kernel.git/arch/arm64/mm, branch linux-5.16.y

arm64: mm: Drop 'const' from conditional arm64_dma_phys_limit definition

2022-04-08T12:06:53Z

commit 770093459b9b333380aa71f2c31c60b14895c1df upstream. Commit 031495635b46 ("arm64: Do not defer reserve_crashkernel() for platforms with no DMA memory zones") introduced different definitions for 'arm64_dma_phys_limit' depending on CONFIG_ZONE_DMA{,32} based on a late suggestion from Pasha. Sadly, this results in a build error when passing W=1: | arch/arm64/mm/init.c:90:19: error: conflicting type qualifiers for 'arm64_dma_phys_limit' Drop the 'const' for now and use '__ro_after_init' consistently. Link: https://lore.kernel.org/r/202203090241.aj7paWeX-lkp@intel.com Link: https://lore.kernel.org/r/CA+CK2bDbbx=8R=UthkMesWOST8eJMtOGJdfMRTFSwVmo0Vn0EA@mail.gmail.com Fixes: 031495635b46 ("arm64: Do not defer reserve_crashkernel() for platforms with no DMA memory zones") Signed-off-by: Will Deacon Signed-off-by: Greg Kroah-Hartman

arm64/mm: avoid fixmap race condition when create pud mapping

2022-04-08T12:05:38Z

[ Upstream commit ee017ee353506fcec58e481673e4331ff198a80e ] The 'fixmap' is a global resource and is used recursively by create pud mapping(), leading to a potential race condition in the presence of a concurrent call to alloc_init_pud(): kernel_init thread virtio-mem workqueue thread ================== =========================== alloc_init_pud(...) alloc_init_pud(...) pudp = pud_set_fixmap_offset(...) pudp = pud_set_fixmap_offset(...) READ_ONCE(*pudp) pud_clear_fixmap(...) READ_ONCE(*pudp) // CRASH! As kernel may sleep during creating pud mapping, introduce a mutex lock to serialise use of the fixmap entries by alloc_init_pud(). However, there is no need for locking in early boot stage and it doesn't work well with KASLR enabled when early boot. So, enable lock when system_state doesn't equal to "SYSTEM_BOOTING". Signed-off-by: Jianyong Wu Reviewed-by: Catalin Marinas Fixes: f4710445458c ("arm64: mm: use fixmap when creating page tables") Link: https://lore.kernel.org/r/20220201114400.56885-1-jianyong.wu@arm.com Signed-off-by: Will Deacon Signed-off-by: Sasha Levin

arm64: Do not defer reserve_crashkernel() for platforms with no DMA memory zones

2022-04-08T12:05:29Z

commit 031495635b4668f94e964e037ca93d0d38bfde58 upstream. The following patches resulted in deferring crash kernel reservation to mem_init(), mainly aimed at platforms with DMA memory zones (no IOMMU), in particular Raspberry Pi 4. commit 1a8e1cef7603 ("arm64: use both ZONE_DMA and ZONE_DMA32") commit 8424ecdde7df ("arm64: mm: Set ZONE_DMA size based on devicetree's dma-ranges") commit 0a30c53573b0 ("arm64: mm: Move reserve_crashkernel() into mem_init()") commit 2687275a5843 ("arm64: Force NO_BLOCK_MAPPINGS if crashkernel reservation is required") Above changes introduced boot slowdown due to linear map creation for all the memory banks with NO_BLOCK_MAPPINGS, see discussion[1]. The proposed changes restore crash kernel reservation to earlier behavior thus avoids slow boot, particularly for platforms with IOMMU (no DMA memory zones). Tested changes to confirm no ~150ms boot slowdown on our SoC with IOMMU and 8GB memory. Also tested with ZONE_DMA and/or ZONE_DMA32 configs to confirm no regression to deferring scheme of crash kernel memory reservation. In both cases successfully collected kernel crash dump. [1] https://lore.kernel.org/all/9436d033-579b-55fa-9b00-6f4b661c2dd7@linux.microsoft.com/ Signed-off-by: Vijay Balakrishna Cc: stable@vger.kernel.org Reviewed-by: Pasha Tatashin Link: https://lore.kernel.org/r/1646242689-20744-1-git-send-email-vijayb@linux.microsoft.com [will: Add #ifdef CONFIG_KEXEC_CORE guards to fix 'crashk_res' references in allnoconfig build] Signed-off-by: Will Deacon Signed-off-by: Greg Kroah-Hartman

