[uml-devel] That cowify_req bug again

[Matthew B. <ma...@by...>]

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

Hi there;

When I'd previously reported crashes during cowify_req, someone provided a
patch which added 4K onto a buffer allocation request which by and large
fixed the problem for our customers; ISTR Jeff said it was a bodge, but one
that worked.  However now we have a rather larger sample set to pick from
than a couple of months ago :) and I've got a similar-looking backtrace from
one customer along with the same total kernel freeze:

(gdb) bt
#0  0xa0113ca1 in flush_kernel_vm_range (start=2759852032, end=2826960896)
at tlb.c:88
#1  0xa0113d62 in flush_tlb_kernel_vm_skas () at tlb.c:117
#2  0xa010e885 in flush_tlb_kernel_vm () at tlb.c:25
#3  0xa010ec42 in segv (address=2760822784, ip=1076801926, is_write=0,
is_user=0, sc=0xa1038278) at trap_kern.c:95
#4  0xa010f182 in segv_handler (sig=11, regs=0xa1038278) at trap_user.c:69
#5  0xa0113e6f in sig_handler_common_skas (sig=11, sc_ptr=0xa103b1f0) at
trap_user.c:33
#6  0xa010f1dd in sig_handler (sig=11, sc=
      {gs = 0, __gsh = 0, fs = 0, __fsh = 0, es = 43, __esh = 0, ds = 43,
__dsh = 0, edi = 786428, esi = 7, ebp = 2701374732, esp = 2701374660, ebx =
2760036352, edx = 2760036352, ecx = 2687115288, eax = 196607, trapno = 14,
err = 4, eip = 2685731993, cs = 35, __csh = 0, eflags = 66054, esp_at_signal
= 2701374660, ss = 43, __ssh = 0, fpstate = 0x0, oldmask = 503324672, cr2 =
2760822784}) at trap_user.c:101
#7  <signal handler called>
#8  0xa0150499 in cowify_req (req=0xa103b580, dev=0xa02a2018) at
ubd_kern.c:842
#9  0xa0150676 in prepare_request (req=0xa03e55a0, io_req=0xa103b580) at
ubd_kern.c:890
#10 0xa01506f7 in do_ubd_request (q=0xa0383470) at ubd_kern.c:913
#11 0xa014f8e2 in ubd_handler () at ubd_kern.c:442
#12 0xa014f8f3 in ubd_intr (irq=4, dev=0xa02a1fa0, unused=0xa1038278) at
ubd_kern.c:448
#13 0xa0102f27 in handle_IRQ_event (irq=4, regs=0xa1038278,
action=0xa039aea0) at irq.c:159
#14 0xa01030d3 in do_IRQ (irq=4, regs=0xa1038278) at irq.c:322
#15 0xa010393b in sigio_handler (sig=29, regs=0xa1038278) at irq_user.c:73
#16 0xa0113e6f in sig_handler_common_skas (sig=29, sc_ptr=0xa103b718) at
trap_user.c:33
#17 0xa010f1dd in sig_handler (sig=29, sc=
      {gs = 0, __gsh = 0, fs = 0, __fsh = 0, es = 43, __esh = 0, ds = 43,
__dsh = 0, edi = 1, esi = 8, ebp = 2701375996, esp = 2701375984, ebx = 1,
edx = 0, ecx = 2701376060, eax = 0, trapno = 14, err = 4, eip = 2686438669,
cs = 35, __csh = 0, eflags = 659, esp_at_signal = 2701375984, ss = 43, __ssh
= 0, fpstate = 0x0, oldmask = 0, cr2 = 2760130773}) at trap_user.c:101
#18 <signal handler called>
#19 0xa01fcd0d in sigprocmask ()
#20 0xa0109a60 in change_signals (type=1) at signal_user.c:69
#21 0xa0109aac in unblock_signals () at signal_user.c:80
#22 0xa0108899 in handle_signal (regs=0xa1038278, signr=11, ka=0xa1a9e66c,
info=0xa103bbdc, oldset=0xa10385b4, error=0) at signal_kern.c:97
#23 0xa0108c9a in kern_do_signal (regs=0xa1038278, oldset=0xa10385b4,
error=0) at signal_kern.c:231
#24 0xa0108e39 in do_signal (error=0) at signal_kern.c:262
#25 0xa0105a1f in interrupt_end () at process_kern.c:140
#26 0xa0112f5e in userspace (regs=0xa1038278) at process.c:158
#27 0xa01136f7 in fork_handler (sig=10) at process_kern.c:99
#28 <signal handler called>
#29 0xa01fcd41 in kill ()
#30 0x85ffffb2 in ?? ()
#31 0xc085f045 in ?? ()

If this happens agin can someone advise me which kernel variables & structs
to get a dump of as well in order to diagnose this?

thanks,

- -- 
Matthew Bloch                             Bytemark Hosting
                                  tel. +44 (0) 8707 455026
                        http://www.bytemark-hosting.co.uk/
          Dedicated Linux hosts from 15ukp ($26) per month
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (GNU/Linux)

iD8DBQE+v1LXT2rVDg8aLXQRAnHZAJoCoI9DIygRH2k0sZgfz9Zv/xfHpgCaAnlz
Pr8D8pm1reGAiHqkeyN4o54=
=4kZO
-----END PGP SIGNATURE-----