path: root/arch
AgeCommit message (Collapse)Author
2021-06-26Merge tag 's390-5.13-5' of ↵Linus Torvalds
git:// Pull s390 fixes from Vasily Gorbik: - Fix a couple of late pt_regs flags handling findings of conversion to generic entry. - Fix potential register clobbering in stack switch helper. - Fix thread/group masks for offline cpus. - Fix cleanup of mdev resources when remove callback is invoked in vfio-ap code. * tag 's390-5.13-5' of git:// s390/stack: fix possible register corruption with stack switch helper s390/topology: clear thread/group maps for offline cpus s390/vfio-ap: clean up mdev resources when remove callback invoked s390: clear pt_regs::flags on irq entry s390: fix system call restart with multiple signals
2021-06-25Merge branch 'akpm' (patches from Andrew)Linus Torvalds
Merge misc fixes from Andrew Morton: "24 patches, based on 4a09d388f2ab382f217a764e6a152b3f614246f6. Subsystems affected by this patch series: mm (thp, vmalloc, hugetlb, memory-failure, and pagealloc), nilfs2, kthread, MAINTAINERS, and mailmap" * emailed patches from Andrew Morton <>: (24 commits) mailmap: add Marek's other e-mail address and identity without diacritics MAINTAINERS: fix Marek's identity again mm/page_alloc: do bulk array bounds check after checking populated elements mm/page_alloc: __alloc_pages_bulk(): do bounds check before accessing array mm/hwpoison: do not lock page again when me_huge_page() successfully recovers mm,hwpoison: return -EHWPOISON to denote that the page has already been poisoned mm/memory-failure: use a mutex to avoid memory_failure() races mm, futex: fix shared futex pgoff on shmem huge page kthread: prevent deadlock when kthread_mod_delayed_work() races with kthread_cancel_delayed_work_sync() kthread_worker: split code for canceling the delayed work timer mm/vmalloc: unbreak kasan vmalloc support KVM: s390: prepare for hugepage vmalloc mm/vmalloc: add vmalloc_no_huge nilfs2: fix memory leak in nilfs_sysfs_delete_device_group mm/thp: another PVMW_SYNC fix in page_vma_mapped_walk() mm/thp: fix page_vma_mapped_walk() if THP mapped by ptes mm: page_vma_mapped_walk(): get vma_address_end() earlier mm: page_vma_mapped_walk(): use goto instead of while (1) mm: page_vma_mapped_walk(): add a level of indentation mm: page_vma_mapped_walk(): crossing page table boundary ...
2021-06-25Merge tag 'x86_urgent_for_v5.13' of ↵Linus Torvalds
git:// Pull x86 fixes from Borislav Petkov: "Two more urgent FPU fixes: - prevent unprivileged userspace from reinitializing supervisor states - prepare init_fpstate, which is the buffer used when initializing FPU state, properly in case the skip-writing-state-components XSAVE* variants are used" * tag 'x86_urgent_for_v5.13' of git:// x86/fpu: Make init_fpstate correct with optimized XSAVE x86/fpu: Preserve supervisor states in sanitize_restored_user_xstate()
2021-06-24KVM: s390: prepare for hugepage vmallocClaudio Imbrenda
The Create Secure Configuration Ultravisor Call does not support using large pages for the virtual memory area. This is a hardware limitation. This patch replaces the vzalloc call with an almost equivalent call to the newly introduced vmalloc_no_huge function, which guarantees that only small pages will be used for the backing. The new call will not clear the allocated memory, but that has never been an actual requirement. Link: Fixes: 121e6f3258fe3 ("mm/vmalloc: hugepage vmalloc mappings") Signed-off-by: Claudio Imbrenda <> Reviewed-by: Janosch Frank <> Acked-by: Christian Borntraeger <> Acked-by: Nicholas Piggin <> Reviewed-by: David Hildenbrand <> Cc: Nicholas Piggin <> Cc: Uladzislau Rezki (Sony) <> Cc: Catalin Marinas <> Cc: Thomas Gleixner <> Cc: Ingo Molnar <> Cc: David Rientjes <> Cc: Christoph Hellwig <> Signed-off-by: Andrew Morton <> Signed-off-by: Linus Torvalds <>
2021-06-24Merge tag 'perf-urgent-2021-06-24' of ↵Linus Torvalds
git:// Pull x86 perf fix from Ingo Molnar: "An LBR buffer fix for code that probably only worked accidentally" * tag 'perf-urgent-2021-06-24' of git:// perf/x86/intel/lbr: Zero the xstate buffer on allocation
2021-06-24Merge tag 'objtool-urgent-2021-06-24' of ↵Linus Torvalds
git:// Pull objtool fixes from Ingo Molnar: "Address a number of objtool warnings that got reported. No change in behavior intended, but code generation might be impacted by commit 1f008d46f124 ("x86: Always inline task_size_max()")" * tag 'objtool-urgent-2021-06-24' of git:// locking/lockdep: Improve noinstr vs errors x86: Always inline task_size_max() x86/xen: Fix noinstr fail in exc_xen_unknown_trap() x86/xen: Fix noinstr fail in xen_pv_evtchn_do_upcall() x86/entry: Fix noinstr fail in __do_fast_syscall_32() objtool/x86: Ignore __x86_indirect_alt_* symbols
2021-06-24perf/x86/intel/lbr: Zero the xstate buffer on allocationThomas Gleixner
XRSTORS requires a valid xstate buffer to work correctly. XSAVES does not guarantee to write a fully valid buffer according to the SDM: "XSAVES does not write to any parts of the XSAVE header other than the XSTATE_BV and XCOMP_BV fields." XRSTORS triggers a #GP: "If bytes 63:16 of the XSAVE header are not all zero." It's dubious at best how this can work at all when the buffer is not zeroed before use. Allocate the buffers with __GFP_ZERO to prevent XRSTORS failure. Fixes: ce711ea3cab9 ("perf/x86/intel/lbr: Support XSAVES/XRSTORS for LBR context switch") Signed-off-by: Thomas Gleixner <> Signed-off-by: Ingo Molnar <> Link:
2021-06-22x86: Always inline task_size_max()Peter Zijlstra
Fix: vmlinux.o: warning: objtool: handle_bug()+0x10: call to task_size_max() leaves .noinstr.text section When #UD isn't a BUG, we shouldn't violate noinstr (we'll still probably die, but that's another story). Fixes: 025768a966a3 ("x86/cpu: Use alternative to generate the TASK_SIZE_MAX constant") Signed-off-by: Peter Zijlstra (Intel) <> Signed-off-by: Ingo Molnar <> Link:
2021-06-22x86/xen: Fix noinstr fail in exc_xen_unknown_trap()Peter Zijlstra
