kernel.git/arch/arm64/mm, branch linux-5.12.y

arm64: mm: don't use CON and BLK mapping if KFENCE is enabled

2021-06-03T07:09:52Z

commit e69012400b0cb42b2070748322cb72f9effec00f upstream. When we added KFENCE support for arm64, we intended that it would force the entire linear map to be mapped at page granularity, but we only enforced this in arch_add_memory() and not in map_mem(), so memory mapped at boot time can be mapped at a larger granularity. When booting a kernel with KFENCE=y and RODATA_FULL=n, this results in the following WARNING at boot: [ 0.000000] ------------[ cut here ]------------ [ 0.000000] WARNING: CPU: 0 PID: 0 at mm/memory.c:2462 apply_to_pmd_range+0xec/0x190 [ 0.000000] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 5.13.0-rc1+ #10 [ 0.000000] Hardware name: linux,dummy-virt (DT) [ 0.000000] pstate: 600000c5 (nZCv daIF -PAN -UAO -TCO BTYPE=--) [ 0.000000] pc : apply_to_pmd_range+0xec/0x190 [ 0.000000] lr : __apply_to_page_range+0x94/0x170 [ 0.000000] sp : ffffffc010573e20 [ 0.000000] x29: ffffffc010573e20 x28: ffffff801f400000 x27: ffffff801f401000 [ 0.000000] x26: 0000000000000001 x25: ffffff801f400fff x24: ffffffc010573f28 [ 0.000000] x23: ffffffc01002b710 x22: ffffffc0105fa450 x21: ffffffc010573ee4 [ 0.000000] x20: ffffff801fffb7d0 x19: ffffff801f401000 x18: 00000000fffffffe [ 0.000000] x17: 000000000000003f x16: 000000000000000a x15: ffffffc01060b940 [ 0.000000] x14: 0000000000000000 x13: 0098968000000000 x12: 0000000098968000 [ 0.000000] x11: 0000000000000000 x10: 0000000098968000 x9 : 0000000000000001 [ 0.000000] x8 : 0000000000000000 x7 : ffffffc010573ee4 x6 : 0000000000000001 [ 0.000000] x5 : ffffffc010573f28 x4 : ffffffc01002b710 x3 : 0000000040000000 [ 0.000000] x2 : ffffff801f5fffff x1 : 0000000000000001 x0 : 007800005f400705 [ 0.000000] Call trace: [ 0.000000] apply_to_pmd_range+0xec/0x190 [ 0.000000] __apply_to_page_range+0x94/0x170 [ 0.000000] apply_to_page_range+0x10/0x20 [ 0.000000] __change_memory_common+0x50/0xdc [ 0.000000] set_memory_valid+0x30/0x40 [ 0.000000] kfence_init_pool+0x9c/0x16c [ 0.000000] kfence_init+0x20/0x98 [ 0.000000] start_kernel+0x284/0x3f8 Fixes: 840b23986344 ("arm64, kfence: enable KFENCE for ARM64") Cc: # 5.12.x Signed-off-by: Jisheng Zhang Acked-by: Mark Rutland Acked-by: Marco Elver Tested-by: Marco Elver Link: https://lore.kernel.org/r/20210525104551.2ec37f77@xhacker.debian Signed-off-by: Catalin Marinas Signed-off-by: Greg Kroah-Hartman

arm64: Fix race condition on PG_dcache_clean in __sync_icache_dcache()

2021-05-19T08:56:34Z

commit 588a513d34257fdde95a9f0df0202e31998e85c6 upstream. To ensure that instructions are observable in a new mapping, the arm64 set_pte_at() implementation cleans the D-cache and invalidates the I-cache to the PoU. As an optimisation, this is only done on executable mappings and the PG_dcache_clean page flag is set to avoid future cache maintenance on the same page. When two different processes map the same page (e.g. private executable file or shared mapping) there's a potential race on checking and setting PG_dcache_clean via set_pte_at() -> __sync_icache_dcache(). While on the fault paths the page is locked (PG_locked), mprotect() does not take the page lock. The result is that one process may see the PG_dcache_clean flag set but the I/D cache maintenance not yet performed. Avoid test_and_set_bit(PG_dcache_clean) in favour of separate test_bit() and set_bit(). In the rare event of a race, the cache maintenance is done twice. Signed-off-by: Catalin Marinas Cc: Cc: Will Deacon Cc: Steven Price Reviewed-by: Steven Price Acked-by: Will Deacon Link: https://lore.kernel.org/r/20210514095001.13236-1-catalin.marinas@arm.com Signed-off-by: Catalin Marinas Signed-off-by: Greg Kroah-Hartman

