From: Beata Michalska <beata.michalska@arm.com>
To: linux-arm-kernel@lists.infradead.org
Cc: Vladimir.Murzin@arm.com, catalin.marinas@arm.com,
will@kernel.org, suzuki.poulose@arm.com, xuewen.yan94@gmail.com,
linux-kernel@vger.kernel.org
Subject: [PATCH 2/3] arm64: errata: Add Cortex-A715/A720 AMU core counter workaround
Date: Wed, 11 Nov 2026 13:39:28 +0100 [thread overview]
Message-ID: <20261111124005.2688820-3-beata.michalska@arm.com> (raw)
In-Reply-To: <20261111124005.2688820-1-beata.michalska@arm.com>
On Cortex-A715 (erratum 2738462) and Cortex-A720 (erratum 2738461),
the AMEVCNTR00 counter may continue incrementing while the core
remains in WFE with the clocks being enabled. The resulting core
cycle count may become unreliable.
Add a workaround that effectively treats the AMU counter
as unavailable.
Cc: stable@vger.kernel.org
Signed-off-by: Beata Michalska <beata.michalska@arm.com>
---
Documentation/arch/arm64/silicon-errata.rst | 4 +++
arch/arm64/Kconfig | 32 +++++++++++++++++++++
arch/arm64/kernel/cpu_errata.c | 20 +++++++++++++
arch/arm64/kernel/cpufeature.c | 3 +-
arch/arm64/kernel/topology.c | 3 +-
arch/arm64/tools/cpucaps | 1 +
6 files changed, 61 insertions(+), 2 deletions(-)
diff --git a/Documentation/arch/arm64/silicon-errata.rst b/Documentation/arch/arm64/silicon-errata.rst
index 09fe032e1aea..5801597a387c 100644
--- a/Documentation/arch/arm64/silicon-errata.rst
+++ b/Documentation/arch/arm64/silicon-errata.rst
@@ -171,8 +171,12 @@ stable kernels.
+----------------+-----------------+-----------------+-----------------------------+
| ARM | Cortex-A715 | #2645198 | ARM64_ERRATUM_2645198 |
+----------------+-----------------+-----------------+-----------------------------+
+| ARM | Cortex-A715 | #2738462 | ARM64_ERRATUM_2738461 |
++----------------+-----------------+-----------------+-----------------------------+
| ARM | Cortex-A715 | #3456084 | ARM64_ERRATUM_3194386 |
+----------------+-----------------+-----------------+-----------------------------+
+| ARM | Cortex-A720 | #2738461 | ARM64_ERRATUM_2738461 |
++----------------+-----------------+-----------------+-----------------------------+
| ARM | Cortex-A720 | #3456091 | ARM64_ERRATUM_3194386 |
+----------------+-----------------+-----------------+-----------------------------+
| ARM | Cortex-A720 | #4689339 | ARM64_ERRATUM_3821522 |
diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig
index cf50f3051edc..afc6226689be 100644
--- a/arch/arm64/Kconfig
+++ b/arch/arm64/Kconfig
@@ -1051,6 +1051,38 @@ config ARM64_WORKAROUND_BROKEN_AMU_CONSTCNT
# making it unreliable. As a result, AMU users should treat it as
# unavailable rather than relying on misleading counter values.
+config ARM64_WORKAROUND_BROKEN_AMU_CORECNT
+ bool
+ # Targeting errata affecting AMEVCNTR00, the AMU core cycle counter.
+ # On affected CPUs the counter may continue incrementing while core
+ # remains in WFE with the clocks enabled.
+ # This may lead to readings of said counter being unreliable.
+ # How much though depends on the sampling intervals.
+
+config ARM64_ERRATUM_2738461
+ bool "Cortex-A715/Cortex-A720: workaround for AMEVCNTR00 incrementing incorrectly"
+ depends on ARM64_AMU_EXTN
+ select ARM64_WORKAROUND_BROKEN_AMU_CORECNT
+ default y
+ help
+ This option adds the workaround for the following errata:
+
+ * ARM Cortex-A715 erratum 2738462
+ * ARM Cortex-A720 erratum 2738461
+
+ On affected cores the AMEVCNTR00 counter (AMU core frequency cycles
+ counter) may continue incrementing while the core remains in WFE
+ with the clocks being enabled.
+ This may result in inaccurate counter readings, with the magnitude
+ of the discrepancy depending on the sampling interval.
+
+ With this workaround, reads of the affected counter return 0 in
+ the relevant paths. This causes users of the counter to treat it
+ as unavailable and is functionally equivalent to firmware disabling
+ the affected counter.
+
+ If unsure, say Y.
