* [PATCH v2 1/8] arm64: cpufeature: Relax check for IESB support
2020-04-21 14:29 [PATCH v2 0/8] Relax sanity checking for mismatched AArch32 EL1 Will Deacon
@ 2020-04-21 14:29 ` Will Deacon
2020-04-21 14:29 ` [PATCH v2 2/8] arm64: cpufeature: Spell out register fields for ID_ISAR4 and ID_PFR1 Will Deacon
` (6 subsequent siblings)
7 siblings, 0 replies; 10+ messages in thread
From: Will Deacon @ 2020-04-21 14:29 UTC (permalink / raw)
To: linux-arm-kernel, kvmarm
Cc: linux-kernel, Will Deacon, Suzuki K Poulose, Mark Rutland,
Marc Zyngier, Anshuman Khandual, Catalin Marinas,
Sai Prakash Ranjan, Doug Anderson, kernel-team
From: Sai Prakash Ranjan <saiprakash.ranjan@codeaurora.org>
We don't care if IESB is supported or not as we always set
SCTLR_ELx.IESB and, if it works, that's really great.
Relax the ID_AA64MMFR2.IESB cpufeature check so that we don't warn and
taint if it's mismatched.
Reviewed-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Signed-off-by: Sai Prakash Ranjan <saiprakash.ranjan@codeaurora.org>
[will: rewrote commit message]
Signed-off-by: Will Deacon <will@kernel.org>
---
arch/arm64/kernel/cpufeature.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/arch/arm64/kernel/cpufeature.c b/arch/arm64/kernel/cpufeature.c
index 9fac745aa7bb..63df28e6a425 100644
--- a/arch/arm64/kernel/cpufeature.c
+++ b/arch/arm64/kernel/cpufeature.c
@@ -247,7 +247,7 @@ static const struct arm64_ftr_bits ftr_id_aa64mmfr2[] = {
ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_AA64MMFR2_FWB_SHIFT, 4, 0),
ARM64_FTR_BITS(FTR_VISIBLE, FTR_STRICT, FTR_LOWER_SAFE, ID_AA64MMFR2_AT_SHIFT, 4, 0),
ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_AA64MMFR2_LVA_SHIFT, 4, 0),
- ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_AA64MMFR2_IESB_SHIFT, 4, 0),
+ ARM64_FTR_BITS(FTR_HIDDEN, FTR_NONSTRICT, FTR_LOWER_SAFE, ID_AA64MMFR2_IESB_SHIFT, 4, 0),
ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_AA64MMFR2_LSM_SHIFT, 4, 0),
ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_AA64MMFR2_UAO_SHIFT, 4, 0),
ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_AA64MMFR2_CNP_SHIFT, 4, 0),
--
2.26.1.301.g55bc3eb7cb9-goog
^ permalink raw reply related [flat|nested] 10+ messages in thread* [PATCH v2 2/8] arm64: cpufeature: Spell out register fields for ID_ISAR4 and ID_PFR1
2020-04-21 14:29 [PATCH v2 0/8] Relax sanity checking for mismatched AArch32 EL1 Will Deacon
2020-04-21 14:29 ` [PATCH v2 1/8] arm64: cpufeature: Relax check for IESB support Will Deacon
@ 2020-04-21 14:29 ` Will Deacon
2020-04-21 14:29 ` [PATCH v2 3/8] arm64: cpufeature: Add CPU capability for AArch32 EL1 support Will Deacon
` (5 subsequent siblings)
7 siblings, 0 replies; 10+ messages in thread
From: Will Deacon @ 2020-04-21 14:29 UTC (permalink / raw)
To: linux-arm-kernel, kvmarm
Cc: linux-kernel, Will Deacon, Suzuki K Poulose, Mark Rutland,
Marc Zyngier, Anshuman Khandual, Catalin Marinas,
Sai Prakash Ranjan, Doug Anderson, kernel-team
In preparation for runtime updates to the strictness of some AArch32
features, spell out the register fields for ID_ISAR4 and ID_PFR1 to make
things clearer to read. Note that this isn't functionally necessary, as
the feature arrays themselves are not modified dynamically and remain
'const'.
Reviewed-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Tested-by: Sai Prakash Ranjan <saiprakash.ranjan@codeaurora.org>
Signed-off-by: Will Deacon <will@kernel.org>
---
arch/arm64/include/asm/sysreg.h | 17 +++++++++++++++++
arch/arm64/kernel/cpufeature.c | 28 ++++++++++++++++++++++++++--
2 files changed, 43 insertions(+), 2 deletions(-)
diff --git a/arch/arm64/include/asm/sysreg.h b/arch/arm64/include/asm/sysreg.h
index c4ac0ac25a00..d7181972d28d 100644
--- a/arch/arm64/include/asm/sysreg.h
+++ b/arch/arm64/include/asm/sysreg.h
@@ -752,6 +752,15 @@
#define ID_DFR0_PERFMON_8_1 0x4
+#define ID_ISAR4_SWP_FRAC_SHIFT 28
+#define ID_ISAR4_PSR_M_SHIFT 24
+#define ID_ISAR4_SYNCH_PRIM_FRAC_SHIFT 20
+#define ID_ISAR4_BARRIER_SHIFT 16
+#define ID_ISAR4_SMC_SHIFT 12
+#define ID_ISAR4_WRITEBACK_SHIFT 8
+#define ID_ISAR4_WITHSHIFTS_SHIFT 4
+#define ID_ISAR4_UNPRIV_SHIFT 0
+
#define ID_ISAR5_RDM_SHIFT 24
#define ID_ISAR5_CRC32_SHIFT 16
#define ID_ISAR5_SHA2_SHIFT 12
@@ -785,6 +794,14 @@
#define MVFR1_FPDNAN_SHIFT 4
#define MVFR1_FPFTZ_SHIFT 0
+#define ID_PFR1_GIC_SHIFT 28
+#define ID_PFR1_VIRT_FRAC_SHIFT 24
+#define ID_PFR1_SEC_FRAC_SHIFT 20
+#define ID_PFR1_GENTIMER_SHIFT 16
+#define ID_PFR1_VIRTUALIZATION_SHIFT 12
