Linux-Fsdevel Archive on lore.kernel.org
 help / color / Atom feed
From: Muchun Song <songmuchun@bytedance.com>
To: Oscar Salvador <osalvador@suse.de>
Cc: Jonathan Corbet <corbet@lwn.net>,
	Mike Kravetz <mike.kravetz@oracle.com>,
	Thomas Gleixner <tglx@linutronix.de>,
	mingo@redhat.com, bp@alien8.de, x86@kernel.org, hpa@zytor.com,
	dave.hansen@linux.intel.com, luto@kernel.org,
	Peter Zijlstra <peterz@infradead.org>,
	viro@zeniv.linux.org.uk,
	Andrew Morton <akpm@linux-foundation.org>,
	paulmck@kernel.org, mchehab+huawei@kernel.org,
	pawan.kumar.gupta@linux.intel.com,
	Randy Dunlap <rdunlap@infradead.org>,
	oneukum@suse.com, anshuman.khandual@arm.com, jroedel@suse.de,
	Mina Almasry <almasrymina@google.com>,
	David Rientjes <rientjes@google.com>,
	Matthew Wilcox <willy@infradead.org>,
	Xiongchun duan <duanxiongchun@bytedance.com>,
	linux-doc@vger.kernel.org, LKML <linux-kernel@vger.kernel.org>,
	Linux Memory Management List <linux-mm@kvack.org>,
	linux-fsdevel <linux-fsdevel@vger.kernel.org>
Subject: Re: [External] Re: [PATCH v2 05/19] mm/hugetlb: Introduce pgtable allocation/freeing helpers
Date: Fri, 6 Nov 2020 00:08:22 +0800
Message-ID: <CAMZfGtXwKJ3uCuNC3mxHQLNJqTcUzj7Gd2-JRuOWEjZ1C7Oh=A@mail.gmail.com> (raw)
In-Reply-To: <20201105132337.GA7552@linux>

On Thu, Nov 5, 2020 at 9:23 PM Oscar Salvador <osalvador@suse.de> wrote:
>
> On Mon, Oct 26, 2020 at 10:51:00PM +0800, Muchun Song wrote:
> > +#ifdef CONFIG_HUGETLB_PAGE_FREE_VMEMMAP
> > +#define VMEMMAP_HPAGE_SHIFT                  PMD_SHIFT
> > +#define arch_vmemmap_support_huge_mapping()  boot_cpu_has(X86_FEATURE_PSE)
>
> I do not think you need this.
> We already have hugepages_supported().

Maybe some architectures support hugepage, but the vmemmap do not
use the hugepage map. In  this case, we need it. But I am not sure if it
exists in the real world. At least, x86 can reuse hugepages_supported.

>
> > +#ifdef CONFIG_HUGETLB_PAGE_FREE_VMEMMAP
> > +#ifndef arch_vmemmap_support_huge_mapping
> > +static inline bool arch_vmemmap_support_huge_mapping(void)
> > +{
> > +     return false;
> > +}
>
> Same as above
>
> >  static inline unsigned int nr_free_vmemmap(struct hstate *h)
> >  {
> >       return h->nr_free_vmemmap_pages;
> >  }
> >
> > +static inline unsigned int nr_vmemmap(struct hstate *h)
> > +{
> > +     return nr_free_vmemmap(h) + RESERVE_VMEMMAP_NR;
> > +}
> > +
> > +static inline unsigned long nr_vmemmap_size(struct hstate *h)
> > +{
> > +     return (unsigned long)nr_vmemmap(h) << PAGE_SHIFT;
> > +}
> > +
> > +static inline unsigned int nr_pgtable(struct hstate *h)
> > +{
> > +     unsigned long vmemmap_size = nr_vmemmap_size(h);
> > +
> > +     if (!arch_vmemmap_support_huge_mapping())
> > +             return 0;
> > +
> > +     /*
> > +      * No need pre-allocate page tabels when there is no vmemmap pages
> > +      * to free.
> > +      */
> > +     if (!nr_free_vmemmap(h))
> > +             return 0;
> > +
> > +     return ALIGN(vmemmap_size, VMEMMAP_HPAGE_SIZE) >> VMEMMAP_HPAGE_SHIFT;
> > +}
>
> IMHO, Mike's naming suggestion fit much better.

I will do that.

>
> > +static void vmemmap_pgtable_deposit(struct page *page, pte_t *pte_p)
> > +{
> > +     pgtable_t pgtable = virt_to_page(pte_p);
> > +
> > +     /* FIFO */
> > +     if (!page_huge_pte(page))
> > +             INIT_LIST_HEAD(&pgtable->lru);
> > +     else
> > +             list_add(&pgtable->lru, &page_huge_pte(page)->lru);
> > +     page_huge_pte(page) = pgtable;
> > +}
>
> I think it would make more sense if this took a pgtable argument
> instead of a pte_t *.

