kernel.git/arch/arm64/mm, branch linux-6.0.y

arm64: mm: kfence: only handle translation faults

2022-12-31T12:25:38Z

[ Upstream commit 0bb1fbffc631064db567ccaeb9ed6b6df6342b66 ] Alexander noted that KFENCE only expects to handle faults from invalid page table entries (i.e. translation faults), but arm64's fault handling logic will call kfence_handle_page_fault() for other types of faults, including alignment faults caused by unaligned atomics. This has the unfortunate property of causing those other faults to be reported as "KFENCE: use-after-free", which is misleading and hinders debugging. Fix this by only forwarding unhandled translation faults to the KFENCE code, similar to what x86 does already. Alexander has verified that this passes all the tests in the KFENCE test suite and avoids bogus reports on misaligned atomics. Link: https://lore.kernel.org/all/20221102081620.1465154-1-zhongbaisong@huawei.com/ Fixes: 840b23986344 ("arm64, kfence: enable KFENCE for ARM64") Signed-off-by: Mark Rutland Reviewed-by: Alexander Potapenko Tested-by: Alexander Potapenko Cc: Catalin Marinas Cc: Marco Elver Cc: Will Deacon Link: https://lore.kernel.org/r/20221114104411.2853040-1-mark.rutland@arm.com Signed-off-by: Will Deacon Signed-off-by: Sasha Levin

arm64: fix rodata=full again

2022-11-26T08:27:22Z

[ Upstream commit 2081b3bd0c11757725dcab9ba5d38e1bddb03459 ] Commit 2e8cff0a0eee87b2 ("arm64: fix rodata=full") addressed a couple of issues with the rodata= kernel command line option, which is not a simple boolean on arm64, and inadvertently got broken due to changes in the generic bool handling. Unfortunately, the resulting code never clears the rodata_full boolean variable if it defaults to true and rodata=on or rodata=off is passed, as the generic code is not aware of the existence of this variable. Given the way this code is plumbed together, clearing rodata_full when returning false from arch_parse_debug_rodata() may result in inconsistencies if the generic code decides that it cannot parse the right hand side, so the best way to deal with this is to only take rodata_full in account if rodata_enabled is also true. Fixes: 2e8cff0a0eee ("arm64: fix rodata=full") Cc: # 6.0.x Signed-off-by: Ard Biesheuvel Acked-by: Will Deacon Link: https://lore.kernel.org/r/20221103170015.4124426-1-ardb@kernel.org Signed-off-by: Catalin Marinas Signed-off-by: Sasha Levin

arm64/mm: fold check for KFENCE into can_set_direct_map()

2022-11-26T08:27:22Z

[ Upstream commit b9dd04a20f81333e4b99662f1bbaf7c9e3a1e137 ] KFENCE requires linear map to be mapped at page granularity, so that it is possible to protect/unprotect single pages, just like with rodata_full and DEBUG_PAGEALLOC. Instead of repating can_set_direct_map() || IS_ENABLED(CONFIG_KFENCE) make can_set_direct_map() handle the KFENCE case. This also prevents potential false positives in kernel_page_present() that may return true for non-present page if CONFIG_KFENCE is enabled. Signed-off-by: Mike Rapoport Reviewed-by: Anshuman Khandual Link: https://lore.kernel.org/r/20220921074841.382615-1-rppt@kernel.org Signed-off-by: Catalin Marinas Stable-dep-of: 2081b3bd0c11 ("arm64: fix rodata=full again") Signed-off-by: Sasha Levin