+#define ID_PFR1_MPROGMOD_SHIFT 8
+#define ID_PFR1_SECURITY_SHIFT 4
+#define ID_PFR1_PROGMOD_SHIFT 0
#define ID_AA64MMFR0_TGRAN4_SHIFT 28
#define ID_AA64MMFR0_TGRAN64_SHIFT 24
diff --git a/arch/arm64/kernel/cpufeature.c b/arch/arm64/kernel/cpufeature.c
index 63df28e6a425..b143f8bc6c52 100644
--- a/arch/arm64/kernel/cpufeature.c
+++ b/arch/arm64/kernel/cpufeature.c
@@ -332,6 +332,18 @@ static const struct arm64_ftr_bits ftr_id_mmfr4[] = {
ARM64_FTR_END,
};
+static const struct arm64_ftr_bits ftr_id_isar4[] = {
+ ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_ISAR4_SWP_FRAC_SHIFT, 4, 0),
+ ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_ISAR4_PSR_M_SHIFT, 4, 0),
+ ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_ISAR4_SYNCH_PRIM_FRAC_SHIFT, 4, 0),
+ ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_ISAR4_BARRIER_SHIFT, 4, 0),
+ ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_ISAR4_SMC_SHIFT, 4, 0),
+ ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_ISAR4_WRITEBACK_SHIFT, 4, 0),
+ ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_ISAR4_WITHSHIFTS_SHIFT, 4, 0),
+ ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_ISAR4_UNPRIV_SHIFT, 4, 0),
+ ARM64_FTR_END,
+};
+
static const struct arm64_ftr_bits ftr_id_isar6[] = {
ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_ISAR6_I8MM_SHIFT, 4, 0),
ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_ISAR6_BF16_SHIFT, 4, 0),
@@ -351,6 +363,18 @@ static const struct arm64_ftr_bits ftr_id_pfr0[] = {
ARM64_FTR_END,
};
+static const struct arm64_ftr_bits ftr_id_pfr1[] = {
+ ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_PFR1_GIC_SHIFT, 4, 0),
+ ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_PFR1_VIRT_FRAC_SHIFT, 4, 0),
+ ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_PFR1_SEC_FRAC_SHIFT, 4, 0),
+ ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_PFR1_GENTIMER_SHIFT, 4, 0),
+ ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_PFR1_VIRTUALIZATION_SHIFT, 4, 0),
+ ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_PFR1_MPROGMOD_SHIFT, 4, 0),
+ ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_PFR1_SECURITY_SHIFT, 4, 0),
+ ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_PFR1_PROGMOD_SHIFT, 4, 0),
+ ARM64_FTR_END,
+};
+
static const struct arm64_ftr_bits ftr_id_dfr0[] = {
ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, 28, 4, 0),
S_ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, 24, 4, 0xf), /* PerfMon */
@@ -411,7 +435,7 @@ static const struct __ftr_reg_entry {
/* Op1 = 0, CRn = 0, CRm = 1 */
ARM64_FTR_REG(SYS_ID_PFR0_EL1, ftr_id_pfr0),
- ARM64_FTR_REG(SYS_ID_PFR1_EL1, ftr_generic_32bits),
+ ARM64_FTR_REG(SYS_ID_PFR1_EL1, ftr_id_pfr1),
ARM64_FTR_REG(SYS_ID_DFR0_EL1, ftr_id_dfr0),
ARM64_FTR_REG(SYS_ID_MMFR0_EL1, ftr_id_mmfr0),
ARM64_FTR_REG(SYS_ID_MMFR1_EL1, ftr_generic_32bits),
@@ -423,7 +447,7 @@ static const struct __ftr_reg_entry {
ARM64_FTR_REG(SYS_ID_ISAR1_EL1, ftr_generic_32bits),
ARM64_FTR_REG(SYS_ID_ISAR2_EL1, ftr_generic_32bits),
ARM64_FTR_REG(SYS_ID_ISAR3_EL1, ftr_generic_32bits),
- ARM64_FTR_REG(SYS_ID_ISAR4_EL1, ftr_generic_32bits),
+ ARM64_FTR_REG(SYS_ID_ISAR4_EL1, ftr_id_isar4),
ARM64_FTR_REG(SYS_ID_ISAR5_EL1, ftr_id_isar5),
ARM64_FTR_REG(SYS_ID_MMFR4_EL1, ftr_id_mmfr4),
ARM64_FTR_REG(SYS_ID_ISAR6_EL1, ftr_id_isar6),
--
2.26.1.301.g55bc3eb7cb9-goog
^ permalink raw reply related [flat|nested] 10+ messages in thread* [PATCH v2 3/8] arm64: cpufeature: Add CPU capability for AArch32 EL1 support
2020-04-21 14:29 [PATCH v2 0/8] Relax sanity checking for mismatched AArch32 EL1 Will Deacon
2020-04-21 14:29 ` [PATCH v2 1/8] arm64: cpufeature: Relax check for IESB support Will Deacon
2020-04-21 14:29 ` [PATCH v2 2/8] arm64: cpufeature: Spell out register fields for ID_ISAR4 and ID_PFR1 Will Deacon
@ 2020-04-21 14:29 ` Will Deacon
2020-04-21 14:29 ` [PATCH v2 4/8] arm64: cpufeature: Remove redundant call to id_aa64pfr0_32bit_el0() Will Deacon
` (4 subsequent siblings)
7 siblings, 0 replies; 10+ messages in thread
From: Will Deacon @ 2020-04-21 14:29 UTC (permalink / raw)
To: linux-arm-kernel, kvmarm
Cc: linux-kernel, Will Deacon, Suzuki K Poulose, Mark Rutland,
Marc Zyngier, Anshuman Khandual, Catalin Marinas,
Sai Prakash Ranjan, Doug Anderson, kernel-team
Although we emit a "SANITY CHECK" warning and taint the kernel if we
detect a CPU mismatch for AArch32 support at EL1, we still online the
CPU with disastrous consequences for any running 32-bit VMs.
