I think that lio_listio() is supposed to do operations in-order, so
Phil's implementation, while perfect for us, doesn't behave as it should
for POSIX. This is probably also why the glibc code only posts one
operation on a FD at a time.
Rob
Phil Carns wrote:
Sam Lang wrote:
[snip]
Does it make sense to factor that out so that it could be used as a
separate implementation for others (LD_PRELOAD=libfastaio.so :-))? I
guess the only advantage pvfs would gain from that would be the
performance comparisons of the two different lio_listio
implementations from all the aio tests out already out there. A
disadvantage might be that you can't switch between one and the other
dynamically at runtime.
I would imagine this is possible, but probably more work than I have
time to tinker with :) You could probably even play some trick to make
it controllable at runtime (looking at the value of a global variable if
defined, or maybe latching hold of the aio_init() function to control it).
_______________________________________________
Pvfs2-developers mailing list
[email protected]
http://www.beowulf-underground.org/mailman/listinfo/pvfs2-developers