Fix: vmlinux.o: warning: objtool: exc_xen_unknown_trap()+0x7: call to printk() leaves .noinstr.text section Fixes: 2e92493637a0 ("x86/xen: avoid warning in Xen pv guest with CONFIG_AMD_MEM_ENCRYPT enabled") Signed-off-by: Peter Zijlstra (Intel) <> Signed-off-by: Ingo Molnar <> Link:
2021-06-22x86/xen: Fix noinstr fail in xen_pv_evtchn_do_upcall()Peter Zijlstra
Fix: vmlinux.o: warning: objtool: xen_pv_evtchn_do_upcall()+0x23: call to irq_enter_rcu() leaves .noinstr.text section Fixes: 359f01d1816f ("x86/entry: Use run_sysvec_on_irqstack_cond() for XEN upcall") Signed-off-by: Peter Zijlstra (Intel) <> Signed-off-by: Ingo Molnar <> Link:
2021-06-22x86/entry: Fix noinstr fail in __do_fast_syscall_32()Peter Zijlstra
Fix: vmlinux.o: warning: objtool: __do_fast_syscall_32()+0xf5: call to trace_hardirqs_off() leaves .noinstr.text section Fixes: 5d5675df792f ("x86/entry: Fix entry/exit mismatch on failed fast 32-bit syscalls") Signed-off-by: Peter Zijlstra (Intel) <> Signed-off-by: Ingo Molnar <> Link:
2021-06-22x86/fpu: Make init_fpstate correct with optimized XSAVEThomas Gleixner
The XSAVE init code initializes all enabled and supported components with XRSTOR(S) to init state. Then it XSAVEs the state of the components back into init_fpstate which is used in several places to fill in the init state of components. This works correctly with XSAVE, but not with XSAVEOPT and XSAVES because those use the init optimization and skip writing state of components which are in init state. So init_fpstate.xsave still contains all zeroes after this operation. There are two ways to solve that: 1) Use XSAVE unconditionally, but that requires to reshuffle the buffer when XSAVES is enabled because XSAVES uses compacted format. 2) Save the components which are known to have a non-zero init state by other means. Looking deeper, #2 is the right thing to do because all components the kernel supports have all-zeroes init state except the legacy features (FP, SSE). Those cannot be hard coded because the states are not identical on all CPUs, but they can be saved with FXSAVE which avoids all conditionals. Use FXSAVE to save the legacy FP/SSE components in init_fpstate along with a BUILD_BUG_ON() which reminds developers to validate that a newly added component has all zeroes init state. As a bonus remove the now unused copy_xregs_to_kernel_booting() crutch. The XSAVE and reshuffle method can still be implemented in the unlikely case that components are added which have a non-zero init state and no other means to save them. For now, FXSAVE is just simple and good enough. [ bp: Fix a typo or two in the text. ] Fixes: 6bad06b76892 ("x86, xsave: Use xsaveopt in context-switch path when supported") Signed-off-by: Thomas Gleixner <> Signed-off-by: Borislav Petkov <> Reviewed-by: Borislav Petkov <> Cc: Link:
2021-06-22x86/fpu: Preserve supervisor states in sanitize_restored_user_xstate()Thomas Gleixner
sanitize_restored_user_xstate() preserves the supervisor states only when the fx_only argument is zero, which allows unprivileged user space to put supervisor states back into init state. Preserve them unconditionally. [ bp: Fix a typo or two in the text. ] Fixes: 5d6b6a6f9b5c ("x86/fpu/xstate: Update sanitize_restored_xstate() for supervisor xstates") Signed-off-by: Thomas Gleixner <> Signed-off-by: Borislav Petkov <> Cc: Link:
2021-06-21Merge tag 'for-linus' of git:// Torvalds
Pull ARM fix from Russell King: - fix gcc 10 compiler regression with cpu_init() * tag 'for-linus' of git:// ARM: 9081/1: fix gcc-10 thumb2-kernel regression
2021-06-21objtool/x86: Ignore __x86_indirect_alt_* symbolsPeter Zijlstra
Because the __x86_indirect_alt* symbols are just that, objtool will try and validate them as regular symbols, instead of the alternative replacements that they are. This goes sideways for FRAME_POINTER=y builds; which generate a fair amount of warnings. Fixes: 9bc0bb50727c ("objtool/x86: Rewrite retpoline thunk calls") Signed-off-by: Peter Zijlstra (Intel) <> Signed-off-by: Ingo Molnar <> Link:
2021-06-21s390/stack: fix possible register corruption with stack switch helperHeiko Carstens
The CALL_ON_STACK macro is used to call a C function from inline assembly, and therefore must consider the C ABI, which says that only registers 6-13, and 15 are non-volatile (restored by the called function). The inline assembly incorrectly marks all registers used to pass parameters to the called function as read-only input operands, instead of operands that are read and written to. This might result in register corruption depending on usage, compiler, and compile options. Fix this by marking all operands used to pass parameters as read/write operands. To keep the code simple even register 6, if used, is marked as read-write operand. Fixes: ff340d2472ec ("s390: add stack switch helper") Cc: <> # 4.20 Reviewed-by: Vasily Gorbik <> Signed-off-by: Heiko Carstens <> Signed-off-by: Vasily Gorbik <>
2021-06-21s390/topology: clear thread/group maps for offline cpusSven Schnelle
The current code doesn't clear the thread/group maps for offline CPUs. This may cause kernel crashes like the one bewlow in common code that assumes if a CPU has sibblings it is online. Unable to handle kernel pointer dereference in virtual kernel address space Call Trace: [<000000013a4b8c3c>] blk_mq_map_swqueue+0x10c/0x388 ([<000000013a4b8bcc>] blk_mq_map_swqueue+0x9c/0x388) [<000000013a4b9300>] blk_mq_init_allocated_queue+0x448/0x478 [<000000013a4b9416>] blk_mq_init_queue+0x4e/0x90 [<000003ff8019d3e6>] loop_add+0x106/0x278 [loop] [<000003ff801b8148>] loop_init+0x148/0x1000 [loop] [<0000000139de4924>] do_one_initcall+0x3c/0x1e0 [<0000000139ef449a>] do_init_module+0x6a/0x2a0 [<0000000139ef61bc>] __do_sys_finit_module+0xa4/0xc0 [<0000000139de9e6e>] do_syscall+0x7e/0xd0 [<000000013a8e0aec>] __do_syscall+0xbc/0x110 [<000000013a8ee2e8>] system_call+0x78/0xa0 Fixes: 52aeda7accb6 ("s390/topology: remove offline CPUs from CPU topology masks") Cc: <> # 5.7+ Reported-by: Marius Hillenbrand <> Signed-off-by: Sven Schnelle <> Reviewed-by: Heiko Carstens <> Signed-off-by: Vasily Gorbik <>
