From: Ryan Roberts <ryan.roberts@arm.com>
To: David Hildenbrand <david@redhat.com>,
Catalin Marinas <catalin.marinas@arm.com>,
Will Deacon <will@kernel.org>, Ard Biesheuvel <ardb@kernel.org>,
Marc Zyngier <maz@kernel.org>, James Morse <james.morse@arm.com>,
Andrey Ryabinin <ryabinin.a.a@gmail.com>,
Andrew Morton <akpm@linux-foundation.org>,
Matthew Wilcox <willy@infradead.org>,
Mark Rutland <mark.rutland@arm.com>,
Kefeng Wang <wangkefeng.wang@huawei.com>,
John Hubbard <jhubbard@nvidia.com>, Zi Yan <ziy@nvidia.com>,
Barry Song <21cnbao@gmail.com>,
Alistair Popple <apopple@nvidia.com>,
Yang Shi <shy828301@gmail.com>,
Nicholas Piggin <npiggin@gmail.com>,
Christophe Leroy <christophe.leroy@csgroup.eu>,
"Aneesh Kumar K.V" <aneesh.kumar@kernel.org>,
"Naveen N. Rao" <naveen.n.rao@linux.ibm.com>,
Thomas Gleixner <tglx@linutronix.de>,
Ingo Molnar <mingo@redhat.com>, Borislav Petkov <bp@alien8.de>,
Dave Hansen <dave.hansen@linux.intel.com>,
"H. Peter Anvin" <hpa@zytor.com>
Cc: linux-arm-kernel@lists.infradead.org, x86@kernel.org,
linuxppc-dev@lists.ozlabs.org, linux-mm@kvack.org,
linux-kernel@vger.kernel.org
Subject: Re: [PATCH v5 22/25] mm: Add pte_batch_hint() to reduce scanning in folio_pte_batch()
Date: Mon, 12 Feb 2024 15:47:44 +0000 [thread overview]
Message-ID: <82c59a7f-328e-4521-8855-ccacc3dc4ce5@arm.com> (raw)
In-Reply-To: <6d452a1a-1edc-4e97-8b39-99dc48315bb8@redhat.com>
On 12/02/2024 13:43, David Hildenbrand wrote:
> On 02.02.24 09:07, Ryan Roberts wrote:
>> Some architectures (e.g. arm64) can tell from looking at a pte, if some
>> follow-on ptes also map contiguous physical memory with the same pgprot.
>> (for arm64, these are contpte mappings).
>>
>> Take advantage of this knowledge to optimize folio_pte_batch() so that
>> it can skip these ptes when scanning to create a batch. By default, if
>> an arch does not opt-in, folio_pte_batch() returns a compile-time 1, so
>> the changes are optimized out and the behaviour is as before.
>>
>> arm64 will opt-in to providing this hint in the next patch, which will
>> greatly reduce the cost of ptep_get() when scanning a range of contptes.
>>
>> Tested-by: John Hubbard <jhubbard@nvidia.com>
>> Signed-off-by: Ryan Roberts <ryan.roberts@arm.com>
>> ---
>> include/linux/pgtable.h | 18 ++++++++++++++++++
>> mm/memory.c | 20 +++++++++++++-------
>> 2 files changed, 31 insertions(+), 7 deletions(-)
>>
>> diff --git a/include/linux/pgtable.h b/include/linux/pgtable.h
>> index 50f32cccbd92..cba31f177d27 100644
>> --- a/include/linux/pgtable.h
>> +++ b/include/linux/pgtable.h
>> @@ -212,6 +212,24 @@ static inline int pmd_dirty(pmd_t pmd)
>> #define arch_flush_lazy_mmu_mode() do {} while (0)
>> #endif
>> +#ifndef pte_batch_hint
>> +/**
>> + * pte_batch_hint - Number of pages that can be added to batch without scanning.
>> + * @ptep: Page table pointer for the entry.
>> + * @pte: Page table entry.
