2.4.26: kernel BUG at page_alloc.c:235

Linux 2.4.26 compiled with gcc-2.96 20000731 (RH7.1 2.96-85), running on an Athlon XP2200+ clocked at 1800MHz.

Any ideas?

--J.

Jul 11 11:03:26 screech kernel: kernel BUG at page_alloc.c:235!
Jul 11 11:03:26 screech kernel: invalid operand: 0000
Jul 11 11:03:26 screech kernel: CPU:    0
Jul 11 11:03:26 screech kernel: EIP:    0010:[<c012f69e>]    Not tainted
Using defaults from ksymoops -t elf32-i386 -a i386
Jul 11 11:03:26 screech kernel: EFLAGS: 00010202
Jul 11 11:03:26 screech kernel: eax: 00000040   ebx: c125ca44   ecx: 00001000   edx: 0000dbde
Jul 11 11:03:26 screech kernel: esi: c02b86b8   edi: 00000001   ebp: 00000000   esp: d062be68
Jul 11 11:03:26 screech kernel: ds: 0018   es: 0018   ss: 0018
Jul 11 11:03:26 screech kernel: Process imapd (pid: 31910, stackpage=d062b000)
Jul 11 11:03:26 screech kernel: Stack: 00001000 0000cbde 00000282 00000000 c02b86b8 c02b888c 00000001 0000000c 
Jul 11 11:03:26 screech kernel:        00000001 c012f964 c02b86b8 c012e0b5 c125ca70 00000001 c02b86b8 c02b8888 
Jul 11 11:03:26 screech kernel:        00000000 000001d2 c125ca70 cffb3b34 00000010 00007637 df68c1a8 c0127cf1 
Jul 11 11:03:26 screech kernel: Call Trace:    [<c012f964>] [<c012e0b5>] [<c0127cf1>] [<c0128367>] [<c01285b8>]
Jul 11 11:03:26 screech kernel:   [<c0128bc6>] [<c0128ac0>] [<c0135975>] [<c0135640>] [<c01357ee>] [<c01086a3>]
Jul 11 11:03:26 screech kernel: Code: 0f 0b eb 00 73 ff 26 c0 8b 43 18 a9 80 00 00 00 74 08 0f 0b 

Trace; c012f964 <__alloc_pages+74/2c0>
Trace; c012e0b5 <lru_cache_add+65/70>
Trace; c0127cf1 <page_cache_read+61/b0>
Trace; c0128367 <generic_file_readahead+117/150>
Trace; c01285b8 <do_generic_file_read+1d8/450>
Trace; c0128bc6 <generic_file_read+86/160>
Trace; c0128ac0 <file_read_actor+0/80>
Trace; c0135975 <sys_read+95/f0>
Trace; c0135640 <generic_file_llseek+0/b0>
Trace; c01357ee <sys_lseek+6e/80>
Trace; c01086a3 <system_call+33/38>
Code; c012f69e <rmqueue+1fe/240>
00000000 <_EIP>:
Code; c012f69e <rmqueue+1fe/240> <=====
0: 0f 0b ud2a <=====
Code; c012f6a0 <rmqueue+200/240>
2: eb 00 jmp 4 <_EIP+0x4> c012f6a2 <rmqueue+202/240>
Code; c012f6a2 <rmqueue+202/240>
4: 73 ff jae 5 <_EIP+0x5> c012f6a3 <rmqueue+203/240>
Code; c012f6a4 <rmqueue+204/240>
6: 26 c0 8b 43 18 a9 80 rorb $0x0,%es:0x80a91843(%ebx)
Code; c012f6ab <rmqueue+20b/240>
d: 00
Code; c012f6ac <rmqueue+20c/240>
e: 00 00 add %al,(%eax)
Code; c012f6ae <rmqueue+20e/240>
10: 74 08 je 1a <_EIP+0x1a> c012f6b8 <rmqueue+218/240>
Code; c012f6b0 <rmqueue+210/240>
12: 0f 0b ud2a