arm64: mte: initialize RGSR_EL1.SEED in __cpu_setup

2021-05-19T08:56:34Z

commit 37a8024d265564eba680575df6421f19db21dfce upstream. A valid implementation choice for the ChooseRandomNonExcludedTag() pseudocode function used by IRG is to behave in the same way as with GCR_EL1.RRND=0. This would mean that RGSR_EL1.SEED is used as an LFSR which must have a non-zero value in order for IRG to properly produce pseudorandom numbers. However, RGSR_EL1 is reset to an UNKNOWN value on soft reset and thus may reset to 0. Therefore we must initialize RGSR_EL1.SEED to a non-zero value in order to ensure that IRG behaves as expected. Signed-off-by: Peter Collingbourne Fixes: 3b714d24ef17 ("arm64: mte: CPU feature detection and initial sysreg configuration") Cc: # 5.10 Link: https://linux-review.googlesource.com/id/I2b089b6c7d6f17ee37e2f0db7df5ad5bcc04526c Acked-by: Mark Rutland Link: https://lore.kernel.org/r/20210507185905.1745402-1-pcc@google.com Signed-off-by: Catalin Marinas Signed-off-by: Greg Kroah-Hartman

arm64: mm: correct the inside linear map range during hotplug check

2021-03-22T12:47:40Z

Memory hotplug may fail on systems with CONFIG_RANDOMIZE_BASE because the linear map range is not checked correctly. The start physical address that linear map covers can be actually at the end of the range because of randomization. Check that and if so reduce it to 0. This can be verified on QEMU with setting kaslr-seed to ~0ul: memstart_offset_seed = 0xffff START: __pa(_PAGE_OFFSET(vabits_actual)) = ffff9000c0000000 END: __pa(PAGE_END - 1) = 1000bfffffff Signed-off-by: Pavel Tatashin Fixes: 58284a901b42 ("arm64/mm: Validate hotplug range before creating linear mapping") Tested-by: Tyler Hicks Reviewed-by: Anshuman Khandual Link: https://lore.kernel.org/r/20210216150351.129018-2-pasha.tatashin@soleen.com Signed-off-by: Will Deacon

arm64: mm: use a 48-bit ID map when possible on 52-bit VA builds

2021-03-11T13:04:28Z

52-bit VA kernels can run on hardware that is only 48-bit capable, but configure the ID map as 52-bit by default. This was not a problem until recently, because the special T0SZ value for a 52-bit VA space was never programmed into the TCR register anwyay, and because a 52-bit ID map happens to use the same number of translation levels as a 48-bit one. This behavior was changed by commit 1401bef703a4 ("arm64: mm: Always update TCR_EL1 from __cpu_set_tcr_t0sz()"), which causes the unsupported T0SZ value for a 52-bit VA to be programmed into TCR_EL1. While some hardware simply ignores this, Mark reports that Amberwing systems choke on this, resulting in a broken boot. But even before that commit, the unsupported idmap_t0sz value was exposed to KVM and used to program TCR_EL2 incorrectly as well. Given that we already have to deal with address spaces being either 48-bit or 52-bit in size, the cleanest approach seems to be to simply default to a 48-bit VA ID map, and only switch to a 52-bit one if the placement of the kernel in DRAM requires it. This is guaranteed not to happen unless the system is actually 52-bit VA capable. Fixes: 90ec95cda91a ("arm64: mm: Introduce VA_BITS_MIN") Reported-by: Mark Salter Link: http://lore.kernel.org/r/20210310003216.410037-1-msalter@redhat.com Signed-off-by: Ard Biesheuvel Link: https://lore.kernel.org/r/20210310171515.416643-2-ardb@kernel.org Signed-off-by: Will Deacon