Introduce a capability for AArch32 support at EL1 so that late onlining
of incompatible CPUs is forbidden.
Acked-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Tested-by: Sai Prakash Ranjan <saiprakash.ranjan@codeaurora.org>
Signed-off-by: Will Deacon <will@kernel.org>
---
arch/arm64/include/asm/cpucaps.h | 3 ++-
arch/arm64/include/asm/sysreg.h | 1 +
arch/arm64/kernel/cpufeature.c | 12 ++++++++++++
arch/arm64/kvm/reset.c | 12 ++----------
4 files changed, 17 insertions(+), 11 deletions(-)
diff --git a/arch/arm64/include/asm/cpucaps.h b/arch/arm64/include/asm/cpucaps.h
index 8eb5a088ae65..c54c674e6c21 100644
--- a/arch/arm64/include/asm/cpucaps.h
+++ b/arch/arm64/include/asm/cpucaps.h
@@ -61,7 +61,8 @@
#define ARM64_HAS_AMU_EXTN 51
#define ARM64_HAS_ADDRESS_AUTH 52
#define ARM64_HAS_GENERIC_AUTH 53
+#define ARM64_HAS_32BIT_EL1 54
-#define ARM64_NCAPS 54
+#define ARM64_NCAPS 55
#endif /* __ASM_CPUCAPS_H */
diff --git a/arch/arm64/include/asm/sysreg.h b/arch/arm64/include/asm/sysreg.h
index d7181972d28d..c4e896bf77f3 100644
--- a/arch/arm64/include/asm/sysreg.h
+++ b/arch/arm64/include/asm/sysreg.h
@@ -655,6 +655,7 @@
#define ID_AA64PFR0_ASIMD_NI 0xf
#define ID_AA64PFR0_ASIMD_SUPPORTED 0x0
#define ID_AA64PFR0_EL1_64BIT_ONLY 0x1
+#define ID_AA64PFR0_EL1_32BIT_64BIT 0x2
#define ID_AA64PFR0_EL0_64BIT_ONLY 0x1
#define ID_AA64PFR0_EL0_32BIT_64BIT 0x2
diff --git a/arch/arm64/kernel/cpufeature.c b/arch/arm64/kernel/cpufeature.c
index b143f8bc6c52..838fe5cc8d7e 100644
--- a/arch/arm64/kernel/cpufeature.c
+++ b/arch/arm64/kernel/cpufeature.c
@@ -1535,6 +1535,18 @@ static const struct arm64_cpu_capabilities arm64_features[] = {
.field_pos = ID_AA64PFR0_EL0_SHIFT,
.min_field_value = ID_AA64PFR0_EL0_32BIT_64BIT,
},
+#ifdef CONFIG_KVM
+ {
+ .desc = "32-bit EL1 Support",
+ .capability = ARM64_HAS_32BIT_EL1,
+ .type = ARM64_CPUCAP_SYSTEM_FEATURE,
+ .matches = has_cpuid_feature,
+ .sys_reg = SYS_ID_AA64PFR0_EL1,
+ .sign = FTR_UNSIGNED,
+ .field_pos = ID_AA64PFR0_EL1_SHIFT,
+ .min_field_value = ID_AA64PFR0_EL1_32BIT_64BIT,
+ },
+#endif
{
.desc = "Kernel page table isolation (KPTI)",
.capability = ARM64_UNMAP_KERNEL_AT_EL0,
diff --git a/arch/arm64/kvm/reset.c b/arch/arm64/kvm/reset.c
index 30b7ea680f66..102e5c4e01a0 100644
--- a/arch/arm64/kvm/reset.c
+++ b/arch/arm64/kvm/reset.c
@@ -46,14 +46,6 @@ static const struct kvm_regs default_regs_reset32 = {
PSR_AA32_I_BIT | PSR_AA32_F_BIT),
};
-static bool cpu_has_32bit_el1(void)
-{
- u64 pfr0;
-
- pfr0 = read_sanitised_ftr_reg(SYS_ID_AA64PFR0_EL1);
- return !!(pfr0 & 0x20);
-}
-
/**
* kvm_arch_vm_ioctl_check_extension
*
@@ -66,7 +58,7 @@ int kvm_arch_vm_ioctl_check_extension(struct kvm *kvm, long ext)
switch (ext) {
case KVM_CAP_ARM_EL1_32BIT:
- r = cpu_has_32bit_el1();
+ r = cpus_have_const_cap(ARM64_HAS_32BIT_EL1);
break;
case KVM_CAP_GUEST_DEBUG_HW_BPS:
r = get_num_brps();
@@ -288,7 +280,7 @@ int kvm_reset_vcpu(struct kvm_vcpu *vcpu)
switch (vcpu->arch.target) {
default:
if (test_bit(KVM_ARM_VCPU_EL1_32BIT, vcpu->arch.features)) {
- if (!cpu_has_32bit_el1())
+ if (!cpus_have_const_cap(ARM64_HAS_32BIT_EL1))
goto out;
cpu_reset = &default_regs_reset32;
} else {
--
2.26.1.301.g55bc3eb7cb9-goog
^ permalink raw reply related [flat|nested] 10+ messages in thread* [PATCH v2 4/8] arm64: cpufeature: Remove redundant call to id_aa64pfr0_32bit_el0()
2020-04-21 14:29 [PATCH v2 0/8] Relax sanity checking for mismatched AArch32 EL1 Will Deacon
` (2 preceding siblings ...)
