kernel/linux.git/include/asm-generic, branch v5.10.257

mm/hugetlb: fix excessive IPI broadcasts when unsharing PMD tables using mmu_gather

2026-04-18T08:31:14+00:00

commit 8ce720d5bd91e9dc16db3604aa4b1bf76770a9a1 upstream. As reported, ever since commit 1013af4f585f ("mm/hugetlb: fix huge_pmd_unshare() vs GUP-fast race") we can end up in some situations where we perform so many IPI broadcasts when unsharing hugetlb PMD page tables that it severely regresses some workloads. In particular, when we fork()+exit(), or when we munmap() a large area backed by many shared PMD tables, we perform one IPI broadcast per unshared PMD table. There are two optimizations to be had: (1) When we process (unshare) multiple such PMD tables, such as during exit(), it is sufficient to send a single IPI broadcast (as long as we respect locking rules) instead of one per PMD table. Locking prevents that any of these PMD tables could get reused before we drop the lock. (2) When we are not the last sharer (> 2 users including us), there is no need to send the IPI broadcast. The shared PMD tables cannot become exclusive (fully unshared) before an IPI will be broadcasted by the last sharer. Concurrent GUP-fast could walk into a PMD table just before we unshared it. It could then succeed in grabbing a page from the shared page table even after munmap() etc succeeded (and supressed an IPI). But there is not difference compared to GUP-fast just sleeping for a while after grabbing the page and re-enabling IRQs. Most importantly, GUP-fast will never walk into page tables that are no-longer shared, because the last sharer will issue an IPI broadcast. (if ever required, checking whether the PUD changed in GUP-fast after grabbing the page like we do in the PTE case could handle this) So let's rework PMD sharing TLB flushing + IPI sync to use the mmu_gather infrastructure so we can implement these optimizations and demystify the code at least a bit. Extend the mmu_gather infrastructure to be able to deal with our special hugetlb PMD table sharing implementation. To make initialization of the mmu_gather easier when working on a single VMA (in particular, when dealing with hugetlb), provide tlb_gather_mmu_vma(). We'll consolidate the handling for (full) unsharing of PMD tables in tlb_unshare_pmd_ptdesc() and tlb_flush_unshared_tables(), and track in "struct mmu_gather" whether we had (full) unsharing of PMD tables. Because locking is very special (concurrent unsharing+reuse must be prevented), we disallow deferring flushing to tlb_finish_mmu() and instead require an explicit earlier call to tlb_flush_unshared_tables(). >From hugetlb code, we call huge_pmd_unshare_flush() where we make sure that the expected lock protecting us from concurrent unsharing+reuse is still held. Check with a VM_WARN_ON_ONCE() in tlb_finish_mmu() that tlb_flush_unshared_tables() was properly called earlier. Document it all properly. Notes about tlb_remove_table_sync_one() interaction with unsharing: There are two fairly tricky things: (1) tlb_remove_table_sync_one() is a NOP on architectures without CONFIG_MMU_GATHER_RCU_TABLE_FREE. Here, the assumption is that the previous TLB flush would send an IPI to all relevant CPUs. Careful: some architectures like x86 only send IPIs to all relevant CPUs when tlb->freed_tables is set. The relevant architectures should be selecting MMU_GATHER_RCU_TABLE_FREE, but x86 might not do that in stable kernels and it might have been problematic before this patch. Also, the arch flushing behavior (independent of IPIs) is different when tlb->freed_tables is set. Do we have to enlighten them to also take care of