arm64: mte: Map hotplugged memory as Normal Tagged

2021-03-10T10:56:46Z

In a system supporting MTE, the linear map must allow reading/writing allocation tags by setting the memory type as Normal Tagged. Currently, this is only handled for memory present at boot. Hotplugged memory uses Normal non-Tagged memory. Introduce pgprot_mhp() for hotplugged memory and use it in add_memory_resource(). The arm64 code maps pgprot_mhp() to pgprot_tagged(). Note that ZONE_DEVICE memory should not be mapped as Tagged and therefore setting the memory type in arch_add_memory() is not feasible. Signed-off-by: Catalin Marinas Fixes: 0178dc761368 ("arm64: mte: Use Normal Tagged attributes for the linear map") Reported-by: Patrick Daly Tested-by: Patrick Daly Link: https://lore.kernel.org/r/1614745263-27827-1-git-send-email-pdaly@codeaurora.org Cc: # 5.10.x Cc: Will Deacon Cc: Andrew Morton Cc: Vincenzo Frascino Cc: David Hildenbrand Reviewed-by: David Hildenbrand Reviewed-by: Vincenzo Frascino Reviewed-by: Anshuman Khandual Link: https://lore.kernel.org/r/20210309122601.5543-1-catalin.marinas@arm.com Signed-off-by: Will Deacon

arm64/mm: Reorganize pfn_valid()

2021-03-08T18:04:00Z

There are multiple instances of pfn_to_section_nr() and __pfn_to_section() when CONFIG_SPARSEMEM is enabled. This can be optimized if memory section is fetched earlier. This replaces the open coded PFN and ADDR conversion with PFN_PHYS() and PHYS_PFN() helpers. While there, also add a comment. This does not cause any functional change. Cc: Catalin Marinas Cc: Will Deacon Cc: Ard Biesheuvel Cc: linux-arm-kernel@lists.infradead.org Cc: linux-kernel@vger.kernel.org Reviewed-by: David Hildenbrand Signed-off-by: Anshuman Khandual Acked-by: Catalin Marinas Link: https://lore.kernel.org/r/1614921898-4099-3-git-send-email-anshuman.khandual@arm.com Signed-off-by: Will Deacon

arm64/mm: Fix pfn_valid() for ZONE_DEVICE based memory

2021-03-08T18:04:00Z

pfn_valid() validates a pfn but basically it checks for a valid struct page backing for that pfn. It should always return positive for memory ranges backed with struct page mapping. But currently pfn_valid() fails for all ZONE_DEVICE based memory types even though they have struct page mapping. pfn_valid() asserts that there is a memblock entry for a given pfn without MEMBLOCK_NOMAP flag being set. The problem with ZONE_DEVICE based memory is that they do not have memblock entries. Hence memblock_is_map_memory() will invariably fail via memblock_search() for a ZONE_DEVICE based address. This eventually fails pfn_valid() which is wrong. memblock_is_map_memory() needs to be skipped for such memory ranges. As ZONE_DEVICE memory gets hotplugged into the system via memremap_pages() called from a driver, their respective memory sections will not have SECTION_IS_EARLY set. Normal hotplug memory will never have MEMBLOCK_NOMAP set in their memblock regions. Because the flag MEMBLOCK_NOMAP was specifically designed and set for firmware reserved memory regions. memblock_is_map_memory() can just be skipped as its always going to be positive and that will be an optimization for the normal hotplug memory. Like ZONE_DEVICE based memory, all normal hotplugged memory too will not have SECTION_IS_EARLY set for their sections Skipping memblock_is_map_memory() for all non early memory sections would fix pfn_valid() problem for ZONE_DEVICE based memory and also improve its performance for normal hotplug memory as well. Cc: Catalin Marinas Cc: Will Deacon Cc: Ard Biesheuvel Cc: Robin Murphy Cc: linux-arm-kernel@lists.infradead.org Cc: linux-kernel@vger.kernel.org Acked-by: David Hildenbrand Fixes: 73b20c84d42d ("arm64: mm: implement pte_devmap support") Signed-off-by: Anshuman Khandual Acked-by: Catalin Marinas Link: https://lore.kernel.org/r/1614921898-4099-2-git-send-email-anshuman.khandual@arm.com Signed-off-by: Will Deacon

