Re: Linux 2.4.5-ac9

David Ford (david@blue-labs.org)
Tue, 05 Jun 2001 21:32:48 -0700


2.4.5-ac8 has a brokenness about it.

sshd stalled in [down] with the following, subsequent sshd attempts
which needed a tty resulted in D state the same as the first:

invalid operand: 0000
CPU: 0
EIP: 0010:[<c01269f9>]
Using defaults from ksymoops -t elf32-i386 -a i386
EFLAGS: 00010086
eax: 0000001b ebx: c13bf768 ecx: c0345060 edx: 00002c76
esi: c0a54000 edi: c0a549aa ebp: c0a549aa esp: c57afe54
ds: 0018 es: 0018 ss: 0018
Process sshd (pid: 3627, stackpage=c57af000)
Stack: c02cab45 000004dc 00000000 0000800a c03dc900 c03dc900 00001000
00000246
c01b1021 00000c3c 00000007 c03dc900 c01b1c43 0000000a c03dc900
c03dc900
00000000 c13c9e50 0000000a 00000000 00000000 00000000 00000000
c13c9e50
Call Trace: [<c01b1021>] [<c01b1c43>] [<c01b2643>] [<c0137fc0>]
[<c0138871>]
[<c0167ccb>] [<c012e389>] [<c012e2c2>] [<c012e5b0>] [<c0106a93>]
Code: 0f 0b 83 c4 08 89 f6 f6 43 11 04 74 51 b8 a5 c2 0f 17 87 06

>>EIP; c01269f9 <proc_getdata+4d/154> <=====
Trace; c01b1021 <read_eeprom+131/1a8>
Trace; c01b1c43 <rtl8139_tx_timeout+143/148>
Trace; c01b2643 <rtl8139_interrupt+5f/170>
Trace; c0137fc0 <__emul_lookup_dentry+a4/fc>
Trace; c0138871 <open_namei+4d1/560>
Trace; c0167ccb <leaf_define_dest_src_infos+1a7/1ac>
Trace; c012e389 <do_readv_writev+15d/228>
Trace; c012e2c2 <do_readv_writev+96/228>
Trace; c012e5b0 <sys_pread+bc/d0>
Trace; c0106a93 <system_call+23/40>
Code; c01269f9 <proc_getdata+4d/154>
00000000 <_EIP>:
Code; c01269f9 <proc_getdata+4d/154> <=====
0: 0f 0b ud2a <=====
Code; c01269fb <proc_getdata+4f/154>
2: 83 c4 08 add $0x8,%esp
Code; c01269fe <proc_getdata+52/154>
5: 89 f6 mov %esi,%esi
Code; c0126a00 <proc_getdata+54/154>
7: f6 43 11 04 testb $0x4,0x11(%ebx)
Code; c0126a04 <proc_getdata+58/154>
b: 74 51 je 5e <_EIP+0x5e> c0126a57
<proc_getdata+ab/154>
Code; c0126a06 <proc_getdata+5a/154>
d: b8 a5 c2 0f 17 mov $0x170fc2a5,%eax
Code; c0126a0b <proc_getdata+5f/154>
12: 87 06 xchg %eax,(%esi)

Alan Cox wrote:

> ftp://ftp.kernel.org/pub/linux/kernel/people/alan/2.4/
>
> Intermediate diffs are available from
> http://www.bzimage.org
>
>In terms of going through the code audit almost all the sound drivers still
>need fixing to lock against format changes during a read/write. Poll creating
>and starting a buffer as write does and also mmap during write, write during
>an mmap.
>
>2.4.5-ac9
>

-
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/