On an unrelated note, it looks like cm.c would benefit from an irqsave
diet: it seems to perform work almost exclusively from thread
context, so just spin_lock_irq is sure to be enough.

I don't think everything is done at thread context - that depends on the ULP, but it could definitely replace irqsave with just irq in any of the message processing code.

- Sean
_______________________________________________
general mailing list
[email protected]
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general

To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general

Reply via email to