Merge tag 'riscv-for-linus-5.12-mw0' of git://git.kernel.org/pub/scm/linux/kernel/git/riscv/linux

2021-02-26T18:28:35Z

Pull RISC-V updates from Palmer Dabbelt: "A handful of new RISC-V related patches for this merge window: - A check to ensure drivers are properly using uaccess. This isn't manifesting with any of the drivers I'm currently using, but may catch errors in new drivers. - Some preliminary support for the FU740, along with the HiFive Unleashed it will appear on. - NUMA support for RISC-V, which involves making the arm64 code generic. - Support for kasan on the vmalloc region. - A handful of new drivers for the Kendryte K210, along with the DT plumbing required to boot on a handful of K210-based boards. - Support for allocating ASIDs. - Preliminary support for kernels larger than 128MiB. - Various other improvements to our KASAN support, including the utilization of huge pages when allocating the KASAN regions. We may have already found a bug with the KASAN_VMALLOC code, but it's passing my tests. There's a fix in the works, but that will probably miss the merge window. * tag 'riscv-for-linus-5.12-mw0' of git://git.kernel.org/pub/scm/linux/kernel/git/riscv/linux: (75 commits) riscv: Improve kasan population by using hugepages when possible riscv: Improve kasan population function riscv: Use KASAN_SHADOW_INIT define for kasan memory initialization riscv: Improve kasan definitions riscv: Get rid of MAX_EARLY_MAPPING_SIZE soc: canaan: Sort the Makefile alphabetically riscv: Disable KSAN_SANITIZE for vDSO riscv: Remove unnecessary declaration riscv: Add Canaan Kendryte K210 SD card defconfig riscv: Update Canaan Kendryte K210 defconfig riscv: Add Kendryte KD233 board device tree riscv: Add SiPeed MAIXDUINO board device tree riscv: Add SiPeed MAIX GO board device tree riscv: Add SiPeed MAIX DOCK board device tree riscv: Add SiPeed MAIX BiT board device tree riscv: Update Canaan Kendryte K210 device tree dt-bindings: add resets property to dw-apb-timer dt-bindings: fix sifive gpio properties dt-bindings: update sifive uart compatible string dt-bindings: update sifive clint compatible string ...

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

2021-02-26T18:19:03Z

Pull arm64 fixes from Will Deacon: "The big one is a fix for the VHE enabling path during early boot, where the code enabling the MMU wasn't necessarily in the identity map of the new page-tables, resulting in a consistent crash with 64k pages. In fixing that, we noticed some missing barriers too, so we added those for the sake of architectural compliance. Other than that, just the usual merge window trickle. There'll be more to come, too. Summary: - Fix lockdep false alarm on resume-from-cpuidle path - Fix memory leak in kexec_file - Fix module linker script to work with GDB - Fix error code when trying to use uprobes with AArch32 instructions - Fix late VHE enabling with 64k pages - Add missing ISBs after TLB invalidation - Fix seccomp when tracing syscall -1 - Fix stacktrace return code at end of stack - Fix inconsistent whitespace for pointer return values - Fix compiler warnings when building with W=1" * tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux: arm64: stacktrace: Report when we reach the end of the stack arm64: ptrace: Fix seccomp of traced syscall -1 (NO_SYSCALL) arm64: Add missing ISB after invalidating TLB in enter_vhe arm64: Add missing ISB after invalidating TLB in __primary_switch arm64: VHE: Enable EL2 MMU from the idmap KVM: arm64: make the hyp vector table entries local arm64/mm: Fixed some coding style issues arm64: uprobe: Return EOPNOTSUPP for AARCH32 instruction probing kexec: move machine_kexec_post_load() to public interface arm64 module: set plt* section addresses to 0x0 arm64: kexec_file: fix memory leakage in create_dtb() when fdt_open_into() fails arm64: spectre: Prevent lockdep splat on v4 mitigation enable path