2019-06-03 08:44:50 +03:00
// SPDX-License-Identifier: GPL-2.0-only
2015-03-27 16:09:23 +03:00
/*
* Contains CPU feature definitions
*
* Copyright ( C ) 2015 ARM Ltd .
2020-04-21 17:29:22 +03:00
*
* 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 .
2020-05-05 13:45:21 +03:00
*
* - If the arm64_ftr_bits [ ] for a register has a missing field , then this
* field is treated as STRICT RES0 , including for read_sanitised_ftr_reg ( ) .
* This is stronger than FTR_HIDDEN and can be used to hide features from
* KVM guests .
2015-03-27 16:09:23 +03:00
*/
2015-10-19 16:24:41 +03:00
# define pr_fmt(fmt) "CPU features: " fmt
2015-03-27 16:09:23 +03:00
2015-10-19 16:24:45 +03:00
# include <linux/bsearch.h>
2016-10-18 13:27:46 +03:00
# include <linux/cpumask.h>
2018-07-31 16:08:56 +03:00
# include <linux/crash_dump.h>
2015-10-19 16:24:45 +03:00
# include <linux/sort.h>
2016-10-18 13:27:46 +03:00
# include <linux/stop_machine.h>
2021-07-30 14:24:40 +03:00
# include <linux/sysfs.h>
2015-03-27 16:09:23 +03:00
# include <linux/types.h>
2021-04-29 23:50:46 +03:00
# include <linux/minmax.h>
2017-01-11 00:35:49 +03:00
# include <linux/mm.h>
2019-04-12 23:39:32 +03:00
# include <linux/cpu.h>
2020-12-22 23:02:10 +03:00
# include <linux/kasan.h>
2021-11-23 21:29:25 +03:00
# include <linux/percpu.h>
2015-03-27 16:09:23 +03:00
# include <asm/cpu.h>
# include <asm/cpufeature.h>
arm64: Delay cpu feature capability checks
At the moment we run through the arm64_features capability list for
each CPU and set the capability if one of the CPU supports it. This
could be problematic in a heterogeneous system with differing capabilities.
Delay the CPU feature checks until all the enabled CPUs are up(i.e,
smp_cpus_done(), so that we can make better decisions based on the
overall system capability. Once we decide and advertise the capabilities
the alternatives can be applied. From this state, we cannot roll back
a feature to disabled based on the values from a new hotplugged CPU,
due to the runtime patching and other reasons. So, for all new CPUs,
we need to make sure that they have the established system capabilities.
Failing which, we bring the CPU down, preventing it from turning online.
Once the capabilities are decided, any new CPU booting up goes through
verification to ensure that it has all the enabled capabilities and also
invokes the respective enable() method on the CPU.
The CPU errata checks are not delayed and is still executed per-CPU
to detect the respective capabilities. If we ever come across a non-errata
capability that needs to be checked on each-CPU, we could introduce them via
a new capability table(or introduce a flag), which can be processed per CPU.
The next patch will make the feature checks use the system wide
safe value of a feature register.
NOTE: The enable() methods associated with the capability is scheduled
on all the CPUs (which is the only use case at the moment). If we need
a different type of 'enable()' which only needs to be run once on any CPU,
we should be able to handle that when needed.
Signed-off-by: Suzuki K. Poulose <suzuki.poulose@arm.com>
Tested-by: Dave Martin <Dave.Martin@arm.com>
[catalin.marinas@arm.com: static variable and coding style fixes]
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2015-10-19 16:24:50 +03:00
# include <asm/cpu_ops.h>
2017-10-31 18:51:10 +03:00
# include <asm/fpsimd.h>
2022-07-14 19:15:23 +03:00
# include <asm/hwcap.h>
2021-06-09 13:23:01 +03:00
# include <asm/insn.h>
2020-12-02 21:40:58 +03:00
# include <asm/kvm_host.h>
2016-02-23 13:31:45 +03:00
# include <asm/mmu_context.h>
2020-05-04 16:42:36 +03:00
# include <asm/mte.h>
2015-07-22 21:05:54 +03:00
# include <asm/processor.h>
2021-07-08 14:15:42 +03:00
# include <asm/smp.h>
2015-10-19 16:24:42 +03:00
# include <asm/sysreg.h>
2017-01-09 20:28:31 +03:00
# include <asm/traps.h>
2021-11-23 21:29:25 +03:00
# include <asm/vectors.h>
2015-01-29 14:24:05 +03:00
# include <asm/virt.h>
2015-03-27 16:09:23 +03:00
2019-04-09 12:52:41 +03:00
/* Kernel representation of AT_HWCAP and AT_HWCAP2 */
2022-07-07 13:36:31 +03:00
static DECLARE_BITMAP ( elf_hwcap , MAX_CPU_FEATURES ) __read_mostly ;
2015-10-19 16:24:41 +03:00
# ifdef CONFIG_COMPAT
# define COMPAT_ELF_HWCAP_DEFAULT \
( COMPAT_HWCAP_HALF | COMPAT_HWCAP_THUMB | \
COMPAT_HWCAP_FAST_MULT | COMPAT_HWCAP_EDSP | \
2020-01-14 02:30:20 +03:00
COMPAT_HWCAP_TLS | COMPAT_HWCAP_IDIV | \
2015-10-19 16:24:41 +03:00
COMPAT_HWCAP_LPAE )
unsigned int compat_elf_hwcap __read_mostly = COMPAT_ELF_HWCAP_DEFAULT ;
unsigned int compat_elf_hwcap2 __read_mostly ;
# endif
DECLARE_BITMAP ( cpu_hwcaps , ARM64_NCAPS ) ;
2016-07-01 18:53:00 +03:00
EXPORT_SYMBOL ( cpu_hwcaps ) ;
arm64: capabilities: Speed up capability lookup
We maintain two separate tables of capabilities, errata and features,
which decide the system capabilities. We iterate over each of these
tables for various operations (e.g, detection, verification etc.).
We do not have a way to map a system "capability" to its entry,
(i.e, cap -> struct arm64_cpu_capabilities) which is needed for
this_cpu_has_cap(). So we iterate over the table one by one to
find the entry and then do the operation. Also, this prevents
us from optimizing the way we "enable" the capabilities on the
CPUs, where we now issue a stop_machine() for each available
capability.
One solution is to merge the two tables into a single table,
sorted by the capability. But this is has the following
disadvantages:
- We loose the "classification" of an errata vs. feature
- It is quite easy to make a mistake when adding an entry,
unless we sort the table at runtime.
So we maintain a list of pointers to the capability entry, sorted
by the "cap number" in a separate array, initialized at boot time.
The only restriction is that we can have one "entry" per capability.
While at it, remove the duplicate declaration of arm64_errata table.
Reviewed-by: Vladimir Murzin <vladimir.murzin@arm.com>
Tested-by: Vladimir Murzin <vladimir.murzin@arm.com>
Signed-off-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
2018-11-30 20:18:03 +03:00
static struct arm64_cpu_capabilities const __ro_after_init * cpu_hwcaps_ptrs [ ARM64_NCAPS ] ;
2015-10-19 16:24:41 +03:00
2022-09-12 19:22:08 +03:00
DECLARE_BITMAP ( boot_capabilities , ARM64_NCAPS ) ;
2019-01-31 17:58:53 +03:00
2019-12-09 21:12:17 +03:00
bool arm64_use_ng_mappings = false ;
EXPORT_SYMBOL ( arm64_use_ng_mappings ) ;
2021-11-23 21:29:25 +03:00
DEFINE_PER_CPU_READ_MOSTLY ( const char * , this_cpu_vector ) = vectors ;
2021-06-08 21:02:55 +03:00
/*
* Permit PER_LINUX32 and execve ( ) of 32 - bit binaries even if not all CPUs
* support it ?
*/
static bool __read_mostly allow_mismatched_32bit_el0 ;
/*
* Static branch enabled only if allow_mismatched_32bit_el0 is set and we have
* seen at least one CPU capable of 32 - bit EL0 .
*/
DEFINE_STATIC_KEY_FALSE ( arm64_mismatched_32bit_el0 ) ;
/*
* Mask of CPUs supporting 32 - bit EL0 .
* Only valid if arm64_mismatched_32bit_el0 is enabled .
*/
static cpumask_var_t cpu_32bit_el0_mask __cpumask_var_read_mostly ;
2020-06-29 07:38:31 +03:00
void dump_cpu_features ( void )
2017-06-21 20:11:23 +03:00
{
/* file-wide pr_fmt adds "CPU features: " prefix */
pr_emerg ( " 0x%*pb \n " , ARM64_NCAPS , & cpu_hwcaps ) ;
}
2017-01-09 20:28:30 +03:00
# define __ARM64_FTR_BITS(SIGNED, VISIBLE, STRICT, TYPE, SHIFT, WIDTH, SAFE_VAL) \
2015-10-19 16:24:45 +03:00
{ \
2015-11-18 20:08:57 +03:00
. sign = SIGNED , \
2017-01-09 20:28:30 +03:00
. visible = VISIBLE , \
2015-10-19 16:24:45 +03:00
. strict = STRICT , \
. type = TYPE , \
. shift = SHIFT , \
. width = WIDTH , \
. safe_val = SAFE_VAL , \
}
arm64: cpufeature: Fix the sign of feature bits
There is a confusion on whether the values of a feature are signed
or not in ARM. This is not clearly mentioned in the ARM ARM either.
We have dealt most of the bits as signed so far, and marked the
rest as unsigned explicitly. This fixed in ARM ARM and will be rolled
out soon.
Here is the criteria in a nutshell:
1) The fields, which are either signed or unsigned, use increasing
numerical values to indicate an increase in functionality. Thus, if a value
of 0x1 indicates the presence of some instructions, then the 0x2 value will
indicate the presence of those instructions plus some additional instructions
or functionality.
2) For ID field values where the value 0x0 defines that a feature is not present,
the number is an unsigned value.
3) For some features where the feature was made optional or removed after the
start of the definition of the architecture, the value 0x0 is used to
indicate the presence of a feature, and 0xF indicates the absence of the
feature. In these cases, the fields are, in effect, holding signed values.
So with these rules applied, we have only the following fields which are signed and
the rest are unsigned.
a) ID_AA64PFR0_EL1: {FP, ASIMD}
b) ID_AA64MMFR0_EL1: {TGran4K, TGran64K}
c) ID_AA64DFR0_EL1: PMUVer (0xf - PMUv3 not implemented)
d) ID_DFR0_EL1: PerfMon
e) ID_MMFR0_EL1: {InnerShr, OuterShr}
Signed-off-by: Suzuki K. Poulose <suzuki.poulose@arm.com>
Acked-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2016-01-26 13:58:14 +03:00
/* Define a feature with unsigned values */
2017-01-09 20:28:30 +03:00
# define ARM64_FTR_BITS(VISIBLE, STRICT, TYPE, SHIFT, WIDTH, SAFE_VAL) \
__ARM64_FTR_BITS ( FTR_UNSIGNED , VISIBLE , STRICT , TYPE , SHIFT , WIDTH , SAFE_VAL )
2015-11-18 20:08:57 +03:00
arm64: cpufeature: Fix the sign of feature bits
There is a confusion on whether the values of a feature are signed
or not in ARM. This is not clearly mentioned in the ARM ARM either.
We have dealt most of the bits as signed so far, and marked the
rest as unsigned explicitly. This fixed in ARM ARM and will be rolled
out soon.
Here is the criteria in a nutshell:
1) The fields, which are either signed or unsigned, use increasing
numerical values to indicate an increase in functionality. Thus, if a value
of 0x1 indicates the presence of some instructions, then the 0x2 value will
indicate the presence of those instructions plus some additional instructions
or functionality.
2) For ID field values where the value 0x0 defines that a feature is not present,
the number is an unsigned value.
3) For some features where the feature was made optional or removed after the
start of the definition of the architecture, the value 0x0 is used to
indicate the presence of a feature, and 0xF indicates the absence of the
feature. In these cases, the fields are, in effect, holding signed values.
So with these rules applied, we have only the following fields which are signed and
the rest are unsigned.
a) ID_AA64PFR0_EL1: {FP, ASIMD}
b) ID_AA64MMFR0_EL1: {TGran4K, TGran64K}
c) ID_AA64DFR0_EL1: PMUVer (0xf - PMUv3 not implemented)
d) ID_DFR0_EL1: PerfMon
e) ID_MMFR0_EL1: {InnerShr, OuterShr}
Signed-off-by: Suzuki K. Poulose <suzuki.poulose@arm.com>
Acked-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2016-01-26 13:58:14 +03:00
/* Define a feature with a signed value */
2017-01-09 20:28:30 +03:00
# define S_ARM64_FTR_BITS(VISIBLE, STRICT, TYPE, SHIFT, WIDTH, SAFE_VAL) \
__ARM64_FTR_BITS ( FTR_SIGNED , VISIBLE , STRICT , TYPE , SHIFT , WIDTH , SAFE_VAL )
arm64: cpufeature: Fix the sign of feature bits
There is a confusion on whether the values of a feature are signed
or not in ARM. This is not clearly mentioned in the ARM ARM either.
We have dealt most of the bits as signed so far, and marked the
rest as unsigned explicitly. This fixed in ARM ARM and will be rolled
out soon.
Here is the criteria in a nutshell:
1) The fields, which are either signed or unsigned, use increasing
numerical values to indicate an increase in functionality. Thus, if a value
of 0x1 indicates the presence of some instructions, then the 0x2 value will
indicate the presence of those instructions plus some additional instructions
or functionality.
2) For ID field values where the value 0x0 defines that a feature is not present,
the number is an unsigned value.
3) For some features where the feature was made optional or removed after the
start of the definition of the architecture, the value 0x0 is used to
indicate the presence of a feature, and 0xF indicates the absence of the
feature. In these cases, the fields are, in effect, holding signed values.
So with these rules applied, we have only the following fields which are signed and
the rest are unsigned.
a) ID_AA64PFR0_EL1: {FP, ASIMD}
b) ID_AA64MMFR0_EL1: {TGran4K, TGran64K}
c) ID_AA64DFR0_EL1: PMUVer (0xf - PMUv3 not implemented)
d) ID_DFR0_EL1: PerfMon
e) ID_MMFR0_EL1: {InnerShr, OuterShr}
Signed-off-by: Suzuki K. Poulose <suzuki.poulose@arm.com>
Acked-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2016-01-26 13:58:14 +03:00
2015-10-19 16:24:45 +03:00
# define ARM64_FTR_END \
{ \
. width = 0 , \
}
2018-07-31 16:08:56 +03:00
static void cpu_enable_cnp ( struct arm64_cpu_capabilities const * cap ) ;
2016-02-05 17:58:50 +03:00
2020-03-13 12:04:48 +03:00
static bool __system_matches_cap ( unsigned int n ) ;
2017-01-09 20:28:32 +03:00
/*
* NOTE : Any changes to the visibility of features should be kept in
* sync with the documentation of the CPU feature register ABI .
*/
2016-08-31 13:31:08 +03:00
static const struct arm64_ftr_bits ftr_id_aa64isar0 [ ] = {
2022-05-03 20:02:28 +03:00
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64ISAR0_EL1_RNDR_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64ISAR0_EL1_TLB_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64ISAR0_EL1_TS_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64ISAR0_EL1_FHM_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64ISAR0_EL1_DP_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64ISAR0_EL1_SM4_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64ISAR0_EL1_SM3_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64ISAR0_EL1_SHA3_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64ISAR0_EL1_RDM_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64ISAR0_EL1_ATOMIC_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64ISAR0_EL1_CRC32_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64ISAR0_EL1_SHA2_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64ISAR0_EL1_SHA1_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64ISAR0_EL1_AES_SHIFT , 4 , 0 ) ,
2015-10-19 16:24:45 +03:00
ARM64_FTR_END ,
} ;
2017-03-14 21:13:25 +03:00
static const struct arm64_ftr_bits ftr_id_aa64isar1 [ ] = {
2022-07-04 20:02:49 +03:00
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64ISAR1_EL1_I8MM_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64ISAR1_EL1_DGH_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64ISAR1_EL1_BF16_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64ISAR1_EL1_SPECRES_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64ISAR1_EL1_SB_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64ISAR1_EL1_FRINTTS_SHIFT , 4 , 0 ) ,
2018-12-07 21:39:24 +03:00
ARM64_FTR_BITS ( FTR_VISIBLE_IF_IS_ENABLED ( CONFIG_ARM64_PTR_AUTH ) ,
2022-07-04 20:02:49 +03:00
FTR_STRICT , FTR_LOWER_SAFE , ID_AA64ISAR1_EL1_GPI_SHIFT , 4 , 0 ) ,
2018-12-07 21:39:24 +03:00
ARM64_FTR_BITS ( FTR_VISIBLE_IF_IS_ENABLED ( CONFIG_ARM64_PTR_AUTH ) ,
2022-07-04 20:02:49 +03:00
FTR_STRICT , FTR_LOWER_SAFE , ID_AA64ISAR1_EL1_GPA_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64ISAR1_EL1_LRCPC_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64ISAR1_EL1_FCMA_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64ISAR1_EL1_JSCVT_SHIFT , 4 , 0 ) ,
2018-12-07 21:39:24 +03:00
ARM64_FTR_BITS ( FTR_VISIBLE_IF_IS_ENABLED ( CONFIG_ARM64_PTR_AUTH ) ,
2022-07-04 20:02:49 +03:00
FTR_STRICT , FTR_EXACT , ID_AA64ISAR1_EL1_API_SHIFT , 4 , 0 ) ,
2018-12-07 21:39:24 +03:00
ARM64_FTR_BITS ( FTR_VISIBLE_IF_IS_ENABLED ( CONFIG_ARM64_PTR_AUTH ) ,
2022-07-04 20:02:49 +03:00
FTR_STRICT , FTR_EXACT , ID_AA64ISAR1_EL1_APA_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64ISAR1_EL1_DPB_SHIFT , 4 , 0 ) ,
2017-03-14 21:13:25 +03:00
ARM64_FTR_END ,
} ;
2021-12-10 19:54:31 +03:00
static const struct arm64_ftr_bits ftr_id_aa64isar2 [ ] = {
2022-07-04 20:02:50 +03:00
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_HIGHER_SAFE , ID_AA64ISAR2_EL1_BC_SHIFT , 4 , 0 ) ,
2022-02-24 15:49:52 +03:00
ARM64_FTR_BITS ( FTR_VISIBLE_IF_IS_ENABLED ( CONFIG_ARM64_PTR_AUTH ) ,
2022-07-04 20:02:50 +03:00
FTR_STRICT , FTR_EXACT , ID_AA64ISAR2_EL1_APA3_SHIFT , 4 , 0 ) ,
2022-02-24 15:49:52 +03:00
ARM64_FTR_BITS ( FTR_VISIBLE_IF_IS_ENABLED ( CONFIG_ARM64_PTR_AUTH ) ,
2022-07-04 20:02:50 +03:00
FTR_STRICT , FTR_LOWER_SAFE , ID_AA64ISAR2_EL1_GPA3_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_NONSTRICT , FTR_LOWER_SAFE , ID_AA64ISAR2_EL1_RPRES_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_NONSTRICT , FTR_LOWER_SAFE , ID_AA64ISAR2_EL1_WFxT_SHIFT , 4 , 0 ) ,
2021-12-10 19:54:31 +03:00
ARM64_FTR_END ,
} ;
2016-08-31 13:31:08 +03:00
static const struct arm64_ftr_bits ftr_id_aa64pfr0 [ ] = {
2022-09-06 01:54:03 +03:00
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_NONSTRICT , FTR_LOWER_SAFE , ID_AA64PFR0_EL1_CSV3_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_NONSTRICT , FTR_LOWER_SAFE , ID_AA64PFR0_EL1_CSV2_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64PFR0_EL1_DIT_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_NONSTRICT , FTR_LOWER_SAFE , ID_AA64PFR0_EL1_AMU_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64PFR0_EL1_MPAM_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_NONSTRICT , FTR_LOWER_SAFE , ID_AA64PFR0_EL1_SEL2_SHIFT , 4 , 0 ) ,
2017-12-14 17:03:44 +03:00
ARM64_FTR_BITS ( FTR_VISIBLE_IF_IS_ENABLED ( CONFIG_ARM64_SVE ) ,
2022-09-06 01:54:03 +03:00
FTR_STRICT , FTR_LOWER_SAFE , ID_AA64PFR0_EL1_SVE_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64PFR0_EL1_RAS_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64PFR0_EL1_GIC_SHIFT , 4 , 0 ) ,
2022-09-06 01:54:11 +03:00
S_ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64PFR0_EL1_AdvSIMD_SHIFT , 4 , ID_AA64PFR0_EL1_AdvSIMD_NI ) ,
2022-09-06 01:54:03 +03:00
S_ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64PFR0_EL1_FP_SHIFT , 4 , ID_AA64PFR0_EL1_FP_NI ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_NONSTRICT , FTR_LOWER_SAFE , ID_AA64PFR0_EL1_EL3_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_NONSTRICT , FTR_LOWER_SAFE , ID_AA64PFR0_EL1_EL2_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_NONSTRICT , FTR_LOWER_SAFE , ID_AA64PFR0_EL1_EL1_SHIFT , 4 , ID_AA64PFR0_EL1_ELx_64BIT_ONLY ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_NONSTRICT , FTR_LOWER_SAFE , ID_AA64PFR0_EL1_EL0_SHIFT , 4 , ID_AA64PFR0_EL1_ELx_64BIT_ONLY ) ,
2015-10-19 16:24:45 +03:00
ARM64_FTR_END ,
} ;
2018-06-15 13:37:34 +03:00
static const struct arm64_ftr_bits ftr_id_aa64pfr1 [ ] = {
2022-04-19 14:22:16 +03:00
ARM64_FTR_BITS ( FTR_VISIBLE_IF_IS_ENABLED ( CONFIG_ARM64_SME ) ,
2022-09-06 01:54:04 +03:00
FTR_STRICT , FTR_LOWER_SAFE , ID_AA64PFR1_EL1_SME_SHIFT , 4 , 0 ) ,
2022-09-06 01:54:14 +03:00
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64PFR1_EL1_MPAM_frac_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64PFR1_EL1_RAS_frac_SHIFT , 4 , 0 ) ,
2019-09-06 12:58:01 +03:00
ARM64_FTR_BITS ( FTR_VISIBLE_IF_IS_ENABLED ( CONFIG_ARM64_MTE ) ,
2022-09-06 01:54:04 +03:00
FTR_STRICT , FTR_LOWER_SAFE , ID_AA64PFR1_EL1_MTE_SHIFT , 4 , ID_AA64PFR1_EL1_MTE_NI ) ,
2022-09-06 01:54:12 +03:00
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_NONSTRICT , FTR_LOWER_SAFE , ID_AA64PFR1_EL1_SSBS_SHIFT , 4 , ID_AA64PFR1_EL1_SSBS_NI ) ,
2020-03-16 19:50:45 +03:00
ARM64_FTR_BITS ( FTR_VISIBLE_IF_IS_ENABLED ( CONFIG_ARM64_BTI ) ,
2022-09-06 01:54:04 +03:00
FTR_STRICT , FTR_LOWER_SAFE , ID_AA64PFR1_EL1_BT_SHIFT , 4 , 0 ) ,
2018-06-15 13:37:34 +03:00
ARM64_FTR_END ,
} ;
2019-04-18 20:41:38 +03:00
static const struct arm64_ftr_bits ftr_id_aa64zfr0 [ ] = {
2019-12-16 14:33:37 +03:00
ARM64_FTR_BITS ( FTR_VISIBLE_IF_IS_ENABLED ( CONFIG_ARM64_SVE ) ,
2022-07-04 20:02:47 +03:00
FTR_STRICT , FTR_LOWER_SAFE , ID_AA64ZFR0_EL1_F64MM_SHIFT , 4 , 0 ) ,
2019-12-16 14:33:37 +03:00
ARM64_FTR_BITS ( FTR_VISIBLE_IF_IS_ENABLED ( CONFIG_ARM64_SVE ) ,
2022-07-04 20:02:47 +03:00
FTR_STRICT , FTR_LOWER_SAFE , ID_AA64ZFR0_EL1_F32MM_SHIFT , 4 , 0 ) ,
2019-12-16 14:33:37 +03:00
ARM64_FTR_BITS ( FTR_VISIBLE_IF_IS_ENABLED ( CONFIG_ARM64_SVE ) ,
2022-07-04 20:02:47 +03:00
FTR_STRICT , FTR_LOWER_SAFE , ID_AA64ZFR0_EL1_I8MM_SHIFT , 4 , 0 ) ,
2019-10-14 13:21:13 +03:00
ARM64_FTR_BITS ( FTR_VISIBLE_IF_IS_ENABLED ( CONFIG_ARM64_SVE ) ,
2022-07-04 20:02:47 +03:00
FTR_STRICT , FTR_LOWER_SAFE , ID_AA64ZFR0_EL1_SM4_SHIFT , 4 , 0 ) ,
2019-10-14 13:21:13 +03:00
ARM64_FTR_BITS ( FTR_VISIBLE_IF_IS_ENABLED ( CONFIG_ARM64_SVE ) ,
2022-07-04 20:02:47 +03:00
FTR_STRICT , FTR_LOWER_SAFE , ID_AA64ZFR0_EL1_SHA3_SHIFT , 4 , 0 ) ,
2019-12-16 14:33:37 +03:00
ARM64_FTR_BITS ( FTR_VISIBLE_IF_IS_ENABLED ( CONFIG_ARM64_SVE ) ,
2022-07-04 20:02:47 +03:00
FTR_STRICT , FTR_LOWER_SAFE , ID_AA64ZFR0_EL1_BF16_SHIFT , 4 , 0 ) ,
2019-10-14 13:21:13 +03:00
ARM64_FTR_BITS ( FTR_VISIBLE_IF_IS_ENABLED ( CONFIG_ARM64_SVE ) ,
2022-07-04 20:02:47 +03:00
FTR_STRICT , FTR_LOWER_SAFE , ID_AA64ZFR0_EL1_BitPerm_SHIFT , 4 , 0 ) ,
2019-10-14 13:21:13 +03:00
ARM64_FTR_BITS ( FTR_VISIBLE_IF_IS_ENABLED ( CONFIG_ARM64_SVE ) ,
2022-07-04 20:02:47 +03:00
FTR_STRICT , FTR_LOWER_SAFE , ID_AA64ZFR0_EL1_AES_SHIFT , 4 , 0 ) ,
2019-10-14 13:21:13 +03:00
ARM64_FTR_BITS ( FTR_VISIBLE_IF_IS_ENABLED ( CONFIG_ARM64_SVE ) ,
2022-07-04 20:02:47 +03:00
FTR_STRICT , FTR_LOWER_SAFE , ID_AA64ZFR0_EL1_SVEver_SHIFT , 4 , 0 ) ,
2019-04-18 20:41:38 +03:00
ARM64_FTR_END ,
} ;
2022-04-19 14:22:16 +03:00
static const struct arm64_ftr_bits ftr_id_aa64smfr0 [ ] = {
ARM64_FTR_BITS ( FTR_VISIBLE_IF_IS_ENABLED ( CONFIG_ARM64_SME ) ,
2022-07-04 20:02:46 +03:00
FTR_STRICT , FTR_EXACT , ID_AA64SMFR0_EL1_FA64_SHIFT , 1 , 0 ) ,
2022-04-19 14:22:16 +03:00
ARM64_FTR_BITS ( FTR_VISIBLE_IF_IS_ENABLED ( CONFIG_ARM64_SME ) ,
2022-07-04 20:02:46 +03:00
FTR_STRICT , FTR_EXACT , ID_AA64SMFR0_EL1_I16I64_SHIFT , 4 , 0 ) ,
2022-04-19 14:22:16 +03:00
ARM64_FTR_BITS ( FTR_VISIBLE_IF_IS_ENABLED ( CONFIG_ARM64_SME ) ,
2022-07-04 20:02:46 +03:00
FTR_STRICT , FTR_EXACT , ID_AA64SMFR0_EL1_F64F64_SHIFT , 1 , 0 ) ,
2022-04-19 14:22:16 +03:00
ARM64_FTR_BITS ( FTR_VISIBLE_IF_IS_ENABLED ( CONFIG_ARM64_SME ) ,
2022-07-04 20:02:46 +03:00
FTR_STRICT , FTR_EXACT , ID_AA64SMFR0_EL1_I8I32_SHIFT , 4 , 0 ) ,
2022-04-19 14:22:16 +03:00
ARM64_FTR_BITS ( FTR_VISIBLE_IF_IS_ENABLED ( CONFIG_ARM64_SME ) ,
2022-07-04 20:02:46 +03:00
FTR_STRICT , FTR_EXACT , ID_AA64SMFR0_EL1_F16F32_SHIFT , 1 , 0 ) ,
2022-04-19 14:22:16 +03:00
ARM64_FTR_BITS ( FTR_VISIBLE_IF_IS_ENABLED ( CONFIG_ARM64_SME ) ,
2022-07-04 20:02:46 +03:00
FTR_STRICT , FTR_EXACT , ID_AA64SMFR0_EL1_B16F32_SHIFT , 1 , 0 ) ,
2022-04-19 14:22:16 +03:00
ARM64_FTR_BITS ( FTR_VISIBLE_IF_IS_ENABLED ( CONFIG_ARM64_SME ) ,
2022-07-04 20:02:46 +03:00
FTR_STRICT , FTR_EXACT , ID_AA64SMFR0_EL1_F32F32_SHIFT , 1 , 0 ) ,
2022-04-19 14:22:16 +03:00
ARM64_FTR_END ,
} ;
2016-08-31 13:31:08 +03:00
static const struct arm64_ftr_bits ftr_id_aa64mmfr0 [ ] = {
2022-09-06 01:54:01 +03:00
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64MMFR0_EL1_ECV_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64MMFR0_EL1_FGT_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64MMFR0_EL1_EXS_SHIFT , 4 , 0 ) ,
2020-05-28 16:12:58 +03:00
/*
* Page size not being supported at Stage - 2 is not fatal . You
* just give up KVM if PAGE_SIZE isn ' t supported there . Go fix
* your favourite nesting hypervisor .
*
* There is a small corner case where the hypervisor explicitly
* advertises a given granule size at Stage - 2 ( value 2 ) on some
* vCPUs , and uses the fallback to Stage - 1 ( value 0 ) for other
* vCPUs . Although this is not forbidden by the architecture , it
* indicates that the hypervisor is being silly ( or buggy ) .
*
* We make no effort to cope with this and pretend that if these
* fields are inconsistent across vCPUs , then it isn ' t worth
* trying to bring KVM up .
*/
2022-09-06 01:54:01 +03:00
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_NONSTRICT , FTR_EXACT , ID_AA64MMFR0_EL1_TGRAN4_2_SHIFT , 4 , 1 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_NONSTRICT , FTR_EXACT , ID_AA64MMFR0_EL1_TGRAN64_2_SHIFT , 4 , 1 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_NONSTRICT , FTR_EXACT , ID_AA64MMFR0_EL1_TGRAN16_2_SHIFT , 4 , 1 ) ,
2019-08-12 18:02:25 +03:00
/*
* We already refuse to boot CPUs that don ' t support our configured
* page size , so we can only detect mismatches for a page size other
* than the one we ' re currently using . Unfortunately , SoCs like this
* exist in the wild so , even though we don ' t like it , we ' ll have to go
* along with it and treat them as non - strict .
*/
2022-09-06 01:54:01 +03:00
S_ARM64_FTR_BITS ( FTR_HIDDEN , FTR_NONSTRICT , FTR_LOWER_SAFE , ID_AA64MMFR0_EL1_TGRAN4_SHIFT , 4 , ID_AA64MMFR0_EL1_TGRAN4_NI ) ,
S_ARM64_FTR_BITS ( FTR_HIDDEN , FTR_NONSTRICT , FTR_LOWER_SAFE , ID_AA64MMFR0_EL1_TGRAN64_SHIFT , 4 , ID_AA64MMFR0_EL1_TGRAN64_NI ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_NONSTRICT , FTR_LOWER_SAFE , ID_AA64MMFR0_EL1_TGRAN16_SHIFT , 4 , ID_AA64MMFR0_EL1_TGRAN16_NI ) ,
2019-08-12 18:02:25 +03:00
2022-09-06 01:54:01 +03:00
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64MMFR0_EL1_BIGENDEL0_SHIFT , 4 , 0 ) ,
2015-10-19 16:24:45 +03:00
/* Linux shouldn't care about secure memory */
2022-09-06 01:54:01 +03:00
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_NONSTRICT , FTR_LOWER_SAFE , ID_AA64MMFR0_EL1_SNSMEM_SHIFT , 4 , 0 ) ,
2022-09-06 01:54:05 +03:00
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64MMFR0_EL1_BIGEND_SHIFT , 4 , 0 ) ,
2022-09-06 01:54:06 +03:00
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64MMFR0_EL1_ASIDBITS_SHIFT , 4 , 0 ) ,
2015-10-19 16:24:45 +03:00
/*
* Differing PARange is fine as long as all peripherals and memory are mapped
* within the minimum PARange of all CPUs
*/
2022-09-06 01:54:01 +03:00
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_NONSTRICT , FTR_LOWER_SAFE , ID_AA64MMFR0_EL1_PARANGE_SHIFT , 4 , 0 ) ,
2015-10-19 16:24:45 +03:00
ARM64_FTR_END ,
} ;
2016-08-31 13:31:08 +03:00
static const struct arm64_ftr_bits ftr_id_aa64mmfr1 [ ] = {
2022-09-06 01:54:07 +03:00
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_NONSTRICT , FTR_LOWER_SAFE , ID_AA64MMFR1_EL1_TIDCP1_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64MMFR1_EL1_AFP_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64MMFR1_EL1_ETS_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64MMFR1_EL1_TWED_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64MMFR1_EL1_XNX_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_HIGHER_SAFE , ID_AA64MMFR1_EL1_SpecSEI_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64MMFR1_EL1_PAN_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64MMFR1_EL1_LO_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64MMFR1_EL1_HPDS_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64MMFR1_EL1_VH_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64MMFR1_EL1_VMIDBits_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64MMFR1_EL1_HAFDBS_SHIFT , 4 , 0 ) ,
2015-10-19 16:24:45 +03:00
ARM64_FTR_END ,
} ;
2016-08-31 13:31:08 +03:00
static const struct arm64_ftr_bits ftr_id_aa64mmfr2 [ ] = {
2022-09-06 01:54:02 +03:00
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_NONSTRICT , FTR_LOWER_SAFE , ID_AA64MMFR2_EL1_E0PD_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64MMFR2_EL1_EVT_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64MMFR2_EL1_BBM_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64MMFR2_EL1_TTL_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64MMFR2_EL1_FWB_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64MMFR2_EL1_IDS_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64MMFR2_EL1_AT_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64MMFR2_EL1_ST_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64MMFR2_EL1_NV_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64MMFR2_EL1_CCIDX_SHIFT , 4 , 0 ) ,
2022-09-06 01:54:08 +03:00
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64MMFR2_EL1_VARange_SHIFT , 4 , 0 ) ,
2022-09-06 01:54:02 +03:00
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_NONSTRICT , FTR_LOWER_SAFE , ID_AA64MMFR2_EL1_IESB_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64MMFR2_EL1_LSM_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64MMFR2_EL1_UAO_SHIFT , 4 , 0 ) ,
2022-09-06 01:54:09 +03:00
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64MMFR2_EL1_CnP_SHIFT , 4 , 0 ) ,
2016-02-05 17:58:47 +03:00
ARM64_FTR_END ,
} ;
2016-08-31 13:31:08 +03:00
static const struct arm64_ftr_bits ftr_ctr [ ] = {
2018-03-07 18:00:08 +03:00
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_EXACT , 31 , 1 , 1 ) , /* RES1 */
2022-07-04 20:02:40 +03:00
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , CTR_EL0_DIC_SHIFT , 1 , 1 ) ,
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , CTR_EL0_IDC_SHIFT , 1 , 1 ) ,
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_HIGHER_OR_ZERO_SAFE , CTR_EL0_CWG_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_HIGHER_OR_ZERO_SAFE , CTR_EL0_ERG_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , CTR_EL0_DminLine_SHIFT , 4 , 1 ) ,
2015-10-19 16:24:45 +03:00
/*
* Linux can handle differing I - cache policies . Userspace JITs will
2016-09-09 16:07:08 +03:00
* make use of * minLine .
2017-03-10 23:32:22 +03:00
* If we have differing I - cache policies , report it as the weakest - VIPT .
