kernel.git/arch/powerpc/kvm, branch linux-5.13.y

KVM: PPC: Fix clearing never mapped TCEs in realmode

2021-09-18T11:41:56Z

[ Upstream commit 1d78dfde33a02da1d816279c2e3452978b7abd39 ] Since commit e1a1ef84cd07 ("KVM: PPC: Book3S: Allocate guest TCEs on demand too"), pages for TCE tables for KVM guests are allocated only when needed. This allows skipping any update when clearing TCEs. This works mostly fine as TCE updates are handled when the MMU is enabled. The realmode handlers fail with H_TOO_HARD when pages are not yet allocated, except when clearing a TCE in which case KVM prints a warning and proceeds to dereference a NULL pointer, which crashes the host OS. This has not been caught so far as the change in commit e1a1ef84cd07 is reasonably new, and POWER9 runs mostly radix which does not use realmode handlers. With hash, the default TCE table is memset() by QEMU when the machine is reset which triggers page faults and the KVM TCE device's kvm_spapr_tce_fault() handles those with MMU on. And the huge DMA windows are not cleared by VMs which instead successfully create a DMA window big enough to map the VM memory 1:1 and then VMs just map everything without clearing. This started crashing now as commit 381ceda88c4c ("powerpc/pseries/iommu: Make use of DDW for indirect mapping") added a mode when a dymanic DMA window not big enough to map the VM memory 1:1 but it is used anyway, and the VM now is the first (i.e. not QEMU) to clear a just created table. Note that upstream QEMU needs to be modified to trigger the VM to trigger the host OS crash. This replaces WARN_ON_ONCE_RM() with a check and return, and adds another warning if TCE is not being cleared. Fixes: e1a1ef84cd07 ("KVM: PPC: Book3S: Allocate guest TCEs on demand too") Signed-off-by: Alexey Kardashevskiy Signed-off-by: Michael Ellerman Link: https://lore.kernel.org/r/20210827040706.517652-1-aik@ozlabs.ru Signed-off-by: Sasha Levin

KVM: PPC: Book3S HV Nested: Reflect guest PMU in-use to L0 when guest SPRs are live

2021-09-18T11:41:55Z

[ Upstream commit 1782663897945a5cf28e564ba5eed730098e9aa4 ] After the L1 saves its PMU SPRs but before loading the L2's PMU SPRs, switch the pmcregs_in_use field in the L1 lppaca to the value advertised by the L2 in its VPA. On the way out of the L2, set it back after saving the L2 PMU registers (if they were in-use). This transfers the PMU liveness indication between the L1 and L2 at the points where the registers are not live. This fixes the nested HV bug for which a workaround was added to the L0 HV by commit 63279eeb7f93a ("KVM: PPC: Book3S HV: Always save guest pmu for guest capable of nesting"), which explains the problem in detail. That workaround is no longer required for guests that include this bug fix. Fixes: 360cae313702 ("KVM: PPC: Book3S HV: Nested guest entry via hypercall") Signed-off-by: Nicholas Piggin Signed-off-by: Michael Ellerman Reviewed-by: Fabiano Rosas Link: https://lore.kernel.org/r/20210811160134.904987-10-npiggin@gmail.com Signed-off-by: Sasha Levin

KVM: PPC: Book3S HV: Fix copy_tofrom_guest routines

2021-09-18T11:41:55Z

[ Upstream commit 5d7d6dac8fe99ed59eee2300e4a03370f94d5222 ] The __kvmhv_copy_tofrom_guest_radix function was introduced along with nested HV guest support. It uses the platform's Radix MMU quadrants to provide a nested hypervisor with fast access to its nested guests memory (H_COPY_TOFROM_GUEST hypercall). It has also since been added as a fast path for the kvmppc_ld/st routines which are used during instruction emulation. The commit def0bfdbd603 ("powerpc: use probe_user_read() and probe_user_write()") changed the low level copy function from raw_copy_from_user to probe_user_read, which adds a check to access_ok. In powerpc that is: static inline bool __access_ok(unsigned long addr, unsigned long size) { return addr < TASK_SIZE_MAX && size <= TASK_SIZE_MAX - addr; } and TASK_SIZE_MAX is 0x0010000000000000UL for 64-bit, which means that setting the two MSBs of the effective address (which correspond to the quadrant) now cause access_ok to reject the access. This was not caught earlier because the most common code path via kvmppc_ld/st contains a fallback (kvm_read_guest) that is likely to succeed for L1 guests. For nested guests there is no fallback. Another issue is that probe_user_read (now __copy_from_user_nofault) does not return the number of bytes not copied in case of failure, so the destination memory is not being cleared anymore in kvmhv_copy_from_guest_radix: ret = kvmhv_copy_tofrom_guest_radix(vcpu, eaddr, to, NULL, n); if (ret > 0) <-- always false! memset(to + (n - ret), 0, ret); This patch fixes both issues by skipping access_ok and open-coding the low level __copy_to/from_user_inatomic. Fixes: def0bfdbd603 ("powerpc: use probe_user_read() and probe_user_write()") Signed-off-by: Fabiano Rosas Reviewed-by: Nicholas Piggin Signed-off-by: Michael Ellerman Link: https://lore.kernel.org/r/20210805212616.2641017-2-farosas@linux.ibm.com Signed-off-by: Sasha Levin

