[PATCH] target/arm: Implement FEAT_HPMN0

Peter Maydell posted 1 patch 7 months, 2 weeks ago
Patches applied successfully (tree, apply log)
git fetch https://github.com/patchew-project/qemu tags/patchew/20230921185445.3339214-1-peter.maydell@linaro.org
Maintainers: Peter Maydell <peter.maydell@linaro.org>
docs/system/arm/emulation.rst | 1 +
target/arm/helper.c           | 2 +-
target/arm/tcg/cpu32.c        | 4 ++++
target/arm/tcg/cpu64.c        | 1 +
4 files changed, 7 insertions(+), 1 deletion(-)
[PATCH] target/arm: Implement FEAT_HPMN0
Posted by Peter Maydell 7 months, 2 weeks ago
FEAT_HPMN0 is a small feature which defines that it is valid for
MDCR_EL2.HPMN to be set to 0, meaning "no PMU event counters provided
to an EL1 guest" (previously this setting was reserved). QEMU's
implementation almost gets HPMN == 0 right, but we need to fix
one check in pmevcntr_is_64_bit(). That is enough for us to
advertise the feature in the 'max' CPU.

(We don't need to make the behaviour conditional on feature
presence, because the FEAT_HPMN0 behaviour is within the range
of permitted UNPREDICTABLE behaviour for a non-FEAT_HPMN0
implementation.)

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
---
 docs/system/arm/emulation.rst | 1 +
 target/arm/helper.c           | 2 +-
 target/arm/tcg/cpu32.c        | 4 ++++
 target/arm/tcg/cpu64.c        | 1 +
 4 files changed, 7 insertions(+), 1 deletion(-)

diff --git a/docs/system/arm/emulation.rst b/docs/system/arm/emulation.rst
index 3df936fc356..b19ea198c24 100644
--- a/docs/system/arm/emulation.rst
+++ b/docs/system/arm/emulation.rst
@@ -45,6 +45,7 @@ the following architecture extensions:
 - FEAT_HCX (Support for the HCRX_EL2 register)
 - FEAT_HPDS (Hierarchical permission disables)
 - FEAT_HPDS2 (Translation table page-based hardware attributes)
+- FEAT_HPMN0 (Setting of MDCR_EL2.HPMN to zero)
 - FEAT_I8MM (AArch64 Int8 matrix multiplication instructions)
 - FEAT_IDST (ID space trap handling)
 - FEAT_IESB (Implicit error synchronization event)
diff --git a/target/arm/helper.c b/target/arm/helper.c
index 3b22596eabf..ea3e5c6fd0f 100644
--- a/target/arm/helper.c
+++ b/target/arm/helper.c
@@ -1283,7 +1283,7 @@ static bool pmevcntr_is_64_bit(CPUARMState *env, int counter)
         bool hlp = env->cp15.mdcr_el2 & MDCR_HLP;
         int hpmn = env->cp15.mdcr_el2 & MDCR_HPMN;
 
-        if (hpmn != 0 && counter >= hpmn) {
+        if (counter >= hpmn) {
             return hlp;
         }
     }
diff --git a/target/arm/tcg/cpu32.c b/target/arm/tcg/cpu32.c
index 1f918ff5375..0d5d8e307dd 100644
--- a/target/arm/tcg/cpu32.c
+++ b/target/arm/tcg/cpu32.c
@@ -89,6 +89,10 @@ void aa32_max_features(ARMCPU *cpu)
     t = FIELD_DP32(t, ID_DFR0, COPSDBG, 9);       /* FEAT_Debugv8p4 */
     t = FIELD_DP32(t, ID_DFR0, PERFMON, 6);       /* FEAT_PMUv3p5 */
     cpu->isar.id_dfr0 = t;
+
+    t = cpu->isar.id_dfr1;
+    t = FIELD_DP32(t, ID_DFR1, HPMN0, 1);         /* FEAT_HPMN0 */
+    cpu->isar.id_dfr1 = t;
 }
 
 /* CPU models. These are not needed for the AArch64 linux-user build. */
diff --git a/target/arm/tcg/cpu64.c b/target/arm/tcg/cpu64.c
index 7264ab5ead1..ee369f10db6 100644
--- a/target/arm/tcg/cpu64.c
+++ b/target/arm/tcg/cpu64.c
@@ -1104,6 +1104,7 @@ void aarch64_max_tcg_initfn(Object *obj)
     t = cpu->isar.id_aa64dfr0;
     t = FIELD_DP64(t, ID_AA64DFR0, DEBUGVER, 9);  /* FEAT_Debugv8p4 */
     t = FIELD_DP64(t, ID_AA64DFR0, PMUVER, 6);    /* FEAT_PMUv3p5 */
+    t = FIELD_DP64(t, ID_AA64DFR0, HPMN0, 1);     /* FEAT_HPMN0 */
     cpu->isar.id_aa64dfr0 = t;
 
     t = cpu->isar.id_aa64smfr0;
-- 
2.34.1
Re: [PATCH] target/arm: Implement FEAT_HPMN0
Posted by Richard Henderson 7 months, 1 week ago
On 9/21/23 14:54, Peter Maydell wrote:
> FEAT_HPMN0 is a small feature which defines that it is valid for
> MDCR_EL2.HPMN to be set to 0, meaning "no PMU event counters provided
> to an EL1 guest" (previously this setting was reserved). QEMU's
> implementation almost gets HPMN == 0 right, but we need to fix
> one check in pmevcntr_is_64_bit(). That is enough for us to
> advertise the feature in the 'max' CPU.
> 
> (We don't need to make the behaviour conditional on feature
> presence, because the FEAT_HPMN0 behaviour is within the range
> of permitted UNPREDICTABLE behaviour for a non-FEAT_HPMN0
> implementation.)
> 
> Signed-off-by: Peter Maydell<peter.maydell@linaro.org>
> ---
>   docs/system/arm/emulation.rst | 1 +
>   target/arm/helper.c           | 2 +-
>   target/arm/tcg/cpu32.c        | 4 ++++
>   target/arm/tcg/cpu64.c        | 1 +
>   4 files changed, 7 insertions(+), 1 deletion(-)

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>

r~