The patch titled Subject: riscv: pgtable: move pagetable_dtor() to __tlb_remove_table() has been added to the -mm mm-unstable branch. Its filename is riscv-pgtable-move-pagetable_dtor-to-__tlb_remove_table.patch This patch will shortly appear at https://git.kernel.org/pub/scm/linux/kernel/git/akpm/25-new.git/tree/patches/riscv-pgtable-move-pagetable_dtor-to-__tlb_remove_table.patch This patch will later appear in the mm-unstable branch at git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm Before you just go and hit "reply", please: a) Consider who else should be cc'ed b) Prefer to cc a suitable mailing list as well c) Ideally: find the original patch on the mailing list and do a reply-to-all to that, adding suitable additional cc's *** Remember to use Documentation/process/submit-checklist.rst when testing your code *** The -mm tree is included into linux-next via the mm-everything branch at git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm and is updated there every 2-3 working days ------------------------------------------------------ From: Qi Zheng <zhengqi.arch@xxxxxxxxxxxxx> Subject: riscv: pgtable: move pagetable_dtor() to __tlb_remove_table() Date: Mon, 23 Dec 2024 17:40:56 +0800 Move pagetable_dtor() to __tlb_remove_table(), so that ptlock and page table pages can be freed together (regardless of whether RCU is used). This prevents the use-after-free problem where the ptlock is freed immediately but the page table pages is freed later via RCU. Page tables shouldn't have swap cache, so use pagetable_free() instead of free_page_and_swap_cache() to free page table pages. By the way, move the comment above __tlb_remove_table() to riscv_tlb_remove_ptdesc(), it will be more appropriate. Link: https://lkml.kernel.org/r/0e8f0b3835c15e99145e0006ac1020ae45a2b166.1734945104.git.zhengqi.arch@xxxxxxxxxxxxx Signed-off-by: Qi Zheng <zhengqi.arch@xxxxxxxxxxxxx> Suggested-by: Peter Zijlstra (Intel) <peterz@xxxxxxxxxxxxx> Cc: Alexander Gordeev <agordeev@xxxxxxxxxxxxx> Cc: Aneesh Kumar K.V (Arm) <aneesh.kumar@xxxxxxxxxx> Cc: Arnd Bergmann <arnd@xxxxxxxx> Cc: Dave Hansen <dave.hansen@xxxxxxxxxxxxxxx> Cc: David Hildenbrand <david@xxxxxxxxxx> Cc: David Rientjes <rientjes@xxxxxxxxxx> Cc: Hugh Dickins <hughd@xxxxxxxxxx> Cc: Jann Horn <jannh@xxxxxxxxxx> Cc: Kevin Brodsky <kevin.brodsky@xxxxxxx> Cc: Lorenzo Stoakes <lorenzo.stoakes@xxxxxxxxxx> Cc: Matthew Wilcox (Oracle) <willy@xxxxxxxxxxxxx> Cc: Mike Rapoport (Microsoft) <rppt@xxxxxxxxxx> Cc: Muchun Song <muchun.song@xxxxxxxxx> Cc: Nicholas Piggin <npiggin@xxxxxxxxx> Cc: Ryan Roberts <ryan.roberts@xxxxxxx> Cc: Thomas Gleixner <tglx@xxxxxxxxxxxxx> Cc: Vishal Moola (Oracle) <vishal.moola@xxxxxxxxx> Cc: Will Deacon <will@xxxxxxxxxx> Cc: Yu Zhao <yuzhao@xxxxxxxxxx> Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> --- arch/riscv/include/asm/pgalloc.h | 38 ++++++++++++----------------- arch/riscv/include/asm/tlb.h | 14 +++------- 2 files changed, 21 insertions(+), 31 deletions(-) --- a/arch/riscv/include/asm/pgalloc.h~riscv-pgtable-move-pagetable_dtor-to-__tlb_remove_table +++ a/arch/riscv/include/asm/pgalloc.h @@ -15,12 +15,22 @@ #define __HAVE_ARCH_PUD_FREE #include <asm-generic/pgalloc.h> +/* + * While riscv platforms with riscv_ipi_for_rfence as true require an IPI to + * perform TLB shootdown, some platforms with riscv_ipi_for_rfence as false use + * SBI to perform TLB shootdown. To keep software pagetable walkers safe in this + * case we switch to RCU based table free (MMU_GATHER_RCU_TABLE_FREE). See the + * comment below 'ifdef CONFIG_MMU_GATHER_RCU_TABLE_FREE' in include/asm-generic/tlb.h + * for more details. + */ static inline void riscv_tlb_remove_ptdesc(struct mmu_gather *tlb, void *pt) { - if (riscv_use_sbi_for_rfence()) + if (riscv_use_sbi_for_rfence()) { tlb_remove_ptdesc(tlb, pt); - else + } else { + pagetable_dtor(pt); tlb_remove_page_ptdesc(tlb, pt); + } } static inline void pmd_populate_kernel(struct mm_struct *mm, @@ -97,23 +107,15 @@ static inline void pud_free(struct mm_st static inline void __pud_free_tlb(struct mmu_gather *tlb, pud_t *pud, unsigned long addr) { - if (pgtable_l4_enabled) { - struct ptdesc *ptdesc = virt_to_ptdesc(pud); - - pagetable_dtor(ptdesc); - riscv_tlb_remove_ptdesc(tlb, ptdesc); - } + if (pgtable_l4_enabled) + riscv_tlb_remove_ptdesc(tlb, virt_to_ptdesc(pud)); } static inline void __p4d_free_tlb(struct mmu_gather *tlb, p4d_t *p4d, unsigned long addr) { - if (pgtable_l5_enabled) { - struct ptdesc *ptdesc = virt_to_ptdesc(p4d); - - pagetable_dtor(ptdesc); + if (pgtable_l5_enabled) riscv_tlb_remove_ptdesc(tlb, virt_to_ptdesc(p4d)); - } } #endif /* __PAGETABLE_PMD_FOLDED */ @@ -142,10 +144,7 @@ static inline pgd_t *pgd_alloc(struct mm static inline void __pmd_free_tlb(struct mmu_gather *tlb, pmd_t *pmd, unsigned long addr) { - struct ptdesc *ptdesc = virt_to_ptdesc(pmd); - - pagetable_dtor(ptdesc); - riscv_tlb_remove_ptdesc(tlb, ptdesc); + riscv_tlb_remove_ptdesc(tlb, virt_to_ptdesc(pmd)); } #endif /* __PAGETABLE_PMD_FOLDED */ @@ -153,10 +152,7 @@ static inline void __pmd_free_tlb(struct static inline void __pte_free_tlb(struct mmu_gather *tlb, pgtable_t pte, unsigned long addr) { - struct ptdesc *ptdesc = page_ptdesc(pte); - - pagetable_dtor(ptdesc); - riscv_tlb_remove_ptdesc(tlb, ptdesc); + riscv_tlb_remove_ptdesc(tlb, page_ptdesc(pte)); } #endif /* CONFIG_MMU */ --- a/arch/riscv/include/asm/tlb.h~riscv-pgtable-move-pagetable_dtor-to-__tlb_remove_table +++ a/arch/riscv/include/asm/tlb.h @@ -11,19 +11,13 @@ struct mmu_gather; static void tlb_flush(struct mmu_gather *tlb); #ifdef CONFIG_MMU -#include <linux/swap.h> -/* - * While riscv platforms with riscv_ipi_for_rfence as true require an IPI to - * perform TLB shootdown, some platforms with riscv_ipi_for_rfence as false use - * SBI to perform TLB shootdown. To keep software pagetable walkers safe in this - * case we switch to RCU based table free (MMU_GATHER_RCU_TABLE_FREE). See the - * comment below 'ifdef CONFIG_MMU_GATHER_RCU_TABLE_FREE' in include/asm-generic/tlb.h - * for more details. - */ static inline void __tlb_remove_table(void *table) { - free_page_and_swap_cache(table); + struct ptdesc *ptdesc = (struct ptdesc *)table; + + pagetable_dtor(ptdesc); + pagetable_free(ptdesc); } #endif /* CONFIG_MMU */ _ Patches currently in -mm which might be from zhengqi.arch@xxxxxxxxxxxxx are mm-pgtable-make-ptep_clear-non-atomic.patch mm-khugepaged-recheck-pmd-state-in-retract_page_tables.patch mm-userfaultfd-recheck-dst_pmd-entry-in-move_pages_pte.patch mm-userfaultfd-recheck-dst_pmd-entry-in-move_pages_pte-fix.patch mm-introduce-zap_nonpresent_ptes.patch mm-introduce-do_zap_pte_range.patch mm-skip-over-all-consecutive-none-ptes-in-do_zap_pte_range.patch mm-zap_install_uffd_wp_if_needed-return-whether-uffd-wp-pte-has-been-re-installed.patch mm-do_zap_pte_range-return-any_skipped-information-to-the-caller.patch mm-make-zap_pte_range-handle-full-within-pmd-range.patch mm-pgtable-reclaim-empty-pte-page-in-madvisemadv_dontneed.patch mm-pgtable-reclaim-empty-pte-page-in-madvisemadv_dontneed-fix.patch x86-mm-free-page-table-pages-by-rcu-instead-of-semi-rcu.patch mm-pgtable-make-ptlock-be-freed-by-rcu.patch x86-select-arch_supports_pt_reclaim-if-x86_64.patch revert-mm-pgtable-make-ptlock-be-freed-by-rcu.patch mm-pgtable-add-statistics-for-p4d-level-page-table.patch arm64-pgtable-use-mmu-gather-to-free-p4d-level-page-table.patch s390-pgtable-add-statistics-for-pud-and-p4d-level-page-table.patch mm-pgtable-introduce-pagetable_dtor.patch arm-pgtable-move-pagetable_dtor-to-__tlb_remove_table.patch arm64-pgtable-move-pagetable_dtor-to-__tlb_remove_table.patch riscv-pgtable-move-pagetable_dtor-to-__tlb_remove_table.patch x86-pgtable-move-pagetable_dtor-to-__tlb_remove_table.patch s390-pgtable-also-move-pagetable_dtor-of-pxd-to-__tlb_remove_table.patch mm-pgtable-introduce-generic-__tlb_remove_table.patch mm-pgtable-move-__tlb_remove_table_one-in-x86-to-generic-file.patch mm-pgtable-remove-tlb_remove_page_ptdesc.patch mm-pgtable-remove-tlb_remove_ptdesc.patch mm-pgtable-introduce-generic-pagetable_dtor_free.patch