[PATCH v3 0/2] riscv: support new isa extension detection devicetree properties

[PATCH v3 0/2] riscv: support new isa extension detection devicetree properties

[Conor Dooley]

From: Conor Dooley <conor.dooley@microchip.com>

Making it a series to keep the standalone change to riscv_isa_string()
that Drew reported separate.

Changes in v3:
- g_free() isa_extensions too
- use misa_mxl_max rather than the compile target for the base isa
- add a new patch changing riscv_isa_string() to do the same
- drop a null check that cannot be null
- rebased on top of Alistair's next branch

Changes in v2:
- use g_strdup() for multiletter extension string copying
- wrap stuff in #ifndef to prevent breaking the user mode build
- rename riscv_isa_set_props() -> riscv_isa_write_fdt()

CC: Alistair Francis <Alistair.Francis@wdc.com>
CC: Bin Meng <bin.meng@windriver.com>
CC: Palmer Dabbelt <palmer@dabbelt.com>
CC: Weiwei Li <liwei1518@gmail.com>
CC: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
CC: Andrew Jones <ajones@ventanamicro.com>
CC: Liu Zhiwei <zhiwei_liu@linux.alibaba.com>
CC: qemu-riscv@nongnu.org
CC: qemu-devel@nongnu.org

Conor Dooley (2):
  target/riscv: use misa_mxl_max to populate isa string rather than
    TARGET_LONG_BITS
  target/riscv: support new isa extension detection devicetree
    properties

 hw/riscv/sifive_u.c |  7 ++----
 hw/riscv/spike.c    |  6 ++---
 hw/riscv/virt.c     |  6 ++---
 target/riscv/cpu.c  | 57 ++++++++++++++++++++++++++++++++++++++++++++-
 target/riscv/cpu.h  |  1 +
 5 files changed, 63 insertions(+), 14 deletions(-)

-- 
2.39.2

[PATCH v3 1/2] target/riscv: use misa_mxl_max to populate isa string rather than TARGET_LONG_BITS

[Conor Dooley]

From: Conor Dooley <conor.dooley@microchip.com>

A cpu may not have the same xlen as the compile time target, and
misa_mxl_max is the source of truth for what the hart supports.

Reported-by: Andrew Jones <ajones@ventanamicro.com>
Link: https://lore.kernel.org/qemu-riscv/20240108-efa3f83dcd3997dc0af458d7@orel/
Signed-off-by: Conor Dooley <conor.dooley@microchip.com>
---
Perhaps this misa_mxl_max -> width conversion should exist as a macro?
There's now 3 individual conversions of this type - two I added and one
in the gdb code.
---
 target/riscv/cpu.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/target/riscv/cpu.c b/target/riscv/cpu.c
index 8cbfc7e781..5b5da970f2 100644
--- a/target/riscv/cpu.c
+++ b/target/riscv/cpu.c
@@ -1860,7 +1860,9 @@ char *riscv_isa_string(RISCVCPU *cpu)
     int i;
     const size_t maxlen = sizeof("rv128") + sizeof(riscv_single_letter_exts);
     char *isa_str = g_new(char, maxlen);
-    char *p = isa_str + snprintf(isa_str, maxlen, "rv%d", TARGET_LONG_BITS);
+    int xlen = 16 << cpu->env.misa_mxl_max;
+    char *p = isa_str + snprintf(isa_str, maxlen, "rv%d", xlen);
+
     for (i = 0; i < sizeof(riscv_single_letter_exts) - 1; i++) {
         if (cpu->env.misa_ext & RV(riscv_single_letter_exts[i])) {
             *p++ = qemu_tolower(riscv_single_letter_exts[i]);
-- 
2.39.2

Re: [PATCH v3 1/2] target/riscv: use misa_mxl_max to populate isa string rather than TARGET_LONG_BITS

[Andrew Jones]

On Wed, Jan 10, 2024 at 10:25:36AM +0000, Conor Dooley wrote:
> From: Conor Dooley <conor.dooley@microchip.com>
> 
> A cpu may not have the same xlen as the compile time target, and
> misa_mxl_max is the source of truth for what the hart supports.
> 
> Reported-by: Andrew Jones <ajones@ventanamicro.com>
> Link: https://lore.kernel.org/qemu-riscv/20240108-efa3f83dcd3997dc0af458d7@orel/
> Signed-off-by: Conor Dooley <conor.dooley@microchip.com>
> ---
> Perhaps this misa_mxl_max -> width conversion should exist as a macro?
> There's now 3 individual conversions of this type - two I added and one
> in the gdb code.

A macro is a good idea. I had to go look at Table 3.1 of the priv spec to
understand the 16 shift stuff.

