Re: hash table sizes

[Martin J. Bligh]

>> It would be very nice to have some confirmation that the size of these
>> tables is being appropriately chosen, too.  Maybe we should shrink 'em 32x
>> and see who complains...
> 
> Why dont we just do node round robin allocations during boot? This
> should mean the static boot time hashes would at least end up on
> different nodes.

We could probably implement a generic striped allocate, which would
do a vmalloc or similar on 64 bit, and either the magic boottime
node-alloc hack, or just a straight node 0 alloc on 32 bit (ie use
vmalloc where needed, without crippling other platforms).

Someone had a patch to do round-robin already (Manfred?) - IMHO doing
it from the node with the most free mem each time would be better, if
we're not going to stripe. 
 
> 0 248652
> 1 7374

...

but yes, that does look utterly screwed ;-)

M.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/