Jul 11 11:03:29 screech kernel:  kernel BUG at page_alloc.c:235!
Jul 11 11:03:29 screech kernel: invalid operand: 0000
Jul 11 11:03:29 screech kernel: CPU:    0
Jul 11 11:03:29 screech kernel: EIP:    0010:[<c012f69e>]    Not tainted
Jul 11 11:03:29 screech kernel: EFLAGS: 00010202
Jul 11 11:03:29 screech kernel: eax: 00000040   ebx: c128cc18   ecx: 00001000   edx: 0000ed5d
Jul 11 11:03:29 screech kernel: esi: c02b86b8   edi: 00000001   ebp: 00000000   esp: d062be40
Jul 11 11:03:29 screech kernel: ds: 0018   es: 0018   ss: 0018
Jul 11 11:03:29 screech kernel: Process imapd (pid: 31916, stackpage=d062b000)
Jul 11 11:03:29 screech kernel: Stack: 00001000 0000dd5d 00000282 00000000 c02b86b8 c02b888c 00000001 0000000c 
Jul 11 11:03:29 screech kernel:        00000001 c012f964 c02b86b8 db31f7c0 ce38d8c0 ce38d8c0 c02b86b8 c02b8888 
Jul 11 11:03:29 screech kernel:        00000000 000001d2 ce38dd40 c138f4f4 00000001 ce38dd40 db9c98c0 c012590d 
Jul 11 11:03:29 screech kernel: Call Trace:    [<c012f964>] [<c012590d>] [<c0125eb4>] [<c0125a58>] [<c01264af>]
Jul 11 11:03:29 screech kernel:   [<c0114806>] [<c010d45a>] [<c012b48d>] [<c013550d>] [<c0114690>] [<c0108794>]
Jul 11 11:03:29 screech kernel: Code: 0f 0b eb 00 73 ff 26 c0 8b 43 18 a9 80 00 00 00 74 08 0f 0b 

Trace; c012f964 <__alloc_pages+74/2c0>
Trace; c012590d <do_no_page+8d/180>
Trace; c0125eb4 <__vma_link+64/c0>
Trace; c0125a58 <handle_mm_fault+58/c0>
Trace; c01264af <do_mmap_pgoff+4af/560>
Trace; c0114806 <do_page_fault+176/4ab>
Trace; c010d45a <old_mmap+ea/120>
Trace; c012b48d <sys_mprotect+11d/208>
Trace; c013550d <filp_close+4d/60>
Trace; c0114690 <do_page_fault+0/4ab>
Trace; c0108794 <error_code+34/3c>
Code; c012f69e <rmqueue+1fe/240>
00000000 <_EIP>:
Code; c012f69e <rmqueue+1fe/240> <=====
0: 0f 0b ud2a <=====
Code; c012f6a0 <rmqueue+200/240>
2: eb 00 jmp 4 <_EIP+0x4> c012f6a2 <rmqueue+202/240>
Code; c012f6a2 <rmqueue+202/240>
4: 73 ff jae 5 <_EIP+0x5> c012f6a3 <rmqueue+203/240>
Code; c012f6a4 <rmqueue+204/240>
6: 26 c0 8b 43 18 a9 80 rorb $0x0,%es:0x80a91843(%ebx)
Code; c012f6ab <rmqueue+20b/240>
d: 00
Code; c012f6ac <rmqueue+20c/240>
e: 00 00 add %al,(%eax)
Code; c012f6ae <rmqueue+20e/240>
10: 74 08 je 1a <_EIP+0x1a> c012f6b8 <rmqueue+218/240>
Code; c012f6b0 <rmqueue+210/240>
12: 0f 0b ud2a

18 warnings issued. Results may not be reliable.

Hi Jan,

The BUG() is at rmqueue(), "if (PageLRU(page)) BUG".

The same BUG has been seen before. The BUG means an LRU page was found in a zone's free list (this is not a valid thing to happen).

This could be either a driver bug (which uses a page after freeing it, as wli smartly pointed out at the time) or faulty hardware, as it showed to be in that occasion.

What drivers are you using and, just as matter of "safety", have you ran memtest on this box?

Marcelo> Hi Jan,

Marcelo> The BUG() is at rmqueue(), "if (PageLRU(page)) BUG".

Marcelo> The same BUG has been seen before. The BUG means an LRU page Marcelo> was found in a zone's free list (this is not a valid thing to Marcelo> happen).

Marcelo> This could be either a driver bug (which uses a page after Marcelo> freeing it, as wli smartly pointed out at the time) or faulty Marcelo> hardware, as it showed to be in that occasion.

Marcelo> What drivers are you using

8139too for networking, reiserfs and ext2 for filesystems. Not much more, really. IDE drives. iptable_nat + ip_conntrack. TUN/TAP. That's pretty much it.

Marcelo> and, just as matter of "safety", have you ran memtest on this Marcelo> box?

Unfortunately, no -- and it isn't all that easy for me to do so (the box is remotely hosted). I will try to do it...

--J.