> ---
>  target/riscv/cpu.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/target/riscv/cpu.c b/target/riscv/cpu.c
> index 8cbfc7e781..5b5da970f2 100644
> --- a/target/riscv/cpu.c
> +++ b/target/riscv/cpu.c
> @@ -1860,7 +1860,9 @@ char *riscv_isa_string(RISCVCPU *cpu)
>      int i;
>      const size_t maxlen = sizeof("rv128") + sizeof(riscv_single_letter_exts);
>      char *isa_str = g_new(char, maxlen);
> -    char *p = isa_str + snprintf(isa_str, maxlen, "rv%d", TARGET_LONG_BITS);
> +    int xlen = 16 << cpu->env.misa_mxl_max;
> +    char *p = isa_str + snprintf(isa_str, maxlen, "rv%d", xlen);
> +
>      for (i = 0; i < sizeof(riscv_single_letter_exts) - 1; i++) {
>          if (cpu->env.misa_ext & RV(riscv_single_letter_exts[i])) {
>              *p++ = qemu_tolower(riscv_single_letter_exts[i]);
> -- 
> 2.39.2
>

Reviewed-by: Andrew Jones <ajones@ventanamicro.com>

Re: [PATCH v3 1/2] target/riscv: use misa_mxl_max to populate isa string rather than TARGET_LONG_BITS

[Alistair Francis]

On Wed, Jan 10, 2024 at 9:34 PM Conor Dooley <conor@kernel.org> wrote:
>
> From: Conor Dooley <conor.dooley@microchip.com>
>
> A cpu may not have the same xlen as the compile time target, and
> misa_mxl_max is the source of truth for what the hart supports.
>
> Reported-by: Andrew Jones <ajones@ventanamicro.com>
> Link: https://lore.kernel.org/qemu-riscv/20240108-efa3f83dcd3997dc0af458d7@orel/
> Signed-off-by: Conor Dooley <conor.dooley@microchip.com>

Reviewed-by: Alistair Francis <alistair.francis@wdc.com>

Alistair

> ---
> Perhaps this misa_mxl_max -> width conversion should exist as a macro?
> There's now 3 individual conversions of this type - two I added and one
> in the gdb code.
> ---
>  target/riscv/cpu.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/target/riscv/cpu.c b/target/riscv/cpu.c
> index 8cbfc7e781..5b5da970f2 100644
> --- a/target/riscv/cpu.c
> +++ b/target/riscv/cpu.c
> @@ -1860,7 +1860,9 @@ char *riscv_isa_string(RISCVCPU *cpu)
>      int i;
>      const size_t maxlen = sizeof("rv128") + sizeof(riscv_single_letter_exts);
>      char *isa_str = g_new(char, maxlen);
> -    char *p = isa_str + snprintf(isa_str, maxlen, "rv%d", TARGET_LONG_BITS);
> +    int xlen = 16 << cpu->env.misa_mxl_max;
> +    char *p = isa_str + snprintf(isa_str, maxlen, "rv%d", xlen);
> +
>      for (i = 0; i < sizeof(riscv_single_letter_exts) - 1; i++) {
>          if (cpu->env.misa_ext & RV(riscv_single_letter_exts[i])) {
>              *p++ = qemu_tolower(riscv_single_letter_exts[i]);
> --
> 2.39.2
>
>

[PATCH v3 2/2] target/riscv: support new isa extension detection devicetree properties

[Conor Dooley]

From: Conor Dooley <conor.dooley@microchip.com>

A few months ago I submitted a patch to various lists, deprecating
"riscv,isa" with a lengthy commit message [0] that is now commit
aeb71e42caae ("dt-bindings: riscv: deprecate riscv,isa") in the Linux
kernel tree. Primarily, the goal was to replace "riscv,isa" with a new
set of properties that allowed for strictly defining the meaning of
various extensions, where "riscv,isa" was tied to whatever definitions
inflicted upon us by the ISA manual, which have seen some variance over
time.

Two new properties were introduced: "riscv,isa-base" and
"riscv,isa-extensions". The former is a simple string to communicate the
base ISA implemented by a hart and the latter an array of strings used
to communicate the set of ISA extensions supported, per the definitions
of each substring in extensions.yaml [1]. A beneficial side effect was
also the ability to define vendor extensions in a more "official" way,
as the ISA manual and other RVI specifications only covered the format
for vendor extensions in the ISA string, but not the meaning of vendor
extensions, for obvious reasons.

Add support for setting these two new properties in the devicetrees for
the various devicetree platforms supported by QEMU for RISC-V. The Linux
kernel already supports parsing ISA extensions from these new
properties, and documenting them in the dt-binding is a requirement for
new extension detection being added to the kernel.