2020-04-21 14:29 ` [PATCH v2 3/8] arm64: cpufeature: Add CPU capability for AArch32 EL1 support Will Deacon
@ 2020-04-21 14:29 ` Will Deacon
2020-04-21 14:29 ` [PATCH v2 5/8] arm64: cpufeature: Factor out checking of AArch32 features Will Deacon
` (3 subsequent siblings)
7 siblings, 0 replies; 10+ messages in thread
From: Will Deacon @ 2020-04-21 14:29 UTC (permalink / raw)
To: linux-arm-kernel, kvmarm
Cc: linux-kernel, Will Deacon, Suzuki K Poulose, Mark Rutland,
Marc Zyngier, Anshuman Khandual, Catalin Marinas,
Sai Prakash Ranjan, Doug Anderson, kernel-team
There's no need to call id_aa64pfr0_32bit_el0() twice because the
sanitised value of ID_AA64PFR0_EL1 has already been updated for the CPU
being onlined.
Remove the redundant function call.
Reviewed-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Tested-by: Sai Prakash Ranjan <saiprakash.ranjan@codeaurora.org>
Signed-off-by: Will Deacon <will@kernel.org>
---
arch/arm64/kernel/cpufeature.c | 4 +---
1 file changed, 1 insertion(+), 3 deletions(-)
diff --git a/arch/arm64/kernel/cpufeature.c b/arch/arm64/kernel/cpufeature.c
index 838fe5cc8d7e..7dfcdd9e75c1 100644
--- a/arch/arm64/kernel/cpufeature.c
+++ b/arch/arm64/kernel/cpufeature.c
@@ -792,9 +792,7 @@ void update_cpu_features(int cpu,
* If we have AArch32, we care about 32-bit features for compat.
* If the system doesn't support AArch32, don't update them.
*/
- if (id_aa64pfr0_32bit_el0(read_sanitised_ftr_reg(SYS_ID_AA64PFR0_EL1)) &&
- id_aa64pfr0_32bit_el0(info->reg_id_aa64pfr0)) {
-
+ if (id_aa64pfr0_32bit_el0(read_sanitised_ftr_reg(SYS_ID_AA64PFR0_EL1))) {
taint |= check_update_ftr_reg(SYS_ID_DFR0_EL1, cpu,
info->reg_id_dfr0, boot->reg_id_dfr0);
taint |= check_update_ftr_reg(SYS_ID_ISAR0_EL1, cpu,
--
2.26.1.301.g55bc3eb7cb9-goog
^ permalink raw reply related [flat|nested] 10+ messages in thread* [PATCH v2 5/8] arm64: cpufeature: Factor out checking of AArch32 features
2020-04-21 14:29 [PATCH v2 0/8] Relax sanity checking for mismatched AArch32 EL1 Will Deacon
` (3 preceding siblings ...)
2020-04-21 14:29 ` [PATCH v2 4/8] arm64: cpufeature: Remove redundant call to id_aa64pfr0_32bit_el0() Will Deacon
@ 2020-04-21 14:29 ` Will Deacon
2020-04-21 14:29 ` [PATCH v2 6/8] arm64: cpufeature: Relax AArch32 system checks if EL1 is 64-bit only Will Deacon
` (2 subsequent siblings)
7 siblings, 0 replies; 10+ messages in thread
From: Will Deacon @ 2020-04-21 14:29 UTC (permalink / raw)
To: linux-arm-kernel, kvmarm
Cc: linux-kernel, Will Deacon, Suzuki K Poulose, Mark Rutland,
Marc Zyngier, Anshuman Khandual, Catalin Marinas,
Sai Prakash Ranjan, Doug Anderson, kernel-team
update_cpu_features() is pretty large, so split out the checking of the
AArch32 features into a separate function and call it after checking the
AArch64 features.
Reviewed-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Tested-by: Sai Prakash Ranjan <saiprakash.ranjan@codeaurora.org>
Signed-off-by: Will Deacon <will@kernel.org>
---
arch/arm64/kernel/cpufeature.c | 112 +++++++++++++++++++--------------
1 file changed, 65 insertions(+), 47 deletions(-)
diff --git a/arch/arm64/kernel/cpufeature.c b/arch/arm64/kernel/cpufeature.c
index 7dfcdd9e75c1..6892b2440676 100644
--- a/arch/arm64/kernel/cpufeature.c
+++ b/arch/arm64/kernel/cpufeature.c
@@ -715,6 +715,65 @@ static int check_update_ftr_reg(u32 sys_id, int cpu, u64 val, u64 boot)
return 1;
}
+static int update_32bit_cpu_features(int cpu, struct cpuinfo_arm64 *info,
+ struct cpuinfo_arm64 *boot)
+{
+ int taint = 0;
+ u64 pfr0 = read_sanitised_ftr_reg(SYS_ID_AA64PFR0_EL1);
+
+ /*
+ * If we don't have AArch32 at all then skip the checks entirely
+ * as the register values may be UNKNOWN and we're not going to be
+ * using them for anything.
+ */
+ if (!id_aa64pfr0_32bit_el0(pfr0))
+ return taint;
+
+ taint |= check_update_ftr_reg(SYS_ID_DFR0_EL1, cpu,
+ info->reg_id_dfr0, boot->reg_id_dfr0);
+ taint |= check_update_ftr_reg(SYS_ID_ISAR0_EL1, cpu,
+ info->reg_id_isar0, boot->reg_id_isar0);
+ taint |= check_update_ftr_reg(SYS_ID_ISAR1_EL1, cpu,
+ info->reg_id_isar1, boot->reg_id_isar1);
+ taint |= check_update_ftr_reg(SYS_ID_ISAR2_EL1, cpu,
+ info->reg_id_isar2, boot->reg_id_isar2);
+ taint |= check_update_ftr_reg(SYS_ID_ISAR3_EL1, cpu,
+ info->reg_id_isar3, boot->reg_id_isar3);
+ taint |= check_update_ftr_reg(SYS_ID_ISAR4_EL1, cpu,
+ info->reg_id_isar4, boot->reg_id_isar4);
+ taint |= check_update_ftr_reg(SYS_ID_ISAR5_EL1, cpu,
+ info->reg_id_isar5, boot->reg_id_isar5);
+ taint |= check_update_ftr_reg(SYS_ID_ISAR6_EL1, cpu,
+ info->reg_id_isar6, boot->reg_id_isar6);
+
+ /*
+ * Regardless of the value of the AuxReg field, the AIFSR, ADFSR, and
+ * ACTLR formats could differ across CPUs and therefore would have to
+ * be trapped for virtualization anyway.
