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 X-Spam-Level: X-Spam-Status: No, score=-13.8 required=3.0 tests=BAYES_00,DKIM_SIGNED, DKIM_VALID,HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 69C79C47080 for ; Tue, 1 Jun 2021 19:30:37 +0000 (UTC) Received: from kanga.kvack.org (kanga.kvack.org [205.233.56.17]) by mail.kernel.org (Postfix) with ESMTP id D5C926136E for ; Tue, 1 Jun 2021 19:30:36 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org D5C926136E Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=intel.com Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=owner-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix) id 6E0336B006C; Tue, 1 Jun 2021 15:30:36 -0400 (EDT) Received: by kanga.kvack.org (Postfix, from userid 40) id 690B06B006E; Tue, 1 Jun 2021 15:30:36 -0400 (EDT) X-Delivered-To: int-list-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix, from userid 63042) id 531946B0070; Tue, 1 Jun 2021 15:30:36 -0400 (EDT) X-Delivered-To: linux-mm@kvack.org Received: from forelay.hostedemail.com (smtprelay0037.hostedemail.com [216.40.44.37]) by kanga.kvack.org (Postfix) with ESMTP id 1E7F76B006C for ; Tue, 1 Jun 2021 15:30:36 -0400 (EDT) Received: from smtpin10.hostedemail.com (10.5.19.251.rfc1918.com [10.5.19.251]) by forelay05.hostedemail.com (Postfix) with ESMTP id A0B98181AEF07 for ; Tue, 1 Jun 2021 19:30:35 +0000 (UTC) X-FDA: 78206146830.10.351605F Received: from mail-pj1-f53.google.com (mail-pj1-f53.google.com [209.85.216.53]) by imf08.hostedemail.com (Postfix) with ESMTP id 07EB08019352 for ; Tue, 1 Jun 2021 19:30:20 +0000 (UTC) Received: by mail-pj1-f53.google.com with SMTP id m13-20020a17090b068db02901656cc93a75so2040139pjz.3 for ; Tue, 01 Jun 2021 12:30:32 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=intel-com.20150623.gappssmtp.com; s=20150623; h=mime-version:references:in-reply-to:from:date:message-id:subject:to :cc; bh=NP8oMHG5bQo51HvT2u0k4h1m9ygxJ6vIvzhY8PC1eaY=; b=tjqacv1sMpi9foss0VSmv3HCnlDqDAtH/DuUQ0bqVX6KTidhMGccd6fi7R9HRHfTDL 52dwV7fqbrFpD9jLdJpqxJMXPUdduFCe087hk8d/W6d8LDfWnlexpjgCDQ67KzwyaZpA AA7H3uwDButu3BQQrqplGo7u1cmjYOgUNdfMWfPJ9Lhm+CRcR4x9x5Mq7SfPi4kBC8Vh 9Gbhk53hPcbdjAGddvsDbjtLC/hwKTIj+bwEi6sCuNRdIi3sn22OsbRRENWlCFb23eV2 x65kwLAPal+MPpr3z3H1041Vwu7WUNxbvMUKZA24Gzve5BFNlU/ok+Ybzb+XKD8V85/k 4prg== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:mime-version:references:in-reply-to:from:date :message-id:subject:to:cc; bh=NP8oMHG5bQo51HvT2u0k4h1m9ygxJ6vIvzhY8PC1eaY=; b=q8eUYcyUbva3LGk10I4Jjxt9XKrvysjTbqrafkm1Smkn657fbm5vqKwfTfuHOx+vTn YFZYXWUGGMIBGIosgcVguq4bxVtDksRJr6VTXnsLcvOeAdaYO8L8THGWWNIZEcKCdWNR JmRRLaDCb/Q4dEzYIQYR/VLtypZ9ptjdmdTgndvVP5KPYQwJIPq+YNL6xvgzMvIuciiL 7Yql4VxlXwOqI0Uzlrvz2VDNPpfTAivAfpFZeYGq1AAqzCpt9u8RoBQ4PHtsK7p+Msid npm0YpYJ5kkkP7bP/C6jZHjJYeqWpaIPo611z63MthIVxfjPLkUFI23/+oiubz8krs1n 4RUg== X-Gm-Message-State: AOAM532sYhr02Wp4Uv0299FRtXlQQ/Am+72G3WcxYYKtgYIcWz1P/0sc fqfRlpHtzRhK/TKOpmyzte0YrCJ1CMwwZVTtCnX0ow== X-Google-Smtp-Source: ABdhPJxUuo7nshnzaSlqRQFEgaEmdIi3vzzj7qrOeAzMW+kfVTmq9bvMbE362zINEg12QouusnurNKfIErPcb05LaSI= X-Received: by 2002:a17:902:8203:b029:101:26a8:764d with SMTP id x3-20020a1709028203b029010126a8764dmr21436994pln.79.1622575831834; Tue, 01 Jun 2021 12:30:31 -0700 (PDT) MIME-Version: 1.0 References: <20210325230938.30752-1-joao.m.martins@oracle.com> <20210325230938.30752-9-joao.m.martins@oracle.com> In-Reply-To: <20210325230938.30752-9-joao.m.martins@oracle.com> From: Dan Williams Date: Tue, 1 Jun 2021 12:30:21 -0700 Message-ID: Subject: Re: [PATCH v1 08/11] mm/sparse-vmemmap: use hugepages for PUD compound pagemaps To: Joao Martins Cc: Linux MM , Ira Weiny , linux-nvdimm , Matthew Wilcox , Jason Gunthorpe , Jane Chu , Muchun Song , Mike Kravetz , Andrew Morton Content-Type: text/plain; charset="UTF-8" X-Rspamd-Queue-Id: 07EB08019352 Authentication-Results: imf08.hostedemail.com; dkim=pass header.d=intel-com.20150623.gappssmtp.com header.s=20150623 header.b=tjqacv1s; dmarc=fail reason="No valid SPF, DKIM not aligned (relaxed)" header.from=intel.com (policy=none); spf=none (imf08.hostedemail.com: domain of dan.j.williams@intel.com has no SPF policy when checking 209.85.216.53) smtp.mailfrom=dan.j.williams@intel.com X-Rspamd-Server: rspam03 X-Stat-Signature: hgo8cnywbs5dxdjqn6z46cxmwzre9a8a X-HE-Tag: 1622575820-186400 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: Sorry for the delay, and the sync up on IRC. I think this subject needs to change to "optimize memory savings for compound pud memmap geometry", and move this to the back of the series to make it clear which patches are base functionality and which extend the idea further. As far as I can see this patch can move to the end of the series. Some additional changelog feedback below: On Thu, Mar 25, 2021 at 4:10 PM Joao Martins wrote: > > Right now basepages are used to populate the PUD tail pages, and it > picks the address of the previous page of the subsection that preceeds s/preceeds/precedes/ > the memmap we are initializing. This is done when a given memmap > address isn't aligned to the pgmap @align (which is safe to do because > @ranges are guaranteed to be aligned to @align). You know what would help is if you could draw an ascii art picture of the before and after of the head page vs tail page arrangement. I can see how the words line up to the code, but it takes a while to get the picture in my head and I think future work in this area will benefit from having a place in Documentation that draws a picture of the layout of the various geometries. I've used asciiflow.com for these types of diagrams in the past. > > For pagemaps with an align which spans various sections, this means > that PMD pages are unnecessarily allocated for reusing the same tail > pages. Effectively, on x86 a PUD can span 8 sections (depending on > config), and a page is being allocated a page for the PMD to reuse > the tail vmemmap across the rest of the PTEs. In short effecitvely the > PMD cover the tail vmemmap areas all contain the same PFN. So instead > of doing this way, populate a new PMD on the second section of the > compound page (tail vmemmap PMD), and then the following sections > utilize the preceding PMD we previously populated which only contain > tail pages). > > After this scheme for an 1GB pagemap aligned area, the first PMD > (section) would contain head page and 32767 tail pages, where the > second PMD contains the full 32768 tail pages. The latter page gets > its PMD reused across future section mapping of the same pagemap. > > Besides fewer pagetable entries allocated, keeping parity with > hugepages in the directmap (as done by vmemmap_populate_hugepages()), > this further increases savings per compound page. For each PUD-aligned > pagemap we go from 40960 bytes down to 16384 bytes. Rather than > requiring 8 PMD page allocations we only need 2 (plus two base pages > allocated for head and tail areas for the first PMD). 2M pages still > require using base pages, though. I would front load the above savings to the top of this discussion and say something like: "Currently, for compound PUD mappings, the implementation consumes X GB per TB this can be optimized to Y GB per TB with the approach detailed below." > > Signed-off-by: Joao Martins > --- > include/linux/mm.h | 3 +- > mm/sparse-vmemmap.c | 79 ++++++++++++++++++++++++++++++++++----------- > 2 files changed, 63 insertions(+), 19 deletions(-) > > diff --git a/include/linux/mm.h b/include/linux/mm.h > index 49d717ae40ae..9c1a676d6b95 100644 > --- a/include/linux/mm.h > +++ b/include/linux/mm.h > @@ -3038,7 +3038,8 @@ struct page * __populate_section_memmap(unsigned long pfn, > pgd_t *vmemmap_pgd_populate(unsigned long addr, int node); > p4d_t *vmemmap_p4d_populate(pgd_t *pgd, unsigned long addr, int node); > pud_t *vmemmap_pud_populate(p4d_t *p4d, unsigned long addr, int node); > -pmd_t *vmemmap_pmd_populate(pud_t *pud, unsigned long addr, int node); > +pmd_t *vmemmap_pmd_populate(pud_t *pud, unsigned long addr, int node, > + void *block); > pte_t *vmemmap_pte_populate(pmd_t *pmd, unsigned long addr, int node, > struct vmem_altmap *altmap, void *block); > void *vmemmap_alloc_block(unsigned long size, int node); > diff --git a/mm/sparse-vmemmap.c b/mm/sparse-vmemmap.c > index f57c5eada099..291a8a32480a 100644 > --- a/mm/sparse-vmemmap.c > +++ b/mm/sparse-vmemmap.c > @@ -172,13 +172,20 @@ static void * __meminit vmemmap_alloc_block_zero(unsigned long size, int node) > return p; > } > > -pmd_t * __meminit vmemmap_pmd_populate(pud_t *pud, unsigned long addr, int node) > +pmd_t * __meminit vmemmap_pmd_populate(pud_t *pud, unsigned long addr, int node, > + void *block) > { > pmd_t *pmd = pmd_offset(pud, addr); > if (pmd_none(*pmd)) { > - void *p = vmemmap_alloc_block_zero(PAGE_SIZE, node); > - if (!p) > - return NULL; > + void *p = block; > + > + if (!block) { > + p = vmemmap_alloc_block_zero(PAGE_SIZE, node); > + if (!p) > + return NULL; > + } else { > + get_page(virt_to_page(block)); > + } > pmd_populate_kernel(&init_mm, pmd, p); > } > return pmd; > @@ -220,15 +227,14 @@ pgd_t * __meminit vmemmap_pgd_populate(unsigned long addr, int node) > return pgd; > } > > -static int __meminit vmemmap_populate_address(unsigned long addr, int node, > - struct vmem_altmap *altmap, > - void *page, void **ptr) > +static int __meminit vmemmap_populate_pmd_address(unsigned long addr, int node, > + struct vmem_altmap *altmap, > + void *page, pmd_t **ptr) > { > pgd_t *pgd; > p4d_t *p4d; > pud_t *pud; > pmd_t *pmd; > - pte_t *pte; > > pgd = vmemmap_pgd_populate(addr, node); > if (!pgd) > @@ -239,9 +245,24 @@ static int __meminit vmemmap_populate_address(unsigned long addr, int node, > pud = vmemmap_pud_populate(p4d, addr, node); > if (!pud) > return -ENOMEM; > - pmd = vmemmap_pmd_populate(pud, addr, node); > + pmd = vmemmap_pmd_populate(pud, addr, node, page); > if (!pmd) > return -ENOMEM; > + if (ptr) > + *ptr = pmd; > + return 0; > +} > + > +static int __meminit vmemmap_populate_address(unsigned long addr, int node, > + struct vmem_altmap *altmap, > + void *page, void **ptr) > +{ > + pmd_t *pmd; > + pte_t *pte; > + > + if (vmemmap_populate_pmd_address(addr, node, altmap, NULL, &pmd)) > + return -ENOMEM; > + > pte = vmemmap_pte_populate(pmd, addr, node, altmap, page); > if (!pte) > return -ENOMEM; > @@ -285,13 +306,26 @@ static inline int __meminit vmemmap_populate_page(unsigned long addr, int node, > return vmemmap_populate_address(addr, node, NULL, NULL, ptr); > } > > -static pte_t * __meminit vmemmap_lookup_address(unsigned long addr) > +static int __meminit vmemmap_populate_pmd_range(unsigned long start, > + unsigned long end, > + int node, void *page) > +{ > + unsigned long addr = start; > + > + for (; addr < end; addr += PMD_SIZE) { > + if (vmemmap_populate_pmd_address(addr, node, NULL, page, NULL)) > + return -ENOMEM; > + } > + > + return 0; > +} > + > +static pmd_t * __meminit vmemmap_lookup_address(unsigned long addr) > { > pgd_t *pgd; > p4d_t *p4d; > pud_t *pud; > pmd_t *pmd; > - pte_t *pte; > > pgd = pgd_offset_k(addr); > if (pgd_none(*pgd)) > @@ -309,11 +343,7 @@ static pte_t * __meminit vmemmap_lookup_address(unsigned long addr) > if (pmd_none(*pmd)) > return NULL; > > - pte = pte_offset_kernel(pmd, addr); > - if (pte_none(*pte)) > - return NULL; > - > - return pte; > + return pmd; > } > > static int __meminit vmemmap_populate_compound_pages(unsigned long start_pfn, > @@ -335,9 +365,22 @@ static int __meminit vmemmap_populate_compound_pages(unsigned long start_pfn, > offset = PFN_PHYS(start_pfn) - pgmap->ranges[pgmap->nr_range].start; > if (!IS_ALIGNED(offset, pgmap_align(pgmap)) && > pgmap_align(pgmap) > SUBSECTION_SIZE) { > - pte_t *ptep = vmemmap_lookup_address(start - PAGE_SIZE); > + pmd_t *pmdp; > + pte_t *ptep; > + > + addr = start - PAGE_SIZE; > + pmdp = vmemmap_lookup_address(addr); > + if (!pmdp) > + return -ENOMEM; > + > + /* Reuse the tail pages vmemmap pmd page */ This comment really wants to be "See layout diagram in Documentation/vm/compound_pagemaps.rst", because the reuse algorithm is not immediately obvious. Other than that the implementation looks ok to me, modulo previous comments about @block type and the use of the "geometry" term.