Unfortunately, I seem to be able to reliably oops ARMLinux 2.0.35. Here's
a description of the envelope of the problem in case it jogs anyone's
mind:
I have 16+16+2MB RAM and a 20MB swap partition. /u is NFS-mounted
from an IRIC 6.5.2m fileserver. I run `make install' in glibc-2.0.112
(previously cross-compiled on an linux/x86). It gets up to the stdlib
directory and then ..
make -C stdlib subdir_lib
make[2]: Entering directory `/u/mrw/Linux/gcc/glibc-2.0.112/stdlib'
make[2]: *** virtual memory exhausted. Stop.
make[2]: Leaving directory `/u/mrw/Linux/gcc/glibc-2.0.112/stdlib'
make[1]: *** [stdlib/subdir_lib] Error 2
make[1]: Leaving directory `/u/mrw/Linux/gcc/glibc-2.0.112'
make: *** [install] Error 2
root@kell:/u/mrw/Linux/gcc/glibc-build-arm# ldd `which make`
statically linked
(so I'm not banging my head on libc 4)
I add a 50MB swap file and try again. Then I get an oops, reliably, when I
get to this directory again. Here's an example one:
Unable to handle kernel paging request at virtual address 6000fbf8
current->tss.memmap = 10224000
*pgd = 00000000 *pmd = 00000000
Internal error: Oops: 2
CPU: 0
pc: [<c0014c58>]
lr: [<c0014ee0>]
sp: [<c09aed00>] ip: [<c09aed28>] fp: [<c09aed24>]
r10: [<00000000>] r9: [<c00d4f44>] r8: [<c00d39e8>]
r7: [<c09aeda4>] r6: [<c09aeda8>] r5: [<c00f05bc>] r4: [<4b534154>]
r3: [<a0000093>] r2: [<00000080>] r1: [<6000fbf8>] r0: [<c09aeda8>]
flags nzCv IRQs off FIQs on Mode SVC32
Process rpc.nfsd (pid 178, stackpage c09ae000)
Stack
Backtrace
Function entered at c0014c14 from c0014ee0
r9: c000a588
r8: 20000013
r7: c01dd218
r6: 00000001
r5: c00f05bc
r4: c09aed9c
Function entered at c0014ea0 from c0095ea0
code: e3510000 0a000030 e5913000 0a000002
Aiee, killing interrupt handler.
Of course, the machine dies at this point and the oops never makes it to
the syslog, so this is typed in by hand. Some relevant entries from the
System.map:
c000a544 t srepeat
c000a5a8 t recheck_bh2
[...]
c0014b94 T _sys_pause
c0014c14 T _wake_up
c0014d6c T _wake_up_interruptible
c0014d6c T _wake_up_interruptible
c0014ea0 T ___up
c0014ee8 T ___do_down
c0015048 T ___down
c0015060 T ___down_interruptible
[...]
c0095d38 T _ide_end_request
c0095eb4 t _ide_hwif_to_major
[...]
c00d3998 D _init_task
c00d3c64 D _jiffies
[...]
c00d4f44 D _nr_running
[...]
c00effec b _all_requests
c00f0bec b _ro_bits
Russell, if you want a syslog full of errors from the ether3 driver,
you need only ask; I can reproduce those to order too :-(
--
Matthew Wilcox <[EMAIL PROTECTED]>
"I decry the current tendency to seek patents on algorithms. There are
better ways to earn a living than to prevent other people from making use of
one's contributions to computer science." -- Donald E. Knuth, TAoCP vol 3
unsubscribe: body of `unsubscribe linux-arm' to [EMAIL PROTECTED]
