From 8df728e1ae614f592961e51f65d3e3212ede5a75 Mon Sep 17 00:00:00 2001 From: Robin Murphy Date: Fri, 12 May 2017 13:48:41 +0100 Subject: [PATCH 1/3] arm64: Remove redundant mov from LL/SC cmpxchg The cmpxchg implementation introduced by commit c342f78217e8 ("arm64: cmpxchg: patch in lse instructions when supported by the CPU") performs an apparently redundant register move of [old] to [oldval] in the success case - it always uses the same register width as [oldval] was originally loaded with, and is only executed when [old] and [oldval] are known to be equal anyway. The only effect it seemingly does have is to take up a surprising amount of space in the kernel text, as removing it reveals: text data bss dec hex filename 12426658 1348614 4499749 18275021 116dacd vmlinux.o.new 12429238 1348614 4499749 18277601 116e4e1 vmlinux.o.old Reviewed-by: Will Deacon Signed-off-by: Robin Murphy Signed-off-by: Catalin Marinas --- arch/arm64/include/asm/atomic_ll_sc.h | 1 - 1 file changed, 1 deletion(-) diff --git a/arch/arm64/include/asm/atomic_ll_sc.h b/arch/arm64/include/asm/atomic_ll_sc.h index f819fdcff1accf..f5a2d09afb3841 100644 --- a/arch/arm64/include/asm/atomic_ll_sc.h +++ b/arch/arm64/include/asm/atomic_ll_sc.h @@ -264,7 +264,6 @@ __LL_SC_PREFIX(__cmpxchg_case_##name(volatile void *ptr, \ " st" #rel "xr" #sz "\t%w[tmp], %" #w "[new], %[v]\n" \ " cbnz %w[tmp], 1b\n" \ " " #mb "\n" \ - " mov %" #w "[oldval], %" #w "[old]\n" \ "2:" \ : [tmp] "=&r" (tmp), [oldval] "=&r" (oldval), \ [v] "+Q" (*(unsigned long *)ptr) \ From 78a19cfdf37d19002c83c8790853c1cc10feccdc Mon Sep 17 00:00:00 2001 From: Ganapatrao Kulkarni Date: Tue, 2 May 2017 21:59:34 +0530 Subject: [PATCH 2/3] arm64: perf: Ignore exclude_hv when kernel is running in HYP commit d98ecdaca296 ("arm64: perf: Count EL2 events if the kernel is running in HYP") returns -EINVAL when perf system call perf_event_open is called with exclude_hv != exclude_kernel. This change breaks applications on VHE enabled ARMv8.1 platforms. The issue was observed with HHVM application, which calls perf_event_open with exclude_hv = 1 and exclude_kernel = 0. There is no separate hypervisor privilege level when VHE is enabled, the host kernel runs at EL2. So when VHE is enabled, we should ignore exclude_hv from the application. This behaviour is consistent with PowerPC where the exclude_hv is ignored when the hypervisor is not present and with x86 where this flag is ignored. Signed-off-by: Ganapatrao Kulkarni [will: added comment to justify the behaviour of exclude_hv] Signed-off-by: Will Deacon Signed-off-by: Catalin Marinas --- arch/arm64/kernel/perf_event.c | 23 ++++++++++++++++------- 1 file changed, 16 insertions(+), 7 deletions(-) diff --git a/arch/arm64/kernel/perf_event.c b/arch/arm64/kernel/perf_event.c index bcc79471b38e24..83a1b1ad189f51 100644 --- a/arch/arm64/kernel/perf_event.c +++ b/arch/arm64/kernel/perf_event.c @@ -877,15 +877,24 @@ static int armv8pmu_set_event_filter(struct hw_perf_event *event, if (attr->exclude_idle) return -EPERM; - if (is_kernel_in_hyp_mode() && - attr->exclude_kernel != attr->exclude_hv) - return -EINVAL; + + /* + * If we're running in hyp mode, then we *are* the hypervisor. + * Therefore we ignore exclude_hv in this configuration, since + * there's no hypervisor to sample anyway. This is consistent + * with other architectures (x86 and Power). + */ + if (is_kernel_in_hyp_mode()) { + if (!attr->exclude_kernel) + config_base |= ARMV8_PMU_INCLUDE_EL2; + } else { + if (attr->exclude_kernel) + config_base |= ARMV8_PMU_EXCLUDE_EL1; + if (!attr->exclude_hv) + config_base |= ARMV8_PMU_INCLUDE_EL2; + } if (attr->exclude_user) config_base |= ARMV8_PMU_EXCLUDE_EL0; - if (!is_kernel_in_hyp_mode() && attr->exclude_kernel) - config_base |= ARMV8_PMU_EXCLUDE_EL1; - if (!attr->exclude_hv) - config_base |= ARMV8_PMU_INCLUDE_EL2; /* * Install the filter into config_base as this is used to From 63a1e1c95e60e798fa09ab3c536fb555aa5bbf2b Mon Sep 17 00:00:00 2001 From: Mark Rutland Date: Tue, 16 May 2017 15:18:05 +0100 Subject: [PATCH 3/3] arm64/cpufeature: don't use mutex in bringup path Currently, cpus_set_cap() calls static_branch_enable_cpuslocked(), which must take the jump_label mutex. We call cpus_set_cap() in the secondary bringup path, from the idle thread where interrupts are disabled. Taking a mutex in this path "is a NONO" regardless of whether it's contended, and something we must avoid. We didn't spot this until recently, as ___might_sleep() won't warn for this case until all CPUs have been brought up. This patch avoids taking the mutex in the secondary bringup path. The poking of static keys is deferred until enable_cpu_capabilities(), which runs in a suitable context on the boot CPU. To account for the static keys being set later, cpus_have_const_cap() is updated to use