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 25D5FC282EC for ; Tue, 11 Mar 2025 21:00:30 +0000 (UTC) Received: by kanga.kvack.org (Postfix) id 320C7280003; Tue, 11 Mar 2025 17:00:28 -0400 (EDT) Received: by kanga.kvack.org (Postfix, from userid 40) id 2A95F280001; Tue, 11 Mar 2025 17:00:28 -0400 (EDT) X-Delivered-To: int-list-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix, from userid 63042) id 17161280003; Tue, 11 Mar 2025 17:00:28 -0400 (EDT) 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 E65CA280001 for ; Tue, 11 Mar 2025 17:00:27 -0400 (EDT) Received: from smtpin23.hostedemail.com (a10.router.float.18 [10.200.18.1]) by unirelay04.hostedemail.com (Postfix) with ESMTP id C36601A0690 for ; Tue, 11 Mar 2025 21:00:28 +0000 (UTC) X-FDA: 83210488536.23.8AD3416 Received: from nyc.source.kernel.org (nyc.source.kernel.org [147.75.193.91]) by imf11.hostedemail.com (Postfix) with ESMTP id 09F6A40010 for ; Tue, 11 Mar 2025 21:00:26 +0000 (UTC) Authentication-Results: imf11.hostedemail.com; dkim=pass header.d=kernel.org header.s=k20201202 header.b=EybcrF6m; dmarc=pass (policy=quarantine) header.from=kernel.org; spf=pass (imf11.hostedemail.com: domain of sj@kernel.org designates 147.75.193.91 as permitted sender) smtp.mailfrom=sj@kernel.org ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=hostedemail.com; s=arc-20220608; t=1741726827; 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-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references:dkim-signature; bh=apUjkLps3PAsd0xxMcDStWEq2JOZjEXGpLmbWK6/4Nc=; b=UG6DILx8IlAb/unooh7N1rf2AqxQdb/RM3j2MsBsmUjgyYwnQdKCmTcePmU0su+k+ImuAp qQfq9nUKj0XFtjLqrEXo/huRUuptIOrxweNxD+ZitJ0xsfOwMb0CAmiM9820bjJyxD/LH0 NF5csiT48Tg6yJ0D9sB+zQeArzBdkTw= ARC-Seal: i=1; s=arc-20220608; d=hostedemail.com; t=1741726827; a=rsa-sha256; cv=none; b=S36Bc0MjefSqJz6k8sWdSU4Apy5zDDZd/73eUTp7lJOMpCohnpQPCeMLP1cFSbx2hJFc6E Wd7O3VrSOYyYsoqMm7792SsjEj9RJ6GuXyYVUvd+vJzcmOcHShzm+W6Sy8w/mpxn1u8Mei ErbSghe0R9GrsuqvYWeC73uESdVXx2k= ARC-Authentication-Results: i=1; imf11.hostedemail.com; dkim=pass header.d=kernel.org header.s=k20201202 header.b=EybcrF6m; dmarc=pass (policy=quarantine) header.from=kernel.org; spf=pass (imf11.hostedemail.com: domain of sj@kernel.org designates 147.75.193.91 as permitted sender) smtp.mailfrom=sj@kernel.org Received: from smtp.kernel.org (transwarp.subspace.kernel.org [100.75.92.58]) by nyc.source.kernel.org (Postfix) with ESMTP id 0F057A4696D; Tue, 11 Mar 2025 20:54:56 +0000 (UTC) Received: by smtp.kernel.org (Postfix) with ESMTPSA id 7D71AC4CEE9; Tue, 11 Mar 2025 21:00:25 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=k20201202; t=1741726825; bh=QLkONEElsgGKxIzNwS3VferQTigrHPHxuZz5dFwdp5s=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=EybcrF6mztt/bLYNcuSF2rCGMEPvxvMLV/UEQGBqRccO0fFU0RoFznkIYcX/c+NLf Q0kxEjONBkyhvtip0iJkEaZihjA0e2TENb/PBuBlNPhVcbVgilBsFMz4HAAxwoL2m4 ZaJXyj8rW2EO9iTwjElzPDOnK+R743BcX7ouTbaNdgTS9VF6rcjmYrVcAmkQ+eRf79 Xm37L0xrPhd0glAOPz6Py3uVce2Ync3qlXsGPxqEEKkLVeyLPIOpdg6u7fYLw0wvAh 8rIe4nKRawf+/SBEhgyvxZGxR57THsyCpMkHtlzGY1CZ/lvOmqSYrMFqCy0cVth/s8 vi0dZ6rNs0PIQ== From: SeongJae Park To: Lorenzo Stoakes Cc: SeongJae Park , Andrew Morton , "Liam R. Howlett" , David Hildenbrand , Shakeel Butt , Vlastimil Babka , kernel-team@meta.com, linux-kernel@vger.kernel.org, linux-mm@kvack.org Subject: Re: [PATCH 7/9] mm/madvise: let madvise_{dontneed,free}_single_vma() caller batches tlb flushes Date: Tue, 11 Mar 2025 14:00:23 -0700 Message-Id: <20250311210023.85435-1-sj@kernel.org> X-Mailer: git-send-email 2.39.5 In-Reply-To: References: MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Rspam-User: