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 D0B7DE77182 for ; Thu, 12 Dec 2024 09:15:38 +0000 (UTC) Received: by kanga.kvack.org (Postfix) id 67C496B0083; Thu, 12 Dec 2024 04:15:38 -0500 (EST) Received: by kanga.kvack.org (Postfix, from userid 40) id 62A6B6B0085; Thu, 12 Dec 2024 04:15:38 -0500 (EST) X-Delivered-To: int-list-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix, from userid 63042) id 4A32E6B0089; Thu, 12 Dec 2024 04:15:38 -0500 (EST) X-Delivered-To: linux-mm@kvack.org Received: from relay.hostedemail.com (smtprelay0011.hostedemail.com [216.40.44.11]) by kanga.kvack.org (Postfix) with ESMTP id 2CB246B0083 for ; Thu, 12 Dec 2024 04:15:38 -0500 (EST) Received: from smtpin04.hostedemail.com (a10.router.float.18 [10.200.18.1]) by unirelay01.hostedemail.com (Postfix) with ESMTP id D48D21C891B for ; Thu, 12 Dec 2024 09:15:37 +0000 (UTC) X-FDA: 82885748652.04.E56B098 Received: from smtp-out1.suse.de (smtp-out1.suse.de [195.135.223.130]) by imf07.hostedemail.com (Postfix) with ESMTP id 71EB040010 for ; Thu, 12 Dec 2024 09:15:07 +0000 (UTC) Authentication-Results: imf07.hostedemail.com; dkim=pass header.d=suse.cz header.s=susede2_rsa header.b="SP/h1UKN"; dkim=pass header.d=suse.cz header.s=susede2_ed25519 header.b=oteZM1W7; dkim=pass header.d=suse.cz header.s=susede2_rsa header.b=ldSXcSXS; dkim=pass header.d=suse.cz header.s=susede2_ed25519 header.b="4rW/o5t9"; spf=pass (imf07.hostedemail.com: domain of vbabka@suse.cz designates 195.135.223.130 as permitted sender) smtp.mailfrom=vbabka@suse.cz; dmarc=none ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=hostedemail.com; s=arc-20220608; t=1733994917; 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=R7fVrjcp/UTDDfv/utL2sG0J56iJc0yfUCHiiOAS7hk=; b=GdUqGTLyylknwdSUmPteCSkGNVPwXa+4MPrl2TguidPwxiYndUrNN5x6cRFr/yC1+lYdSp s0J5S/mbXObLsrvLxI2ib907k9xghHUt4DIzgfVMYIKeOKFYAFMLAkdbYlXBRpTIbYlLA3 lUiWDI3EtKiTEKs0yFCJXZE5r6mcfgo= ARC-Authentication-Results: i=1; imf07.hostedemail.com; dkim=pass header.d=suse.cz header.s=susede2_rsa header.b="SP/h1UKN"; dkim=pass header.d=suse.cz header.s=susede2_ed25519 header.b=oteZM1W7; dkim=pass header.d=suse.cz header.s=susede2_rsa header.b=ldSXcSXS; dkim=pass header.d=suse.cz header.s=susede2_ed25519 header.b="4rW/o5t9"; spf=pass (imf07.hostedemail.com: domain of vbabka@suse.cz designates 195.135.223.130 as permitted sender) smtp.mailfrom=vbabka@suse.cz; dmarc=none ARC-Seal: i=1; s=arc-20220608; d=hostedemail.com; t=1733994917; a=rsa-sha256; cv=none; b=bzo5GMJJ5luFZxihCmUHMNN20nLdpv9D8SlYutxsvzCRq+oPNalbgbXDttT+9p7CJwlnqR +d+0Twi/SX3dPOmZ032UVbkWhSmD1E0VudQurGaOv+G3uuWCcKlIR/RO5gDlbnVqiAVuwN WyD1pBIAevRCany4SxWK/9b454RFqk8= Received: from imap1.dmz-prg2.suse.org (imap1.dmz-prg2.suse.org [IPv6:2a07:de40:b281:104:10:150:64:97]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (4096 bits) server-digest SHA256) (No client certificate requested) by smtp-out1.suse.de (Postfix) with ESMTPS id AC7F021119; Thu, 12 Dec 2024 09:15:32 