arm64: Ensure execute-only permissions are not allowed without EPAN

2022-03-16T13:26:50Z

commit 6e2edd6371a497a6350bb735534c9bda2a31f43d upstream. Commit 18107f8a2df6 ("arm64: Support execute-only permissions with Enhanced PAN") re-introduced execute-only permissions when EPAN is available. When EPAN is not available, arch_filter_pgprot() is supposed to change a PAGE_EXECONLY permission into PAGE_READONLY_EXEC. However, if BTI or MTE are present, such check does not detect the execute-only pgprot in the presence of PTE_GP (BTI) or MT_NORMAL_TAGGED (MTE), allowing the user to request PROT_EXEC with PROT_BTI or PROT_MTE. Remove the arch_filter_pgprot() function, change the default VM_EXEC permissions to PAGE_READONLY_EXEC and update the protection_map[] array at core_initcall() if EPAN is detected. Signed-off-by: Catalin Marinas Fixes: 18107f8a2df6 ("arm64: Support execute-only permissions with Enhanced PAN") Cc: # 5.13.x Acked-by: Will Deacon Reviewed-by: Vladimir Murzin Tested-by: Vladimir Murzin Signed-off-by: Greg Kroah-Hartman

arm64: entry: Allow the trampoline text to occupy multiple pages

2022-03-11T11:42:07Z

commit a9c406e6462ff14956d690de7bbe5131a5677dc9 upstream. Adding a second set of vectors to .entry.tramp.text will make it larger than a single 4K page. Allow the trampoline text to occupy up to three pages by adding two more fixmap slots. Previous changes to tramp_valias allowed it to reach beyond a single page. Reviewed-by: Catalin Marinas Signed-off-by: James Morse Signed-off-by: Greg Kroah-Hartman

arm64: extable: fix load_unaligned_zeropad() reg indices

2022-02-01T16:29:06Z

commit 3758a6c74e08bdc15ccccd6872a6ad37d165239a upstream. In ex_handler_load_unaligned_zeropad() we erroneously extract the data and addr register indices from ex->type rather than ex->data. As ex->type will contain EX_TYPE_LOAD_UNALIGNED_ZEROPAD (i.e. 4): * We'll always treat X0 as the address register, since EX_DATA_REG_ADDR is extracted from bits [9:5]. Thus, we may attempt to dereference an arbitrary address as X0 may hold an arbitrary value. * We'll always treat X4 as the data register, since EX_DATA_REG_DATA is extracted from bits [4:0]. Thus we will corrupt X4 and cause arbitrary behaviour within load_unaligned_zeropad() and its caller. Fix this by extracting both values from ex->data as originally intended. On an MTE-enabled QEMU image we are hitting the following crash: Unable to handle kernel NULL pointer dereference at virtual address 0000000000000000 Call trace: fixup_exception+0xc4/0x108 __do_kernel_fault+0x3c/0x268 do_tag_check_fault+0x3c/0x104 do_mem_abort+0x44/0xf4 el1_abort+0x40/0x64 el1h_64_sync_handler+0x60/0xa0 el1h_64_sync+0x7c/0x80 link_path_walk+0x150/0x344 path_openat+0xa0/0x7dc do_filp_open+0xb8/0x168 do_sys_openat2+0x88/0x17c __arm64_sys_openat+0x74/0xa0 invoke_syscall+0x48/0x148 el0_svc_common+0xb8/0xf8 do_el0_svc+0x28/0x88 el0_svc+0x24/0x84 el0t_64_sync_handler+0x88/0xec el0t_64_sync+0x1b4/0x1b8 Code: f8695a69 71007d1f 540000e0 927df12a (f940014a) Fixes: 753b32368705 ("arm64: extable: add load_unaligned_zeropad() handler") Cc: # 5.16.x Reviewed-by: Mark Rutland Signed-off-by: Evgenii Stepanov Link: https://lore.kernel.org/r/20220125182217.2605202-1-eugenis@google.com Signed-off-by: Catalin Marinas Signed-off-by: Greg Kroah-Hartman

