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 kanga.kvack.org (kanga.kvack.org [205.233.56.17]) by smtp.lore.kernel.org (Postfix) with ESMTP id 3451AC4345F for ; Thu, 18 Apr 2024 19:40:06 +0000 (UTC) Received: by kanga.kvack.org (Postfix) id A96F66B0092; Thu, 18 Apr 2024 15:40:05 -0400 (EDT) Received: by kanga.kvack.org (Postfix, from userid 40) id A472A6B0093; Thu, 18 Apr 2024 15:40:05 -0400 (EDT) X-Delivered-To: int-list-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix, from userid 63042) id 90F346B0096; Thu, 18 Apr 2024 15:40:05 -0400 (EDT) X-Delivered-To: linux-mm@kvack.org Received: from relay.hostedemail.com (smtprelay0016.hostedemail.com [216.40.44.16]) by kanga.kvack.org (Postfix) with ESMTP id 73B9C6B0092 for ; Thu, 18 Apr 2024 15:40:05 -0400 (EDT) Received: from smtpin08.hostedemail.com (a10.router.float.18 [10.200.18.1]) by unirelay10.hostedemail.com (Postfix) with ESMTP id E5A27C114B for ; Thu, 18 Apr 2024 19:40:04 +0000 (UTC) X-FDA: 82023668328.08.E0AAEEF Received: from dfw.source.kernel.org (dfw.source.kernel.org [139.178.84.217]) by imf26.hostedemail.com (Postfix) with ESMTP id 136BF140004 for ; Thu, 18 Apr 2024 19:40:02 +0000 (UTC) Authentication-Results: imf26.hostedemail.com; dkim=pass header.d=linux-foundation.org header.s=korg header.b=fK4+rhQG; dmarc=none; spf=pass (imf26.hostedemail.com: domain of akpm@linux-foundation.org designates 139.178.84.217 as permitted sender) smtp.mailfrom=akpm@linux-foundation.org ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=hostedemail.com; s=arc-20220608; t=1713469203; h=from:from:sender:reply-to:subject:subject:date:date: message-id:message-id:to:to:cc:cc:mime-version:mime-version: content-type:content-type: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references:dkim-signature; bh=/09oZhHAHOvIqFxSujKGO2dSewus+quoY8lu1inVrYY=; b=H27x02Tl1U4HbFg7MagblC/zLygavH7ZBmbz897jGWZ+Uyiwvj5G3SHaCsppbHUn+53eon +xF0DTaGx+tlNAYXYyVQVDGY7grr7kb/GPtE6jKLS9abfwTOUKlpvRtlQR/ZL+nt8iNIxP 17bRkXNl2cDt7T3rUjyje5Cf0lC5fUE= ARC-Authentication-Results: i=1; imf26.hostedemail.com; dkim=pass header.d=linux-foundation.org header.s=korg header.b=fK4+rhQG; dmarc=none; spf=pass (imf26.hostedemail.com: domain of akpm@linux-foundation.org designates 139.178.84.217 as permitted sender) smtp.mailfrom=akpm@linux-foundation.org ARC-Seal: i=1; s=arc-20220608; d=hostedemail.com; t=1713469203; a=rsa-sha256; cv=none; b=ufLOt4T5UhNDmzG3CztmcJIBozNl3Bb0c4jUZuXG9RNrSX4AvgEZO3PYqG2q/oi/FV3r8I qgiQL1wUXorn1RFHvsii72zAwWm1gZK+iEu65b5K9aLRtT72Dn6YivDDKUymT+fnAT4k9B ECYlfhKHUUwU7CPjoa5xN8xwn8/NXYA= Received: from smtp.kernel.org (transwarp.subspace.kernel.org [100.75.92.58]) by dfw.source.kernel.org (Postfix) with ESMTP id E61C661808; Thu, 18 Apr 2024 19:40:01 +0000 (UTC) Received: by smtp.kernel.org (Postfix) with ESMTPSA id 15D52C113CC; Thu, 18 Apr 2024 19:40:01 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=linux-foundation.org; s=korg; t=1713469201; bh=c+2Am6FmwvmgOAIMm5qyWEX2cWFI5nJZcTTExzVFlr4=; h=Date:From:To:Cc:Subject:In-Reply-To:References:From; b=fK4+rhQGDmnzaFkK6/D0dn6YdSRYFho6z8svfL2L9PlTAebnkFW27qjR3LaFGDOcd xSS3Bymo7Z4k6IWpZHM+Lkh1fNZP27YvcVyRiAXpOnRKmWDVE8bk5Rb4iRrQzO9ah4 JvpvjVmqvf54AgJJyMRvxmJVSvBHG2SIaWl0zbj8= Date: Thu, 18 Apr 2024 12:40:00 -0700 From: Andrew Morton To: Peng Zhang Cc: , , , , , , , , , , , , Subject: Re: [RFC PATCH v2 1/2] percpu_counter: introduce atomic mode for percpu_counter Message-Id: <20240418124000.ce4606dad982d7f31fc0d117@linux-foundation.org> In-Reply-To: <20240418142008.2775308-2-zhangpeng362@huawei.com> References: <20240418142008.2775308-1-zhangpeng362@huawei.com> <20240418142008.2775308-2-zhangpeng362@huawei.com> X-Mailer: Sylpheed 3.8.0beta1 (GTK+ 2.24.33; x86_64-pc-linux-gnu) Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit X-Rspam-User: X-Rspamd-Server: rspam08 X-Rspamd-Queue-Id: 136BF140004 X-Stat-Signature: epr3r9zqehb7k1wipgz4fm38bqij559i X-HE-Tag: 1713469202-394767 X-HE-Meta: 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 JNl3CB0a 3kb902RRqjDS7kvgOol8F3W+LSL/cmJFGoFePI4mLasOn6WOLtgs83yFrW5sJSYhSJyqma0DrPCMH+79y87ASNMTJooXkldMge65QA03aox65ci5yEAd/kInyhuWJZydU4oqZyUBNi5IOO1/Ca67veJEDLRspXy1cy5NwJhelhEy6xligmP8QzE2v1KCkp+VzGk98l+cHk2cee97Qmt3aFDNfk1Aszvc9c69SH0qDWVrvO4lQlLsJLLX0tT5tOei+8lVihurD7RAzUN1DW4OqGCqahq1YsCH4NsiFw2v7Xv5E17w= X-Bogosity: Ham, tests=bogofilter, spamicity=0.000000, version=1.2.4 Sender: owner-linux-mm@kvack.org Precedence: bulk X-Loop: owner-majordomo@kvack.org List-ID: List-Subscribe: List-Unsubscribe: On Thu, 18 Apr 2024 22:20:07 +0800 Peng Zhang wrote: > From: ZhangPeng > > Depending on whether counters is NULL, we can support two modes: > atomic mode and perpcu mode. We implement both modes by grouping > the s64 count and atomic64_t count_atomic in a union. At the same time, > we create the interface for adding and reading in atomic mode and for > switching atomic mode to percpu mode. > I think it would be better if we had a detailed code comment in an appropriate place which fully describes the tradeoffs here. Tell people when they would benefit from using one mode versus the other. > --- a/lib/percpu_counter.c > +++ b/lib/percpu_counter.c > @@ -153,7 +153,7 @@ EXPORT_SYMBOL(__percpu_counter_sum); > > int __percpu_counter_init_many(struct percpu_counter *fbc, s64 amount, > gfp_t gfp, u32 nr_counters, > - struct lock_class_key *key) > + struct lock_class_key *key, bool switch_mode) > { > unsigned long flags __maybe_unused; > size_t counter_size; > @@ -174,7 +174,8 @@ int __percpu_counter_init_many(struct percpu_counter *fbc, s64 amount, > #ifdef CONFIG_HOTPLUG_CPU > INIT_LIST_HEAD(&fbc[i].list); > #endif > - fbc[i].count = amount; > + if (likely(!switch_mode)) > + fbc[i].count = amount; > fbc[i].counters = (void *)counters + (i * counter_size); > > debug_percpu_counter_activate(&fbc[i]); > @@ -357,6 +358,32 @@ bool __percpu_counter_limited_add(struct percpu_counter *fbc, > return good; > } > > +/* > + * percpu_counter_switch_to_pcpu_many: Converts struct percpu_counters from > + * atomic mode to percpu mode. Describe what happens if that GFP_ATOMIC allocation fails. We remain in atomic mode, yes? > + */ > +int percpu_counter_switch_to_pcpu_many(struct percpu_counter *fbc, > + u32 nr_counters) > +{ > + static struct lock_class_key __key; > + unsigned long flags; > + bool ret = 0; > + > + if (percpu_counter_initialized(fbc)) > + return 0; > + > + preempt_disable(); > + local_irq_save(flags); Do we need both? Does local_irq_save() always disable preemption? This might not be the case for RT kernels, I always forget. > + if (likely(!percpu_counter_initialized(fbc))) > + ret = __percpu_counter_init_many(fbc, 0, > + GFP_ATOMIC|__GFP_NOWARN|__GFP_ZERO, > + nr_counters, &__key, true); > + local_irq_restore(flags); > + preempt_enable(); > + > + return ret; > +} > + Why is there no API for switching back to atomic mode?