tlb->unshared_tables? So far we didn't care, so hopefully we are fine. Of course, we could be setting tlb->freed_tables as well, but that might then unnecessarily flush too much, because the semantics of tlb->freed_tables are a bit fuzzy. This patch changes nothing in this regard. (2) tlb_remove_table_sync_one() is not a NOP on architectures with CONFIG_MMU_GATHER_RCU_TABLE_FREE that actually don't need a sync. Take x86 as an example: in the common case (!pv, !X86_FEATURE_INVLPGB) we still issue IPIs during TLB flushes and don't actually need the second tlb_remove_table_sync_one(). This optimized can be implemented on top of this, by checking e.g., in tlb_remove_table_sync_one() whether we really need IPIs. But as described in (1), it really must honor tlb->freed_tables then to send IPIs to all relevant CPUs. Notes on TLB flushing changes: (1) Flushing for non-shared PMD tables We're converting from flush_hugetlb_tlb_range() to tlb_remove_huge_tlb_entry(). Given that we properly initialize the MMU gather in tlb_gather_mmu_vma() to be hugetlb aware, similar to __unmap_hugepage_range(), that should be fine. (2) Flushing for shared PMD tables We're converting from various things (flush_hugetlb_tlb_range(), tlb_flush_pmd_range(), flush_tlb_range()) to tlb_flush_pmd_range(). tlb_flush_pmd_range() achieves the same that tlb_remove_huge_tlb_entry() would achieve in these scenarios. Note that tlb_remove_huge_tlb_entry() also calls __tlb_remove_tlb_entry(), however that is only implemented on powerpc, which does not support PMD table sharing. Similar to (1), tlb_gather_mmu_vma() should make sure that TLB flushing keeps on working as expected. Further, note that the ptdesc_pmd_pts_dec() in huge_pmd_share() is not a concern, as we are holding the i_mmap_lock the whole time, preventing concurrent unsharing. That ptdesc_pmd_pts_dec() usage will be removed separately as a cleanup later. There are plenty more cleanups to be had, but they have to wait until this is fixed. [david@kernel.org: fix kerneldoc] Link: https://lkml.kernel.org/r/f223dd74-331c-412d-93fc-69e360a5006c@kernel.org Link: https://lkml.kernel.org/r/20251223214037.580860-5-david@kernel.org Fixes: 1013af4f585f ("mm/hugetlb: fix huge_pmd_unshare() vs GUP-fast race") Signed-off-by: David Hildenbrand (Red Hat) Reported-by: "Uschakow, Stanislav" Closes: https://lore.kernel.org/all/4d3878531c76479d9f8ca9789dc6485d@amazon.de/ Tested-by: Laurence Oberman Acked-by: Harry Yoo Reviewed-by: Lorenzo Stoakes Cc: Lance Yang Cc: Liu Shixin Cc: Oscar Salvador Cc: Rik van Riel Cc: Signed-off-by: Andrew Morton (cherry picked from commit 8ce720d5bd91e9dc16db3604aa4b1bf76770a9a1) [ David: We don't have ptdesc and the wrappers, so work directly on page->pt_share_count and pass "struct page" instead of "struct ptdesc". CONFIG_HUGETLB_PMD_PAGE_TABLE_SHARING is still called CONFIG_ARCH_WANT_HUGE_PMD_SHARE and is set even without CONFIG_HUGETLB_PAGE. We don't have 550a7d60bd5e ("mm, hugepages: add mremap() support for hugepage backed vma"), so move_hugetlb_page_tables() does not exist. We don't have 40549ba8f8e0 ("hugetlb: use new vma_lock for pmd sharing synchronization") and a98a2f0c8ce1 ("mm/rmap: split migration into its own so changes in mm/rmap.c looks quite different. We don't have 4ddb4d91b82f ("hugetlb: do not update address in huge_pmd_unshare"), so huge_pmd_unshare() still gets a pointer to an address. tlb_gather_mmu() + tlb_finish_mmu() still consume ranges, so also teach tlb_gather_mmu_vma() to forward ranges. Some smaller contextual stuff, in particular, around tlb_gather_mmu_full(). ] Signed-off-by: David Hildenbrand (Arm) Signed-off-by: Sasha Levin

