Skip to contentSkip to conversation

target/riscv: add missing named features

PATCH New Daniel Henrique Barboza 1 participant 4 messages Jun 5, 2025 RISC-V

lore
Share & tools
lore
#0 lore raw

v3 cover target/riscv: add missing named features

Hi,

New version where typos in patch 1 were fixed. No other changes made.

All patches acked.

Changes from v2:
- patch 1
- fixed typos dince -> since and specd -> spec
- v2 link: https://lore.kernel.org/qemu-riscv/[email protected]/

Daniel Henrique Barboza (3):
target/riscv/cpu.c: add 'sdtrig' in riscv,isa target/riscv/cpu.c: add 'ssstrict' to riscv,isa target/riscv/cpu.c: do better with 'named features' doc

 target/riscv/cpu.c                |  16 ++++++++++++++--
 target/riscv/tcg/tcg-cpu.c        |   9 +++++++++
 tests/data/acpi/riscv64/virt/RHCT | Bin 400 -> 416 bytes
 3 files changed, 23 insertions(+), 2 deletions(-)
Signature · 1 line
2.49.0
#1 lore raw

PATCH v3 1/3 target/riscv/cpu.c: add 'sdtrig' in riscv,isa

We have support for sdtrig for awhile but we are not advertising it. It is enabled by default via the 'debug' flag. Use the same flag to also advertise sdtrig.

Add an exception in disable_priv_spec_isa_exts() to avoid spamming warnings for 'sdtrig' for vendor CPUs like sifive_u.

---
 target/riscv/cpu.c                |   1 +
 target/riscv/tcg/tcg-cpu.c        |   9 +++++++++
 tests/data/acpi/riscv64/virt/RHCT | Bin 400 -> 406 bytes
 3 files changed, 10 insertions(+)
target/riscv/cpu.c +1 -0
index 629ac37501..3c1b1dca47 100644
@@ -189,6 +189,7 @@ const RISCVIsaExtData isa_edata_arr[] = {
     ISA_EXT_DATA_ENTRY(zvkt, PRIV_VERSION_1_12_0, ext_zvkt),
     ISA_EXT_DATA_ENTRY(zhinx, PRIV_VERSION_1_12_0, ext_zhinx),
     ISA_EXT_DATA_ENTRY(zhinxmin, PRIV_VERSION_1_12_0, ext_zhinxmin),
+    ISA_EXT_DATA_ENTRY(sdtrig, PRIV_VERSION_1_12_0, debug),
     ISA_EXT_DATA_ENTRY(shcounterenw, PRIV_VERSION_1_12_0, has_priv_1_12),
     ISA_EXT_DATA_ENTRY(sha, PRIV_VERSION_1_12_0, ext_sha),
     ISA_EXT_DATA_ENTRY(shgatpa, PRIV_VERSION_1_12_0, has_priv_1_12),
target/riscv/tcg/tcg-cpu.c +9 -0
index 55fd9e5584..e9d80a13ef 100644
@@ -451,6 +451,15 @@ static void riscv_cpu_disable_priv_spec_isa_exts(RISCVCPU *cpu)
                 continue;
             }
 
+            /*
+             * cpu.debug = true is marked as 'sdtrig', priv spec 1.12.
+             * Skip this warning since existing CPUs with older priv
+             * spec and debug = true will be impacted.
+             */
+            if (!strcmp(edata->name, "sdtrig")) {
+                continue;
+            }
+
             isa_ext_update_enabled(cpu, edata->ext_enable_offset, false);
 
             /*
tests/data/acpi/riscv64/virt/RHCT binary +0 -0
index 13c8025b868051485be5ba62974a22971a07bc6a..156607dec45b0e63e5b3ebed62e81076dacd80d0 100644
GIT binary patch
Signature · 1 line
2.49.0
#2 lore raw

PATCH v3 2/3 target/riscv/cpu.c: add 'ssstrict' to riscv,isa

In reply to Daniel Henrique Barboza

'ssstrict' is a RVA23 profile-defined extension defined as follows:

"No non-conforming extensions are present. Attempts to execute unimplemented opcodes or access unimplemented CSRs in the standard or reserved encoding spaces raises an illegal instruction exception that results in a contained trap to the supervisor-mode trap handler."