A side effect of the implementation is that the meaning for elements in
"riscv,isa" and in "riscv,isa-extensions" are now tied together as they
are constructed from the same source. The same applies to the ISA string
provided in ACPI tables, but there does not appear to be any strict
definitions of meanings in ACPI land either.

Link: https://lore.kernel.org/qemu-riscv/20230702-eats-scorebook-c951f170d29f@spud/ [0]
Link: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/devicetree/bindings/riscv/extensions.yaml [1]
Signed-off-by: Conor Dooley <conor.dooley@microchip.com>
---
 hw/riscv/sifive_u.c |  7 ++----
 hw/riscv/spike.c    |  6 ++---
 hw/riscv/virt.c     |  6 ++---
 target/riscv/cpu.c  | 53 +++++++++++++++++++++++++++++++++++++++++++++
 target/riscv/cpu.h  |  1 +
 5 files changed, 60 insertions(+), 13 deletions(-)

diff --git a/hw/riscv/sifive_u.c b/hw/riscv/sifive_u.c
index ec76dce6c9..2f227f15bc 100644
--- a/hw/riscv/sifive_u.c
+++ b/hw/riscv/sifive_u.c
@@ -171,7 +171,6 @@ static void create_fdt(SiFiveUState *s, const MemMapEntry *memmap,
         int cpu_phandle = phandle++;
         nodename = g_strdup_printf("/cpus/cpu@%d", cpu);
         char *intc = g_strdup_printf("/cpus/cpu@%d/interrupt-controller", cpu);
-        char *isa;
         qemu_fdt_add_subnode(fdt, nodename);
         /* cpu 0 is the management hart that does not have mmu */
         if (cpu != 0) {
@@ -180,11 +179,10 @@ static void create_fdt(SiFiveUState *s, const MemMapEntry *memmap,
             } else {
                 qemu_fdt_setprop_string(fdt, nodename, "mmu-type", "riscv,sv48");
             }
-            isa = riscv_isa_string(&s->soc.u_cpus.harts[cpu - 1]);
+            riscv_isa_write_fdt(&s->soc.u_cpus.harts[cpu - 1], fdt, nodename);
         } else {
-            isa = riscv_isa_string(&s->soc.e_cpus.harts[0]);
+            riscv_isa_write_fdt(&s->soc.e_cpus.harts[0], fdt, nodename);
         }
-        qemu_fdt_setprop_string(fdt, nodename, "riscv,isa", isa);
         qemu_fdt_setprop_string(fdt, nodename, "compatible", "riscv");
         qemu_fdt_setprop_string(fdt, nodename, "status", "okay");
         qemu_fdt_setprop_cell(fdt, nodename, "reg", cpu);
@@ -194,7 +192,6 @@ static void create_fdt(SiFiveUState *s, const MemMapEntry *memmap,
         qemu_fdt_setprop_string(fdt, intc, "compatible", "riscv,cpu-intc");
         qemu_fdt_setprop(fdt, intc, "interrupt-controller", NULL, 0);
         qemu_fdt_setprop_cell(fdt, intc, "#interrupt-cells", 1);
-        g_free(isa);
         g_free(intc);
         g_free(nodename);
     }
diff --git a/hw/riscv/spike.c b/hw/riscv/spike.c
index 81f7e53aed..64074395bc 100644
--- a/hw/riscv/spike.c
+++ b/hw/riscv/spike.c
@@ -59,7 +59,7 @@ static void create_fdt(SpikeState *s, const MemMapEntry *memmap,
     MachineState *ms = MACHINE(s);
     uint32_t *clint_cells;
     uint32_t cpu_phandle, intc_phandle, phandle = 1;
-    char *name, *mem_name, *clint_name, *clust_name;
+    char *mem_name, *clint_name, *clust_name;
     char *core_name, *cpu_name, *intc_name;
     static const char * const clint_compat[2] = {
         "sifive,clint0", "riscv,clint0"
@@ -113,9 +113,7 @@ static void create_fdt(SpikeState *s, const MemMapEntry *memmap,
             } else {
                 qemu_fdt_setprop_string(fdt, cpu_name, "mmu-type", "riscv,sv48");
             }
-            name = riscv_isa_string(&s->soc[socket].harts[cpu]);
-            qemu_fdt_setprop_string(fdt, cpu_name, "riscv,isa", name);
-            g_free(name);
+            riscv_isa_write_fdt(&s->soc[socket].harts[cpu], fdt, cpu_name);
             qemu_fdt_setprop_string(fdt, cpu_name, "compatible", "riscv");
             qemu_fdt_setprop_string(fdt, cpu_name, "status", "okay");
             qemu_fdt_setprop_cell(fdt, cpu_name, "reg",
diff --git a/hw/riscv/virt.c b/hw/riscv/virt.c
index f9fd1341fc..c47b2d397a 100644
--- a/hw/riscv/virt.c
+++ b/hw/riscv/virt.c
@@ -215,7 +215,7 @@ static void create_fdt_socket_cpus(RISCVVirtState *s, int socket,
     int cpu;
     uint32_t cpu_phandle;
     MachineState *ms = MACHINE(s);
-    char *name, *cpu_name, *core_name, *intc_name, *sv_name;
+    char *cpu_name, *core_name, *intc_name, *sv_name;
     bool is_32_bit = riscv_is_32bit(&s->soc[0]);
     uint8_t satp_mode_max;
 
