From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S262341AbVATQfX (ORCPT ); Thu, 20 Jan 2005 11:35:23 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S262200AbVATQfM (ORCPT ); Thu, 20 Jan 2005 11:35:12 -0500 Received: from fw.osdl.org ([65.172.181.6]:30674 "EHLO mail.osdl.org") by vger.kernel.org with ESMTP id S262270AbVATQbU (ORCPT ); Thu, 20 Jan 2005 11:31:20 -0500 Date: Thu, 20 Jan 2005 08:31:11 -0800 (PST) From: Linus Torvalds To: Ingo Molnar cc: Peter Chubb , Chris Wedgwood , Andrew Morton , paulus@samba.org, linux-kernel@vger.kernel.org, tony.luck@intel.com, dsw@gelato.unsw.edu.au, benh@kernel.crashing.org, linux-ia64@vger.kernel.org, hch@infradead.org, wli@holomorphy.com, jbarnes@sgi.com Subject: Re: [patch 1/3] spinlock fix #1, *_can_lock() primitives In-Reply-To: <20050120160839.GA13067@elte.hu> Message-ID: References: <16877.42598.336096.561224@wombat.chubb.wattle.id.au> <20050119080403.GB29037@elte.hu> <16878.9678.73202.771962@wombat.chubb.wattle.id.au> <20050119092013.GA2045@elte.hu> <16878.54402.344079.528038@cargo.ozlabs.ibm.com> <20050120023445.GA3475@taniwha.stupidest.org> <20050119190104.71f0a76f.akpm@osdl.org> <20050120031854.GA8538@taniwha.stupidest.org> <16879.29449.734172.893834@wombat.chubb.wattle.id.au> <20050120160839.GA13067@elte.hu> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Thu, 20 Jan 2005, Ingo Molnar wrote: > > anyway, here's my first patch again, with s/trylock_test/can_lock/. I don't want to break all the other architectures. Or at least not most of them. Especially since I was hoping to do a -pre2 soon (well, like today, but I guess that's out..) and make the 2.6.11 cycle shorter than 2.6.10. So I'd like to now _first_ get > spin_can_lock(lock) > read_can_lock(lock) > write_can_lock(lock) for at least most architectures (ie for me at a minimum that is x86, x86-64, ia64 and ppc64 - and obviously the "always true" cases for the UP version). Ok? Also, I've already made sure that I can't apply any half-measures by mistake by undoing the mess that it was before, and making sure that any patches I get have to be "clean slate". That said, I like how just the _renaming_ of the thing (and making them all consistent) made your BUILD_LOCK_OPS() helper macro much simpler. So I'm convinced that this is the right solution - I just want to not screw up other architectures. I can do ppc64 myself, can others fix the other architectures (Ingo, shouldn't the UP case have the read/write_can_lock() cases too? And wouldn't you agree that it makes more sense to have the rwlock test variants in asm/rwlock.h?): Linus > --- linux/include/linux/spinlock.h.orig > +++ linux/include/linux/spinlock.h > @@ -584,4 +584,10 @@ static inline int bit_spin_is_locked(int > #define DEFINE_SPINLOCK(x) spinlock_t x = SPIN_LOCK_UNLOCKED > #define DEFINE_RWLOCK(x) rwlock_t x = RW_LOCK_UNLOCKED > > +/** > + * spin_can_lock - would spin_trylock() succeed? > + * @lock: the spinlock in question. > + */ > +#define spin_can_lock(lock) (!spin_is_locked(lock)) > + > #endif /* __LINUX_SPINLOCK_H */ > --- linux/include/asm-i386/spinlock.h.orig > +++ linux/include/asm-i386/spinlock.h > @@ -188,6 +188,18 @@ typedef struct { > > #define rwlock_is_locked(x) ((x)->lock != RW_LOCK_BIAS) > > +/** > + * read_can_lock - would read_trylock() succeed? > + * @lock: the rwlock in question. > + */ > +#define read_can_lock(x) (atomic_read((atomic_t *)&(x)->lock) > 0) > + > +/** > + * write_can_lock - would write_trylock() succeed? > + * @lock: the rwlock in question. > + */ > +#define write_can_lock(x) ((x)->lock == RW_LOCK_BIAS) > + > /* > * On x86, we implement read-write locks as a 32-bit counter > * with the high bit (sign) being the "contended" bit. >