2015-10-19 16:24:45 +03:00
*/
2022-07-04 20:02:40 +03:00
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_NONSTRICT , FTR_EXACT , CTR_EL0_L1Ip_SHIFT , 2 , CTR_EL0_L1Ip_VIPT ) , /* L1Ip */
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , CTR_EL0_IminLine_SHIFT , 4 , 0 ) ,
2015-10-19 16:24:45 +03:00
ARM64_FTR_END ,
} ;
2021-02-08 12:57:19 +03:00
static struct arm64_ftr_override __ro_after_init no_override = { } ;
2016-08-31 13:31:10 +03:00
struct arm64_ftr_reg arm64_ftr_reg_ctrel0 = {
. name = " SYS_CTR_EL0 " ,
2021-02-08 12:57:19 +03:00
. ftr_bits = ftr_ctr ,
. override = & no_override ,
2016-08-31 13:31:10 +03:00
} ;
2016-08-31 13:31:08 +03:00
static const struct arm64_ftr_bits ftr_id_mmfr0 [ ] = {
2020-07-03 06:51:37 +03:00
S_ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_MMFR0_INNERSHR_SHIFT , 4 , 0xf ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_MMFR0_FCSE_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_NONSTRICT , FTR_LOWER_SAFE , ID_MMFR0_AUXREG_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_MMFR0_TCM_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_MMFR0_SHARELVL_SHIFT , 4 , 0 ) ,
S_ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_MMFR0_OUTERSHR_SHIFT , 4 , 0xf ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_MMFR0_PMSA_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_MMFR0_VMSA_SHIFT , 4 , 0 ) ,
2015-10-19 16:24:45 +03:00
ARM64_FTR_END ,
} ;
2016-08-31 13:31:08 +03:00
static const struct arm64_ftr_bits ftr_id_aa64dfr0 [ ] = {
2022-09-10 19:33:50 +03:00
S_ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64DFR0_EL1_DoubleLock_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_NONSTRICT , FTR_LOWER_SAFE , ID_AA64DFR0_EL1_PMSVer_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64DFR0_EL1_CTX_CMPs_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64DFR0_EL1_WRPs_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_AA64DFR0_EL1_BRPs_SHIFT , 4 , 0 ) ,
2016-07-25 18:17:52 +03:00
/*
* We can instantiate multiple PMU instances with different levels
* of support .
2017-01-09 20:28:30 +03:00
*/
2022-09-10 19:33:50 +03:00
S_ARM64_FTR_BITS ( FTR_HIDDEN , FTR_NONSTRICT , FTR_EXACT , ID_AA64DFR0_EL1_PMUVer_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_EXACT , ID_AA64DFR0_EL1_DebugVer_SHIFT , 4 , 0x6 ) ,
2015-10-19 16:24:45 +03:00
ARM64_FTR_END ,
} ;
2022-11-03 11:22:32 +03:00
static const struct arm64_ftr_bits ftr_mvfr0 [ ] = {
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , MVFR0_FPROUND_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , MVFR0_FPSHVEC_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , MVFR0_FPSQRT_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , MVFR0_FPDIVIDE_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , MVFR0_FPTRAP_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , MVFR0_FPDP_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , MVFR0_FPSP_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , MVFR0_SIMD_SHIFT , 4 , 0 ) ,
ARM64_FTR_END ,
} ;
static const struct arm64_ftr_bits ftr_mvfr1 [ ] = {
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , MVFR1_SIMDFMAC_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , MVFR1_FPHP_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , MVFR1_SIMDHP_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , MVFR1_SIMDSP_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , MVFR1_SIMDINT_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , MVFR1_SIMDLS_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , MVFR1_FPDNAN_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , MVFR1_FPFTZ_SHIFT , 4 , 0 ) ,
ARM64_FTR_END ,
} ;
2016-08-31 13:31:08 +03:00
static const struct arm64_ftr_bits ftr_mvfr2 [ ] = {
2020-07-03 06:51:37 +03:00
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , MVFR2_FPMISC_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , MVFR2_SIMDMISC_SHIFT , 4 , 0 ) ,
2015-10-19 16:24:45 +03:00
ARM64_FTR_END ,
} ;
2016-08-31 13:31:08 +03:00
static const struct arm64_ftr_bits ftr_dczid [ ] = {
2022-07-04 20:02:41 +03:00
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_EXACT , DCZID_EL0_DZP_SHIFT , 1 , 1 ) ,
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , DCZID_EL0_BS_SHIFT , 4 , 0 ) ,
2015-10-19 16:24:45 +03:00
ARM64_FTR_END ,
} ;
2021-05-26 22:36:21 +03:00
static const struct arm64_ftr_bits ftr_gmid [ ] = {
2022-07-04 20:02:42 +03:00
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , GMID_EL1_BS_SHIFT , 4 , 0 ) ,
2021-05-26 22:36:21 +03:00
ARM64_FTR_END ,
} ;
2020-05-19 12:40:38 +03:00
static const struct arm64_ftr_bits ftr_id_isar0 [ ] = {
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_ISAR0_DIVIDE_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_ISAR0_DEBUG_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_ISAR0_COPROC_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_ISAR0_CMPBRANCH_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_ISAR0_BITFIELD_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_ISAR0_BITCOUNT_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_ISAR0_SWAP_SHIFT , 4 , 0 ) ,
ARM64_FTR_END ,
} ;
2015-10-19 16:24:45 +03:00
2016-08-31 13:31:08 +03:00
static const struct arm64_ftr_bits ftr_id_isar5 [ ] = {
2017-10-19 18:39:02 +03:00
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_ISAR5_RDM_SHIFT , 4 , 0 ) ,
2022-11-03 11:22:32 +03:00
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , ID_ISAR5_CRC32_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , ID_ISAR5_SHA2_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , ID_ISAR5_SHA1_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_VISIBLE , FTR_STRICT , FTR_LOWER_SAFE , ID_ISAR5_AES_SHIFT , 4 , 0 ) ,
2017-10-19 18:39:02 +03:00
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_ISAR5_SEVL_SHIFT , 4 , 0 ) ,
2015-10-19 16:24:45 +03:00
ARM64_FTR_END ,
} ;
2016-08-31 13:31:08 +03:00
static const struct arm64_ftr_bits ftr_id_mmfr4 [ ] = {
2020-05-19 12:40:45 +03:00
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_MMFR4_EVT_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_MMFR4_CCIDX_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_MMFR4_LSM_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_MMFR4_HPDS_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_MMFR4_CNP_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_MMFR4_XNX_SHIFT , 4 , 0 ) ,
2020-07-03 06:51:37 +03:00
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_MMFR4_AC2_SHIFT , 4 , 0 ) ,
2020-05-19 12:40:45 +03:00
/*
* SpecSEI = 1 indicates that the PE might generate an SError on an
* external abort on speculative read . It is safe to assume that an
* SError might be generated than it will not be . Hence it has been
* classified as FTR_HIGHER_SAFE .
*/
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_HIGHER_SAFE , ID_MMFR4_SPECSEI_SHIFT , 4 , 0 ) ,
2015-10-19 16:24:45 +03:00
ARM64_FTR_END ,
} ;
2020-04-21 17:29:16 +03:00
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 ,
} ;
2020-05-19 12:40:43 +03:00
static const struct arm64_ftr_bits ftr_id_mmfr5 [ ] = {
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_MMFR5_ETS_SHIFT , 4 , 0 ) ,
ARM64_FTR_END ,
} ;
2019-12-17 17:47:32 +03:00
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 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_ISAR6_SPECRES_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_ISAR6_SB_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_ISAR6_FHM_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_ISAR6_DP_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_ISAR6_JSCVT_SHIFT , 4 , 0 ) ,
ARM64_FTR_END ,
} ;
2016-08-31 13:31:08 +03:00
static const struct arm64_ftr_bits ftr_id_pfr0 [ ] = {
2020-05-19 12:40:44 +03:00
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_PFR0_DIT_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_NONSTRICT , FTR_LOWER_SAFE , ID_PFR0_CSV2_SHIFT , 4 , 0 ) ,
2020-07-03 06:51:37 +03:00
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_PFR0_STATE3_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_PFR0_STATE2_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_PFR0_STATE1_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_PFR0_STATE0_SHIFT , 4 , 0 ) ,
2015-10-19 16:24:45 +03:00
ARM64_FTR_END ,
} ;
2020-04-21 17:29:16 +03:00
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 ,
} ;
2020-05-19 12:40:41 +03:00
static const struct arm64_ftr_bits ftr_id_pfr2 [ ] = {
2020-09-16 01:56:12 +03:00
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_NONSTRICT , FTR_LOWER_SAFE , ID_PFR2_SSBS_SHIFT , 4 , 0 ) ,
2020-05-19 12:40:41 +03:00
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_NONSTRICT , FTR_LOWER_SAFE , ID_PFR2_CSV3_SHIFT , 4 , 0 ) ,
ARM64_FTR_END ,
} ;
2016-08-31 13:31:08 +03:00
static const struct arm64_ftr_bits ftr_id_dfr0 [ ] = {
2020-05-19 12:40:39 +03:00
/* [31:28] TraceFilt */
arm64: cpufeature: Allow different PMU versions in ID_DFR0_EL1
Commit b20d1ba3cf4b ("arm64: cpufeature: allow for version discrepancy in
PMU implementations") made it possible to run Linux on a machine with PMUs
with different versions without tainting the kernel. The patch relaxed the
restriction only for the ID_AA64DFR0_EL1.PMUVer field, and missed doing the
same for ID_DFR0_EL1.PerfMon , which also reports the PMU version, but for
the AArch32 state.
For example, with Linux running on two clusters with different PMU
versions, the kernel is tainted when bringing up secondaries with the
following message:
[ 0.097027] smp: Bringing up secondary CPUs ...
[..]
[ 0.142805] Detected PIPT I-cache on CPU4
[ 0.142805] CPU features: SANITY CHECK: Unexpected variation in SYS_ID_DFR0_EL1. Boot CPU: 0x00000004011088, CPU4: 0x00000005011088
[ 0.143555] CPU features: Unsupported CPU feature variation detected.
[ 0.143702] GICv3: CPU4: found redistributor 10000 region 0:0x000000002f180000
[ 0.143702] GICv3: CPU4: using allocated LPI pending table @0x00000008800d0000
[ 0.144888] CPU4: Booted secondary processor 0x0000010000 [0x410fd0f0]
The boot CPU implements FEAT_PMUv3p1 (ID_DFR0_EL1.PerfMon, bits 27:24, is
0b0100), but CPU4, part of the other cluster, implements FEAT_PMUv3p4
(ID_DFR0_EL1.PerfMon = 0b0101).
Treat the PerfMon field as FTR_NONSTRICT and FTR_EXACT to pass the sanity
check and to match how PMUVer is treated for the 64bit ID register.
Fixes: b20d1ba3cf4b ("arm64: cpufeature: allow for version discrepancy in PMU implementations")
Signed-off-by: Alexandru Elisei <alexandru.elisei@arm.com>
Link: https://lore.kernel.org/r/20220617111332.203061-1-alexandru.elisei@arm.com
Signed-off-by: Will Deacon <will@kernel.org>
2022-06-17 14:13:32 +03:00
S_ARM64_FTR_BITS ( FTR_HIDDEN , FTR_NONSTRICT , FTR_EXACT , ID_DFR0_PERFMON_SHIFT , 4 , 0 ) ,
2020-07-03 06:51:37 +03:00
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_DFR0_MPROFDBG_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_DFR0_MMAPTRC_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_DFR0_COPTRC_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_DFR0_MMAPDBG_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_DFR0_COPSDBG_SHIFT , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_DFR0_COPDBG_SHIFT , 4 , 0 ) ,
2016-01-26 13:58:13 +03:00
ARM64_FTR_END ,
} ;
2020-05-19 12:40:42 +03:00
static const struct arm64_ftr_bits ftr_id_dfr1 [ ] = {
S_ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , ID_DFR1_MTPMU_SHIFT , 4 , 0 ) ,
ARM64_FTR_END ,
} ;
2017-10-31 18:51:10 +03:00
static const struct arm64_ftr_bits ftr_zcr [ ] = {
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_NONSTRICT , FTR_LOWER_SAFE ,
2022-05-10 19:11:58 +03:00
ZCR_ELx_LEN_SHIFT , ZCR_ELx_LEN_WIDTH , 0 ) , /* LEN */
2017-10-31 18:51:10 +03:00
ARM64_FTR_END ,
} ;
2022-04-19 14:22:17 +03:00
static const struct arm64_ftr_bits ftr_smcr [ ] = {
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_NONSTRICT , FTR_LOWER_SAFE ,
2022-05-10 19:11:58 +03:00
SMCR_ELx_LEN_SHIFT , SMCR_ELx_LEN_WIDTH , 0 ) , /* LEN */
2017-10-31 18:51:10 +03:00
ARM64_FTR_END ,
} ;
2015-10-19 16:24:45 +03:00
/*
* Common ftr bits for a 32 bit register with all hidden , strict
* attributes , with 4 bit feature fields and a default safe value of
* 0. Covers the following 32 bit registers :
2022-11-03 11:22:32 +03:00
* id_isar [ 1 - 3 ] , id_mmfr [ 1 - 3 ]
2015-10-19 16:24:45 +03:00
*/
2016-08-31 13:31:08 +03:00
static const struct arm64_ftr_bits ftr_generic_32bits [ ] = {
2017-01-09 20:28:30 +03:00
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , 28 , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , 24 , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , 20 , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , 16 , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , 12 , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , 8 , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , 4 , 4 , 0 ) ,
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_LOWER_SAFE , 0 , 4 , 0 ) ,
2015-10-19 16:24:45 +03:00
ARM64_FTR_END ,
} ;
2017-01-09 20:28:26 +03:00
/* Table for a single 32bit feature value */
static const struct arm64_ftr_bits ftr_single32 [ ] = {
2017-01-09 20:28:30 +03:00
ARM64_FTR_BITS ( FTR_HIDDEN , FTR_STRICT , FTR_EXACT , 0 , 32 , 0 ) ,
2015-10-19 16:24:45 +03:00
ARM64_FTR_END ,
} ;
2017-01-09 20:28:26 +03:00
static const struct arm64_ftr_bits ftr_raz [ ] = {
2015-10-19 16:24:45 +03:00
ARM64_FTR_END ,
} ;
2021-11-01 07:54:21 +03:00
# define __ARM64_FTR_REG_OVERRIDE(id_str, id, table, ovr) { \
2021-02-08 12:57:19 +03:00
. sys_id = id , \
. reg = & ( struct arm64_ftr_reg ) { \
2021-11-01 07:54:21 +03:00
. name = id_str , \
2021-02-08 12:57:19 +03:00
. override = ( ovr ) , \
. ftr_bits = & ( ( table ) [ 0 ] ) , \
2016-08-31 13:31:09 +03:00
} }
2015-10-19 16:24:45 +03:00
2021-11-01 07:54:21 +03:00
# define ARM64_FTR_REG_OVERRIDE(id, table, ovr) \
__ARM64_FTR_REG_OVERRIDE ( # id , id , table , ovr )
# define ARM64_FTR_REG(id, table) \
__ARM64_FTR_REG_OVERRIDE ( # id , id , table , & no_override )
2021-02-08 12:57:19 +03:00
2021-02-08 12:57:23 +03:00
struct arm64_ftr_override __ro_after_init id_aa64mmfr1_override ;
2022-06-30 19:04:59 +03:00
struct arm64_ftr_override __ro_after_init id_aa64pfr0_override ;
2021-02-08 12:57:29 +03:00
struct arm64_ftr_override __ro_after_init id_aa64pfr1_override ;
2022-06-30 19:04:59 +03:00
struct arm64_ftr_override __ro_after_init id_aa64zfr0_override ;
2022-06-30 19:04:58 +03:00
struct arm64_ftr_override __ro_after_init id_aa64smfr0_override ;
2021-02-08 12:57:31 +03:00
struct arm64_ftr_override __ro_after_init id_aa64isar1_override ;
2022-02-24 15:49:52 +03:00
struct arm64_ftr_override __ro_after_init id_aa64isar2_override ;
2021-02-08 12:57:23 +03:00
2016-08-31 13:31:09 +03:00
static const struct __ftr_reg_entry {
u32 sys_id ;
struct arm64_ftr_reg * reg ;
} arm64_ftr_regs [ ] = {
2015-10-19 16:24:45 +03:00
/* Op1 = 0, CRn = 0, CRm = 1 */
ARM64_FTR_REG ( SYS_ID_PFR0_EL1 , ftr_id_pfr0 ) ,
2020-04-21 17:29:16 +03:00
ARM64_FTR_REG ( SYS_ID_PFR1_EL1 , ftr_id_pfr1 ) ,
2016-01-26 13:58:13 +03:00
ARM64_FTR_REG ( SYS_ID_DFR0_EL1 , ftr_id_dfr0 ) ,
2015-10-19 16:24:45 +03:00
ARM64_FTR_REG ( SYS_ID_MMFR0_EL1 , ftr_id_mmfr0 ) ,
ARM64_FTR_REG ( SYS_ID_MMFR1_EL1 , ftr_generic_32bits ) ,
ARM64_FTR_REG ( SYS_ID_MMFR2_EL1 , ftr_generic_32bits ) ,
ARM64_FTR_REG ( SYS_ID_MMFR3_EL1 , ftr_generic_32bits ) ,
/* Op1 = 0, CRn = 0, CRm = 2 */
2020-05-19 12:40:38 +03:00
ARM64_FTR_REG ( SYS_ID_ISAR0_EL1 , ftr_id_isar0 ) ,
2015-10-19 16:24:45 +03:00
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 ) ,
2020-04-21 17:29:16 +03:00
ARM64_FTR_REG ( SYS_ID_ISAR4_EL1 , ftr_id_isar4 ) ,
2015-10-19 16:24:45 +03:00
ARM64_FTR_REG ( SYS_ID_ISAR5_EL1 , ftr_id_isar5 ) ,
ARM64_FTR_REG ( SYS_ID_MMFR4_EL1 , ftr_id_mmfr4 ) ,
2019-12-17 17:47:32 +03:00
ARM64_FTR_REG ( SYS_ID_ISAR6_EL1 , ftr_id_isar6 ) ,
2015-10-19 16:24:45 +03:00
/* Op1 = 0, CRn = 0, CRm = 3 */
2022-11-03 11:22:32 +03:00
ARM64_FTR_REG ( SYS_MVFR0_EL1 , ftr_mvfr0 ) ,
ARM64_FTR_REG ( SYS_MVFR1_EL1 , ftr_mvfr1 ) ,
2015-10-19 16:24:45 +03:00
ARM64_FTR_REG ( SYS_MVFR2_EL1 , ftr_mvfr2 ) ,
2020-05-19 12:40:41 +03:00
ARM64_FTR_REG ( SYS_ID_PFR2_EL1 , ftr_id_pfr2 ) ,
2020-05-19 12:40:42 +03:00
ARM64_FTR_REG ( SYS_ID_DFR1_EL1 , ftr_id_dfr1 ) ,
2020-05-19 12:40:43 +03:00
ARM64_FTR_REG ( SYS_ID_MMFR5_EL1 , ftr_id_mmfr5 ) ,
2015-10-19 16:24:45 +03:00
/* Op1 = 0, CRn = 0, CRm = 4 */
2022-06-30 19:04:59 +03:00
ARM64_FTR_REG_OVERRIDE ( SYS_ID_AA64PFR0_EL1 , ftr_id_aa64pfr0 ,
& id_aa64pfr0_override ) ,
2021-02-08 12:57:29 +03:00
ARM64_FTR_REG_OVERRIDE ( SYS_ID_AA64PFR1_EL1 , ftr_id_aa64pfr1 ,
& id_aa64pfr1_override ) ,
2022-06-30 19:04:59 +03:00
ARM64_FTR_REG_OVERRIDE ( SYS_ID_AA64ZFR0_EL1 , ftr_id_aa64zfr0 ,
& id_aa64zfr0_override ) ,
2022-06-30 19:04:58 +03:00
ARM64_FTR_REG_OVERRIDE ( SYS_ID_AA64SMFR0_EL1 , ftr_id_aa64smfr0 ,
& id_aa64smfr0_override ) ,
2015-10-19 16:24:45 +03:00
/* Op1 = 0, CRn = 0, CRm = 5 */
ARM64_FTR_REG ( SYS_ID_AA64DFR0_EL1 , ftr_id_aa64dfr0 ) ,
2017-01-09 20:28:26 +03:00
ARM64_FTR_REG ( SYS_ID_AA64DFR1_EL1 , ftr_raz ) ,
2015-10-19 16:24:45 +03:00
/* Op1 = 0, CRn = 0, CRm = 6 */
ARM64_FTR_REG ( SYS_ID_AA64ISAR0_EL1 , ftr_id_aa64isar0 ) ,
2021-02-08 12:57:31 +03:00
ARM64_FTR_REG_OVERRIDE ( SYS_ID_AA64ISAR1_EL1 , ftr_id_aa64isar1 ,
& id_aa64isar1_override ) ,
2022-02-24 15:49:52 +03:00
ARM64_FTR_REG_OVERRIDE ( SYS_ID_AA64ISAR2_EL1 , ftr_id_aa64isar2 ,
& id_aa64isar2_override ) ,
2015-10-19 16:24:45 +03:00
/* Op1 = 0, CRn = 0, CRm = 7 */
ARM64_FTR_REG ( SYS_ID_AA64MMFR0_EL1 , ftr_id_aa64mmfr0 ) ,
2021-02-08 12:57:23 +03:00
ARM64_FTR_REG_OVERRIDE ( SYS_ID_AA64MMFR1_EL1 , ftr_id_aa64mmfr1 ,
& id_aa64mmfr1_override ) ,
2016-02-05 17:58:47 +03:00
ARM64_FTR_REG ( SYS_ID_AA64MMFR2_EL1 , ftr_id_aa64mmfr2 ) ,
2015-10-19 16:24:45 +03:00
2017-10-31 18:51:10 +03:00
/* Op1 = 0, CRn = 1, CRm = 2 */
ARM64_FTR_REG ( SYS_ZCR_EL1 , ftr_zcr ) ,
2022-04-19 14:22:17 +03:00
ARM64_FTR_REG ( SYS_SMCR_EL1 , ftr_smcr ) ,
2017-10-31 18:51:10 +03:00
2021-05-26 22:36:21 +03:00
/* Op1 = 1, CRn = 0, CRm = 0 */
ARM64_FTR_REG ( SYS_GMID_EL1 , ftr_gmid ) ,
2015-10-19 16:24:45 +03:00
/* Op1 = 3, CRn = 0, CRm = 0 */
2016-08-31 13:31:10 +03:00
{ SYS_CTR_EL0 , & arm64_ftr_reg_ctrel0 } ,
2015-10-19 16:24:45 +03:00
ARM64_FTR_REG ( SYS_DCZID_EL0 , ftr_dczid ) ,
/* Op1 = 3, CRn = 14, CRm = 0 */
2017-01-09 20:28:26 +03:00
ARM64_FTR_REG ( SYS_CNTFRQ_EL0 , ftr_single32 ) ,
2015-10-19 16:24:45 +03:00
} ;
static int search_cmp_ftr_reg ( const void * id , const void * regp )
{
2016-08-31 13:31:09 +03:00
return ( int ) ( unsigned long ) id - ( int ) ( ( const struct __ftr_reg_entry * ) regp ) - > sys_id ;
2015-10-19 16:24:45 +03:00
}
/*
2020-05-27 13:04:36 +03:00
* get_arm64_ftr_reg_nowarn - Looks up a feature register entry using
* its sys_reg ( ) encoding . With the array arm64_ftr_regs sorted in the
* ascending order of sys_id , we use binary search to find a matching
2015-10-19 16:24:45 +03:00
* entry .
*
* returns - Upon success , matching ftr_reg entry for id .
* - NULL on failure . It is upto the caller to decide
* the impact of a failure .
*/
2020-05-27 13:04:36 +03:00
static struct arm64_ftr_reg * get_arm64_ftr_reg_nowarn ( u32 sys_id )
2015-10-19 16:24:45 +03:00
{
2016-08-31 13:31:09 +03:00
const struct __ftr_reg_entry * ret ;
ret = bsearch ( ( const void * ) ( unsigned long ) sys_id ,
2015-10-19 16:24:45 +03:00
arm64_ftr_regs ,
ARRAY_SIZE ( arm64_ftr_regs ) ,
sizeof ( arm64_ftr_regs [ 0 ] ) ,
search_cmp_ftr_reg ) ;
2016-08-31 13:31:09 +03:00
if ( ret )
return ret - > reg ;
return NULL ;
2015-10-19 16:24:45 +03:00
}
2020-05-27 13:04:36 +03:00
/*
* get_arm64_ftr_reg - Looks up a feature register entry using
* its sys_reg ( ) encoding . This calls get_arm64_ftr_reg_nowarn ( ) .
*
* returns - Upon success , matching ftr_reg entry for id .
* - NULL on failure but with an WARN_ON ( ) .
*/
2022-09-09 19:59:37 +03:00
struct arm64_ftr_reg * get_arm64_ftr_reg ( u32 sys_id )
2020-05-27 13:04:36 +03:00
{
struct arm64_ftr_reg * reg ;
reg = get_arm64_ftr_reg_nowarn ( sys_id ) ;
/*
* Requesting a non - existent register search is an error . Warn
* and let the caller handle it .
*/
WARN_ON ( ! reg ) ;
return reg ;
}
2016-08-31 13:31:08 +03:00
static u64 arm64_ftr_set_value ( const struct arm64_ftr_bits * ftrp , s64 reg ,
s64 ftr_val )
2015-10-19 16:24:45 +03:00
{
u64 mask = arm64_ftr_mask ( ftrp ) ;
reg & = ~ mask ;
reg | = ( ftr_val < < ftrp - > shift ) & mask ;
return reg ;
}
2016-08-31 13:31:08 +03:00
static s64 arm64_ftr_safe_value ( const struct arm64_ftr_bits * ftrp , s64 new ,
s64 cur )
2015-10-19 16:24:45 +03:00
{
s64 ret = 0 ;
switch ( ftrp - > type ) {
case FTR_EXACT :
ret = ftrp - > safe_val ;
break ;
case FTR_LOWER_SAFE :
2021-04-29 23:50:46 +03:00
ret = min ( new , cur ) ;
2015-10-19 16:24:45 +03:00
break ;
2019-07-30 17:40:20 +03:00
case FTR_HIGHER_OR_ZERO_SAFE :
if ( ! cur | | ! new )
break ;
2020-08-24 01:36:59 +03:00
fallthrough ;
2015-10-19 16:24:45 +03:00
case FTR_HIGHER_SAFE :
2021-04-29 23:50:46 +03:00
ret = max ( new , cur ) ;
2015-10-19 16:24:45 +03:00
break ;
default :
BUG ( ) ;
}
return ret ;
}
static void __init sort_ftr_regs ( void )
{
2020-07-07 17:23:13 +03:00
unsigned int i ;
for ( i = 0 ; i < ARRAY_SIZE ( arm64_ftr_regs ) ; i + + ) {
const struct arm64_ftr_reg * ftr_reg = arm64_ftr_regs [ i ] . reg ;
const struct arm64_ftr_bits * ftr_bits = ftr_reg - > ftr_bits ;
unsigned int j = 0 ;
/*
* Features here must be sorted in descending order with respect
* to their shift values and should not overlap with each other .
*/
for ( ; ftr_bits - > width ! = 0 ; ftr_bits + + , j + + ) {
unsigned int width = ftr_reg - > ftr_bits [ j ] . width ;
unsigned int shift = ftr_reg - > ftr_bits [ j ] . shift ;
unsigned int prev_shift ;
WARN ( ( shift + width ) > 64 ,
" %s has invalid feature at shift %d \n " ,
ftr_reg - > name , shift ) ;
/*
* Skip the first feature . There is nothing to
* compare against for now .
*/
if ( j = = 0 )
continue ;
prev_shift = ftr_reg - > ftr_bits [ j - 1 ] . shift ;
WARN ( ( shift + width ) > prev_shift ,
" %s has feature overlap at shift %d \n " ,
ftr_reg - > name , shift ) ;
}
2016-08-31 13:31:09 +03:00
2020-07-07 17:23:13 +03:00
/*
* Skip the first register . There is nothing to
* compare against for now .
*/
if ( i = = 0 )
continue ;
/*
* Registers here must be sorted in ascending order with respect
* to sys_id for subsequent binary search in get_arm64_ftr_reg ( )
* to work correctly .
*/
2022-05-11 19:20:30 +03:00
BUG_ON ( arm64_ftr_regs [ i ] . sys_id < = arm64_ftr_regs [ i - 1 ] . sys_id ) ;
2020-07-07 17:23:13 +03:00
}
2015-10-19 16:24:45 +03:00
}
/*
* Initialise the CPU feature register from Boot CPU values .
* Also initiliases the strict_mask for the register .
2017-01-09 20:28:24 +03:00
* Any bits that are not covered by an arm64_ftr_bits entry are considered
* RES0 for the system - wide value , and must strictly match .
2015-10-19 16:24:45 +03:00
*/
2021-06-08 21:02:55 +03:00
static void init_cpu_ftr_reg ( u32 sys_reg , u64 new )
2015-10-19 16:24:45 +03:00
{
u64 val = 0 ;
u64 strict_mask = ~ 0x0ULL ;
2017-01-09 20:28:30 +03:00
u64 user_mask = 0 ;
2017-01-09 20:28:24 +03:00
u64 valid_mask = 0 ;
2016-08-31 13:31:08 +03:00
const struct arm64_ftr_bits * ftrp ;
2015-10-19 16:24:45 +03:00
struct arm64_ftr_reg * reg = get_arm64_ftr_reg ( sys_reg ) ;
2020-05-27 13:04:36 +03:00
if ( ! reg )
return ;
2015-10-19 16:24:45 +03:00
2020-03-11 09:52:49 +03:00
for ( ftrp = reg - > ftr_bits ; ftrp - > width ; ftrp + + ) {
2017-01-09 20:28:24 +03:00
u64 ftr_mask = arm64_ftr_mask ( ftrp ) ;
2015-10-19 16:24:45 +03:00
s64 ftr_new = arm64_ftr_value ( ftrp , new ) ;
2021-02-08 12:57:19 +03:00
s64 ftr_ovr = arm64_ftr_value ( ftrp , reg - > override - > val ) ;
if ( ( ftr_mask & reg - > override - > mask ) = = ftr_mask ) {
s64 tmp = arm64_ftr_safe_value ( ftrp , ftr_ovr , ftr_new ) ;
char * str = NULL ;
if ( ftr_ovr ! = tmp ) {
/* Unsafe, remove the override */
reg - > override - > mask & = ~ ftr_mask ;
reg - > override - > val & = ~ ftr_mask ;
tmp = ftr_ovr ;
str = " ignoring override " ;
} else if ( ftr_new ! = tmp ) {
/* Override was valid */
ftr_new = tmp ;
str = " forced " ;
} else if ( ftr_ovr = = tmp ) {
/* Override was the safe value */
str = " already set " ;
}
if ( str )
pr_warn ( " %s[%d:%d]: %s to %llx \n " ,
reg - > name ,
ftrp - > shift + ftrp - > width - 1 ,
ftrp - > shift , str , tmp ) ;
2021-04-08 16:10:08 +03:00
} else if ( ( ftr_mask & reg - > override - > val ) = = ftr_mask ) {
reg - > override - > val & = ~ ftr_mask ;
pr_warn ( " %s[%d:%d]: impossible override, ignored \n " ,
reg - > name ,
ftrp - > shift + ftrp - > width - 1 ,
ftrp - > shift ) ;
2021-02-08 12:57:19 +03:00
}
2015-10-19 16:24:45 +03:00
val = arm64_ftr_set_value ( ftrp , val , ftr_new ) ;
2017-01-09 20:28:24 +03:00
valid_mask | = ftr_mask ;
2015-10-19 16:24:45 +03:00
if ( ! ftrp - > strict )
2017-01-09 20:28:24 +03:00
strict_mask & = ~ ftr_mask ;
2017-01-09 20:28:30 +03:00
if ( ftrp - > visible )
user_mask | = ftr_mask ;
else
reg - > user_val = arm64_ftr_set_value ( ftrp ,
reg - > user_val ,
ftrp - > safe_val ) ;
2015-10-19 16:24:45 +03:00
}
2017-01-09 20:28:24 +03:00
val & = valid_mask ;
2015-10-19 16:24:45 +03:00
reg - > sys_val = val ;
reg - > strict_mask = strict_mask ;
2017-01-09 20:28:30 +03:00
reg - > user_mask = user_mask ;
2015-10-19 16:24:45 +03:00
}
2018-03-26 17:12:30 +03:00
extern const struct arm64_cpu_capabilities arm64_errata [ ] ;
arm64: capabilities: Speed up capability lookup
We maintain two separate tables of capabilities, errata and features,
which decide the system capabilities. We iterate over each of these
tables for various operations (e.g, detection, verification etc.).
We do not have a way to map a system "capability" to its entry,
(i.e, cap -> struct arm64_cpu_capabilities) which is needed for
this_cpu_has_cap(). So we iterate over the table one by one to
find the entry and then do the operation. Also, this prevents
us from optimizing the way we "enable" the capabilities on the
CPUs, where we now issue a stop_machine() for each available
capability.
One solution is to merge the two tables into a single table,
sorted by the capability. But this is has the following
disadvantages:
- We loose the "classification" of an errata vs. feature
- It is quite easy to make a mistake when adding an entry,
unless we sort the table at runtime.
So we maintain a list of pointers to the capability entry, sorted
by the "cap number" in a separate array, initialized at boot time.
The only restriction is that we can have one "entry" per capability.
While at it, remove the duplicate declaration of arm64_errata table.