+ */
+ taint |= check_update_ftr_reg(SYS_ID_MMFR0_EL1, cpu,
+ info->reg_id_mmfr0, boot->reg_id_mmfr0);
+ taint |= check_update_ftr_reg(SYS_ID_MMFR1_EL1, cpu,
+ info->reg_id_mmfr1, boot->reg_id_mmfr1);
+ taint |= check_update_ftr_reg(SYS_ID_MMFR2_EL1, cpu,
+ info->reg_id_mmfr2, boot->reg_id_mmfr2);
+ taint |= check_update_ftr_reg(SYS_ID_MMFR3_EL1, cpu,
+ info->reg_id_mmfr3, boot->reg_id_mmfr3);
+ taint |= check_update_ftr_reg(SYS_ID_PFR0_EL1, cpu,
+ info->reg_id_pfr0, boot->reg_id_pfr0);
+ taint |= check_update_ftr_reg(SYS_ID_PFR1_EL1, cpu,
+ info->reg_id_pfr1, boot->reg_id_pfr1);
+ taint |= check_update_ftr_reg(SYS_MVFR0_EL1, cpu,
+ info->reg_mvfr0, boot->reg_mvfr0);
+ taint |= check_update_ftr_reg(SYS_MVFR1_EL1, cpu,
+ info->reg_mvfr1, boot->reg_mvfr1);
+ taint |= check_update_ftr_reg(SYS_MVFR2_EL1, cpu,
+ info->reg_mvfr2, boot->reg_mvfr2);
+
+ return taint;
+}
+
+
/*
* Update system wide CPU feature registers with the values from a
* non-boot CPU. Also performs SANITY checks to make sure that there
@@ -788,53 +847,6 @@ void update_cpu_features(int cpu,
taint |= check_update_ftr_reg(SYS_ID_AA64ZFR0_EL1, cpu,
info->reg_id_aa64zfr0, boot->reg_id_aa64zfr0);
- /*
- * If we have AArch32, we care about 32-bit features for compat.
- * If the system doesn't support AArch32, don't update them.
- */
- if (id_aa64pfr0_32bit_el0(read_sanitised_ftr_reg(SYS_ID_AA64PFR0_EL1))) {
- taint |= check_update_ftr_reg(SYS_ID_DFR0_EL1, cpu,
- info->reg_id_dfr0, boot->reg_id_dfr0);
- taint |= check_update_ftr_reg(SYS_ID_ISAR0_EL1, cpu,
- info->reg_id_isar0, boot->reg_id_isar0);
- taint |= check_update_ftr_reg(SYS_ID_ISAR1_EL1, cpu,
- info->reg_id_isar1, boot->reg_id_isar1);
- taint |= check_update_ftr_reg(SYS_ID_ISAR2_EL1, cpu,
- info->reg_id_isar2, boot->reg_id_isar2);
- taint |= check_update_ftr_reg(SYS_ID_ISAR3_EL1, cpu,
- info->reg_id_isar3, boot->reg_id_isar3);
- taint |= check_update_ftr_reg(SYS_ID_ISAR4_EL1, cpu,
- info->reg_id_isar4, boot->reg_id_isar4);
- taint |= check_update_ftr_reg(SYS_ID_ISAR5_EL1, cpu,
- info->reg_id_isar5, boot->reg_id_isar5);
- taint |= check_update_ftr_reg(SYS_ID_ISAR6_EL1, cpu,
- info->reg_id_isar6, boot->reg_id_isar6);
-
- /*
- * Regardless of the value of the AuxReg field, the AIFSR, ADFSR, and
- * ACTLR formats could differ across CPUs and therefore would have to
- * be trapped for virtualization anyway.
- */
- taint |= check_update_ftr_reg(SYS_ID_MMFR0_EL1, cpu,
- info->reg_id_mmfr0, boot->reg_id_mmfr0);
- taint |= check_update_ftr_reg(SYS_ID_MMFR1_EL1, cpu,
- info->reg_id_mmfr1, boot->reg_id_mmfr1);
- taint |= check_update_ftr_reg(SYS_ID_MMFR2_EL1, cpu,
- info->reg_id_mmfr2, boot->reg_id_mmfr2);
- taint |= check_update_ftr_reg(SYS_ID_MMFR3_EL1, cpu,
- info->reg_id_mmfr3, boot->reg_id_mmfr3);
- taint |= check_update_ftr_reg(SYS_ID_PFR0_EL1, cpu,
- info->reg_id_pfr0, boot->reg_id_pfr0);
- taint |= check_update_ftr_reg(SYS_ID_PFR1_EL1, cpu,
- info->reg_id_pfr1, boot->reg_id_pfr1);
- taint |= check_update_ftr_reg(SYS_MVFR0_EL1, cpu,
- info->reg_mvfr0, boot->reg_mvfr0);
- taint |= check_update_ftr_reg(SYS_MVFR1_EL1, cpu,
- info->reg_mvfr1, boot->reg_mvfr1);
- taint |= check_update_ftr_reg(SYS_MVFR2_EL1, cpu,
- info->reg_mvfr2, boot->reg_mvfr2);
- }
-
if (id_aa64pfr0_sve(info->reg_id_aa64pfr0)) {
taint |= check_update_ftr_reg(SYS_ZCR_EL1, cpu,
info->reg_zcr, boot->reg_zcr);
@@ -845,6 +857,12 @@ void update_cpu_features(int cpu,
sve_update_vq_map();
}
+ /*
+ * This relies on a sanitised view of the AArch64 ID registers
+ * (e.g. SYS_ID_AA64PFR0_EL1), so we call it last.