2021-06-21s390: clear pt_regs::flags on irq entrySven Schnelle
The current irq entry code doesn't initialize pt_regs::flags. On exit to user mode arch_do_signal_or_restart() tests whether PIF_SYSCALL is set, which might yield wrong results. Fix this by clearing pt_regs::flags in the entry.S irq handler code. Reported-by: Heiko Carstens <> Signed-off-by: Sven Schnelle <> Reviewed-by: Heiko Carstens <> Fixes: 56e62a737028 ("s390: convert to generic entry") Cc: <> # 5.12 Signed-off-by: Vasily Gorbik <>
2021-06-21s390: fix system call restart with multiple signalsSven Schnelle
glibc complained with "The futex facility returned an unexpected error code.". It turned out that the futex syscall returned -ERESTARTSYS because a signal is pending. arch_do_signal_or_restart() restored the syscall parameters (nameley regs->gprs[2]) and set PIF_SYSCALL_RESTART. When another signal is made pending later in the exit loop arch_do_signal_or_restart() is called again. This function clears PIF_SYSCALL_RESTART and checks the return code which is set in regs->gprs[2]. However, regs->gprs[2] was restored in the previous run and no longer contains -ERESTARTSYS, so PIF_SYSCALL_RESTART isn't set again and the syscall is skipped. Fix this by not clearing PIF_SYSCALL_RESTART - it is already cleared in __do_syscall() when the syscall is restarted. Reported-by: Bjoern Walk <> Signed-off-by: Sven Schnelle <> Reviewed-by: Heiko Carstens <> Fixes: 56e62a737028 ("s390: convert to generic entry") Cc: <> # 5.12 Signed-off-by: Vasily Gorbik <>
2021-06-20Merge tag 'x86_urgent_for_v5.13_rc6' of ↵Linus Torvalds
git:// Pull x86 fixes from Borislav Petkov: "A first set of urgent fixes to the FPU/XSTATE handling mess^W code. (There's a lot more in the pipe): - Prevent corruption of the XSTATE buffer in signal handling by validating what is being copied from userspace first. - Invalidate other task's preserved FPU registers on XRSTOR failure (#PF) because latter can still modify some of them. - Restore the proper PKRU value in case userspace modified it - Reset FPU state when signal restoring fails Other: - Map EFI boot services data memory as encrypted in a SEV guest so that the guest can access it and actually boot properly - Two SGX correctness fixes: proper resources freeing and a NUMA fix" * tag 'x86_urgent_for_v5.13_rc6' of git:// x86/mm: Avoid truncating memblocks for SGX memory x86/sgx: Add missing xa_destroy() when virtual EPC is destroyed x86/fpu: Reset state for all signal restore failures x86/pkru: Write hardware init value to PKRU when xstate is init x86/process: Check PF_KTHREAD and not current->mm for kernel threads x86/fpu: Invalidate FPU state after a failed XRSTOR from a user buffer x86/fpu: Prevent state corruption in __fpu__restore_sig() x86/ioremap: Map EFI-reserved memory as encrypted for SEV
2021-06-19Merge tag 'powerpc-5.13-6' of ↵Linus Torvalds
git:// Pull powerpc fixes from Michael Ellerman: "Fix initrd corruption caused by our recent change to use relative jump labels. Fix a crash using perf record on systems without a hardware PMU backend. Rework our 64-bit signal handling slighty to make it more closely match the old behaviour, after the recent change to use unsafe user accessors. Thanks to Anastasia Kovaleva, Athira Rajeev, Christophe Leroy, Daniel Axtens, Greg Kurz, and Roman Bolshakov" * tag 'powerpc-5.13-6' of git:// powerpc/perf: Fix crash in perf_instruction_pointer() when ppmu is not set powerpc: Fix initrd corruption with relative jump labels powerpc/signal64: Copy siginfo before changing regs->nip powerpc/mem: Add back missing header to fix 'no previous prototype' error
2021-06-19Merge tag 'riscv-for-linus-5.13-rc7' of ↵Linus Torvalds
git:// Pull RISC-V fixes from Palmer Dabbelt: - A build fix to always build modules with the 'medany' code model, as the module loader doesn't support 'medlow'. - A Kconfig warning fix for the SiFive errata. - A pair of fixes that for regressions to the recent memory layout changes. - A fix for the FU740 device tree. * tag 'riscv-for-linus-5.13-rc7' of git:// riscv: dts: fu740: fix cache-controller interrupts riscv: Ensure BPF_JIT_REGION_START aligned with PMD size riscv: kasan: Fix MODULES_VADDR evaluation due to local variables' name riscv: sifive: fix Kconfig errata warning riscv32: Use medany C model for modules
2021-06-19Merge tag 's390-5.13-4' of ↵Linus Torvalds
git:// Pull s390 fixes from Vasily Gorbik: - Fix zcrypt ioctl hang due to AP queue msg counter dropping below 0 when pending requests are purged. - Two fixes for the machine check handler in the entry code. * tag 's390-5.13-4' of git:// s390/ap: Fix hanging ioctl caused by wrong msg counter s390/mcck: fix invalid KVM guest condition check s390/mcck: fix calculation of SIE critical section size
2021-06-19riscv: dts: fu740: fix cache-controller interruptsDavid Abdurachmanov
The order of interrupt numbers is incorrect. The order for FU740 is: DirError, DataError, DataFail, DirFail From SiFive FU740-C000 Manual: 19 - L2 Cache DirError 20 - L2 Cache DirFail 21 - L2 Cache DataError 22 - L2 Cache DataFail Signed-off-by: David Abdurachmanov <> Signed-off-by: Palmer Dabbelt <>
2021-06-18riscv: Ensure BPF_JIT_REGION_START aligned with PMD sizeJisheng Zhang