Reviewed-by: Vladimir Murzin <vladimir.murzin@arm.com>
Tested-by: Vladimir Murzin <vladimir.murzin@arm.com>
Signed-off-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
2018-11-30 20:18:03 +03:00
static const struct arm64_cpu_capabilities arm64_features [ ] ;
static void __init
init_cpu_hwcaps_indirect_list_from_array ( const struct arm64_cpu_capabilities * caps )
{
for ( ; caps - > matches ; caps + + ) {
if ( WARN ( caps - > capability > = ARM64_NCAPS ,
" Invalid capability %d \n " , caps - > capability ) )
continue ;
if ( WARN ( cpu_hwcaps_ptrs [ caps - > capability ] ,
" Duplicate entry for capability %d \n " ,
caps - > capability ) )
continue ;
cpu_hwcaps_ptrs [ caps - > capability ] = caps ;
}
}
static void __init init_cpu_hwcaps_indirect_list ( void )
{
init_cpu_hwcaps_indirect_list_from_array ( arm64_features ) ;
init_cpu_hwcaps_indirect_list_from_array ( arm64_errata ) ;
}
2018-03-26 17:12:41 +03:00
static void __init setup_boot_cpu_capabilities ( void ) ;
2018-03-26 17:12:30 +03:00
2021-06-08 21:02:55 +03:00
static void init_32bit_cpu_features ( struct cpuinfo_32bit * info )
2021-06-08 21:02:54 +03:00
{
init_cpu_ftr_reg ( SYS_ID_DFR0_EL1 , info - > reg_id_dfr0 ) ;
init_cpu_ftr_reg ( SYS_ID_DFR1_EL1 , info - > reg_id_dfr1 ) ;
init_cpu_ftr_reg ( SYS_ID_ISAR0_EL1 , info - > reg_id_isar0 ) ;
init_cpu_ftr_reg ( SYS_ID_ISAR1_EL1 , info - > reg_id_isar1 ) ;
init_cpu_ftr_reg ( SYS_ID_ISAR2_EL1 , info - > reg_id_isar2 ) ;
init_cpu_ftr_reg ( SYS_ID_ISAR3_EL1 , info - > reg_id_isar3 ) ;
init_cpu_ftr_reg ( SYS_ID_ISAR4_EL1 , info - > reg_id_isar4 ) ;
init_cpu_ftr_reg ( SYS_ID_ISAR5_EL1 , info - > reg_id_isar5 ) ;
init_cpu_ftr_reg ( SYS_ID_ISAR6_EL1 , info - > reg_id_isar6 ) ;
init_cpu_ftr_reg ( SYS_ID_MMFR0_EL1 , info - > reg_id_mmfr0 ) ;
init_cpu_ftr_reg ( SYS_ID_MMFR1_EL1 , info - > reg_id_mmfr1 ) ;
init_cpu_ftr_reg ( SYS_ID_MMFR2_EL1 , info - > reg_id_mmfr2 ) ;
init_cpu_ftr_reg ( SYS_ID_MMFR3_EL1 , info - > reg_id_mmfr3 ) ;
init_cpu_ftr_reg ( SYS_ID_MMFR4_EL1 , info - > reg_id_mmfr4 ) ;
init_cpu_ftr_reg ( SYS_ID_MMFR5_EL1 , info - > reg_id_mmfr5 ) ;
init_cpu_ftr_reg ( SYS_ID_PFR0_EL1 , info - > reg_id_pfr0 ) ;
init_cpu_ftr_reg ( SYS_ID_PFR1_EL1 , info - > reg_id_pfr1 ) ;
init_cpu_ftr_reg ( SYS_ID_PFR2_EL1 , info - > reg_id_pfr2 ) ;
init_cpu_ftr_reg ( SYS_MVFR0_EL1 , info - > reg_mvfr0 ) ;
init_cpu_ftr_reg ( SYS_MVFR1_EL1 , info - > reg_mvfr1 ) ;
init_cpu_ftr_reg ( SYS_MVFR2_EL1 , info - > reg_mvfr2 ) ;
}
2015-10-19 16:24:45 +03:00
void __init init_cpu_features ( struct cpuinfo_arm64 * info )
{
/* Before we start using the tables, make sure it is sorted */
sort_ftr_regs ( ) ;
init_cpu_ftr_reg ( SYS_CTR_EL0 , info - > reg_ctr ) ;
init_cpu_ftr_reg ( SYS_DCZID_EL0 , info - > reg_dczid ) ;
init_cpu_ftr_reg ( SYS_CNTFRQ_EL0 , info - > reg_cntfrq ) ;
init_cpu_ftr_reg ( SYS_ID_AA64DFR0_EL1 , info - > reg_id_aa64dfr0 ) ;
init_cpu_ftr_reg ( SYS_ID_AA64DFR1_EL1 , info - > reg_id_aa64dfr1 ) ;
init_cpu_ftr_reg ( SYS_ID_AA64ISAR0_EL1 , info - > reg_id_aa64isar0 ) ;
init_cpu_ftr_reg ( SYS_ID_AA64ISAR1_EL1 , info - > reg_id_aa64isar1 ) ;
2021-12-10 19:54:31 +03:00
init_cpu_ftr_reg ( SYS_ID_AA64ISAR2_EL1 , info - > reg_id_aa64isar2 ) ;
2015-10-19 16:24:45 +03:00
init_cpu_ftr_reg ( SYS_ID_AA64MMFR0_EL1 , info - > reg_id_aa64mmfr0 ) ;
init_cpu_ftr_reg ( SYS_ID_AA64MMFR1_EL1 , info - > reg_id_aa64mmfr1 ) ;
2016-02-05 17:58:47 +03:00
init_cpu_ftr_reg ( SYS_ID_AA64MMFR2_EL1 , info - > reg_id_aa64mmfr2 ) ;
2015-10-19 16:24:45 +03:00
init_cpu_ftr_reg ( SYS_ID_AA64PFR0_EL1 , info - > reg_id_aa64pfr0 ) ;
init_cpu_ftr_reg ( SYS_ID_AA64PFR1_EL1 , info - > reg_id_aa64pfr1 ) ;
2017-10-31 18:51:10 +03:00
init_cpu_ftr_reg ( SYS_ID_AA64ZFR0_EL1 , info - > reg_id_aa64zfr0 ) ;
2022-04-19 14:22:16 +03:00
init_cpu_ftr_reg ( SYS_ID_AA64SMFR0_EL1 , info - > reg_id_aa64smfr0 ) ;
2016-04-18 12:28:35 +03:00
2021-06-08 21:02:54 +03:00
if ( id_aa64pfr0_32bit_el0 ( info - > reg_id_aa64pfr0 ) )
init_32bit_cpu_features ( & info - > aarch32 ) ;
2016-04-18 12:28:35 +03:00
2022-07-20 13:52:19 +03:00
if ( IS_ENABLED ( CONFIG_ARM64_SVE ) & &
id_aa64pfr0_sve ( read_sanitised_ftr_reg ( SYS_ID_AA64PFR0_EL1 ) ) ) {
info - > reg_zcr = read_zcr_features ( ) ;
2017-10-31 18:51:10 +03:00
init_cpu_ftr_reg ( SYS_ZCR_EL1 , info - > reg_zcr ) ;
2021-10-19 20:22:12 +03:00
vec_init_vq_map ( ARM64_VEC_SVE ) ;
2017-10-31 18:51:10 +03:00
}
2018-03-26 17:12:29 +03:00
2022-07-20 13:52:19 +03:00
if ( IS_ENABLED ( CONFIG_ARM64_SME ) & &
id_aa64pfr1_sme ( read_sanitised_ftr_reg ( SYS_ID_AA64PFR1_EL1 ) ) ) {
info - > reg_smcr = read_smcr_features ( ) ;
/*
* We mask out SMPS since even if the hardware
* supports priorities the kernel does not at present
* and we block access to them .
*/
info - > reg_smidr = read_cpuid ( SMIDR_EL1 ) & ~ SMIDR_EL1_SMPS ;
2022-04-19 14:22:17 +03:00
init_cpu_ftr_reg ( SYS_SMCR_EL1 , info - > reg_smcr ) ;
2022-07-20 13:52:19 +03:00
vec_init_vq_map ( ARM64_VEC_SME ) ;
2022-04-19 14:22:17 +03:00
}
2021-05-26 22:36:21 +03:00
if ( id_aa64pfr1_mte ( info - > reg_id_aa64pfr1 ) )
init_cpu_ftr_reg ( SYS_GMID_EL1 , info - > reg_gmid ) ;
arm64: capabilities: Speed up capability lookup
We maintain two separate tables of capabilities, errata and features,
which decide the system capabilities. We iterate over each of these
tables for various operations (e.g, detection, verification etc.).
We do not have a way to map a system "capability" to its entry,
(i.e, cap -> struct arm64_cpu_capabilities) which is needed for
this_cpu_has_cap(). So we iterate over the table one by one to
find the entry and then do the operation. Also, this prevents
us from optimizing the way we "enable" the capabilities on the
CPUs, where we now issue a stop_machine() for each available
capability.
One solution is to merge the two tables into a single table,
sorted by the capability. But this is has the following
disadvantages:
- We loose the "classification" of an errata vs. feature
- It is quite easy to make a mistake when adding an entry,
unless we sort the table at runtime.
So we maintain a list of pointers to the capability entry, sorted
by the "cap number" in a separate array, initialized at boot time.
The only restriction is that we can have one "entry" per capability.
While at it, remove the duplicate declaration of arm64_errata table.
Reviewed-by: Vladimir Murzin <vladimir.murzin@arm.com>
Tested-by: Vladimir Murzin <vladimir.murzin@arm.com>
Signed-off-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
2018-11-30 20:18:03 +03:00
/*
* Initialize the indirect array of CPU hwcaps capabilities pointers
* before we handle the boot CPU below .
*/
init_cpu_hwcaps_indirect_list ( ) ;
2018-03-26 17:12:29 +03:00
/*
2018-03-26 17:12:41 +03:00
* Detect and enable early CPU capabilities based on the boot CPU ,
* after we have initialised the CPU feature infrastructure .
2018-03-26 17:12:29 +03:00
*/
2018-03-26 17:12:41 +03:00
setup_boot_cpu_capabilities ( ) ;
2015-10-19 16:24:45 +03:00
}
2015-10-19 16:24:46 +03:00
static void update_cpu_ftr_reg ( struct arm64_ftr_reg * reg , u64 new )
2015-10-19 16:24:45 +03:00
{
2016-08-31 13:31:08 +03:00
const struct arm64_ftr_bits * ftrp ;
2015-10-19 16:24:45 +03:00
for ( ftrp = reg - > ftr_bits ; ftrp - > width ; ftrp + + ) {
s64 ftr_cur = arm64_ftr_value ( ftrp , reg - > sys_val ) ;
s64 ftr_new = arm64_ftr_value ( ftrp , new ) ;
if ( ftr_cur = = ftr_new )
continue ;
/* Find a safe value */
ftr_new = arm64_ftr_safe_value ( ftrp , ftr_new , ftr_cur ) ;
reg - > sys_val = arm64_ftr_set_value ( ftrp , reg - > sys_val , ftr_new ) ;
}
}
2015-10-19 16:24:46 +03:00
static int check_update_ftr_reg ( u32 sys_id , int cpu , u64 val , u64 boot )
2015-10-19 16:24:42 +03:00
{
2015-10-19 16:24:46 +03:00
struct arm64_ftr_reg * regp = get_arm64_ftr_reg ( sys_id ) ;
2020-05-27 13:04:36 +03:00
if ( ! regp )
return 0 ;
2015-10-19 16:24:46 +03:00
update_cpu_ftr_reg ( regp , val ) ;
if ( ( boot & regp - > strict_mask ) = = ( val & regp - > strict_mask ) )
return 0 ;
pr_warn ( " SANITY CHECK: Unexpected variation in %s. Boot CPU: %#016llx, CPU%d: %#016llx \n " ,
regp - > name , boot , cpu , val ) ;
return 1 ;
}
2020-04-21 17:29:20 +03:00
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 ) ;
2020-05-27 13:04:36 +03:00
if ( ! regp )
2020-04-21 17:29:20 +03:00
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 ) ;
}
2021-06-08 21:02:55 +03:00
static void lazy_init_32bit_cpu_features ( struct cpuinfo_arm64 * info ,
struct cpuinfo_arm64 * boot )
{
static bool boot_cpu_32bit_regs_overridden = false ;
if ( ! allow_mismatched_32bit_el0 | | boot_cpu_32bit_regs_overridden )
return ;
if ( id_aa64pfr0_32bit_el0 ( boot - > reg_id_aa64pfr0 ) )
return ;
boot - > aarch32 = info - > aarch32 ;
init_32bit_cpu_features ( & boot - > aarch32 ) ;
boot_cpu_32bit_regs_overridden = true ;
}
2021-06-08 21:02:54 +03:00
static int update_32bit_cpu_features ( int cpu , struct cpuinfo_32bit * info ,
struct cpuinfo_32bit * boot )
2020-04-21 17:29:19 +03:00
{
int taint = 0 ;
u64 pfr0 = read_sanitised_ftr_reg ( SYS_ID_AA64PFR0_EL1 ) ;
2020-04-21 17:29:20 +03:00
/*
* 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 ) ;
}
2020-04-21 17:29:19 +03:00
taint | = check_update_ftr_reg ( SYS_ID_DFR0_EL1 , cpu ,
info - > reg_id_dfr0 , boot - > reg_id_dfr0 ) ;
2020-05-19 12:40:42 +03:00
taint | = check_update_ftr_reg ( SYS_ID_DFR1_EL1 , cpu ,
info - > reg_id_dfr1 , boot - > reg_id_dfr1 ) ;
2020-04-21 17:29:19 +03:00
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 ) ;
2020-05-19 12:40:54 +03:00
taint | = check_update_ftr_reg ( SYS_ID_MMFR4_EL1 , cpu ,
info - > reg_id_mmfr4 , boot - > reg_id_mmfr4 ) ;
2020-05-19 12:40:43 +03:00
taint | = check_update_ftr_reg ( SYS_ID_MMFR5_EL1 , cpu ,
info - > reg_id_mmfr5 , boot - > reg_id_mmfr5 ) ;
2020-04-21 17:29:19 +03:00
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 ) ;
2020-05-19 12:40:41 +03:00
taint | = check_update_ftr_reg ( SYS_ID_PFR2_EL1 , cpu ,
info - > reg_id_pfr2 , boot - > reg_id_pfr2 ) ;
2020-04-21 17:29:19 +03:00
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 ;
}
2015-10-19 16:24:46 +03:00
/*
* Update system wide CPU feature registers with the values from a
* non - boot CPU . Also performs SANITY checks to make sure that there
* aren ' t any insane variations from that of the boot CPU .
*/
void update_cpu_features ( int cpu ,
struct cpuinfo_arm64 * info ,
struct cpuinfo_arm64 * boot )
{
int taint = 0 ;
/*
* The kernel can handle differing I - cache policies , but otherwise
* caches should look identical . Userspace JITs will make use of
* * minLine .
*/
taint | = check_update_ftr_reg ( SYS_CTR_EL0 , cpu ,
info - > reg_ctr , boot - > reg_ctr ) ;
/*
* Userspace may perform DC ZVA instructions . Mismatched block sizes
* could result in too much or too little memory being zeroed if a
* process is preempted and migrated between CPUs .
*/
taint | = check_update_ftr_reg ( SYS_DCZID_EL0 , cpu ,
info - > reg_dczid , boot - > reg_dczid ) ;
/* If different, timekeeping will be broken (especially with KVM) */
taint | = check_update_ftr_reg ( SYS_CNTFRQ_EL0 , cpu ,
info - > reg_cntfrq , boot - > reg_cntfrq ) ;
/*
* The kernel uses self - hosted debug features and expects CPUs to
* support identical debug features . We presently need CTX_CMPs , WRPs ,
* and BRPs to be identical .
* ID_AA64DFR1 is currently RES0 .
*/
taint | = check_update_ftr_reg ( SYS_ID_AA64DFR0_EL1 , cpu ,
info - > reg_id_aa64dfr0 , boot - > reg_id_aa64dfr0 ) ;
taint | = check_update_ftr_reg ( SYS_ID_AA64DFR1_EL1 , cpu ,
info - > reg_id_aa64dfr1 , boot - > reg_id_aa64dfr1 ) ;
/*
* Even in big . LITTLE , processors should be identical instruction - set
* wise .
*/
taint | = check_update_ftr_reg ( SYS_ID_AA64ISAR0_EL1 , cpu ,
info - > reg_id_aa64isar0 , boot - > reg_id_aa64isar0 ) ;
taint | = check_update_ftr_reg ( SYS_ID_AA64ISAR1_EL1 , cpu ,
info - > reg_id_aa64isar1 , boot - > reg_id_aa64isar1 ) ;
2021-12-10 19:54:31 +03:00
taint | = check_update_ftr_reg ( SYS_ID_AA64ISAR2_EL1 , cpu ,
info - > reg_id_aa64isar2 , boot - > reg_id_aa64isar2 ) ;
2015-10-19 16:24:46 +03:00
/*
* Differing PARange support is fine as long as all peripherals and
* memory are mapped within the minimum PARange of all CPUs .
* Linux should not care about secure memory .
*/
taint | = check_update_ftr_reg ( SYS_ID_AA64MMFR0_EL1 , cpu ,
info - > reg_id_aa64mmfr0 , boot - > reg_id_aa64mmfr0 ) ;
taint | = check_update_ftr_reg ( SYS_ID_AA64MMFR1_EL1 , cpu ,
info - > reg_id_aa64mmfr1 , boot - > reg_id_aa64mmfr1 ) ;
2016-02-05 17:58:47 +03:00
taint | = check_update_ftr_reg ( SYS_ID_AA64MMFR2_EL1 , cpu ,
info - > reg_id_aa64mmfr2 , boot - > reg_id_aa64mmfr2 ) ;
2015-10-19 16:24:46 +03:00
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 ,
info - > reg_id_aa64pfr1 , boot - > reg_id_aa64pfr1 ) ;
2017-10-31 18:51:10 +03:00
taint | = check_update_ftr_reg ( SYS_ID_AA64ZFR0_EL1 , cpu ,
info - > reg_id_aa64zfr0 , boot - > reg_id_aa64zfr0 ) ;
2022-04-19 14:22:17 +03:00
taint | = check_update_ftr_reg ( SYS_ID_AA64SMFR0_EL1 , cpu ,
info - > reg_id_aa64smfr0 , boot - > reg_id_aa64smfr0 ) ;
2022-07-20 13:52:19 +03:00
if ( IS_ENABLED ( CONFIG_ARM64_SVE ) & &
id_aa64pfr0_sve ( read_sanitised_ftr_reg ( SYS_ID_AA64PFR0_EL1 ) ) ) {
info - > reg_zcr = read_zcr_features ( ) ;
2017-10-31 18:51:10 +03:00
taint | = check_update_ftr_reg ( SYS_ZCR_EL1 , cpu ,
info - > reg_zcr , boot - > reg_zcr ) ;
2022-07-20 13:52:19 +03:00
/* Probe vector lengths */
if ( ! system_capabilities_finalized ( ) )
2021-10-19 20:22:12 +03:00
vec_update_vq_map ( ARM64_VEC_SVE ) ;
2017-10-31 18:51:10 +03:00
}
2022-07-20 13:52:19 +03:00
if ( IS_ENABLED ( CONFIG_ARM64_SME ) & &
id_aa64pfr1_sme ( read_sanitised_ftr_reg ( SYS_ID_AA64PFR1_EL1 ) ) ) {
info - > reg_smcr = read_smcr_features ( ) ;
/*
* We mask out SMPS since even if the hardware
* supports priorities the kernel does not at present
* and we block access to them .
*/
info - > reg_smidr = read_cpuid ( SMIDR_EL1 ) & ~ SMIDR_EL1_SMPS ;
2022-04-19 14:22:17 +03:00
taint | = check_update_ftr_reg ( SYS_SMCR_EL1 , cpu ,
info - > reg_smcr , boot - > reg_smcr ) ;
2022-07-20 13:52:19 +03:00
/* Probe vector lengths */
if ( ! system_capabilities_finalized ( ) )
2022-04-19 14:22:17 +03:00
vec_update_vq_map ( ARM64_VEC_SME ) ;
}
2021-05-26 22:36:21 +03:00
/*
* The kernel uses the LDGM / STGM instructions and the number of tags
* they read / write depends on the GMID_EL1 . BS field . Check that the
* value is the same on all CPUs .
*/
if ( IS_ENABLED ( CONFIG_ARM64_MTE ) & &
2021-06-08 21:02:54 +03:00
id_aa64pfr1_mte ( info - > reg_id_aa64pfr1 ) ) {
2021-05-26 22:36:21 +03:00
taint | = check_update_ftr_reg ( SYS_GMID_EL1 , cpu ,
info - > reg_gmid , boot - > reg_gmid ) ;
2021-06-08 21:02:54 +03:00
}
2021-05-26 22:36:21 +03:00
2020-04-21 17:29:19 +03:00
/*
2021-06-08 21:02:54 +03:00
* 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 .
*
2020-04-21 17:29:19 +03:00
* This relies on a sanitised view of the AArch64 ID registers
* ( e . g . SYS_ID_AA64PFR0_EL1 ) , so we call it last .
*/
2021-06-08 21:02:54 +03:00
if ( id_aa64pfr0_32bit_el0 ( info - > reg_id_aa64pfr0 ) ) {
2021-06-08 21:02:55 +03:00
lazy_init_32bit_cpu_features ( info , boot ) ;
2021-06-08 21:02:54 +03:00
taint | = update_32bit_cpu_features ( cpu , & info - > aarch32 ,
& boot - > aarch32 ) ;
}
2020-04-21 17:29:19 +03:00
2015-10-19 16:24:46 +03:00
/*
* Mismatched CPU features are a recipe for disaster . Don ' t even
* pretend to support them .
*/
2017-06-05 13:40:23 +03:00
if ( taint ) {
pr_warn_once ( " Unsupported CPU feature variation detected. \n " ) ;
add_taint ( TAINT_CPU_OUT_OF_SPEC , LOCKDEP_STILL_OK ) ;
}
2015-10-19 16:24:42 +03:00
}
2017-03-23 18:14:39 +03:00
u64 read_sanitised_ftr_reg ( u32 id )
2015-10-19 16:24:47 +03:00
{
struct arm64_ftr_reg * regp = get_arm64_ftr_reg ( id ) ;
2020-05-27 13:04:36 +03:00
if ( ! regp )
return 0 ;
2015-10-19 16:24:47 +03:00
return regp - > sys_val ;
}
2020-09-18 13:18:46 +03:00
EXPORT_SYMBOL_GPL ( read_sanitised_ftr_reg ) ;
2015-03-27 16:09:23 +03:00
2017-02-02 20:32:15 +03:00
# define read_sysreg_case(r) \
2021-02-08 12:57:20 +03:00
case r : val = read_sysreg_s ( r ) ; break ;
2017-02-02 20:32:15 +03:00
2016-04-22 14:25:31 +03:00
/*
2017-03-23 18:14:39 +03:00
* __read_sysreg_by_encoding ( ) - Used by a STARTING cpu before cpuinfo is populated .
2016-04-22 14:25:31 +03:00
* Read the system register on the current CPU
*/
2021-02-08 12:57:20 +03:00
u64 __read_sysreg_by_encoding ( u32 sys_id )
2016-04-22 14:25:31 +03:00
{
2021-02-08 12:57:20 +03:00
struct arm64_ftr_reg * regp ;
u64 val ;
2016-04-22 14:25:31 +03:00
switch ( sys_id ) {
2017-02-02 20:32:15 +03:00
read_sysreg_case ( SYS_ID_PFR0_EL1 ) ;
read_sysreg_case ( SYS_ID_PFR1_EL1 ) ;
2020-05-19 12:40:41 +03:00
read_sysreg_case ( SYS_ID_PFR2_EL1 ) ;
2017-02-02 20:32:15 +03:00
read_sysreg_case ( SYS_ID_DFR0_EL1 ) ;
2020-05-19 12:40:42 +03:00
read_sysreg_case ( SYS_ID_DFR1_EL1 ) ;
2017-02-02 20:32:15 +03:00
read_sysreg_case ( SYS_ID_MMFR0_EL1 ) ;
read_sysreg_case ( SYS_ID_MMFR1_EL1 ) ;
read_sysreg_case ( SYS_ID_MMFR2_EL1 ) ;
read_sysreg_case ( SYS_ID_MMFR3_EL1 ) ;
2020-05-19 12:40:54 +03:00
read_sysreg_case ( SYS_ID_MMFR4_EL1 ) ;
2020-05-19 12:40:43 +03:00
read_sysreg_case ( SYS_ID_MMFR5_EL1 ) ;
2017-02-02 20:32:15 +03:00
read_sysreg_case ( SYS_ID_ISAR0_EL1 ) ;
read_sysreg_case ( SYS_ID_ISAR1_EL1 ) ;
read_sysreg_case ( SYS_ID_ISAR2_EL1 ) ;
read_sysreg_case ( SYS_ID_ISAR3_EL1 ) ;
read_sysreg_case ( SYS_ID_ISAR4_EL1 ) ;
read_sysreg_case ( SYS_ID_ISAR5_EL1 ) ;
2019-12-17 17:47:32 +03:00
read_sysreg_case ( SYS_ID_ISAR6_EL1 ) ;
2017-02-02 20:32:15 +03:00
read_sysreg_case ( SYS_MVFR0_EL1 ) ;
read_sysreg_case ( SYS_MVFR1_EL1 ) ;
read_sysreg_case ( SYS_MVFR2_EL1 ) ;
read_sysreg_case ( SYS_ID_AA64PFR0_EL1 ) ;
read_sysreg_case ( SYS_ID_AA64PFR1_EL1 ) ;
2019-06-03 18:35:02 +03:00
read_sysreg_case ( SYS_ID_AA64ZFR0_EL1 ) ;
2022-04-27 16:08:28 +03:00
read_sysreg_case ( SYS_ID_AA64SMFR0_EL1 ) ;
2017-02-02 20:32:15 +03:00
read_sysreg_case ( SYS_ID_AA64DFR0_EL1 ) ;
read_sysreg_case ( SYS_ID_AA64DFR1_EL1 ) ;
read_sysreg_case ( SYS_ID_AA64MMFR0_EL1 ) ;
read_sysreg_case ( SYS_ID_AA64MMFR1_EL1 ) ;
read_sysreg_case ( SYS_ID_AA64MMFR2_EL1 ) ;
read_sysreg_case ( SYS_ID_AA64ISAR0_EL1 ) ;
read_sysreg_case ( SYS_ID_AA64ISAR1_EL1 ) ;
2021-12-10 19:54:31 +03:00
read_sysreg_case ( SYS_ID_AA64ISAR2_EL1 ) ;
2017-02-02 20:32:15 +03:00
read_sysreg_case ( SYS_CNTFRQ_EL0 ) ;
read_sysreg_case ( SYS_CTR_EL0 ) ;
read_sysreg_case ( SYS_DCZID_EL0 ) ;
2016-04-22 14:25:31 +03:00
default :
BUG ( ) ;
return 0 ;
}
2021-02-08 12:57:20 +03:00
regp = get_arm64_ftr_reg ( sys_id ) ;
if ( regp ) {
val & = ~ regp - > override - > mask ;
val | = ( regp - > override - > val & regp - > override - > mask ) ;
}
return val ;
2016-04-22 14:25:31 +03:00
}
2015-09-30 13:50:04 +03:00
# include <linux/irqchip/arm-gic-v3.h>
2022-09-12 19:22:08 +03:00
static bool
has_always ( const struct arm64_cpu_capabilities * entry , int scope )
{
return true ;
}
2015-07-21 15:23:29 +03:00
static bool
feature_matches ( u64 reg , const struct arm64_cpu_capabilities * entry )
{
2022-02-07 18:20:32 +03:00
int val = cpuid_feature_extract_field_width ( reg , entry - > field_pos ,
entry - > field_width ,
entry - > sign ) ;
2015-07-21 15:23:29 +03:00
return val > = entry - > min_field_value ;
}
2022-09-09 19:59:36 +03:00
static u64
read_scoped_sysreg ( const struct arm64_cpu_capabilities * entry , int scope )
2015-10-19 16:24:51 +03:00
{
2016-04-22 14:25:31 +03:00
WARN_ON ( scope = = SCOPE_LOCAL_CPU & & preemptible ( ) ) ;
if ( scope = = SCOPE_SYSTEM )
2022-09-09 19:59:36 +03:00
return read_sanitised_ftr_reg ( entry - > sys_reg ) ;
2016-04-22 14:25:31 +03:00
else
2022-09-09 19:59:36 +03:00
return __read_sysreg_by_encoding ( entry - > sys_reg ) ;
}
2016-04-22 14:25:31 +03:00
2022-09-09 19:59:36 +03:00
static bool
has_user_cpuid_feature ( const struct arm64_cpu_capabilities * entry , int scope )
{
int mask ;
struct arm64_ftr_reg * regp ;
u64 val = read_scoped_sysreg ( entry , scope ) ;
regp = get_arm64_ftr_reg ( entry - > sys_reg ) ;
if ( ! regp )
return false ;
mask = cpuid_feature_extract_unsigned_field_width ( regp - > user_mask ,
entry - > field_pos ,
entry - > field_width ) ;
if ( ! mask )
return false ;
return feature_matches ( val , entry ) ;
}
2016-04-22 14:25:31 +03:00
2022-09-09 19:59:36 +03:00
static bool
has_cpuid_feature ( const struct arm64_cpu_capabilities * entry , int scope )
{
u64 val = read_scoped_sysreg ( entry , scope ) ;
2015-10-19 16:24:51 +03:00
return feature_matches ( val , entry ) ;
}
2015-07-22 21:05:54 +03:00
2021-06-08 21:02:55 +03:00
const struct cpumask * system_32bit_el0_cpumask ( void )
{
if ( ! system_supports_32bit_el0 ( ) )
return cpu_none_mask ;
if ( static_branch_unlikely ( & arm64_mismatched_32bit_el0 ) )
return cpu_32bit_el0_mask ;
return cpu_possible_mask ;
}
2021-07-30 14:24:41 +03:00
static int __init parse_32bit_el0_param ( char * str )
{
allow_mismatched_32bit_el0 = true ;
return 0 ;
}
early_param ( " allow_mismatched_32bit_el0 " , parse_32bit_el0_param ) ;
2021-07-30 14:24:40 +03:00
static ssize_t aarch32_el0_show ( struct device * dev ,
struct device_attribute * attr , char * buf )
{
const struct cpumask * mask = system_32bit_el0_cpumask ( ) ;
return sysfs_emit ( buf , " %*pbl \n " , cpumask_pr_args ( mask ) ) ;
}
static const DEVICE_ATTR_RO ( aarch32_el0 ) ;
static int __init aarch32_el0_sysfs_init ( void )
{
if ( ! allow_mismatched_32bit_el0 )
return 0 ;
return device_create_file ( cpu_subsys . dev_root , & dev_attr_aarch32_el0 ) ;
}
device_initcall ( aarch32_el0_sysfs_init ) ;
2021-06-08 21:02:55 +03:00
static bool has_32bit_el0 ( const struct arm64_cpu_capabilities * entry , int scope )
{
if ( ! has_cpuid_feature ( entry , scope ) )
return allow_mismatched_32bit_el0 ;
if ( scope = = SCOPE_SYSTEM )
pr_info ( " detected: 32-bit EL0 Support \n " ) ;
return true ;
}
2016-04-22 14:25:31 +03:00
static bool has_useable_gicv3_cpuif ( const struct arm64_cpu_capabilities * entry , int scope )
2015-09-30 13:50:04 +03:00
{
bool has_sre ;
2016-04-22 14:25:31 +03:00
if ( ! has_cpuid_feature ( entry , scope ) )
2015-09-30 13:50:04 +03:00
return false ;
has_sre = gic_enable_sre ( ) ;
if ( ! has_sre )
pr_warn_once ( " %s present but disabled by higher exception level \n " ,
entry - > desc ) ;
return has_sre ;
}
2016-04-22 14:25:31 +03:00
static bool has_no_hw_prefetch ( const struct arm64_cpu_capabilities * entry , int __unused )
2016-02-02 15:46:24 +03:00
{
u32 midr = read_cpuid_id ( ) ;
/* Cavium ThunderX pass 1.x and 2.x */
2019-08-06 06:05:03 +03:00
return midr_is_cpu_model_range ( midr , MIDR_THUNDERX ,
2017-01-13 16:12:09 +03:00
MIDR_CPU_VAR_REV ( 0 , 0 ) ,
MIDR_CPU_VAR_REV ( 1 , MIDR_REVISION_MASK ) ) ;
2016-02-02 15:46:24 +03:00
}
2016-11-08 16:56:21 +03:00
static bool has_no_fpsimd ( const struct arm64_cpu_capabilities * entry , int __unused )
{
2017-03-23 18:14:39 +03:00
u64 pfr0 = read_sanitised_ftr_reg ( SYS_ID_AA64PFR0_EL1 ) ;
2016-11-08 16:56:21 +03:00
return cpuid_feature_extract_signed_field ( pfr0 ,
2022-09-06 01:54:03 +03:00
ID_AA64PFR0_EL1_FP_SHIFT ) < 0 ;
2016-11-08 16:56:21 +03:00
}
2018-03-07 18:00:08 +03:00
static bool has_cache_idc ( const struct arm64_cpu_capabilities * entry ,
2018-10-09 16:47:05 +03:00
int scope )
2018-03-07 18:00:08 +03:00
{
2018-10-09 16:47:05 +03:00
u64 ctr ;
if ( scope = = SCOPE_SYSTEM )
ctr = arm64_ftr_reg_ctrel0 . sys_val ;
else
2018-10-09 16:47:06 +03:00
ctr = read_cpuid_effective_cachetype ( ) ;
2018-10-09 16:47:05 +03:00
2022-07-04 20:02:40 +03:00
return ctr & BIT ( CTR_EL0_IDC_SHIFT ) ;
2018-03-07 18:00:08 +03:00
}
2018-10-09 16:47:06 +03:00
static void cpu_emulate_effective_ctr ( const struct arm64_cpu_capabilities * __unused )
{
/*
* If the CPU exposes raw CTR_EL0 . IDC = 0 , while effectively
* CTR_EL0 . IDC = 1 ( from CLIDR values ) , we need to trap accesses
* to the CTR_EL0 on this CPU and emulate it with the real / safe
* value .
*/
2022-07-04 20:02:40 +03:00
if ( ! ( read_cpuid_cachetype ( ) & BIT ( CTR_EL0_IDC_SHIFT ) ) )
2018-10-09 16:47:06 +03:00
sysreg_clear_set ( sctlr_el1 , SCTLR_EL1_UCT , 0 ) ;
}
2018-03-07 18:00:08 +03:00
static bool has_cache_dic ( const struct arm64_cpu_capabilities * entry ,
2018-10-09 16:47:05 +03:00
int scope )
2018-03-07 18:00:08 +03:00
{
2018-10-09 16:47:05 +03:00
u64 ctr ;
if ( scope = = SCOPE_SYSTEM )
ctr = arm64_ftr_reg_ctrel0 . sys_val ;
else
ctr = read_cpuid_cachetype ( ) ;
2022-07-04 20:02:40 +03:00
return ctr & BIT ( CTR_EL0_DIC_SHIFT ) ;
2018-03-07 18:00:08 +03:00
}
2018-07-31 16:08:56 +03:00
static bool __maybe_unused
has_useable_cnp ( const struct arm64_cpu_capabilities * entry , int scope )
{
/*
* Kdump isn ' t guaranteed to power - off all secondary CPUs , CNP
* may share TLB entries with a CPU stuck in the crashed
* kernel .
*/
2021-03-24 03:28:09 +03:00
if ( is_kdump_kernel ( ) )
return false ;
if ( cpus_have_const_cap ( ARM64_WORKAROUND_NVIDIA_CARMEL_CNP ) )
2018-07-31 16:08:56 +03:00
return false ;
return has_cpuid_feature ( entry , scope ) ;
}
2019-12-09 21:12:17 +03:00
/*
* This check is triggered during the early boot before the cpufeature
* is initialised . Checking the status on the local CPU allows the boot
* CPU to detect the need for non - global mappings and thus avoiding a
* pagetable re - write after all the CPUs are booted . This check will be
* anyway run on individual CPUs , allowing us to get the consistent
* state once the SMP CPUs are up and thus make the switch to non - global
* mappings if required .
*/
bool kaslr_requires_kpti ( void )
{
if ( ! IS_ENABLED ( CONFIG_RANDOMIZE_BASE ) )
return false ;
/*
* E0PD does a similar job to KPTI so can be used instead
* where available .
*/
if ( IS_ENABLED ( CONFIG_ARM64_E0PD ) ) {
2020-01-15 17:06:37 +03:00
u64 mmfr2 = read_sysreg_s ( SYS_ID_AA64MMFR2_EL1 ) ;
if ( cpuid_feature_extract_unsigned_field ( mmfr2 ,
2022-09-06 01:54:02 +03:00
ID_AA64MMFR2_EL1_E0PD_SHIFT ) )
2019-12-09 21:12:17 +03:00
return false ;
}
/*
* Systems affected by Cavium erratum 24756 are incompatible
* with KPTI .
*/
2020-01-15 16:59:58 +03:00
if ( IS_ENABLED ( CONFIG_CAVIUM_ERRATUM_27456 ) ) {
2019-12-09 21:12:17 +03:00
extern const struct midr_range cavium_erratum_27456_cpus [ ] ;
2020-01-15 16:59:58 +03:00
if ( is_midr_in_range_list ( read_cpuid_id ( ) ,
cavium_erratum_27456_cpus ) )
return false ;
2019-12-09 21:12:17 +03:00
}
return kaslr_offset ( ) > 0 ;
}
2019-04-16 00:21:22 +03:00
static bool __meltdown_safe = true ;
2017-11-14 17:38:19 +03:00
static int __kpti_forced ; /* 0: not forced, >0: forced on, <0: forced off */
static bool unmap_kernel_at_el0 ( const struct arm64_cpu_capabilities * entry ,
2018-03-26 17:12:40 +03:00
int scope )
2017-11-14 17:38:19 +03:00
{
2018-03-26 17:12:45 +03:00
/* List of CPUs that are not vulnerable and don't need KPTI */
static const struct midr_range kpti_safe_list [ ] = {
MIDR_ALL_VERSIONS ( MIDR_CAVIUM_THUNDERX2 ) ,
MIDR_ALL_VERSIONS ( MIDR_BRCM_VULCAN ) ,
2020-01-07 01:54:12 +03:00
MIDR_ALL_VERSIONS ( MIDR_BRAHMA_B53 ) ,
2018-12-13 16:47:38 +03:00
MIDR_ALL_VERSIONS ( MIDR_CORTEX_A35 ) ,
MIDR_ALL_VERSIONS ( MIDR_CORTEX_A53 ) ,
MIDR_ALL_VERSIONS ( MIDR_CORTEX_A55 ) ,
MIDR_ALL_VERSIONS ( MIDR_CORTEX_A57 ) ,
MIDR_ALL_VERSIONS ( MIDR_CORTEX_A72 ) ,
MIDR_ALL_VERSIONS ( MIDR_CORTEX_A73 ) ,
2019-03-05 16:40:58 +03:00
MIDR_ALL_VERSIONS ( MIDR_HISI_TSV110 ) ,
2019-11-05 21:45:10 +03:00
MIDR_ALL_VERSIONS ( MIDR_NVIDIA_CARMEL ) ,
2020-11-05 02:22:11 +03:00
MIDR_ALL_VERSIONS ( MIDR_QCOM_KRYO_2XX_GOLD ) ,
MIDR_ALL_VERSIONS ( MIDR_QCOM_KRYO_2XX_SILVER ) ,
2020-06-24 15:34:06 +03:00
MIDR_ALL_VERSIONS ( MIDR_QCOM_KRYO_3XX_SILVER ) ,
MIDR_ALL_VERSIONS ( MIDR_QCOM_KRYO_4XX_SILVER ) ,
2018-04-23 13:41:33 +03:00
{ /* sentinel */ }
2018-03-26 17:12:45 +03:00
} ;
2019-04-12 23:39:32 +03:00
char const * str = " kpti command line option " ;
2019-04-16 00:21:22 +03:00
bool meltdown_safe ;
meltdown_safe = is_midr_in_range_list ( read_cpuid_id ( ) , kpti_safe_list ) ;
/* Defer to CPU feature registers */
if ( has_cpuid_feature ( entry , scope ) )
meltdown_safe = true ;
if ( ! meltdown_safe )
__meltdown_safe = false ;
2017-11-27 21:29:30 +03:00
2018-01-29 14:59:56 +03:00
/*
* For reasons that aren ' t entirely clear , enabling KPTI on Cavium
* ThunderX leads to apparent I - cache corruption of kernel text , which
2021-09-23 17:50:02 +03:00
* ends as well as you might imagine . Don ' t even try . We cannot rely
* on the cpus_have_ * cap ( ) helpers here to detect the CPU erratum
* because cpucap detection order may change . However , since we know
* affected CPUs are always in a homogeneous configuration , it is
* safe to rely on this_cpu_has_cap ( ) here .