Will do. Thanks for your suggestions.

>
> > +static pte_t *vmemmap_pgtable_withdraw(struct page *page)
> > +{
> > +     pgtable_t pgtable;
> > +
> > +     /* FIFO */
> > +     pgtable = page_huge_pte(page);
> > +     if (unlikely(!pgtable))
> > +             return NULL;
>
> AFAICS, above check only needs to be run once.
> It think we can move it to vmemmap_pgtable_free, can't we?

Yeah, we can. Thanks.

>
> > +     page_huge_pte(page) = list_first_entry_or_null(&pgtable->lru,
> > +                                                    struct page, lru);
> > +     if (page_huge_pte(page))
> > +             list_del(&pgtable->lru);
> > +     return page_to_virt(pgtable);
> > +}
>
> At the risk of adding more code, I think it would be nice to return a
> pagetable_t?
> So it is more coherent with the name of the function and with what
> we are doing.

Yeah.

>
> It is a pity we cannot converge these and pgtable_trans_huge_*.
> They share some code but it is different enough.
>
> > +static int vmemmap_pgtable_prealloc(struct hstate *h, struct page *page)
> > +{
> > +     int i;
> > +     pte_t *pte_p;
> > +     unsigned int nr = nr_pgtable(h);
> > +
> > +     if (!nr)
> > +             return 0;
> > +
> > +     vmemmap_pgtable_init(page);
>
> Maybe just open code this one?

Sorry. I don't quite understand what it means. Could you explain?


>
> > +static inline void vmemmap_pgtable_free(struct hstate *h, struct page *page)
> > +{
> > +     pte_t *pte_p;
> > +
> > +     if (!nr_pgtable(h))
> > +             return;
> > +
> > +     while ((pte_p = vmemmap_pgtable_withdraw(page)))
> > +             pte_free_kernel(&init_mm, pte_p);
>
> As mentioned above, move the pgtable_t check from vmemmap_pgtable_withdraw
> in here.

OK.

>
>
> >  static void prep_new_huge_page(struct hstate *h, struct page *page, int nid)
> >  {
> > +     /* Must be called before the initialization of @page->lru */
> > +     vmemmap_pgtable_free(h, page);
> > +
> >       INIT_LIST_HEAD(&page->lru);
> >       set_compound_page_dtor(page, HUGETLB_PAGE_DTOR);
> >       set_hugetlb_cgroup(page, NULL);
> > @@ -1783,6 +1892,14 @@ static struct page *alloc_fresh_huge_page(struct hstate *h,
> >       if (!page)
> >               return NULL;
> >
> > +     if (vmemmap_pgtable_prealloc(h, page)) {
> > +             if (hstate_is_gigantic(h))
> > +                     free_gigantic_page(page, huge_page_order(h));
> > +             else
> > +                     put_page(page);
> > +             return NULL;
> > +     }
> > +
>
> I must confess I am bit puzzled.
>
> IIUC, in this patch we are just adding the helpers to create/tear the page
> tables.
> I do not think we actually need to call vmemmap_pgtable_prealloc/vmemmap_pgtable_free, do we?
> In the end, we are just allocating pages for pagetables and then free them shortly.
>
> I think it would make more sense to add the calls when they need to be?

OK,  will do. Thanks very much.

