Re: kernel BUG at net/ipv4/tcp_output.c:1006!

From: Eric Dumazet
Date: Fri May 13 2011 - 15:47:48 EST


Le vendredi 13 mai 2011 Ã 15:30 -0400, TB a Ãcrit :
> On 11-05-13 01:27 PM, Eric Dumazet wrote:
> > Le vendredi 13 mai 2011 Ã 13:11 -0400, TB a Ãcrit :
> >> This is the 2.6.38.5 kernel with the patch in
> >> [PATCH] tcp_cubic: limit delayed_ack ratio to prevent divide error
> >>
> >
> > Please send us full disassembly of tcp_fragment (from vmlinux file)
>
>
> GCC is debian 4.3.2-1.1
> AS 2.18.0.20080103
>
> CPU is Intel Xeon E5620
> Kernel CPU is set to MCORE2 (Core 2/newer Xeon)
>
>
> ffffffff814e7eb0 <tcp_fragment>:
> ffffffff814e7eb0: 41 57 push %r15
> ffffffff814e7eb2: 49 89 ff mov %rdi,%r15
> ffffffff814e7eb5: 41 56 push %r14
> ffffffff814e7eb7: 41 55 push %r13
> ffffffff814e7eb9: 41 89 d5 mov %edx,%r13d
> ffffffff814e7ebc: 41 54 push %r12
> ffffffff814e7ebe: 55 push %rbp
> ffffffff814e7ebf: 53 push %rbx
> ffffffff814e7ec0: 48 89 f3 mov %rsi,%rbx
> ffffffff814e7ec3: 48 83 ec 18 sub $0x18,%rsp
> ffffffff814e7ec7: 89 4c 24 0c mov %ecx,0xc(%rsp)
> ffffffff814e7ecb: 8b 6e 68 mov 0x68(%rsi),%ebp
> ffffffff814e7ece: 39 ea cmp %ebp,%edx
> ffffffff814e7ed0: 76 04 jbe ffffffff814e7ed6
> <tcp_fragment+0x26>
> ffffffff814e7ed2: 0f 0b ud2a
> ffffffff814e7ed4: eb fe jmp ffffffff814e7ed4
> <tcp_fragment+0x24>



So skb->len = 0x1540 and len = 0x1708

I suspect we should push commit 2fceec13375e5d98 (tcp: len check is
unnecessarily devastating, change to WARN_ON) to stable if not already
done...

http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=2fceec13375e5d98

David, is this commit in your stable queue ?

Thanks !


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