2018-01-29 14:59:56 +03:00
*/
2021-09-23 17:50:02 +03:00
if ( this_cpu_has_cap ( ARM64_WORKAROUND_CAVIUM_27456 ) ) {
2018-01-29 14:59:56 +03:00
str = " ARM64_WORKAROUND_CAVIUM_27456 " ;
__kpti_forced = - 1 ;
}
2019-04-16 00:21:22 +03:00
/* Useful for KASLR robustness */
2019-12-09 21:12:15 +03:00
if ( kaslr_requires_kpti ( ) ) {
2019-04-16 00:21:22 +03:00
if ( ! __kpti_forced ) {
str = " KASLR " ;
__kpti_forced = 1 ;
}
}
2019-04-12 23:39:32 +03:00
if ( cpu_mitigations_off ( ) & & ! __kpti_forced ) {
str = " mitigations=off " ;
__kpti_forced = - 1 ;
}
2019-04-16 00:21:22 +03:00
if ( ! IS_ENABLED ( CONFIG_UNMAP_KERNEL_AT_EL0 ) ) {
pr_info_once ( " kernel page table isolation disabled by kernel configuration \n " ) ;
return false ;
}
2018-01-29 14:59:56 +03:00
/* Forced? */
2017-11-14 17:38:19 +03:00
if ( __kpti_forced ) {
2018-01-29 14:59:56 +03:00
pr_info_once ( " kernel page table isolation forced %s by %s \n " ,
__kpti_forced > 0 ? " ON " : " OFF " , str ) ;
2017-11-14 17:38:19 +03:00
return __kpti_forced > 0 ;
}
2019-04-16 00:21:22 +03:00
return ! meltdown_safe ;
2017-11-14 17:38:19 +03:00
}
2019-04-16 00:21:22 +03:00
# ifdef CONFIG_UNMAP_KERNEL_AT_EL0
arm64: mm: install KPTI nG mappings with MMU enabled
In cases where we unmap the kernel while running in user space, we rely
on ASIDs to distinguish the minimal trampoline from the full kernel
mapping, and this means we must use non-global attributes for those
mappings, to ensure they are scoped by ASID and will not hit in the TLB
inadvertently.
We only do this when needed, as this is generally more costly in terms
of TLB pressure, and so we boot without these non-global attributes, and
apply them to all existing kernel mappings once all CPUs are up and we
know whether or not the non-global attributes are needed. At this point,
we cannot simply unmap and remap the entire address space, so we have to
update all existing block and page descriptors in place.
Currently, we go through a lot of trouble to perform these updates with
the MMU and caches off, to avoid violating break before make (BBM) rules
imposed by the architecture. Since we make changes to page tables that
are not covered by the ID map, we gain access to those descriptors by
disabling translations altogether. This means that the stores to memory
are issued with device attributes, and require extra care in terms of
coherency, which is costly. We also rely on the ID map to access a
shared flag, which requires the ID map to be executable and writable at
the same time, which is another thing we'd prefer to avoid.
So let's switch to an approach where we replace the kernel mapping with
a minimal mapping of a few pages that can be used for a minimal, ad-hoc
fixmap that we can use to map each page table in turn as we traverse the
hierarchy.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Link: https://lore.kernel.org/r/20220609174320.4035379-3-ardb@kernel.org
Signed-off-by: Will Deacon <will@kernel.org>
2022-06-09 20:43:20 +03:00
# define KPTI_NG_TEMP_VA (-(1UL << PMD_SHIFT))
extern
void create_kpti_ng_temp_pgd ( pgd_t * pgdir , phys_addr_t phys , unsigned long virt ,
phys_addr_t size , pgprot_t prot ,
phys_addr_t ( * pgtable_alloc ) ( int ) , int flags ) ;
static phys_addr_t kpti_ng_temp_alloc ;
static phys_addr_t kpti_ng_pgd_alloc ( int shift )
{
kpti_ng_temp_alloc - = PAGE_SIZE ;
return kpti_ng_temp_alloc ;
}
2022-09-09 00:54:53 +03:00
static void
2018-03-26 17:12:28 +03:00
kpti_install_ng_mappings ( const struct arm64_cpu_capabilities * __unused )
2018-02-07 01:22:50 +03:00
{
arm64: mm: install KPTI nG mappings with MMU enabled
In cases where we unmap the kernel while running in user space, we rely
on ASIDs to distinguish the minimal trampoline from the full kernel
mapping, and this means we must use non-global attributes for those
mappings, to ensure they are scoped by ASID and will not hit in the TLB
inadvertently.
We only do this when needed, as this is generally more costly in terms
of TLB pressure, and so we boot without these non-global attributes, and
apply them to all existing kernel mappings once all CPUs are up and we
know whether or not the non-global attributes are needed. At this point,
we cannot simply unmap and remap the entire address space, so we have to
update all existing block and page descriptors in place.
Currently, we go through a lot of trouble to perform these updates with
the MMU and caches off, to avoid violating break before make (BBM) rules
imposed by the architecture. Since we make changes to page tables that
are not covered by the ID map, we gain access to those descriptors by
disabling translations altogether. This means that the stores to memory
are issued with device attributes, and require extra care in terms of
coherency, which is costly. We also rely on the ID map to access a
shared flag, which requires the ID map to be executable and writable at
the same time, which is another thing we'd prefer to avoid.
So let's switch to an approach where we replace the kernel mapping with
a minimal mapping of a few pages that can be used for a minimal, ad-hoc
fixmap that we can use to map each page table in turn as we traverse the
hierarchy.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Link: https://lore.kernel.org/r/20220609174320.4035379-3-ardb@kernel.org
Signed-off-by: Will Deacon <will@kernel.org>
2022-06-09 20:43:20 +03:00
typedef void ( kpti_remap_fn ) ( int , int , phys_addr_t , unsigned long ) ;
2018-02-07 01:22:50 +03:00
extern kpti_remap_fn idmap_kpti_install_ng_mappings ;
kpti_remap_fn * remap_fn ;
int cpu = smp_processor_id ( ) ;
arm64: mm: install KPTI nG mappings with MMU enabled
In cases where we unmap the kernel while running in user space, we rely
on ASIDs to distinguish the minimal trampoline from the full kernel
mapping, and this means we must use non-global attributes for those
mappings, to ensure they are scoped by ASID and will not hit in the TLB
inadvertently.
We only do this when needed, as this is generally more costly in terms
of TLB pressure, and so we boot without these non-global attributes, and
apply them to all existing kernel mappings once all CPUs are up and we
know whether or not the non-global attributes are needed. At this point,
we cannot simply unmap and remap the entire address space, so we have to
update all existing block and page descriptors in place.
Currently, we go through a lot of trouble to perform these updates with
the MMU and caches off, to avoid violating break before make (BBM) rules
imposed by the architecture. Since we make changes to page tables that
are not covered by the ID map, we gain access to those descriptors by
disabling translations altogether. This means that the stores to memory
are issued with device attributes, and require extra care in terms of
coherency, which is costly. We also rely on the ID map to access a
shared flag, which requires the ID map to be executable and writable at
the same time, which is another thing we'd prefer to avoid.
So let's switch to an approach where we replace the kernel mapping with
a minimal mapping of a few pages that can be used for a minimal, ad-hoc
fixmap that we can use to map each page table in turn as we traverse the
hierarchy.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Link: https://lore.kernel.org/r/20220609174320.4035379-3-ardb@kernel.org
Signed-off-by: Will Deacon <will@kernel.org>
2022-06-09 20:43:20 +03:00
int levels = CONFIG_PGTABLE_LEVELS ;
int order = order_base_2 ( levels ) ;
u64 kpti_ng_temp_pgd_pa = 0 ;
pgd_t * kpti_ng_temp_pgd ;
u64 alloc = 0 ;
2018-02-07 01:22:50 +03:00
2021-11-23 21:29:25 +03:00
if ( __this_cpu_read ( this_cpu_vector ) = = vectors ) {
const char * v = arm64_get_bp_hardening_vector ( EL1_VECTOR_KPTI ) ;
__this_cpu_write ( this_cpu_vector , v ) ;
}
2019-01-08 19:19:01 +03:00
/*
* We don ' t need to rewrite the page - tables if either we ' ve done
* it already or we have KASLR enabled and therefore have not
* created any global mappings at all .
*/
2019-12-09 21:12:17 +03:00
if ( arm64_use_ng_mappings )
2018-03-26 17:12:28 +03:00
return ;
2018-02-07 01:22:50 +03:00
2022-09-09 00:54:55 +03:00
remap_fn = ( void * ) __pa_symbol ( idmap_kpti_install_ng_mappings ) ;
2018-02-07 01:22:50 +03:00
arm64: mm: install KPTI nG mappings with MMU enabled
In cases where we unmap the kernel while running in user space, we rely
on ASIDs to distinguish the minimal trampoline from the full kernel
mapping, and this means we must use non-global attributes for those
mappings, to ensure they are scoped by ASID and will not hit in the TLB
inadvertently.
We only do this when needed, as this is generally more costly in terms
of TLB pressure, and so we boot without these non-global attributes, and
apply them to all existing kernel mappings once all CPUs are up and we
know whether or not the non-global attributes are needed. At this point,
we cannot simply unmap and remap the entire address space, so we have to
update all existing block and page descriptors in place.
Currently, we go through a lot of trouble to perform these updates with
the MMU and caches off, to avoid violating break before make (BBM) rules
imposed by the architecture. Since we make changes to page tables that
are not covered by the ID map, we gain access to those descriptors by
disabling translations altogether. This means that the stores to memory
are issued with device attributes, and require extra care in terms of
coherency, which is costly. We also rely on the ID map to access a
shared flag, which requires the ID map to be executable and writable at
the same time, which is another thing we'd prefer to avoid.
So let's switch to an approach where we replace the kernel mapping with
a minimal mapping of a few pages that can be used for a minimal, ad-hoc
fixmap that we can use to map each page table in turn as we traverse the
hierarchy.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Link: https://lore.kernel.org/r/20220609174320.4035379-3-ardb@kernel.org
Signed-off-by: Will Deacon <will@kernel.org>
2022-06-09 20:43:20 +03:00
if ( ! cpu ) {
alloc = __get_free_pages ( GFP_ATOMIC | __GFP_ZERO , order ) ;
kpti_ng_temp_pgd = ( pgd_t * ) ( alloc + ( levels - 1 ) * PAGE_SIZE ) ;
kpti_ng_temp_alloc = kpti_ng_temp_pgd_pa = __pa ( kpti_ng_temp_pgd ) ;
//
// Create a minimal page table hierarchy that permits us to map
// the swapper page tables temporarily as we traverse them.
//
// The physical pages are laid out as follows:
//
// +--------+-/-------+-/------ +-\\--------+
// : PTE[] : | PMD[] : | PUD[] : || PGD[] :
// +--------+-\-------+-\------ +-//--------+
// ^
// The first page is mapped into this hierarchy at a PMD_SHIFT
// aligned virtual address, so that we can manipulate the PTE
// level entries while the mapping is active. The first entry
// covers the PTE[] page itself, the remaining entries are free
// to be used as a ad-hoc fixmap.
//
create_kpti_ng_temp_pgd ( kpti_ng_temp_pgd , __pa ( alloc ) ,
KPTI_NG_TEMP_VA , PAGE_SIZE , PAGE_KERNEL ,
kpti_ng_pgd_alloc , 0 ) ;
}
2018-02-07 01:22:50 +03:00
cpu_install_idmap ( ) ;
arm64: mm: install KPTI nG mappings with MMU enabled
In cases where we unmap the kernel while running in user space, we rely
on ASIDs to distinguish the minimal trampoline from the full kernel
mapping, and this means we must use non-global attributes for those
mappings, to ensure they are scoped by ASID and will not hit in the TLB
inadvertently.
We only do this when needed, as this is generally more costly in terms
of TLB pressure, and so we boot without these non-global attributes, and
apply them to all existing kernel mappings once all CPUs are up and we
know whether or not the non-global attributes are needed. At this point,
we cannot simply unmap and remap the entire address space, so we have to
update all existing block and page descriptors in place.
Currently, we go through a lot of trouble to perform these updates with
the MMU and caches off, to avoid violating break before make (BBM) rules
imposed by the architecture. Since we make changes to page tables that
are not covered by the ID map, we gain access to those descriptors by
disabling translations altogether. This means that the stores to memory
are issued with device attributes, and require extra care in terms of
coherency, which is costly. We also rely on the ID map to access a
shared flag, which requires the ID map to be executable and writable at
the same time, which is another thing we'd prefer to avoid.
So let's switch to an approach where we replace the kernel mapping with
a minimal mapping of a few pages that can be used for a minimal, ad-hoc
fixmap that we can use to map each page table in turn as we traverse the
hierarchy.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Link: https://lore.kernel.org/r/20220609174320.4035379-3-ardb@kernel.org
Signed-off-by: Will Deacon <will@kernel.org>
2022-06-09 20:43:20 +03:00
remap_fn ( cpu , num_online_cpus ( ) , kpti_ng_temp_pgd_pa , KPTI_NG_TEMP_VA ) ;
2018-02-07 01:22:50 +03:00
cpu_uninstall_idmap ( ) ;
arm64: mm: install KPTI nG mappings with MMU enabled
In cases where we unmap the kernel while running in user space, we rely
on ASIDs to distinguish the minimal trampoline from the full kernel
mapping, and this means we must use non-global attributes for those
mappings, to ensure they are scoped by ASID and will not hit in the TLB
inadvertently.
We only do this when needed, as this is generally more costly in terms
of TLB pressure, and so we boot without these non-global attributes, and
apply them to all existing kernel mappings once all CPUs are up and we
know whether or not the non-global attributes are needed. At this point,
we cannot simply unmap and remap the entire address space, so we have to
update all existing block and page descriptors in place.
Currently, we go through a lot of trouble to perform these updates with
the MMU and caches off, to avoid violating break before make (BBM) rules
imposed by the architecture. Since we make changes to page tables that
are not covered by the ID map, we gain access to those descriptors by
disabling translations altogether. This means that the stores to memory
are issued with device attributes, and require extra care in terms of
coherency, which is costly. We also rely on the ID map to access a
shared flag, which requires the ID map to be executable and writable at
the same time, which is another thing we'd prefer to avoid.
So let's switch to an approach where we replace the kernel mapping with
a minimal mapping of a few pages that can be used for a minimal, ad-hoc
fixmap that we can use to map each page table in turn as we traverse the
hierarchy.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Link: https://lore.kernel.org/r/20220609174320.4035379-3-ardb@kernel.org
Signed-off-by: Will Deacon <will@kernel.org>
2022-06-09 20:43:20 +03:00
if ( ! cpu ) {
free_pages ( alloc , order ) ;
2019-12-09 21:12:17 +03:00
arm64_use_ng_mappings = true ;
arm64: mm: install KPTI nG mappings with MMU enabled
In cases where we unmap the kernel while running in user space, we rely
on ASIDs to distinguish the minimal trampoline from the full kernel
mapping, and this means we must use non-global attributes for those
mappings, to ensure they are scoped by ASID and will not hit in the TLB
inadvertently.
We only do this when needed, as this is generally more costly in terms
of TLB pressure, and so we boot without these non-global attributes, and
apply them to all existing kernel mappings once all CPUs are up and we
know whether or not the non-global attributes are needed. At this point,
we cannot simply unmap and remap the entire address space, so we have to
update all existing block and page descriptors in place.
Currently, we go through a lot of trouble to perform these updates with
the MMU and caches off, to avoid violating break before make (BBM) rules
imposed by the architecture. Since we make changes to page tables that
are not covered by the ID map, we gain access to those descriptors by
disabling translations altogether. This means that the stores to memory
are issued with device attributes, and require extra care in terms of
coherency, which is costly. We also rely on the ID map to access a
shared flag, which requires the ID map to be executable and writable at
the same time, which is another thing we'd prefer to avoid.
So let's switch to an approach where we replace the kernel mapping with
a minimal mapping of a few pages that can be used for a minimal, ad-hoc
fixmap that we can use to map each page table in turn as we traverse the
hierarchy.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Link: https://lore.kernel.org/r/20220609174320.4035379-3-ardb@kernel.org
Signed-off-by: Will Deacon <will@kernel.org>
2022-06-09 20:43:20 +03:00
}
2018-02-07 01:22:50 +03:00
}
2019-04-16 00:21:22 +03:00
# else
static void
kpti_install_ng_mappings ( const struct arm64_cpu_capabilities * __unused )
{
}
# endif /* CONFIG_UNMAP_KERNEL_AT_EL0 */
2018-02-07 01:22:50 +03:00
2017-11-14 17:38:19 +03:00
static int __init parse_kpti ( char * str )
{
bool enabled ;
int ret = strtobool ( str , & enabled ) ;
if ( ret )
return ret ;
__kpti_forced = enabled ? 1 : - 1 ;
return 0 ;
}
2018-06-22 12:25:25 +03:00
early_param ( " kpti " , parse_kpti ) ;
2017-11-14 17:38:19 +03:00
2018-03-26 17:12:48 +03:00
# ifdef CONFIG_ARM64_HW_AFDBM
static inline void __cpu_enable_hw_dbm ( void )
{
u64 tcr = read_sysreg ( tcr_el1 ) | TCR_HD ;
write_sysreg ( tcr , tcr_el1 ) ;
isb ( ) ;
2020-10-01 11:48:21 +03:00
local_flush_tlb_all ( ) ;
2018-03-26 17:12:48 +03:00
}
2018-03-26 17:12:49 +03:00
static bool cpu_has_broken_dbm ( void )
{
/* List of CPUs which have broken DBM support. */
static const struct midr_range cpus [ ] = {
# ifdef CONFIG_ARM64_ERRATUM_1024718
2021-02-04 02:00:57 +03:00
MIDR_ALL_VERSIONS ( MIDR_CORTEX_A55 ) ,
2020-06-30 21:00:55 +03:00
/* Kryo4xx Silver (rdpe => r1p0) */
MIDR_REV ( MIDR_QCOM_KRYO_4XX_SILVER , 0xd , 0xe ) ,
2022-01-25 18:40:40 +03:00
# endif
# ifdef CONFIG_ARM64_ERRATUM_2051678
MIDR_REV_RANGE ( MIDR_CORTEX_A510 , 0 , 0 , 2 ) ,
2018-03-26 17:12:49 +03:00
# endif
{ } ,
} ;
return is_midr_in_range_list ( read_cpuid_id ( ) , cpus ) ;
}
2018-03-26 17:12:48 +03:00
static bool cpu_can_use_dbm ( const struct arm64_cpu_capabilities * cap )
{
2018-03-26 17:12:49 +03:00
return has_cpuid_feature ( cap , SCOPE_LOCAL_CPU ) & &
! cpu_has_broken_dbm ( ) ;
2018-03-26 17:12:48 +03:00
}
static void cpu_enable_hw_dbm ( struct arm64_cpu_capabilities const * cap )
{
if ( cpu_can_use_dbm ( cap ) )
__cpu_enable_hw_dbm ( ) ;
}
static bool has_hw_dbm ( const struct arm64_cpu_capabilities * cap ,
int __unused )
{
static bool detected = false ;
/*
* DBM is a non - conflicting feature . i . e , the kernel can safely
* run a mix of CPUs with and without the feature . So , we
* unconditionally enable the capability to allow any late CPU
* to use the feature . We only enable the control bits on the
* CPU , if it actually supports .
*
* We have to make sure we print the " feature " detection only
* when at least one CPU actually uses it . So check if this CPU
* can actually use it and print the message exactly once .
*
* This is safe as all CPUs ( including secondary CPUs - due to the
* LOCAL_CPU scope - and the hotplugged CPUs - via verification )
* goes through the " matches " check exactly once . Also if a CPU
* matches the criteria , it is guaranteed that the CPU will turn
* the DBM on , as the capability is unconditionally enabled .
*/
if ( ! detected & & cpu_can_use_dbm ( cap ) ) {
detected = true ;
pr_info ( " detected: Hardware dirty bit management \n " ) ;
}
return true ;
}
# endif
2020-03-05 12:06:21 +03:00
# ifdef CONFIG_ARM64_AMU_EXTN
/*
* The " amu_cpus " cpumask only signals that the CPU implementation for the
* flagged CPUs supports the Activity Monitors Unit ( AMU ) but does not provide
* information regarding all the events that it supports . When a CPU bit is
* set in the cpumask , the user of this feature can only rely on the presence
* of the 4 fixed counters for that CPU . But this does not guarantee that the
* counters are enabled or access to these counters is enabled by code
* executed at higher exception levels ( firmware ) .
*/
static struct cpumask amu_cpus __read_mostly ;
bool cpu_has_amu_feat ( int cpu )
{
return cpumask_test_cpu ( cpu , & amu_cpus ) ;
}
2020-11-06 15:53:34 +03:00
int get_cpu_with_amu_feat ( void )
{
return cpumask_any ( & amu_cpus ) ;
}
arm64: use activity monitors for frequency invariance
The Frequency Invariance Engine (FIE) is providing a frequency
scaling correction factor that helps achieve more accurate
load-tracking.
So far, for arm and arm64 platforms, this scale factor has been
obtained based on the ratio between the current frequency and the
maximum supported frequency recorded by the cpufreq policy. The
setting of this scale factor is triggered from cpufreq drivers by
calling arch_set_freq_scale. The current frequency used in computation
is the frequency requested by a governor, but it may not be the
frequency that was implemented by the platform.
This correction factor can also be obtained using a core counter and a
constant counter to get information on the performance (frequency based
only) obtained in a period of time. This will more accurately reflect
the actual current frequency of the CPU, compared with the alternative
implementation that reflects the request of a performance level from
the OS.
Therefore, implement arch_scale_freq_tick to use activity monitors, if
present, for the computation of the frequency scale factor.
The use of AMU counters depends on:
- CONFIG_ARM64_AMU_EXTN - depents on the AMU extension being present
- CONFIG_CPU_FREQ - the current frequency obtained using counter
information is divided by the maximum frequency obtained from the
cpufreq policy.
While it is possible to have a combination of CPUs in the system with
and without support for activity monitors, the use of counters for
frequency invariance is only enabled for a CPU if all related CPUs
(CPUs in the same frequency domain) support and have enabled the core
and constant activity monitor counters. In this way, there is a clear
separation between the policies for which arch_set_freq_scale (cpufreq
based FIE) is used, and the policies for which arch_scale_freq_tick
(counter based FIE) is used to set the frequency scale factor. For
this purpose, a late_initcall_sync is registered to trigger validation
work for policies that will enable or disable the use of AMU counters
for frequency invariance. If CONFIG_CPU_FREQ is not defined, the use
of counters is enabled on all CPUs only if all possible CPUs correctly
support the necessary counters.
Signed-off-by: Ionela Voinescu <ionela.voinescu@arm.com>
Reviewed-by: Lukasz Luba <lukasz.luba@arm.com>
Acked-by: Sudeep Holla <sudeep.holla@arm.com>
Cc: Sudeep Holla <sudeep.holla@arm.com>
Cc: Will Deacon <will@kernel.org>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2020-03-05 12:06:26 +03:00
2020-03-05 12:06:21 +03:00
static void cpu_amu_enable ( struct arm64_cpu_capabilities const * cap )
{
if ( has_cpuid_feature ( cap , SCOPE_LOCAL_CPU ) ) {
pr_info ( " detected CPU%d: Activity Monitors Unit (AMU) \n " ,
smp_processor_id ( ) ) ;
cpumask_set_cpu ( smp_processor_id ( ) , & amu_cpus ) ;
2022-08-19 13:30:50 +03:00
/* 0 reference values signal broken/disabled counters */
if ( ! this_cpu_has_cap ( ARM64_WORKAROUND_2457168 ) )
update_freq_counters_refs ( ) ;
2020-03-05 12:06:21 +03:00
}
}
static bool has_amu ( const struct arm64_cpu_capabilities * cap ,
int __unused )
{
/*
* The AMU extension is a non - conflicting feature : the kernel can
* safely run a mix of CPUs with and without support for the
* activity monitors extension . Therefore , unconditionally enable
* the capability to allow any late CPU to use the feature .
*
* With this feature unconditionally enabled , the cpu_enable
* function will be called for all CPUs that match the criteria ,
* including secondary and hotplugged , marking this feature as
* present on that respective CPU . The enable function will also
* print a detection message .
*/
return true ;
}
2020-11-06 15:53:34 +03:00
# else
int get_cpu_with_amu_feat ( void )
{
return nr_cpu_ids ;
}
2020-03-05 12:06:21 +03:00
# endif
2018-03-27 13:51:12 +03:00
static bool runs_at_el2 ( const struct arm64_cpu_capabilities * entry , int __unused )
{
return is_kernel_in_hyp_mode ( ) ;
}
2018-03-26 17:12:28 +03:00
static void cpu_copy_el2regs ( const struct arm64_cpu_capabilities * __unused )
2018-01-08 18:38:06 +03:00
{
/*
* Copy register values that aren ' t redirected by hardware .
*
* Before code patching , we only set tpidr_el1 , all CPUs need to copy
* this value to tpidr_el2 before we patch the code . Once we ' ve done
* that , freshly - onlined CPUs will set tpidr_el2 , so we don ' t need to
* do anything here .
*/
2019-01-31 17:58:52 +03:00
if ( ! alternative_is_applied ( ARM64_HAS_VIRT_HOST_EXTN ) )
2018-01-08 18:38:06 +03:00
write_sysreg ( read_sysreg ( tpidr_el1 ) , tpidr_el2 ) ;
}
2018-08-07 15:53:41 +03:00
# ifdef CONFIG_ARM64_PAN
static void cpu_enable_pan ( const struct arm64_cpu_capabilities * __unused )
{
/*
* We modify PSTATE . This won ' t work from irq context as the PSTATE
* is discarded once we return from the exception .
*/
WARN_ON_ONCE ( in_interrupt ( ) ) ;
sysreg_clear_set ( sctlr_el1 , SCTLR_EL1_SPAN , 0 ) ;
2020-11-13 15:49:22 +03:00
set_pstate_pan ( 1 ) ;
2018-08-07 15:53:41 +03:00
}
# endif /* CONFIG_ARM64_PAN */
# ifdef CONFIG_ARM64_RAS_EXTN
static void cpu_clear_disr ( const struct arm64_cpu_capabilities * __unused )
{
/* Firmware may have left a deferred SError in this register. */
write_sysreg_s ( 0 , SYS_DISR_EL1 ) ;
}
# endif /* CONFIG_ARM64_RAS_EXTN */
2018-12-07 21:39:24 +03:00
# ifdef CONFIG_ARM64_PTR_AUTH
arm64: cpufeature: Modify address authentication cpufeature to exact
The current address authentication cpufeature levels are set as LOWER_SAFE
which is not compatible with the different configurations added for Armv8.3
ptrauth enhancements as the different levels have different behaviour and
there is no tunable to enable the lower safe versions. This is rectified
by setting those cpufeature type as EXACT.
The current cpufeature framework also does not interfere in the booting of
non-exact secondary cpus but rather marks them as tainted. As a workaround
this is fixed by replacing the generic match handler with a new handler
specific to ptrauth.
After this change, if there is any variation in ptrauth configurations in
secondary cpus from boot cpu then those mismatched cpus are parked in an
infinite loop.
Following ptrauth crash log is observed in Arm fastmodel with simulated
mismatched cpus without this fix,
CPU features: SANITY CHECK: Unexpected variation in SYS_ID_AA64ISAR1_EL1. Boot CPU: 0x11111110211402, CPU4: 0x11111110211102
CPU features: Unsupported CPU feature variation detected.
GICv3: CPU4: found redistributor 100 region 0:0x000000002f180000
CPU4: Booted secondary processor 0x0000000100 [0x410fd0f0]
Unable to handle kernel paging request at virtual address bfff800010dadf3c
Mem abort info:
ESR = 0x86000004
EC = 0x21: IABT (current EL), IL = 32 bits
SET = 0, FnV = 0
EA = 0, S1PTW = 0
[bfff800010dadf3c] address between user and kernel address ranges
Internal error: Oops: 86000004 [#1] PREEMPT SMP
Modules linked in:
CPU: 4 PID: 29 Comm: migration/4 Tainted: G S 5.8.0-rc4-00005-ge658591d66d1-dirty #158
Hardware name: Foundation-v8A (DT)
pstate: 60000089 (nZCv daIf -PAN -UAO BTYPE=--)
pc : 0xbfff800010dadf3c
lr : __schedule+0x2b4/0x5a8
sp : ffff800012043d70
x29: ffff800012043d70 x28: 0080000000000000
x27: ffff800011cbe000 x26: ffff00087ad37580
x25: ffff00087ad37000 x24: ffff800010de7d50
x23: ffff800011674018 x22: 0784800010dae2a8
x21: ffff00087ad37000 x20: ffff00087acb8000
x19: ffff00087f742100 x18: 0000000000000030
x17: 0000000000000000 x16: 0000000000000000
x15: ffff800011ac1000 x14: 00000000000001bd
x13: 0000000000000000 x12: 0000000000000000
x11: 0000000000000000 x10: 71519a147ddfeb82
x9 : 825d5ec0fb246314 x8 : ffff00087ad37dd8
x7 : 0000000000000000 x6 : 00000000fffedb0e
x5 : 00000000ffffffff x4 : 0000000000000000
x3 : 0000000000000028 x2 : ffff80086e11e000
x1 : ffff00087ad37000 x0 : ffff00087acdc600
Call trace:
0xbfff800010dadf3c
schedule+0x78/0x110
schedule_preempt_disabled+0x24/0x40
__kthread_parkme+0x68/0xd0
kthread+0x138/0x160
ret_from_fork+0x10/0x34
Code: bad PC value
After this fix, the mismatched CPU4 is parked as,
CPU features: CPU4: Detected conflict for capability 39 (Address authentication (IMP DEF algorithm)), System: 1, CPU: 0
CPU4: will not boot
CPU4: failed to come online
CPU4: died during early boot
[Suzuki: Introduce new matching function for address authentication]
Suggested-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Signed-off-by: Amit Daniel Kachhap <amit.kachhap@arm.com>
Reviewed-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Link: https://lore.kernel.org/r/20200914083656.21428-5-amit.kachhap@arm.com
Signed-off-by: Will Deacon <will@kernel.org>
2020-09-14 11:36:54 +03:00
static bool has_address_auth_cpucap ( const struct arm64_cpu_capabilities * entry , int scope )
2020-03-13 12:04:49 +03:00
{
arm64: cpufeature: Modify address authentication cpufeature to exact
The current address authentication cpufeature levels are set as LOWER_SAFE
which is not compatible with the different configurations added for Armv8.3
ptrauth enhancements as the different levels have different behaviour and
there is no tunable to enable the lower safe versions. This is rectified
by setting those cpufeature type as EXACT.
The current cpufeature framework also does not interfere in the booting of
non-exact secondary cpus but rather marks them as tainted. As a workaround
this is fixed by replacing the generic match handler with a new handler
specific to ptrauth.
After this change, if there is any variation in ptrauth configurations in
secondary cpus from boot cpu then those mismatched cpus are parked in an
infinite loop.
Following ptrauth crash log is observed in Arm fastmodel with simulated
mismatched cpus without this fix,
CPU features: SANITY CHECK: Unexpected variation in SYS_ID_AA64ISAR1_EL1. Boot CPU: 0x11111110211402, CPU4: 0x11111110211102
CPU features: Unsupported CPU feature variation detected.
GICv3: CPU4: found redistributor 100 region 0:0x000000002f180000
CPU4: Booted secondary processor 0x0000000100 [0x410fd0f0]
Unable to handle kernel paging request at virtual address bfff800010dadf3c
Mem abort info:
ESR = 0x86000004
EC = 0x21: IABT (current EL), IL = 32 bits
SET = 0, FnV = 0
EA = 0, S1PTW = 0
[bfff800010dadf3c] address between user and kernel address ranges
Internal error: Oops: 86000004 [#1] PREEMPT SMP
Modules linked in:
CPU: 4 PID: 29 Comm: migration/4 Tainted: G S 5.8.0-rc4-00005-ge658591d66d1-dirty #158
Hardware name: Foundation-v8A (DT)
pstate: 60000089 (nZCv daIf -PAN -UAO BTYPE=--)
pc : 0xbfff800010dadf3c
lr : __schedule+0x2b4/0x5a8
sp : ffff800012043d70
x29: ffff800012043d70 x28: 0080000000000000
x27: ffff800011cbe000 x26: ffff00087ad37580
x25: ffff00087ad37000 x24: ffff800010de7d50
x23: ffff800011674018 x22: 0784800010dae2a8
x21: ffff00087ad37000 x20: ffff00087acb8000
x19: ffff00087f742100 x18: 0000000000000030
x17: 0000000000000000 x16: 0000000000000000
x15: ffff800011ac1000 x14: 00000000000001bd
x13: 0000000000000000 x12: 0000000000000000
x11: 0000000000000000 x10: 71519a147ddfeb82
x9 : 825d5ec0fb246314 x8 : ffff00087ad37dd8
x7 : 0000000000000000 x6 : 00000000fffedb0e
x5 : 00000000ffffffff x4 : 0000000000000000
x3 : 0000000000000028 x2 : ffff80086e11e000
x1 : ffff00087ad37000 x0 : ffff00087acdc600
Call trace:
0xbfff800010dadf3c
schedule+0x78/0x110
schedule_preempt_disabled+0x24/0x40
__kthread_parkme+0x68/0xd0
kthread+0x138/0x160
ret_from_fork+0x10/0x34
Code: bad PC value
After this fix, the mismatched CPU4 is parked as,
CPU features: CPU4: Detected conflict for capability 39 (Address authentication (IMP DEF algorithm)), System: 1, CPU: 0
CPU4: will not boot
CPU4: failed to come online
CPU4: died during early boot
[Suzuki: Introduce new matching function for address authentication]
Suggested-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Signed-off-by: Amit Daniel Kachhap <amit.kachhap@arm.com>
Reviewed-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Link: https://lore.kernel.org/r/20200914083656.21428-5-amit.kachhap@arm.com
Signed-off-by: Will Deacon <will@kernel.org>
2020-09-14 11:36:54 +03:00
int boot_val , sec_val ;
/* We don't expect to be called with SCOPE_SYSTEM */
WARN_ON ( scope = = SCOPE_SYSTEM ) ;
/*
* The ptr - auth feature levels are not intercompatible with lower
* levels . Hence we must match ptr - auth feature level of the secondary
* CPUs with that of the boot CPU . The level of boot cpu is fetched
* from the sanitised register whereas direct register read is done for
* the secondary CPUs .
* The sanitised feature state is guaranteed to match that of the
* boot CPU as a mismatched secondary CPU is parked before it gets
* a chance to update the state , with the capability .
