mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Ryan Roberts <ryan.roberts@arm.com>
To: Christophe Leroy <christophe.leroy@csgroup.eu>,
	Peter Xu <peterx@redhat.com>
Cc: Matthew Wilcox <willy@infradead.org>,
	Christoph Hellwig <hch@infradead.org>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	"linux-mm@kvack.org" <linux-mm@kvack.org>,
	Andrea Arcangeli <aarcange@redhat.com>,
	James Houghton <jthoughton@google.com>,
	Lorenzo Stoakes <lstoakes@gmail.com>,
	David Hildenbrand <david@redhat.com>,
	Vlastimil Babka <vbabka@suse.cz>,
	John Hubbard <jhubbard@nvidia.com>,
	Yang Shi <shy828301@gmail.com>, Rik van Riel <riel@surriel.com>,
	Hugh Dickins <hughd@google.com>, Jason Gunthorpe <jgg@nvidia.com>,
	Axel Rasmussen <axelrasmussen@google.com>,
	"Kirill A . Shutemov" <kirill@shutemov.name>,
	Andrew Morton <akpm@linux-foundation.org>,
	"linuxppc-dev@lists.ozlabs.org" <linuxppc-dev@lists.ozlabs.org>,
	Mike Rapoport <rppt@kernel.org>,
	Mike Kravetz <mike.kravetz@oracle.com>
Subject: Re: [PATCH RFC 06/12] mm/gup: Drop folio_fast_pin_allowed() in hugepd processing
Date: Mon, 4 Dec 2023 12:02:38 +0000	[thread overview]
Message-ID: <11601011-07af-4662-8ee4-f98785f75e29@arm.com> (raw)
In-Reply-To: <8c7fe945-ee34-4eb6-b466-5707660c7723@csgroup.eu>

