[PATCH][2.5] smp_call_function needs mb()

[PATCH][2.5] smp_call_function needs mb()

[Zwane Mwaikambo]

Tested on afflicted 3way P133 system for 3days and on 8way P3 700 for 
regression.

Unable to handle kernel NULL pointer dereference at virtual address 00000200
 printing eip:
00000200
*pde = 00000000
Oops: 0000 [#1]
CPU:    2
EIP:    0060:[<00000200>]    Not tainted
EFLAGS: 00210082
EIP is at 0x200
eax: 00000026   ebx: c7666000   ecx: c7666000   edx: c7666000
esi: 00000200   edi: c034ad74   ebp: c0bc8000   esp: c7667fa8
ds: 007b   es: 007b   ss: 0068
Process rhn-applet (pid: 1654, threadinfo=c7666000 task=c0785340)
Stack: c0116407 c034ad74 40a2d760 08458490 00000004 bfffec78 c010a41a 40a2d760 
       00000000 00000000 08458490 00000004 bfffec78 0830d800 0000007b 0000007b 
       fffffffb 40a23da4 00000073 00200202 bfffec48 0000007b 
Call Trace:
 [<c0116407>] smp_call_function_interrupt+0x57/0xb0
 [<c034ad74>] sr_do_ioctl+0x124/0x250
 [<c010a41a>] call_function_interrupt+0x1a/0x20

Code:  Bad EIP value.
 <0>Kernel panic: Aiee, killing interrupt handler!

Index: linux-2.5.66/arch/i386/kernel/smp.c
===================================================================
RCS file: /build/cvsroot/linux-2.5.66/arch/i386/kernel/smp.c,v
retrieving revision 1.1.1.1
diff -u -p -B -r1.1.1.1 smp.c
--- linux-2.5.66/arch/i386/kernel/smp.c	24 Mar 2003 23:40:27 -0000	1.1.1.1
+++ linux-2.5.66/arch/i386/kernel/smp.c	28 Mar 2003 05:08:54 -0000
@@ -522,7 +521,8 @@ int smp_call_function (void (*func) (voi
 
 	spin_lock(&call_lock);
 	call_data = &data;
-	wmb();
+	mb();
+	
 	/* Send a message to all other CPUs and wait for them to respond */
 	send_IPI_allbutself(CALL_FUNCTION_VECTOR);
 

Re: [PATCH][2.5] smp_call_function needs mb()

[Zwane Mwaikambo]

On Tue, 1 Apr 2003, Joe Korty wrote:

> This should be sent to Marcello as well.
> Joe

I'll try oops 2.4.20 and send.

-- 
function.linuxpower.ca

Re: [PATCH][2.5] smp_call_function needs mb()

[Zwane Mwaikambo]

On Tue, 1 Apr 2003, Zwane Mwaikambo wrote:

> On Tue, 1 Apr 2003, Joe Korty wrote:
> 
> > This should be sent to Marcello as well.
> > Joe
> 
> I'll try oops 2.4.20 and send.

2.4.21-pre5-ac3 Nothing happened.. mainly because the whole box was 
completely useless during the whole test run. There was no real progress 
during the entire run so i stoppped it. I'll send to Alan anyway

305485 total                                      0.1583
246390 default_idle                             3849.8438
 11361 .text.lock.inode                          50.9462
  9027 generic_file_write                         4.7411
  1374 invalidate_bdev                            2.5257
  1166 do_get_write_access                        0.7288
  1116 journal_add_journal_head                   2.2500
  1061 __make_request                             0.6197
   913 shrink_cache                               0.5764
   872 kmem_cache_reap                            0.5737
   848 unlock_buffer                             10.6000
   846 journal_dirty_metadata                     1.7625
   796 .text.lock.sched                           1.5191
   792 do_anonymous_page                          1.7679
   755 kmem_cache_alloc                           1.1509
   738 get_hash_table                             3.8438
   688 .text.lock.buffer                          1.1486
   630 rmqueue                                    0.7572
   620 ext3_do_update_inode                       0.7045
   613 __rdtsc_delay                             19.1562
   588 do_wp_page                                 0.6682
   574 write_some_buffers                         1.8882
   560 journal_get_write_access                   5.8333
   542 journal_unlock_journal_head                2.8229
   524 journal_cancel_revoke                      2.5192
   519 schedule                                   0.4569
   502 try_to_free_buffers                        0.9508
   460 __wake_up                                  2.6136
   454 kmalloc                                    0.6168
   399 ext3_new_block                             0.1651
   379 .text.lock.ioctl                           9.0238

Re: [PATCH][2.5] smp_call_function needs mb()

[Zwane Mwaikambo]

Some vmstat too...

 procs                      memory      swap          io     system      cpu
 r  b  w   swpd   free   buff  cache   si   so    bi    bo   in    cs  us sy id
 1 129  3  25416   3348   7504  96220    0   76  3356  1252  414   760 42 14 44
17 115  3  25416   2368   7592  97004   32  156  2080  7420  453   690 21 54 26
 2 128  2  25416   3484   7780  96768    0   16  1108   988  296   572 39 41 20
 2 128  3  25416   3980   8284  93964    0    0   524   504  443   873 43 21 36
 2 127  2  25416   3488   8816  91616   32   44   500  1944  468   752 35 44 21
 4 127  2  25568   3956   9208  90544    0  308   340  1420  374   457 14 79  7


-- 
function.linuxpower.ca