@@ -236,9 +236,7 @@ static void create_fdt_socket_cpus(RISCVVirtState *s, int socket,
             g_free(sv_name);
         }
 
-        name = riscv_isa_string(cpu_ptr);
-        qemu_fdt_setprop_string(ms->fdt, cpu_name, "riscv,isa", name);
-        g_free(name);
+        riscv_isa_write_fdt(cpu_ptr, ms->fdt, cpu_name);
 
         if (cpu_ptr->cfg.ext_zicbom) {
             qemu_fdt_setprop_cell(ms->fdt, cpu_name, "riscv,cbom-block-size",
diff --git a/target/riscv/cpu.c b/target/riscv/cpu.c
index 5b5da970f2..1c8c81ca4c 100644
--- a/target/riscv/cpu.c
+++ b/target/riscv/cpu.c
@@ -31,6 +31,7 @@
 #include "hw/qdev-properties.h"
 #include "migration/vmstate.h"
 #include "fpu/softfloat-helpers.h"
+#include "sysemu/device_tree.h"
 #include "sysemu/kvm.h"
 #include "sysemu/tcg.h"
 #include "kvm/kvm_riscv.h"
@@ -1875,6 +1876,58 @@ char *riscv_isa_string(RISCVCPU *cpu)
     return isa_str;
 }
 
+#ifndef CONFIG_USER_ONLY
+static char **riscv_isa_extensions_list(RISCVCPU *cpu, int *count)
+{
+    int maxlen = ARRAY_SIZE(riscv_single_letter_exts) + ARRAY_SIZE(isa_edata_arr);
+    char **extensions = g_new(char *, maxlen);
+
+    for (int i = 0; i < sizeof(riscv_single_letter_exts) - 1; i++) {
+        if (cpu->env.misa_ext & RV(riscv_single_letter_exts[i])) {
+            extensions[*count] = g_new(char, 2);
+            snprintf(extensions[*count], 2, "%c",
+                     qemu_tolower(riscv_single_letter_exts[i]));
+            (*count)++;
+        }
+    }
+
+    for (const RISCVIsaExtData *edata = isa_edata_arr; edata->name; edata++) {
+        if (isa_ext_is_enabled(cpu, edata->ext_enable_offset)) {
+            extensions[*count] = g_strdup(edata->name);
+            (*count)++;
+        }
+    }
+
+    return extensions;
+}
+
+void riscv_isa_write_fdt(RISCVCPU *cpu, void *fdt, char *nodename)
+{
+    const size_t maxlen = sizeof("rv128i");
+    g_autofree char *isa_base = g_new(char, maxlen);
+    g_autofree char *riscv_isa;
+    char **isa_extensions;
+    int count = 0;
+    int xlen = 16 << cpu->env.misa_mxl_max;
+
+    riscv_isa = riscv_isa_string(cpu);
+    qemu_fdt_setprop_string(fdt, nodename, "riscv,isa", riscv_isa);
+
+    snprintf(isa_base, maxlen, "rv%di", xlen);
+    qemu_fdt_setprop_string(fdt, nodename, "riscv,isa-base", isa_base);
+
+    isa_extensions = riscv_isa_extensions_list(cpu, &count);
+    qemu_fdt_setprop_string_array(fdt, nodename, "riscv,isa-extensions",
+                                  isa_extensions, count);
+
+    for (int i = 0; i < count; i++) {
+        g_free(isa_extensions[i]);
+    }
+
+    g_free(isa_extensions);
+}
+#endif
+
 #define DEFINE_CPU(type_name, initfn)      \
     {                                      \
         .name = type_name,                 \
diff --git a/target/riscv/cpu.h b/target/riscv/cpu.h
index 5f3955c38d..192d0c2d31 100644
--- a/target/riscv/cpu.h
+++ b/target/riscv/cpu.h
@@ -510,6 +510,7 @@ char *riscv_isa_string(RISCVCPU *cpu);
 #define cpu_mmu_index riscv_cpu_mmu_index
 
 #ifndef CONFIG_USER_ONLY
+void riscv_isa_write_fdt(RISCVCPU *cpu, void *fdt, char *nodename);
 void riscv_cpu_do_transaction_failed(CPUState *cs, hwaddr physaddr,
                                      vaddr addr, unsigned size,
                                      MMUAccessType access_type,
-- 
2.39.2