>> + *
>> + * Some architectures know that a set of contiguous ptes all map the same
>> + * contiguous memory with the same permissions. In this case, it can provide a
>> + * hint to aid pte batching without the core code needing to scan every pte.
>
> I think we might want to document here the expectation regarding
> dirty/accessed bits. folio_pte_batch() will ignore dirty bits only with
> FPB_IGNORE_DIRTY. But especially for arm64, it makes sense to ignore them
> always when batching, because the dirty bit may target any pte part of the
> cont-pte group either way.
>
> Maybe something like:
>
> "
> An architecture implementation may only ignore the PTE accessed and dirty bits.
> Further, it may only ignore the dirty bit if that bit is already not
> maintained with precision per PTE inside the hinted batch, and ptep_get()
> would already have to collect it from various PTEs.
> "
I'm proposing to simplify this to:
"
An architecture implementation may ignore the PTE accessed state. Further, the
dirty state must apply atomically to all the PTEs described by the hint.
"
Which I think more accurately describes the requirement. Shout if you disagree.
>
> I think there are some more details to it, but I'm hoping something along
> the lines above is sufficient.
>
>
>> +
>> #ifndef pte_advance_pfn
>> static inline pte_t pte_advance_pfn(pte_t pte, unsigned long nr)
>> {
>> diff --git a/mm/memory.c b/mm/memory.c
>> index 65fbe4f886c1..902665b27702 100644
>> --- a/mm/memory.c
>> +++ b/mm/memory.c
>> @@ -988,16 +988,21 @@ static inline int folio_pte_batch(struct folio *folio,
>> unsigned long addr,
>> {
>> unsigned long folio_end_pfn = folio_pfn(folio) + folio_nr_pages(folio);
>> const pte_t *end_ptep = start_ptep + max_nr;
>> - pte_t expected_pte = __pte_batch_clear_ignored(pte_advance_pfn(pte, 1),
>> flags);
>> - pte_t *ptep = start_ptep + 1;
>> + pte_t expected_pte = __pte_batch_clear_ignored(pte, flags);
>> + pte_t *ptep = start_ptep;
>> bool writable;
>> + int nr;
>> if (any_writable)
>> *any_writable = false;
>> VM_WARN_ON_FOLIO(!pte_present(pte), folio);
>> - while (ptep != end_ptep) {
>> + nr = pte_batch_hint(ptep, pte);
>> + expected_pte = pte_advance_pfn(expected_pte, nr);
>> + ptep += nr;
>> +
>
> *Maybe* it's easier to get when initializing expected_pte+ptep only once.
>
> Like:
>
> [...]
> pte_t expected_pte, *ptep;
> [...]
>
> nr = pte_batch_hint(start_ptep, pte);
> expected_pte = __pte_batch_clear_ignored(pte_advance_pfn(pte, nr), flags);
> ptep = start_ptep + nr;
>
>> + while (ptep < end_ptep) {
>> pte = ptep_get(ptep);
>> if (any_writable)
>> writable = !!pte_write(pte);
>> @@ -1011,17 +1016,18 @@ static inline int folio_pte_batch(struct folio *folio,
>> unsigned long addr,
>> * corner cases the next PFN might fall into a different
>> * folio.
