From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 360D0CA0FE1 for ; Fri, 1 Sep 2023 06:18:51 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S237610AbjIAGSw convert rfc822-to-8bit (ORCPT ); Fri, 1 Sep 2023 02:18:52 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:43934 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S230011AbjIAGSw (ORCPT ); Fri, 1 Sep 2023 02:18:52 -0400 Received: from us-smtp-delivery-124.mimecast.com (us-smtp-delivery-124.mimecast.com [170.10.129.124]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id C62A391 for ; Thu, 31 Aug 2023 23:18:04 -0700 (PDT) Received: from mail-qt1-f198.google.com (mail-qt1-f198.google.com [209.85.160.198]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.3, cipher=TLS_AES_256_GCM_SHA384) id us-mta-637-tEbdxz8NNyujic8afw2ZIg-1; Fri, 01 Sep 2023 02:18:02 -0400 X-MC-Unique: tEbdxz8NNyujic8afw2ZIg-1 Received: by mail-qt1-f198.google.com with SMTP id d75a77b69052e-4136ec108ecso17287651cf.2 for ; Thu, 31 Aug 2023 23:18:02 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20221208; t=1693549082; x=1694153882; h=content-transfer-encoding:cc:to:subject:message-id:date:from :in-reply-to:references:mime-version:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=NCoHUy21Y+HIbtl00/+o/mR3wL3CsEO4zSXXLcm78Tg=; b=bKEES/GUuNhzvOHqoywF7zJVvsueQ6ob417TprA9Eb93jZYHZnd7mP79epS+SKvJib Nz/DtZE62LUGd0MXByg5u+3bPqxSAEv3OyjeNbwRuHsJ37i9KRw181BErbz1Eeds8m/8 G3es3IA1EZN89gmq9Mxsl4QWq7En6jIoGyOhGBB/RjWH6AKjeSCT9o+NGzkpRT9ZtcNg UIL3hVBUrxnDeVVN7Td2aacYvIMLDdg/QyAljv3nbbmzDkowkUaaxgv03whgywLnbubq KXRtRYHk407gZYCxV9WHAio3Hp3/LTfLsxycPdtpoomRJmtIcfzYpCxivrZ0bTqsNBuX stEw== X-Gm-Message-State: AOJu0Ywktro2J7HDu+nZAGtvVu2awg3CQcbRWDpH/TrtALTv6Rk/m2OW Qwo8WenwFX1JEEqPYzeyywlQjt3M1JLCAsVsenQHoGZcUAruq8cbHcRGqVXKmIu5WrRf19dXGvN pRFLnp8mP6V0AGAIO2LUBuT6FcrYQTzD8l6/7etTmD40tGTaD X-Received: by 2002:ac8:7d56:0:b0:412:1e4c:e858 with SMTP id h22-20020ac87d56000000b004121e4ce858mr1914052qtb.36.1693549082435; Thu, 31 Aug 2023 23:18:02 -0700 (PDT) X-Google-Smtp-Source: AGHT+IGF8Qj0L34On7VCvINdaprG6njP7EEdfXIQ5lwObVY15Tfc/HZL0XMnvJG3rnOKeETsZkTiyUuElX2TEmOia18= X-Received: by 2002:ac8:7d56:0:b0:412:1e4c:e858 with SMTP id h22-20020ac87d56000000b004121e4ce858mr1914039qtb.36.1693549082189; Thu, 31 Aug 2023 23:18:02 -0700 (PDT) MIME-Version: 1.0 References: <20230831063129.335425-1-leobras@redhat.com> In-Reply-To: From: Leonardo Bras Soares Passos Date: Fri, 1 Sep 2023 03:17:50 -0300 Message-ID: Subject: Re: [RFC PATCH 1/1] smp: Change function signatures to use call_single_data_t To: Guo Ren Cc: Steven Rostedt , Masami Hiramatsu , Peter Zijlstra , Josh Poimboeuf , Valentin Schneider , "Paul E. McKenney" , Juergen Gross , Yury Norov , Imran Khan , linux-kernel@vger.kernel.org, linux-trace-kernel@vger.kernel.org X-Mimecast-Spam-Score: 0 X-Mimecast-Originator: redhat.com Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 8BIT Precedence: bulk List-ID: X-Mailing-List: linux-trace-kernel@vger.kernel.org On Thu, Aug 31, 2023 at 3:58 AM Guo Ren wrote: > > On Thu, Aug 31, 2023 at 2:31 PM Leonardo Bras wrote: > > > > call_single_data_t is a size-aligned typedef of struct __call_single_data. > > > > This alignment is desirable in order to have smp_call_function*() avoid > > bouncing an extra cacheline in case of an unaligned csd, given this > > would hurt performance. > > > > Since the removal of struct request->csd in commit 660e802c76c8 > > ("blk-mq: use percpu csd to remote complete instead of per-rq csd") there > > are no current users of smp_call_function*() with unaligned csd. > > > > Change every 'struct __call_single_data' function parameter to > > 'call_single_data_t', so we have warnings if any new code tries to > > introduce an smp_call_function*() call with unaligned csd. > I agree to prevent __call_single_data usage. > > Reviewed-by: Guo Ren Thanks for reviewing! > > /* > * structure shares (partial) layout with struct irq_work > */ > struct __call_single_data { > struct __call_single_node node; > smp_call_func_t func; > void *info; > }; > > #define CSD_INIT(_func, _info) \ > (struct __call_single_data){ .func = (_func), .info = (_info), } > > /* Use __aligned() to avoid to use 2 cache lines for 1 csd */ > typedef struct __call_single_data call_single_data_t > __aligned(sizeof(struct __call_single_data)); > > > > > Signed-off-by: Leonardo Bras > > --- > > include/linux/smp.h | 2 +- > > include/trace/events/csd.h | 8 ++++---- > > kernel/smp.c | 26 +++++++++++++------------- > > kernel/up.c | 2 +- > > 4 files changed, 19 insertions(+), 19 deletions(-) > > > > diff --git a/include/linux/smp.h b/include/linux/smp.h > > index 91ea4a67f8ca..e87520dc2959 100644 > > --- a/include/linux/smp.h > > +++ b/include/linux/smp.h > > @@ -53,7 +53,7 @@ int smp_call_function_single(int cpuid, smp_call_func_t func, void *info, > > void on_each_cpu_cond_mask(smp_cond_func_t cond_func, smp_call_func_t func, > > void *info, bool wait, const struct cpumask *mask); > > > > -int smp_call_function_single_async(int cpu, struct __call_single_data *csd); > > +int smp_call_function_single_async(int cpu, call_single_data_t *csd); > > > > /* > > * Cpus stopping functions in panic. All have default weak definitions. > > diff --git a/include/trace/events/csd.h b/include/trace/events/csd.h > > index 67e9d01f80c2..58cc83b99c34 100644 > > --- a/include/trace/events/csd.h > > +++ b/include/trace/events/csd.h > > @@ -12,7 +12,7 @@ TRACE_EVENT(csd_queue_cpu, > > TP_PROTO(const unsigned int cpu, > > unsigned long callsite, > > smp_call_func_t func, > > - struct __call_single_data *csd), > > + call_single_data_t *csd), > > > > TP_ARGS(cpu, callsite, func, csd), > > > > @@ -39,7 +39,7 @@ TRACE_EVENT(csd_queue_cpu, > > */ > > DECLARE_EVENT_CLASS(csd_function, > > > > - TP_PROTO(smp_call_func_t func, struct __call_single_data *csd), > > + TP_PROTO(smp_call_func_t func, call_single_data_t *csd), > > > > TP_ARGS(func, csd), > > > > @@ -57,12 +57,12 @@ DECLARE_EVENT_CLASS(csd_function, > > ); > > > > DEFINE_EVENT(csd_function, csd_function_entry, > > - TP_PROTO(smp_call_func_t func, struct __call_single_data *csd), > > + TP_PROTO(smp_call_func_t func, call_single_data_t *csd), > > TP_ARGS(func, csd) > > ); > > > > DEFINE_EVENT(csd_function, csd_function_exit, > > - TP_PROTO(smp_call_func_t func, struct __call_single_data *csd), > > + TP_PROTO(smp_call_func_t func, call_single_data_t *csd), > > TP_ARGS(func, csd) > > ); > > > > diff --git a/kernel/smp.c b/kernel/smp.c > > index 8455a53465af..8c714583786b 100644 > > --- a/kernel/smp.c > > +++ b/kernel/smp.c > > @@ -127,7 +127,7 @@ send_call_function_ipi_mask(struct cpumask *mask) > > } > > > > static __always_inline void > > -csd_do_func(smp_call_func_t func, void *info, struct __call_single_data *csd) > > +csd_do_func(smp_call_func_t func, void *info, call_single_data_t *csd) > > { > > trace_csd_function_entry(func, csd); > > func(info); > > @@ -174,7 +174,7 @@ module_param(csd_lock_timeout, ulong, 0444); > > static atomic_t csd_bug_count = ATOMIC_INIT(0); > > > > /* Record current CSD work for current CPU, NULL to erase. */ > > -static void __csd_lock_record(struct __call_single_data *csd) > > +static void __csd_lock_record(call_single_data_t *csd) > > { > > if (!csd) { > > smp_mb(); /* NULL cur_csd after unlock. */ > > @@ -189,13 +189,13 @@ static void __csd_lock_record(struct __call_single_data *csd) > > /* Or before unlock, as the case may be. */ > > } > > > > -static __always_inline void csd_lock_record(struct __call_single_data *csd) > > +static __always_inline void csd_lock_record(call_single_data_t *csd) > > { > > if (static_branch_unlikely(&csdlock_debug_enabled)) > > __csd_lock_record(csd); > > } > > > > -static