From: Peter Zijlstra <peterz@infradead.org> To: Ingo Molnar <mingo@kernel.org> Cc: Mark Rutland <mark.rutland@arm.com>, linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org, aryabinin@virtuozzo.com, boqun.feng@gmail.com, catalin.marinas@arm.com, dvyukov@google.com, will.deacon@arm.com Subject: Re: [PATCH] locking/atomics: Clean up the atomic.h maze of #defines Date: Sat, 5 May 2018 11:24:38 +0200 [thread overview] Message-ID: <20180505092438.GB12217@hirez.programming.kicks-ass.net> (raw) In-Reply-To: <20180505090403.p2ywuen42rnlwizq@gmail.com> On Sat, May 05, 2018 at 11:04:03AM +0200, Ingo Molnar wrote: > > * Peter Zijlstra <peterz@infradead.org> wrote: > > > > So we could do the following simplification on top of that: > > > > > > #ifndef atomic_fetch_dec_relaxed > > > # ifndef atomic_fetch_dec > > > # define atomic_fetch_dec(v) atomic_fetch_sub(1, (v)) > > > # define atomic_fetch_dec_relaxed(v) atomic_fetch_sub_relaxed(1, (v)) > > > # define atomic_fetch_dec_acquire(v) atomic_fetch_sub_acquire(1, (v)) > > > # define atomic_fetch_dec_release(v) atomic_fetch_sub_release(1, (v)) > > > # else > > > # define atomic_fetch_dec_relaxed atomic_fetch_dec > > > # define atomic_fetch_dec_acquire atomic_fetch_dec > > > # define atomic_fetch_dec_release atomic_fetch_dec > > > # endif > > > #else > > > # ifndef atomic_fetch_dec > > > # define atomic_fetch_dec(...) __atomic_op_fence(atomic_fetch_dec, __VA_ARGS__) > > > # define atomic_fetch_dec_acquire(...) __atomic_op_acquire(atomic_fetch_dec, __VA_ARGS__) > > > # define atomic_fetch_dec_release(...) __atomic_op_release(atomic_fetch_dec, __VA_ARGS__) > > > # endif > > > #endif > > > > This would disallow an architecture to override just fetch_dec_release for > > instance. > > Couldn't such a crazy arch just define _all_ the 3 APIs in this group? > That's really a small price and makes the place pay the complexity > price that does the weirdness... I would expect the pattern where it can do all 'release' and/or all 'acquire' variants special but cannot use the __atomic_op_*() wrappery. > > I don't think there currently is any architecture that does that, but the > > intent was to allow it to override anything and only provide defaults where it > > does not. > > I'd argue that if a new arch only defines one of these APIs that's probably a bug. > If they absolutely want to do it, they still can - by defining all 3 APIs. > > So there's no loss in arch flexibility. Ideally we'd generate the whole mess.. and then allowing these extra few overrides is not a problem at all. > > None of this takes away the giant trainwreck that is the annotated atomic stuff > > though. > > > > And I seriously hate this one: > > > > ba1c9f83f633 ("locking/atomic/x86: Un-macro-ify atomic ops implementation") > > > > and will likely undo that the moment I need to change anything there. > > If it makes the code more readable then I don't object - the problem was that the > instrumentation indirection made all that code much harder to follow. Thing is, it is all the exact same loop, and bitrot mandates they drift over time. When I cleaned up all the architectures I found plenty cases where there were spurious differences between things.
WARNING: multiple messages have this Message-ID (diff)
From: peterz@infradead.org (Peter Zijlstra) To: linux-arm-kernel@lists.infradead.org Subject: [PATCH] locking/atomics: Clean up the atomic.h maze of #defines Date: Sat, 5 May 2018 11:24:38 +0200 [thread overview] Message-ID: <20180505092438.GB12217@hirez.programming.kicks-ass.net> (raw) In-Reply-To: <20180505090403.p2ywuen42rnlwizq@gmail.com> On Sat, May 05, 2018 at 11:04:03AM +0200, Ingo Molnar wrote: > > * Peter Zijlstra <peterz@infradead.org> wrote: > > > > So we could do the following simplification on top of that: > > > > > > #ifndef atomic_fetch_dec_relaxed > > > # ifndef atomic_fetch_dec > > > # define atomic_fetch_dec(v) atomic_fetch_sub(1, (v)) > > > # define