X-Rspamd-Server: rspam04 X-Rspamd-Queue-Id: 09F6A40010 X-Stat-Signature: o6anjujfwrsnmwkhsmfujpfinfw136nf X-HE-Tag: 1741726826-307792 X-HE-Meta: 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 AgFGHQH+ Z1orRjsyR/wALNLcGVmdTCGAQ/uJUke0Z9nZUrlzD3mMQYOoi5Ne8fN6w8Z0pN3eQddd5hT5IgvMNH3uo+bmhEIKER7hb1G9JG1k2A2aubfcmK1e6z84Mr7ScygbXCW1EyJ5YbAdRyWF0A4sf7OHrgd4v4RsHBTja3koPGnHgx58p8K9HWndj7KeNcqirW0FAAAwcINxaK/z5wI2NOa7yDuSnA9LKVOJqI1oN7QBZ1zVKUg3dF9h/BUeDq28kyjOt+wUMNIgo/PTMIKyb9zZziufbycTLZB6R8PBloHFas10xGz0= 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 Tue, 11 Mar 2025 13:07:25 +0000 Lorenzo Stoakes wrote: > Super super UBER nitty but... pretty sure the subject here should be <= 75 > chars right? :P I believe that's not a hard limit, but I will try to make it shorter in the next spin. > > On Mon, Mar 10, 2025 at 10:23:16AM -0700, SeongJae Park wrote: > > Update madvise_dontneed_single_vma() and madvise_free_single_vma() > > functions so that the caller can pass an mmu_gather object that should > > be initialized and will be finished outside, for batched tlb flushes. > > Also modify their internal code to support such usage by skipping the > > initialization and finishing of self-allocated mmu_gather object if it > > received a valid mmu_gather object. > > > > Signed-off-by: SeongJae Park > > --- > > mm/internal.h | 3 +++ > > mm/madvise.c | 37 +++++++++++++++++++++++++------------ > > mm/memory.c | 16 +++++++++++++--- > > 3 files changed, 41 insertions(+), 15 deletions(-) > > > > diff --git a/mm/internal.h b/mm/internal.h > > index 0caa64dc2cb7..ce7fb2383f65 100644 > > --- a/mm/internal.h > > +++ b/mm/internal.h > > @@ -438,6 +438,9 @@ void unmap_page_range(struct mmu_gather *tlb, > > struct vm_area_struct *vma, > > unsigned long addr, unsigned long end, > > struct zap_details *details); > > +void unmap_vma_single(struct mmu_gather *tlb, struct vm_area_struct *vma, > > + unsigned long addr, unsigned long size, > > + struct zap_details *details); > > int folio_unmap_invalidate(struct address_space *mapping, struct folio *folio, > > gfp_t gfp); > > > > diff --git a/mm/madvise.c b/mm/madvise.c > > index ba2a78795207..d7ea71c6422c 100644 > > --- a/mm/madvise.c > > +++ b/mm/madvise.c > > @@ -794,12 +794,19 @@ static const struct mm_walk_ops madvise_free_walk_ops = { > > .walk_lock = PGWALK_RDLOCK, > > }; > > > > -static int madvise_free_single_vma(struct vm_area_struct *vma, > > - unsigned long start_addr, unsigned long end_addr) > > +static int madvise_free_single_vma( > > + struct mmu_gather *caller_tlb, struct vm_area_struct *vma, > > I find this interface horrible, and super confusing. It's not clear at all > what's going on here. > > Why not use your new helper struct to add a field you can thread through > here? I will do so in the next spin. > > > + unsigned long start_addr, unsigned long end_addr) > > { > > struct mm_struct *mm = vma->vm_mm; > > struct mmu_notifier_range range; > > - struct mmu_gather tlb; > > + struct mmu_gather self_tlb; > > + struct mmu_gather *tlb; > > + > > + if (caller_tlb) > > + tlb = caller_tlb; > > + else > > + tlb = &self_tlb; > > > > /* MADV_FREE works for only anon vma at the moment */ > > if (!vma_is_anonymous(vma)) > > @@ -815,16 +822,18 @@ static int madvise_free_single_vma(struct vm_area_struct *vma, > > range.start, range.end); > > > > lru_add_drain(); > > - tlb_gather_mmu(&tlb, mm); > > + if (!caller_tlb) > > + tlb_gather_mmu(tlb, mm); > > Yeah really don't like this. > > Ideally we'd abstract the mmu_gather struct to the helper struct (which I > see you do in a subsequent patch anyway) would