another static key to check whether the const cap keys have been initialised, falling back to the caps bitmap until this is the case. This means that users of cpus_have_const_cap() gain should only gain a single additional NOP in the fast path once the const caps are initialised, but should always see the current cap value. The hyp code should never dereference the caps array, since the caps are initialized before we run the module initcall to initialise hyp. A check is added to the hyp init code to document this requirement. This change will sidestep a number of issues when the upcoming hotplug locking rework is merged. Signed-off-by: Mark Rutland Reviewed-by: Marc Zyniger Reviewed-by: Suzuki Poulose Acked-by: Will Deacon Cc: Christoffer Dall Cc: Peter Zijlstra Cc: Sebastian Sewior Cc: Thomas Gleixner Signed-off-by: Catalin Marinas --- arch/arm64/include/asm/cpufeature.h | 12 ++++++++++-- arch/arm64/include/asm/kvm_host.h | 8 ++++++-- arch/arm64/kernel/cpufeature.c | 23 +++++++++++++++++++++-- 3 files changed, 37 insertions(+), 6 deletions(-) diff --git a/arch/arm64/include/asm/cpufeature.h b/arch/arm64/include/asm/cpufeature.h index e7f84a7b44658d..428ee1f2468c55 100644 --- a/arch/arm64/include/asm/cpufeature.h +++ b/arch/arm64/include/asm/cpufeature.h @@ -115,6 +115,7 @@ struct arm64_cpu_capabilities { extern DECLARE_BITMAP(cpu_hwcaps, ARM64_NCAPS); extern struct static_key_false cpu_hwcap_keys[ARM64_NCAPS]; +extern struct static_key_false arm64_const_caps_ready; bool this_cpu_has_cap(unsigned int cap); @@ -124,7 +125,7 @@ static inline bool cpu_have_feature(unsigned int num) } /* System capability check for constant caps */ -static inline bool cpus_have_const_cap(int num) +static inline bool __cpus_have_const_cap(int num) { if (num >= ARM64_NCAPS) return false; @@ -138,6 +139,14 @@ static inline bool cpus_have_cap(unsigned int num) return test_bit(num, cpu_hwcaps); } +static inline bool cpus_have_const_cap(int num) +{ + if (static_branch_likely(&arm64_const_caps_ready)) + return __cpus_have_const_cap(num); + else + return cpus_have_cap(num); +} + static inline void cpus_set_cap(unsigned int num) { if (num >= ARM64_NCAPS) { @@ -145,7 +154,6 @@ static inline void cpus_set_cap(unsigned int num) num, ARM64_NCAPS); } else { __set_bit(num, cpu_hwcaps); - static_branch_enable(&cpu_hwcap_keys[num]); } } diff --git a/arch/arm64/include/asm/kvm_host.h b/arch/arm64/include/asm/kvm_host.h index 5e19165c5fa8b8..1f252a95bc0297 100644 --- a/arch/arm64/include/asm/kvm_host.h +++ b/arch/arm64/include/asm/kvm_host.h @@ -24,6 +24,7 @@ #include #include +#include #include #include #include @@ -355,9 +356,12 @@ static inline void __cpu_init_hyp_mode(phys_addr_t pgd_ptr, unsigned long vector_ptr) { /* - * Call initialization code, and switch to the full blown - * HYP code. + * Call initialization code, and switch to the full blown HYP code. + * If the cpucaps haven't been finalized yet, something has gone very + * wrong, and hyp will crash and burn when it uses any + * cpus_have_const_cap() wrapper. */ + BUG_ON(!static_branch_likely(&arm64_const_caps_ready)); __kvm_call_hyp((void *)pgd_ptr, hyp_stack_ptr, vector_ptr); } diff --git a/arch/arm64/kernel/cpufeature.c b/arch/arm64/kernel/cpufeature.c index 94b8f7fc33100f..817ce3365e200d 100644 --- a/arch/arm64/kernel/cpufeature.c +++ b/arch/arm64/kernel/cpufeature.c @@ -985,8 +985,16 @@ void update_cpu_capabilities(const struct arm64_cpu_capabilities *caps, */ void __init enable_cpu_capabilities(const struct arm64_cpu_capabilities *caps) { - for (; caps->matches; caps++) - if (caps->enable && cpus_have_cap(caps->capability)) + for (; caps->matches; caps++) { + unsigned int num = caps->capability; + + if (!cpus_have_cap(num)) + continue; + + /* Ensure cpus_have_const_cap(num) works */ + static_branch_enable(&cpu_hwcap_keys[num]); + + if (caps->enable) { /* * Use stop_machine() as it schedules the work allowing * us to modify PSTATE, instead of on_each_cpu() which @@ -994,6 +1002,8 @@ void __init enable_cpu_capabilities(const struct arm64_cpu_capabilities *caps) * we return. */ stop_machine(caps->enable, NULL, cpu_online_mask); + } + } } /* @@ -1096,6 +1106,14 @@ static void __init setup_feature_capabilities(void) enable_cpu_capabilities(arm64_features); } +DEFINE_STATIC_KEY_FALSE(arm64_const_caps_ready); +EXPORT_SYMBOL(arm64_const_caps_ready); + +static void __init mark_const_caps_ready(void) +{ + static_branch_enable(&arm64_const_caps_ready); +} + /* * Check if the current CPU has a given feature capability. * Should be called from non-preemptible context. @@ -1131,6 +1149,7 @@ void __init setup_cpu_features(void) /* Set the CPU feature capabilies */ setup_feature_capabilities(); enable_errata_workarounds(); + mark_const_caps_ready(); setup_elf_hwcaps(arm64_elf_hwcaps); if (system_supports_32bit_el0())