On 04/12/2023 11:57, Christophe Leroy wrote:
> 
> 
> Le 04/12/2023 à 12:46, Ryan Roberts a écrit :
>> On 04/12/2023 11:25, Christophe Leroy wrote:
>>>
>>>
>>> Le 04/12/2023 à 12:11, Ryan Roberts a écrit :
>>>> On 03/12/2023 13:33, Christophe Leroy wrote:
>>>>>
>>>>>
>>>>> Le 30/11/2023 à 22:30, Peter Xu a écrit :
>>>>>> On Fri, Nov 24, 2023 at 11:07:51AM -0500, Peter Xu wrote:
>>>>>>> On Fri, Nov 24, 2023 at 09:06:01AM +0000, Ryan Roberts wrote:
>>>>>>>> I don't have any micro-benchmarks for GUP though, if that's your question. Is
>>>>>>>> there an easy-to-use test I can run to get some numbers? I'd be happy to try it out.
>>>>>>>
>>>>>>> Thanks Ryan.  Then nothing is needed to be tested if gup is not yet touched
>>>>>>> from your side, afaict.  I'll see whether I can provide some rough numbers
>>>>>>> instead in the next post (I'll probably only be able to test it in a VM,
>>>>>>> though, but hopefully that should still reflect mostly the truth).
>>>>>>
>>>>>> An update: I finished a round of 64K cont_pte test, in the slow gup micro
>>>>>> benchmark I see ~15% perf degrade with this patchset applied on a VM on top
>>>>>> of Apple M1.
>>>>>>
>>>>>> Frankly that's even less than I expected, considering not only how slow gup
>>>>>> THP used to be, but also on the fact that that's a tight loop over slow
>>>>>> gup, which in normal cases shouldn't happen: "present" ptes normally goes
>>>>>> to fast-gup, while !present goes into a fault following it.  I assume
>>>>>> that's why nobody cared slow gup for THP before.  I think adding cont_pte
>>>>>> support shouldn't be very hard, but that will include making cont_pte idea
>>>>>> global just for arm64 and riscv Svnapot.
>>>>>
>>>>> Is there any documentation on what cont_pte is ? I have always wondered
>>>>> if it could also fit powerpc 8xx need ?
>>>>
>>>> pte_cont() (and pte_mkcont() and pte_mknoncont()) test and manipulte the
>>>> "contiguous bit" in the arm64 PTE entries. Those helpers are arm64-specific
>>>> (AFAIK). The contiguous bit is a hint to the HW to tell it that a block of PTEs
>>>> are mapping a physically contiguous and naturally aligned piece of memory. The
>>>> HW can use this to coalesce entries in the TLB. When using 4K base pages, the
>>>> contpte size is 64K (16 PTEs). For 16K base pages, its 2M (128 PTEs) and for 64K
>>>> base pages, its 2M (32 PTEs).
>>>>
>>>>>
>>>>> On powerpc, for 16k pages, we have to define 4 consecutive PTEs. All 4
>>>>> PTE are flagged with the SPS bit telling it's a 16k pages, but for TLB
>>>>> misses the HW needs one entrie for each 4k fragment.
>>>>
>>>>   From that description, it sounds like the SPS bit might be similar to arm64
>>>> contiguous bit? Although sounds like you are currently using it in a slightly
>>>> different way - telling kernel that the base page is 16K but mapping each 16K
>>>> page with 4x 4K entries (plus the SPS bit set)?
>>>
>>> Yes it's both.
>>>
>>> When the base page is 16k, there are 4x 4k entries (with SPS bit set) in
>>> the page table, and pte_t is a table of 4 'unsigned long'
>>>
>>> When the base page is 4k, there is a 16k hugepage size, which is the
>>> same 4x 4k entries with SPS bit set.
>>>
>>> So it looks similar to the contiguous bit.
>>>
>>>
>>> And by extension, the same principle is used for 512k hugepages, the bit
>>> _PAGE_HUGE is copied by the TLB miss handler into the lower bit of PS,
>>> PS being as follows:
>>> - 00 Small (4 Kbyte or 16 Kbyte)
>>> - 01 512 Kbyte
>>> - 10 Reserved
>>> - 11 8 Mbyte
>>>
>>> So as PMD size is 4M, 512k pages are 128 identical consecutive entries
>>> in the page table.
>>>
>>> I which I could have THP with 16k or 512k pages.
>>
>> Then you have come to the right place! :)
>>
>> https://lore.kernel.org/linux-mm/20231204102027.57185-1-ryan.roberts@arm.com/
>>
> 
> That looks great. That series only modifies core mm/ .
> No changes needed in arch ? Will it work on powerpc without any 
> change/additions to arch code ?

Yes there are also changes needed in arch; I have a separate series for arm64,
which transparently manages the contiguous bit when it sees appropriate PTEs:

https://lore.kernel.org/linux-arm-kernel/20231204105440.61448-1-ryan.roberts@arm.com/

> 
> Well, I'll try it soon.
> 
> Christophe


  reply	other threads:[~2023-12-04 12:03 UTC|newest]