In short, we need to throw an exception when accessing unimplemented CSRs or opcodes. We do that, so let's advertise it.

---
 target/riscv/cpu.c                |   1 +
 tests/data/acpi/riscv64/virt/RHCT | Bin 406 -> 416 bytes
 2 files changed, 1 insertion(+)
target/riscv/cpu.c +1 -0
index 3c1b1dca47..058387ac48 100644
@@ -217,6 +217,7 @@ const RISCVIsaExtData isa_edata_arr[] = {
     ISA_EXT_DATA_ENTRY(ssnpm, PRIV_VERSION_1_13_0, ext_ssnpm),
     ISA_EXT_DATA_ENTRY(sspm, PRIV_VERSION_1_13_0, ext_sspm),
     ISA_EXT_DATA_ENTRY(ssstateen, PRIV_VERSION_1_12_0, ext_ssstateen),
+    ISA_EXT_DATA_ENTRY(ssstrict, PRIV_VERSION_1_12_0, has_priv_1_12),
     ISA_EXT_DATA_ENTRY(sstc, PRIV_VERSION_1_12_0, ext_sstc),
     ISA_EXT_DATA_ENTRY(sstvala, PRIV_VERSION_1_12_0, has_priv_1_12),
     ISA_EXT_DATA_ENTRY(sstvecd, PRIV_VERSION_1_12_0, has_priv_1_12),
tests/data/acpi/riscv64/virt/RHCT binary +0 -0
index 156607dec45b0e63e5b3ebed62e81076dacd80d0..52a4cc4b6380eee3299b965271a39e9e01f5a698 100644
GIT binary patch
Signature · 1 line
2.49.0
#3 lore raw

PATCH v3 3/3 target/riscv/cpu.c: do better with 'named features' doc

In reply to Daniel Henrique Barboza

Most of the named features are added directly in isa_edata_arr[], some of them are also added in riscv_cpu_named_features(). There is a reason for that, and the existing docs can do better explaining it.

---
 target/riscv/cpu.c | 14 ++++++++++++--
 1 file changed, 12 insertions(+), 2 deletions(-)
target/riscv/cpu.c +12 -2
index 058387ac48..798bfed6b7 100644
@@ -1377,13 +1377,23 @@ const RISCVCPUMultiExtConfig riscv_cpu_experimental_exts[] = {
  * 'Named features' is the name we give to extensions that we
  * don't want to expose to users. They are either immutable
  * (always enabled/disable) or they'll vary depending on
- * the resulting CPU state. They have riscv,isa strings
- * and priv_ver like regular extensions.
+ * the resulting CPU state.
+ *
+ * Some of them are always enabled depending on priv version
+ * of the CPU and are declared directly in isa_edata_arr[].
+ * The ones listed here have special checks during finalize()
+ * time and require their own flags like regular extensions.
+ * See riscv_cpu_update_named_features() for more info.
  */
 const RISCVCPUMultiExtConfig riscv_cpu_named_features[] = {
     MULTI_EXT_CFG_BOOL("zic64b", ext_zic64b, true),
     MULTI_EXT_CFG_BOOL("ssstateen", ext_ssstateen, true),
     MULTI_EXT_CFG_BOOL("sha", ext_sha, true),
+
+    /*
+     * 'ziccrse' has its own flag because the KVM driver
+     * wants to enable/disable it on its own accord.
+     */
     MULTI_EXT_CFG_BOOL("ziccrse", ext_ziccrse, true),
 
     { },
Signature · 1 line
2.49.0