[Gc] patch to handle EINTR

Paolo Molaro lupus at debian.org
Wed Jan 19 07:47:02 PST 2005


On 01/18/05 Boehm, Hans wrote:
> I actually already had a more verbose version of this in the gc7 tree.
> I put your version into the gc6.5 tree.
> 
> Is this actually supposed to happen somewhere?  I observed this on Linux
> when running under gdb.  But the sem_wait documentation on Linux very
> explicitly states that this can't happen.  Hence the current code
> complains before ignoring it.

I noticed the issue when using gdb, too.
I think the documentation in the manpage is wrong and we shouldn't
rely on it: any syscall could be interrupted and needs to deal with
EINTR (I think this was mentioned recently on linux-kernel, too).
Thanks.

lupus

-- 
-----------------------------------------------------------------
lupus at debian.org                                     debian/rules
lupus at ximian.com                             Monkeys do it better


More information about the Gc mailing list