[PATCH] target/i386: remove PCOMMIT from TCG, deprecate property

Paolo Bonzini posted 1 patch 6 months, 2 weeks ago
Patches applied successfully (tree, apply log)
git fetch https://github.com/patchew-project/qemu tags/patchew/20240508154421.61419-1-pbonzini@redhat.com
Maintainers: Paolo Bonzini <pbonzini@redhat.com>, Richard Henderson <richard.henderson@linaro.org>, Eduardo Habkost <eduardo@habkost.net>
docs/about/deprecated.rst   |  8 ++++++++
target/i386/cpu.h           |  2 --
target/i386/cpu.c           |  2 +-
target/i386/tcg/translate.c | 12 +-----------
4 files changed, 10 insertions(+), 14 deletions(-)
[PATCH] target/i386: remove PCOMMIT from TCG, deprecate property
Posted by Paolo Bonzini 6 months, 2 weeks ago
The PCOMMIT instruction was never included in any physical processor.
TCG implements it as a no-op instruction, but its utility is debatable
to say the least.  Drop it from the decoder since it is only available
with "-cpu max", which does not guarantee migration compatibility
across versions, and deprecate the property just in case someone is
using it as "pcommit=off".

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
 docs/about/deprecated.rst   |  8 ++++++++
 target/i386/cpu.h           |  2 --
 target/i386/cpu.c           |  2 +-
 target/i386/tcg/translate.c | 12 +-----------
 4 files changed, 10 insertions(+), 14 deletions(-)

diff --git a/docs/about/deprecated.rst b/docs/about/deprecated.rst
index 03f8b1b655e..b87fe70be47 100644
--- a/docs/about/deprecated.rst
+++ b/docs/about/deprecated.rst
@@ -418,6 +418,14 @@ Backend ``memory`` (since 9.0)
 CPU device properties
 '''''''''''''''''''''
 
+``pcommit`` on x86 (since 9.1)
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+The PCOMMIT instruction was never included in any physical processor.
+It was implemented as a no-op instruction in TCG up to QEMU 9.0, but
+only with ``-cpu max`` (which does not guarantee migration compatibility
+across versions).
+
 ``pmu-num=n`` on RISC-V CPUs (since 8.2)
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
 
diff --git a/target/i386/cpu.h b/target/i386/cpu.h
index 1e0d2c915f5..ccccb62fc35 100644
--- a/target/i386/cpu.h
+++ b/target/i386/cpu.h
@@ -816,8 +816,6 @@ uint64_t x86_cpu_get_supported_feature_word(FeatureWord w,
 #define CPUID_7_0_EBX_SMAP              (1U << 20)
 /* AVX-512 Integer Fused Multiply Add */
 #define CPUID_7_0_EBX_AVX512IFMA        (1U << 21)
-/* Persistent Commit */
-#define CPUID_7_0_EBX_PCOMMIT           (1U << 22)
 /* Flush a Cache Line Optimized */
 #define CPUID_7_0_EBX_CLFLUSHOPT        (1U << 23)
 /* Cache Line Write Back */
diff --git a/target/i386/cpu.c b/target/i386/cpu.c
index 1058b6803fd..79372de8c5a 100644
--- a/target/i386/cpu.c
+++ b/target/i386/cpu.c
@@ -713,7 +713,7 @@ void x86_cpu_vendor_words2str(char *dst, uint32_t vendor1,
 #endif
 #define TCG_7_0_EBX_FEATURES (CPUID_7_0_EBX_SMEP | CPUID_7_0_EBX_SMAP | \
           CPUID_7_0_EBX_BMI1 | CPUID_7_0_EBX_BMI2 | CPUID_7_0_EBX_ADX | \
-          CPUID_7_0_EBX_PCOMMIT | CPUID_7_0_EBX_CLFLUSHOPT |            \
+          CPUID_7_0_EBX_CLFLUSHOPT |            \
           CPUID_7_0_EBX_CLWB | CPUID_7_0_EBX_MPX | CPUID_7_0_EBX_FSGSBASE | \
           CPUID_7_0_EBX_ERMS | CPUID_7_0_EBX_AVX2 | CPUID_7_0_EBX_RDSEED | \
           CPUID_7_0_EBX_SHA_NI | CPUID_7_0_EBX_KERNEL_FEATURES)
diff --git a/target/i386/tcg/translate.c b/target/i386/tcg/translate.c
index 3842b294842..7d9f6b5c55b 100644
--- a/target/i386/tcg/translate.c
+++ b/target/i386/tcg/translate.c
@@ -4487,17 +4487,7 @@ static void disas_insn_old(DisasContext *s, CPUState *cpu, int b)
             }
             goto unknown_op;
 
-        case 0xf8: /* sfence / pcommit */
-            if (prefixes & PREFIX_DATA) {
-                /* pcommit */
-                if (!(s->cpuid_7_0_ebx_features & CPUID_7_0_EBX_PCOMMIT)
-                    || (prefixes & PREFIX_LOCK)) {
-                    goto illegal_op;
-                }
-                break;
-            }
-            /* fallthru */
-        case 0xf9 ... 0xff: /* sfence */
+        case 0xf8 ... 0xff: /* sfence */
             if (!(s->cpuid_features & CPUID_SSE)
                 || (prefixes & PREFIX_LOCK)) {
                 goto illegal_op;
-- 
2.45.0
Re: [PATCH] target/i386: remove PCOMMIT from TCG, deprecate property
Posted by Zhao Liu 6 months, 2 weeks ago
On Wed, May 08, 2024 at 05:44:21PM +0200, Paolo Bonzini wrote:
> Date: Wed,  8 May 2024 17:44:21 +0200
> From: Paolo Bonzini <pbonzini@redhat.com>
> Subject: [PATCH] target/i386: remove PCOMMIT from TCG, deprecate property
> X-Mailer: git-send-email 2.45.0
> 
> The PCOMMIT instruction was never included in any physical processor.
> TCG implements it as a no-op instruction, but its utility is debatable
> to say the least.  Drop it from the decoder since it is only available
> with "-cpu max", which does not guarantee migration compatibility
> across versions, and deprecate the property just in case someone is
> using it as "pcommit=off".
> 
> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> ---
>  docs/about/deprecated.rst   |  8 ++++++++
>  target/i386/cpu.h           |  2 --
>  target/i386/cpu.c           |  2 +-
>  target/i386/tcg/translate.c | 12 +-----------
>  4 files changed, 10 insertions(+), 14 deletions(-)

Reviewed-by: Zhao Liu <zhao1.liu@intel.com>
Re: [PATCH] target/i386: remove PCOMMIT from TCG, deprecate property
Posted by Richard Henderson 6 months, 2 weeks ago
On 5/8/24 17:44, Paolo Bonzini wrote:
> The PCOMMIT instruction was never included in any physical processor.
> TCG implements it as a no-op instruction, but its utility is debatable
> to say the least.  Drop it from the decoder since it is only available
> with "-cpu max", which does not guarantee migration compatibility
> across versions, and deprecate the property just in case someone is
> using it as "pcommit=off".
> 
> Signed-off-by: Paolo Bonzini<pbonzini@redhat.com>
> ---
>   docs/about/deprecated.rst   |  8 ++++++++
>   target/i386/cpu.h           |  2 --
>   target/i386/cpu.c           |  2 +-
>   target/i386/tcg/translate.c | 12 +-----------
>   4 files changed, 10 insertions(+), 14 deletions(-)

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

r~