atomic_fetch_dec_relaxed(v) atomic_fetch_sub_relaxed(1, (v)) > > > # define atomic_fetch_dec_acquire(v) atomic_fetch_sub_acquire(1, (v)) > > > # define atomic_fetch_dec_release(v) atomic_fetch_sub_release(1, (v)) > > > # else > > > # define atomic_fetch_dec_relaxed atomic_fetch_dec > > > # define atomic_fetch_dec_acquire atomic_fetch_dec > > > # define atomic_fetch_dec_release atomic_fetch_dec > > > # endif > > > #else > > > # ifndef atomic_fetch_dec > > > # define atomic_fetch_dec(...) __atomic_op_fence(atomic_fetch_dec, __VA_ARGS__) > > > # define atomic_fetch_dec_acquire(...) __atomic_op_acquire(atomic_fetch_dec, __VA_ARGS__) > > > # define atomic_fetch_dec_release(...) __atomic_op_release(atomic_fetch_dec, __VA_ARGS__) > > > # endif > > > #endif > > > > This would disallow an architecture to override just fetch_dec_release for > > instance. > > Couldn't such a crazy arch just define _all_ the 3 APIs in this group? > That's really a small price and makes the place pay the complexity > price that does the weirdness... I would expect the pattern where it can do all 'release' and/or all 'acquire' variants special but cannot use the __atomic_op_*() wrappery. > > I don't think there currently is any architecture that does that, but the > > intent was to allow it to override anything and only provide defaults where it > > does not. > > I'd argue that if a new arch only defines one of these APIs that's probably a bug. > If they absolutely want to do it, they still can - by defining all 3 APIs. > > So there's no loss in arch flexibility. Ideally we'd generate the whole mess.. and then allowing these extra few overrides is not a problem at all. > > None of this takes away the giant trainwreck that is the annotated atomic stuff > > though. > > > > And I seriously hate this one: > > > > ba1c9f83f633 ("locking/atomic/x86: Un-macro-ify atomic ops implementation") > > > > and will likely undo that the moment I need to change anything there. > > If it makes the code more readable then I don't object - the problem was that the > instrumentation indirection made all that code much harder to follow. Thing is, it is all the exact same loop, and bitrot mandates they drift over time. When I cleaned up all the architectures I found plenty cases where there were spurious differences between things.
next prev parent reply other threads:[~2018-05-05 9:24 UTC|newest] Thread overview: 103+ messages / expand[flat|nested] mbox.gz Atom feed top 2018-05-04 17:39 [PATCH 0/6] arm64: add instrumented atomics Mark Rutland 2018-05-04 17:39 ` Mark Rutland 2018-05-04 17:39 ` [PATCH 1/6] locking/atomic, asm-generic: instrument ordering variants Mark Rutland 2018-05-04 17:39 ` Mark Rutland 2018-05-04 18:01 ` Peter Zijlstra 2018-05-04 18:01 ` Peter Zijlstra 2018-05-04 18:09 ` Mark Rutland 2018-05-04 18:09 ` Mark Rutland 2018-05-04 18:24 ` Peter Zijlstra 2018-05-04 18:24 ` Peter Zijlstra 2018-05-05 9:12 ` Mark Rutland 2018-05-05 9:12 ` Mark Rutland 2018-05-05 8:11 ` [PATCH] locking/atomics: Clean up the atomic.h maze of #defines Ingo Molnar 2018-05-05 8:11 ` Ingo Molnar 2018-05-05 8:36 ` [PATCH] locking/atomics: Simplify the op definitions in atomic.h some more Ingo Molnar 2018-05-05 8:36 ` Ingo Molnar 2018-05-05 8:54 ` [PATCH] locking/atomics: Combine the atomic_andnot() and atomic64_andnot() API definitions Ingo Molnar 2018-05-05 8:54 ` Ingo Molnar 2018-05-06 12:15 ` [tip:locking/core] " tip-bot for Ingo Molnar 2018-05-06 14:15 ` [PATCH] " Andrea Parri 2018-05-06 14:15 ` Andrea Parri 2018-05-06 12:14 ` [tip:locking/core] locking/atomics: Simplify the op definitions in atomic.h some more tip-bot for Ingo Molnar 2018-05-09 7:33 ` Peter Zijlstra 2018-05-09 13:03 ` Will Deacon 2018-05-15 8:54 ` Ingo Molnar 2018-05-15 8:35 ` Ingo Molnar 2018-05-15 11:41 ` Peter Zijlstra 2018-05-15 12:13 ` Peter Zijlstra 2018-05-15 15:43 ` Mark Rutland 2018-05-15 17:10 ` Peter Zijlstra 2018-05-15 17:53 ` Mark Rutland 2018-05-15 