>
>
> --
> Oscar Salvador
> SUSE L3



--
Yours,
Muchun

  reply index

Thread overview: 43+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-10-26 14:50 [PATCH v2 00/19] Free some vmemmap pages of hugetlb page Muchun Song
2020-10-26 14:50 ` [PATCH v2 01/19] mm/memory_hotplug: Move bootmem info registration API to bootmem_info.c Muchun Song
2020-10-26 14:50 ` [PATCH v2 02/19] mm/memory_hotplug: Move {get,put}_page_bootmem() " Muchun Song
2020-10-26 14:50 ` [PATCH v2 03/19] mm/hugetlb: Introduce a new config HUGETLB_PAGE_FREE_VMEMMAP Muchun Song
2020-10-29 10:29   ` Oscar Salvador
2020-10-29 13:34     ` [External] " Muchun Song
2020-10-26 14:50 ` [PATCH v2 04/19] mm/hugetlb: Introduce nr_free_vmemmap_pages in the struct hstate Muchun Song
2020-10-27 22:03   ` Mike Kravetz
2020-10-29 13:26   ` Oscar Salvador
2020-10-29 13:41     ` [External] " Muchun Song
2020-10-26 14:51 ` [PATCH v2 05/19] mm/hugetlb: Introduce pgtable allocation/freeing helpers Muchun Song
2020-10-28  0:32   ` Mike Kravetz
2020-10-28  7:26     ` [External] " Muchun Song
2020-10-28 23:42       ` Mike Kravetz
2020-11-05 13:23   ` Oscar Salvador
2020-11-05 16:08     ` Muchun Song [this message]
2020-11-06  9:46       ` [External] " Oscar Salvador
2020-11-06 16:43         ` Muchun Song
2020-10-26 14:51 ` [PATCH v2 06/19] mm/bootmem_info: Introduce {free,prepare}_vmemmap_page() Muchun Song
2020-10-26 14:51 ` [PATCH v2 07/19] mm/hugetlb: Free the vmemmap pages associated with each hugetlb page Muchun Song
2020-10-26 16:01   ` Matthew Wilcox
2020-10-27  2:58     ` [External] " Muchun Song
2020-10-28 23:42   ` Mike Kravetz
2020-10-29  6:13     ` [External] " Muchun Song
2020-10-29 21:59       ` Mike Kravetz
2020-10-30  2:58         ` Muchun Song
2020-10-26 14:51 ` [PATCH v2 08/19] mm/hugetlb: Defer freeing of hugetlb pages Muchun Song
2020-10-26 14:51 ` [PATCH v2 09/19] mm/hugetlb: Allocate the vmemmap pages associated with each hugetlb page Muchun Song
2020-10-26 14:51 ` [PATCH v2 10/19] mm/hugetlb: Introduce remap_huge_page_pmd_vmemmap helper Muchun Song
2020-10-26 14:51 ` [PATCH v2 11/19] mm/hugetlb: Use PG_slab to indicate split pmd Muchun Song
2020-10-26 14:51 ` [PATCH v2 12/19] mm/hugetlb: Support freeing vmemmap pages of gigantic page Muchun Song
2020-10-26 14:51 ` [PATCH v2 13/19] mm/hugetlb: Add a BUILD_BUG_ON to check if struct page size is a power of two Muchun Song
2020-10-26 14:51 ` [PATCH v2 14/19] mm/hugetlb: Clear PageHWPoison on the non-error memory page Muchun Song
2020-10-26 14:51 ` [PATCH v2 15/19] mm/hugetlb: Flush work when dissolving hugetlb page Muchun Song
2020-10-26 14:51 ` [PATCH v2 16/19] mm/hugetlb: Add a kernel parameter hugetlb_free_vmemmap Muchun Song
2020-10-26 14:51 ` [PATCH v2 17/19] mm/hugetlb: Merge pte to huge pmd only for gigantic page Muchun Song
2020-10-26 14:51 ` [PATCH v2 18/19] mm/hugetlb: Gather discrete indexes of tail page Muchun Song
2020-10-26 14:51 ` [PATCH v2 19/19] mm/hugetlb: Add BUILD_BUG_ON to catch invalid usage of tail struct page Muchun Song
2020-10-26 15:53 ` [PATCH v2 00/19] Free some vmemmap pages of hugetlb page Matthew Wilcox
2020-10-27  2:54   ` [External] " Muchun Song
2020-10-30  9:14 ` Michal Hocko
2020-10-30 10:24   ` [External] " Muchun Song
2020-10-30 15:19     ` Michal Hocko

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to='CAMZfGtXwKJ3uCuNC3mxHQLNJqTcUzj7Gd2-JRuOWEjZ1C7Oh=A@mail.gmail.com' \
    --to=songmuchun@bytedance.com \
    --cc=akpm@linux-foundation.org \
    --cc=almasrymina@google.com \
    --cc=anshuman.khandual@arm.com \
    --cc=bp@alien8.de \
    --cc=corbet@lwn.net \
    --cc=dave.hansen@linux.intel.com \
    --cc=duanxiongchun@bytedance.com \
    --cc=hpa@zytor.com \
    --cc=jroedel@suse.de \
    --cc=linux-doc@vger.kernel.org \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=luto@kernel.org \
    --cc=mchehab+huawei@kernel.org \
    --cc=mike.kravetz@oracle.com \
    --cc=mingo@redhat.com \
    --cc=oneukum@suse.com \
    --cc=osalvador@suse.de \
    --cc=paulmck@kernel.org \
    --cc=pawan.kumar.gupta@linux.intel.com \
    --cc=peterz@infradead.org \
    --cc=rdunlap@infradead.org \
    --cc=rientjes@google.com \
    --cc=tglx@linutronix.de \
    --cc=viro@zeniv.linux.org.uk \
    --cc=willy@infradead.org \
    --cc=x86@kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

Linux-Fsdevel Archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/linux-fsdevel/0 linux-fsdevel/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 linux-fsdevel linux-fsdevel/ https://lore.kernel.org/linux-fsdevel \
		linux-fsdevel@vger.kernel.org
	public-inbox-index linux-fsdevel

Example config snippet for mirrors

Newsgroup available over NNTP:
	nntp://nntp.lore.kernel.org/org.kernel.vger.linux-fsdevel


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git