arm64: mte: Avoid setting PG_mte_tagged if no tags cleared or restored

2022-10-21T10:37:42Z

commit a8e5e5146ad08d794c58252bab00b261045ef16d upstream. Prior to commit 69e3b846d8a7 ("arm64: mte: Sync tags for pages where PTE is untagged"), mte_sync_tags() was only called for pte_tagged() entries (those mapped with PROT_MTE). Therefore mte_sync_tags() could safely use test_and_set_bit(PG_mte_tagged, &page->flags) without inadvertently setting PG_mte_tagged on an untagged page. The above commit was required as guests may enable MTE without any control at the stage 2 mapping, nor a PROT_MTE mapping in the VMM. However, the side-effect was that any page with a PTE that looked like swap (or migration) was getting PG_mte_tagged set automatically. A subsequent page copy (e.g. migration) copied the tags to the destination page even if the tags were owned by KASAN. This issue was masked by the page_kasan_tag_reset() call introduced in commit e5b8d9218951 ("arm64: mte: reset the page tag in page->flags"). When this commit was reverted (20794545c146), KASAN started reporting access faults because the overriding tags in a page did not match the original page->flags (with CONFIG_KASAN_HW_TAGS=y): BUG: KASAN: invalid-access in copy_page+0x10/0xd0 arch/arm64/lib/copy_page.S:26 Read at addr f5ff000017f2e000 by task syz-executor.1/2218 Pointer tag: [f5], memory tag: [f2] Move the PG_mte_tagged bit setting from mte_sync_tags() to the actual place where tags are cleared (mte_sync_page_tags()) or restored (mte_restore_tags()). Signed-off-by: Catalin Marinas Reported-by: syzbot+c2c79c6d6eddc5262b77@syzkaller.appspotmail.com Fixes: 69e3b846d8a7 ("arm64: mte: Sync tags for pages where PTE is untagged") Cc: # 5.14.x Cc: Steven Price Cc: Andrey Konovalov Cc: Vincenzo Frascino Cc: Will Deacon Link: https://lore.kernel.org/r/0000000000004387dc05e5888ae5@google.com/ Reviewed-by: Steven Price Link: https://lore.kernel.org/r/20221006163354.3194102-1-catalin.marinas@arm.com Signed-off-by: Catalin Marinas Signed-off-by: Greg Kroah-Hartman

arm64: mte: move register initialization to C

2022-10-21T10:37:42Z

commit 973b9e37330656dec719ede508e4dc40e5c2d80c upstream. If FEAT_MTE2 is disabled via the arm64.nomte command line argument on a CPU that claims to support FEAT_MTE2, the kernel will use Tagged Normal in the MAIR. If we interpret arm64.nomte to mean that the CPU does not in fact implement FEAT_MTE2, setting the system register like this may lead to UNSPECIFIED behavior. Fix it by arranging for MAIR to be set in the C function cpu_enable_mte which is called based on the sanitized version of the system register. There is no need for the rest of the MTE-related system register initialization to happen from assembly, with the exception of TCR_EL1, which must be set to include at least TBI1 because the secondary CPUs access KASan-allocated data structures early. Therefore, make the TCR_EL1 initialization unconditional and move the rest of the initialization to cpu_enable_mte so that we no longer have a dependency on the unsanitized ID register value. Co-developed-by: Evgenii Stepanov Signed-off-by: Peter Collingbourne Signed-off-by: Evgenii Stepanov Suggested-by: Catalin Marinas Reported-by: kernel test robot Fixes: 3b714d24ef17 ("arm64: mte: CPU feature detection and initial sysreg configuration") Cc: # 5.10.x Link: https://lore.kernel.org/r/20220915222053.3484231-1-eugenis@google.com Signed-off-by: Catalin Marinas Signed-off-by: Greg Kroah-Hartman

arm64: mm: don't acquire mutex when rewriting swapper

2022-09-22T11:55:39Z

Since commit: 47546a1912fc4a03 ("arm64: mm: install KPTI nG mappings with MMU enabled)" ... when building with CONFIG_DEBUG_ATOMIC_SLEEP=y and booting under QEMU TCG with '-cpu max', there's a boot-time splat: | BUG: sleeping function called from invalid context at kernel/locking/mutex.c:580 | in_atomic(): 1, irqs_disabled(): 128, non_block: 0, pid: 15, name: migration/0 | preempt_count: 1, expected: 0 | RCU nest depth: 0, expected: 0 | no locks held by migration/0/15. | irq event stamp: 28 | hardirqs last enabled at (27): [] _raw_spin_unlock_irq+0x3c/0x7c | hardirqs last disabled at (28): [] multi_cpu_stop+0x150/0x18c | softirqs last enabled at (0): [] copy_process+0x594/0x1964 | softirqs last disabled at (0): [<0000000000000000>] 0x0 | CPU: 0 PID: 15 Comm: migration/0 Not tainted 6.0.0-rc3-00002-g419b42ff7eef #3 | Hardware name: linux,dummy-virt (DT) | Stopper: multi_cpu_stop+0x0/0x18c <- stop_cpus.constprop.0+0xa0/0xfc | Call trace: | dump_backtrace.part.0+0xd0/0xe0 | show_stack+0x1c/0x5c | dump_stack_lvl+0x88/0xb4 | dump_stack+0x1c/0x38 | __might_resched+0x180/0x230 | __might_sleep+0x4c/0xa0 | __mutex_lock+0x5c/0x450 | mutex_lock_nested+0x30/0x40 | create_kpti_ng_temp_pgd+0x4fc/0x6d0 | kpti_install_ng_mappings+0x2b8/0x3b0 | cpu_enable_non_boot_scope_capabilities+0x7c/0xd0 | multi_cpu_stop+0xa0/0x18c | cpu_stopper_thread+0x88/0x11c | smpboot_thread_fn+0x1ec/0x290 | kthread+0x118/0x120 | ret_from_fork+0x10/0x20 Since commit: ee017ee353506fce ("arm64/mm: avoid fixmap race condition when create pud mapping") ... once the kernel leave the SYSTEM_BOOTING state, the fixmap pagetable entries are protected by the fixmap_lock mutex. The new KPTI rewrite code uses __create_pgd_mapping() to create a temporary pagetable. This happens in atomic context, after secondary CPUs are brought up and the kernel has left the SYSTEM_BOOTING state. Hence we try to acquire a mutex in atomic context, which is generally unsound (though benign in this case as the mutex should be free and all other CPUs are quiescent). This patch avoids the issue by pulling the mutex out of alloc_init_pud() and calling it at a higher level in the pagetable manipulation code. This allows it to be used without locking where one CPU is known to be in exclusive control of the machine, even after having left the SYSTEM_BOOTING state. Fixes: 47546a1912fc ("arm64: mm: install KPTI nG mappings with MMU enabled") Signed-off-by: Mark Rutland Cc: Ard Biesheuvel Cc: Catalin Marinas Cc: Will Deacon Reviewed-by: Ard Biesheuvel Link: https://lore.kernel.org/r/20220920134731.1625740-1-mark.rutland@arm.com Signed-off-by: Will Deacon