Andreas reported commit fc8504765ec5 ("riscv: bpf: Avoid breaking W^X") breaks booting with one kind of defconfig, I reproduced a kernel panic with the defconfig: [ 0.138553] Unable to handle kernel paging request at virtual address ffffffff81201220 [ 0.139159] Oops [#1] [ 0.139303] Modules linked in: [ 0.139601] CPU: 0 PID: 1 Comm: swapper/0 Not tainted 5.13.0-rc5-default+ #1 [ 0.139934] Hardware name: riscv-virtio,qemu (DT) [ 0.140193] epc : __memset+0xc4/0xfc [ 0.140416] ra : skb_flow_dissector_init+0x1e/0x82 [ 0.140609] epc : ffffffff8029806c ra : ffffffff8033be78 sp : ffffffe001647da0 [ 0.140878] gp : ffffffff81134b08 tp : ffffffe001654380 t0 : ffffffff81201158 [ 0.141156] t1 : 0000000000000002 t2 : 0000000000000154 s0 : ffffffe001647dd0 [ 0.141424] s1 : ffffffff80a43250 a0 : ffffffff81201220 a1 : 0000000000000000 [ 0.141654] a2 : 000000000000003c a3 : ffffffff81201258 a4 : 0000000000000064 [ 0.141893] a5 : ffffffff8029806c a6 : 0000000000000040 a7 : ffffffffffffffff [ 0.142126] s2 : ffffffff81201220 s3 : 0000000000000009 s4 : ffffffff81135088 [ 0.142353] s5 : ffffffff81135038 s6 : ffffffff8080ce80 s7 : ffffffff80800438 [ 0.142584] s8 : ffffffff80bc6578 s9 : 0000000000000008 s10: ffffffff806000ac [ 0.142810] s11: 0000000000000000 t3 : fffffffffffffffc t4 : 0000000000000000 [ 0.143042] t5 : 0000000000000155 t6 : 00000000000003ff [ 0.143220] status: 0000000000000120 badaddr: ffffffff81201220 cause: 000000000000000f [ 0.143560] [<ffffffff8029806c>] __memset+0xc4/0xfc [ 0.143859] [<ffffffff8061e984>] init_default_flow_dissectors+0x22/0x60 [ 0.144092] [<ffffffff800010fc>] do_one_initcall+0x3e/0x168 [ 0.144278] [<ffffffff80600df0>] kernel_init_freeable+0x1c8/0x224 [ 0.144479] [<ffffffff804868a8>] kernel_init+0x12/0x110 [ 0.144658] [<ffffffff800022de>] ret_from_exception+0x0/0xc [ 0.145124] ---[ end trace f1e9643daa46d591 ]--- After some investigation, I think I found the root cause: commit 2bfc6cd81bd ("move kernel mapping outside of linear mapping") moves BPF JIT region after the kernel: | #define BPF_JIT_REGION_START PFN_ALIGN((unsigned long)&_end) The &_end is unlikely aligned with PMD size, so the front bpf jit region sits with part of kernel .data section in one PMD size mapping. But kernel is mapped in PMD SIZE, when bpf_jit_binary_lock_ro() is called to make the first bpf jit prog ROX, we will make part of kernel .data section RO too, so when we write to, for example memset the .data section, MMU will trigger a store page fault. To fix the issue, we need to ensure the BPF JIT region is PMD size aligned. This patch acchieve this goal by restoring the BPF JIT region to original position, I.E the 128MB before kernel .text section. The modification to kasan_init.c is inspired by Alexandre. Fixes: fc8504765ec5 ("riscv: bpf: Avoid breaking W^X") Reported-by: Andreas Schwab <> Signed-off-by: Jisheng Zhang <> Signed-off-by: Palmer Dabbelt <>
2021-06-18riscv: kasan: Fix MODULES_VADDR evaluation due to local variables' nameJisheng Zhang
commit 2bfc6cd81bd1 ("riscv: Move kernel mapping outside of linear mapping") makes use of MODULES_VADDR to populate kernel, BPF, modules mapping. Currently, MODULES_VADDR is defined as below for RV64: | #define MODULES_VADDR (PFN_ALIGN((unsigned long)&_end) - SZ_2G) But kasan_init() has two local variables which are also named as _start, _end, so MODULES_VADDR is evaluated with the local variable _end rather than the global "_end" as we expected. Fix this issue by renaming the two local variables. Fixes: 2bfc6cd81bd1 ("riscv: Move kernel mapping outside of linear mapping") Signed-off-by: Jisheng Zhang <> Signed-off-by: Palmer Dabbelt <>
2021-06-18Merge tag 'pci-v5.13-fixes-2' of ↵Linus Torvalds
git:// Pull PCI fixes from Bjorn Helgaas: - Clear 64-bit flag for host bridge windows below 4GB to fix a resource allocation regression added in -rc1 (Punit Agrawal) - Fix tegra194 MCFG quirk build regressions added in -rc1 (Jon Hunter) - Avoid secondary bus resets on TI KeyStone C667X devices (Antti Järvinen) - Avoid secondary bus resets on some NVIDIA GPUs (Shanker Donthineni) - Work around FLR erratum on Huawei Intelligent NIC VF (Chiqijun) - Avoid broken ATS on AMD Navi14 GPU (Evan Quan) - Trust Broadcom BCM57414 NIC to isolate functions even though it doesn't advertise ACS support (Sriharsha Basavapatna) - Work around AMD RS690 BIOSes that don't configure DMA above 4GB (Mikel Rychliski) - Fix panic during PIO transfer on Aardvark controller (Pali Rohár) * tag 'pci-v5.13-fixes-2' of git:// PCI: aardvark: Fix kernel panic during PIO transfer PCI: Add AMD RS690 quirk to enable 64-bit DMA PCI: Add ACS quirk for Broadcom BCM57414 NIC PCI: Mark AMD Navi14 GPU ATS as broken PCI: Work around Huawei Intelligent NIC VF FLR erratum PCI: Mark some NVIDIA GPUs to avoid bus reset PCI: Mark TI C667X to avoid bus reset PCI: tegra194: Fix MCFG quirk build regressions PCI: of: Clear 64-bit flag for non-prefetchable memory below 4GB
2021-06-18Merge tag 'arc-5.13-rc7-fixes' of ↵Linus Torvalds
git:// Pull ARC fixes from Vineet Gupta: - ARCv2 userspace ABI not populating a few registers - Unbork CONFIG_HARDENED_USERCOPY for ARC * tag 'arc-5.13-rc7-fixes' of git:// ARC: fix CONFIG_HARDENED_USERCOPY ARCv2: save ABI registers across signal handling
2021-06-18x86/mm: Avoid truncating memblocks for SGX memoryFan Du
tl;dr: Several SGX users reported seeing the following message on NUMA systems: sgx: [Firmware Bug]: Unable to map EPC section to online node. Fallback to the NUMA node 0. This turned out to be the memblock code mistakenly throwing away SGX memory. === Full Changelog === The 'max_pfn' variable represents the highest known RAM address. It can be used, for instance, to quickly determine for which physical addresses there is mem_map[] space allocated. The numa_meminfo code makes an effort to throw out ("trim") all memory blocks which are above 'max_pfn'. SGX memory is not considered RAM (it is marked as "Reserved" in the e820) and is not taken into account by max_pfn. Despite this, SGX memory areas have NUMA affinity and are enumerated in the ACPI SRAT table. The existing SGX code uses the numa_meminfo mechanism to look up the NUMA affinity for its memory areas. In cases where SGX memory was above max_pfn (usually just the one EPC section in the last highest NUMA node), the numa_memblock is truncated at 'max_pfn', which is below the SGX memory. When the SGX code tries to look up the affinity of this memory, it fails and produces an error message: sgx: [Firmware Bug]: Unable to map EPC section to online