>> Later, Jukka reported that running his application with an X server
>> that does not use mmap/munmap, made things go much faster.  I suspect
>> that the application is still doing small writes, but the round-trip
>> time between the application and the X-server is now low enough
>> that the small packet size is masked.
>
>Yes, I believe this is the case.  I still haven't tried building a
>profiled version of the X server to see why it runs so slow with
>mmap/mumnap enabled, but I've been satisfied with the X performance
>with the old memory allocation scheme.

I don't think that is the case.  The real problem (as I saw it) was that
the mmap allocator was just taking a long time to run.  The fact that
it was allocating memory for _every request_ is actually a bit crazy.

--Ken