Re: [2.6.9-rc1-bk14 Oops] In groups_search()

[Peter Williams <pwil3058@bigpond.net.au>]

Peter Williams wrote:
> Andrew Morton wrote:
> 
>>
>> Could you see if this patch fixes the above crash?
>>
>> --- 25/fs/isofs/rock.c~rock-kludge    2004-09-10 00:52:30.394468656 -0700
>> +++ 25-akpm/fs/isofs/rock.c    2004-09-10 00:53:14.544756792 -0700
>> @@ -62,7 +62,7 @@
>>  }                                      
>>  #define MAYBE_CONTINUE(LABEL,DEV) \
>> -  {if (buffer) kfree(buffer); \
>> +  {if (buffer) { kfree(buffer); buffer = NULL; } \
>>    if (cont_extent){ \
>>      int block, offset, offset1; \
>>      struct buffer_head * pbh; \
>> _
>>
>>
>> I sure hope it does, so I don't have to look at rock.c again.
> 
> 
> It does and no sign of the oops or scheduling while atomic messages.

I was mistaken about the schedule while atomic messages.  They showed up 
again when I did the "make install".

>  I 
> still have the original four patches applied.  I'll try again with an 
> unpatched bk16 and let you know the results shortly.
> 

With out of the box bk16 plus your rock.c patch and with 
CONFIG_DEBUG_SLAB and CONFIG_DEBUG_PAGEALLOC selected I get no oops in 
in_groupse_p() or kfree() but I still get the scheduling while atomic 
messages when I do "make install".

Peter
-- 
Peter Williams                                   pwil3058@bigpond.net.au

"Learning, n. The kind of ignorance distinguishing the studious."
  -- Ambrose Bierce