node. Fallback to the NUMA node 0. and assigns the memory to NUMA node 0. Instead of silently truncating the memory block at 'max_pfn' and dropping the SGX memory, add the truncated portion to 'numa_reserved_meminfo'. This allows the SGX code to later determine the NUMA affinity of its 'Reserved' area. Before, numa_meminfo looked like this (from 'crash'): blk = { start = 0x0, end = 0x2080000000, nid = 0x0 } { start = 0x2080000000, end = 0x4000000000, nid = 0x1 } numa_reserved_meminfo is empty. With this, numa_meminfo looks like this: blk = { start = 0x0, end = 0x2080000000, nid = 0x0 } { start = 0x2080000000, end = 0x4000000000, nid = 0x1 } and numa_reserved_meminfo has an entry for node 1's SGX memory: blk = { start = 0x4000000000, end = 0x4080000000, nid = 0x1 } [ daveh: completely rewrote/reworked changelog ] Fixes: 5d30f92e7631 ("x86/NUMA: Provide a range-to-target_node lookup facility") Reported-by: Reinette Chatre <> Signed-off-by: Fan Du <> Signed-off-by: Dave Hansen <> Signed-off-by: Borislav Petkov <> Reviewed-by: Jarkko Sakkinen <> Reviewed-by: Dan Williams <> Reviewed-by: Dave Hansen <> Cc: <> Link:
2021-06-18PCI: Add AMD RS690 quirk to enable 64-bit DMAMikel Rychliski
Although the AMD RS690 chipset has 64-bit DMA support, BIOS implementations sometimes fail to configure the memory limit registers correctly. The Acer F690GVM mainboard uses this chipset and a Marvell 88E8056 NIC. The sky2 driver programs the NIC to use 64-bit DMA, which will not work: sky2 0000:02:00.0: error interrupt status=0x8 sky2 0000:02:00.0 eth0: tx timeout sky2 0000:02:00.0 eth0: transmit ring 0 .. 22 report=0 done=0 Other drivers required by this mainboard either don't support 64-bit DMA, or have it disabled using driver specific quirks. For example, the ahci driver has quirks to enable or disable 64-bit DMA depending on the BIOS version (see ahci_sb600_enable_64bit() in ahci.c). This ahci quirk matches against the SB600 SATA controller, but the real issue is almost certainly with the RS690 PCI host that it was commonly attached to. To avoid this issue in all drivers with 64-bit DMA support, fix the configuration of the PCI host. If the kernel is aware of physical memory above 4GB, but the BIOS never configured the PCI host with this information, update the registers with our values. [bhelgaas: drop PCI_DEVICE_ID_ATI_RS690 definition] Link: Signed-off-by: Mikel Rychliski <> Signed-off-by: Bjorn Helgaas <>
2021-06-18powerpc/perf: Fix crash in perf_instruction_pointer() when ppmu is not setAthira Rajeev
On systems without any specific PMU driver support registered, running perf record causes Oops. The relevant portion from call trace: BUG: Kernel NULL pointer dereference on read at 0x00000040 Faulting instruction address: 0xc0021f0c Oops: Kernel access of bad area, sig: 11 [#1] BE PAGE_SIZE=4K PREEMPT CMPCPRO SAF3000 DIE NOTIFICATION CPU: 0 PID: 442 Comm: null_syscall Not tainted 5.13.0-rc6-s3k-dev-01645-g7649ee3d2957 #5164 NIP: c0021f0c LR: c00e8ad8 CTR: c00d8a5c NIP perf_instruction_pointer+0x10/0x60 LR perf_prepare_sample+0x344/0x674 Call Trace: perf_prepare_sample+0x7c/0x674 (unreliable) perf_event_output_forward+0x3c/0x94 __perf_event_overflow+0x74/0x14c perf_swevent_hrtimer+0xf8/0x170 __hrtimer_run_queues.constprop.0+0x160/0x318 hrtimer_interrupt+0x148/0x3b0 timer_interrupt+0xc4/0x22c Decrementer_virt+0xb8/0xbc During perf record session, perf_instruction_pointer() is called to capture the sample IP. This function in core-book3s accesses ppmu->flags. If a platform specific PMU driver is not registered, ppmu is set to NULL and accessing its members results in a crash. Fix this crash by checking if ppmu is set. Fixes: 2ca13a4cc56c ("powerpc/perf: Use regs->nip when SIAR is zero") Cc: # v5.11+ Reported-by: Christophe Leroy <> Signed-off-by: Athira Rajeev <> Tested-by: Christophe Leroy <> Signed-off-by: Michael Ellerman <> Link:
2021-06-17Merge tag 'for-linus' of git:// Torvalds
Pull kvm fixes from Paolo Bonzini: "Miscellaneous bugfixes. The main interesting one is a NULL pointer dereference reported by syzkaller ("KVM: x86: Immediately reset the MMU context when the SMM flag is cleared")" * tag 'for-linus' of git:// KVM: selftests: Fix kvm_check_cap() assertion KVM: x86/mmu: Calculate and check "full" mmu_role for nested MMU KVM: X86: Fix x86_emulator slab cache leak KVM: SVM: Call SEV Guest Decommission if ASID binding fails KVM: x86: Immediately reset the MMU context when the SMM flag is cleared KVM: x86: Fix fall-through warnings for Clang KVM: SVM: fix doc warnings KVM: selftests: Fix compiling errors when initializing the static structure kvm: LAPIC: Restore guard to prevent illegal APIC register access
2021-06-15x86/sgx: Add missing xa_destroy() when virtual EPC is destroyedKai Huang
xa_destroy() needs to be called to destroy a virtual EPC's page array before calling kfree() to free the virtual EPC. Currently it is not called so add the missing xa_destroy(). Fixes: 540745ddbc70 ("x86/sgx: Introduce virtual EPC for use by KVM guests") Signed-off-by: Kai Huang <> Signed-off-by: Borislav Petkov <> Acked-by: Dave Hansen <> Tested-by: Yang Zhong <> Link:
2021-06-15powerpc: Fix initrd corruption with relative jump labelsMichael Ellerman