audit: add missing syscalls to read class

2026-03-04T12:19:51+00:00

[ Upstream commit bcb90a2834c7393c26df9609b889a3097b7700cd ] The "at" variant of getxattr() and listxattr() are missing from the audit read class. Calling getxattrat() or listxattrat() on a file to read its extended attributes will bypass audit rules such as: -w /tmp/test -p rwa -k test_rwa The current patch adds missing syscalls to the audit read class. Signed-off-by: Jeffrey Bencteux Signed-off-by: Paul Moore Signed-off-by: Sasha Levin

audit: add fchmodat2() to change attributes class

2026-03-04T12:19:50+00:00

[ Upstream commit 4f493a6079b588cf1f04ce5ed6cdad45ab0d53dc ] fchmodat2(), introduced in version 6.6 is currently not in the change attribute class of audit. Calling fchmodat2() to change a file attribute in the same fashion than chmod() or fchmodat() will bypass audit rules such as: -w /tmp/test -p rwa -k test_rwa The current patch adds fchmodat2() to the change attributes class. Signed-off-by: Jeffrey Bencteux Signed-off-by: Paul Moore Signed-off-by: Sasha Levin

asm-generic: Add memory barrier dma_mb()

2025-08-28T14:22:57+00:00

[ Upstream commit ed59dfd9509d172e4920994ed9cbebf93b0050cc ] The memory barrier dma_mb() is introduced by commit a76a37777f2c ("iommu/arm-smmu-v3: Ensure queue is read after updating prod pointer"), which is used to ensure that prior (both reads and writes) accesses to memory by a CPU are ordered w.r.t. a subsequent MMIO write. Reviewed-by: Arnd Bergmann # for asm-generic Signed-off-by: Kefeng Wang Reviewed-by: Marco Elver Link: https://lore.kernel.org/r/20220523113126.171714-2-wangkefeng.wang@huawei.com Signed-off-by: Will Deacon Stable-dep-of: aa6956150f82 ("wifi: ath11k: fix dest ring-buffer corruption when ring is full") Signed-off-by: Sasha Levin Signed-off-by: Greg Kroah-Hartman

locking/barriers, kcsan: Support generic instrumentation

2025-08-28T14:22:57+00:00

[ Upstream commit 2505a51ac6f249956735e0a369e2404f96eebef0 ] Thus far only smp_*() barriers had been defined by asm-generic/barrier.h based on __smp_*() barriers, because the !SMP case is usually generic. With the introduction of instrumentation, it also makes sense to have asm-generic/barrier.h assist in the definition of instrumented versions of mb(), rmb(), wmb(), dma_rmb(), and dma_wmb(). Because there is no requirement to distinguish the !SMP case, the definition can be simpler: we can avoid also providing fallbacks for the __ prefixed cases, and only check if `defined(__)`, to finally define the KCSAN-instrumented versions. This also allows for the compiler to complain if an architecture accidentally defines both the normal and __ prefixed variant. Signed-off-by: Marco Elver Signed-off-by: Paul E. McKenney Stable-dep-of: aa6956150f82 ("wifi: ath11k: fix dest ring-buffer corruption when ring is full") Signed-off-by: Sasha Levin Signed-off-by: Greg Kroah-Hartman

vmlinux.lds: Ensure that const vars with relocations are mapped R/O

2025-03-13T11:47:34+00:00

commit 68f3ea7ee199ef77551e090dfef5a49046ea8443 upstream. In the kernel, there are architectures (x86, arm64) that perform boot-time relocation (for KASLR) without relying on PIE codegen. In this case, all const global objects are emitted into .rodata, including const objects with fields that will be fixed up by the boot-time relocation code. This implies that .rodata (and .text in some cases) need to be writable at boot, but they will usually be mapped read-only as soon as the boot completes. When using PIE codegen, the compiler will emit const global objects into .data.rel.ro rather than .rodata if the object contains fields that need such fixups at boot-time. This permits the linker to annotate such regions as requiring read-write access only at load time, but not at execution time (in user space), while keeping .rodata truly const (in user space, this is important for reducing the CoW footprint of dynamic executables). This distinction does not matter for the kernel, but it does imply that const data will end up in writable memory if the .data.rel.ro sections are not treated in a special way, as they will end up in the writable .data segment by default. So emit .data.rel.ro into the .rodata segment. Cc: stable@vger.kernel.org Signed-off-by: Ard Biesheuvel Link: https://lore.kernel.org/r/20250221135704.431269-5-ardb+git@google.com Signed-off-by: Josh Poimboeuf Signed-off-by: Greg Kroah-Hartman

