Subject: Re: alternate rough SA patch (works on SMP)
To: Christian Limpach <firstname.lastname@example.org>
From: Stephan Uphoff <email@example.com>
Date: 06/27/2003 15:28:02
Christian Limpach wrote:
> I've also come up with a patch for pthread/SMP.
> Is this going in the right direction?
I finally had some time to look at your patch.
Here are a few comments:
- With your patch a LWP that had been blocked (block upcall received by
libpthread) can return later to user space without the associated
unblock upcall or holding the virtual processor.
This will confuse libpthread.
- ltsleep and therefore also copyout can not be called with SCHED_LOCK held.
- as a target for mi_switch the lwp has to be in memory (l_flag & L_INMEM)