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=suse.cz; s=susede2_rsa; t=1733994933; h=from:from:reply-to: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:autocrypt:autocrypt; bh=R7fVrjcp/UTDDfv/utL2sG0J56iJc0yfUCHiiOAS7hk=; b=SP/h1UKNFO2G3e4CYg3KAYpe5TJtbOiW0vw3ugVHdCo6fYSdVTJ4jp7maHKASeR8LJ3onH rI6T++ih/XBiZGO+6kFnv/7qHvMD4n7DgskCBBKQP+bxpnin8XTlS38usphefLPaiB7IBm kigJQKVojktiWAthi6E3SDths4zqXTw= DKIM-Signature: v=1; a=ed25519-sha256; c=relaxed/relaxed; d=suse.cz; s=susede2_ed25519; t=1733994933; h=from:from:reply-to: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:autocrypt:autocrypt; bh=R7fVrjcp/UTDDfv/utL2sG0J56iJc0yfUCHiiOAS7hk=; b=oteZM1W7LjoMROSvuKU6eFxZuOuKt317SaphnOOcifWXP3wJjZ40FCJUoyd/1JX+nQvZ0M 741tXn0k+fnM7vDA== DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=suse.cz; s=susede2_rsa; t=1733994932; h=from:from:reply-to: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:autocrypt:autocrypt; bh=R7fVrjcp/UTDDfv/utL2sG0J56iJc0yfUCHiiOAS7hk=; b=ldSXcSXSCIA7tIh9TDFBv6CTB7UgTuIq9X1tD2XR8I/wmQTNYH40KKuEl8+yiXjxO1U56E lAlTMPCOqju0od3oiNUwhAgeXIPIO0dIsd4Cp1gKkDrZbBzAepC1XL/qWx5Mj9uATYFggH qkOUWWw0BMaZMJtwdB6NOVEfVch22tA= DKIM-Signature: v=1; a=ed25519-sha256; c=relaxed/relaxed; d=suse.cz; s=susede2_ed25519; t=1733994932; h=from:from:reply-to: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:autocrypt:autocrypt; bh=R7fVrjcp/UTDDfv/utL2sG0J56iJc0yfUCHiiOAS7hk=; b=4rW/o5t9tXlxIDQ7tYXA3j8TEkC7zk8hwOsaWxf1r675RmumoDhpBsuYLNPI7nUxt+rXuG xn47DLy4nbw/K5Bw== Received: from imap1.dmz-prg2.suse.org (localhost [127.0.0.1]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (4096 bits) server-digest SHA256) (No client certificate requested) by imap1.dmz-prg2.suse.org (Postfix) with ESMTPS id 86DF013508; Thu, 12 Dec 2024 09:15:32 +0000 (UTC) Received: from dovecot-director2.suse.de ([2a07:de40:b281:106:10:150:64:167]) by imap1.dmz-prg2.suse.org with ESMTPSA id 7yihILSpWmfQCwAAD6G6ig (envelope-from ); Thu, 12 Dec 2024 09:15:32 +0000 Message-ID: <60fadf5a-4c96-43b3-8cc2-baf71eb93e3f@suse.cz> Date: Thu, 12 Dec 2024 10:15:32 +0100 MIME-Version: 1.0 User-Agent: Mozilla Thunderbird Subject: Re: [PATCH bpf-next v2 3/6] locking/local_lock: Introduce local_trylock_irqsave() Content-Language: en-US To: Alexei Starovoitov Cc: bpf , Andrii Nakryiko , Kumar Kartikeya Dwivedi , Andrew Morton , Peter Zijlstra , Sebastian Sewior , Steven Rostedt , Hou Tao , Johannes Weiner , shakeel.butt@linux.dev, Michal Hocko , Matthew Wilcox , Thomas Gleixner , Tejun Heo , linux-mm , Kernel Team , Jann Horn References: <20241210023936.46871-1-alexei.starovoitov@gmail.com> <20241210023936.46871-4-alexei.starovoitov@gmail.com> <1c760bf1-14a4-42e4-a55b-438a29987aef@suse.cz> <9e5bdef1-a692-47d5-82b9-96a4f2c68463@suse.cz> From: Vlastimil Babka Autocrypt: addr=vbabka@suse.cz; keydata= xsFNBFZdmxYBEADsw/SiUSjB0dM+vSh95UkgcHjzEVBlby/Fg+g42O7LAEkCYXi/vvq31JTB KxRWDHX0R2tgpFDXHnzZcQywawu8eSq0LxzxFNYMvtB7sV1pxYwej2qx9B75qW2plBs+7+YB 87tMFA+u+L4Z5xAzIimfLD5EKC56kJ1CsXlM8S/LHcmdD9Ctkn3trYDNnat0eoAcfPIP2OZ+ 9oe9IF/R28zmh0ifLXyJQQz5ofdj4bPf8ecEW0rhcqHfTD8k4yK0xxt3xW+6Exqp9n9bydiy