Re: Bug with shared memory.

[Martin J. Bligh]

>> I've been benchmarking rmap 13 against mainline (2.4.19-pre7)
>> and with the latest lock breakup changes performance now seems
>> to be about equivalent to mainline (for kernel compile on NUMA-Q).
>> Those changes reduced system time from 650s to 160s. The only
> 
> How much are you swapping in your workload? (as said the fast paths are
> hurted a little so it's expected that it's almost as fast as mainline
> with a kernel compile, similar to the fact we also add anon pages to the
> lru list). I think you're only exercising the fast paths in your
> workload, not the memory balancing that is the whole point of the change.

No swapping. We fixed the horrendous locking problem we were seeing,
but this was only one test - obviously others are needed. But I think we're
in agreement that it's time to give it a beating and see what happens ;-)

M.