APPLIED: [SRU][N/J:linux-azure][PATCH 0/3] [Arm64] High reboot time for 96 core Linux VMs
John Cabaj
john.cabaj at canonical.com
Fri Jun 27 22:41:34 UTC 2025
On 6/23/25 10:22 AM, John Cabaj wrote:
> BugLink: https://bugs.launchpad.net/bugs/2114218
>
> SRU Justification
>
> [ Impact ]
>
> * Microsoft request to fix long boot times on arm64 VMs with 96 cores.
>
> [ Fix ]
>
> * Clean cherry-pick of the following patches from upstream:
> * 5c63db59c5f8: "arm64: mm: Don't remap pgtables per-cont(pte|pmd) block"
> * 1fcb7cea8a5f: "arm64: mm: Batch dsb and isb when populating pgtables"
> * 0e9df1c905d8: "arm64: mm: Don't remap pgtables for allocate vs populat"
>
> [ Test Plan ]
>
> * Compile tested
> * Microsoft tested
>
> [ Where problems could occur ]
>
> * Changes made to memory-management, which could have impacts across the memory subsystems.
> * Backport involved stripping out portions of 5-level paging, which was missing prerequisite
> commits. This has been validated, but if logic was not well-separated in the upstream commits,
> the changes for present paging might have some reliance on the 5-level paging changes.
>
> [ Other info ]
>
> * SF#00412644
>
> Ryan Roberts (3):
> arm64: mm: Don't remap pgtables per-cont(pte|pmd) block
> arm64: mm: Batch dsb and isb when populating pgtables
> arm64: mm: Don't remap pgtables for allocate vs populate
>
> arch/arm64/include/asm/pgtable.h | 7 ++-
> arch/arm64/mm/mmu.c | 92 ++++++++++++++++++--------------
> 2 files changed, 57 insertions(+), 42 deletions(-)
>
Applied to Noble and Jammy linxu-azure master-next branches.
Thanks,
John
More information about the kernel-team
mailing list