arm64/bpf: Remove 128MB limit for BPF JIT programs

2022-01-29T09:59:24Z

commit b89ddf4cca43f1269093942cf5c4e457fd45c335 upstream. Commit 91fc957c9b1d ("arm64/bpf: don't allocate BPF JIT programs in module memory") restricts BPF JIT program allocation to a 128MB region to ensure BPF programs are still in branching range of each other. However this restriction should not apply to the aarch64 JIT, since BPF_JMP | BPF_CALL are implemented as a 64-bit move into a register and then a BLR instruction - which has the effect of being able to call anything without proximity limitation. The practical reason to relax this restriction on JIT memory is that 128MB of JIT memory can be quickly exhausted, especially where PAGE_SIZE is 64KB - one page is needed per program. In cases where seccomp filters are applied to multiple VMs on VM launch - such filters are classic BPF but converted to BPF - this can severely limit the number of VMs that can be launched. In a world where we support BPF JIT always on, turning off the JIT isn't always an option either. Fixes: 91fc957c9b1d ("arm64/bpf: don't allocate BPF JIT programs in module memory") Suggested-by: Ard Biesheuvel Signed-off-by: Russell King Signed-off-by: Daniel Borkmann Tested-by: Alan Maguire Link: https://lore.kernel.org/bpf/1636131046-5982-2-git-send-email-alan.maguire@oracle.com Signed-off-by: Greg Kroah-Hartman

kasan: add kasan mode messages when kasan init

2021-11-11T17:34:35Z

There are multiple kasan modes. It makes sense that we add some messages to know which kasan mode is active when booting up [1]. Link: https://bugzilla.kernel.org/show_bug.cgi?id=212195 [1] Link: https://lkml.kernel.org/r/20211020094850.4113-1-Kuan-Ying.Lee@mediatek.com Signed-off-by: Kuan-Ying Lee Reviewed-by: Marco Elver Reviewed-by: David Hildenbrand Cc: Andrey Ryabinin Cc: Alexander Potapenko Cc: Andrey Konovalov Cc: Dmitry Vyukov Cc: Catalin Marinas Cc: Will Deacon Cc: Matthias Brugger Cc: Chinwen Chang Cc: Yee Lee Cc: Nicholas Tang Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds

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

2021-11-10T19:29:30Z

Pull arm64 fixes from Will Deacon: - Fix double-evaluation of 'pte' macro argument when using 52-bit PAs - Fix signedness of some MTE prctl PR_* constants - Fix kmemleak memory usage by skipping early pgtable allocations - Fix printing of CPU feature register strings - Remove redundant -nostdlib linker flag for vDSO binaries * tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux: arm64: pgtable: make __pte_to_phys/__phys_to_pte_val inline functions arm64: Track no early_pgtable_alloc() for kmemleak arm64: mte: change PR_MTE_TCF_NONE back into an unsigned long arm64: vdso: remove -nostdlib compiler flag arm64: arm64_ftr_reg->name may not be a human-readable string

arm64: Track no early_pgtable_alloc() for kmemleak

2021-11-08T10:05:22Z

After switched page size from 64KB to 4KB on several arm64 servers here, kmemleak starts to run out of early memory pool due to a huge number of those early_pgtable_alloc() calls: kmemleak_alloc_phys() memblock_alloc_range_nid() memblock_phys_alloc_range() early_pgtable_alloc() init_pmd() alloc_init_pud() __create_pgd_mapping() __map_memblock() paging_init() setup_arch() start_kernel() Increased the default value of DEBUG_KMEMLEAK_MEM_POOL_SIZE by 4 times won't be enough for a server with 200GB+ memory. There isn't much interesting to check memory leaks for those early page tables and those early memory mappings should not reference to other memory. Hence, no kmemleak false positives, and we can safely skip tracking those early allocations from kmemleak like we did in the commit fed84c785270 ("mm/memblock.c: skip kmemleak for kasan_init()") without needing to introduce complications to automatically scale the value depends on the runtime memory size etc. After the patch, the default value of DEBUG_KMEMLEAK_MEM_POOL_SIZE becomes sufficient again. Signed-off-by: Qian Cai Reviewed-by: Catalin Marinas Reviewed-by: Mike Rapoport Link: https://lore.kernel.org/r/20211105150509.7826-1-quic_qiancai@quicinc.com Signed-off-by: Will Deacon