KVM: PPC: Book3S HV Nested: Sanitise H_ENTER_NESTED TM state

2021-07-28T12:37:36Z

commit d9c57d3ed52a92536f5fa59dc5ccdd58b4875076 upstream. The H_ENTER_NESTED hypercall is handled by the L0, and it is a request by the L1 to switch the context of the vCPU over to that of its L2 guest, and return with an interrupt indication. The L1 is responsible for switching some registers to guest context, and the L0 switches others (including all the hypervisor privileged state). If the L2 MSR has TM active, then the L1 is responsible for recheckpointing the L2 TM state. Then the L1 exits to L0 via the H_ENTER_NESTED hcall, and the L0 saves the TM state as part of the exit, and then it recheckpoints the TM state as part of the nested entry and finally HRFIDs into the L2 with TM active MSR. Not efficient, but about the simplest approach for something that's horrendously complicated. Problems arise if the L1 exits to the L0 with a TM state which does not match the L2 TM state being requested. For example if the L1 is transactional but the L2 MSR is non-transactional, or vice versa. The L0's HRFID can take a TM Bad Thing interrupt and crash. Fix this by disallowing H_ENTER_NESTED in TM[T] state entirely, and then ensuring that if the L1 is suspended then the L2 must have TM active, and if the L1 is not suspended then the L2 must not have TM active. Fixes: 360cae313702 ("KVM: PPC: Book3S HV: Nested guest entry via hypercall") Cc: stable@vger.kernel.org # v4.20+ Reported-by: Alexey Kardashevskiy Acked-by: Michael Neuling Signed-off-by: Nicholas Piggin Signed-off-by: Michael Ellerman Signed-off-by: Greg Kroah-Hartman

KVM: PPC: Book3S: Fix H_RTAS rets buffer overflow

2021-07-28T12:37:36Z

commit f62f3c20647ebd5fb6ecb8f0b477b9281c44c10a upstream. The kvmppc_rtas_hcall() sets the host rtas_args.rets pointer based on the rtas_args.nargs that was provided by the guest. That guest nargs value is not range checked, so the guest can cause the host rets pointer to be pointed outside the args array. The individual rtas function handlers check the nargs and nrets values to ensure they are correct, but if they are not, the handlers store a -3 (0xfffffffd) failure indication in rets[0] which corrupts host memory. Fix this by testing up front whether the guest supplied nargs and nret would exceed the array size, and fail the hcall directly without storing a failure indication to rets[0]. Also expand on a comment about why we kill the guest and try not to return errors directly if we have a valid rets[0] pointer. Fixes: 8e591cb72047 ("KVM: PPC: Book3S: Add infrastructure to implement kernel-side RTAS calls") Cc: stable@vger.kernel.org # v3.10+ Reported-by: Alexey Kardashevskiy Signed-off-by: Nicholas Piggin Signed-off-by: Michael Ellerman Signed-off-by: Greg Kroah-Hartman

KVM: PPC: Fix kvm_arch_vcpu_ioctl vcpu_load leak

2021-07-28T12:37:29Z

[ Upstream commit bc4188a2f56e821ea057aca6bf444e138d06c252 ] vcpu_put is not called if the user copy fails. This can result in preempt notifier corruption and crashes, among other issues. Fixes: b3cebfe8c1ca ("KVM: PPC: Move vcpu_load/vcpu_put down to each ioctl case in kvm_arch_vcpu_ioctl") Reported-by: Alexey Kardashevskiy Signed-off-by: Nicholas Piggin Signed-off-by: Michael Ellerman Link: https://lore.kernel.org/r/20210716024310.164448-2-npiggin@gmail.com Signed-off-by: Sasha Levin

