On Thu, Feb 07, 2002 at 03:08:02PM -0500, Robert Love wrote:
> On Thu, 2002-02-07 at 14:58, yodaiken@fsmlabs.com wrote:
>
> > On Thu, Feb 07, 2002 at 01:40:59PM -0500, Robert Love wrote:
> > > We shouldn't engage in wholesale changing of spinlocks to semaphores
> > > without a priority-inheritance mechanism. And _that_ is the bigger
> > > issue ...
> >
> > Cool. We can then have the Solaris "this usually doesn't fail on test" priority
> > inherit read/write lock. I can hardly wait.
>
> Or, we could do things right and not.
I'd love to hear how things could be done right here.
There seem to be 3 choices for reader writer locks
1. Do the right thing and say no to inheritance: and this
means no inheritance on mutexes either.
2. Use the Solaris - "sometimes kinda works" method.
3. Make readers/writer locks very slow and expensive e.g
a complete list of reader identities that with atomic insert/delete
and with check for uniqueness on insert! Not to mention the write
promotion, any interactions between the "favor writes" design it should
have and inheritance, links for a mutex inheriting lock to follow down
the complete tree of paths from the r/w lock ...
-- --------------------------------------------------------- Victor Yodaiken Finite State Machine Labs: The RTLinux Company. www.fsmlabs.com www.rtlinux.com- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Thu Feb 07 2002 - 21:01:06 EST