmlinux (specified)
-K (specified)
-L (specified)
-O (specified)
-m /usr/src/linux/System.map (specified)

swapper(0): Instruction fault 4
pc = [<fffffc000053f944>] ra = [<fffffc000038c8b8>] ps = 0000
Using defaults from ksymoops -t elf64-alpha -a alpha
r0 = fffffc00008fe1e0 r1 = 0000000000000000 r2 = fffffc00008fe160
r3 = 0000000000000000 r4 = fffffc00008fe1e0 r5 = fffffc00008fe1e0
r6 = fffffc00008fe1e0 r7 = fffffc00005d6f60 r8 = fffffc0000588000
r16= fffffc00008fe1e0 r17= fffffc00005967d1 r18= 0000000000000007
r19= fffffc0000611988 r20= f0f0f0f0f0f0f0f1 r21= 0000000000000001
r22= 0000000000000000 r23= 000000000000047f r24= 0000000000000000
r25= f0f0f0f0f0f0f0f1 r27= fffffc000053f840 r28= fffffc0000611988
gp = fffffc0000606480 sp = fffffc000058bf60
Code: 42411412 ee400006 28310000 <42403532> 42203411 38240000 40803404 fe5ffffa 6bfa8001

>>PC; fffffc000053f944 <memcpy+104/120> <=====
Code; fffffc000053f938 <memcpy+f8/120>
0000000000000000 <_PC>:
Code; fffffc000053f938 <memcpy+f8/120>
0: 12 14 41 42 addq a2,0x8,a2
Code; fffffc000053f93c <memcpy+fc/120>
4: 06 00 40 ee ble a2,20 <_PC+0x20> fffffc000053f958 <memcpy+118/120>
Code; fffffc000053f940 <memcpy+100/120>
8: 00 00 31 28 ldbu t0,0(a1)
^^^^^^^^^^^^^^^^
Code; fffffc000053f944 <memcpy+104/120> <=====
c: 32 35 40 42 subq a2,0x1,a2 <=====
Code; fffffc000053f948 <memcpy+108/120>
10: 11 34 20 42 addq a1,0x1,a1
Code; fffffc000053f94c <memcpy+10c/120>
14: 00 00 24 38 stb t0,0(t3)
Code; fffffc000053f950 <memcpy+110/120>
18: 04 34 80 40 addq t3,0x1,t3
Code; fffffc000053f954 <memcpy+114/120>
1c: fa ff 5f fe bgt a2,8 <_PC+0x8> fffffc000053f940 <memcpy+100/120>
Code; fffffc000053f958 <memcpy+118/120>
20: 01 80 fa 6b ret zero,(ra),0x1

Trace: [<fffffc000038fc7c>] [<fffffc000038ea20>] [<fffffc000031001c>]
Kernel panic: Attempted to kill the idle task!
In swapper task - not syncing
Warning (Oops_read): Code line not seen, dumping what data is available

Trace; fffffc000038fc7c <proc_misc_init+5c/140>
Trace; fffffc000038ea20 <uptime_read_proc+0/140>
Trace; fffffc000031001c <_stext+1c/20>

1 warning issued. Results may not be reliable.

I'm not very experienced in alpha assembler, yet. But isn't the LDBU insn
a 'load byte from memory to register' that doesn't exist on ev4 cpu's?
If so I should look where it originates from....

Thorsten

-- 
| Thorsten Kranzkowski         Internet: dl8bcu@gmx.net                        |
| Mobile: ++49 161 7210230        Snail: Niemannsweg 30, 49201 Dissen, Germany |
| Ampr: dl8bcu@db0nei.#nrw.deu.eu, dl8bcu@marvin.dl8bcu.ampr.org [44.130.8.19] |


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/