arm64: fix rodata=full

2022-08-23T10:02:02Z

On arm64, "rodata=full" has been suppored (but not documented) since commit: c55191e96caa9d78 ("arm64: mm: apply r/o permissions of VM areas to its linear alias as well") As it's necessary to determine the rodata configuration early during boot, arm64 has an early_param() handler for this, whereas init/main.c has a __setup() handler which is run later. Unfortunately, this split meant that since commit: f9a40b0890658330 ("init/main.c: return 1 from handled __setup() functions") ... passing "rodata=full" would result in a spurious warning from the __setup() handler (though RO permissions would be configured appropriately). Further, "rodata=full" has been broken since commit: 0d6ea3ac94ca77c5 ("lib/kstrtox.c: add "false"/"true" support to kstrtobool()") ... which caused strtobool() to parse "full" as false (in addition to many other values not documented for the "rodata=" kernel parameter. This patch fixes this breakage by: * Moving the core parameter parser to an __early_param(), such that it is available early. * Adding an (optional) arch hook which arm64 can use to parse "full". * Updating the documentation to mention that "full" is valid for arm64. * Having the core parameter parser handle "on" and "off" explicitly, such that any undocumented values (e.g. typos such as "ful") are reported as errors rather than being silently accepted. Note that __setup() and early_param() have opposite conventions for their return values, where __setup() uses 1 to indicate a parameter was handled and early_param() uses 0 to indicate a parameter was handled. Fixes: f9a40b089065 ("init/main.c: return 1 from handled __setup() functions") Fixes: 0d6ea3ac94ca ("lib/kstrtox.c: add "false"/"true" support to kstrtobool()") Signed-off-by: Mark Rutland Cc: Andy Shevchenko Cc: Ard Biesheuvel Cc: Catalin Marinas Cc: Jagdish Gediya Cc: Matthew Wilcox Cc: Randy Dunlap Cc: Will Deacon Reviewed-by: Ard Biesheuvel Link: https://lore.kernel.org/r/20220817154022.3974645-1-mark.rutland@arm.com Signed-off-by: Will Deacon

mm: hugetlb_vmemmap: delete hugetlb_optimize_vmemmap_enabled()

2022-08-09T01:06:42Z

Patch series "Simplify hugetlb vmemmap and improve its readability", v2. This series aims to simplify hugetlb vmemmap and improve its readability. This patch (of 8): The name hugetlb_optimize_vmemmap_enabled() a bit confusing as it tests two conditions (enabled and pages in use). Instead of coming up to an appropriate name, we could just delete it. There is already a discussion about deleting it in thread [1]. There is only one user of hugetlb_optimize_vmemmap_enabled() outside of hugetlb_vmemmap, that is flush_dcache_page() in arch/arm64/mm/flush.c. However, it does not need to call hugetlb_optimize_vmemmap_enabled() in flush_dcache_page() since HugeTLB pages are always fully mapped and only head page will be set PG_dcache_clean meaning only head page's flag may need to be cleared (see commit cf5a501d985b). So it is easy to remove hugetlb_optimize_vmemmap_enabled(). Link: https://lore.kernel.org/all/c77c61c8-8a5a-87e8-db89-d04d8aaab4cc@oracle.com/ [1] Link: https://lkml.kernel.org/r/20220628092235.91270-2-songmuchun@bytedance.com Signed-off-by: Muchun Song Reviewed-by: Oscar Salvador Reviewed-by: Mike Kravetz Reviewed-by: Catalin Marinas Cc: Will Deacon Cc: Anshuman Khandual Cc: David Hildenbrand Cc: Jonathan Corbet Cc: Xiongchun Duan Signed-off-by: Andrew Morton