+ */
+ taint |= update_32bit_cpu_features(cpu, info, boot);
+
/*
* Mismatched CPU features are a recipe for disaster. Don't even
* pretend to support them.
--
2.26.1.301.g55bc3eb7cb9-goog
^ permalink raw reply related [flat|nested] 10+ messages in thread* [PATCH v2 6/8] arm64: cpufeature: Relax AArch32 system checks if EL1 is 64-bit only
2020-04-21 14:29 [PATCH v2 0/8] Relax sanity checking for mismatched AArch32 EL1 Will Deacon
` (4 preceding siblings ...)
2020-04-21 14:29 ` [PATCH v2 5/8] arm64: cpufeature: Factor out checking of AArch32 features Will Deacon
@ 2020-04-21 14:29 ` Will Deacon
2020-04-21 14:29 ` [PATCH v2 7/8] arm64: cpufeature: Relax checks for AArch32 support at EL[0-2] Will Deacon
2020-04-21 14:29 ` [PATCH v2 8/8] arm64: cpufeature: Add an overview comment for the cpufeature framework Will Deacon
7 siblings, 0 replies; 10+ messages in thread
From: Will Deacon @ 2020-04-21 14:29 UTC (permalink / raw)
To: linux-arm-kernel, kvmarm
Cc: linux-kernel, Will Deacon, Suzuki K Poulose, Mark Rutland,
Marc Zyngier, Anshuman Khandual, Catalin Marinas,
Sai Prakash Ranjan, Doug Anderson, kernel-team
If AArch32 is not supported at EL1, the AArch32 feature register fields
no longer advertise support for some system features:
* ISAR4.SMC
* PFR1.{Virt_frac, Sec_frac, Virtualization, Security, ProgMod}
In which case, we don't need to emit "SANITY CHECK" failures for all of
them.
Add logic to relax the strictness of individual feature register fields
at runtime and use this for the fields above if 32-bit EL1 is not
supported.
Reviewed-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Tested-by: Sai Prakash Ranjan <saiprakash.ranjan@codeaurora.org>
Signed-off-by: Will Deacon <will@kernel.org>
---
arch/arm64/include/asm/cpufeature.h | 7 ++++++
arch/arm64/kernel/cpufeature.c | 33 ++++++++++++++++++++++++++++-
2 files changed, 39 insertions(+), 1 deletion(-)
diff --git a/arch/arm64/include/asm/cpufeature.h b/arch/arm64/include/asm/cpufeature.h
index afe08251ff95..f5c4672e498b 100644
--- a/arch/arm64/include/asm/cpufeature.h
+++ b/arch/arm64/include/asm/cpufeature.h
@@ -551,6 +551,13 @@ static inline bool id_aa64mmfr0_mixed_endian_el0(u64 mmfr0)
cpuid_feature_extract_unsigned_field(mmfr0, ID_AA64MMFR0_BIGENDEL0_SHIFT) == 0x1;
}
+static inline bool id_aa64pfr0_32bit_el1(u64 pfr0)
+{
+ u32 val = cpuid_feature_extract_unsigned_field(pfr0, ID_AA64PFR0_EL1_SHIFT);
+
+ return val == ID_AA64PFR0_EL1_32BIT_64BIT;
+}
+
static inline bool id_aa64pfr0_32bit_el0(u64 pfr0)
{
u32 val = cpuid_feature_extract_unsigned_field(pfr0, ID_AA64PFR0_EL0_SHIFT);
diff --git a/arch/arm64/kernel/cpufeature.c b/arch/arm64/kernel/cpufeature.c
index 6892b2440676..7e0dbe2a2f2d 100644
--- a/arch/arm64/kernel/cpufeature.c
+++ b/arch/arm64/kernel/cpufeature.c
@@ -715,6 +715,25 @@ static int check_update_ftr_reg(u32 sys_id, int cpu, u64 val, u64 boot)
return 1;
}
+static void relax_cpu_ftr_reg(u32 sys_id, int field)
+{
+ const struct arm64_ftr_bits *ftrp;
+ struct arm64_ftr_reg *regp = get_arm64_ftr_reg(sys_id);
+
+ if (WARN_ON(!regp))
+ return;
+
+ for (ftrp = regp->ftr_bits; ftrp->width; ftrp++) {
+ if (ftrp->shift == field) {
+ regp->strict_mask &= ~arm64_ftr_mask(ftrp);
+ break;
+ }
+ }
+
+ /* Bogus field? */
+ WARN_ON(!ftrp->width);
+}
+
static int update_32bit_cpu_features(int cpu, struct cpuinfo_arm64 *info,
struct cpuinfo_arm64 *boot)
{
@@ -729,6 +748,19 @@ static int update_32bit_cpu_features(int cpu, struct cpuinfo_arm64 *info,
if (!id_aa64pfr0_32bit_el0(pfr0))
return taint;
+ /*
+ * If we don't have AArch32 at EL1, then relax the strictness of
+ * EL1-dependent register fields to avoid spurious sanity check fails.