Re: [PATCH v3 2/2] target/riscv: support new isa extension detection devicetree properties

[Andrew Jones]

On Wed, Jan 10, 2024 at 10:25:37AM +0000, Conor Dooley wrote:
> From: Conor Dooley <conor.dooley@microchip.com>
> 
> A few months ago I submitted a patch to various lists, deprecating
> "riscv,isa" with a lengthy commit message [0] that is now commit
> aeb71e42caae ("dt-bindings: riscv: deprecate riscv,isa") in the Linux
> kernel tree. Primarily, the goal was to replace "riscv,isa" with a new
> set of properties that allowed for strictly defining the meaning of
> various extensions, where "riscv,isa" was tied to whatever definitions
> inflicted upon us by the ISA manual, which have seen some variance over
> time.
> 
> Two new properties were introduced: "riscv,isa-base" and
> "riscv,isa-extensions". The former is a simple string to communicate the
> base ISA implemented by a hart and the latter an array of strings used
> to communicate the set of ISA extensions supported, per the definitions
> of each substring in extensions.yaml [1]. A beneficial side effect was
> also the ability to define vendor extensions in a more "official" way,
> as the ISA manual and other RVI specifications only covered the format
> for vendor extensions in the ISA string, but not the meaning of vendor
> extensions, for obvious reasons.
> 
> Add support for setting these two new properties in the devicetrees for
> the various devicetree platforms supported by QEMU for RISC-V. The Linux
> kernel already supports parsing ISA extensions from these new
> properties, and documenting them in the dt-binding is a requirement for
> new extension detection being added to the kernel.
> 
> A side effect of the implementation is that the meaning for elements in
> "riscv,isa" and in "riscv,isa-extensions" are now tied together as they
> are constructed from the same source. The same applies to the ISA string
> provided in ACPI tables, but there does not appear to be any strict
> definitions of meanings in ACPI land either.
> 
> Link: https://lore.kernel.org/qemu-riscv/20230702-eats-scorebook-c951f170d29f@spud/ [0]
> Link: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/devicetree/bindings/riscv/extensions.yaml [1]
> Signed-off-by: Conor Dooley <conor.dooley@microchip.com>
> ---
>  hw/riscv/sifive_u.c |  7 ++----
>  hw/riscv/spike.c    |  6 ++---
>  hw/riscv/virt.c     |  6 ++---
>  target/riscv/cpu.c  | 53 +++++++++++++++++++++++++++++++++++++++++++++
>  target/riscv/cpu.h  |  1 +
>  5 files changed, 60 insertions(+), 13 deletions(-)
> 
> diff --git a/hw/riscv/sifive_u.c b/hw/riscv/sifive_u.c
> index ec76dce6c9..2f227f15bc 100644
> --- a/hw/riscv/sifive_u.c
> +++ b/hw/riscv/sifive_u.c
> @@ -171,7 +171,6 @@ static void create_fdt(SiFiveUState *s, const MemMapEntry *memmap,
>          int cpu_phandle = phandle++;
>          nodename = g_strdup_printf("/cpus/cpu@%d", cpu);
>          char *intc = g_strdup_printf("/cpus/cpu@%d/interrupt-controller", cpu);
> -        char *isa;
>          qemu_fdt_add_subnode(fdt, nodename);
>          /* cpu 0 is the management hart that does not have mmu */
>          if (cpu != 0) {
> @@ -180,11 +179,10 @@ static void create_fdt(SiFiveUState *s, const MemMapEntry *memmap,
>              } else {
>                  qemu_fdt_setprop_string(fdt, nodename, "mmu-type", "riscv,sv48");
>              }
> -            isa = riscv_isa_string(&s->soc.u_cpus.harts[cpu - 1]);
> +            riscv_isa_write_fdt(&s->soc.u_cpus.harts[cpu - 1], fdt, nodename);
>          } else {
> -            isa = riscv_isa_string(&s->soc.e_cpus.harts[0]);
> +            riscv_isa_write_fdt(&s->soc.e_cpus.harts[0], fdt, nodename);
>          }
> -        qemu_fdt_setprop_string(fdt, nodename, "riscv,isa", isa);
>          qemu_fdt_setprop_string(fdt, nodename, "compatible", "riscv");
>          qemu_fdt_setprop_string(fdt, nodename, "status", "okay");
>          qemu_fdt_setprop_cell(fdt, nodename, "reg", cpu);
> @@ -194,7 +192,6 @@ static void create_fdt(SiFiveUState *s, const MemMapEntry *memmap,
>          qemu_fdt_setprop_string(fdt, intc, "compatible", "riscv,cpu-intc");
>          qemu_fdt_setprop(fdt, intc, "interrupt-controller", NULL, 0);
>          qemu_fdt_setprop_cell(fdt, intc, "#interrupt-cells", 1);
> -        g_free(isa);
>          g_free(intc);
>          g_free(nodename);
>      }
> diff --git a/hw/riscv/spike.c b/hw/riscv/spike.c
> index 81f7e53aed..64074395bc 100644
> --- a/hw/riscv/spike.c
> +++ b/hw/riscv/spike.c
> @@ -59,7 +59,7 @@ static void create_fdt(SpikeState *s, const MemMapEntry *memmap,
>      MachineState *ms = MACHINE(s);
>      uint32_t *clint_cells;
>      uint32_t cpu_phandle, intc_phandle, phandle = 1;
> -    char *name, *mem_name, *clint_name, *clust_name;
> +    char *mem_name, *clint_name, *clust_name;