Commit b0b3b2c78ec0 ("powerpc: Switch to relative jump labels") switched us to using relative jump labels. That involves changing the code, target and key members in struct jump_entry to be relative to the address of the jump_entry, rather than absolute addresses. We have two static inlines that create a struct jump_entry, arch_static_branch() and arch_static_branch_jump(), as well as an asm macro ARCH_STATIC_BRANCH, which is used by the pseries-only hypervisor tracing code. Unfortunately we missed updating the key to be a relative reference in ARCH_STATIC_BRANCH. That causes a pseries kernel to have a handful of jump_entry structs with bad key values. Instead of being a relative reference they instead hold the full address of the key. However the code doesn't expect that, it still adds the key value to the address of the jump_entry (see jump_entry_key()) expecting to get a pointer to a key somewhere in kernel data. The table of jump_entry structs sits in rodata, which comes after the kernel text. In a typical build this will be somewhere around 15MB. The address of the key will be somewhere in data, typically around 20MB. Adding the two values together gets us a pointer somewhere around 45MB. We then call static_key_set_entries() with that bad pointer and modify some members of the struct static_key we think we are pointing at. A pseries kernel is typically ~30MB in size, so writing to ~45MB won't corrupt the kernel itself. However if we're booting with an initrd, depending on the size and exact location of the initrd, we can corrupt the initrd. Depending on how exactly we corrupt the initrd it can either cause the system to not boot, or just corrupt one of the files in the initrd. The fix is simply to make the key value relative to the jump_entry struct in the ARCH_STATIC_BRANCH macro. Fixes: b0b3b2c78ec0 ("powerpc: Switch to relative jump labels") Reported-by: Anastasia Kovaleva <> Reported-by: Roman Bolshakov <> Reported-by: Greg Kurz <> Reported-by: Daniel Axtens <> Signed-off-by: Michael Ellerman <> Tested-by: Daniel Axtens <> Tested-by: Greg Kurz <> Signed-off-by: Michael Ellerman <> Link:
2021-06-14powerpc/signal64: Copy siginfo before changing regs->nipMichael Ellerman
In commit 96d7a4e06fab ("powerpc/signal64: Rewrite handle_rt_signal64() to minimise uaccess switches") the 64-bit signal code was rearranged to use user_write_access_begin/end(). As part of that change the call to copy_siginfo_to_user() was moved later in the function, so that it could be done after the user_write_access_end(). In particular it was moved after we modify regs->nip to point to the signal trampoline. That means if copy_siginfo_to_user() fails we exit handle_rt_signal64() with an error but with regs->nip modified, whereas previously we would not modify regs->nip until the copy succeeded. Returning an error from signal delivery but with regs->nip updated leaves the process in a sort of half-delivered state. We do immediately force a SEGV in signal_setup_done(), called from do_signal(), so the process should never run in the half-delivered state. However that SEGV is not delivered until we've gone around to do_notify_resume() again, so it's possible some tracing could observe the half-delivered state. There are other cases where we fail signal delivery with regs partly updated, eg. the write to newsp and SA_SIGINFO, but the latter at least is very unlikely to fail as it reads back from the frame we just wrote to. Looking at other arches they seem to be more careful about leaving regs unchanged until the copy operations have succeeded, and in general that seems like good hygenie. So although the current behaviour is not cleary buggy, it's also not clearly correct. So move the call to copy_siginfo_to_user() up prior to the modification of regs->nip, which is closer to the old behaviour, and easier to reason about. Fixes: 96d7a4e06fab ("powerpc/signal64: Rewrite handle_rt_signal64() to minimise uaccess switches") Signed-off-by: Michael Ellerman <> Link:
2021-06-12riscv: sifive: fix Kconfig errata warningRandy Dunlap
The SOC_SIFIVE Kconfig entry unconditionally selects ERRATA_SIFIVE. However, ERRATA_SIFIVE depends on RISCV_ERRATA_ALTERNATIVE, which is not set, so SOC_SIFIVE should either depend on or select RISCV_ERRATA_ALTERNATIVE. Use 'select' here to quieten the Kconfig warning. WARNING: unmet direct dependencies detected for ERRATA_SIFIVE Depends on [n]: RISCV_ERRATA_ALTERNATIVE [=n] Selected by [y]: - SOC_SIFIVE [=y] Fixes: 1a0e5dbd3723 ("riscv: sifive: Add SiFive alternative ports") Signed-off-by: Randy Dunlap <> Cc: Paul Walmsley <> Cc: Palmer Dabbelt <> Cc: Albert Ou <> Cc: Cc: Vincent Chen <> Signed-off-by: Palmer Dabbelt <>
2021-06-12riscv32: Use medany C model for modulesKhem Raj
When CONFIG_CMODEL_MEDLOW is used it ends up generating riscv_hi20_rela relocations in modules which are not resolved during runtime and following errors would be seen [ 4.802714] virtio_input: target 00000000c1539090 can not be addressed by the 32-bit offset from PC = 39148b7b [ 4.854800] virtio_input: target 00000000c1539090 can not be addressed by the 32-bit offset from PC = 9774456d Signed-off-by: Khem Raj <> Signed-off-by: Palmer Dabbelt <>
2021-06-12Merge tag 'riscv-for-linus-5.13-rc6' of ↵Linus Torvalds
git:// Pull RISC-V fixes from Palmer Dabbelt: - A pair of XIP fixes: one to fix alternatives, and one to turn off the rest of the features that require code modification - A fix to a type that was causing some alternatives to break - A build fix for BUILTIN_DTB * tag 'riscv-for-linus-5.13-rc6' of git:// riscv: Fix BUILTIN_DTB for sifive and microchip soc riscv: alternative: fix typo in macro name riscv: code patching only works on !XIP_KERNEL riscv: xip: support runtime trap patching
2021-06-12Merge tag 'perf-urgent-2021-06-12' of ↵Linus Torvalds
git:// Pull perf fixes from Ingo Molnar: "Misc fixes: - Fix the NMI watchdog on ancient Intel CPUs - Remove a misguided, NMI-unsafe KASAN callback from the NMI-safe irq_work path used by perf. - Fix uncore events on Ice Lake servers. - Someone booted maxcpus=1 on an SNB-EP, and the uncore driver emitted warnings and was probably buggy. Fix it. - KCSAN found a genuine data race in the core perf code. Somewhat ironically the bug was introduced through a recent race fix. :-/ In our defense, the new race window was much more narrow. Fix it" * tag 'perf-urgent-2021-06-12' of git:// x86/nmi_watchdog: Fix old-style NMI watchdog regression on old Intel CPUs irq_work: Make irq_work_queue() NMI-safe again perf/x86/intel/uncore: Fix M2M event umask for Ice Lake server perf/x86/intel/uncore: Fix a kernel WARNING triggered by maxcpus=1 perf: Fix data race between pin_count increment/decrement