asm-generic: qspinlock: fix queued_spin_value_unlocked() implementation

2023-12-20T14:44:35+00:00

[ Upstream commit 125b0bb95dd6bec81b806b997a4ccb026eeecf8f ] We really don't want to do atomic_read() or anything like that, since we already have the value, not the lock. The whole point of this is that we've loaded the lock from memory, and we want to check whether the value we loaded was a locked one or not. The main use of this is the lockref code, which loads both the lock and the reference count in one atomic operation, and then works on that combined value. With the atomic_read(), the compiler would pointlessly spill the value to the stack, in order to then be able to read it back "atomically". This is the qspinlock version of commit c6f4a9002252 ("asm-generic: ticket-lock: Optimize arch_spin_value_unlocked()") which fixed this same bug for ticket locks. Cc: Guo Ren Cc: Ingo Molnar Cc: Waiman Long Link: https://lore.kernel.org/all/CAHk-=whNRv0v6kQiV5QO6DJhjH4KEL36vWQ6Re8Csrnh4zbRkQ@mail.gmail.com/ Signed-off-by: Linus Torvalds Signed-off-by: Sasha Levin

word-at-a-time: use the same return type for has_zero regardless of endianness

2023-08-11T09:57:47+00:00

[ Upstream commit 79e8328e5acbe691bbde029a52c89d70dcbc22f3 ] Compiling big-endian targets with Clang produces the diagnostic: fs/namei.c:2173:13: warning: use of bitwise '|' with boolean operands [-Wbitwise-instead-of-logical] } while (!(has_zero(a, &adata, &constants) | has_zero(b, &bdata, &constants))); ~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ || fs/namei.c:2173:13: note: cast one or both operands to int to silence this warning It appears that when has_zero was introduced, two definitions were produced with different signatures (in particular different return types). Looking at the usage in hash_name() in fs/namei.c, I suspect that has_zero() is meant to be invoked twice per while loop iteration; using logical-or would not update `bdata` when `a` did not have zeros. So I think it's preferred to always return an unsigned long rather than a bool than update the while loop in hash_name() to use a logical-or rather than bitwise-or. [ Also changed powerpc version to do the same - Linus ] Link: https://github.com/ClangBuiltLinux/linux/issues/1832 Link: https://lore.kernel.org/lkml/20230801-bitwise-v1-1-799bec468dc4@google.com/ Fixes: 36126f8f2ed8 ("word-at-a-time: make the interfaces truly generic") Debugged-by: Nathan Chancellor Signed-off-by: Nick Desaulniers Acked-by: Heiko Carstens Cc: Arnd Bergmann Signed-off-by: Linus Torvalds Signed-off-by: Sasha Levin

init: Remove check_bugs() leftovers

2023-08-08T17:57:37+00:00

commit 61235b24b9cb37c13fcad5b9596d59a1afdcec30 upstream Everything is converted over to arch_cpu_finalize_init(). Remove the check_bugs() leftovers including the empty stubs in asm-generic, alpha, parisc, powerpc and xtensa. Signed-off-by: Thomas Gleixner Reviewed-by: Richard Henderson Link: https://lore.kernel.org/r/20230613224545.553215951@linutronix.de Signed-off-by: Daniel Sneddon Signed-off-by: Greg Kroah-Hartman

mm: rename p4d_page_vaddr to p4d_pgtable and make it return pud_t *

2023-07-27T06:43:58+00:00

[ Upstream commit dc4875f0e791de554bdc45aa1dbd6e45e107e50f ] No functional change in this patch. [aneesh.kumar@linux.ibm.com: m68k build error reported by kernel robot] Link: https://lkml.kernel.org/r/87tulxnb2v.fsf@linux.ibm.com Link: https://lkml.kernel.org/r/20210615110859.320299-2-aneesh.kumar@linux.ibm.com Link: https://lore.kernel.org/linuxppc-dev/CAHk-=wi+J+iodze9FtjM3Zi4j4OeS+qqbKxME9QN4roxPEXH9Q@mail.gmail.com/ Signed-off-by: Aneesh Kumar K.V Cc: Christophe Leroy Cc: Hugh Dickins Cc: Joel Fernandes Cc: Kalesh Singh Cc: Kirill A. Shutemov Cc: Michael Ellerman Cc: Nicholas Piggin Cc: Stephen Rothwell Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds Stable-dep-of: 0da90af431ab ("powerpc/book3s64/mm: Fix DirectMap stats in /proc/meminfo") Signed-off-by: Sasha Levin