>      char *core_name, *cpu_name, *intc_name;
>      static const char * const clint_compat[2] = {
>          "sifive,clint0", "riscv,clint0"
> @@ -113,9 +113,7 @@ static void create_fdt(SpikeState *s, const MemMapEntry *memmap,
>              } else {
>                  qemu_fdt_setprop_string(fdt, cpu_name, "mmu-type", "riscv,sv48");
>              }
> -            name = riscv_isa_string(&s->soc[socket].harts[cpu]);
> -            qemu_fdt_setprop_string(fdt, cpu_name, "riscv,isa", name);
> -            g_free(name);
> +            riscv_isa_write_fdt(&s->soc[socket].harts[cpu], fdt, cpu_name);
>              qemu_fdt_setprop_string(fdt, cpu_name, "compatible", "riscv");
>              qemu_fdt_setprop_string(fdt, cpu_name, "status", "okay");
>              qemu_fdt_setprop_cell(fdt, cpu_name, "reg",
> diff --git a/hw/riscv/virt.c b/hw/riscv/virt.c
> index f9fd1341fc..c47b2d397a 100644
> --- a/hw/riscv/virt.c
> +++ b/hw/riscv/virt.c
> @@ -215,7 +215,7 @@ static void create_fdt_socket_cpus(RISCVVirtState *s, int socket,
>      int cpu;
>      uint32_t cpu_phandle;
>      MachineState *ms = MACHINE(s);
> -    char *name, *cpu_name, *core_name, *intc_name, *sv_name;
> +    char *cpu_name, *core_name, *intc_name, *sv_name;
>      bool is_32_bit = riscv_is_32bit(&s->soc[0]);
>      uint8_t satp_mode_max;
>  
> @@ -236,9 +236,7 @@ static void create_fdt_socket_cpus(RISCVVirtState *s, int socket,
>              g_free(sv_name);
>          }
>  
> -        name = riscv_isa_string(cpu_ptr);
> -        qemu_fdt_setprop_string(ms->fdt, cpu_name, "riscv,isa", name);
> -        g_free(name);
> +        riscv_isa_write_fdt(cpu_ptr, ms->fdt, cpu_name);
>  
>          if (cpu_ptr->cfg.ext_zicbom) {
>              qemu_fdt_setprop_cell(ms->fdt, cpu_name, "riscv,cbom-block-size",
> diff --git a/target/riscv/cpu.c b/target/riscv/cpu.c
> index 5b5da970f2..1c8c81ca4c 100644
> --- a/target/riscv/cpu.c
> +++ b/target/riscv/cpu.c
> @@ -31,6 +31,7 @@
>  #include "hw/qdev-properties.h"
>  #include "migration/vmstate.h"
>  #include "fpu/softfloat-helpers.h"
> +#include "sysemu/device_tree.h"
>  #include "sysemu/kvm.h"
>  #include "sysemu/tcg.h"
>  #include "kvm/kvm_riscv.h"
> @@ -1875,6 +1876,58 @@ char *riscv_isa_string(RISCVCPU *cpu)
>      return isa_str;
>  }
>  
> +#ifndef CONFIG_USER_ONLY
> +static char **riscv_isa_extensions_list(RISCVCPU *cpu, int *count)
> +{
> +    int maxlen = ARRAY_SIZE(riscv_single_letter_exts) + ARRAY_SIZE(isa_edata_arr);
> +    char **extensions = g_new(char *, maxlen);
> +
> +    for (int i = 0; i < sizeof(riscv_single_letter_exts) - 1; i++) {
> +        if (cpu->env.misa_ext & RV(riscv_single_letter_exts[i])) {
> +            extensions[*count] = g_new(char, 2);
> +            snprintf(extensions[*count], 2, "%c",
> +                     qemu_tolower(riscv_single_letter_exts[i]));
> +            (*count)++;
> +        }
> +    }
> +
> +    for (const RISCVIsaExtData *edata = isa_edata_arr; edata->name; edata++) {
> +        if (isa_ext_is_enabled(cpu, edata->ext_enable_offset)) {
> +            extensions[*count] = g_strdup(edata->name);
> +            (*count)++;
> +        }
> +    }
> +
> +    return extensions;
> +}
> +
> +void riscv_isa_write_fdt(RISCVCPU *cpu, void *fdt, char *nodename)
> +{
> +    const size_t maxlen = sizeof("rv128i");
> +    g_autofree char *isa_base = g_new(char, maxlen);
> +    g_autofree char *riscv_isa;
> +    char **isa_extensions;
> +    int count = 0;
> +    int xlen = 16 << cpu->env.misa_mxl_max;
> +
> +    riscv_isa = riscv_isa_string(cpu);
> +    qemu_fdt_setprop_string(fdt, nodename, "riscv,isa", riscv_isa);
> +
> +    snprintf(isa_base, maxlen, "rv%di", xlen);
> +    qemu_fdt_setprop_string(fdt, nodename, "riscv,isa-base", isa_base);
> +
> +    isa_extensions = riscv_isa_extensions_list(cpu, &count);
> +    qemu_fdt_setprop_string_array(fdt, nodename, "riscv,isa-extensions",
> +                                  isa_extensions, count);
> +
> +    for (int i = 0; i < count; i++) {
> +        g_free(isa_extensions[i]);
> +    }
> +
> +    g_free(isa_extensions);
> +}
> +#endif
> +
>  #define DEFINE_CPU(type_name, initfn)      \
>      {                                      \
>          .name = type_name,                 \
> diff --git a/target/riscv/cpu.h b/target/riscv/cpu.h
> index 5f3955c38d..192d0c2d31 100644
> --- a/target/riscv/cpu.h
> +++ b/target/riscv/cpu.h
> @@ -510,6 +510,7 @@ char *riscv_isa_string(RISCVCPU *cpu);
>  #define cpu_mmu_index riscv_cpu_mmu_index
>  
>  #ifndef CONFIG_USER_ONLY
> +void riscv_isa_write_fdt(RISCVCPU *cpu, void *fdt, char *nodename);
>  void riscv_cpu_do_transaction_failed(CPUState *cs, hwaddr physaddr,
>                                       vaddr addr, unsigned size,
>                                       MMUAccessType access_type,
> -- 
> 2.39.2
>

