NACK: [PULL v2][Noble] backport arm64 THP improvements from 6.9
Roxana Nicolescu
roxana.nicolescu at canonical.com
Wed May 1 14:56:54 UTC 2024
On 03/04/2024 15:20, dann frazier wrote:
> BugLink: https://bugs.launchpad.net/bugs/2059316
>
> These are all clean cherry-picks from upstream, save one that required
> minor backporting due to an API change. This does touch generic code.
> I've regression tested as much as I've had time to given I just
> learned about this patchset last week. But with the performance
> improvements I've measured, and the GPU workload improvements reported
> upstream, I think this is something we need to seriously consider
> for our next LTS.
>
> I plan to monitor upstream for Fixes commits during the 6.9
> development cycle, in case regressions are found/fixed.
>
> Changes from v1:
> * Relabeled as SAUCE
>
> The following changes since commit 4427c45609f6faf3b5b15e9e9246caa87894c36f:
>
> UBUNTU: Ubuntu-6.8.0-20.20 (2024-03-18 11:08:14 +0100)
>
> are available in the Git repository at:
>
> git://git.launchpad.net/~dannf/ubuntu/+source/linux/+git/linux noble-mthp-sauce
>
> for you to fetch changes up to 9ddb0a4d82ff0cf89be70d20c26ebfab83b9d3b0:
>
> UBUNTU: SAUCE: arm64/mm: improve comment in contpte_ptep_get_lockless() (2024-04-03 06:55:48 -0600)
>
> ----------------------------------------------------------------
> David Hildenbrand (14):
> UBUNTU: SAUCE: arm/pgtable: define PFN_PTE_SHIFT
> UBUNTU: SAUCE: nios2/pgtable: define PFN_PTE_SHIFT
> UBUNTU: SAUCE: powerpc/pgtable: define PFN_PTE_SHIFT
> UBUNTU: SAUCE: riscv/pgtable: define PFN_PTE_SHIFT
> UBUNTU: SAUCE: s390/pgtable: define PFN_PTE_SHIFT
> UBUNTU: SAUCE: sparc/pgtable: define PFN_PTE_SHIFT
> UBUNTU: SAUCE: mm/pgtable: make pte_next_pfn() independent of set_ptes()
> UBUNTU: SAUCE: arm/mm: use pte_next_pfn() in set_ptes()
> UBUNTU: SAUCE: powerpc/mm: use pte_next_pfn() in set_ptes()
> UBUNTU: SAUCE: mm/memory: factor out copying the actual PTE in copy_present_pte()
> UBUNTU: SAUCE: mm/memory: pass PTE to copy_present_pte()
> UBUNTU: SAUCE: mm/memory: optimize fork() with PTE-mapped THP
> UBUNTU: SAUCE: mm/memory: ignore dirty/accessed/soft-dirty bits in folio_pte_batch()
> UBUNTU: SAUCE: mm/memory: ignore writable bit in folio_pte_batch()
>
> Ryan Roberts (21):
> UBUNTU: SAUCE: arm64/mm: make set_ptes() robust when OAs cross 48-bit boundary
> UBUNTU: SAUCE: mm: clarify the spec for set_ptes()
> UBUNTU: SAUCE: mm: thp: batch-collapse PMD with set_ptes()
> UBUNTU: SAUCE: mm: introduce pte_advance_pfn() and use for pte_next_pfn()
> UBUNTU: SAUCE: arm64/mm: convert pte_next_pfn() to pte_advance_pfn()
> UBUNTU: SAUCE: x86/mm: convert pte_next_pfn() to pte_advance_pfn()
> UBUNTU: SAUCE: mm: tidy up pte_next_pfn() definition
> UBUNTU: SAUCE: arm64/mm: convert READ_ONCE(*ptep) to ptep_get(ptep)
> UBUNTU: SAUCE: arm64/mm: convert set_pte_at() to set_ptes(..., 1)
> UBUNTU: SAUCE: arm64/mm: convert ptep_clear() to ptep_get_and_clear()
> UBUNTU: SAUCE: arm64/mm: new ptep layer to manage contig bit
> UBUNTU: SAUCE: arm64/mm: dplit __flush_tlb_range() to elide trailing DSB
> UBUNTU: SAUCE: arm64/mm: wire up PTE_CONT for user mappings
> UBUNTU: SAUCE: arm64/mm: implement new wrprotect_ptes() batch API
> UBUNTU: SAUCE: arm64/mm: implement new [get_and_]clear_full_ptes() batch APIs
> UBUNTU: SAUCE: mm: add pte_batch_hint() to reduce scanning in folio_pte_batch()
> UBUNTU: SAUCE: arm64/mm: implement pte_batch_hint()
> UBUNTU: SAUCE: arm64/mm: __always_inline to improve fork() perf
> UBUNTU: SAUCE: arm64/mm: automatically fold contpte mappings
> UBUNTU: SAUCE: arm64/mm: export contpte symbols only to GPL users
> UBUNTU: SAUCE: arm64/mm: improve comment in contpte_ptep_get_lockless()
>
> dann frazier (1):
> UBUNTU: [Config] arm64: ARM64_CONTPTE=y
>
> arch/arm/include/asm/pgtable.h | 2 +
> arch/arm/mm/mmu.c | 2 +-
> arch/arm64/Kconfig | 9 +
> arch/arm64/include/asm/pgtable.h | 431 +++++++++++++++++++++++++++++++-----
> arch/arm64/include/asm/tlbflush.h | 13 +-
> arch/arm64/kernel/efi.c | 4 +-
> arch/arm64/kernel/mte.c | 2 +-
> arch/arm64/kvm/guest.c | 2 +-
> arch/arm64/mm/Makefile | 1 +
> arch/arm64/mm/contpte.c | 408 ++++++++++++++++++++++++++++++++++
> arch/arm64/mm/fault.c | 12 +-
> arch/arm64/mm/fixmap.c | 4 +-
> arch/arm64/mm/hugetlbpage.c | 40 ++--
> arch/arm64/mm/kasan_init.c | 6 +-
> arch/arm64/mm/mmu.c | 16 +-
> arch/arm64/mm/pageattr.c | 6 +-
> arch/arm64/mm/trans_pgd.c | 6 +-
> arch/nios2/include/asm/pgtable.h | 2 +
> arch/powerpc/include/asm/pgtable.h | 2 +
> arch/powerpc/mm/pgtable.c | 5 +-
> arch/riscv/include/asm/pgtable.h | 2 +
> arch/s390/include/asm/pgtable.h | 2 +
> arch/sparc/include/asm/pgtable_64.h | 2 +
> arch/x86/include/asm/pgtable.h | 8 +-
> debian.master/config/annotations | 3 +
> include/linux/efi.h | 5 +
> include/linux/pgtable.h | 65 +++++-
> mm/huge_memory.c | 58 ++---
> mm/memory.c | 219 ++++++++++++++----
> 29 files changed, 1151 insertions(+), 186 deletions(-)
> create mode 100644 arch/arm64/mm/contpte.c
>
This was dropped from the noble:linux master-next branch. And it won't
be part of the upcoming stable releases.
Having these many sauce patches makes keeping up with upstream harder
that it already is.
In a few months a new noble:linux-hwe will be available and it will
contain these improvements.
Roxana
More information about the kernel-team
mailing list