Thread overview: 57+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-11-16  1:28 [PATCH RFC 00/12] mm/gup: Unify hugetlb, part 2 Peter Xu
2023-11-16  1:28 ` [PATCH RFC 01/12] mm/hugetlb: Export hugetlbfs_pagecache_present() Peter Xu
2023-11-23  7:23   ` Christoph Hellwig
2023-11-23 16:05     ` Peter Xu
2023-11-16  1:28 ` [PATCH RFC 02/12] mm: Provide generic pmd_thp_or_huge() Peter Xu
2023-11-16  1:28 ` [PATCH RFC 03/12] mm: Export HPAGE_PXD_* macros even if !THP Peter Xu
2023-11-23  7:23   ` Christoph Hellwig
2023-11-23  9:53     ` Mike Rapoport
2023-11-23 15:27       ` Peter Xu
2023-11-16  1:29 ` [PATCH RFC 04/12] mm: Introduce vma_pgtable_walk_{begin|end}() Peter Xu
2023-11-23  7:24   ` Christoph Hellwig
2023-11-23 16:11     ` Peter Xu
     [not found]   ` <874jhb94u2.fsf@kernel.org>
2023-11-24 15:34     ` Peter Xu
2023-11-16  1:29 ` [PATCH RFC 05/12] mm/gup: Fix follow_devmap_p[mu]d() to return even if NULL Peter Xu
2023-11-23  7:25   ` Christoph Hellwig
2023-11-23 17:59     ` Peter Xu
2023-11-16  1:29 ` [PATCH RFC 06/12] mm/gup: Drop folio_fast_pin_allowed() in hugepd processing Peter Xu
2023-11-20  8:26   ` Christoph Hellwig
2023-11-21 15:59     ` Peter Xu
2023-11-22  8:00       ` Christoph Hellwig
2023-11-22 15:22         ` Peter Xu
2023-11-23  7:21           ` Christoph Hellwig
2023-11-23 16:10             ` Peter Xu
2023-11-23 18:22           ` Christophe Leroy
2023-11-23 19:37             ` Peter Xu
2023-11-24  7:03               ` Christophe Leroy
2023-11-24 18:16                 ` Peter Xu
2023-11-24  1:06           ` Michael Ellerman
2023-11-23 15:47         ` Matthew Wilcox
2023-11-23 17:22           ` Peter Xu
2023-11-23 19:11             ` Ryan Roberts
2023-11-23 19:46               ` Peter Xu
2023-11-24  9:06                 ` Ryan Roberts
2023-11-24 16:07                   ` Peter Xu
2023-11-30 21:30                     ` Peter Xu
2023-12-03 13:33                       ` Christophe Leroy
2023-12-04 11:11                         ` Ryan Roberts
2023-12-04 11:25                           ` Christophe Leroy
2023-12-04 11:46                             ` Ryan Roberts
2023-12-04 11:57                               ` Christophe Leroy
2023-12-04 12:02                                 ` Ryan Roberts [this message]
2023-12-04 16:48                           ` Peter Xu
2023-11-16  1:29 ` [PATCH RFC 07/12] mm/gup: Refactor record_subpages() to find 1st small page Peter Xu
2023-11-16 14:51   ` Matthew Wilcox
2023-11-16 19:40     ` Peter Xu
2023-11-16 19:41       ` Matthew Wilcox
2023-11-16  1:29 ` [PATCH RFC 08/12] mm/gup: Handle hugetlb for no_page_table() Peter Xu
2023-11-23  7:26   ` Christoph Hellwig
2023-11-16  1:29 ` [PATCH RFC 09/12] mm/gup: Handle huge pud for follow_pud_mask() Peter Xu
2023-11-23  7:28   ` Christoph Hellwig
2023-11-23 16:19     ` Peter Xu
2023-11-16  1:29 ` [PATCH RFC 10/12] mm/gup: Handle huge pmd for follow_pmd_mask() Peter Xu
2023-11-16  1:29 ` [PATCH RFC 11/12] mm/gup: Handle hugepd for follow_page() Peter Xu
2023-11-16  1:29 ` [PATCH RFC 12/12] mm/gup: Merge hugetlb into generic mm code Peter Xu
2023-11-23  7:29   ` Christoph Hellwig
2023-11-23 16:21     ` Peter Xu
2023-11-22 14:51 ` [PATCH RFC 00/12] mm/gup: Unify hugetlb, part 2 Jason Gunthorpe

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=11601011-07af-4662-8ee4-f98785f75e29@arm.com \
    --to=ryan.roberts@arm.com \
    --cc=aarcange@redhat.com \
    --cc=akpm@linux-foundation.org \
    --cc=axelrasmussen@google.com \
    --cc=christophe.leroy@csgroup.eu \
    --cc=david@redhat.com \
    --cc=hch@infradead.org \
    --cc=hughd@google.com \
    --cc=jgg@nvidia.com \
    --cc=jhubbard@nvidia.com \
    --cc=jthoughton@google.com \
    --cc=kirill@shutemov.name \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=linuxppc-dev@lists.ozlabs.org \
    --cc=lstoakes@gmail.com \
    --cc=mike.kravetz@oracle.com \
    --cc=peterx@redhat.com \
    --cc=riel@surriel.com \
    --cc=rppt@kernel.org \
    --cc=shy828301@gmail.com \
    --cc=vbabka@suse.cz \
    --cc=willy@infradead.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
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®