Reviewed-by: Andrew Jones <ajones@ventanamicro.com>

Re: [PATCH v3 2/2] target/riscv: support new isa extension detection devicetree properties

[Alistair Francis]

On Wed, Jan 10, 2024 at 8:28 PM Conor Dooley <conor@kernel.org> wrote:
>
> From: Conor Dooley <conor.dooley@microchip.com>
>
> A few months ago I submitted a patch to various lists, deprecating
> "riscv,isa" with a lengthy commit message [0] that is now commit
> aeb71e42caae ("dt-bindings: riscv: deprecate riscv,isa") in the Linux
> kernel tree. Primarily, the goal was to replace "riscv,isa" with a new
> set of properties that allowed for strictly defining the meaning of
> various extensions, where "riscv,isa" was tied to whatever definitions
> inflicted upon us by the ISA manual, which have seen some variance over
> time.
>
> Two new properties were introduced: "riscv,isa-base" and
> "riscv,isa-extensions". The former is a simple string to communicate the
> base ISA implemented by a hart and the latter an array of strings used
> to communicate the set of ISA extensions supported, per the definitions
> of each substring in extensions.yaml [1]. A beneficial side effect was
> also the ability to define vendor extensions in a more "official" way,
> as the ISA manual and other RVI specifications only covered the format
> for vendor extensions in the ISA string, but not the meaning of vendor
> extensions, for obvious reasons.
>
> Add support for setting these two new properties in the devicetrees for
> the various devicetree platforms supported by QEMU for RISC-V. The Linux
> kernel already supports parsing ISA extensions from these new
> properties, and documenting them in the dt-binding is a requirement for
> new extension detection being added to the kernel.
>
> A side effect of the implementation is that the meaning for elements in
> "riscv,isa" and in "riscv,isa-extensions" are now tied together as they
> are constructed from the same source. The same applies to the ISA string
> provided in ACPI tables, but there does not appear to be any strict
> definitions of meanings in ACPI land either.
>
> Link: https://lore.kernel.org/qemu-riscv/20230702-eats-scorebook-c951f170d29f@spud/ [0]
> Link: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/devicetree/bindings/riscv/extensions.yaml [1]
> Signed-off-by: Conor Dooley <conor.dooley@microchip.com>

