Re: jprobe bug?

[William Cohen <wcohen at redhat dot com>]

William Cohen wrote:
> Martin Hunt wrote:
>
> > > To avoid having to mess with the various address I converted the 
> > > example back to a dtr script. I have attached the script to this 
> > > email. The script seemed to install okay. Does the address used in 
> > > the module look okay? I don't have a version of that kernel built.
> >
> >
> >
> > I started with a dtr script like the one you attached.  It causes the
> > crash too. Then I tried just the basic C code I sent and it still
> > happened.
> >
> > I've been running more tests this morning. It looks like you must have 2
> > CPUs (or 1 with hyperthreading enabled) and run the SMP kernel to
> > reproduce.
> >
> > CPU           kernel   bug?
> > ------------------------
> > Xeon/EM64T    UP       NO
> > 2 Xeon/EM64T  SMP      YES
> > P4/HT         UP       NO
> > P4/HT         SMP      YES
> > P4/NoHT       SMP      NO
>
>
> That explains why I didn't encounter the problem on my AMD64 machine. It 
> only has one processor.  I have a memory impoverished dual processor 
> pentium III machine. It is running a 2.6.10-1.741_FCsmp kernel. I tried 
> the same experiment and it seems to be hung.  For a while the shiftlock 
> worked and the update applet flashed, but it seems to be stuck at this 
> point.
>
> -Will

I was able to get an oops that looks like it is related to the problem 
you are talking about.

Jan 31 17:25:42 trek kernel: ------------[ cut here ]------------
Jan 31 17:25:42 trek kernel: kernel BUG at include/asm/spinlock.h:109!
Jan 31 17:25:42 trek kernel: invalid operand: 0000 [#1]
Jan 31 17:25:42 trek kernel: SMP 
Jan 31 17:25:42 trek kernel: Modules linked in: dtr(U) md5 ipv6 parport_pc lp parport autofs4 sunrpc ipt_REJECT ipt_state ip_conntrack iptable_filter ip_tables dm_mod video button battery ac uhci_hcd hw_random i2c_i801 i2c_core emu10k1_gp snd_cs46xx gameport snd_emu10k1 snd_rawmidi snd_seq_device snd_ac97_codec snd_pcm_oss snd_mixer_oss snd_pcm snd_timer snd_page_alloc snd_util_mem snd_hwdep snd soundcore 3c59x floppy ext3 jbd
Jan 31 17:25:42 trek kernel: CPU:    1
Jan 31 17:25:42 trek kernel: EIP:    0060:[<c02bb726>]    Not tainted VLI
Jan 31 17:25:42 trek kernel: EFLAGS: 00210002   (2.6.10-1.741_FC3smp) 
Jan 31 17:25:42 trek kernel: EIP is at _spin_unlock_irqrestore+0x1f/0x2d
Jan 31 17:25:42 trek kernel: eax: 00000001   ebx: 00200246   ecx: c6709f08   edx: 00200201
Jan 31 17:25:42 trek kernel: esi: 00000001   edi: c04148d8   ebp: c04148d8   esp: c5383f48
Jan 31 17:25:42 trek kernel: ds: 007b   es: 007b   ss: 0068
Jan 31 17:25:42 trek kernel: Process firefox-bin (pid: 5284, threadinfo=c5383000 task=c6ebcaa0)
Jan 31 17:25:42 trek kernel: Stack: c6709f00 c012e99f 00000000 c04148cc c012ea31 00000001 09019000 c77e6820 
Jan 31 17:25:42 trek kernel:        00000064 00000000 09019064 00000001 00000000 c5383000 c012f0f7 09019064 
Jan 31 17:25:42 trek kernel:        09019064 00000000 c012f1dc 7fffffff 00000000 00000000 b5f6c30c 00000000 
Jan 31 17:25:42 trek kernel: Call Trace:
Jan 31 17:25:42 trek kernel:  [<c012e99f>] wake_futex+0x3a/0x44
Jan 31 17:25:42 trek kernel:  [<c012ea31>] futex_wake+0x88/0xbe
Jan 31 17:25:42 trek kernel:  [<c012f0f7>] do_futex+0x32/0x5a
Jan 31 17:25:42 trek kernel:  [<c012f1dc>] sys_futex+0xbd/0xcc
Jan 31 17:25:42 trek kernel:  [<c012194c>] sys_gettimeofday+0x25/0x55
Jan 31 17:25:42 trek kernel:  [<c0103c97>] syscall_call+0x7/0xb
Jan 31 17:25:42 trek kernel: Code: 00 00 00 01 74 05 e8 02 eb ff ff c3 53 89 c1 89 d3 81 78 04 ad 4e ad de b2 01 74 08 0f 0b 6c 00 9d d4 2c c0 0f b6 01 84 c0 7e 08 <0f> 0b 6d 00 9d d4 2c c0 86 11 53 9d 5b c3 81 78 04 ad 4e ad de