tcSAw/TahjW6yrA+6JhSBv1v2tIm+itQc073zjSX8OFL51qQVzRFr7H2UQG33lw2QrvHRXqD Ot7ViKam7v0Ho9wEWiQOOZlHItOOXFphWb2yq3nzrKe45oWoSgkxKb97MVsQ+q2SYjJRBBH4 8qKhphADYxkIP6yut/eaj9ImvRUZZRi0DTc8xfnvHGTjKbJzC2xpFcY0DQbZzuwsIZ8OPJCc LM4S7mT25NE5kUTG/TKQCk922vRdGVMoLA7dIQrgXnRXtyT61sg8PG4wcfOnuWf8577aXP1x 6mzw3/jh3F+oSBHb/GcLC7mvWreJifUL2gEdssGfXhGWBo6zLS3qhgtwjay0Jl+kza1lo+Cv BB2T79D4WGdDuVa4eOrQ02TxqGN7G0Biz5ZLRSFzQSQwLn8fbwARAQABzSBWbGFzdGltaWwg QmFia2EgPHZiYWJrYUBzdXNlLmN6PsLBlAQTAQoAPgIbAwULCQgHAwUVCgkICwUWAgMBAAIe AQIXgBYhBKlA1DSZLC6OmRA9UCJPp+fMgqZkBQJkBREIBQkRadznAAoJECJPp+fMgqZkNxIQ ALZRqwdUGzqL2aeSavbum/VF/+td+nZfuH0xeWiO2w8mG0+nPd5j9ujYeHcUP1edE7uQrjOC Gs9sm8+W1xYnbClMJTsXiAV88D2btFUdU1mCXURAL9wWZ8Jsmz5ZH2V6AUszvNezsS/VIT87 AmTtj31TLDGwdxaZTSYLwAOOOtyqafOEq+gJB30RxTRE3h3G1zpO7OM9K6ysLdAlwAGYWgJJ V4JqGsQ/lyEtxxFpUCjb5Pztp7cQxhlkil0oBYHkudiG8j1U3DG8iC6rnB4yJaLphKx57NuQ PIY0Bccg+r9gIQ4XeSK2PQhdXdy3UWBr913ZQ9AI2usid3s5vabo4iBvpJNFLgUmxFnr73SJ KsRh/2OBsg1XXF/wRQGBO9vRuJUAbnaIVcmGOUogdBVS9Sun/Sy4GNA++KtFZK95U7J417/J Hub2xV6Ehc7UGW6fIvIQmzJ3zaTEfuriU1P8ayfddrAgZb25JnOW7L1zdYL8rXiezOyYZ8Fm ZyXjzWdO0RpxcUEp6GsJr11Bc4F3aae9OZtwtLL/jxc7y6pUugB00PodgnQ6CMcfR/HjXlae h2VS3zl9+tQWHu6s1R58t5BuMS2FNA58wU/IazImc/ZQA+slDBfhRDGYlExjg19UXWe/gMcl De3P1kxYPgZdGE2eZpRLIbt+rYnqQKy8UxlszsBNBFsZNTUBCACfQfpSsWJZyi+SHoRdVyX5 J6rI7okc4+b571a7RXD5UhS9dlVRVVAtrU9ANSLqPTQKGVxHrqD39XSw8hxK61pw8p90pg4G /N3iuWEvyt+t0SxDDkClnGsDyRhlUyEWYFEoBrrCizbmahOUwqkJbNMfzj5Y7n7OIJOxNRkB IBOjPdF26dMP69BwePQao1M8Acrrex9sAHYjQGyVmReRjVEtv9iG4DoTsnIR3amKVk6si4Ea X/mrapJqSCcBUVYUFH8M7bsm4CSxier5ofy8jTEa/CfvkqpKThTMCQPNZKY7hke5qEq1CBk2 wxhX48ZrJEFf1v3NuV3OimgsF2odzieNABEBAAHCwXwEGAEKACYCGwwWIQSpQNQ0mSwujpkQ PVAiT6fnzIKmZAUCZAUSmwUJDK5EZgAKCRAiT6fnzIKmZOJGEACOKABgo9wJXsbWhGWYO7mD 8R8mUyJHqbvaz+yTLnvRwfe/VwafFfDMx5GYVYzMY9TWpA8psFTKTUIIQmx2scYsRBUwm5VI EurRWKqENcDRjyo+ol59j0FViYysjQQeobXBDDE31t5SBg++veI6tXfpco/UiKEsDswL1WAr tEAZaruo7254TyH+gydURl2wJuzo/aZ7Y7PpqaODbYv727Dvm5eX64HCyyAH0s6sOCyGF5/p eIhrOn24oBf67KtdAN3H9JoFNUVTYJc1VJU3R1JtVdgwEdr+NEciEfYl0O19VpLE/PZxP4wX PWnhf5WjdoNI1Xec+RcJ5p/pSel0jnvBX8L2cmniYnmI883NhtGZsEWj++wyKiS4NranDFlA HdDM3b4lUth1pTtABKQ1YuTvehj7EfoWD3bv9kuGZGPrAeFNiHPdOT7DaXKeHpW9homgtBxj 8aX/UkSvEGJKUEbFL9cVa5tzyialGkSiZJNkWgeHe+jEcfRT6pJZOJidSCdzvJpbdJmm+eED w9XOLH1IIWh7RURU7G1iOfEfmImFeC3cbbS73LQEFGe1urxvIH5K/7vX+FkNcr9ujwWuPE9b 1C2o4i/yZPLXIVy387EjA6GZMqvQUFuSTs/GeBcv0NjIQi8867H3uLjz+mQy63fAitsDwLmR EP+ylKVEKb0Q2A== In-Reply-To: Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit X-Rspamd-Action: no action X-Rspamd-Queue-Id: 71EB040010 X-Rspamd-Server: rspam12 X-Stat-Signature: zqdkf1ncfgf3e5hr5dd1btxizjmbyg9n X-Rspam-User: X-HE-Tag: 1733994907-576927 X-HE-Meta: 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 ZxuVAvo4 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 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 12/12/24 03:49, Alexei Starovoitov wrote: > On Wed, Dec 11, 2024 at 3:55 AM Vlastimil Babka wrote: >> >> On 12/11/24 11:53, Vlastimil Babka wrote: >> > On 12/10/24 03:39, Alexei Starovoitov wrote: >> >> From: Alexei Starovoitov >> >> >> >> Similar to local_lock_irqsave() introduce local_trylock_irqsave(). >> >> It uses spin_trylock in PREEMPT_RT and always succeeds when !RT. >> > >> > Hmm but is that correct to always succeed? If we're in an nmi, we might be >> > preempting an existing local_(try)lock_irqsave() critical section because >> > disabling irqs doesn't stop NMI's, right? >> >> So unless I'm missing something, it would need to be a new kind of local >> lock to support this trylock operation on !RT? > > Ohh. Correct. Forgot about nmi interrupting local_lock_irqsave region in !RT. > >> Perhaps based on the same >> principle of a simple active/locked flag that I tried in my sheaves RFC? [1] >> There could be also the advantage that if all (potentially) irq contexts >> (not just nmi) used trylock, it would be sufficient to disable preeemption >> and not interrupts, which is cheaper. > > I don't think it's the case. > pushf was slow on old x86. > According to https://www.agner.org/optimize/instruction_tables.pdf > it's 3 uops on skylake. > That could be faster than preempt_disable (incl %gs:addr) > which is 3-4 uops assuming cache hit. I think the costly ones are not pushf, but cli and popf. I did some microbenchmark in the kernel (Ryzen 2700, not really latest thing but anyway) and IIRC it was twice slower to do the irqsave/restore than preempt only. >> The RT variant could work as you proposed here, that was wrong in my RFC as >> you already pointed out when we discussed v1 of this series. >> >> [1] >> https://lore.kernel.org/all/20241112-slub-percpu-caches-v1-5-ddc0bdc27e05@suse.cz/ > > I like your > +struct local_tryirq_lock > approach, but let's put it in local_lock.h ? Sure, that was a proof of concept so kept it local. > and it probably needs local_inc_return() instead of READ/WRITE_ONCE. > With irq and nmis it's racy. Hm guess you are right, thanks! > In the meantime I think I will fix below: > >> >> +#define __local_trylock_irqsave(lock, flags) \ >> >> + ({ \ >> >> + local_irq_save(flags); \ >> >> + local_trylock_acquire(this_cpu_ptr(lock)); \ >> >> + 1; \ >> >> + }) > > as > #define __local_trylock_irqsave(lock, flags) \ > ({ \ > local_irq_save(flags); \ > local_trylock_acquire(this_cpu_ptr(lock)); \ > !in_nmi(); \ > }) > > I think that's good enough for memcg patch 4 and > doesn't grow local_lock_t on !RT. But that means you'll never succeed in nmi, doesn't that limit the bpf use case? > We can introduce > > typedef struct { > int count; > #ifdef CONFIG_DEBUG_LOCK_ALLOC > struct lockdep_map dep_map; > struct task_struct *owner; > #endif > } local_trylock_t; > > and the whole set of local_trylock_lock, local_trylock_unlock,... > But that's quite a bit of code. Feels a bit overkill for memcg patch 4. SLUB also uses local_locks so it would be needed there later too. > At this point it feels that adding 'int count' to existing local_lock_t > is cleaner. We have Peter and Thomas in Cc let's see what they think :)