*/
boot_val = cpuid_feature_extract_field ( read_sanitised_ftr_reg ( entry - > sys_reg ) ,
entry - > field_pos , entry - > sign ) ;
if ( scope & SCOPE_BOOT_CPU )
return boot_val > = entry - > min_field_value ;
/* Now check for the secondary CPUs with SCOPE_LOCAL_CPU scope */
sec_val = cpuid_feature_extract_field ( __read_sysreg_by_encoding ( entry - > sys_reg ) ,
entry - > field_pos , entry - > sign ) ;
2022-02-24 15:49:50 +03:00
return ( sec_val > = entry - > min_field_value ) & & ( sec_val = = boot_val ) ;
arm64: cpufeature: Modify address authentication cpufeature to exact
The current address authentication cpufeature levels are set as LOWER_SAFE
which is not compatible with the different configurations added for Armv8.3
ptrauth enhancements as the different levels have different behaviour and
there is no tunable to enable the lower safe versions. This is rectified
by setting those cpufeature type as EXACT.
The current cpufeature framework also does not interfere in the booting of
non-exact secondary cpus but rather marks them as tainted. As a workaround
this is fixed by replacing the generic match handler with a new handler
specific to ptrauth.
After this change, if there is any variation in ptrauth configurations in
secondary cpus from boot cpu then those mismatched cpus are parked in an
infinite loop.
Following ptrauth crash log is observed in Arm fastmodel with simulated
mismatched cpus without this fix,
CPU features: SANITY CHECK: Unexpected variation in SYS_ID_AA64ISAR1_EL1. Boot CPU: 0x11111110211402, CPU4: 0x11111110211102
CPU features: Unsupported CPU feature variation detected.
GICv3: CPU4: found redistributor 100 region 0:0x000000002f180000
CPU4: Booted secondary processor 0x0000000100 [0x410fd0f0]
Unable to handle kernel paging request at virtual address bfff800010dadf3c
Mem abort info:
ESR = 0x86000004
EC = 0x21: IABT (current EL), IL = 32 bits
SET = 0, FnV = 0
EA = 0, S1PTW = 0
[bfff800010dadf3c] address between user and kernel address ranges
Internal error: Oops: 86000004 [#1] PREEMPT SMP
Modules linked in:
CPU: 4 PID: 29 Comm: migration/4 Tainted: G S 5.8.0-rc4-00005-ge658591d66d1-dirty #158
Hardware name: Foundation-v8A (DT)
pstate: 60000089 (nZCv daIf -PAN -UAO BTYPE=--)
pc : 0xbfff800010dadf3c
lr : __schedule+0x2b4/0x5a8
sp : ffff800012043d70
x29: ffff800012043d70 x28: 0080000000000000
x27: ffff800011cbe000 x26: ffff00087ad37580
x25: ffff00087ad37000 x24: ffff800010de7d50
x23: ffff800011674018 x22: 0784800010dae2a8
x21: ffff00087ad37000 x20: ffff00087acb8000
x19: ffff00087f742100 x18: 0000000000000030
x17: 0000000000000000 x16: 0000000000000000
x15: ffff800011ac1000 x14: 00000000000001bd
x13: 0000000000000000 x12: 0000000000000000
x11: 0000000000000000 x10: 71519a147ddfeb82
x9 : 825d5ec0fb246314 x8 : ffff00087ad37dd8
x7 : 0000000000000000 x6 : 00000000fffedb0e
x5 : 00000000ffffffff x4 : 0000000000000000
x3 : 0000000000000028 x2 : ffff80086e11e000
x1 : ffff00087ad37000 x0 : ffff00087acdc600
Call trace:
0xbfff800010dadf3c
schedule+0x78/0x110
schedule_preempt_disabled+0x24/0x40
__kthread_parkme+0x68/0xd0
kthread+0x138/0x160
ret_from_fork+0x10/0x34
Code: bad PC value
After this fix, the mismatched CPU4 is parked as,
CPU features: CPU4: Detected conflict for capability 39 (Address authentication (IMP DEF algorithm)), System: 1, CPU: 0
CPU4: will not boot
CPU4: failed to come online
CPU4: died during early boot
[Suzuki: Introduce new matching function for address authentication]
Suggested-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Signed-off-by: Amit Daniel Kachhap <amit.kachhap@arm.com>
Reviewed-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Link: https://lore.kernel.org/r/20200914083656.21428-5-amit.kachhap@arm.com
Signed-off-by: Will Deacon <will@kernel.org>
2020-09-14 11:36:54 +03:00
}
static bool has_address_auth_metacap ( const struct arm64_cpu_capabilities * entry ,
int scope )
{
2022-02-24 15:49:51 +03:00
bool api = has_address_auth_cpucap ( cpu_hwcaps_ptrs [ ARM64_HAS_ADDRESS_AUTH_IMP_DEF ] , scope ) ;
bool apa = has_address_auth_cpucap ( cpu_hwcaps_ptrs [ ARM64_HAS_ADDRESS_AUTH_ARCH_QARMA5 ] , scope ) ;
2022-02-24 15:49:52 +03:00
bool apa3 = has_address_auth_cpucap ( cpu_hwcaps_ptrs [ ARM64_HAS_ADDRESS_AUTH_ARCH_QARMA3 ] , scope ) ;
2022-02-24 15:49:51 +03:00
2022-02-24 15:49:52 +03:00
return apa | | apa3 | | api ;
2020-03-13 12:04:49 +03:00
}
static bool has_generic_auth ( const struct arm64_cpu_capabilities * entry ,
int __unused )
arm64: add basic pointer authentication support
This patch adds basic support for pointer authentication, allowing
userspace to make use of APIAKey, APIBKey, APDAKey, APDBKey, and
APGAKey. The kernel maintains key values for each process (shared by all
threads within), which are initialised to random values at exec() time.
The ID_AA64ISAR1_EL1.{APA,API,GPA,GPI} fields are exposed to userspace,
to describe that pointer authentication instructions are available and
that the kernel is managing the keys. Two new hwcaps are added for the
same reason: PACA (for address authentication) and PACG (for generic
authentication).
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Kristina Martsenko <kristina.martsenko@arm.com>
Tested-by: Adam Wallis <awallis@codeaurora.org>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
Cc: Suzuki K Poulose <suzuki.poulose@arm.com>
Cc: Will Deacon <will.deacon@arm.com>
[will: Fix sizeof() usage and unroll address key initialisation]
Signed-off-by: Will Deacon <will.deacon@arm.com>
2018-12-07 21:39:25 +03:00
{
2022-02-24 15:49:51 +03:00
bool gpi = __system_matches_cap ( ARM64_HAS_GENERIC_AUTH_IMP_DEF ) ;
bool gpa = __system_matches_cap ( ARM64_HAS_GENERIC_AUTH_ARCH_QARMA5 ) ;
2022-02-24 15:49:52 +03:00
bool gpa3 = __system_matches_cap ( ARM64_HAS_GENERIC_AUTH_ARCH_QARMA3 ) ;
2022-02-24 15:49:51 +03:00
2022-02-24 15:49:52 +03:00
return gpa | | gpa3 | | gpi ;
arm64: add basic pointer authentication support
This patch adds basic support for pointer authentication, allowing
userspace to make use of APIAKey, APIBKey, APDAKey, APDBKey, and
APGAKey. The kernel maintains key values for each process (shared by all
threads within), which are initialised to random values at exec() time.
The ID_AA64ISAR1_EL1.{APA,API,GPA,GPI} fields are exposed to userspace,
to describe that pointer authentication instructions are available and
that the kernel is managing the keys. Two new hwcaps are added for the
same reason: PACA (for address authentication) and PACG (for generic
authentication).
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Kristina Martsenko <kristina.martsenko@arm.com>
Tested-by: Adam Wallis <awallis@codeaurora.org>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
Cc: Suzuki K Poulose <suzuki.poulose@arm.com>
Cc: Will Deacon <will.deacon@arm.com>
[will: Fix sizeof() usage and unroll address key initialisation]
Signed-off-by: Will Deacon <will.deacon@arm.com>
2018-12-07 21:39:25 +03:00
}
2018-12-07 21:39:24 +03:00
# endif /* CONFIG_ARM64_PTR_AUTH */
2019-12-09 21:12:14 +03:00
# ifdef CONFIG_ARM64_E0PD
static void cpu_enable_e0pd ( struct arm64_cpu_capabilities const * cap )
{
if ( this_cpu_has_cap ( ARM64_HAS_E0PD ) )
sysreg_clear_set ( tcr_el1 , 0 , TCR_E0PD1 ) ;
}
# endif /* CONFIG_ARM64_E0PD */
2019-01-31 17:58:42 +03:00
# ifdef CONFIG_ARM64_PSEUDO_NMI
2019-01-31 17:59:03 +03:00
static bool enable_pseudo_nmi ;
static int __init early_enable_pseudo_nmi ( char * p )
{
return strtobool ( p , & enable_pseudo_nmi ) ;
}
early_param ( " irqchip.gicv3_pseudo_nmi " , early_enable_pseudo_nmi ) ;
2019-01-31 17:58:42 +03:00
static bool can_use_gic_priorities ( const struct arm64_cpu_capabilities * entry ,
int scope )
{
2019-01-31 17:59:03 +03:00
return enable_pseudo_nmi & & has_useable_gicv3_cpuif ( entry , scope ) ;
2019-01-31 17:58:42 +03:00
}
# endif
2020-03-16 19:50:45 +03:00
# ifdef CONFIG_ARM64_BTI
static void bti_enable ( const struct arm64_cpu_capabilities * __unused )
{
/*
* Use of X16 / X17 for tail - calls and trampolines that jump to
* function entry points using BR is a requirement for
* marking binaries with GNU_PROPERTY_AARCH64_FEATURE_1_BTI .
* So , be strict and forbid other BRs using other registers to
* jump onto a PACIxSP instruction :
*/
sysreg_clear_set ( sctlr_el1 , 0 , SCTLR_EL1_BT0 | SCTLR_EL1_BT1 ) ;
isb ( ) ;
}
# endif /* CONFIG_ARM64_BTI */
2020-05-04 16:42:36 +03:00
# ifdef CONFIG_ARM64_MTE
static void cpu_enable_mte ( struct arm64_cpu_capabilities const * cap )
{
2021-08-03 10:08:22 +03:00
sysreg_clear_set ( sctlr_el1 , 0 , SCTLR_ELx_ATA | SCTLR_EL1_ATA0 ) ;
2022-09-16 01:20:53 +03:00
mte_cpu_setup ( ) ;
2021-08-03 10:08:22 +03:00
2020-05-04 16:42:36 +03:00
/*
* Clear the tags in the zero page . This needs to be done via the
* linear map which has the Tagged attribute .
*/
2021-02-10 21:03:16 +03:00
if ( ! test_and_set_bit ( PG_mte_tagged , & ZERO_PAGE ( 0 ) - > flags ) )
2020-05-04 16:42:36 +03:00
mte_clear_page_tags ( lm_alias ( empty_zero_page ) ) ;
2020-12-22 23:02:10 +03:00
kasan_init_hw_tags_cpu ( ) ;
2020-05-04 16:42:36 +03:00
}
# endif /* CONFIG_ARM64_MTE */
2022-07-14 19:15:23 +03:00
static void elf_hwcap_fixup ( void )
{
# ifdef CONFIG_ARM64_ERRATUM_1742098
if ( cpus_have_const_cap ( ARM64_WORKAROUND_1742098 ) )
compat_elf_hwcap2 & = ~ COMPAT_HWCAP2_AES ;
# endif /* ARM64_ERRATUM_1742098 */
}
2020-12-02 21:40:58 +03:00
# ifdef CONFIG_KVM
static bool is_kvm_protected_mode ( const struct arm64_cpu_capabilities * entry , int __unused )
{
2022-06-09 15:12:20 +03:00
return kvm_get_mode ( ) = = KVM_MODE_PROTECTED ;
2020-12-02 21:40:58 +03:00
}
# endif /* CONFIG_KVM */
2022-06-22 14:54:24 +03:00
static void cpu_trap_el0_impdef ( const struct arm64_cpu_capabilities * __unused )
{
sysreg_clear_set ( sctlr_el1 , 0 , SCTLR_EL1_TIDCP ) ;
}
2020-03-13 12:04:53 +03:00
/* Internal helper functions to match cpu capability type */
static bool
cpucap_late_cpu_optional ( const struct arm64_cpu_capabilities * cap )
{
return ! ! ( cap - > type & ARM64_CPUCAP_OPTIONAL_FOR_LATE_CPU ) ;
}
static bool
cpucap_late_cpu_permitted ( const struct arm64_cpu_capabilities * cap )
{
return ! ! ( cap - > type & ARM64_CPUCAP_PERMITTED_FOR_LATE_CPU ) ;
}
2020-03-13 12:04:54 +03:00
static bool
cpucap_panic_on_conflict ( const struct arm64_cpu_capabilities * cap )
{
return ! ! ( cap - > type & ARM64_CPUCAP_PANIC_ON_CONFLICT ) ;
}
2015-03-27 16:09:23 +03:00
static const struct arm64_cpu_capabilities arm64_features [ ] = {
2022-09-12 19:22:08 +03:00
{
. capability = ARM64_ALWAYS_BOOT ,
. type = ARM64_CPUCAP_BOOT_CPU_FEATURE ,
. matches = has_always ,
} ,
{
. capability = ARM64_ALWAYS_SYSTEM ,
. type = ARM64_CPUCAP_SYSTEM_FEATURE ,
. matches = has_always ,
} ,
2015-06-12 14:06:36 +03:00
{
. desc = " GIC system register CPU interface " ,
. capability = ARM64_HAS_SYSREG_GIC_CPUIF ,
2019-01-31 17:58:41 +03:00
. type = ARM64_CPUCAP_STRICT_BOOT_CPU_FEATURE ,
2015-09-30 13:50:04 +03:00
. matches = has_useable_gicv3_cpuif ,
2015-10-19 16:24:51 +03:00
. sys_reg = SYS_ID_AA64PFR0_EL1 ,
2022-09-06 01:54:03 +03:00
. field_pos = ID_AA64PFR0_EL1_GIC_SHIFT ,
2022-03-02 16:42:25 +03:00
. field_width = 4 ,
2016-01-26 13:58:15 +03:00
. sign = FTR_UNSIGNED ,
2015-07-21 15:23:29 +03:00
. min_field_value = 1 ,
2015-06-12 14:06:36 +03:00
} ,
2021-10-17 15:42:22 +03:00
{
. desc = " Enhanced Counter Virtualization " ,
. capability = ARM64_HAS_ECV ,
. type = ARM64_CPUCAP_SYSTEM_FEATURE ,
. matches = has_cpuid_feature ,
. sys_reg = SYS_ID_AA64MMFR0_EL1 ,
2022-09-06 01:54:01 +03:00
. field_pos = ID_AA64MMFR0_EL1_ECV_SHIFT ,
2022-02-07 18:20:32 +03:00
. field_width = 4 ,
2021-10-17 15:42:22 +03:00
. sign = FTR_UNSIGNED ,
. min_field_value = 1 ,
} ,
2015-07-22 21:05:54 +03:00
# ifdef CONFIG_ARM64_PAN
{
. desc = " Privileged Access Never " ,
. capability = ARM64_HAS_PAN ,
2018-03-26 17:12:32 +03:00
. type = ARM64_CPUCAP_SYSTEM_FEATURE ,
2015-10-19 16:24:51 +03:00
. matches = has_cpuid_feature ,
. sys_reg = SYS_ID_AA64MMFR1_EL1 ,
2022-09-06 01:54:07 +03:00
. field_pos = ID_AA64MMFR1_EL1_PAN_SHIFT ,
2022-02-07 18:20:32 +03:00
. field_width = 4 ,
2016-01-26 13:58:15 +03:00
. sign = FTR_UNSIGNED ,
2015-07-22 21:05:54 +03:00
. min_field_value = 1 ,
2018-03-26 17:12:28 +03:00
. cpu_enable = cpu_enable_pan ,
2015-07-22 21:05:54 +03:00
} ,
# endif /* CONFIG_ARM64_PAN */
2021-03-12 20:38:10 +03:00
# ifdef CONFIG_ARM64_EPAN
{
. desc = " Enhanced Privileged Access Never " ,
. capability = ARM64_HAS_EPAN ,
. type = ARM64_CPUCAP_SYSTEM_FEATURE ,
. matches = has_cpuid_feature ,
. sys_reg = SYS_ID_AA64MMFR1_EL1 ,
2022-09-06 01:54:07 +03:00
. field_pos = ID_AA64MMFR1_EL1_PAN_SHIFT ,
2022-02-07 18:20:32 +03:00
. field_width = 4 ,
2021-03-12 20:38:10 +03:00
. sign = FTR_UNSIGNED ,
. min_field_value = 3 ,
} ,
# endif /* CONFIG_ARM64_EPAN */
2020-01-15 14:30:08 +03:00
# ifdef CONFIG_ARM64_LSE_ATOMICS
2015-07-27 18:23:58 +03:00
{
. desc = " LSE atomic instructions " ,
. capability = ARM64_HAS_LSE_ATOMICS ,
2018-03-26 17:12:32 +03:00
. type = ARM64_CPUCAP_SYSTEM_FEATURE ,
2015-10-19 16:24:51 +03:00
. matches = has_cpuid_feature ,
. sys_reg = SYS_ID_AA64ISAR0_EL1 ,
2022-05-03 20:02:28 +03:00
. field_pos = ID_AA64ISAR0_EL1_ATOMIC_SHIFT ,
2022-02-07 18:20:32 +03:00
. field_width = 4 ,
2016-01-26 13:58:15 +03:00
. sign = FTR_UNSIGNED ,
2015-07-27 18:23:58 +03:00
. min_field_value = 2 ,
} ,
2020-01-15 14:30:08 +03:00
# endif /* CONFIG_ARM64_LSE_ATOMICS */
2016-02-02 15:46:24 +03:00
{
. desc = " Software prefetching using PRFM " ,
. capability = ARM64_HAS_NO_HW_PREFETCH ,
2018-03-26 17:12:39 +03:00
. type = ARM64_CPUCAP_WEAK_LOCAL_CPU_FEATURE ,
2016-02-02 15:46:24 +03:00
. matches = has_no_hw_prefetch ,
} ,
2015-01-29 14:24:05 +03:00
{
. desc = " Virtualization Host Extensions " ,
. capability = ARM64_HAS_VIRT_HOST_EXTN ,
2018-03-26 17:12:42 +03:00
. type = ARM64_CPUCAP_STRICT_BOOT_CPU_FEATURE ,
2015-01-29 14:24:05 +03:00
. matches = runs_at_el2 ,
2018-03-26 17:12:28 +03:00
. cpu_enable = cpu_copy_el2regs ,
2015-01-29 14:24:05 +03:00
} ,
2016-04-18 12:28:36 +03:00
{
2021-06-08 21:02:55 +03:00
. capability = ARM64_HAS_32BIT_EL0_DO_NOT_USE ,
2018-03-26 17:12:32 +03:00
. type = ARM64_CPUCAP_SYSTEM_FEATURE ,
2021-06-08 21:02:55 +03:00
. matches = has_32bit_el0 ,
2016-04-18 12:28:36 +03:00
. sys_reg = SYS_ID_AA64PFR0_EL1 ,
. sign = FTR_UNSIGNED ,
2022-09-06 01:54:03 +03:00
. field_pos = ID_AA64PFR0_EL1_EL0_SHIFT ,
2022-02-07 18:20:32 +03:00
. field_width = 4 ,
2022-09-06 01:54:03 +03:00
. min_field_value = ID_AA64PFR0_EL1_ELx_32BIT_64BIT ,
2016-04-18 12:28:36 +03:00
} ,
2020-04-21 17:29:17 +03:00
# 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 ,
2022-09-06 01:54:03 +03:00
. field_pos = ID_AA64PFR0_EL1_EL1_SHIFT ,
2022-02-07 18:20:32 +03:00
. field_width = 4 ,
2022-09-06 01:54:03 +03:00
. min_field_value = ID_AA64PFR0_EL1_ELx_32BIT_64BIT ,
2020-04-21 17:29:17 +03:00
} ,
2020-12-02 21:40:58 +03:00
{
. desc = " Protected KVM " ,
. capability = ARM64_KVM_PROTECTED_MODE ,
. type = ARM64_CPUCAP_SYSTEM_FEATURE ,
. matches = is_kvm_protected_mode ,
} ,
2020-04-21 17:29:17 +03:00
# endif
2017-11-14 17:38:19 +03:00
{
2017-11-27 21:29:30 +03:00
. desc = " Kernel page table isolation (KPTI) " ,
2017-11-14 17:38:19 +03:00
. capability = ARM64_UNMAP_KERNEL_AT_EL0 ,
2018-03-26 17:12:40 +03:00
. type = ARM64_CPUCAP_BOOT_RESTRICTED_CPU_LOCAL_FEATURE ,
/*
* The ID feature fields below are used to indicate that
* the CPU doesn ' t need KPTI . See unmap_kernel_at_el0 for
* more details .
*/
. sys_reg = SYS_ID_AA64PFR0_EL1 ,
2022-09-06 01:54:03 +03:00
. field_pos = ID_AA64PFR0_EL1_CSV3_SHIFT ,
2022-02-07 18:20:32 +03:00
. field_width = 4 ,
2018-03-26 17:12:40 +03:00
. min_field_value = 1 ,
2017-11-14 17:38:19 +03:00
. matches = unmap_kernel_at_el0 ,
2018-03-26 17:12:28 +03:00
. cpu_enable = kpti_install_ng_mappings ,
2017-11-14 17:38:19 +03:00
} ,
2016-11-08 16:56:21 +03:00
{
/* FP/SIMD is not implemented */
. capability = ARM64_HAS_NO_FPSIMD ,
2020-01-14 02:30:19 +03:00
. type = ARM64_CPUCAP_BOOT_RESTRICTED_CPU_LOCAL_FEATURE ,
2016-11-08 16:56:21 +03:00
. min_field_value = 0 ,
. matches = has_no_fpsimd ,
} ,
2017-07-25 13:55:42 +03:00
# ifdef CONFIG_ARM64_PMEM
{
. desc = " Data cache clean to Point of Persistence " ,
. capability = ARM64_HAS_DCPOP ,
2018-03-26 17:12:32 +03:00
. type = ARM64_CPUCAP_SYSTEM_FEATURE ,
2017-07-25 13:55:42 +03:00
. matches = has_cpuid_feature ,
. sys_reg = SYS_ID_AA64ISAR1_EL1 ,
2022-07-04 20:02:49 +03:00
. field_pos = ID_AA64ISAR1_EL1_DPB_SHIFT ,
2022-02-07 18:20:32 +03:00
. field_width = 4 ,
2017-07-25 13:55:42 +03:00
. min_field_value = 1 ,
} ,
2019-04-09 12:52:45 +03:00
{
. desc = " Data cache clean to Point of Deep Persistence " ,
. capability = ARM64_HAS_DCPODP ,
. type = ARM64_CPUCAP_SYSTEM_FEATURE ,
. matches = has_cpuid_feature ,
. sys_reg = SYS_ID_AA64ISAR1_EL1 ,
. sign = FTR_UNSIGNED ,
2022-07-04 20:02:49 +03:00
. field_pos = ID_AA64ISAR1_EL1_DPB_SHIFT ,
2022-02-07 18:20:32 +03:00
. field_width = 4 ,
2019-04-09 12:52:45 +03:00
. min_field_value = 2 ,
} ,
2017-07-25 13:55:42 +03:00
# endif
2017-10-31 18:51:19 +03:00
# ifdef CONFIG_ARM64_SVE
{
. desc = " Scalable Vector Extension " ,
2018-03-26 17:12:32 +03:00
. type = ARM64_CPUCAP_SYSTEM_FEATURE ,
2017-10-31 18:51:19 +03:00
. capability = ARM64_SVE ,
. sys_reg = SYS_ID_AA64PFR0_EL1 ,
. sign = FTR_UNSIGNED ,
2022-09-06 01:54:03 +03:00
. field_pos = ID_AA64PFR0_EL1_SVE_SHIFT ,
2022-02-07 18:20:32 +03:00
. field_width = 4 ,
2022-09-06 01:54:10 +03:00
. min_field_value = ID_AA64PFR0_EL1_SVE_IMP ,
2017-10-31 18:51:19 +03:00
. matches = has_cpuid_feature ,
2018-03-26 17:12:28 +03:00
. cpu_enable = sve_kernel_enable ,
2017-10-31 18:51:19 +03:00
} ,
# endif /* CONFIG_ARM64_SVE */
2018-01-15 22:38:56 +03:00
# ifdef CONFIG_ARM64_RAS_EXTN
{
. desc = " RAS Extension Support " ,
. capability = ARM64_HAS_RAS_EXTN ,
2018-03-26 17:12:32 +03:00
. type = ARM64_CPUCAP_SYSTEM_FEATURE ,
2018-01-15 22:38:56 +03:00
. matches = has_cpuid_feature ,
. sys_reg = SYS_ID_AA64PFR0_EL1 ,
. sign = FTR_UNSIGNED ,
2022-09-06 01:54:03 +03:00
. field_pos = ID_AA64PFR0_EL1_RAS_SHIFT ,
2022-02-07 18:20:32 +03:00
. field_width = 4 ,
2022-09-06 01:54:10 +03:00
. min_field_value = ID_AA64PFR0_EL1_RAS_IMP ,
2018-03-26 17:12:28 +03:00
. cpu_enable = cpu_clear_disr ,
2018-01-15 22:38:56 +03:00
} ,
# endif /* CONFIG_ARM64_RAS_EXTN */
2020-03-05 12:06:21 +03:00
# ifdef CONFIG_ARM64_AMU_EXTN
{
/*
* The feature is enabled by default if CONFIG_ARM64_AMU_EXTN = y .
* Therefore , don ' t provide . desc as we don ' t want the detection
* message to be shown until at least one CPU is detected to
* support the feature .
*/
. capability = ARM64_HAS_AMU_EXTN ,
. type = ARM64_CPUCAP_WEAK_LOCAL_CPU_FEATURE ,
. matches = has_amu ,
. sys_reg = SYS_ID_AA64PFR0_EL1 ,
. sign = FTR_UNSIGNED ,
2022-09-06 01:54:03 +03:00
. field_pos = ID_AA64PFR0_EL1_AMU_SHIFT ,
2022-02-07 18:20:32 +03:00
. field_width = 4 ,
2022-09-06 01:54:10 +03:00
. min_field_value = ID_AA64PFR0_EL1_AMU_IMP ,
2020-03-05 12:06:21 +03:00
. cpu_enable = cpu_amu_enable ,
} ,
# endif /* CONFIG_ARM64_AMU_EXTN */
2018-03-07 18:00:08 +03:00
{
. desc = " Data cache clean to the PoU not required for I/D coherence " ,
. capability = ARM64_HAS_CACHE_IDC ,
2018-03-26 17:12:32 +03:00
. type = ARM64_CPUCAP_SYSTEM_FEATURE ,
2018-03-07 18:00:08 +03:00
. matches = has_cache_idc ,
2018-10-09 16:47:06 +03:00
. cpu_enable = cpu_emulate_effective_ctr ,
2018-03-07 18:00:08 +03:00
} ,
{
. desc = " Instruction cache invalidation not required for I/D coherence " ,
. capability = ARM64_HAS_CACHE_DIC ,
2018-03-26 17:12:32 +03:00
. type = ARM64_CPUCAP_SYSTEM_FEATURE ,
2018-03-07 18:00:08 +03:00
. matches = has_cache_dic ,
} ,
2018-04-06 14:27:28 +03:00
{
. desc = " Stage-2 Force Write-Back " ,
. type = ARM64_CPUCAP_SYSTEM_FEATURE ,
. capability = ARM64_HAS_STAGE2_FWB ,
. sys_reg = SYS_ID_AA64MMFR2_EL1 ,
. sign = FTR_UNSIGNED ,
2022-09-06 01:54:02 +03:00
. field_pos = ID_AA64MMFR2_EL1_FWB_SHIFT ,
2022-02-07 18:20:32 +03:00
. field_width = 4 ,
2018-04-06 14:27:28 +03:00
. min_field_value = 1 ,
. matches = has_cpuid_feature ,
} ,
2018-12-22 15:00:10 +03:00
{
. desc = " ARMv8.4 Translation Table Level " ,
. type = ARM64_CPUCAP_SYSTEM_FEATURE ,
. capability = ARM64_HAS_ARMv8_4_TTL ,
. sys_reg = SYS_ID_AA64MMFR2_EL1 ,
. sign = FTR_UNSIGNED ,
2022-09-06 01:54:02 +03:00
. field_pos = ID_AA64MMFR2_EL1_TTL_SHIFT ,
2022-02-07 18:20:32 +03:00
. field_width = 4 ,
2018-12-22 15:00:10 +03:00
. min_field_value = 1 ,
. matches = has_cpuid_feature ,
} ,
2020-07-15 10:19:43 +03:00
{
. desc = " TLB range maintenance instructions " ,
. capability = ARM64_HAS_TLB_RANGE ,
. type = ARM64_CPUCAP_SYSTEM_FEATURE ,
. matches = has_cpuid_feature ,
. sys_reg = SYS_ID_AA64ISAR0_EL1 ,
2022-05-03 20:02:28 +03:00
. field_pos = ID_AA64ISAR0_EL1_TLB_SHIFT ,
2022-02-07 18:20:32 +03:00
. field_width = 4 ,
2020-07-15 10:19:43 +03:00
. sign = FTR_UNSIGNED ,
2022-05-03 20:02:28 +03:00
. min_field_value = ID_AA64ISAR0_EL1_TLB_RANGE ,
2020-07-15 10:19:43 +03:00
} ,
2018-03-26 17:12:48 +03:00
# ifdef CONFIG_ARM64_HW_AFDBM
{
/*
* Since we turn this on always , we don ' t want the user to
* think that the feature is available when it may not be .
* So hide the description .
*
* . desc = " Hardware pagetable Dirty Bit Management " ,
*
*/
. type = ARM64_CPUCAP_WEAK_LOCAL_CPU_FEATURE ,
. capability = ARM64_HW_DBM ,
. sys_reg = SYS_ID_AA64MMFR1_EL1 ,
. sign = FTR_UNSIGNED ,
2022-09-06 01:54:07 +03:00
. field_pos = ID_AA64MMFR1_EL1_HAFDBS_SHIFT ,
2022-02-07 18:20:32 +03:00
. field_width = 4 ,
2018-03-26 17:12:48 +03:00
. min_field_value = 2 ,
. matches = has_hw_dbm ,
. cpu_enable = cpu_enable_hw_dbm ,
} ,
# endif
2018-08-27 14:02:43 +03:00
{
. desc = " CRC32 instructions " ,
. capability = ARM64_HAS_CRC32 ,
. type = ARM64_CPUCAP_SYSTEM_FEATURE ,
. matches = has_cpuid_feature ,
. sys_reg = SYS_ID_AA64ISAR0_EL1 ,
2022-05-03 20:02:28 +03:00
. field_pos = ID_AA64ISAR0_EL1_CRC32_SHIFT ,
2022-02-07 18:20:32 +03:00
. field_width = 4 ,
2018-08-27 14:02:43 +03:00
. min_field_value = 1 ,
} ,
2018-06-15 13:37:34 +03:00
{
. desc = " Speculative Store Bypassing Safe (SSBS) " ,
. capability = ARM64_SSBS ,
2020-09-16 01:56:12 +03:00
. type = ARM64_CPUCAP_SYSTEM_FEATURE ,
2018-06-15 13:37:34 +03:00
. matches = has_cpuid_feature ,
. sys_reg = SYS_ID_AA64PFR1_EL1 ,
2022-09-06 01:54:04 +03:00
. field_pos = ID_AA64PFR1_EL1_SSBS_SHIFT ,
2022-02-07 18:20:32 +03:00
. field_width = 4 ,
2018-06-15 13:37:34 +03:00
. sign = FTR_UNSIGNED ,
2022-09-06 01:54:12 +03:00
. min_field_value = ID_AA64PFR1_EL1_SSBS_IMP ,
2018-06-15 13:37:34 +03:00
} ,
2018-07-31 16:08:56 +03:00
# ifdef CONFIG_ARM64_CNP
{
. desc = " Common not Private translations " ,
. capability = ARM64_HAS_CNP ,
. type = ARM64_CPUCAP_SYSTEM_FEATURE ,
. matches = has_useable_cnp ,
. sys_reg = SYS_ID_AA64MMFR2_EL1 ,
. sign = FTR_UNSIGNED ,
2022-09-06 01:54:09 +03:00
. field_pos = ID_AA64MMFR2_EL1_CnP_SHIFT ,
2022-02-07 18:20:32 +03:00
. field_width = 4 ,
2018-07-31 16:08:56 +03:00
. min_field_value = 1 ,
. cpu_enable = cpu_enable_cnp ,
} ,
2018-08-07 15:47:06 +03:00
# endif
2018-06-14 13:21:34 +03:00
{
. desc = " Speculation barrier (SB) " ,
. capability = ARM64_HAS_SB ,
. type = ARM64_CPUCAP_SYSTEM_FEATURE ,
. matches = has_cpuid_feature ,
. sys_reg = SYS_ID_AA64ISAR1_EL1 ,
2022-07-04 20:02:49 +03:00
. field_pos = ID_AA64ISAR1_EL1_SB_SHIFT ,
2022-02-07 18:20:32 +03:00
. field_width = 4 ,
2018-06-14 13:21:34 +03:00
. sign = FTR_UNSIGNED ,
. min_field_value = 1 ,
} ,
2018-12-07 21:39:24 +03:00
# ifdef CONFIG_ARM64_PTR_AUTH
{
2022-02-24 15:49:51 +03:00
. desc = " Address authentication (architected QARMA5 algorithm) " ,
. capability = ARM64_HAS_ADDRESS_AUTH_ARCH_QARMA5 ,
2020-03-13 12:04:55 +03:00
. type = ARM64_CPUCAP_BOOT_CPU_FEATURE ,
2018-12-07 21:39:24 +03:00
. sys_reg = SYS_ID_AA64ISAR1_EL1 ,
. sign = FTR_UNSIGNED ,
2022-07-04 20:02:49 +03:00
. field_pos = ID_AA64ISAR1_EL1_APA_SHIFT ,
2022-02-07 18:20:32 +03:00
. field_width = 4 ,
2022-07-04 20:02:49 +03:00
. min_field_value = ID_AA64ISAR1_EL1_APA_PAuth ,
arm64: cpufeature: Modify address authentication cpufeature to exact
The current address authentication cpufeature levels are set as LOWER_SAFE
which is not compatible with the different configurations added for Armv8.3
ptrauth enhancements as the different levels have different behaviour and
there is no tunable to enable the lower safe versions. This is rectified
by setting those cpufeature type as EXACT.
The current cpufeature framework also does not interfere in the booting of
non-exact secondary cpus but rather marks them as tainted. As a workaround
this is fixed by replacing the generic match handler with a new handler
specific to ptrauth.
After this change, if there is any variation in ptrauth configurations in
secondary cpus from boot cpu then those mismatched cpus are parked in an
infinite loop.
Following ptrauth crash log is observed in Arm fastmodel with simulated
mismatched cpus without this fix,
CPU features: SANITY CHECK: Unexpected variation in SYS_ID_AA64ISAR1_EL1. Boot CPU: 0x11111110211402, CPU4: 0x11111110211102
CPU features: Unsupported CPU feature variation detected.
