Skip to content

Commit

Permalink
arm64: errata: add detection for AMEVCNTR01 incrementing incorrectly
Browse files Browse the repository at this point in the history
commit e89d120 upstream.

The AMU counter AMEVCNTR01 (constant counter) should increment at the same
rate as the system counter. On affected Cortex-A510 cores, AMEVCNTR01
increments incorrectly giving a significantly higher output value. This
results in inaccurate task scheduler utilization tracking and incorrect
feedback on CPU frequency.

Work around this problem by returning 0 when reading the affected counter
in key locations that results in disabling all users of this counter from
using it either for frequency invariance or as FFH reference counter. This
effect is the same to firmware disabling affected counters.

Details on how the two features are affected by this erratum:

 - AMU counters will not be used for frequency invariance for affected
   CPUs and CPUs in the same cpufreq policy. AMUs can still be used for
   frequency invariance for unaffected CPUs in the system. Although
   unlikely, if no alternative method can be found to support frequency
   invariance for affected CPUs (cpufreq based or solution based on
   platform counters) frequency invariance will be disabled. Please check
   the chapter on frequency invariance at
   Documentation/scheduler/sched-capacity.rst for details of its effect.

 - Given that FFH can be used to fetch either the core or constant counter
   values, restrictions are lifted regarding any of these counters
   returning a valid (!0) value. Therefore FFH is considered supported
   if there is a least one CPU that support AMUs, independent of any
   counters being disabled or affected by this erratum. Clarifying
   comments are now added to the cpc_ffh_supported(), cpu_read_constcnt()
   and cpu_read_corecnt() functions.

The above is achieved through adding a new erratum: ARM64_ERRATUM_2457168.

Signed-off-by: Ionela Voinescu <ionela.voinescu@arm.com>
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Will Deacon <will@kernel.org>
Cc: James Morse <james.morse@arm.com>
Link: https://lore.kernel.org/r/20220819103050.24211-1-ionela.voinescu@arm.com
Signed-off-by: Will Deacon <will@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
  • Loading branch information
Ionela Voinescu authored and Greg Kroah-Hartman committed Sep 15, 2022
1 parent 2023413 commit 71d3adb
Show file tree
Hide file tree
Showing 5 changed files with 35 additions and 2 deletions.
2 changes: 2 additions & 0 deletions Documentation/arm64/silicon-errata.rst
Original file line number Diff line number Diff line change
Expand Up @@ -92,6 +92,8 @@ stable kernels.
+----------------+-----------------+-----------------+-----------------------------+
| ARM | Cortex-A77 | #1508412 | ARM64_ERRATUM_1508412 |
+----------------+-----------------+-----------------+-----------------------------+
| ARM | Cortex-A510 | #2457168 | ARM64_ERRATUM_2457168 |
+----------------+-----------------+-----------------+-----------------------------+
| ARM | Neoverse-N1 | #1188873,1418040| ARM64_ERRATUM_1418040 |
+----------------+-----------------+-----------------+-----------------------------+
| ARM | Neoverse-N1 | #1349291 | N/A |
Expand Down
18 changes: 18 additions & 0 deletions arch/arm64/Kconfig
Original file line number Diff line number Diff line change
Expand Up @@ -657,6 +657,24 @@ config ARM64_ERRATUM_1508412

If unsure, say Y.

config ARM64_ERRATUM_2457168
bool "Cortex-A510: 2457168: workaround for AMEVCNTR01 incrementing incorrectly"
depends on ARM64_AMU_EXTN
default y
help
This option adds the workaround for ARM Cortex-A510 erratum 2457168.

The AMU counter AMEVCNTR01 (constant counter) should increment at the same rate
as the system counter. On affected Cortex-A510 cores AMEVCNTR01 increments
incorrectly giving a significantly higher output value.

Work around this problem by keeping the reference values of affected counters
to 0 thus signaling an error case. This effect is the same to firmware disabling
affected counters, in which case 0 will be returned when reading the disabled
counters.

If unsure, say Y.

config CAVIUM_ERRATUM_22375
bool "Cavium erratum 22375, 24313"
default y
Expand Down
3 changes: 2 additions & 1 deletion arch/arm64/include/asm/cpucaps.h
Original file line number Diff line number Diff line change
Expand Up @@ -67,7 +67,8 @@
#define ARM64_MTE 57
#define ARM64_WORKAROUND_1508412 58
#define ARM64_SPECTRE_BHB 59
#define ARM64_WORKAROUND_2457168 60

#define ARM64_NCAPS 60
#define ARM64_NCAPS 61

#endif /* __ASM_CPUCAPS_H */
9 changes: 9 additions & 0 deletions arch/arm64/kernel/cpu_errata.c
Original file line number Diff line number Diff line change
Expand Up @@ -545,6 +545,15 @@ const struct arm64_cpu_capabilities arm64_errata[] = {
0, 0,
1, 0),
},
#endif
#ifdef CONFIG_ARM64_ERRATUM_2457168
{
.desc = "ARM erratum 2457168",
.capability = ARM64_WORKAROUND_2457168,
.type = ARM64_CPUCAP_WEAK_LOCAL_CPU_FEATURE,
/* Cortex-A510 r0p0-r1p1 */
CAP_MIDR_RANGE(MIDR_CORTEX_A510, 0, 0, 1, 1)
},
#endif
{
}
Expand Down
5 changes: 4 additions & 1 deletion arch/arm64/kernel/cpufeature.c
Original file line number Diff line number Diff line change
Expand Up @@ -1559,7 +1559,10 @@ static void cpu_amu_enable(struct arm64_cpu_capabilities const *cap)
pr_info("detected CPU%d: Activity Monitors Unit (AMU)\n",
smp_processor_id());
cpumask_set_cpu(smp_processor_id(), &amu_cpus);
init_cpu_freq_invariance_counters();

/* 0 reference values signal broken/disabled counters */
if (!this_cpu_has_cap(ARM64_WORKAROUND_2457168))
init_cpu_freq_invariance_counters();
}
}

Expand Down

0 comments on commit 71d3adb

Please sign in to comment.