> I figured out why I was hitting the BUG_ON() in page_remove_rmap(). Its > because my library was destroying the QP or CQ object _before_ unmaping > the objects. I changed it to unmap first, and things work as expected.
I guess kernel side also needs to be fixed, otherwise buggy userspace can trigger BUG_ON's? -- MST _______________________________________________ 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