>> */
>> - if (pte_pfn(pte) == folio_end_pfn)
>> + if (pte_pfn(pte) >= folio_end_pfn)
>> break;
>> if (any_writable)
>> *any_writable |= writable;
>> - expected_pte = pte_advance_pfn(expected_pte, 1);
>> - ptep++;
>> + nr = pte_batch_hint(ptep, pte);
>> + expected_pte = pte_advance_pfn(expected_pte, nr);
>> + ptep += nr;
>> }
>> - return ptep - start_ptep;
>> + return min(ptep - start_ptep, max_nr);
>> }
>
> Acked-by: David Hildenbrand <david@redhat.com>
>
next prev parent reply other threads:[~2024-02-12 15:47 UTC|newest]
Thread overview: 80+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-02-02 8:07 [PATCH v5 00/25] Transparent Contiguous PTEs for User Mappings Ryan Roberts
2024-02-02 8:07 ` [PATCH v5 01/25] mm: Clarify the spec for set_ptes() Ryan Roberts
2024-02-12 12:03 ` David Hildenbrand
2024-02-02 8:07 ` [PATCH v5 02/25] mm: thp: Batch-collapse PMD with set_ptes() Ryan Roberts
2024-02-02 8:07 ` [PATCH v5 03/25] mm: Make pte_next_pfn() a wrapper around pte_advance_pfn() Ryan Roberts
2024-02-12 12:14 ` David Hildenbrand
2024-02-12 14:10 ` Ryan Roberts
2024-02-12 14:29 ` David Hildenbrand
2024-02-12 21:34 ` Ryan Roberts
2024-02-13 9:54 ` David Hildenbrand
2024-02-02 8:07 ` [PATCH v5 04/25] arm/mm: Convert pte_next_pfn() to pte_advance_pfn() Ryan Roberts
2024-02-02 8:07 ` [PATCH v5 05/25] arm64/mm: " Ryan Roberts
2024-02-02 8:07 ` [PATCH v5 06/25] powerpc/mm: " Ryan Roberts
2024-02-02 8:07 ` [PATCH v5 07/25] x86/mm: " Ryan Roberts
2024-02-02 8:07 ` [PATCH v5 08/25] mm: Remove pte_next_pfn() and replace with pte_advance_pfn() Ryan Roberts
2024-02-02 8:07 ` [PATCH v5 09/25] arm64/mm: set_pte(): New layer to manage contig bit Ryan Roberts
2024-02-02 8:07 ` [PATCH v5 10/25] arm64/mm: set_ptes()/set_pte_at(): " Ryan Roberts
2024-02-02 8:07 ` [PATCH v5 11/25] arm64/mm: pte_clear(): " Ryan Roberts
2024-02-02 8:07 ` [PATCH v5 12/25] arm64/mm: ptep_get_and_clear(): " Ryan Roberts
2024-02-02 8:07 ` [PATCH v5 13/25] arm64/mm: ptep_test_and_clear_young(): " Ryan Roberts
2024-02-02 8:07 ` [PATCH v5 14/25] arm64/mm: ptep_clear_flush_young(): " Ryan Roberts
2024-02-02 8:07 ` [PATCH v5 15/25] arm64/mm: ptep_set_wrprotect(): " Ryan Roberts
2024-02-02 8:07 ` [PATCH v5 16/25] arm64/mm: ptep_set_access_flags(): " Ryan Roberts
2024-02-02 8:07 ` [PATCH v5 17/25] arm64/mm: ptep_get(): " Ryan Roberts
2024-02-02 8:07 ` [PATCH v5 18/25] arm64/mm: Split __flush_tlb_range() to elide trailing DSB Ryan Roberts
2024-02-12 12:44 ` David Hildenbrand
2024-02-12 13:05 ` Ryan Roberts
2024-02-12 13:15 ` David Hildenbrand
2024-02-12 13:27 ` Ryan Roberts
2024-02-02 8:07 ` [PATCH v5 19/25] arm64/mm: Wire up PTE_CONT for user mappings Ryan Roberts
2024-02-12 12:00 ` Mark Rutland
2024-02-12 12:59 ` Ryan Roberts
2024-02-12 13:54 ` David Hildenbrand
2024-02-12 14:45 ` Ryan Roberts
2024-02-12 15:26 ` David Hildenbrand
2024-02-12 15:34 ` Ryan Roberts
2024-02-12 16:24 ` David Hildenbrand
