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 6940EC5321E for ; Mon, 26 Aug 2024 09:25:37 +0000 (UTC) Received: by kanga.kvack.org (Postfix) id 03B886B028D; Mon, 26 Aug 2024 05:25:37 -0400 (EDT) Received: by kanga.kvack.org (Postfix, from userid 40) id F2BAB6B04CD; Mon, 26 Aug 2024 05:25:36 -0400 (EDT) X-Delivered-To: int-list-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix, from userid 63042) id E1A9E6B04CE; Mon, 26 Aug 2024 05:25:36 -0400 (EDT) X-Delivered-To: linux-mm@kvack.org Received: from relay.hostedemail.com (smtprelay0017.hostedemail.com [216.40.44.17]) by kanga.kvack.org (Postfix) with ESMTP id C5E366B04CC for ; Mon, 26 Aug 2024 05:25:36 -0400 (EDT) Received: from smtpin27.hostedemail.com (a10.router.float.18 [10.200.18.1]) by unirelay03.hostedemail.com (Postfix) with ESMTP id 6C573A0A50 for ; Mon, 26 Aug 2024 09:25:36 +0000 (UTC) X-FDA: 82493863872.27.E7434BC Received: from nyc.source.kernel.org (nyc.source.kernel.org [147.75.193.91]) by imf18.hostedemail.com (Postfix) with ESMTP id CFA151C000F for ; Mon, 26 Aug 2024 09:25:34 +0000 (UTC) Authentication-Results: imf18.hostedemail.com; dkim=pass header.d=kernel.org header.s=k20201202 header.b=h59wbWnV; dmarc=pass (policy=quarantine) header.from=kernel.org; spf=pass (imf18.hostedemail.com: domain of brauner@kernel.org designates 147.75.193.91 as permitted sender) smtp.mailfrom=brauner@kernel.org ARC-Seal: i=1; s=arc-20220608; d=hostedemail.com; t=1724664271; a=rsa-sha256; cv=none; b=2mbhqCap/ptR1SjtHC8PHl/POIF00MblRgSwgeniD+L1d3GlzVAXcSfgWr/01XHdHhW/M2 4ZmjEwsuxdlAZt+xOxcIa0Xwvw7135u7npr9Y1jxBpgrIKd38MYQSdq1Y04sbxIznQt1OL eo3ZFYib+7kR4VRQ/jeeCY8v9IKcScs= ARC-Authentication-Results: i=1; imf18.hostedemail.com; dkim=pass header.d=kernel.org header.s=k20201202 header.b=h59wbWnV; dmarc=pass (policy=quarantine) header.from=kernel.org; spf=pass (imf18.hostedemail.com: domain of brauner@kernel.org designates 147.75.193.91 as permitted sender) smtp.mailfrom=brauner@kernel.org ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=hostedemail.com; s=arc-20220608; t=1724664271; 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: in-reply-to:in-reply-to:references:references:dkim-signature; bh=jfEqq+XrzbMgH/PA5j2gQJj9l1FTEIseey30IA980g8=; b=1SzmwQZTyjNEEtyO57vnLCzEU1Jx6FNYuK5dEWHQb/4qISTST7DdL/hrRjTtLsJidmfOza LPcSjS/OnPlEseVuPoS7fkCGU6XkSwqWu4RfFTakpKEFaYBtvxCLebjdTNCfx4vWMyDU8z OMXhesmKQZO/WeD/dLkhg1KdZx1BOMI= Received: from smtp.kernel.org (transwarp.subspace.kernel.org [100.75.92.58]) by nyc.source.kernel.org (Postfix) with ESMTP id 5A84AA41D8A; Mon, 26 Aug 2024 09:25:27 +0000 (UTC) Received: by smtp.kernel.org (Postfix) with ESMTPSA id 2CBD6C5812B; Mon, 26 Aug 2024 09:25:30 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=k20201202; t=1724664333; bh=yR5dBquNpjlZ++EesY9D5XF2kbdtsmQUbpZyrMOeLyY=; h=Date:From:To:Cc:Subject:References:In-Reply-To:From; b=h59wbWnVWnkJJZffVobkoaGnY9shwl2c/vKofj8cXV2HnmS2LBT/avy2y+QI++8tG hYNSCgfMB08KboNxjVB5WpjyGHNs3ccT7m1K/1fJzwYvWIpfZyOxNBbYC3TCusF0Lg ToKQTB67d+cSmAyja7phY/G2WQj7rbAAkLvGJR6egLy/Duvh4vwxH6pQrksIPJ6NZ+ 8SwgoFIkbpSOeiwdiFbxyM6lLdXFaPwWQLA9pM+afj+ghudCxQkFgTg6vcbBBk693N ilshaYPirUC5XcwhTwmwLuOe/YnvCNrGtUzNkh8YUW9wuBk76cAz3jHF56stvPoqzT +e2fUlrOeNpQg== Date: Mon, 26 Aug 2024 11:25:28 +0200 From: Christian Brauner To: Vlastimil Babka , linux-mm@kvack.org Cc: Jens Axboe , "Paul E. McKenney" , Roman Gushchin , Linus Torvalds , Jann Horn