[parisc-linux] Re: [patch] spinlock consolidation, v2

[Ingo Molnar <mingo@elte.hu>]

* Grant Grundler <grundler@parisc-linux.org> wrote:

> IPI and TLB locks can be normal locks.
> I've dropped the parts that changed those.
> 
> New (smaller) patch parked at:
>     ftp://ftp.parisc-linux.org/patches/consolidate-spinlocks-parisc.patch-02

thanks. I'm still wondering about the fundamental question though: why 
doesnt the box boot without your patch? Why does the __atomic_hash and 
the pending_lock have to be converted to a raw_spinlock? This patch does 
not change anything fundamental - you are still getting the same 
spinlock primitives from spinlock_t. The only thing that changes is with 
spinlock debugging: there another (generic) piece of code kicks in. The 
only problem could be with assembly code that 'knows' the layout of the 
spinlocks but doesnt take debugging into account - but is this possible 
in the PARISC case? (Yet another possibility would be if i messed up the 
raw type completely, but then your box wouldnt boot at all.)

so this is still quite much of a mystery to me!

	Ingo
_______________________________________________
parisc-linux mailing list
parisc-linux@lists.parisc-linux.org
http://lists.parisc-linux.org/mailman/listinfo/parisc-linux