GICv3: CPU4: found redistributor 100 region 0:0x000000002f180000
CPU4: Booted secondary processor 0x0000000100 [0x410fd0f0]
Unable to handle kernel paging request at virtual address bfff800010dadf3c
Mem abort info:
ESR = 0x86000004
EC = 0x21: IABT (current EL), IL = 32 bits
SET = 0, FnV = 0
EA = 0, S1PTW = 0
[bfff800010dadf3c] address between user and kernel address ranges
Internal error: Oops: 86000004 [#1] PREEMPT SMP
Modules linked in:
CPU: 4 PID: 29 Comm: migration/4 Tainted: G S 5.8.0-rc4-00005-ge658591d66d1-dirty #158
Hardware name: Foundation-v8A (DT)
pstate: 60000089 (nZCv daIf -PAN -UAO BTYPE=--)
pc : 0xbfff800010dadf3c
lr : __schedule+0x2b4/0x5a8
sp : ffff800012043d70
x29: ffff800012043d70 x28: 0080000000000000
x27: ffff800011cbe000 x26: ffff00087ad37580
x25: ffff00087ad37000 x24: ffff800010de7d50
x23: ffff800011674018 x22: 0784800010dae2a8
x21: ffff00087ad37000 x20: ffff00087acb8000
x19: ffff00087f742100 x18: 0000000000000030
x17: 0000000000000000 x16: 0000000000000000
x15: ffff800011ac1000 x14: 00000000000001bd
x13: 0000000000000000 x12: 0000000000000000
x11: 0000000000000000 x10: 71519a147ddfeb82
x9 : 825d5ec0fb246314 x8 : ffff00087ad37dd8
x7 : 0000000000000000 x6 : 00000000fffedb0e
x5 : 00000000ffffffff x4 : 0000000000000000
x3 : 0000000000000028 x2 : ffff80086e11e000
x1 : ffff00087ad37000 x0 : ffff00087acdc600
Call trace:
0xbfff800010dadf3c
schedule+0x78/0x110
schedule_preempt_disabled+0x24/0x40
__kthread_parkme+0x68/0xd0
kthread+0x138/0x160
ret_from_fork+0x10/0x34
Code: bad PC value
After this fix, the mismatched CPU4 is parked as,
CPU features: CPU4: Detected conflict for capability 39 (Address authentication (IMP DEF algorithm)), System: 1, CPU: 0
CPU4: will not boot
CPU4: failed to come online
CPU4: died during early boot
[Suzuki: Introduce new matching function for address authentication]
Suggested-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Signed-off-by: Amit Daniel Kachhap <amit.kachhap@arm.com>
Reviewed-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Link: https://lore.kernel.org/r/20200914083656.21428-5-amit.kachhap@arm.com
Signed-off-by: Will Deacon <will@kernel.org>
2020-09-14 11:36:54 +03:00
. matches = has_address_auth_cpucap ,
2018-12-07 21:39:24 +03:00
} ,
2022-02-24 15:49:52 +03:00
{
. desc = " Address authentication (architected QARMA3 algorithm) " ,
. capability = ARM64_HAS_ADDRESS_AUTH_ARCH_QARMA3 ,
. type = ARM64_CPUCAP_BOOT_CPU_FEATURE ,
. sys_reg = SYS_ID_AA64ISAR2_EL1 ,
. sign = FTR_UNSIGNED ,
2022-07-04 20:02:50 +03:00
. field_pos = ID_AA64ISAR2_EL1_APA3_SHIFT ,
2022-03-14 22:04:22 +03:00
. field_width = 4 ,
2022-07-04 20:02:50 +03:00
. min_field_value = ID_AA64ISAR2_EL1_APA3_PAuth ,
2022-02-24 15:49:52 +03:00
. matches = has_address_auth_cpucap ,
} ,
2018-12-07 21:39:24 +03:00
{
. desc = " Address authentication (IMP DEF algorithm) " ,
. capability = ARM64_HAS_ADDRESS_AUTH_IMP_DEF ,
2020-03-13 12:04:55 +03:00
. type = ARM64_CPUCAP_BOOT_CPU_FEATURE ,
2018-12-07 21:39:24 +03:00
. sys_reg = SYS_ID_AA64ISAR1_EL1 ,
. sign = FTR_UNSIGNED ,
2022-07-04 20:02:49 +03:00
. field_pos = ID_AA64ISAR1_EL1_API_SHIFT ,
2022-02-07 18:20:32 +03:00
. field_width = 4 ,
2022-07-04 20:02:49 +03:00
. min_field_value = ID_AA64ISAR1_EL1_API_PAuth ,
arm64: cpufeature: Modify address authentication cpufeature to exact
The current address authentication cpufeature levels are set as LOWER_SAFE
which is not compatible with the different configurations added for Armv8.3
ptrauth enhancements as the different levels have different behaviour and
there is no tunable to enable the lower safe versions. This is rectified
by setting those cpufeature type as EXACT.
The current cpufeature framework also does not interfere in the booting of
non-exact secondary cpus but rather marks them as tainted. As a workaround
this is fixed by replacing the generic match handler with a new handler
specific to ptrauth.
After this change, if there is any variation in ptrauth configurations in
secondary cpus from boot cpu then those mismatched cpus are parked in an
infinite loop.
Following ptrauth crash log is observed in Arm fastmodel with simulated
mismatched cpus without this fix,
CPU features: SANITY CHECK: Unexpected variation in SYS_ID_AA64ISAR1_EL1. Boot CPU: 0x11111110211402, CPU4: 0x11111110211102
CPU features: Unsupported CPU feature variation detected.
GICv3: CPU4: found redistributor 100 region 0:0x000000002f180000
CPU4: Booted secondary processor 0x0000000100 [0x410fd0f0]
Unable to handle kernel paging request at virtual address bfff800010dadf3c
Mem abort info:
ESR = 0x86000004
EC = 0x21: IABT (current EL), IL = 32 bits
SET = 0, FnV = 0
EA = 0, S1PTW = 0
[bfff800010dadf3c] address between user and kernel address ranges
Internal error: Oops: 86000004 [#1] PREEMPT SMP
Modules linked in:
CPU: 4 PID: 29 Comm: migration/4 Tainted: G S 5.8.0-rc4-00005-ge658591d66d1-dirty #158
Hardware name: Foundation-v8A (DT)
pstate: 60000089 (nZCv daIf -PAN -UAO BTYPE=--)
pc : 0xbfff800010dadf3c
lr : __schedule+0x2b4/0x5a8
sp : ffff800012043d70
x29: ffff800012043d70 x28: 0080000000000000
x27: ffff800011cbe000 x26: ffff00087ad37580
x25: ffff00087ad37000 x24: ffff800010de7d50
x23: ffff800011674018 x22: 0784800010dae2a8
x21: ffff00087ad37000 x20: ffff00087acb8000
x19: ffff00087f742100 x18: 0000000000000030
x17: 0000000000000000 x16: 0000000000000000
x15: ffff800011ac1000 x14: 00000000000001bd
x13: 0000000000000000 x12: 0000000000000000
x11: 0000000000000000 x10: 71519a147ddfeb82
x9 : 825d5ec0fb246314 x8 : ffff00087ad37dd8
x7 : 0000000000000000 x6 : 00000000fffedb0e
x5 : 00000000ffffffff x4 : 0000000000000000
x3 : 0000000000000028 x2 : ffff80086e11e000
x1 : ffff00087ad37000 x0 : ffff00087acdc600
Call trace:
0xbfff800010dadf3c
schedule+0x78/0x110
schedule_preempt_disabled+0x24/0x40
__kthread_parkme+0x68/0xd0
kthread+0x138/0x160
ret_from_fork+0x10/0x34
Code: bad PC value
After this fix, the mismatched CPU4 is parked as,
CPU features: CPU4: Detected conflict for capability 39 (Address authentication (IMP DEF algorithm)), System: 1, CPU: 0
CPU4: will not boot
CPU4: failed to come online
CPU4: died during early boot
[Suzuki: Introduce new matching function for address authentication]
Suggested-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Signed-off-by: Amit Daniel Kachhap <amit.kachhap@arm.com>
Reviewed-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Link: https://lore.kernel.org/r/20200914083656.21428-5-amit.kachhap@arm.com
Signed-off-by: Will Deacon <will@kernel.org>
2020-09-14 11:36:54 +03:00
. matches = has_address_auth_cpucap ,
2020-03-13 12:04:49 +03:00
} ,
{
. capability = ARM64_HAS_ADDRESS_AUTH ,
2020-03-13 12:04:55 +03:00
. type = ARM64_CPUCAP_BOOT_CPU_FEATURE ,
arm64: cpufeature: Modify address authentication cpufeature to exact
The current address authentication cpufeature levels are set as LOWER_SAFE
which is not compatible with the different configurations added for Armv8.3
ptrauth enhancements as the different levels have different behaviour and
there is no tunable to enable the lower safe versions. This is rectified
by setting those cpufeature type as EXACT.
The current cpufeature framework also does not interfere in the booting of
non-exact secondary cpus but rather marks them as tainted. As a workaround
this is fixed by replacing the generic match handler with a new handler
specific to ptrauth.
After this change, if there is any variation in ptrauth configurations in
secondary cpus from boot cpu then those mismatched cpus are parked in an
infinite loop.
Following ptrauth crash log is observed in Arm fastmodel with simulated
mismatched cpus without this fix,
CPU features: SANITY CHECK: Unexpected variation in SYS_ID_AA64ISAR1_EL1. Boot CPU: 0x11111110211402, CPU4: 0x11111110211102
CPU features: Unsupported CPU feature variation detected.
GICv3: CPU4: found redistributor 100 region 0:0x000000002f180000
CPU4: Booted secondary processor 0x0000000100 [0x410fd0f0]
Unable to handle kernel paging request at virtual address bfff800010dadf3c
Mem abort info:
ESR = 0x86000004
EC = 0x21: IABT (current EL), IL = 32 bits
SET = 0, FnV = 0
EA = 0, S1PTW = 0
[bfff800010dadf3c] address between user and kernel address ranges
Internal error: Oops: 86000004 [#1] PREEMPT SMP
Modules linked in:
CPU: 4 PID: 29 Comm: migration/4 Tainted: G S 5.8.0-rc4-00005-ge658591d66d1-dirty #158
Hardware name: Foundation-v8A (DT)
pstate: 60000089 (nZCv daIf -PAN -UAO BTYPE=--)
pc : 0xbfff800010dadf3c
lr : __schedule+0x2b4/0x5a8
sp : ffff800012043d70
x29: ffff800012043d70 x28: 0080000000000000
x27: ffff800011cbe000 x26: ffff00087ad37580
x25: ffff00087ad37000 x24: ffff800010de7d50
x23: ffff800011674018 x22: 0784800010dae2a8
x21: ffff00087ad37000 x20: ffff00087acb8000
x19: ffff00087f742100 x18: 0000000000000030
x17: 0000000000000000 x16: 0000000000000000
x15: ffff800011ac1000 x14: 00000000000001bd
x13: 0000000000000000 x12: 0000000000000000
x11: 0000000000000000 x10: 71519a147ddfeb82
x9 : 825d5ec0fb246314 x8 : ffff00087ad37dd8
x7 : 0000000000000000 x6 : 00000000fffedb0e
x5 : 00000000ffffffff x4 : 0000000000000000
x3 : 0000000000000028 x2 : ffff80086e11e000
x1 : ffff00087ad37000 x0 : ffff00087acdc600
Call trace:
0xbfff800010dadf3c
schedule+0x78/0x110
schedule_preempt_disabled+0x24/0x40
__kthread_parkme+0x68/0xd0
kthread+0x138/0x160
ret_from_fork+0x10/0x34
Code: bad PC value
After this fix, the mismatched CPU4 is parked as,
CPU features: CPU4: Detected conflict for capability 39 (Address authentication (IMP DEF algorithm)), System: 1, CPU: 0
CPU4: will not boot
CPU4: failed to come online
CPU4: died during early boot
[Suzuki: Introduce new matching function for address authentication]
Suggested-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Signed-off-by: Amit Daniel Kachhap <amit.kachhap@arm.com>
Reviewed-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Link: https://lore.kernel.org/r/20200914083656.21428-5-amit.kachhap@arm.com
Signed-off-by: Will Deacon <will@kernel.org>
2020-09-14 11:36:54 +03:00
. matches = has_address_auth_metacap ,
2018-12-07 21:39:24 +03:00
} ,
{
2022-02-24 15:49:51 +03:00
. desc = " Generic authentication (architected QARMA5 algorithm) " ,
. capability = ARM64_HAS_GENERIC_AUTH_ARCH_QARMA5 ,
2018-12-07 21:39:24 +03:00
. type = ARM64_CPUCAP_SYSTEM_FEATURE ,
. sys_reg = SYS_ID_AA64ISAR1_EL1 ,
. sign = FTR_UNSIGNED ,
2022-07-04 20:02:49 +03:00
. field_pos = ID_AA64ISAR1_EL1_GPA_SHIFT ,
2022-02-07 18:20:32 +03:00
. field_width = 4 ,
2022-07-04 20:02:49 +03:00
. min_field_value = ID_AA64ISAR1_EL1_GPA_IMP ,
2018-12-07 21:39:24 +03:00
. matches = has_cpuid_feature ,
} ,
2022-02-24 15:49:52 +03:00
{
. desc = " Generic authentication (architected QARMA3 algorithm) " ,
. capability = ARM64_HAS_GENERIC_AUTH_ARCH_QARMA3 ,
. type = ARM64_CPUCAP_SYSTEM_FEATURE ,
. sys_reg = SYS_ID_AA64ISAR2_EL1 ,
. sign = FTR_UNSIGNED ,
2022-07-04 20:02:50 +03:00
. field_pos = ID_AA64ISAR2_EL1_GPA3_SHIFT ,
2022-03-14 22:04:22 +03:00
. field_width = 4 ,
2022-07-04 20:02:50 +03:00
. min_field_value = ID_AA64ISAR2_EL1_GPA3_IMP ,
2022-02-24 15:49:52 +03:00
. matches = has_cpuid_feature ,
} ,
2018-12-07 21:39:24 +03:00
{
. desc = " Generic authentication (IMP DEF algorithm) " ,
. capability = ARM64_HAS_GENERIC_AUTH_IMP_DEF ,
. type = ARM64_CPUCAP_SYSTEM_FEATURE ,
. sys_reg = SYS_ID_AA64ISAR1_EL1 ,
. sign = FTR_UNSIGNED ,
2022-07-04 20:02:49 +03:00
. field_pos = ID_AA64ISAR1_EL1_GPI_SHIFT ,
2022-02-07 18:20:32 +03:00
. field_width = 4 ,
2022-07-04 20:02:49 +03:00
. min_field_value = ID_AA64ISAR1_EL1_GPI_IMP ,
2018-12-07 21:39:24 +03:00
. matches = has_cpuid_feature ,
} ,
2020-03-13 12:04:49 +03:00
{
. capability = ARM64_HAS_GENERIC_AUTH ,
. type = ARM64_CPUCAP_SYSTEM_FEATURE ,
. matches = has_generic_auth ,
} ,
2018-12-07 21:39:24 +03:00
# endif /* CONFIG_ARM64_PTR_AUTH */
2019-01-31 17:58:42 +03:00
# ifdef CONFIG_ARM64_PSEUDO_NMI
{
/*
* Depends on having GICv3
*/
. desc = " IRQ priority masking " ,
. capability = ARM64_HAS_IRQ_PRIO_MASKING ,
. type = ARM64_CPUCAP_STRICT_BOOT_CPU_FEATURE ,
. matches = can_use_gic_priorities ,
. sys_reg = SYS_ID_AA64PFR0_EL1 ,
2022-09-06 01:54:03 +03:00
. field_pos = ID_AA64PFR0_EL1_GIC_SHIFT ,
2022-02-07 18:20:32 +03:00
. field_width = 4 ,
2019-01-31 17:58:42 +03:00
. sign = FTR_UNSIGNED ,
. min_field_value = 1 ,
} ,
2019-12-09 21:12:14 +03:00
# endif
# ifdef CONFIG_ARM64_E0PD
{
. desc = " E0PD " ,
. capability = ARM64_HAS_E0PD ,
. type = ARM64_CPUCAP_SYSTEM_FEATURE ,
. sys_reg = SYS_ID_AA64MMFR2_EL1 ,
. sign = FTR_UNSIGNED ,
2022-02-07 18:20:32 +03:00
. field_width = 4 ,
2022-09-06 01:54:02 +03:00
. field_pos = ID_AA64MMFR2_EL1_E0PD_SHIFT ,
2019-12-09 21:12:14 +03:00
. matches = has_cpuid_feature ,
. min_field_value = 1 ,
. cpu_enable = cpu_enable_e0pd ,
} ,
2020-01-22 14:38:53 +03:00
# endif
2020-01-21 15:58:52 +03:00
{
. desc = " Random Number Generator " ,
. capability = ARM64_HAS_RNG ,
. type = ARM64_CPUCAP_SYSTEM_FEATURE ,
. matches = has_cpuid_feature ,
. sys_reg = SYS_ID_AA64ISAR0_EL1 ,
2022-05-03 20:02:28 +03:00
. field_pos = ID_AA64ISAR0_EL1_RNDR_SHIFT ,
2022-02-07 18:20:32 +03:00
. field_width = 4 ,
2020-01-21 15:58:52 +03:00
. sign = FTR_UNSIGNED ,
. min_field_value = 1 ,
} ,
2020-03-16 19:50:45 +03:00
# ifdef CONFIG_ARM64_BTI
{
. desc = " Branch Target Identification " ,
. capability = ARM64_BTI ,
2020-05-06 22:51:31 +03:00
# ifdef CONFIG_ARM64_BTI_KERNEL
. type = ARM64_CPUCAP_STRICT_BOOT_CPU_FEATURE ,
# else
2020-03-16 19:50:45 +03:00
. type = ARM64_CPUCAP_SYSTEM_FEATURE ,
2020-05-06 22:51:31 +03:00
# endif
2020-03-16 19:50:45 +03:00
. matches = has_cpuid_feature ,
. cpu_enable = bti_enable ,
. sys_reg = SYS_ID_AA64PFR1_EL1 ,
2022-09-06 01:54:04 +03:00
. field_pos = ID_AA64PFR1_EL1_BT_SHIFT ,
2022-02-07 18:20:32 +03:00
. field_width = 4 ,
2022-09-06 01:54:15 +03:00
. min_field_value = ID_AA64PFR1_EL1_BT_IMP ,
2020-03-16 19:50:45 +03:00
. sign = FTR_UNSIGNED ,
} ,
2019-01-31 17:58:42 +03:00
# endif
2019-09-06 12:58:01 +03:00
# ifdef CONFIG_ARM64_MTE
{
. desc = " Memory Tagging Extension " ,
. capability = ARM64_MTE ,
. type = ARM64_CPUCAP_STRICT_BOOT_CPU_FEATURE ,
. matches = has_cpuid_feature ,
. sys_reg = SYS_ID_AA64PFR1_EL1 ,
2022-09-06 01:54:04 +03:00
. field_pos = ID_AA64PFR1_EL1_MTE_SHIFT ,
2022-02-07 18:20:32 +03:00
. field_width = 4 ,
2022-09-06 01:54:13 +03:00
. min_field_value = ID_AA64PFR1_EL1_MTE_MTE2 ,
2019-09-06 12:58:01 +03:00
. sign = FTR_UNSIGNED ,
2020-05-04 16:42:36 +03:00
. cpu_enable = cpu_enable_mte ,
2019-09-06 12:58:01 +03:00
} ,
2021-10-06 18:47:49 +03:00
{
. desc = " Asymmetric MTE Tag Check Fault " ,
. capability = ARM64_MTE_ASYMM ,
. type = ARM64_CPUCAP_BOOT_CPU_FEATURE ,
. matches = has_cpuid_feature ,
. sys_reg = SYS_ID_AA64PFR1_EL1 ,
2022-09-06 01:54:04 +03:00
. field_pos = ID_AA64PFR1_EL1_MTE_SHIFT ,
2022-02-07 18:20:32 +03:00
. field_width = 4 ,
2022-09-06 01:54:13 +03:00
. min_field_value = ID_AA64PFR1_EL1_MTE_MTE3 ,
2021-10-06 18:47:49 +03:00
. sign = FTR_UNSIGNED ,
} ,
2019-09-06 12:58:01 +03:00
# endif /* CONFIG_ARM64_MTE */
2020-06-30 16:02:22 +03:00
{
. desc = " RCpc load-acquire (LDAPR) " ,
. capability = ARM64_HAS_LDAPR ,
. type = ARM64_CPUCAP_SYSTEM_FEATURE ,
. sys_reg = SYS_ID_AA64ISAR1_EL1 ,
. sign = FTR_UNSIGNED ,
2022-07-04 20:02:49 +03:00
. field_pos = ID_AA64ISAR1_EL1_LRCPC_SHIFT ,
2022-02-07 18:20:32 +03:00
. field_width = 4 ,
2020-06-30 16:02:22 +03:00
. matches = has_cpuid_feature ,
. min_field_value = 1 ,
} ,
2022-04-19 14:22:16 +03:00
# ifdef CONFIG_ARM64_SME
{
. desc = " Scalable Matrix Extension " ,
. type = ARM64_CPUCAP_SYSTEM_FEATURE ,
. capability = ARM64_SME ,
. sys_reg = SYS_ID_AA64PFR1_EL1 ,
. sign = FTR_UNSIGNED ,
2022-09-06 01:54:04 +03:00
. field_pos = ID_AA64PFR1_EL1_SME_SHIFT ,
2022-04-19 14:22:16 +03:00
. field_width = 4 ,
2022-09-06 01:54:16 +03:00
. min_field_value = ID_AA64PFR1_EL1_SME_IMP ,
2022-04-19 14:22:16 +03:00
. matches = has_cpuid_feature ,
. cpu_enable = sme_kernel_enable ,
} ,
/* FA64 should be sorted after the base SME capability */
{
. desc = " FA64 " ,
. type = ARM64_CPUCAP_SYSTEM_FEATURE ,
. capability = ARM64_SME_FA64 ,
. sys_reg = SYS_ID_AA64SMFR0_EL1 ,
. sign = FTR_UNSIGNED ,
2022-07-04 20:02:46 +03:00
. field_pos = ID_AA64SMFR0_EL1_FA64_SHIFT ,
2022-04-19 14:22:16 +03:00
. field_width = 1 ,
2022-07-04 20:02:46 +03:00
. min_field_value = ID_AA64SMFR0_EL1_FA64_IMP ,
2022-04-19 14:22:16 +03:00
. matches = has_cpuid_feature ,
. cpu_enable = fa64_kernel_enable ,
} ,
# endif /* CONFIG_ARM64_SME */
2022-04-19 21:27:52 +03:00
{
. desc = " WFx with timeout " ,
. capability = ARM64_HAS_WFXT ,
. type = ARM64_CPUCAP_SYSTEM_FEATURE ,
. sys_reg = SYS_ID_AA64ISAR2_EL1 ,
. sign = FTR_UNSIGNED ,
2022-07-04 20:02:50 +03:00
. field_pos = ID_AA64ISAR2_EL1_WFxT_SHIFT ,
2022-04-19 21:27:52 +03:00
. field_width = 4 ,
. matches = has_cpuid_feature ,
2022-07-04 20:02:50 +03:00
. min_field_value = ID_AA64ISAR2_EL1_WFxT_IMP ,
2022-04-19 21:27:52 +03:00
} ,
2022-06-22 14:54:24 +03:00
{
. desc = " Trap EL0 IMPLEMENTATION DEFINED functionality " ,
. capability = ARM64_HAS_TIDCP1 ,
. type = ARM64_CPUCAP_SYSTEM_FEATURE ,
. sys_reg = SYS_ID_AA64MMFR1_EL1 ,
. sign = FTR_UNSIGNED ,
2022-09-06 01:54:07 +03:00
. field_pos = ID_AA64MMFR1_EL1_TIDCP1_SHIFT ,
2022-06-22 14:54:24 +03:00
. field_width = 4 ,
2022-09-06 01:54:07 +03:00
. min_field_value = ID_AA64MMFR1_EL1_TIDCP1_IMP ,
2022-06-22 14:54:24 +03:00
. matches = has_cpuid_feature ,
. cpu_enable = cpu_trap_el0_impdef ,
} ,
2015-03-27 16:09:23 +03:00
{ } ,
} ;
2022-02-07 18:20:32 +03:00
# define HWCAP_CPUID_MATCH(reg, field, width, s, min_value) \
2022-09-09 19:59:36 +03:00
. matches = has_user_cpuid_feature , \
2018-12-12 18:53:54 +03:00
. sys_reg = reg , \
. field_pos = field , \
2022-02-07 18:20:32 +03:00
. field_width = width , \
2018-12-12 18:53:54 +03:00
. sign = s , \
. min_field_value = min_value ,
# define __HWCAP_CAP(name, cap_type, cap) \
. desc = name , \
. type = ARM64_CPUCAP_SYSTEM_FEATURE , \
. hwcap_type = cap_type , \
. hwcap = cap , \
2022-02-07 18:20:32 +03:00
# define HWCAP_CAP(reg, field, width, s, min_value, cap_type, cap) \
2018-12-12 18:53:54 +03:00
{ \
__HWCAP_CAP ( # cap , cap_type , cap ) \
2022-02-07 18:20:32 +03:00
HWCAP_CPUID_MATCH ( reg , field , width , s , min_value ) \
2015-10-19 16:24:52 +03:00
}
2018-12-12 18:53:54 +03:00
# define HWCAP_MULTI_CAP(list, cap_type, cap) \
{ \
__HWCAP_CAP ( # cap , cap_type , cap ) \
. matches = cpucap_multi_entry_cap_matches , \
. match_list = list , \
}
2020-01-14 02:30:20 +03:00
# define HWCAP_CAP_MATCH(match, cap_type, cap) \
{ \
__HWCAP_CAP ( # cap , cap_type , cap ) \
. matches = match , \
}
2018-12-12 18:53:54 +03:00
# ifdef CONFIG_ARM64_PTR_AUTH
static const struct arm64_cpu_capabilities ptr_auth_hwcap_addr_matches [ ] = {
{
2022-07-04 20:02:49 +03:00
HWCAP_CPUID_MATCH ( SYS_ID_AA64ISAR1_EL1 , ID_AA64ISAR1_EL1_APA_SHIFT ,
2022-02-07 18:20:32 +03:00
4 , FTR_UNSIGNED ,
2022-07-04 20:02:49 +03:00
ID_AA64ISAR1_EL1_APA_PAuth )
2018-12-12 18:53:54 +03:00
} ,
2022-02-24 15:49:52 +03:00
{
2022-07-04 20:02:50 +03:00
HWCAP_CPUID_MATCH ( SYS_ID_AA64ISAR2_EL1 , ID_AA64ISAR2_EL1_APA3_SHIFT ,
4 , FTR_UNSIGNED , ID_AA64ISAR2_EL1_APA3_PAuth )
2022-02-24 15:49:52 +03:00
} ,
2018-12-12 18:53:54 +03:00
{
2022-07-04 20:02:49 +03:00
HWCAP_CPUID_MATCH ( SYS_ID_AA64ISAR1_EL1 , ID_AA64ISAR1_EL1_API_SHIFT ,
4 , FTR_UNSIGNED , ID_AA64ISAR1_EL1_API_PAuth )
2018-12-12 18:53:54 +03:00
} ,
{ } ,
} ;
static const struct arm64_cpu_capabilities ptr_auth_hwcap_gen_matches [ ] = {
{
2022-07-04 20:02:49 +03:00
HWCAP_CPUID_MATCH ( SYS_ID_AA64ISAR1_EL1 , ID_AA64ISAR1_EL1_GPA_SHIFT ,
4 , FTR_UNSIGNED , ID_AA64ISAR1_EL1_GPA_IMP )
2018-12-12 18:53:54 +03:00
} ,
2022-02-24 15:49:52 +03:00
{
2022-07-04 20:02:50 +03:00
HWCAP_CPUID_MATCH ( SYS_ID_AA64ISAR2_EL1 , ID_AA64ISAR2_EL1_GPA3_SHIFT ,
4 , FTR_UNSIGNED , ID_AA64ISAR2_EL1_GPA3_IMP )
2022-02-24 15:49:52 +03:00
} ,
2018-12-12 18:53:54 +03:00
{
2022-07-04 20:02:49 +03:00
HWCAP_CPUID_MATCH ( SYS_ID_AA64ISAR1_EL1 , ID_AA64ISAR1_EL1_GPI_SHIFT ,
4 , FTR_UNSIGNED , ID_AA64ISAR1_EL1_GPI_IMP )
2018-12-12 18:53:54 +03:00
} ,
{ } ,
} ;
# endif
2016-04-18 12:28:32 +03:00
static const struct arm64_cpu_capabilities arm64_elf_hwcaps [ ] = {
2022-05-03 20:02:28 +03:00
HWCAP_CAP ( SYS_ID_AA64ISAR0_EL1 , ID_AA64ISAR0_EL1_AES_SHIFT , 4 , FTR_UNSIGNED , 2 , CAP_HWCAP , KERNEL_HWCAP_PMULL ) ,
HWCAP_CAP ( SYS_ID_AA64ISAR0_EL1 , ID_AA64ISAR0_EL1_AES_SHIFT , 4 , FTR_UNSIGNED , 1 , CAP_HWCAP , KERNEL_HWCAP_AES ) ,
HWCAP_CAP ( SYS_ID_AA64ISAR0_EL1 , ID_AA64ISAR0_EL1_SHA1_SHIFT , 4 , FTR_UNSIGNED , 1 , CAP_HWCAP , KERNEL_HWCAP_SHA1 ) ,
HWCAP_CAP ( SYS_ID_AA64ISAR0_EL1 , ID_AA64ISAR0_EL1_SHA2_SHIFT , 4 , FTR_UNSIGNED , 1 , CAP_HWCAP , KERNEL_HWCAP_SHA2 ) ,
HWCAP_CAP ( SYS_ID_AA64ISAR0_EL1 , ID_AA64ISAR0_EL1_SHA2_SHIFT , 4 , FTR_UNSIGNED , 2 , CAP_HWCAP , KERNEL_HWCAP_SHA512 ) ,
HWCAP_CAP ( SYS_ID_AA64ISAR0_EL1 , ID_AA64ISAR0_EL1_CRC32_SHIFT , 4 , FTR_UNSIGNED , 1 , CAP_HWCAP , KERNEL_HWCAP_CRC32 ) ,
HWCAP_CAP ( SYS_ID_AA64ISAR0_EL1 , ID_AA64ISAR0_EL1_ATOMIC_SHIFT , 4 , FTR_UNSIGNED , 2 , CAP_HWCAP , KERNEL_HWCAP_ATOMICS ) ,
HWCAP_CAP ( SYS_ID_AA64ISAR0_EL1 , ID_AA64ISAR0_EL1_RDM_SHIFT , 4 , FTR_UNSIGNED , 1 , CAP_HWCAP , KERNEL_HWCAP_ASIMDRDM ) ,
HWCAP_CAP ( SYS_ID_AA64ISAR0_EL1 , ID_AA64ISAR0_EL1_SHA3_SHIFT , 4 , FTR_UNSIGNED , 1 , CAP_HWCAP , KERNEL_HWCAP_SHA3 ) ,
HWCAP_CAP ( SYS_ID_AA64ISAR0_EL1 , ID_AA64ISAR0_EL1_SM3_SHIFT , 4 , FTR_UNSIGNED , 1 , CAP_HWCAP , KERNEL_HWCAP_SM3 ) ,
HWCAP_CAP ( SYS_ID_AA64ISAR0_EL1 , ID_AA64ISAR0_EL1_SM4_SHIFT , 4 , FTR_UNSIGNED , 1 , CAP_HWCAP , KERNEL_HWCAP_SM4 ) ,
HWCAP_CAP ( SYS_ID_AA64ISAR0_EL1 , ID_AA64ISAR0_EL1_DP_SHIFT , 4 , FTR_UNSIGNED , 1 , CAP_HWCAP , KERNEL_HWCAP_ASIMDDP ) ,
HWCAP_CAP ( SYS_ID_AA64ISAR0_EL1 , ID_AA64ISAR0_EL1_FHM_SHIFT , 4 , FTR_UNSIGNED , 1 , CAP_HWCAP , KERNEL_HWCAP_ASIMDFHM ) ,
HWCAP_CAP ( SYS_ID_AA64ISAR0_EL1 , ID_AA64ISAR0_EL1_TS_SHIFT , 4 , FTR_UNSIGNED , 1 , CAP_HWCAP , KERNEL_HWCAP_FLAGM ) ,
HWCAP_CAP ( SYS_ID_AA64ISAR0_EL1 , ID_AA64ISAR0_EL1_TS_SHIFT , 4 , FTR_UNSIGNED , 2 , CAP_HWCAP , KERNEL_HWCAP_FLAGM2 ) ,
HWCAP_CAP ( SYS_ID_AA64ISAR0_EL1 , ID_AA64ISAR0_EL1_RNDR_SHIFT , 4 , FTR_UNSIGNED , 1 , CAP_HWCAP , KERNEL_HWCAP_RNG ) ,
2022-09-06 01:54:03 +03:00
HWCAP_CAP ( SYS_ID_AA64PFR0_EL1 , ID_AA64PFR0_EL1_FP_SHIFT , 4 , FTR_SIGNED , 0 , CAP_HWCAP , KERNEL_HWCAP_FP ) ,
HWCAP_CAP ( SYS_ID_AA64PFR0_EL1 , ID_AA64PFR0_EL1_FP_SHIFT , 4 , FTR_SIGNED , 1 , CAP_HWCAP , KERNEL_HWCAP_FPHP ) ,
2022-09-06 01:54:11 +03:00
HWCAP_CAP ( SYS_ID_AA64PFR0_EL1 , ID_AA64PFR0_EL1_AdvSIMD_SHIFT , 4 , FTR_SIGNED , 0 , CAP_HWCAP , KERNEL_HWCAP_ASIMD ) ,
HWCAP_CAP ( SYS_ID_AA64PFR0_EL1 , ID_AA64PFR0_EL1_AdvSIMD_SHIFT , 4 , FTR_SIGNED , 1 , CAP_HWCAP , KERNEL_HWCAP_ASIMDHP ) ,
2022-09-06 01:54:03 +03:00
HWCAP_CAP ( SYS_ID_AA64PFR0_EL1 , ID_AA64PFR0_EL1_DIT_SHIFT , 4 , FTR_SIGNED , 1 , CAP_HWCAP , KERNEL_HWCAP_DIT ) ,
2022-07-04 20:02:49 +03:00
HWCAP_CAP ( SYS_ID_AA64ISAR1_EL1 , ID_AA64ISAR1_EL1_DPB_SHIFT , 4 , FTR_UNSIGNED , 1 , CAP_HWCAP , KERNEL_HWCAP_DCPOP ) ,
HWCAP_CAP ( SYS_ID_AA64ISAR1_EL1 , ID_AA64ISAR1_EL1_DPB_SHIFT , 4 , FTR_UNSIGNED , 2 , CAP_HWCAP , KERNEL_HWCAP_DCPODP ) ,
HWCAP_CAP ( SYS_ID_AA64ISAR1_EL1 , ID_AA64ISAR1_EL1_JSCVT_SHIFT , 4 , FTR_UNSIGNED , 1 , CAP_HWCAP , KERNEL_HWCAP_JSCVT ) ,
HWCAP_CAP ( SYS_ID_AA64ISAR1_EL1 , ID_AA64ISAR1_EL1_FCMA_SHIFT , 4 , FTR_UNSIGNED , 1 , CAP_HWCAP , KERNEL_HWCAP_FCMA ) ,
HWCAP_CAP ( SYS_ID_AA64ISAR1_EL1 , ID_AA64ISAR1_EL1_LRCPC_SHIFT , 4 , FTR_UNSIGNED , 1 , CAP_HWCAP , KERNEL_HWCAP_LRCPC ) ,
HWCAP_CAP ( SYS_ID_AA64ISAR1_EL1 , ID_AA64ISAR1_EL1_LRCPC_SHIFT , 4 , FTR_UNSIGNED , 2 , CAP_HWCAP , KERNEL_HWCAP_ILRCPC ) ,
HWCAP_CAP ( SYS_ID_AA64ISAR1_EL1 , ID_AA64ISAR1_EL1_FRINTTS_SHIFT , 4 , FTR_UNSIGNED , 1 , CAP_HWCAP , KERNEL_HWCAP_FRINT ) ,
HWCAP_CAP ( SYS_ID_AA64ISAR1_EL1 , ID_AA64ISAR1_EL1_SB_SHIFT , 4 , FTR_UNSIGNED , 1 , CAP_HWCAP , KERNEL_HWCAP_SB ) ,
HWCAP_CAP ( SYS_ID_AA64ISAR1_EL1 , ID_AA64ISAR1_EL1_BF16_SHIFT , 4 , FTR_UNSIGNED , 1 , CAP_HWCAP , KERNEL_HWCAP_BF16 ) ,
2022-07-25 12:58:10 +03:00
HWCAP_CAP ( SYS_ID_AA64ISAR1_EL1 , ID_AA64ISAR1_EL1_BF16_SHIFT , 4 , FTR_UNSIGNED , 2 , CAP_HWCAP , KERNEL_HWCAP_EBF16 ) ,
2022-07-04 20:02:49 +03:00
HWCAP_CAP ( SYS_ID_AA64ISAR1_EL1 , ID_AA64ISAR1_EL1_DGH_SHIFT , 4 , FTR_UNSIGNED , 1 , CAP_HWCAP , KERNEL_HWCAP_DGH ) ,
HWCAP_CAP ( SYS_ID_AA64ISAR1_EL1 , ID_AA64ISAR1_EL1_I8MM_SHIFT , 4 , FTR_UNSIGNED , 1 , CAP_HWCAP , KERNEL_HWCAP_I8MM ) ,
2022-09-06 01:54:02 +03:00
HWCAP_CAP ( SYS_ID_AA64MMFR2_EL1 , ID_AA64MMFR2_EL1_AT_SHIFT , 4 , FTR_UNSIGNED , 1 , CAP_HWCAP , KERNEL_HWCAP_USCAT ) ,
2017-10-31 18:51:19 +03:00
# ifdef CONFIG_ARM64_SVE
2022-09-06 01:54:10 +03:00
HWCAP_CAP ( SYS_ID_AA64PFR0_EL1 , ID_AA64PFR0_EL1_SVE_SHIFT , 4 , FTR_UNSIGNED , ID_AA64PFR0_EL1_SVE_IMP , CAP_HWCAP , KERNEL_HWCAP_SVE ) ,
2022-07-04 20:02:47 +03:00
HWCAP_CAP ( SYS_ID_AA64ZFR0_EL1 , ID_AA64ZFR0_EL1_SVEver_SHIFT , 4 , FTR_UNSIGNED , ID_AA64ZFR0_EL1_SVEver_SVE2 , CAP_HWCAP , KERNEL_HWCAP_SVE2 ) ,
HWCAP_CAP ( SYS_ID_AA64ZFR0_EL1 , ID_AA64ZFR0_EL1_AES_SHIFT , 4 , FTR_UNSIGNED , ID_AA64ZFR0_EL1_AES_IMP , CAP_HWCAP , KERNEL_HWCAP_SVEAES ) ,
HWCAP_CAP ( SYS_ID_AA64ZFR0_EL1 , ID_AA64ZFR0_EL1_AES_SHIFT , 4 , FTR_UNSIGNED , ID_AA64ZFR0_EL1_AES_PMULL128 , CAP_HWCAP , KERNEL_HWCAP_SVEPMULL ) ,
HWCAP_CAP ( SYS_ID_AA64ZFR0_EL1 , ID_AA64ZFR0_EL1_BitPerm_SHIFT , 4 , FTR_UNSIGNED , ID_AA64ZFR0_EL1_BitPerm_IMP , CAP_HWCAP , KERNEL_HWCAP_SVEBITPERM ) ,
HWCAP_CAP ( SYS_ID_AA64ZFR0_EL1 , ID_AA64ZFR0_EL1_BF16_SHIFT , 4 , FTR_UNSIGNED , ID_AA64ZFR0_EL1_BF16_IMP , CAP_HWCAP , KERNEL_HWCAP_SVEBF16 ) ,
2022-08-29 18:48:15 +03:00
HWCAP_CAP ( SYS_ID_AA64ZFR0_EL1 , ID_AA64ZFR0_EL1_BF16_SHIFT , 4 , FTR_UNSIGNED , ID_AA64ZFR0_EL1_BF16_EBF16 , CAP_HWCAP , KERNEL_HWCAP_SVE_EBF16 ) ,
2022-07-04 20:02:47 +03:00
HWCAP_CAP ( SYS_ID_AA64ZFR0_EL1 , ID_AA64ZFR0_EL1_SHA3_SHIFT , 4 , FTR_UNSIGNED , ID_AA64ZFR0_EL1_SHA3_IMP , CAP_HWCAP , KERNEL_HWCAP_SVESHA3 ) ,
HWCAP_CAP ( SYS_ID_AA64ZFR0_EL1 , ID_AA64ZFR0_EL1_SM4_SHIFT , 4 , FTR_UNSIGNED , ID_AA64ZFR0_EL1_SM4_IMP , CAP_HWCAP , KERNEL_HWCAP_SVESM4 ) ,
HWCAP_CAP ( SYS_ID_AA64ZFR0_EL1 , ID_AA64ZFR0_EL1_I8MM_SHIFT , 4 , FTR_UNSIGNED , ID_AA64ZFR0_EL1_I8MM_IMP , CAP_HWCAP , KERNEL_HWCAP_SVEI8MM ) ,
HWCAP_CAP ( SYS_ID_AA64ZFR0_EL1 , ID_AA64ZFR0_EL1_F32MM_SHIFT , 4 , FTR_UNSIGNED , ID_AA64ZFR0_EL1_F32MM_IMP , CAP_HWCAP , KERNEL_HWCAP_SVEF32MM ) ,
HWCAP_CAP ( SYS_ID_AA64ZFR0_EL1 , ID_AA64ZFR0_EL1_F64MM_SHIFT , 4 , FTR_UNSIGNED , ID_AA64ZFR0_EL1_F64MM_IMP , CAP_HWCAP , KERNEL_HWCAP_SVEF64MM ) ,
2017-10-31 18:51:19 +03:00
# endif
2022-09-06 01:54:12 +03:00
HWCAP_CAP ( SYS_ID_AA64PFR1_EL1 , ID_AA64PFR1_EL1_SSBS_SHIFT , 4 , FTR_UNSIGNED , ID_AA64PFR1_EL1_SSBS_SSBS2 , CAP_HWCAP , KERNEL_HWCAP_SSBS ) ,
2020-03-16 19:50:45 +03:00
# ifdef CONFIG_ARM64_BTI
2022-09-06 01:54:15 +03:00
HWCAP_CAP ( SYS_ID_AA64PFR1_EL1 , ID_AA64PFR1_EL1_BT_SHIFT , 4 , FTR_UNSIGNED , ID_AA64PFR1_EL1_BT_IMP , CAP_HWCAP , KERNEL_HWCAP_BTI ) ,
2020-03-16 19:50:45 +03:00
# endif
arm64: add basic pointer authentication support
This patch adds basic support for pointer authentication, allowing
userspace to make use of APIAKey, APIBKey, APDAKey, APDBKey, and
APGAKey. The kernel maintains key values for each process (shared by all
threads within), which are initialised to random values at exec() time.
