tjh.dev / kernel
Linux kernel mirror (for testing) git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
kernel os linux
1
fork

Configure Feed

Select the types of activity you want to include in your feed.

Merge tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux

Pull arm64 fixes from Catalin Marinas:

- Do not make a clean PTE dirty in pte_mkwrite()

The Arm architecture, for backwards compatibility reasons (ARMv8.0
before in-hardware dirty bit management - DBM), uses the PTE_RDONLY
bit to mean !dirty while the PTE_WRITE bit means DBM enabled. The
arm64 pte_mkwrite() simply clears the PTE_RDONLY bit and this
inadvertently makes the PTE pte_hw_dirty(). Most places making a PTE
writable also invoke pte_mkdirty() but do_swap_page() does not and we
end up with dirty, freshly swapped in, writeable pages.

- Do not warn if the destination page is already MTE-tagged in
copy_highpage()

In the majority of the cases, a destination page copied into is
freshly allocated without the PG_mte_tagged flag set. However, the
folio migration may be restarted if __folio_migrate_mapping() failed,
triggering the benign WARN_ON_ONCE().

* tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux:
arm64: mte: Do not warn if the page is already tagged in copy_highpage()
arm64, mm: avoid always making PTE dirty in pte_mkwrite()

Linus Torvalds 266ee584 ab431bc3

+10 -4
+2 -1
arch/arm64/include/asm/pgtable.h
··· 293 293 static inline pte_t pte_mkwrite_novma(pte_t pte) 294 294 { 295 295 pte = set_pte_bit(pte, __pgprot(PTE_WRITE)); 296 - pte = clear_pte_bit(pte, __pgprot(PTE_RDONLY)); 296 + if (pte_sw_dirty(pte)) 297 + pte = clear_pte_bit(pte, __pgprot(PTE_RDONLY)); 297 298 return pte; 298 299 } 299 300
+8 -3
arch/arm64/mm/copypage.c
··· 35 35 from != folio_page(src, 0)) 36 36 return; 37 37 38 - WARN_ON_ONCE(!folio_try_hugetlb_mte_tagging(dst)); 38 + folio_try_hugetlb_mte_tagging(dst); 39 39 40 40 /* 41 41 * Populate tags for all subpages. ··· 51 51 } 52 52 folio_set_hugetlb_mte_tagged(dst); 53 53 } else if (page_mte_tagged(from)) { 54 - /* It's a new page, shouldn't have been tagged yet */ 55 - WARN_ON_ONCE(!try_page_mte_tagging(to)); 54 + /* 55 + * Most of the time it's a new page that shouldn't have been 56 + * tagged yet. However, folio migration can end up reusing the 57 + * same page without untagging it. Ignore the warning if the 58 + * page is already tagged. 59 + */ 60 + try_page_mte_tagging(to); 56 61 57 62 mte_copy_page_tags(kto, kfrom); 58 63 set_page_mte_tagged(to);