KVM: PPC: Book3S: Fix CONFIG_TRANSACTIONAL_MEM=n crash

2021-07-28T12:37:29Z

[ Upstream commit bd31ecf44b8e18ccb1e5f6b50f85de6922a60de3 ] When running CPU_FTR_P9_TM_HV_ASSIST, HFSCR[TM] is set for the guest even if the host has CONFIG_TRANSACTIONAL_MEM=n, which causes it to be unprepared to handle guest exits while transactional. Normal guests don't have a problem because the HTM capability will not be advertised, but a rogue or buggy one could crash the host. Fixes: 4bb3c7a0208f ("KVM: PPC: Book3S HV: Work around transactional memory bugs in POWER9") Reported-by: Alexey Kardashevskiy Signed-off-by: Nicholas Piggin Signed-off-by: Michael Ellerman Link: https://lore.kernel.org/r/20210716024310.164448-1-npiggin@gmail.com Signed-off-by: Sasha Levin

KVM: PPC: Book3S HV: Fix TLB management on SMT8 POWER9 and POWER10 processors

2021-07-14T15:06:27Z

[ Upstream commit 77bbbc0cf84834ed130838f7ac1988567f4d0288 ] The POWER9 vCPU TLB management code assumes all threads in a core share a TLB, and that TLBIEL execued by one thread will invalidate TLBs for all threads. This is not the case for SMT8 capable POWER9 and POWER10 (big core) processors, where the TLB is split between groups of threads. This results in TLB multi-hits, random data corruption, etc. Fix this by introducing cpu_first_tlb_thread_sibling etc., to determine which siblings share TLBs, and use that in the guest TLB flushing code. [npiggin@gmail.com: add changelog and comment] Signed-off-by: Paul Mackerras Signed-off-by: Nicholas Piggin Reviewed-by: Fabiano Rosas Signed-off-by: Michael Ellerman Link: https://lore.kernel.org/r/20210602040441.3984352-1-npiggin@gmail.com Signed-off-by: Sasha Levin

KVM: PPC: Book3S HV: Workaround high stack usage with clang

2021-07-14T15:06:12Z

commit 51696f39cbee5bb684e7959c0c98b5f54548aa34 upstream. LLVM does not emit optimal byteswap assembly, which results in high stack usage in kvmhv_enter_nested_guest() due to the inlining of byteswap_pt_regs(). With LLVM 12.0.0: arch/powerpc/kvm/book3s_hv_nested.c:289:6: error: stack frame size of 2512 bytes in function 'kvmhv_enter_nested_guest' [-Werror,-Wframe-larger-than=] long kvmhv_enter_nested_guest(struct kvm_vcpu *vcpu) ^ 1 error generated. While this gets fixed in LLVM, mark byteswap_pt_regs() as noinline_for_stack so that it does not get inlined and break the build due to -Werror by default in arch/powerpc/. Not inlining saves approximately 800 bytes with LLVM 12.0.0: arch/powerpc/kvm/book3s_hv_nested.c:290:6: warning: stack frame size of 1728 bytes in function 'kvmhv_enter_nested_guest' [-Wframe-larger-than=] long kvmhv_enter_nested_guest(struct kvm_vcpu *vcpu) ^ 1 warning generated. Cc: stable@vger.kernel.org # v4.20+ Reported-by: kernel test robot Signed-off-by: Nathan Chancellor Signed-off-by: Michael Ellerman Link: https://github.com/ClangBuiltLinux/linux/issues/1292 Link: https://bugs.llvm.org/show_bug.cgi?id=49610 Link: https://lore.kernel.org/r/202104031853.vDT0Qjqj-lkp@intel.com/ Link: https://gist.github.com/ba710e3703bf45043a31e2806c843ffd Link: https://lore.kernel.org/r/20210621182440.990242-1-nathan@kernel.org Signed-off-by: Greg Kroah-Hartman

Merge tag 'powerpc-5.13-5' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux

2021-06-06T19:39:36Z

Pull powerpc fixes from Michael Ellerman: "Fix our KVM reverse map real-mode handling since we enabled huge vmalloc (in some configurations). Revert a recent change to our IOMMU code which broke some devices. Fix KVM handling of FSCR on P7/P8, which could have possibly let a guest crash it's Qemu. Fix kprobes validation of prefixed instructions across page boundary. Thanks to Alexey Kardashevskiy, Christophe Leroy, Fabiano Rosas, Frederic Barrat, Naveen N. Rao, and Nicholas Piggin" * tag 'powerpc-5.13-5' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux: Revert "powerpc/kernel/iommu: Align size for IOMMU_PAGE_SIZE() to save TCEs" KVM: PPC: Book3S HV: Save host FSCR in the P7/8 path powerpc: Fix reverse map real-mode address lookup with huge vmalloc powerpc/kprobes: Fix validation of prefixed instructions across page boundary