[openib-general] Re: CMA deadlock

[Michael S. Tsirkin]

Quoting r. Sean Hefty <sean.hefty at intel.com>:
> This should be handled by the code.  See the comment near the bottom of the
> cma_ib_handler() routine.

BTW, am I right in sayng that error handling is mostly untested?
If so, a variant of cmatose that fails the connection randomly at various stages
would be quite useful, wouldn't it?

-- 
Michael S. Tsirkin
Staff Engineer, Mellanox Technologies