Subject: Re: [PATCH] [POC/RFE]: Avoid silently growing struct file due to SLAB_TYPESAFE_BY_RCU Message-ID: <20240826-zweihundert-luftkammer-9b2b47f4338e@brauner> References: <20240826-benahm-sorgenfrei-245a41a9acb4@brauner> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline In-Reply-To: <20240826-benahm-sorgenfrei-245a41a9acb4@brauner> X-Rspamd-Queue-Id: CFA151C000F X-Rspam-User: X-Rspamd-Server: rspam05 X-Stat-Signature: 74dgdq7izdngxh6nfo366zrscyxbg9h7 X-HE-Tag: 1724664334-620591 X-HE-Meta: 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 2Gv49rda AwwBMTCe08es0u5+/VGOrY9EJiLcnzlScmijRdc6ySiB6iwKmmMvPY/3wcZ90fK71eTxPydxxjGoH0BNl6z++SNpDPyqXd8xHIYcK0gdWBqH54ZFNSi0TcuTTImktrJ2SxA2J7G+oXkxbHTT+LPxshDCsBhwIOH7WCzrh0+Q8f0qHHU3lz8WVlDDtjcJAeBLZ93iJSvtPp99EkKovj2OQdU7e7NFnSQj7bVozJgfcF52UfmYm0cANwJ+Ywpf0a1g6FVlaBldiH2om3iYEwXldtXpt0oUBYhdRPrn7YH2VOLMY5CKnTVXUIVzWM6pwE4t3P/2SjnjPr6g2WhypZhy0wQAfWfXUR2o0M7oaTE104zjdWtPJqnZ2TWGiHupVcsTQb62x 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: [+ Jann] On Mon, Aug 26, 2024 at 10:41:07AM GMT, Christian Brauner wrote: > I've managed to shrink struct file to 192 bytes aka three cachelines. > What I wasn't aware of is that SLAB_TYPESAFE_BY_RCU silently grows the > struct by another cacheline to accomodate the freelist pointer as > pointed out by Vlastimil. > > It's not too bad because we mostly care about "runtime" size but it's > still unfortunate that we're wasting a whole additional cacheline for no > good reason. > > So I would like to find a way to avoid that. Below you see a hack > because adding a parameter to kmem_cache_create() seemed to require me > to fiddle with mm/slub.c and I'd rather not unless I'm told how. > > For the POC I'm using the offset into the union at the end of struct > file so that SLUB places the freelist pointer into the union during > kmem_cache_free(). > > I'm just looking for ideas from mm to figure out an API where we can say > "use that as the freelist pointer in the struct" or at least a reason > why that's just not possible. > > Sketched-by: Christian Brauner > --- > fs/file_table.c | 3 ++- > include/linux/fs.h | 2 ++ > 2 files changed, 4 insertions(+), 1 deletion(-) > > diff --git a/fs/file_table.c b/fs/file_table.c > index 694199a1a966..7a5fac7d4bbd 100644 > --- a/fs/file_table.c > +++ b/fs/file_table.c > @@ -514,7 +514,8 @@ EXPORT_SYMBOL(__fput_sync); > > void __init files_init(void) > { > - filp_cachep = kmem_cache_create("filp", sizeof(struct file), 0, > + filp_cachep = kmem_cache_create("filp", > + offsetof(struct file, __f_slab_rcu), 0, > SLAB_TYPESAFE_BY_RCU | SLAB_HWCACHE_ALIGN | > SLAB_PANIC | SLAB_ACCOUNT, NULL); > percpu_counter_init(&nr_files, 0, GFP_KERNEL); > diff --git a/include/linux/fs.h b/include/linux/fs.h > index 36bb8456eb38..6c3664d624a0 100644 > --- a/include/linux/fs.h > +++ b/include/linux/fs.h > @@ -1010,6 +1010,7 @@ static inline int ra_has_index(struct file_ra_state *ra, pgoff_t index) > * @f_task_work: task work entry point > * @f_llist: work queue entrypoint > * @f_ra: file's readahead state > + * @f_slab_rcu: freelist pointer for SLAB_TYPESAFE_BY_RCU > */ > struct file { > atomic_long_t f_count; > @@ -1041,6 +1042,7 @@ struct file { > struct callback_head f_task_work; > struct llist_node f_llist; > struct file_ra_state f_ra; > + void *__f_slab_rcu; > }; > /* --- cacheline 3 boundary (192 bytes) --- */ > } __randomize_layout > -- > 2.45.2