[openib-general] [PATCH] opensm: fixes in signal handling

[Fabian Tillier]

On 2/21/06, Sasha Khapyorsky <sashak at voltaire.com> wrote:
> Hello,
>
> There are fixes for broken signal handling stuff. Finally this should
> prevent some opensm crashes (actually deadlocks), for instance caused
> by such command:
>
>  while [ 1 ] ; do kill -HUP <opensm-pid> ; done
>
>
> Yael, I hope this patch does not broke windows compilations (this masks
> signal related functions under __WIN__), but cannot be absolutely sure -
> please look at this from "under windows" point of view. Thanks.

Haven't we discussed eliminating the signal handling from OpenSM
before, since it was put in place to work around access layer bugs? 
Is there any reason to keep signal handling in OpenSM?

- Fab