Peter Tillemans ([EMAIL PROTECTED]) wrote on 18 December 1998 15:35:
>I have the same error using gcc-2.7.2.3.
Good, I mean, it's not an egcs problem :-)
...
>This morning the same 'Kernel panic hm ??'
Did you see Ingo's reply to the list? I enclose it here anyway. Our
main central server just locked twice a few minutes ago because of
this problem, before I had the chance of installing Ingo's fix. I've
just done it, let's see if it works.
MOLNAR Ingo ([EMAIL PROTECTED]) wrote on 18 December 1998 14:19:
>
>On Fri, 18 Dec 1998, Carlos Carvalho wrote:
>
>> Trying this list first, if it's not raid-related I'll go to linux-kernel...
>>
>> Our raid5 server locked again, with the message "kernel panic hm ??"
>> displayed 3 times on the console. This is the second time.
>
>ugh, sorry, this is my goof. This is in the 2.1.125 patch:
>
>+ if (steps >= NR_REQUEST)
>+ panic("hm??");
>
>it should be '>' instead of '>=' ... sorry! (the fact that it locks up so
>rarely shows that the IO request queue is sized properly and the RAID5
>daemons queue stuff fast enough too, only very rarely filling up the
>request queue ...)
>
>i fully agree that this is not the way we'd like to see this though ...
>
>i have again reviewed that part of the code, and everything else is OK.
>
>-- mingo