+ */
+ if (!id_aa64pfr0_32bit_el1(pfr0)) {
+ relax_cpu_ftr_reg(SYS_ID_ISAR4_EL1, ID_ISAR4_SMC_SHIFT);
+ relax_cpu_ftr_reg(SYS_ID_PFR1_EL1, ID_PFR1_VIRT_FRAC_SHIFT);
+ relax_cpu_ftr_reg(SYS_ID_PFR1_EL1, ID_PFR1_SEC_FRAC_SHIFT);
+ relax_cpu_ftr_reg(SYS_ID_PFR1_EL1, ID_PFR1_VIRTUALIZATION_SHIFT);
+ relax_cpu_ftr_reg(SYS_ID_PFR1_EL1, ID_PFR1_SECURITY_SHIFT);
+ relax_cpu_ftr_reg(SYS_ID_PFR1_EL1, ID_PFR1_PROGMOD_SHIFT);
+ }
+
taint |= check_update_ftr_reg(SYS_ID_DFR0_EL1, cpu,
info->reg_id_dfr0, boot->reg_id_dfr0);
taint |= check_update_ftr_reg(SYS_ID_ISAR0_EL1, cpu,
@@ -773,7 +805,6 @@ static int update_32bit_cpu_features(int cpu, struct cpuinfo_arm64 *info,
return taint;
}
-
/*
* Update system wide CPU feature registers with the values from a
* non-boot CPU. Also performs SANITY checks to make sure that there
--
2.26.1.301.g55bc3eb7cb9-goog
^ permalink raw reply related [flat|nested] 10+ messages in thread* [PATCH v2 7/8] arm64: cpufeature: Relax checks for AArch32 support at EL[0-2]
2020-04-21 14:29 [PATCH v2 0/8] Relax sanity checking for mismatched AArch32 EL1 Will Deacon
` (5 preceding siblings ...)
2020-04-21 14:29 ` [PATCH v2 6/8] arm64: cpufeature: Relax AArch32 system checks if EL1 is 64-bit only Will Deacon
@ 2020-04-21 14:29 ` Will Deacon
2020-04-21 14:29 ` [PATCH v2 8/8] arm64: cpufeature: Add an overview comment for the cpufeature framework Will Deacon
7 siblings, 0 replies; 10+ messages in thread
From: Will Deacon @ 2020-04-21 14:29 UTC (permalink / raw)
To: linux-arm-kernel, kvmarm
Cc: linux-kernel, Will Deacon, Suzuki K Poulose, Mark Rutland,
Marc Zyngier, Anshuman Khandual, Catalin Marinas,
Sai Prakash Ranjan, Doug Anderson, kernel-team
We don't need to be quite as strict about mismatched AArch32 support,
which is good because the friendly hardware folks have been busy
mismatching this to their hearts' content.
* We don't care about EL2 or EL3 (there are silly comments concerning
the latter, so remove those)
* EL1 support is gated by the ARM64_HAS_32BIT_EL1 capability and handled
gracefully when a mismatch occurs
* EL0 support is gated by the ARM64_HAS_32BIT_EL0 capability and handled
gracefully when a mismatch occurs
Relax the AArch32 checks to FTR_NONSTRICT.
Reviewed-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Tested-by: Sai Prakash Ranjan <saiprakash.ranjan@codeaurora.org>
Signed-off-by: Will Deacon <will@kernel.org>
---
arch/arm64/kernel/cpufeature.c | 10 +++-------
1 file changed, 3 insertions(+), 7 deletions(-)
diff --git a/arch/arm64/kernel/cpufeature.c b/arch/arm64/kernel/cpufeature.c
index 7e0dbe2a2f2d..d63653d7c5d0 100644
--- a/arch/arm64/kernel/cpufeature.c
+++ b/arch/arm64/kernel/cpufeature.c
@@ -172,11 +172,10 @@ static const struct arm64_ftr_bits ftr_id_aa64pfr0[] = {
ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_AA64PFR0_GIC_SHIFT, 4, 0),
S_ARM64_FTR_BITS(FTR_VISIBLE, FTR_STRICT, FTR_LOWER_SAFE, ID_AA64PFR0_ASIMD_SHIFT, 4, ID_AA64PFR0_ASIMD_NI),
S_ARM64_FTR_BITS(FTR_VISIBLE, FTR_STRICT, FTR_LOWER_SAFE, ID_AA64PFR0_FP_SHIFT, 4, ID_AA64PFR0_FP_NI),
- /* Linux doesn't care about the EL3 */
ARM64_FTR_BITS(FTR_HIDDEN, FTR_NONSTRICT, FTR_LOWER_SAFE, ID_AA64PFR0_EL3_SHIFT, 4, 0),
- ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_AA64PFR0_EL2_SHIFT, 4, 0),
- ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_AA64PFR0_EL1_SHIFT, 4, ID_AA64PFR0_EL1_64BIT_ONLY),
- ARM64_FTR_BITS(FTR_HIDDEN, FTR_STRICT, FTR_LOWER_SAFE, ID_AA64PFR0_EL0_SHIFT, 4, ID_AA64PFR0_EL0_64BIT_ONLY),
+ ARM64_FTR_BITS(FTR_HIDDEN, FTR_NONSTRICT, FTR_LOWER_SAFE, ID_AA64PFR0_EL2_SHIFT, 4, 0),
+ ARM64_FTR_BITS(FTR_HIDDEN, FTR_NONSTRICT, FTR_LOWER_SAFE, ID_AA64PFR0_EL1_SHIFT, 4, ID_AA64PFR0_EL1_64BIT_ONLY),
+ ARM64_FTR_BITS(FTR_HIDDEN, FTR_NONSTRICT, FTR_LOWER_SAFE, ID_AA64PFR0_EL0_SHIFT, 4, ID_AA64PFR0_EL0_64BIT_ONLY),
ARM64_FTR_END,
};
@@ -867,9 +866,6 @@ void update_cpu_features(int cpu,
taint |= check_update_ftr_reg(SYS_ID_AA64MMFR2_EL1, cpu,
info->reg_id_aa64mmfr2, boot->reg_id_aa64mmfr2);
- /*
- * EL3 is not our concern.