Merge tag 'mm-stable-2022-08-03' of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm

2022-08-05T23:32:45Z

Pull MM updates from Andrew Morton: "Most of the MM queue. A few things are still pending. Liam's maple tree rework didn't make it. This has resulted in a few other minor patch series being held over for next time. Multi-gen LRU still isn't merged as we were waiting for mapletree to stabilize. The current plan is to merge MGLRU into -mm soon and to later reintroduce mapletree, with a view to hopefully getting both into 6.1-rc1. Summary: - The usual batches of cleanups from Baoquan He, Muchun Song, Miaohe Lin, Yang Shi, Anshuman Khandual and Mike Rapoport - Some kmemleak fixes from Patrick Wang and Waiman Long - DAMON updates from SeongJae Park - memcg debug/visibility work from Roman Gushchin - vmalloc speedup from Uladzislau Rezki - more folio conversion work from Matthew Wilcox - enhancements for coherent device memory mapping from Alex Sierra - addition of shared pages tracking and CoW support for fsdax, from Shiyang Ruan - hugetlb optimizations from Mike Kravetz - Mel Gorman has contributed some pagealloc changes to improve latency and realtime behaviour. - mprotect soft-dirty checking has been improved by Peter Xu - Many other singleton patches all over the place" [ XFS merge from hell as per Darrick Wong in https://lore.kernel.org/all/YshKnxb4VwXycPO8@magnolia/ ] * tag 'mm-stable-2022-08-03' of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm: (282 commits) tools/testing/selftests/vm/hmm-tests.c: fix build mm: Kconfig: fix typo mm: memory-failure: convert to pr_fmt() mm: use is_zone_movable_page() helper hugetlbfs: fix inaccurate comment in hugetlbfs_statfs() hugetlbfs: cleanup some comments in inode.c hugetlbfs: remove unneeded header file hugetlbfs: remove unneeded hugetlbfs_ops forward declaration hugetlbfs: use helper macro SZ_1{K,M} mm: cleanup is_highmem() mm/hmm: add a test for cross device private faults selftests: add soft-dirty into run_vmtests.sh selftests: soft-dirty: add test for mprotect mm/mprotect: fix soft-dirty check in can_change_pte_writable() mm: memcontrol: fix potential oom_lock recursion deadlock mm/gup.c: fix formatting in check_and_migrate_movable_page() xfs: fail dax mount if reflink is enabled on a partition mm/memcontrol.c: remove the redundant updating of stats_flush_threshold userfaultfd: don't fail on unrecognized features hugetlb_cgroup: fix wrong hugetlb cgroup numa stat ...

Merge tag 'efi-next-for-v5.20' of git://git.kernel.org/pub/scm/linux/kernel/git/efi/efi

2022-08-03T21:38:02Z

Pull EFI updates from Ard Biesheuvel: - Enable mirrored memory for arm64 - Fix up several abuses of the efivar API - Refactor the efivar API in preparation for moving the 'business logic' part of it into efivarfs - Enable ACPI PRM on arm64 * tag 'efi-next-for-v5.20' of git://git.kernel.org/pub/scm/linux/kernel/git/efi/efi: (24 commits) ACPI: Move PRM config option under the main ACPI config ACPI: Enable Platform Runtime Mechanism(PRM) support on ARM64 ACPI: PRM: Change handler_addr type to void pointer efi: Simplify arch_efi_call_virt() macro drivers: fix typo in firmware/efi/memmap.c efi: vars: Drop __efivar_entry_iter() helper which is no longer used efi: vars: Use locking version to iterate over efivars linked lists efi: pstore: Omit efivars caching EFI varstore access layer efi: vars: Add thin wrapper around EFI get/set variable interface efi: vars: Don't drop lock in the middle of efivar_init() pstore: Add priv field to pstore_record for backend specific use Input: applespi - avoid efivars API and invoke EFI services directly selftests/kexec: remove broken EFI_VARS secure boot fallback check brcmfmac: Switch to appropriate helper to load EFI variable contents iwlwifi: Switch to proper EFI variable store interface media: atomisp_gmin_platform: stop abusing efivar API efi: efibc: avoid efivar API for setting variables efi: avoid efivars layer when loading SSDTs from variables efi: Correct comment on efi_memmap_alloc memblock: Disable mirror feature if kernelcore is not specified ...