2021-06-11riscv: Fix BUILTIN_DTB for sifive and microchip socAlexandre Ghiti
Fix BUILTIN_DTB config which resulted in a dtb that was actually not built into the Linux image: in the same manner as Canaan soc does, create an object file from the dtb file that will get linked into the Linux image. Signed-off-by: Alexandre Ghiti <> Signed-off-by: Palmer Dabbelt <>
2021-06-11x86, lto: Pass -stack-alignment only on LLD < 13.0.0Tor Vic
Since LLVM commit 3787ee4, the '-stack-alignment' flag has been dropped [1], leading to the following error message when building a LTO kernel with Clang-13 and LLD-13: ld.lld: error: -plugin-opt=-: ld.lld: Unknown command line argument '-stack-alignment=8'. Try 'ld.lld --help' ld.lld: Did you mean '--stackrealign=8'? It also appears that the '-code-model' flag is not necessary anymore starting with LLVM-9 [2]. Drop '-code-model' and make '-stack-alignment' conditional on LLD < 13.0.0. These flags were necessary because these flags were not encoded in the IR properly, so the link would restart optimizations without them. Now there are properly encoded in the IR, and these flags exposing implementation details are no longer necessary. [1] [2] Cc: Link: Signed-off-by: Tor Vic <> Reviewed-by: Nathan Chancellor <> Tested-by: Nathan Chancellor <> Signed-off-by: Kees Cook <> Link:
2021-06-11KVM: x86/mmu: Calculate and check "full" mmu_role for nested MMUSean Christopherson
Calculate and check the full mmu_role when initializing the MMU context for the nested MMU, where "full" means the bits and pieces of the role that aren't handled by kvm_calc_mmu_role_common(). While the nested MMU isn't used for shadow paging, things like the number of levels in the guest's page tables are surprisingly important when walking the guest page tables. Failure to reinitialize the nested MMU context if L2's paging mode changes can result in unexpected and/or missed page faults, and likely other explosions. E.g. if an L1 vCPU is running both a 32-bit PAE L2 and a 64-bit L2, the "common" role calculation will yield the same role for both L2s. If the 64-bit L2 is run after the 32-bit PAE L2, L0 will fail to reinitialize the nested MMU context, ultimately resulting in a bad walk of L2's page tables as the MMU will still have a guest root_level of PT32E_ROOT_LEVEL. WARNING: CPU: 4 PID: 167334 at arch/x86/kvm/vmx/vmx.c:3075 ept_save_pdptrs+0x15/0xe0 [kvm_intel] Modules linked in: kvm_intel] CPU: 4 PID: 167334 Comm: CPU 3/KVM Not tainted 5.13.0-rc1-d849817d5673-reqs #185 Hardware name: ASUS Q87M-E/Q87M-E, BIOS 1102 03/03/2014 RIP: 0010:ept_save_pdptrs+0x15/0xe0 [kvm_intel] Code: <0f> 0b c3 f6 87 d8 02 00f RSP: 0018:ffffbba702dbba00 EFLAGS: 00010202 RAX: 0000000000000011 RBX: 0000000000000002 RCX: ffffffff810a2c08 RDX: ffff91d7bc30acc0 RSI: 0000000000000011 RDI: ffff91d7bc30a600 RBP: ffff91d7bc30a600 R08: 0000000000000010 R09: 0000000000000007 R10: 0000000000000000 R11: 0000000000000000 R12: ffff91d7bc30a600 R13: ffff91d7bc30acc0 R14: ffff91d67c123460 R15: 0000000115d7e005 FS: 00007fe8e9ffb700(0000) GS:ffff91d90fb00000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 0000000000000000 CR3: 000000029f15a001 CR4: 00000000001726e0 Call Trace: kvm_pdptr_read+0x3a/0x40 [kvm] paging64_walk_addr_generic+0x327/0x6a0 [kvm] paging64_gva_to_gpa_nested+0x3f/0xb0 [kvm] kvm_fetch_guest_virt+0x4c/0xb0 [kvm] __do_insn_fetch_bytes+0x11a/0x1f0 [kvm] x86_decode_insn+0x787/0x1490 [kvm] x86_decode_emulated_instruction+0x58/0x1e0 [kvm] x86_emulate_instruction+0x122/0x4f0 [kvm] vmx_handle_exit+0x120/0x660 [kvm_intel] kvm_arch_vcpu_ioctl_run+0xe25/0x1cb0 [kvm] kvm_vcpu_ioctl+0x211/0x5a0 [kvm] __x64_sys_ioctl+0x83/0xb0 do_syscall_64+0x40/0xb0 entry_SYSCALL_64_after_hwframe+0x44/0xae Cc: Vitaly Kuznetsov <> Cc: Fixes: bf627a928837 ("x86/kvm/mmu: check if MMU reconfiguration is needed in init_kvm_nested_mmu()") Signed-off-by: Sean Christopherson <> Message-Id: <> Signed-off-by: Paolo Bonzini <>
2021-06-11KVM: X86: Fix x86_emulator slab cache leakWanpeng Li
Commit c9b8b07cded58 (KVM: x86: Dynamically allocate per-vCPU emulation context) tries to allocate per-vCPU emulation context dynamically, however, the x86_emulator slab cache is still exiting after the kvm module is unload as below after destroying the VM and unloading the kvm module. grep x86_emulator /proc/slabinfo x86_emulator 36 36 2672 12 8 : tunables 0 0 0 : slabdata 3 3 0 This patch fixes this slab cache leak by destroying the x86_emulator slab cache when the kvm module is unloaded. Fixes: c9b8b07cded58 (KVM: x86: Dynamically allocate per-vCPU emulation context) Cc: Signed-off-by: Wanpeng Li <> Message-Id: <> Signed-off-by: Paolo Bonzini <>
2021-06-11KVM: SVM: Call SEV Guest Decommission if ASID binding failsAlper Gun
Send SEV_CMD_DECOMMISSION command to PSP firmware if ASID binding fails. If a failure happens after a successful LAUNCH_START command, a decommission command should be executed. Otherwise, guest context will be unfreed inside the AMD SP. After the firmware will not have memory to allocate more SEV guest context, LAUNCH_START command will begin to fail with SEV_RET_RESOURCE_LIMIT error. The existing code calls decommission inside sev_unbind_asid, but it is not called if a failure happens before guest activation succeeds. If sev_bind_asid fails, decommission is never called. PSP firmware has a limit for the number of guests. If sev_asid_binding fails many times, PSP firmware will not have resources to create another guest context. Cc: Fixes: 59414c989220 ("KVM: SVM: Add support for KVM_SEV_LAUNCH_START command") Reported-by: Peter Gonda <> Signed-off-by: Alper Gun <> Reviewed-by: Marc Orr <> Signed-off-by: Paolo Bonzini <> Message-Id: <>
2021-06-10riscv: alternative: fix typo in macro nameVitaly Wool
alternative-macros.h defines ALT_NEW_CONTENT in its assembly part and ALT_NEW_CONSTENT in the C part. Most likely it is the latter that is wrong. Fixes: 6f4eea90465ad (riscv: Introduce alternative mechanism to apply errata solution) Signed-off-by: Vitaly Wool <> Signed-off-by: Palmer Dabbelt <>