int csd_lock_wait_getcpu(struct __call_single_data *csd) > > +static int csd_lock_wait_getcpu(call_single_data_t *csd) > > { > > unsigned int csd_type; > > > > @@ -210,7 +210,7 @@ static int csd_lock_wait_getcpu(struct __call_single_data *csd) > > * the CSD_TYPE_SYNC/ASYNC types provide the destination CPU, > > * so waiting on other types gets much less information. > > */ > > -static bool csd_lock_wait_toolong(struct __call_single_data *csd, u64 ts0, u64 *ts1, int *bug_id) > > +static bool csd_lock_wait_toolong(call_single_data_t *csd, u64 ts0, u64 *ts1, int *bug_id) > > { > > int cpu = -1; > > int cpux; > > @@ -276,7 +276,7 @@ static bool csd_lock_wait_toolong(struct __call_single_data *csd, u64 ts0, u64 * > > * previous function call. For multi-cpu calls its even more interesting > > * as we'll have to ensure no other cpu is observing our csd. > > */ > > -static void __csd_lock_wait(struct __call_single_data *csd) > > +static void __csd_lock_wait(call_single_data_t *csd) > > { > > int bug_id = 0; > > u64 ts0, ts1; > > @@ -290,7 +290,7 @@ static void __csd_lock_wait(struct __call_single_data *csd) > > smp_acquire__after_ctrl_dep(); > > } > > > > -static __always_inline void csd_lock_wait(struct __call_single_data *csd) > > +static __always_inline void csd_lock_wait(call_single_data_t *csd) > > { > > if (static_branch_unlikely(&csdlock_debug_enabled)) { > > __csd_lock_wait(csd); > > @@ -300,17 +300,17 @@ static __always_inline void csd_lock_wait(struct __call_single_data *csd) > > smp_cond_load_acquire(&csd->node.u_flags, !(VAL & CSD_FLAG_LOCK)); > > } > > #else > > -static void csd_lock_record(struct __call_single_data *csd) > > +static void csd_lock_record(call_single_data_t *csd) > > { > > } > > > > -static __always_inline void csd_lock_wait(struct __call_single_data *csd) > > +static __always_inline void csd_lock_wait(call_single_data_t *csd) > > { > > smp_cond_load_acquire(&csd->node.u_flags, !(VAL & CSD_FLAG_LOCK)); > > } > > #endif > > > > -static __always_inline void csd_lock(struct __call_single_data *csd) > > +static __always_inline void csd_lock(call_single_data_t *csd) > > { > > csd_lock_wait(csd); > > csd->node.u_flags |= CSD_FLAG_LOCK; > > @@ -323,7 +323,7 @@ static __always_inline void csd_lock(struct __call_single_data *csd) > > smp_wmb(); > > } > > > > -static __always_inline void csd_unlock(struct __call_single_data *csd) > > +static __always_inline void csd_unlock(call_single_data_t *csd) > > { > > WARN_ON(!(csd->node.u_flags & CSD_FLAG_LOCK)); > > > > @@ -376,7 +376,7 @@ void __smp_call_single_queue(int cpu, struct llist_node *node) > > * for execution on the given CPU. data must already have > > * ->func, ->info, and ->flags set. > > */ > > -static int generic_exec_single(int cpu, struct __call_single_data *csd) > > +static int generic_exec_single(int cpu, call_single_data_t *csd) > > { > > if (cpu == smp_processor_id()) { > > smp_call_func_t func = csd->func; > > @@ -667,7 +667,7 @@ EXPORT_SYMBOL(smp_call_function_single); > > * > > * Return: %0 on success or negative errno value on error > > */ > > -int smp_call_function_single_async(int cpu, struct __call_single_data *csd) > > +int smp_call_function_single_async(int cpu, call_single_data_t *csd) > > { > > int err = 0; > > > > diff --git a/kernel/up.c b/kernel/up.c > > index a38b8b095251..df50828cc2f0 100644 > > --- a/kernel/up.c > > +++ b/kernel/up.c > > @@ -25,7 +25,7 @@ int smp_call_function_single(int cpu, void (*func) (void *info), void *info, > > } > > EXPORT_SYMBOL(smp_call_function_single); > > > > -int smp_call_function_single_async(int cpu, struct __call_single_data *csd) > > +int smp_call_function_single_async(int cpu, call_single_data_t *csd) > > { > > unsigned long flags; > > > > -- > > 2.42.0 > > > > > -- > Best Regards > Guo Ren >