The ID_AA64ISAR1_EL1.{APA,API,GPA,GPI} fields are exposed to userspace,
to describe that pointer authentication instructions are available and
that the kernel is managing the keys. Two new hwcaps are added for the
same reason: PACA (for address authentication) and PACG (for generic
authentication).
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Kristina Martsenko <kristina.martsenko@arm.com>
Tested-by: Adam Wallis <awallis@codeaurora.org>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
Cc: Suzuki K Poulose <suzuki.poulose@arm.com>
Cc: Will Deacon <will.deacon@arm.com>
[will: Fix sizeof() usage and unroll address key initialisation]
Signed-off-by: Will Deacon <will.deacon@arm.com>
2018-12-07 21:39:25 +03:00
# ifdef CONFIG_ARM64_PTR_AUTH
2019-04-09 12:52:40 +03:00
HWCAP_MULTI_CAP ( ptr_auth_hwcap_addr_matches , CAP_HWCAP , KERNEL_HWCAP_PACA ) ,
HWCAP_MULTI_CAP ( ptr_auth_hwcap_gen_matches , CAP_HWCAP , KERNEL_HWCAP_PACG ) ,
arm64: add basic pointer authentication support
This patch adds basic support for pointer authentication, allowing
userspace to make use of APIAKey, APIBKey, APDAKey, APDBKey, and
APGAKey. The kernel maintains key values for each process (shared by all
threads within), which are initialised to random values at exec() time.
The ID_AA64ISAR1_EL1.{APA,API,GPA,GPI} fields are exposed to userspace,
to describe that pointer authentication instructions are available and
that the kernel is managing the keys. Two new hwcaps are added for the
same reason: PACA (for address authentication) and PACG (for generic
authentication).
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Kristina Martsenko <kristina.martsenko@arm.com>
Tested-by: Adam Wallis <awallis@codeaurora.org>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
Cc: Suzuki K Poulose <suzuki.poulose@arm.com>
Cc: Will Deacon <will.deacon@arm.com>
[will: Fix sizeof() usage and unroll address key initialisation]
Signed-off-by: Will Deacon <will.deacon@arm.com>
2018-12-07 21:39:25 +03:00
# endif
2019-09-06 12:58:01 +03:00
# ifdef CONFIG_ARM64_MTE
2022-09-06 01:54:13 +03:00
HWCAP_CAP ( SYS_ID_AA64PFR1_EL1 , ID_AA64PFR1_EL1_MTE_SHIFT , 4 , FTR_UNSIGNED , ID_AA64PFR1_EL1_MTE_MTE2 , CAP_HWCAP , KERNEL_HWCAP_MTE ) ,
HWCAP_CAP ( SYS_ID_AA64PFR1_EL1 , ID_AA64PFR1_EL1_MTE_SHIFT , 4 , FTR_UNSIGNED , ID_AA64PFR1_EL1_MTE_MTE3 , CAP_HWCAP , KERNEL_HWCAP_MTE3 ) ,
2019-09-06 12:58:01 +03:00
# endif /* CONFIG_ARM64_MTE */
2022-09-06 01:54:01 +03:00
HWCAP_CAP ( SYS_ID_AA64MMFR0_EL1 , ID_AA64MMFR0_EL1_ECV_SHIFT , 4 , FTR_UNSIGNED , 1 , CAP_HWCAP , KERNEL_HWCAP_ECV ) ,
2022-09-06 01:54:07 +03:00
HWCAP_CAP ( SYS_ID_AA64MMFR1_EL1 , ID_AA64MMFR1_EL1_AFP_SHIFT , 4 , FTR_UNSIGNED , 1 , CAP_HWCAP , KERNEL_HWCAP_AFP ) ,
2022-07-04 20:02:50 +03:00
HWCAP_CAP ( SYS_ID_AA64ISAR2_EL1 , ID_AA64ISAR2_EL1_RPRES_SHIFT , 4 , FTR_UNSIGNED , 1 , CAP_HWCAP , KERNEL_HWCAP_RPRES ) ,
HWCAP_CAP ( SYS_ID_AA64ISAR2_EL1 , ID_AA64ISAR2_EL1_WFxT_SHIFT , 4 , FTR_UNSIGNED , ID_AA64ISAR2_EL1_WFxT_IMP , CAP_HWCAP , KERNEL_HWCAP_WFXT ) ,
2022-04-19 14:22:16 +03:00
# ifdef CONFIG_ARM64_SME
2022-09-06 01:54:16 +03:00
HWCAP_CAP ( SYS_ID_AA64PFR1_EL1 , ID_AA64PFR1_EL1_SME_SHIFT , 4 , FTR_UNSIGNED , ID_AA64PFR1_EL1_SME_IMP , CAP_HWCAP , KERNEL_HWCAP_SME ) ,
2022-07-04 20:02:46 +03:00
HWCAP_CAP ( SYS_ID_AA64SMFR0_EL1 , ID_AA64SMFR0_EL1_FA64_SHIFT , 1 , FTR_UNSIGNED , ID_AA64SMFR0_EL1_FA64_IMP , CAP_HWCAP , KERNEL_HWCAP_SME_FA64 ) ,
HWCAP_CAP ( SYS_ID_AA64SMFR0_EL1 , ID_AA64SMFR0_EL1_I16I64_SHIFT , 4 , FTR_UNSIGNED , ID_AA64SMFR0_EL1_I16I64_IMP , CAP_HWCAP , KERNEL_HWCAP_SME_I16I64 ) ,
HWCAP_CAP ( SYS_ID_AA64SMFR0_EL1 , ID_AA64SMFR0_EL1_F64F64_SHIFT , 1 , FTR_UNSIGNED , ID_AA64SMFR0_EL1_F64F64_IMP , CAP_HWCAP , KERNEL_HWCAP_SME_F64F64 ) ,
HWCAP_CAP ( SYS_ID_AA64SMFR0_EL1 , ID_AA64SMFR0_EL1_I8I32_SHIFT , 4 , FTR_UNSIGNED , ID_AA64SMFR0_EL1_I8I32_IMP , CAP_HWCAP , KERNEL_HWCAP_SME_I8I32 ) ,
HWCAP_CAP ( SYS_ID_AA64SMFR0_EL1 , ID_AA64SMFR0_EL1_F16F32_SHIFT , 1 , FTR_UNSIGNED , ID_AA64SMFR0_EL1_F16F32_IMP , CAP_HWCAP , KERNEL_HWCAP_SME_F16F32 ) ,
HWCAP_CAP ( SYS_ID_AA64SMFR0_EL1 , ID_AA64SMFR0_EL1_B16F32_SHIFT , 1 , FTR_UNSIGNED , ID_AA64SMFR0_EL1_B16F32_IMP , CAP_HWCAP , KERNEL_HWCAP_SME_B16F32 ) ,
HWCAP_CAP ( SYS_ID_AA64SMFR0_EL1 , ID_AA64SMFR0_EL1_F32F32_SHIFT , 1 , FTR_UNSIGNED , ID_AA64SMFR0_EL1_F32F32_IMP , CAP_HWCAP , KERNEL_HWCAP_SME_F32F32 ) ,
2022-04-19 14:22:16 +03:00
# endif /* CONFIG_ARM64_SME */
2016-04-18 12:28:33 +03:00
{ } ,
} ;
2020-01-14 02:30:20 +03:00
# ifdef CONFIG_COMPAT
static bool compat_has_neon ( const struct arm64_cpu_capabilities * cap , int scope )
{
/*
* Check that all of MVFR1_EL1 . { SIMDSP , SIMDInt , SIMDLS } are available ,
* in line with that of arm32 as in vfp_init ( ) . We make sure that the
* check is future proof , by making sure value is non - zero .
*/
u32 mvfr1 ;
WARN_ON ( scope = = SCOPE_LOCAL_CPU & & preemptible ( ) ) ;
if ( scope = = SCOPE_SYSTEM )
mvfr1 = read_sanitised_ftr_reg ( SYS_MVFR1_EL1 ) ;
else
mvfr1 = read_sysreg_s ( SYS_MVFR1_EL1 ) ;
return cpuid_feature_extract_unsigned_field ( mvfr1 , MVFR1_SIMDSP_SHIFT ) & &
cpuid_feature_extract_unsigned_field ( mvfr1 , MVFR1_SIMDINT_SHIFT ) & &
cpuid_feature_extract_unsigned_field ( mvfr1 , MVFR1_SIMDLS_SHIFT ) ;
}
# endif
2016-04-18 12:28:33 +03:00
static const struct arm64_cpu_capabilities compat_elf_hwcaps [ ] = {
2015-10-19 16:24:52 +03:00
# ifdef CONFIG_COMPAT
2020-01-14 02:30:20 +03:00
HWCAP_CAP_MATCH ( compat_has_neon , CAP_COMPAT_HWCAP , COMPAT_HWCAP_NEON ) ,
2022-02-07 18:20:32 +03:00
HWCAP_CAP ( SYS_MVFR1_EL1 , MVFR1_SIMDFMAC_SHIFT , 4 , FTR_UNSIGNED , 1 , CAP_COMPAT_HWCAP , COMPAT_HWCAP_VFPv4 ) ,
2020-01-14 02:30:20 +03:00
/* Arm v8 mandates MVFR0.FPDP == {0, 2}. So, piggy back on this for the presence of VFP support */
2022-02-07 18:20:32 +03:00
HWCAP_CAP ( SYS_MVFR0_EL1 , MVFR0_FPDP_SHIFT , 4 , FTR_UNSIGNED , 2 , CAP_COMPAT_HWCAP , COMPAT_HWCAP_VFP ) ,
HWCAP_CAP ( SYS_MVFR0_EL1 , MVFR0_FPDP_SHIFT , 4 , FTR_UNSIGNED , 2 , CAP_COMPAT_HWCAP , COMPAT_HWCAP_VFPv3 ) ,
HWCAP_CAP ( SYS_ID_ISAR5_EL1 , ID_ISAR5_AES_SHIFT , 4 , FTR_UNSIGNED , 2 , CAP_COMPAT_HWCAP2 , COMPAT_HWCAP2_PMULL ) ,
HWCAP_CAP ( SYS_ID_ISAR5_EL1 , ID_ISAR5_AES_SHIFT , 4 , FTR_UNSIGNED , 1 , CAP_COMPAT_HWCAP2 , COMPAT_HWCAP2_AES ) ,
HWCAP_CAP ( SYS_ID_ISAR5_EL1 , ID_ISAR5_SHA1_SHIFT , 4 , FTR_UNSIGNED , 1 , CAP_COMPAT_HWCAP2 , COMPAT_HWCAP2_SHA1 ) ,
HWCAP_CAP ( SYS_ID_ISAR5_EL1 , ID_ISAR5_SHA2_SHIFT , 4 , FTR_UNSIGNED , 1 , CAP_COMPAT_HWCAP2 , COMPAT_HWCAP2_SHA2 ) ,
HWCAP_CAP ( SYS_ID_ISAR5_EL1 , ID_ISAR5_CRC32_SHIFT , 4 , FTR_UNSIGNED , 1 , CAP_COMPAT_HWCAP2 , COMPAT_HWCAP2_CRC32 ) ,
2015-10-19 16:24:52 +03:00
# endif
{ } ,
} ;
2021-06-08 21:02:55 +03:00
static void cap_set_elf_hwcap ( const struct arm64_cpu_capabilities * cap )
2015-10-19 16:24:52 +03:00
{
switch ( cap - > hwcap_type ) {
case CAP_HWCAP :
2019-04-09 12:52:40 +03:00
cpu_set_feature ( cap - > hwcap ) ;
2015-10-19 16:24:52 +03:00
break ;
# ifdef CONFIG_COMPAT
case CAP_COMPAT_HWCAP :
compat_elf_hwcap | = ( u32 ) cap - > hwcap ;
break ;
case CAP_COMPAT_HWCAP2 :
compat_elf_hwcap2 | = ( u32 ) cap - > hwcap ;
break ;
# endif
default :
WARN_ON ( 1 ) ;
break ;
}
}
/* Check if we have a particular HWCAP enabled */
2016-04-18 12:28:32 +03:00
static bool cpus_have_elf_hwcap ( const struct arm64_cpu_capabilities * cap )
2015-10-19 16:24:52 +03:00
{
bool rc ;
switch ( cap - > hwcap_type ) {
case CAP_HWCAP :
2019-04-09 12:52:40 +03:00
rc = cpu_have_feature ( cap - > hwcap ) ;
2015-10-19 16:24:52 +03:00
break ;
# ifdef CONFIG_COMPAT
case CAP_COMPAT_HWCAP :
rc = ( compat_elf_hwcap & ( u32 ) cap - > hwcap ) ! = 0 ;
break ;
case CAP_COMPAT_HWCAP2 :
rc = ( compat_elf_hwcap2 & ( u32 ) cap - > hwcap ) ! = 0 ;
break ;
# endif
default :
WARN_ON ( 1 ) ;
rc = false ;
}
return rc ;
}
2021-06-08 21:02:55 +03:00
static void setup_elf_hwcaps ( const struct arm64_cpu_capabilities * hwcaps )
2015-10-19 16:24:52 +03:00
{
2017-01-09 20:28:31 +03:00
/* We support emulation of accesses to CPU ID feature registers */
2019-04-09 12:52:40 +03:00
cpu_set_named_feature ( CPUID ) ;
2016-04-18 12:28:33 +03:00
for ( ; hwcaps - > matches ; hwcaps + + )
arm64: capabilities: Prepare for fine grained capabilities
We use arm64_cpu_capabilities to represent CPU ELF HWCAPs exposed
to the userspace and the CPU hwcaps used by the kernel, which
include cpu features and CPU errata work arounds. Capabilities
have some properties that decide how they should be treated :
1) Detection, i.e scope : A cap could be "detected" either :
- if it is present on at least one CPU (SCOPE_LOCAL_CPU)
Or
- if it is present on all the CPUs (SCOPE_SYSTEM)
2) When is it enabled ? - A cap is treated as "enabled" when the
system takes some action based on whether the capability is detected or
not. e.g, setting some control register, patching the kernel code.
Right now, we treat all caps are enabled at boot-time, after all
the CPUs are brought up by the kernel. But there are certain caps,
which are enabled early during the boot (e.g, VHE, GIC_CPUIF for NMI)
and kernel starts using them, even before the secondary CPUs are brought
up. We would need a way to describe this for each capability.
3) Conflict on a late CPU - When a CPU is brought up, it is checked
against the caps that are known to be enabled on the system (via
verify_local_cpu_capabilities()). Based on the state of the capability
on the CPU vs. that of System we could have the following combinations
of conflict.
x-----------------------------x
| Type | System | Late CPU |
------------------------------|
| a | y | n |
------------------------------|
| b | n | y |
x-----------------------------x
Case (a) is not permitted for caps which are system features, which the
system expects all the CPUs to have (e.g VHE). While (a) is ignored for
all errata work arounds. However, there could be exceptions to the plain
filtering approach. e.g, KPTI is an optional feature for a late CPU as
long as the system already enables it.
Case (b) is not permitted for errata work arounds which requires some
work around, which cannot be delayed. And we ignore (b) for features.
Here, yet again, KPTI is an exception, where if a late CPU needs KPTI we
are too late to enable it (because we change the allocation of ASIDs
etc).
So this calls for a lot more fine grained behavior for each capability.
And if we define all the attributes to control their behavior properly,
we may be able to use a single table for the CPU hwcaps (which cover
errata and features, not the ELF HWCAPs). This is a prepartory step
to get there. More bits would be added for the properties listed above.
We are going to use a bit-mask to encode all the properties of a
capabilities. This patch encodes the "SCOPE" of the capability.
As such there is no change in how the capabilities are treated.
Cc: Mark Rutland <mark.rutland@arm.com>
Reviewed-by: Dave Martin <dave.martin@arm.com>
Signed-off-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
2018-03-26 17:12:31 +03:00
if ( hwcaps - > matches ( hwcaps , cpucap_default_scope ( hwcaps ) ) )
2016-04-18 12:28:33 +03:00
cap_set_elf_hwcap ( hwcaps ) ;
2015-10-19 16:24:52 +03:00
}
2018-11-30 20:18:05 +03:00
static void update_cpu_capabilities ( u16 scope_mask )
2018-01-09 19:12:18 +03:00
{
2018-11-30 20:18:05 +03:00
int i ;
2018-01-09 19:12:18 +03:00
const struct arm64_cpu_capabilities * caps ;
2018-03-26 17:12:34 +03:00
scope_mask & = ARM64_CPUCAP_SCOPE_MASK ;
2018-11-30 20:18:05 +03:00
for ( i = 0 ; i < ARM64_NCAPS ; i + + ) {
caps = cpu_hwcaps_ptrs [ i ] ;
if ( ! caps | | ! ( caps - > type & scope_mask ) | |
cpus_have_cap ( caps - > capability ) | |
2018-03-26 17:12:34 +03:00
! caps - > matches ( caps , cpucap_default_scope ( caps ) ) )
2015-03-27 16:09:23 +03:00
continue ;
2018-11-30 20:18:05 +03:00
if ( caps - > desc )
pr_info ( " detected: %s \n " , caps - > desc ) ;
2016-04-18 12:28:33 +03:00
cpus_set_cap ( caps - > capability ) ;
2019-01-31 17:58:53 +03:00
if ( ( scope_mask & SCOPE_BOOT_CPU ) & & ( caps - > type & SCOPE_BOOT_CPU ) )
set_bit ( caps - > capability , boot_capabilities ) ;
2015-03-27 16:09:23 +03:00
}
2015-10-19 16:24:49 +03:00
}
2018-11-30 20:18:06 +03:00
/*
* Enable all the available capabilities on this CPU . The capabilities
* with BOOT_CPU scope are handled separately and hence skipped here .
*/
static int cpu_enable_non_boot_scope_capabilities ( void * __unused )
2018-03-26 17:12:38 +03:00
{
2018-11-30 20:18:06 +03:00
int i ;
u16 non_boot_scope = SCOPE_ALL & ~ SCOPE_BOOT_CPU ;
2018-03-26 17:12:38 +03:00
2018-11-30 20:18:06 +03:00
for_each_available_cap ( i ) {
const struct arm64_cpu_capabilities * cap = cpu_hwcaps_ptrs [ i ] ;
if ( WARN_ON ( ! cap ) )
continue ;
2018-03-26 17:12:28 +03:00
2018-11-30 20:18:06 +03:00
if ( ! ( cap - > type & non_boot_scope ) )
continue ;
if ( cap - > cpu_enable )
cap - > cpu_enable ( cap ) ;
}
2018-03-26 17:12:28 +03:00
return 0 ;
}
2015-10-19 16:24:49 +03:00
/*
arm64: Delay cpu feature capability checks
At the moment we run through the arm64_features capability list for
each CPU and set the capability if one of the CPU supports it. This
could be problematic in a heterogeneous system with differing capabilities.
Delay the CPU feature checks until all the enabled CPUs are up(i.e,
smp_cpus_done(), so that we can make better decisions based on the
overall system capability. Once we decide and advertise the capabilities
the alternatives can be applied. From this state, we cannot roll back
a feature to disabled based on the values from a new hotplugged CPU,
due to the runtime patching and other reasons. So, for all new CPUs,
we need to make sure that they have the established system capabilities.
Failing which, we bring the CPU down, preventing it from turning online.
Once the capabilities are decided, any new CPU booting up goes through
verification to ensure that it has all the enabled capabilities and also
invokes the respective enable() method on the CPU.
The CPU errata checks are not delayed and is still executed per-CPU
to detect the respective capabilities. If we ever come across a non-errata
capability that needs to be checked on each-CPU, we could introduce them via
a new capability table(or introduce a flag), which can be processed per CPU.
The next patch will make the feature checks use the system wide
safe value of a feature register.
NOTE: The enable() methods associated with the capability is scheduled
on all the CPUs (which is the only use case at the moment). If we need
a different type of 'enable()' which only needs to be run once on any CPU,
we should be able to handle that when needed.
Signed-off-by: Suzuki K. Poulose <suzuki.poulose@arm.com>
Tested-by: Dave Martin <Dave.Martin@arm.com>
[catalin.marinas@arm.com: static variable and coding style fixes]
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2015-10-19 16:24:50 +03:00
* Run through the enabled capabilities and enable ( ) it on all active
* CPUs
2015-10-19 16:24:49 +03:00
*/
2018-11-30 20:18:06 +03:00
static void __init enable_cpu_capabilities ( u16 scope_mask )
2015-10-19 16:24:49 +03:00
{
2018-11-30 20:18:06 +03:00
int i ;
const struct arm64_cpu_capabilities * caps ;
bool boot_scope ;
2018-03-26 17:12:34 +03:00
scope_mask & = ARM64_CPUCAP_SCOPE_MASK ;
2018-11-30 20:18:06 +03:00
boot_scope = ! ! ( scope_mask & SCOPE_BOOT_CPU ) ;
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 <mark.rutland@arm.com>
Reviewed-by: Marc Zyniger <marc.zyngier@arm.com>
Reviewed-by: Suzuki Poulose <suzuki.poulose@arm.com>
Acked-by: Will Deacon <will.deacon@arm.com>
Cc: Christoffer Dall <christoffer.dall@linaro.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Sebastian Sewior <bigeasy@linutronix.de>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2017-05-16 17:18:05 +03:00
2018-11-30 20:18:06 +03:00
for ( i = 0 ; i < ARM64_NCAPS ; i + + ) {
unsigned int num ;
caps = cpu_hwcaps_ptrs [ i ] ;
if ( ! caps | | ! ( caps - > type & scope_mask ) )
continue ;
num = caps - > capability ;
if ( ! cpus_have_cap ( num ) )
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 <mark.rutland@arm.com>
Reviewed-by: Marc Zyniger <marc.zyngier@arm.com>
Reviewed-by: Suzuki Poulose <suzuki.poulose@arm.com>
Acked-by: Will Deacon <will.deacon@arm.com>
Cc: Christoffer Dall <christoffer.dall@linaro.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Sebastian Sewior <bigeasy@linutronix.de>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2017-05-16 17:18:05 +03:00
continue ;
2018-11-30 20:18:06 +03:00
if ( boot_scope & & caps - > cpu_enable )
2016-10-18 13:27:46 +03:00
/*
2018-03-26 17:12:41 +03:00
* Capabilities with SCOPE_BOOT_CPU scope are finalised
* before any secondary CPU boots . Thus , each secondary
* will enable the capability as appropriate via
* check_local_cpu_capabilities ( ) . The only exception is
* the boot CPU , for which the capability must be
* enabled here . This approach avoids costly
* stop_machine ( ) calls for this case .
2016-10-18 13:27:46 +03:00
*/
2018-11-30 20:18:06 +03:00
caps - > cpu_enable ( caps ) ;
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 <mark.rutland@arm.com>
Reviewed-by: Marc Zyniger <marc.zyngier@arm.com>
Reviewed-by: Suzuki Poulose <suzuki.poulose@arm.com>
Acked-by: Will Deacon <will.deacon@arm.com>
Cc: Christoffer Dall <christoffer.dall@linaro.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Sebastian Sewior <bigeasy@linutronix.de>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2017-05-16 17:18:05 +03:00
}
arm64: Delay cpu feature capability checks
At the moment we run through the arm64_features capability list for
each CPU and set the capability if one of the CPU supports it. This
could be problematic in a heterogeneous system with differing capabilities.
Delay the CPU feature checks until all the enabled CPUs are up(i.e,
smp_cpus_done(), so that we can make better decisions based on the
overall system capability. Once we decide and advertise the capabilities
the alternatives can be applied. From this state, we cannot roll back
a feature to disabled based on the values from a new hotplugged CPU,
due to the runtime patching and other reasons. So, for all new CPUs,
we need to make sure that they have the established system capabilities.
Failing which, we bring the CPU down, preventing it from turning online.
Once the capabilities are decided, any new CPU booting up goes through
verification to ensure that it has all the enabled capabilities and also
invokes the respective enable() method on the CPU.
The CPU errata checks are not delayed and is still executed per-CPU
to detect the respective capabilities. If we ever come across a non-errata
capability that needs to be checked on each-CPU, we could introduce them via
a new capability table(or introduce a flag), which can be processed per CPU.
The next patch will make the feature checks use the system wide
safe value of a feature register.
NOTE: The enable() methods associated with the capability is scheduled
on all the CPUs (which is the only use case at the moment). If we need
a different type of 'enable()' which only needs to be run once on any CPU,
we should be able to handle that when needed.
Signed-off-by: Suzuki K. Poulose <suzuki.poulose@arm.com>
Tested-by: Dave Martin <Dave.Martin@arm.com>
[catalin.marinas@arm.com: static variable and coding style fixes]
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2015-10-19 16:24:50 +03:00
2018-11-30 20:18:06 +03:00
/*
* For all non - boot scope capabilities , use stop_machine ( )
* as it schedules the work allowing us to modify PSTATE ,
* instead of on_each_cpu ( ) which uses an IPI , giving us a
* PSTATE that disappears when we return .
*/
if ( ! boot_scope )
stop_machine ( cpu_enable_non_boot_scope_capabilities ,
NULL , cpu_online_mask ) ;
2018-03-26 17:12:38 +03:00
}
2018-03-26 17:12:33 +03:00
/*
* Run through the list of capabilities to check for conflicts .
* If the system has already detected a capability , take necessary
* action on this CPU .
*/
2020-03-13 12:04:54 +03:00
static void verify_local_cpu_caps ( u16 scope_mask )
2018-03-26 17:12:33 +03:00
{
2018-11-30 20:18:05 +03:00
int i ;
2018-03-26 17:12:33 +03:00
bool cpu_has_cap , system_has_cap ;
2018-11-30 20:18:05 +03:00
const struct arm64_cpu_capabilities * caps ;
2018-03-26 17:12:33 +03:00
2018-03-26 17:12:34 +03:00
scope_mask & = ARM64_CPUCAP_SCOPE_MASK ;
2018-11-30 20:18:05 +03:00
for ( i = 0 ; i < ARM64_NCAPS ; i + + ) {
caps = cpu_hwcaps_ptrs [ i ] ;
if ( ! caps | | ! ( caps - > type & scope_mask ) )
2018-03-26 17:12:34 +03:00
continue ;
arm64: capabilities: Handle shared entries
Some capabilities have different criteria for detection and associated
actions based on the matching criteria, even though they all share the
same capability bit. So far we have used multiple entries with the same
capability bit to handle this. This is prone to errors, as the
cpu_enable is invoked for each entry, irrespective of whether the
detection rule applies to the CPU or not. And also this complicates
other helpers, e.g, __this_cpu_has_cap.
This patch adds a wrapper entry to cover all the possible variations
of a capability by maintaining list of matches + cpu_enable callbacks.
To avoid complicating the prototypes for the "matches()", we use
arm64_cpu_capabilities maintain the list and we ignore all the other
fields except the matches & cpu_enable.
This ensures :
1) The capabilitiy is set when at least one of the entry detects
2) Action is only taken for the entries that "matches".
This avoids explicit checks in the cpu_enable() take some action.
The only constraint here is that, all the entries should have the
same "type" (i.e, scope and conflict rules).
If a cpu_enable() method is associated with multiple matches for a
single capability, care should be taken that either the match criteria
are mutually exclusive, or that the method is robust against being
called multiple times.
This also reverts the changes introduced by commit 67948af41f2e6818ed
("arm64: capabilities: Handle duplicate entries for a capability").
Cc: Robin Murphy <robin.murphy@arm.com>
Reviewed-by: Dave Martin <dave.martin@arm.com>
Signed-off-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
2018-03-26 17:12:46 +03:00
cpu_has_cap = caps - > matches ( caps , SCOPE_LOCAL_CPU ) ;
2018-03-26 17:12:33 +03:00
system_has_cap = cpus_have_cap ( caps - > capability ) ;
if ( system_has_cap ) {
/*
* Check if the new CPU misses an advertised feature ,
* which is not safe to miss .
*/
if ( ! cpu_has_cap & & ! cpucap_late_cpu_optional ( caps ) )
break ;
/*
* We have to issue cpu_enable ( ) irrespective of
* whether the CPU has it or not , as it is enabeld
* system wide . It is upto the call back to take
* appropriate action on this CPU .
*/
if ( caps - > cpu_enable )
caps - > cpu_enable ( caps ) ;
} else {
/*
* Check if the CPU has this capability if it isn ' t
* safe to have when the system doesn ' t .
*/
if ( cpu_has_cap & & ! cpucap_late_cpu_permitted ( caps ) )
break ;
}
}
2018-11-30 20:18:05 +03:00
if ( i < ARM64_NCAPS ) {
2018-03-26 17:12:33 +03:00
pr_crit ( " CPU%d: Detected conflict for capability %d (%s), System: %d, CPU: %d \n " ,
smp_processor_id ( ) , caps - > capability ,
caps - > desc , system_has_cap , cpu_has_cap ) ;
2020-03-13 12:04:54 +03:00
if ( cpucap_panic_on_conflict ( caps ) )
cpu_panic_kernel ( ) ;
else
cpu_die_early ( ) ;
}
2018-03-26 17:12:33 +03:00
}
arm64: Delay cpu feature capability checks
At the moment we run through the arm64_features capability list for
each CPU and set the capability if one of the CPU supports it. This
could be problematic in a heterogeneous system with differing capabilities.
Delay the CPU feature checks until all the enabled CPUs are up(i.e,
smp_cpus_done(), so that we can make better decisions based on the
overall system capability. Once we decide and advertise the capabilities
the alternatives can be applied. From this state, we cannot roll back
a feature to disabled based on the values from a new hotplugged CPU,
due to the runtime patching and other reasons. So, for all new CPUs,
we need to make sure that they have the established system capabilities.
Failing which, we bring the CPU down, preventing it from turning online.
Once the capabilities are decided, any new CPU booting up goes through
verification to ensure that it has all the enabled capabilities and also
invokes the respective enable() method on the CPU.
The CPU errata checks are not delayed and is still executed per-CPU
to detect the respective capabilities. If we ever come across a non-errata
capability that needs to be checked on each-CPU, we could introduce them via
a new capability table(or introduce a flag), which can be processed per CPU.
The next patch will make the feature checks use the system wide
safe value of a feature register.
NOTE: The enable() methods associated with the capability is scheduled
on all the CPUs (which is the only use case at the moment). If we need
a different type of 'enable()' which only needs to be run once on any CPU,
we should be able to handle that when needed.
