Re: [patch 0/9] mutex subsystem, -V4

[Christoph Hellwig <hch@infradead.org>]

On Thu, Dec 22, 2005 at 03:30:14PM -0800, Andrew Morton wrote:
> No it does not.
> 
> Ingo's work has shown us two things:
> 
> a) semaphores use more kernel text than they should and
> 
> b) semaphores are less efficient than they could be.
> 
> Fine.  Let's update the semaphore implementation to fix those things. 
> Nobody has addressed this code in several years.  If we conclusively cannot
> fix these things then that's the time to start looking at implementing new
> locking mechanisms.

c) semaphores are total overkill for 99% percent of the users.  Remember
this thing about optimizing for the common case?

Pretty much everywhere we do want mutex semantic.  So let's have a proper
primitive exactly for that, and we can keep the current semaphore
implementation (with a much simpler implementation) for that handfull of
users in the kernel that really want a counting semaphore.

I really don't get why you hate mutex primitives so much.