2024-02-13 15:29 ` Ryan Roberts
2024-02-12 15:30 ` Ryan Roberts
2024-02-12 20:38 ` Ryan Roberts
2024-02-13 10:01 ` David Hildenbrand
2024-02-13 12:06 ` Ryan Roberts
2024-02-13 12:19 ` David Hildenbrand
2024-02-13 13:06 ` Ryan Roberts
2024-02-13 13:13 ` David Hildenbrand
2024-02-13 13:20 ` Ryan Roberts
2024-02-13 13:22 ` David Hildenbrand
2024-02-13 13:24 ` Ryan Roberts
2024-02-13 13:33 ` Ard Biesheuvel
2024-02-13 13:45 ` David Hildenbrand
2024-02-13 14:02 ` Ryan Roberts
2024-02-13 14:05 ` David Hildenbrand
2024-02-13 14:08 ` Ard Biesheuvel
2024-02-13 14:21 ` Ryan Roberts
2024-02-13 12:02 ` Mark Rutland
2024-02-13 13:03 ` Ryan Roberts
2024-02-02 8:07 ` [PATCH v5 20/25] arm64/mm: Implement new wrprotect_ptes() batch API Ryan Roberts
2024-02-13 16:31 ` Mark Rutland
2024-02-13 16:36 ` Ryan Roberts
2024-02-02 8:07 ` [PATCH v5 21/25] arm64/mm: Implement new [get_and_]clear_full_ptes() batch APIs Ryan Roberts
2024-02-13 16:43 ` Mark Rutland
2024-02-13 16:48 ` Ryan Roberts
2024-02-13 16:53 ` Mark Rutland
2024-02-02 8:07 ` [PATCH v5 22/25] mm: Add pte_batch_hint() to reduce scanning in folio_pte_batch() Ryan Roberts
2024-02-12 13:43 ` David Hildenbrand
2024-02-12 15:00 ` Ryan Roberts
2024-02-12 15:47 ` Ryan Roberts [this message]
2024-02-12 16:27 ` David Hildenbrand
2024-02-02 8:07 ` [PATCH v5 23/25] arm64/mm: Implement pte_batch_hint() Ryan Roberts
2024-02-12 13:46 ` David Hildenbrand
2024-02-13 16:54 ` Mark Rutland
2024-02-02 8:07 ` [PATCH v5 24/25] arm64/mm: __always_inline to improve fork() perf Ryan Roberts
2024-02-13 16:55 ` Mark Rutland
2024-02-02 8:07 ` [PATCH v5 25/25] arm64/mm: Automatically fold contpte mappings Ryan Roberts
2024-02-13 17:44 ` Mark Rutland
2024-02-13 18:05 ` Ryan Roberts
2024-02-08 17:34 ` [PATCH v5 00/25] Transparent Contiguous PTEs for User Mappings Mark Rutland
2024-02-09 8:54 ` Ryan Roberts
2024-02-09 22:16 ` David Hildenbrand
2024-02-09 23:52 ` Ryan Roberts
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=82c59a7f-328e-4521-8855-ccacc3dc4ce5@arm.com \
--to=ryan.roberts@arm.com \
--cc=21cnbao@gmail.com \
--cc=akpm@linux-foundation.org \
--cc=aneesh.kumar@kernel.org \
--cc=apopple@nvidia.com \
--cc=ardb@kernel.org \
--cc=bp@alien8.de \
--cc=catalin.marinas@arm.com \
--cc=christophe.leroy@csgroup.eu \
--cc=dave.hansen@linux.intel.com \
--cc=david@redhat.com \
--cc=hpa@zytor.com \
--cc=james.morse@arm.com \
--cc=jhubbard@nvidia.com \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=linuxppc-dev@lists.ozlabs.org \
--cc=mark.rutland@arm.com \
--cc=maz@kernel.org \
--cc=mingo@redhat.com \
--cc=naveen.n.rao@linux.ibm.com \
--cc=npiggin@gmail.com \
--cc=ryabinin.a.a@gmail.com \
--cc=shy828301@gmail.com \
--cc=tglx@linutronix.de \
--cc=wangkefeng.wang@huawei.com \
--cc=will@kernel.org \
--cc=willy@infradead.org \
--cc=x86@kernel.org \
--cc=ziy@nvidia.com \
/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;
as well as URLs for NNTP newsgroup(s).