18:11 ` Peter Zijlstra 2018-05-15 18:15 ` Peter Zijlstra 2018-05-15 18:52 ` Linus Torvalds 2018-05-15 19:39 ` Peter Zijlstra 2018-05-21 17:12 ` Mark Rutland 2018-05-06 14:12 ` [PATCH] " Andrea Parri 2018-05-06 14:12 ` Andrea Parri 2018-05-06 14:57 ` Ingo Molnar 2018-05-06 14:57 ` Ingo Molnar 2018-05-07 9:54 ` Andrea Parri 2018-05-07 9:54 ` Andrea Parri 2018-05-18 18:43 ` Palmer Dabbelt 2018-05-18 18:43 ` Palmer Dabbelt 2018-05-05 8:47 ` [PATCH] locking/atomics: Clean up the atomic.h maze of #defines Peter Zijlstra 2018-05-05 8:47 ` Peter Zijlstra 2018-05-05 9:04 ` Ingo Molnar 2018-05-05 9:04 ` Ingo Molnar 2018-05-05 9:24 ` Peter Zijlstra [this message] 2018-05-05 9:24 ` Peter Zijlstra 2018-05-05 9:38 ` Ingo Molnar 2018-05-05 9:38 ` Ingo Molnar 2018-05-05 10:00 ` [RFC PATCH] locking/atomics/powerpc: Introduce optimized cmpxchg_release() family of APIs for PowerPC Ingo Molnar 2018-05-05 10:00 ` Ingo Molnar 2018-05-05 10:26 ` Boqun Feng 2018-05-05 10:26 ` Boqun Feng 2018-05-06 1:56 ` Benjamin Herrenschmidt 2018-05-06 1:56 ` Benjamin Herrenschmidt 2018-05-05 10:16 ` [PATCH] locking/atomics: Clean up the atomic.h maze of #defines Boqun Feng 2018-05-05 10:16 ` Boqun Feng 2018-05-05 10:35 ` [RFC PATCH] locking/atomics/powerpc: Clarify why the cmpxchg_relaxed() family of APIs falls back to full cmpxchg() Ingo Molnar 2018-05-05 10:35 ` Ingo Molnar 2018-05-05 11:28 ` Boqun Feng 2018-05-05 11:28 ` Boqun Feng 2018-05-05 13:27 ` [PATCH] locking/atomics/powerpc: Move cmpxchg helpers to asm/cmpxchg.h and define the full set of cmpxchg APIs Ingo Molnar 2018-05-05 13:27 ` Ingo Molnar 2018-05-05 14:03 ` Boqun Feng 2018-05-05 14:03 ` Boqun Feng 2018-05-06 12:11 ` Ingo Molnar 2018-05-06 12:11 ` Ingo Molnar 2018-05-07 1:04 ` Boqun Feng 2018-05-07 1:04 ` Boqun Feng 2018-05-07 6:50 ` Ingo Molnar 2018-05-07 6:50 ` Ingo Molnar 2018-05-06 12:13 ` [tip:locking/core] " tip-bot for Boqun Feng 2018-05-07 13:31 ` [PATCH v2] " Boqun Feng 2018-05-07 13:31 ` Boqun Feng 2018-05-05 9:05 ` [PATCH] locking/atomics: Clean up the atomic.h maze of #defines Dmitry Vyukov 2018-05-05 9:05 ` Dmitry Vyukov 2018-05-05 9:32 ` Peter Zijlstra 2018-05-05 9:32 ` Peter Zijlstra 2018-05-07 6:43 ` [RFC PATCH] locking/atomics/x86/64: Clean up and fix details of <asm/atomic64_64.h> Ingo Molnar 2018-05-07 6:43 ` Ingo Molnar 2018-05-05 9:09 ` [PATCH] locking/atomics: Clean up the atomic.h maze of #defines Ingo Molnar 2018-05-05 9:09 ` Ingo Molnar 2018-05-05 9:29 ` Peter Zijlstra 2018-05-05 9:29 ` Peter Zijlstra 2018-05-05 10:48 ` [PATCH] locking/atomics: Shorten the __atomic_op() defines to __op() Ingo Molnar 2018-05-05 10:48 ` Ingo Molnar 2018-05-05 10:59 ` Ingo Molnar 2018-05-05 10:59 ` Ingo Molnar 2018-05-06 12:15 ` [tip:locking/core] " tip-bot for Ingo Molnar 2018-05-06 12:14 ` [tip:locking/core] locking/atomics: Clean up the atomic.h maze of #defines tip-bot for Ingo Molnar 2018-05-04 17:39 ` [PATCH 2/6] locking/atomic, asm-generic: instrument atomic*andnot*() Mark Rutland 2018-05-04 17:39 ` Mark Rutland 2018-05-04 17:39 ` [PATCH 3/6] arm64: use <linux/atomic.h> for cmpxchg Mark Rutland 2018-05-04 17:39 ` Mark Rutland 2018-05-04 17:39 ` [PATCH 4/6] arm64: fix assembly constraints " Mark Rutland 2018-05-04 17:39 ` Mark Rutland 2018-05-04 17:39 ` [PATCH 5/6] arm64: use instrumented atomics Mark Rutland 2018-05-04 17:39 ` Mark Rutland 2018-05-04 17:39 ` [PATCH 6/6] arm64: instrument smp_{load_acquire,store_release} Mark Rutland 2018-05-04 17:39 ` Mark Rutland
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=20180505092438.GB12217@hirez.programming.kicks-ass.net \ --to=peterz@infradead.org \ --cc=aryabinin@virtuozzo.com \ --cc=boqun.feng@gmail.com \ --cc=catalin.marinas@arm.com \ --cc=dvyukov@google.com \ --cc=linux-arm-kernel@lists.infradead.org \ --cc=linux-kernel@vger.kernel.org \ --cc=mark.rutland@arm.com \ --cc=mingo@kernel.org \ --cc=will.deacon@arm.com \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.