- */
taint |= check_update_ftr_reg(SYS_ID_AA64PFR0_EL1, cpu,
info->reg_id_aa64pfr0, boot->reg_id_aa64pfr0);
taint |= check_update_ftr_reg(SYS_ID_AA64PFR1_EL1, cpu,
--
2.26.1.301.g55bc3eb7cb9-goog
^ permalink raw reply related [flat|nested] 10+ messages in thread* [PATCH v2 8/8] arm64: cpufeature: Add an overview comment for the cpufeature framework
2020-04-21 14:29 [PATCH v2 0/8] Relax sanity checking for mismatched AArch32 EL1 Will Deacon
` (6 preceding siblings ...)
2020-04-21 14:29 ` [PATCH v2 7/8] arm64: cpufeature: Relax checks for AArch32 support at EL[0-2] Will Deacon
@ 2020-04-21 14:29 ` Will Deacon
2020-04-21 15:15 ` Suzuki K Poulose
7 siblings, 1 reply; 10+ messages in thread
From: Will Deacon @ 2020-04-21 14:29 UTC (permalink / raw)
To: linux-arm-kernel, kvmarm
Cc: linux-kernel, Will Deacon, Suzuki K Poulose, Mark Rutland,
Marc Zyngier, Anshuman Khandual, Catalin Marinas,
Sai Prakash Ranjan, Doug Anderson, kernel-team
Now that Suzuki isn't within throwing distance, I thought I'd better add
a rough overview comment to cpufeature.c so that it doesn't take me days
to remember how it works next time.
Signed-off-by: Will Deacon <will@kernel.org>
---
arch/arm64/kernel/cpufeature.c | 50 ++++++++++++++++++++++++++++++++++
1 file changed, 50 insertions(+)
diff --git a/arch/arm64/kernel/cpufeature.c b/arch/arm64/kernel/cpufeature.c
index d63653d7c5d0..c1d44d127baa 100644
--- a/arch/arm64/kernel/cpufeature.c
+++ b/arch/arm64/kernel/cpufeature.c
@@ -3,6 +3,56 @@
* Contains CPU feature definitions
*
* Copyright (C) 2015 ARM Ltd.
+ *
+ * A note for the weary kernel hacker: the code here is confusing and hard to
+ * follow! That's partly because it's solving a nasty problem, but also because
+ * there's a little bit of over-abstraction that tends to obscure what's going
+ * on behind a maze of helper functions and macros.
+ *
+ * The basic problem is that hardware folks have started gluing together CPUs
+ * with distinct architectural features; in some cases even creating SoCs where
+ * user-visible instructions are available only on a subset of the available
+ * cores. We try to address this by snapshotting the feature registers of the
+ * boot CPU and comparing these with the feature registers of each secondary
+ * CPU when bringing them up. If there is a mismatch, then we update the
+ * snapshot state to indicate the lowest-common denominator of the feature,
+ * known as the "safe" value. This snapshot state can be queried to view the
+ * "sanitised" value of a feature register.
+ *
+ * The sanitised register values are used to decide which capabilities we
+ * have in the system. These may be in the form of traditional "hwcaps"
+ * advertised to userspace or internal "cpucaps" which are used to configure
+ * things like alternative patching and static keys. While a feature mismatch
+ * may result in a TAINT_CPU_OUT_OF_SPEC kernel taint, a capability mismatch
+ * may prevent a CPU from being onlined at all.
+ *
+ * Some implementation details worth remembering:
+ *
+ * - Mismatched features are *always* sanitised to a "safe" value, which
+ * usually indicates that the feature is not supported.
+ *
+ * - A mismatched feature marked with FTR_STRICT will cause a "SANITY CHECK"
+ * warning when onlining an offending CPU and the kernel will be tainted
+ * with TAINT_CPU_OUT_OF_SPEC.
+ *
+ * - Features marked as FTR_VISIBLE have their sanitised value visible to
+ * userspace. FTR_VISIBLE features in registers that are only visible
+ * to EL0 by trapping *must* have a corresponding HWCAP so that late
+ * onlining of CPUs cannot lead to features disappearing at runtime.
+ *
+ * - A "feature" is typically a 4-bit register field. A "capability" is the
+ * high-level description derived from the sanitised field value.
+ *
+ * - Read the Arm ARM (DDI 0487F.a) section D13.1.3 ("Principles of the ID
+ * scheme for fields in ID registers") to understand when feature fields
+ * may be signed or unsigned (FTR_SIGNED and FTR_UNSIGNED accordingly).
+ *
+ * - KVM exposes its own view of the feature registers to guest operating
+ * systems regardless of FTR_VISIBLE. This is typically driven from the
+ * sanitised register values to allow virtual CPUs to be migrated between
+ * arbitrary physical CPUs, but some features not present on the host are
+ * also advertised and emulated. Look at sys_reg_descs[] for the gory
+ * details.
*/
#define pr_fmt(fmt) "CPU features: " fmt
--
2.26.1.301.g55bc3eb7cb9-goog
^ permalink raw reply related [flat|nested] 10+ messages in thread* Re: [PATCH v2 8/8] arm64: cpufeature: Add an overview comment for the cpufeature framework
2020-04-21 14:29 ` [PATCH v2 8/8] arm64: cpufeature: Add an overview comment for the cpufeature framework Will Deacon
@ 2020-04-21 15:15 ` Suzuki K Poulose
0 siblings, 0 replies; 10+ messages in thread
From: Suzuki K Poulose @ 2020-04-21 15:15 UTC (permalink / raw)
To: will, linux-arm-kernel, kvmarm
Cc: linux-kernel, mark.rutland, maz, anshuman.khandual,
catalin.marinas, saiprakash.ranjan, dianders, kernel-team
On 04/21/2020 03:29 PM, Will Deacon wrote:
> Now that Suzuki isn't within throwing distance, I thought I'd better add
> a rough overview comment to cpufeature.c so that it doesn't take me days
> to remember how it works next time.
>
> Signed-off-by: Will Deacon <will@kernel.org>
> ---
Reviewed-by: Suzuki K Poulose <suzuki.poulose@arm.com>
^ permalink raw reply [flat|nested] 10+ messages in thread