2021-06-10ARC: fix CONFIG_HARDENED_USERCOPYVineet Gupta
Currently enabling this triggers a warning | usercopy: Kernel memory overwrite attempt detected to kernel text (offset 155633, size 11)! | usercopy: BUG: failure at mm/usercopy.c:99/usercopy_abort()! | |gcc generated __builtin_trap |Path: /bin/busybox |CPU: 0 PID: 84 Comm: init Not tainted 5.4.22 | |[ECR ]: 0x00090005 => gcc generated __builtin_trap |[EFA ]: 0x9024fcaa |[BLINK ]: usercopy_abort+0x8a/0x8c |[ERET ]: memfd_fcntl+0x0/0x470 |[STAT32]: 0x80080802 : IE K |... |... |Stack Trace: | memfd_fcntl+0x0/0x470 | usercopy_abort+0x8a/0x8c | __check_object_size+0x10e/0x138 | copy_strings+0x1f4/0x38c | __do_execve_file+0x352/0x848 | EV_Trap+0xcc/0xd0 The issue is triggered by an allocation in "init reclaimed" region. ARC _stext emcompasses the init region (for historical reasons we wanted the init.text to be under .text as well). This however trips up __check_object_size()->check_kernel_text_object() which treats this as object bleeding into kernel text. Fix that by rezoning _stext to start from regular kernel .text and leave out .init altogether. Fixes: Reported-by: Evgeniy Didin <> Reviewed-by: Kees Cook <> Signed-off-by: Vineet Gupta <>
2021-06-10ARCv2: save ABI registers across signal handlingVineet Gupta
ARCv2 has some configuration dependent registers (r30, r58, r59) which could be targetted by the compiler. To keep the ABI stable, these were unconditionally part of the glibc ABI (sysdeps/unix/sysv/linux/arc/sys/ucontext.h:mcontext_t) however we missed populating them (by saving/restoring them across signal handling). This patch fixes the issue by - adding arcv2 ABI regs to kernel struct sigcontext - populating them during signal handling Change to struct sigcontext might seem like a glibc ABI change (although it primarily uses ucontext_t:mcontext_t) but the fact is - it has only been extended (existing fields are not touched) - the old sigcontext was ABI incomplete to begin with anyways Fixes: Cc: <> Tested-by: kernel test robot <> Reported-by: Vladimir Isaev <> Signed-off-by: Vineet Gupta <>
2021-06-10riscv: code patching only works on !XIP_KERNELJisheng Zhang
Some features which need code patching such as KPROBES, DYNAMIC_FTRACE KGDB can only work on !XIP_KERNEL. Add dependencies for these features that rely on code patching. Signed-off-by: Jisheng Zhang <> Signed-off-by: Palmer Dabbelt <>
2021-06-10riscv: xip: support runtime trap patchingVitaly Wool
RISCV_ERRATA_ALTERNATIVE patches text at runtime which is currently not possible when the kernel is executed from the flash in XIP mode. Since runtime patching concerns only traps at the moment, let's just have all the traps reside in RAM anyway if RISCV_ERRATA_ALTERNATIVE is set. Thus, these functions will be patch-able even when the .text section is in flash. Signed-off-by: Vitaly Wool <> Signed-off-by: Palmer Dabbelt <>
2021-06-10KVM: x86: Immediately reset the MMU context when the SMM flag is clearedSean Christopherson
Immediately reset the MMU context when the vCPU's SMM flag is cleared so that the SMM flag in the MMU role is always synchronized with the vCPU's flag. If RSM fails (which isn't correctly emulated), KVM will bail without calling post_leave_smm() and leave the MMU in a bad state. The bad MMU role can lead to a NULL pointer dereference when grabbing a shadow page's rmap for a page fault as the initial lookups for the gfn will happen with the vCPU's SMM flag (=0), whereas the rmap lookup will use the shadow page's SMM flag, which comes from the MMU (=1). SMM has an entirely different set of memslots, and so the initial lookup can find a memslot (SMM=0) and then explode on the rmap memslot lookup (SMM=1). general protection fault, probably for non-canonical address 0xdffffc0000000000: 0000 [#1] PREEMPT SMP KASAN KASAN: null-ptr-deref in range [0x0000000000000000-0x0000000000000007] CPU: 1 PID: 8410 Comm: syz-executor382 Not tainted 5.13.0-rc5-syzkaller #0 Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011 RIP: 0010:__gfn_to_rmap arch/x86/kvm/mmu/mmu.c:935 [inline] RIP: 0010:gfn_to_rmap+0x2b0/0x4d0 arch/x86/kvm/mmu/mmu.c:947 Code: <42> 80 3c 20 00 74 08 4c 89 ff e8 f1 79 a9 00 4c 89 fb 4d 8b 37 44 RSP: 0018:ffffc90000ffef98 EFLAGS: 00010246 RAX: 0000000000000000 RBX: ffff888015b9f414 RCX: ffff888019669c40 RDX: 0000000000000000 RSI: 0000000000000001 RDI: 0000000000000001 RBP: 0000000000000001 R08: ffffffff811d9cdb R09: ffffed10065a6002 R10: ffffed10065a6002 R11: 0000000000000000 R12: dffffc0000000000 R13: 0000000000000003 R14: 0000000000000001 R15: 0000000000000000 FS: 000000000124b300(0000) GS:ffff8880b9b00000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 0000000000000000 CR3: 0000000028e31000 CR4: 00000000001526e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 Call Trace: rmap_add arch/x86/kvm/mmu/mmu.c:965 [inline] mmu_set_spte+0x862/0xe60 arch/x86/kvm/mmu/mmu.c:2604 __direct_map arch/x86/kvm/mmu/mmu.c:2862 [inline] direct_page_fault+0x1f74/0x2b70 arch/x86/kvm/mmu/mmu.c:3769 kvm_mmu_do_page_fault arch/x86/kvm/mmu.h:124 [inline] kvm_mmu_page_fault+0x199/0x1440 arch/x86/kvm/mmu/mmu.c:5065 vmx_handle_exit+0x26/0x160 arch/x86/kvm/vmx/vmx.c:6122 vcpu_enter_guest+0x3bdd/0x9630 arch/x86/kvm/x86.c:9428 vcpu_run+0x416/0xc20 arch/x86/kvm/x86.c:9494 kvm_arch_vcpu_ioctl_run+0x4e8/0xa40 arch/x86/kvm/x86.c:9722 kvm_vcpu_ioctl+0x70f/0xbb0 arch/x86/kvm/../../../virt/kvm/kvm_main.c:3460 vfs_ioctl fs/ioctl.c:51 [inline] __do_sys_ioctl fs/ioctl.c:1069 [inline] __se_sys_ioctl+0xfb/0x170 fs/ioctl.c:1055 do_syscall_64+0x3f/0xb0 arch/x86/entry/common.c:47 entry_SYSCALL_64_after_hwframe+0x44/0xae RIP: 0033:0x440ce9 Cc: Reported-by: Fixes: 9ec19493fb86 ("KVM: x86: clear SMM flags before loading state while leaving SMM") Signed-off-by: Sean Christopherson <> Message-Id: <> Signed-off-by: Paolo Bonzini <>