Signed-off-by: Suzuki K. Poulose <suzuki.poulose@arm.com>
Tested-by: Dave Martin <Dave.Martin@arm.com>
[catalin.marinas@arm.com: static variable and coding style fixes]
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2015-10-19 16:24:50 +03:00
/*
2016-02-23 13:31:45 +03:00
* Check for CPU features that are used in early boot
* based on the Boot CPU value .
arm64: Delay cpu feature capability checks
At the moment we run through the arm64_features capability list for
each CPU and set the capability if one of the CPU supports it. This
could be problematic in a heterogeneous system with differing capabilities.
Delay the CPU feature checks until all the enabled CPUs are up(i.e,
smp_cpus_done(), so that we can make better decisions based on the
overall system capability. Once we decide and advertise the capabilities
the alternatives can be applied. From this state, we cannot roll back
a feature to disabled based on the values from a new hotplugged CPU,
due to the runtime patching and other reasons. So, for all new CPUs,
we need to make sure that they have the established system capabilities.
Failing which, we bring the CPU down, preventing it from turning online.
Once the capabilities are decided, any new CPU booting up goes through
verification to ensure that it has all the enabled capabilities and also
invokes the respective enable() method on the CPU.
The CPU errata checks are not delayed and is still executed per-CPU
to detect the respective capabilities. If we ever come across a non-errata
capability that needs to be checked on each-CPU, we could introduce them via
a new capability table(or introduce a flag), which can be processed per CPU.
The next patch will make the feature checks use the system wide
safe value of a feature register.
NOTE: The enable() methods associated with the capability is scheduled
on all the CPUs (which is the only use case at the moment). If we need
a different type of 'enable()' which only needs to be run once on any CPU,
we should be able to handle that when needed.
Signed-off-by: Suzuki K. Poulose <suzuki.poulose@arm.com>
Tested-by: Dave Martin <Dave.Martin@arm.com>
[catalin.marinas@arm.com: static variable and coding style fixes]
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2015-10-19 16:24:50 +03:00
*/
2016-02-23 13:31:45 +03:00
static void check_early_cpu_features ( void )
arm64: Delay cpu feature capability checks
At the moment we run through the arm64_features capability list for
each CPU and set the capability if one of the CPU supports it. This
could be problematic in a heterogeneous system with differing capabilities.
Delay the CPU feature checks until all the enabled CPUs are up(i.e,
smp_cpus_done(), so that we can make better decisions based on the
overall system capability. Once we decide and advertise the capabilities
the alternatives can be applied. From this state, we cannot roll back
a feature to disabled based on the values from a new hotplugged CPU,
due to the runtime patching and other reasons. So, for all new CPUs,
we need to make sure that they have the established system capabilities.
Failing which, we bring the CPU down, preventing it from turning online.
Once the capabilities are decided, any new CPU booting up goes through
verification to ensure that it has all the enabled capabilities and also
invokes the respective enable() method on the CPU.
The CPU errata checks are not delayed and is still executed per-CPU
to detect the respective capabilities. If we ever come across a non-errata
capability that needs to be checked on each-CPU, we could introduce them via
a new capability table(or introduce a flag), which can be processed per CPU.
The next patch will make the feature checks use the system wide
safe value of a feature register.
NOTE: The enable() methods associated with the capability is scheduled
on all the CPUs (which is the only use case at the moment). If we need
a different type of 'enable()' which only needs to be run once on any CPU,
we should be able to handle that when needed.
Signed-off-by: Suzuki K. Poulose <suzuki.poulose@arm.com>
Tested-by: Dave Martin <Dave.Martin@arm.com>
[catalin.marinas@arm.com: static variable and coding style fixes]
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2015-10-19 16:24:50 +03:00
{
2016-02-23 13:31:45 +03:00
verify_cpu_asid_bits ( ) ;
2020-03-13 12:04:54 +03:00
verify_local_cpu_caps ( SCOPE_BOOT_CPU ) ;
arm64: Delay cpu feature capability checks
At the moment we run through the arm64_features capability list for
each CPU and set the capability if one of the CPU supports it. This
could be problematic in a heterogeneous system with differing capabilities.
Delay the CPU feature checks until all the enabled CPUs are up(i.e,
smp_cpus_done(), so that we can make better decisions based on the
overall system capability. Once we decide and advertise the capabilities
the alternatives can be applied. From this state, we cannot roll back
a feature to disabled based on the values from a new hotplugged CPU,
due to the runtime patching and other reasons. So, for all new CPUs,
we need to make sure that they have the established system capabilities.
Failing which, we bring the CPU down, preventing it from turning online.
Once the capabilities are decided, any new CPU booting up goes through
verification to ensure that it has all the enabled capabilities and also
invokes the respective enable() method on the CPU.
The CPU errata checks are not delayed and is still executed per-CPU
to detect the respective capabilities. If we ever come across a non-errata
capability that needs to be checked on each-CPU, we could introduce them via
a new capability table(or introduce a flag), which can be processed per CPU.
The next patch will make the feature checks use the system wide
safe value of a feature register.
NOTE: The enable() methods associated with the capability is scheduled
on all the CPUs (which is the only use case at the moment). If we need
a different type of 'enable()' which only needs to be run once on any CPU,
we should be able to handle that when needed.
Signed-off-by: Suzuki K. Poulose <suzuki.poulose@arm.com>
Tested-by: Dave Martin <Dave.Martin@arm.com>
[catalin.marinas@arm.com: static variable and coding style fixes]
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2015-10-19 16:24:50 +03:00
}
2015-07-21 15:23:28 +03:00
2016-04-18 12:28:33 +03:00
static void
2021-06-08 21:02:55 +03:00
__verify_local_elf_hwcaps ( const struct arm64_cpu_capabilities * caps )
2016-04-18 12:28:33 +03:00
{
2016-04-22 14:25:31 +03:00
for ( ; caps - > matches ; caps + + )
if ( cpus_have_elf_hwcap ( caps ) & & ! caps - > matches ( caps , SCOPE_LOCAL_CPU ) ) {
2016-04-18 12:28:33 +03:00
pr_crit ( " CPU%d: missing HWCAP: %s \n " ,
smp_processor_id ( ) , caps - > desc ) ;
cpu_die_early ( ) ;
}
}
2021-06-08 21:02:55 +03:00
static void verify_local_elf_hwcaps ( void )
{
__verify_local_elf_hwcaps ( arm64_elf_hwcaps ) ;
if ( id_aa64pfr0_32bit_el0 ( read_cpuid ( ID_AA64PFR0_EL1 ) ) )
__verify_local_elf_hwcaps ( compat_elf_hwcaps ) ;
}
2017-10-31 18:51:10 +03:00
static void verify_sve_features ( void )
{
u64 safe_zcr = read_sanitised_ftr_reg ( SYS_ZCR_EL1 ) ;
u64 zcr = read_zcr_features ( ) ;
unsigned int safe_len = safe_zcr & ZCR_ELx_LEN_MASK ;
unsigned int len = zcr & ZCR_ELx_LEN_MASK ;
2021-10-19 20:22:12 +03:00
if ( len < safe_len | | vec_verify_vq_map ( ARM64_VEC_SVE ) ) {
2018-09-28 16:39:10 +03:00
pr_crit ( " CPU%d: SVE: vector length support mismatch \n " ,
2017-10-31 18:51:10 +03:00
smp_processor_id ( ) ) ;
cpu_die_early ( ) ;
}
/* Add checks on other ZCR bits here if necessary */
}
2022-04-19 14:22:17 +03:00
static void verify_sme_features ( void )
{
u64 safe_smcr = read_sanitised_ftr_reg ( SYS_SMCR_EL1 ) ;
u64 smcr = read_smcr_features ( ) ;
unsigned int safe_len = safe_smcr & SMCR_ELx_LEN_MASK ;
unsigned int len = smcr & SMCR_ELx_LEN_MASK ;
if ( len < safe_len | | vec_verify_vq_map ( ARM64_VEC_SME ) ) {
pr_crit ( " CPU%d: SME: vector length support mismatch \n " ,
smp_processor_id ( ) ) ;
cpu_die_early ( ) ;
}
/* Add checks on other SMCR bits here if necessary */
}
2020-05-12 04:57:27 +03:00
static void verify_hyp_capabilities ( void )
{
u64 safe_mmfr1 , mmfr0 , mmfr1 ;
int parange , ipa_max ;
unsigned int safe_vmid_bits , vmid_bits ;
2021-01-04 14:38:44 +03:00
if ( ! IS_ENABLED ( CONFIG_KVM ) )
2020-05-12 04:57:27 +03:00
return ;
safe_mmfr1 = read_sanitised_ftr_reg ( SYS_ID_AA64MMFR1_EL1 ) ;
mmfr0 = read_cpuid ( ID_AA64MMFR0_EL1 ) ;
mmfr1 = read_cpuid ( ID_AA64MMFR1_EL1 ) ;
/* Verify VMID bits */
safe_vmid_bits = get_vmid_bits ( safe_mmfr1 ) ;
vmid_bits = get_vmid_bits ( mmfr1 ) ;
if ( vmid_bits < safe_vmid_bits ) {
pr_crit ( " CPU%d: VMID width mismatch \n " , smp_processor_id ( ) ) ;
cpu_die_early ( ) ;
}
/* Verify IPA range */
2020-05-13 12:03:34 +03:00
parange = cpuid_feature_extract_unsigned_field ( mmfr0 ,
2022-09-06 01:54:01 +03:00
ID_AA64MMFR0_EL1_PARANGE_SHIFT ) ;
2020-05-12 04:57:27 +03:00
ipa_max = id_aa64mmfr0_parange_to_phys_shift ( parange ) ;
if ( ipa_max < get_kvm_ipa_limit ( ) ) {
pr_crit ( " CPU%d: IPA range mismatch \n " , smp_processor_id ( ) ) ;
cpu_die_early ( ) ;
}
}
2018-03-26 17:12:30 +03:00
arm64: Delay cpu feature capability checks
At the moment we run through the arm64_features capability list for
each CPU and set the capability if one of the CPU supports it. This
could be problematic in a heterogeneous system with differing capabilities.
Delay the CPU feature checks until all the enabled CPUs are up(i.e,
smp_cpus_done(), so that we can make better decisions based on the
overall system capability. Once we decide and advertise the capabilities
the alternatives can be applied. From this state, we cannot roll back
a feature to disabled based on the values from a new hotplugged CPU,
due to the runtime patching and other reasons. So, for all new CPUs,
we need to make sure that they have the established system capabilities.
Failing which, we bring the CPU down, preventing it from turning online.
Once the capabilities are decided, any new CPU booting up goes through
verification to ensure that it has all the enabled capabilities and also
invokes the respective enable() method on the CPU.
The CPU errata checks are not delayed and is still executed per-CPU
to detect the respective capabilities. If we ever come across a non-errata
capability that needs to be checked on each-CPU, we could introduce them via
a new capability table(or introduce a flag), which can be processed per CPU.
The next patch will make the feature checks use the system wide
safe value of a feature register.
NOTE: The enable() methods associated with the capability is scheduled
on all the CPUs (which is the only use case at the moment). If we need
a different type of 'enable()' which only needs to be run once on any CPU,
we should be able to handle that when needed.
Signed-off-by: Suzuki K. Poulose <suzuki.poulose@arm.com>
Tested-by: Dave Martin <Dave.Martin@arm.com>
[catalin.marinas@arm.com: static variable and coding style fixes]
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2015-10-19 16:24:50 +03:00
/*
* Run through the enabled system capabilities and enable ( ) it on this CPU .
* The capabilities were decided based on the available CPUs at the boot time .
* Any new CPU should match the system wide status of the capability . If the
* new CPU doesn ' t have a capability which the system now has enabled , we
* cannot do anything to fix it up and could cause unexpected failures . So
* we park the CPU .
*/
2016-09-09 16:07:10 +03:00
static void verify_local_cpu_capabilities ( void )
arm64: Delay cpu feature capability checks
At the moment we run through the arm64_features capability list for
each CPU and set the capability if one of the CPU supports it. This
could be problematic in a heterogeneous system with differing capabilities.
Delay the CPU feature checks until all the enabled CPUs are up(i.e,
smp_cpus_done(), so that we can make better decisions based on the
overall system capability. Once we decide and advertise the capabilities
the alternatives can be applied. From this state, we cannot roll back
a feature to disabled based on the values from a new hotplugged CPU,
due to the runtime patching and other reasons. So, for all new CPUs,
we need to make sure that they have the established system capabilities.
Failing which, we bring the CPU down, preventing it from turning online.
Once the capabilities are decided, any new CPU booting up goes through
verification to ensure that it has all the enabled capabilities and also
invokes the respective enable() method on the CPU.
The CPU errata checks are not delayed and is still executed per-CPU
to detect the respective capabilities. If we ever come across a non-errata
capability that needs to be checked on each-CPU, we could introduce them via
a new capability table(or introduce a flag), which can be processed per CPU.
The next patch will make the feature checks use the system wide
safe value of a feature register.
NOTE: The enable() methods associated with the capability is scheduled
on all the CPUs (which is the only use case at the moment). If we need
a different type of 'enable()' which only needs to be run once on any CPU,
we should be able to handle that when needed.
Signed-off-by: Suzuki K. Poulose <suzuki.poulose@arm.com>
Tested-by: Dave Martin <Dave.Martin@arm.com>
[catalin.marinas@arm.com: static variable and coding style fixes]
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2015-10-19 16:24:50 +03:00
{
2018-03-26 17:12:41 +03:00
/*
* The capabilities with SCOPE_BOOT_CPU are checked from
* check_early_cpu_features ( ) , as they need to be verified
* on all secondary CPUs .
*/
2020-03-13 12:04:54 +03:00
verify_local_cpu_caps ( SCOPE_ALL & ~ SCOPE_BOOT_CPU ) ;
2021-06-08 21:02:55 +03:00
verify_local_elf_hwcaps ( ) ;
2017-10-31 18:51:10 +03:00
if ( system_supports_sve ( ) )
verify_sve_features ( ) ;
2020-05-12 04:57:27 +03:00
2022-04-19 14:22:17 +03:00
if ( system_supports_sme ( ) )
verify_sme_features ( ) ;
2020-05-12 04:57:27 +03:00
if ( is_hyp_mode_available ( ) )
verify_hyp_capabilities ( ) ;
2016-09-09 16:07:10 +03:00
}
arm64: Delay cpu feature capability checks
At the moment we run through the arm64_features capability list for
each CPU and set the capability if one of the CPU supports it. This
could be problematic in a heterogeneous system with differing capabilities.
Delay the CPU feature checks until all the enabled CPUs are up(i.e,
smp_cpus_done(), so that we can make better decisions based on the
overall system capability. Once we decide and advertise the capabilities
the alternatives can be applied. From this state, we cannot roll back
a feature to disabled based on the values from a new hotplugged CPU,
due to the runtime patching and other reasons. So, for all new CPUs,
we need to make sure that they have the established system capabilities.
Failing which, we bring the CPU down, preventing it from turning online.
Once the capabilities are decided, any new CPU booting up goes through
verification to ensure that it has all the enabled capabilities and also
invokes the respective enable() method on the CPU.
The CPU errata checks are not delayed and is still executed per-CPU
to detect the respective capabilities. If we ever come across a non-errata
capability that needs to be checked on each-CPU, we could introduce them via
a new capability table(or introduce a flag), which can be processed per CPU.
The next patch will make the feature checks use the system wide
safe value of a feature register.
NOTE: The enable() methods associated with the capability is scheduled
on all the CPUs (which is the only use case at the moment). If we need
a different type of 'enable()' which only needs to be run once on any CPU,
we should be able to handle that when needed.
Signed-off-by: Suzuki K. Poulose <suzuki.poulose@arm.com>
Tested-by: Dave Martin <Dave.Martin@arm.com>
[catalin.marinas@arm.com: static variable and coding style fixes]
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2015-10-19 16:24:50 +03:00
2016-09-09 16:07:10 +03:00
void check_local_cpu_capabilities ( void )
{
/*
* All secondary CPUs should conform to the early CPU features
* in use by the kernel based on boot CPU .
*/
2016-02-23 13:31:45 +03:00
check_early_cpu_features ( ) ;
arm64: Delay cpu feature capability checks
At the moment we run through the arm64_features capability list for
each CPU and set the capability if one of the CPU supports it. This
could be problematic in a heterogeneous system with differing capabilities.
Delay the CPU feature checks until all the enabled CPUs are up(i.e,
smp_cpus_done(), so that we can make better decisions based on the
overall system capability. Once we decide and advertise the capabilities
the alternatives can be applied. From this state, we cannot roll back
a feature to disabled based on the values from a new hotplugged CPU,
due to the runtime patching and other reasons. So, for all new CPUs,
we need to make sure that they have the established system capabilities.
Failing which, we bring the CPU down, preventing it from turning online.
Once the capabilities are decided, any new CPU booting up goes through
verification to ensure that it has all the enabled capabilities and also
invokes the respective enable() method on the CPU.
The CPU errata checks are not delayed and is still executed per-CPU
to detect the respective capabilities. If we ever come across a non-errata
capability that needs to be checked on each-CPU, we could introduce them via
a new capability table(or introduce a flag), which can be processed per CPU.
The next patch will make the feature checks use the system wide
safe value of a feature register.
NOTE: The enable() methods associated with the capability is scheduled
on all the CPUs (which is the only use case at the moment). If we need
a different type of 'enable()' which only needs to be run once on any CPU,
we should be able to handle that when needed.
Signed-off-by: Suzuki K. Poulose <suzuki.poulose@arm.com>
Tested-by: Dave Martin <Dave.Martin@arm.com>
[catalin.marinas@arm.com: static variable and coding style fixes]
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2015-10-19 16:24:50 +03:00
/*
2016-09-09 16:07:10 +03:00
* If we haven ' t finalised the system capabilities , this CPU gets
2018-03-26 17:12:37 +03:00
* a chance to update the errata work arounds and local features .
2016-09-09 16:07:10 +03:00
* Otherwise , this CPU should verify that it has all the system
* advertised capabilities .
arm64: Delay cpu feature capability checks
At the moment we run through the arm64_features capability list for
each CPU and set the capability if one of the CPU supports it. This
could be problematic in a heterogeneous system with differing capabilities.
Delay the CPU feature checks until all the enabled CPUs are up(i.e,
smp_cpus_done(), so that we can make better decisions based on the
overall system capability. Once we decide and advertise the capabilities
the alternatives can be applied. From this state, we cannot roll back
a feature to disabled based on the values from a new hotplugged CPU,
due to the runtime patching and other reasons. So, for all new CPUs,
we need to make sure that they have the established system capabilities.
Failing which, we bring the CPU down, preventing it from turning online.
Once the capabilities are decided, any new CPU booting up goes through
verification to ensure that it has all the enabled capabilities and also
invokes the respective enable() method on the CPU.
The CPU errata checks are not delayed and is still executed per-CPU
to detect the respective capabilities. If we ever come across a non-errata
capability that needs to be checked on each-CPU, we could introduce them via
a new capability table(or introduce a flag), which can be processed per CPU.
The next patch will make the feature checks use the system wide
safe value of a feature register.
NOTE: The enable() methods associated with the capability is scheduled
on all the CPUs (which is the only use case at the moment). If we need
a different type of 'enable()' which only needs to be run once on any CPU,
we should be able to handle that when needed.
Signed-off-by: Suzuki K. Poulose <suzuki.poulose@arm.com>
Tested-by: Dave Martin <Dave.Martin@arm.com>
[catalin.marinas@arm.com: static variable and coding style fixes]
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2015-10-19 16:24:50 +03:00
*/
2020-01-14 02:30:17 +03:00
if ( ! system_capabilities_finalized ( ) )
2018-03-26 17:12:38 +03:00
update_cpu_capabilities ( SCOPE_LOCAL_CPU ) ;
else
2016-09-09 16:07:10 +03:00
verify_local_cpu_capabilities ( ) ;
2015-03-27 16:09:23 +03:00
}
2018-03-26 17:12:41 +03:00
static void __init setup_boot_cpu_capabilities ( void )
{
/* Detect capabilities with either SCOPE_BOOT_CPU or SCOPE_LOCAL_CPU */
update_cpu_capabilities ( SCOPE_BOOT_CPU | SCOPE_LOCAL_CPU ) ;
/* Enable the SCOPE_BOOT_CPU capabilities alone right away */
enable_cpu_capabilities ( SCOPE_BOOT_CPU ) ;
}
2018-11-30 20:18:04 +03:00
bool this_cpu_has_cap ( unsigned int n )
2017-01-30 18:39:52 +03:00
{
2018-11-30 20:18:04 +03:00
if ( ! WARN_ON ( preemptible ( ) ) & & n < ARM64_NCAPS ) {
const struct arm64_cpu_capabilities * cap = cpu_hwcaps_ptrs [ n ] ;
if ( cap )
return cap - > matches ( cap , SCOPE_LOCAL_CPU ) ;
}
return false ;
2017-01-30 18:39:52 +03:00
}
2021-11-04 01:12:56 +03:00
EXPORT_SYMBOL_GPL ( this_cpu_has_cap ) ;
2017-01-30 18:39:52 +03:00
2020-03-13 12:04:48 +03:00
/*
* This helper function is used in a narrow window when ,
* - The system wide safe registers are set with all the SMP CPUs and ,
* - The SYSTEM_FEATURE cpu_hwcaps may not have been set .
* In all other cases cpus_have_ { const_ } cap ( ) should be used .
*/
2020-12-03 18:24:03 +03:00
static bool __maybe_unused __system_matches_cap ( unsigned int n )
2020-03-13 12:04:48 +03:00
{
if ( n < ARM64_NCAPS ) {
const struct arm64_cpu_capabilities * cap = cpu_hwcaps_ptrs [ n ] ;
if ( cap )
return cap - > matches ( cap , SCOPE_SYSTEM ) ;
}
return false ;
}
2019-04-09 12:52:41 +03:00
void cpu_set_feature ( unsigned int num )
{
2022-07-07 13:36:31 +03:00
set_bit ( num , elf_hwcap ) ;
2019-04-09 12:52:41 +03:00
}
bool cpu_have_feature ( unsigned int num )
{
2022-07-07 13:36:31 +03:00
return test_bit ( num , elf_hwcap ) ;
2019-04-09 12:52:41 +03:00
}
EXPORT_SYMBOL_GPL ( cpu_have_feature ) ;
unsigned long cpu_get_elf_hwcap ( void )
{
/*
* We currently only populate the first 32 bits of AT_HWCAP . Please
* note that for userspace compatibility we guarantee that bits 62
* and 63 will always be returned as 0.
*/
2022-07-07 13:36:31 +03:00
return elf_hwcap [ 0 ] ;
2019-04-09 12:52:41 +03:00
}
unsigned long cpu_get_elf_hwcap2 ( void )
{
2022-07-07 13:36:31 +03:00
return elf_hwcap [ 1 ] ;
2019-04-09 12:52:41 +03:00
}
2018-03-26 17:12:38 +03:00
static void __init setup_system_capabilities ( void )
{
/*
* We have finalised the system - wide safe feature
* registers , finalise the capabilities that depend
2018-03-26 17:12:41 +03:00
* on it . Also enable all the available capabilities ,
* that are not enabled already .
2018-03-26 17:12:38 +03:00
*/
update_cpu_capabilities ( SCOPE_SYSTEM ) ;
2018-03-26 17:12:41 +03:00
enable_cpu_capabilities ( SCOPE_ALL & ~ SCOPE_BOOT_CPU ) ;
2018-03-26 17:12:38 +03:00
}
2015-10-19 16:24:41 +03:00
void __init setup_cpu_features ( void )
2015-03-27 16:09:23 +03:00
{
2015-10-19 16:24:41 +03:00
u32 cwg ;
2018-03-26 17:12:38 +03:00
setup_system_capabilities ( ) ;
2016-04-18 12:28:33 +03:00
setup_elf_hwcaps ( arm64_elf_hwcaps ) ;
2016-04-18 12:28:37 +03:00
2022-07-14 19:15:23 +03:00
if ( system_supports_32bit_el0 ( ) ) {
2016-04-18 12:28:37 +03:00
setup_elf_hwcaps ( compat_elf_hwcaps ) ;
2022-07-14 19:15:23 +03:00
elf_hwcap_fixup ( ) ;
}
arm64: Delay cpu feature capability checks
At the moment we run through the arm64_features capability list for
each CPU and set the capability if one of the CPU supports it. This
could be problematic in a heterogeneous system with differing capabilities.
Delay the CPU feature checks until all the enabled CPUs are up(i.e,
smp_cpus_done(), so that we can make better decisions based on the
overall system capability. Once we decide and advertise the capabilities
the alternatives can be applied. From this state, we cannot roll back
a feature to disabled based on the values from a new hotplugged CPU,
due to the runtime patching and other reasons. So, for all new CPUs,
we need to make sure that they have the established system capabilities.
Failing which, we bring the CPU down, preventing it from turning online.
Once the capabilities are decided, any new CPU booting up goes through
verification to ensure that it has all the enabled capabilities and also
invokes the respective enable() method on the CPU.
The CPU errata checks are not delayed and is still executed per-CPU
to detect the respective capabilities. If we ever come across a non-errata
capability that needs to be checked on each-CPU, we could introduce them via
a new capability table(or introduce a flag), which can be processed per CPU.
The next patch will make the feature checks use the system wide
safe value of a feature register.
NOTE: The enable() methods associated with the capability is scheduled
on all the CPUs (which is the only use case at the moment). If we need
a different type of 'enable()' which only needs to be run once on any CPU,
we should be able to handle that when needed.
Signed-off-by: Suzuki K. Poulose <suzuki.poulose@arm.com>
Tested-by: Dave Martin <Dave.Martin@arm.com>
[catalin.marinas@arm.com: static variable and coding style fixes]
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2015-10-19 16:24:50 +03:00
2018-02-21 21:18:21 +03:00
if ( system_uses_ttbr0_pan ( ) )
pr_info ( " emulated: Privileged Access Never (PAN) using TTBR0_EL1 switching \n " ) ;
2017-10-31 18:51:10 +03:00
sve_setup ( ) ;
2022-04-19 14:22:17 +03:00
sme_setup ( ) ;
arm64: signal: Report signal frame size to userspace via auxv
Stateful CPU architecture extensions may require the signal frame
to grow to a size that exceeds the arch's MINSIGSTKSZ #define.
However, changing this #define is an ABI break.
To allow userspace the option of determining the signal frame size
in a more forwards-compatible way, this patch adds a new auxv entry
tagged with AT_MINSIGSTKSZ, which provides the maximum signal frame
size that the process can observe during its lifetime.
If AT_MINSIGSTKSZ is absent from the aux vector, the caller can
assume that the MINSIGSTKSZ #define is sufficient. This allows for
a consistent interface with older kernels that do not provide
AT_MINSIGSTKSZ.
The idea is that libc could expose this via sysconf() or some
similar mechanism.
There is deliberately no AT_SIGSTKSZ. The kernel knows nothing
about userspace's own stack overheads and should not pretend to
know.
For arm64:
The primary motivation for this interface is the Scalable Vector
Extension, which can require at least 4KB or so of extra space
in the signal frame for the largest hardware implementations.
To determine the correct value, a "Christmas tree" mode (via the
add_all argument) is added to setup_sigframe_layout(), to simulate
addition of all possible records to the signal frame at maximum
possible size.
If this procedure goes wrong somehow, resulting in a stupidly large
frame layout and hence failure of sigframe_alloc() to allocate a
record to the frame, then this is indicative of a kernel bug. In
this case, we WARN() and no attempt is made to populate
AT_MINSIGSTKSZ for userspace.
For arm64 SVE:
The SVE context block in the signal frame needs to be considered
too when computing the maximum possible signal frame size.
Because the size of this block depends on the vector length, this
patch computes the size based not on the thread's current vector
length but instead on the maximum possible vector length: this
determines the maximum size of SVE context block that can be
observed in any signal frame for the lifetime of the process.
Signed-off-by: Dave Martin <Dave.Martin@arm.com>
Acked-by: Will Deacon <will.deacon@arm.com>
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
2018-06-01 13:10:14 +03:00
minsigstksz_setup ( ) ;
2017-10-31 18:51:10 +03:00
2015-10-19 16:24:41 +03:00
/*
* Check for sane CTR_EL0 . CWG value .
*/
cwg = cache_type_cwg ( ) ;
if ( ! cwg )
2018-05-11 15:33:12 +03:00
pr_warn ( " No Cache Writeback Granule information, assuming %d \n " ,
ARCH_DMA_MINALIGN ) ;
2015-03-27 16:09:23 +03:00
}
2016-02-05 17:58:50 +03:00
2021-06-08 21:02:55 +03:00
static int enable_mismatched_32bit_el0 ( unsigned int cpu )
{
2021-07-30 14:24:39 +03:00
/*
* The first 32 - bit - capable CPU we detected and so can no longer
* be offlined by userspace . - 1 indicates we haven ' t yet onlined
* a 32 - bit - capable CPU .
*/
static int lucky_winner = - 1 ;
2021-06-08 21:02:55 +03:00
struct cpuinfo_arm64 * info = & per_cpu ( cpu_data , cpu ) ;
bool cpu_32bit = id_aa64pfr0_32bit_el0 ( info - > reg_id_aa64pfr0 ) ;
if ( cpu_32bit ) {
cpumask_set_cpu ( cpu , cpu_32bit_el0_mask ) ;
static_branch_enable_cpuslocked ( & arm64_mismatched_32bit_el0 ) ;
}
2021-07-30 14:24:39 +03:00
if ( cpumask_test_cpu ( 0 , cpu_32bit_el0_mask ) = = cpu_32bit )
return 0 ;
if ( lucky_winner > = 0 )
return 0 ;
/*
* We ' ve detected a mismatch . We need to keep one of our CPUs with
* 32 - bit EL0 online so that is_cpu_allowed ( ) doesn ' t end up rejecting
* every CPU in the system for a 32 - bit task .
*/
lucky_winner = cpu_32bit ? cpu : cpumask_any_and ( cpu_32bit_el0_mask ,
cpu_active_mask ) ;
get_cpu_device ( lucky_winner ) - > offline_disabled = true ;
setup_elf_hwcaps ( compat_elf_hwcaps ) ;
2022-07-14 19:15:23 +03:00
elf_hwcap_fixup ( ) ;
2021-07-30 14:24:39 +03:00
pr_info ( " Asymmetric 32-bit EL0 support detected on CPU %u; CPU hot-unplug disabled on CPU %u \n " ,
cpu , lucky_winner ) ;
2021-06-08 21:02:55 +03:00
return 0 ;
}
static int __init init_32bit_el0_mask ( void )
{
if ( ! allow_mismatched_32bit_el0 )
return 0 ;
if ( ! zalloc_cpumask_var ( & cpu_32bit_el0_mask , GFP_KERNEL ) )
return - ENOMEM ;
return cpuhp_setup_state ( CPUHP_AP_ONLINE_DYN ,
" arm64/mismatched_32bit_el0:online " ,
enable_mismatched_32bit_el0 , NULL ) ;
}
subsys_initcall_sync ( init_32bit_el0_mask ) ;
2018-07-31 16:08:56 +03:00
static void __maybe_unused cpu_enable_cnp ( struct arm64_cpu_capabilities const * cap )
{
2022-06-24 18:06:40 +03:00
cpu_replace_ttbr1 ( lm_alias ( swapper_pg_dir ) , idmap_pg_dir ) ;
2018-07-31 16:08:56 +03:00
}
2017-01-09 20:28:31 +03:00
/*
* We emulate only the following system register space .
2022-11-03 11:22:32 +03:00
* Op0 = 0x3 , CRn = 0x0 , Op1 = 0x0 , CRm = [ 0 , 2 - 7 ]
2017-01-09 20:28:31 +03:00
* See Table C5 - 6 System instruction encodings for System register accesses ,
* ARMv8 ARM ( ARM DDI 04 87 A . f ) for more details .
*/
static inline bool __attribute_const__ is_emulated ( u32 id )
{
return ( sys_reg_Op0 ( id ) = = 0x3 & &
sys_reg_CRn ( id ) = = 0x0 & &
sys_reg_Op1 ( id ) = = 0x0 & &
( sys_reg_CRm ( id ) = = 0 | |
2022-11-03 11:22:32 +03:00
( ( sys_reg_CRm ( id ) > = 2 ) & & ( sys_reg_CRm ( id ) < = 7 ) ) ) ) ;
2017-01-09 20:28:31 +03:00
}
/*
* With CRm = = 0 , reg should be one of :
* MIDR_EL1 , MPIDR_EL1 or REVIDR_EL1 .
*/
static inline int emulate_id_reg ( u32 id , u64 * valp )
{
switch ( id ) {
case SYS_MIDR_EL1 :
* valp = read_cpuid_id ( ) ;
break ;
case SYS_MPIDR_EL1 :
* valp = SYS_MPIDR_SAFE_VAL ;
break ;
case SYS_REVIDR_EL1 :
/* IMPLEMENTATION DEFINED values are emulated with 0 */
* valp = 0 ;
break ;
default :
return - EINVAL ;
}
return 0 ;
}
static int emulate_sys_reg ( u32 id , u64 * valp )
{
struct arm64_ftr_reg * regp ;
if ( ! is_emulated ( id ) )
return - EINVAL ;
if ( sys_reg_CRm ( id ) = = 0 )
return emulate_id_reg ( id , valp ) ;
2020-05-27 13:04:36 +03:00
regp = get_arm64_ftr_reg_nowarn ( id ) ;
2017-01-09 20:28:31 +03:00
if ( regp )
* valp = arm64_ftr_reg_user_value ( regp ) ;
else
/*
* The untracked registers are either IMPLEMENTATION DEFINED
* ( e . g , ID_AFR0_EL1 ) or reserved RAZ .
*/
* valp = 0 ;
return 0 ;
}
2018-09-20 07:06:20 +03:00
int do_emulate_mrs ( struct pt_regs * regs , u32 sys_reg , u32 rt )
2017-01-09 20:28:31 +03:00
{
int rc ;
u64 val ;
rc = emulate_sys_reg ( sys_reg , & val ) ;
if ( ! rc ) {
2018-09-20 07:06:20 +03:00
pt_regs_write_reg ( regs , rt , val ) ;
2017-10-25 12:04:33 +03:00
arm64_skip_faulting_instruction ( regs , AARCH64_INSN_SIZE ) ;
2017-01-09 20:28:31 +03:00
}
return rc ;
}
2018-09-20 07:06:20 +03:00
static int emulate_mrs ( struct pt_regs * regs , u32 insn )
{
u32 sys_reg , rt ;
/*
* sys_reg values are defined as used in mrs / msr instruction .
* shift the imm value to get the encoding .
*/
sys_reg = ( u32 ) aarch64_insn_decode_immediate ( AARCH64_INSN_IMM_16 , insn ) < < 5 ;
rt = aarch64_insn_decode_register ( AARCH64_INSN_REGTYPE_RT , insn ) ;
return do_emulate_mrs ( regs , sys_reg , rt ) ;
}
2017-01-09 20:28:31 +03:00
static struct undef_hook mrs_hook = {
2021-05-17 21:02:56 +03:00
. instr_mask = 0xffff0000 ,
. instr_val = 0xd5380000 ,
2018-07-05 17:16:52 +03:00
. pstate_mask = PSR_AA32_MODE_MASK ,
2017-01-09 20:28:31 +03:00
. pstate_val = PSR_MODE_EL0t ,
. fn = emulate_mrs ,
} ;
static int __init enable_mrs_emulation ( void )
{
register_undef_hook ( & mrs_hook ) ;
return 0 ;
}
2017-10-06 16:16:52 +03:00
core_initcall ( enable_mrs_emulation ) ;
2019-04-16 00:21:22 +03:00
2020-11-26 20:25:30 +03:00
enum mitigation_state arm64_get_meltdown_state ( void )
{
if ( __meltdown_safe )
return SPECTRE_UNAFFECTED ;
if ( arm64_kernel_unmapped_at_el0 ( ) )
return SPECTRE_MITIGATED ;
return SPECTRE_VULNERABLE ;
}
2019-04-16 00:21:22 +03:00
ssize_t cpu_show_meltdown ( struct device * dev , struct device_attribute * attr ,
char * buf )
{
2020-11-26 20:25:30 +03:00
switch ( arm64_get_meltdown_state ( ) ) {
case SPECTRE_UNAFFECTED :
2019-04-16 00:21:22 +03:00
return sprintf ( buf , " Not affected \n " ) ;
2020-11-26 20:25:30 +03:00
case SPECTRE_MITIGATED :
2019-04-16 00:21:22 +03:00
return sprintf ( buf , " Mitigation: PTI \n " ) ;
2020-11-26 20:25:30 +03:00
default :
return sprintf ( buf , " Vulnerable \n " ) ;
}
2019-04-16 00:21:22 +03:00
}