Re: Another BUG in page_alloc.c:108

[Zwane Mwaikambo]

On Wed, 3 Apr 2002 [e-mail] wrote:

> In article <20020403035406.GA2925@neon> you wrote:
> > EIP:    0010:[__free_pages_ok+45/688]    Tainted: P 
> 
> Nvidia ? 
> I get the distinct impression that the lastest nvidia drivers
> reintroduced a bug that fubars the page allocator ;(

The latest nvidia stuff definately spews major chunks on its way out...

This is on 2.4.19-pre2-ac3, backing upto the previous release nvidia 
drivers i can't reproduce.

invalid operand: 0000
CPU:    0
EIP:    0010:[<c0130c47>]    Not tainted
EFLAGS: 00013282
eax: 00000000   ebx: c15f94b0   ecx: c100000c   edx: db140eb0
esi: 00000000   edi: 00000000   ebp: 00007000   esp: deb35edc
ds: 0018   es: 0018   ss: 0018
Process X (pid: 1457, stackpage=deb35000)
Stack: c0306380 c15f94b0 00000000 00000001 c15f94b0 00008000 c15f94b0 
c15f94b0 
       00008000 debb5688 00007000 c0124782 c15f94b0 1d68d027 00000008 
00000000 
       425a3000 de793424 4259b000 00000000 425a3000 de793424 dfffcd50 
dec00420 
Call Trace: [<c0124782>] [<c01270ed>] [<c01271a2>] [<c0106f2b>] 

Code: 0f 0b c6 43 24 05 8b 43 18 89 f1 83 e0 eb 89 43 18 c1 e8 18

>>EIP; c0130c47 <__free_pages_ok+97/250>   <=====
Trace; c0124782 <zap_page_range+192/260>
Trace; c01270ed <do_munmap+1ed/270>
Trace; c01271a2 <sys_munmap+32/50>
Trace; c0106f2b <system_call+33/38>
Code;  c0130c47 <__free_pages_ok+97/250>
00000000 <_EIP>:
Code;  c0130c47 <__free_pages_ok+97/250>   <=====
   0:   0f 0b                     ud2a      <=====
Code;  c0130c49 <__free_pages_ok+99/250>
   2:   c6 43 24 05               movb   $0x5,0x24(%ebx)
Code;  c0130c4d <__free_pages_ok+9d/250>
   6:   8b 43 18                  mov    0x18(%ebx),%eax
Code;  c0130c50 <__free_pages_ok+a0/250>
   9:   89 f1                     mov    %esi,%ecx
Code;  c0130c52 <__free_pages_ok+a2/250>
   b:   83 e0 eb                  and    $0xffffffeb,%eax
Code;  c0130c55 <__free_pages_ok+a5/250>
   e:   89 43 18                  mov    %eax,0x18(%ebx)
Code;  c0130c58 <__free_pages_ok+a8/250>
  11:   c1 e8 18                  shr    $0x18,%eax

0xc0130c40 <__free_pages_ok+144>:       mov    0x28(%ebx),%edx
0xc0130c43 <__free_pages_ok+147>:       test   %edx,%edx
0xc0130c45 <__free_pages_ok+149>:       je     0xc0130c49 
<__free_pages_ok+153>
0xc0130c47 <__free_pages_ok+151>:       ud2a

if (page->pte_chain)
                BUG();

-- 
http://function.linuxpower.ca