On Aug 7, 2007, at 3:08 PM, Pete Wyckoff wrote:
The code should check that bmi_max_unexp_payload > small_io_size
before returning.
I will also increase the unex size in bmi_mx to 8 KB to avoid having
to modify PVFS2.
No, do it right in the sys-io code. You can up the BMI_MX limit too
if you think that makes sense, but that's a separate patch to this
bug fix you've found. We have a couple new BMI transports here or
on the way that don't want to have to rediscover this. I certainly
appreciate that you found this!
-- Pete
Pete,
I assumed that small_io_size is fixed in PVFS2 and was greater than 4
KB, which why I volunteered to change bmi_mx. I chose 4 KB for bmi_mx
simply because that was the value I used in Lustre (kernel page
size). I am not wedded to 4 KB.
Scott
_______________________________________________
Pvfs2-developers mailing list
[email protected]
http://www.beowulf-underground.org/mailman/listinfo/pvfs2-developers
