Subject: Re: MP/Threading...
To: Greywolf <firstname.lastname@example.org>
From: Bill Sommerfeld <email@example.com>
Date: 03/20/2001 13:26:08
> I was seriously hoping that BigLock would never see an official release.
> ...or is that entirely unreasonable (in which case I apologise for
> my unreasoning)?
The general feeling is that going straight to fine-grained locking is
just not achievable given the resources available..
Having biglock in -current allows work on the (very significant)
amounts of work needed on each MP platform (for TLB shootdown,
IPI/cross-calls, interrupt model changes, etc.,) to proceed in a way
at least partly decoupled from changes to the MI parts of the kernel..