+
config ARM64_ERRATUM_2457168
bool "Cortex-A510: 2457168: workaround for AMEVCNTR01 incrementing incorrectly"
depends on ARM64_AMU_EXTN
diff --git a/arch/arm64/kernel/cpu_errata.c b/arch/arm64/kernel/cpu_errata.c
index b20da8b24be3..957917cfab70 100644
--- a/arch/arm64/kernel/cpu_errata.c
+++ b/arch/arm64/kernel/cpu_errata.c
@@ -436,6 +436,18 @@ static const struct midr_range workaround_amu_constcnt_list[] = {
};
#endif /* CONFIG_ARM64_WORKAROUND_BROKEN_AMU_CONSTCNT */
+#ifdef CONFIG_ARM64_WORKAROUND_BROKEN_AMU_CORECNT
+static const struct midr_range workaround_amu_corecnt_list[] = {
+#ifdef CONFIG_ARM64_ERRATUM_2738461
+ /* Cortex-A715 r0p0 - r1p3 */
+ MIDR_RANGE(MIDR_CORTEX_A715, 0, 0, 1, 3),
+ /* Cortex-A720 r0p0 - r0p1 */
+ MIDR_RANGE(MIDR_CORTEX_A720, 0, 0, 0, 1),
+#endif
+ {}
+};
+#endif /* CONFIG_ARM64_WORKAROUND_BROKEN_AMU_CORECNT */
+
#ifdef CONFIG_CAVIUM_ERRATUM_23154
static const struct midr_range cavium_erratum_23154_cpus[] = {
MIDR_ALL_VERSIONS(MIDR_THUNDERX),
@@ -979,6 +991,14 @@ const struct arm64_cpu_capabilities arm64_errata[] = {
CAP_MIDR_RANGE_LIST(workaround_amu_constcnt_list)
},
#endif
+#ifdef CONFIG_ARM64_WORKAROUND_BROKEN_AMU_CORECNT
+ {
+ .desc = "Broken AMU AMEVCNTR00 (core cycle counter)",
+ .capability = ARM64_WORKAROUND_BROKEN_AMU_CORECNT,
+ .type = ARM64_CPUCAP_WEAK_LOCAL_CPU_FEATURE,
+ CAP_MIDR_RANGE_LIST(workaround_amu_corecnt_list)
+ },
+#endif
#ifdef CONFIG_ARM64_ERRATUM_2038923
{
.desc = "ARM erratum 2038923",
diff --git a/arch/arm64/kernel/cpufeature.c b/arch/arm64/kernel/cpufeature.c
index 34c5c03c0c9b..0e8a9aedb84b 100644
--- a/arch/arm64/kernel/cpufeature.c
+++ b/arch/arm64/kernel/cpufeature.c
@@ -1938,7 +1938,8 @@ static void cpu_amu_enable(struct arm64_cpu_capabilities const *cap)
cpumask_set_cpu(smp_processor_id(), &amu_cpus);
/* 0 reference values signal broken/disabled counters */
- if (!this_cpu_has_cap(ARM64_WORKAROUND_BROKEN_AMU_CONSTCNT))
+ if (!this_cpu_has_cap(ARM64_WORKAROUND_BROKEN_AMU_CONSTCNT)
+ && !this_cpu_has_cap(ARM64_WORKAROUND_BROKEN_AMU_CORECNT))
update_freq_counters_refs();
}
}
diff --git a/arch/arm64/kernel/topology.c b/arch/arm64/kernel/topology.c
index 44305b997a02..66ac67bd025e 100644
--- a/arch/arm64/kernel/topology.c
+++ b/arch/arm64/kernel/topology.c
@@ -389,7 +389,8 @@ static void cpu_read_corecnt(void *val)
* counter read is properly handled as an error case by the users of the
* counter.
*/
- *(u64 *)val = read_corecnt();
+ *(u64 *)val = this_cpu_has_cap(ARM64_WORKAROUND_BROKEN_AMU_CORECNT) ?
+ 0UL : read_corecnt();
}
static void cpu_read_constcnt(void *val)
diff --git a/arch/arm64/tools/cpucaps b/arch/arm64/tools/cpucaps
index 20cc185649f9..3125695122a0 100644
--- a/arch/arm64/tools/cpucaps
+++ b/arch/arm64/tools/cpucaps
@@ -133,3 +133,4 @@ WORKAROUND_SPECULATIVE_AT
WORKAROUND_SPECULATIVE_SSBS
WORKAROUND_SPECULATIVE_UNPRIV_LOAD
WORKAROUND_BROKEN_AMU_CONSTCNT
+WORKAROUND_BROKEN_AMU_CORECNT
--
2.43.0
next prev parent reply other threads:[~2026-11-11 12:40 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-11-11 12:39 [PATCH 0/3] arm64: errata: Update AMU workarounds Beata Michalska
2026-11-11 12:39 ` [PATCH 1/3] arm64: errata: Extend the workaround for broken AMEVCNTR01 counter Beata Michalska
2026-11-11 12:39 ` Beata Michalska [this message]
2026-11-11 12:39 ` [PATCH 3/3] arm64: topology: Reduce the skew for AMU feedback counter sampling Beata Michalska
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20261111124005.2688820-3-beata.michalska@arm.com \
--to=beata.michalska@arm.com \
--cc=Vladimir.Murzin@arm.com \
--cc=catalin.marinas@arm.com \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=suzuki.poulose@arm.com \
--cc=will@kernel.org \
--cc=xuewen.yan94@gmail.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®