be ideal if you could find a > way to make that work. > > But if not, then: > > if (behavior->batched_tlb) > tlb_gather_mmu(&tlb, mm); > > etc. etc. > > Would work better. Agreed. > > > update_hiwater_rss(mm); > > > > mmu_notifier_invalidate_range_start(&range); > > - tlb_start_vma(&tlb, vma); > > + tlb_start_vma(tlb, vma); > > Also not a fan of making tlb refer to a pointer now when before it > didn't... I mean that's more of a nit and maybe unavoidable, but still! > > I mean yeah ok this is probably unavoidable, ignore. Yeah... I also find no good way to make this very cleaner without the followup cleanup for now. > > > walk_page_range(vma->vm_mm, range.start, range.end, > > - &madvise_free_walk_ops, &tlb); > > - tlb_end_vma(&tlb, vma); > > + &madvise_free_walk_ops, tlb); > > + tlb_end_vma(tlb, vma); > > mmu_notifier_invalidate_range_end(&range); > > - tlb_finish_mmu(&tlb); > > + if (!caller_tlb) > > + tlb_finish_mmu(tlb); > > > > return 0; > > } > > @@ -848,7 +857,8 @@ static int madvise_free_single_vma(struct vm_area_struct *vma, > > * An interface that causes the system to free clean pages and flush > > * dirty pages is already available as msync(MS_INVALIDATE). > > */ > > -static long madvise_dontneed_single_vma(struct vm_area_struct *vma, > > +static long madvise_dontneed_single_vma(struct mmu_gather *tlb, > > + struct vm_area_struct *vma, > > unsigned long start, unsigned long end) > > { > > struct zap_details details = { > > @@ -856,7 +866,10 @@ static long madvise_dontneed_single_vma(struct vm_area_struct *vma, > > .even_cows = true, > > }; > > > > - zap_page_range_single(vma, start, end - start, &details); > > + if (!tlb) > > + zap_page_range_single(vma, start, end - start, &details); > > Please don't put the negation case first, it's confusing. Swap them! Ok, I will do so. > > > > + else > > + unmap_vma_single(tlb, vma, start, end - start, &details); > > return 0; > > } > > > > @@ -951,9 +964,9 @@ static long madvise_dontneed_free(struct vm_area_struct *vma, > > } > > > > if (behavior == MADV_DONTNEED || behavior == MADV_DONTNEED_LOCKED) > > - return madvise_dontneed_single_vma(vma, start, end); > > + return madvise_dontneed_single_vma(NULL, vma, start, end); > > else if (behavior == MADV_FREE) > > - return madvise_free_single_vma(vma, start, end); > > + return madvise_free_single_vma(NULL, vma, start, end); > > Not to labour the point, but this is also horrid, passing a mystery NULL > parameter first... Agreed again. I will just pass the madvise_behavior struct in the next spin. > > > else > > return -EINVAL; > > } > > diff --git a/mm/memory.c b/mm/memory.c > > index 88c478e2ed1a..3256b9713cbd 100644 > > --- a/mm/memory.c > > +++ b/mm/memory.c > > @@ -1995,9 +1995,19 @@ void unmap_vmas(struct mmu_gather *tlb, struct ma_state *mas, > > mmu_notifier_invalidate_range_end(&range); > > } > > > > -static void unmap_vma_single(struct mmu_gather *tlb, > > - struct vm_area_struct *vma, unsigned long address, > > - unsigned long size, struct zap_details *details) > > +/** > > + * unmap_vma_single - remove user pages in a given range > > + * @tlb: pointer to the caller's struct mmu_gather > > + * @vma: vm_area_struct holding the applicable pages > > + * @address: starting address of the pages > > + * @size: number of bytes to remove > > + * @details: details of shared cache invalidation > > + * > > + * @tlb shouldn't be NULL. The range must fit into one VMA. > > Can we add some VM_WARN_ON[_ONCE]()'s for these conditions please? Nice suggestion, I will do so. > > Thanks for documenting! Kudos to Shakeel, who suggested this kerneldoc comment :) Thanks, SJ [...]