Reviewed-by: Alistair Francis <alistair.francis@wdc.com>

Alistair

> ---
>  hw/riscv/sifive_u.c |  7 ++----
>  hw/riscv/spike.c    |  6 ++---
>  hw/riscv/virt.c     |  6 ++---
>  target/riscv/cpu.c  | 53 +++++++++++++++++++++++++++++++++++++++++++++
>  target/riscv/cpu.h  |  1 +
>  5 files changed, 60 insertions(+), 13 deletions(-)
>
> diff --git a/hw/riscv/sifive_u.c b/hw/riscv/sifive_u.c
> index ec76dce6c9..2f227f15bc 100644
> --- a/hw/riscv/sifive_u.c
> +++ b/hw/riscv/sifive_u.c
> @@ -171,7 +171,6 @@ static void create_fdt(SiFiveUState *s, const MemMapEntry *memmap,
>          int cpu_phandle = phandle++;
>          nodename = g_strdup_printf("/cpus/cpu@%d", cpu);
>          char *intc = g_strdup_printf("/cpus/cpu@%d/interrupt-controller", cpu);
> -        char *isa;
>          qemu_fdt_add_subnode(fdt, nodename);
>          /* cpu 0 is the management hart that does not have mmu */
>          if (cpu != 0) {
> @@ -180,11 +179,10 @@ static void create_fdt(SiFiveUState *s, const MemMapEntry *memmap,
>              } else {
>                  qemu_fdt_setprop_string(fdt, nodename, "mmu-type", "riscv,sv48");
>              }
> -            isa = riscv_isa_string(&s->soc.u_cpus.harts[cpu - 1]);
> +            riscv_isa_write_fdt(&s->soc.u_cpus.harts[cpu - 1], fdt, nodename);
>          } else {
> -            isa = riscv_isa_string(&s->soc.e_cpus.harts[0]);
> +            riscv_isa_write_fdt(&s->soc.e_cpus.harts[0], fdt, nodename);
>          }
> -        qemu_fdt_setprop_string(fdt, nodename, "riscv,isa", isa);
>          qemu_fdt_setprop_string(fdt, nodename, "compatible", "riscv");
>          qemu_fdt_setprop_string(fdt, nodename, "status", "okay");
>          qemu_fdt_setprop_cell(fdt, nodename, "reg", cpu);
> @@ -194,7 +192,6 @@ static void create_fdt(SiFiveUState *s, const MemMapEntry *memmap,
>          qemu_fdt_setprop_string(fdt, intc, "compatible", "riscv,cpu-intc");
>          qemu_fdt_setprop(fdt, intc, "interrupt-controller", NULL, 0);
>          qemu_fdt_setprop_cell(fdt, intc, "#interrupt-cells", 1);
> -        g_free(isa);
>          g_free(intc);
>          g_free(nodename);
>      }
> diff --git a/hw/riscv/spike.c b/hw/riscv/spike.c
> index 81f7e53aed..64074395bc 100644
> --- a/hw/riscv/spike.c
> +++ b/hw/riscv/spike.c
> @@ -59,7 +59,7 @@ static void create_fdt(SpikeState *s, const MemMapEntry *memmap,
>      MachineState *ms = MACHINE(s);
>      uint32_t *clint_cells;
>      uint32_t cpu_phandle, intc_phandle, phandle = 1;
> -    char *name, *mem_name, *clint_name, *clust_name;
> +    char *mem_name, *clint_name, *clust_name;
>      char *core_name, *cpu_name, *intc_name;
>      static const char * const clint_compat[2] = {
>          "sifive,clint0", "riscv,clint0"
> @@ -113,9 +113,7 @@ static void create_fdt(SpikeState *s, const MemMapEntry *memmap,
>              } else {
>                  qemu_fdt_setprop_string(fdt, cpu_name, "mmu-type", "riscv,sv48");
>              }
> -            name = riscv_isa_string(&s->soc[socket].harts[cpu]);
> -            qemu_fdt_setprop_string(fdt, cpu_name, "riscv,isa", name);
> -            g_free(name);
> +            riscv_isa_write_fdt(&s->soc[socket].harts[cpu], fdt, cpu_name);
>              qemu_fdt_setprop_string(fdt, cpu_name, "compatible", "riscv");
>              qemu_fdt_setprop_string(fdt, cpu_name, "status", "okay");
>              qemu_fdt_setprop_cell(fdt, cpu_name, "reg",
> diff --git a/hw/riscv/virt.c b/hw/riscv/virt.c
> index f9fd1341fc..c47b2d397a 100644
> --- a/hw/riscv/virt.c
> +++ b/hw/riscv/virt.c
> @@ -215,7 +215,7 @@ static void create_fdt_socket_cpus(RISCVVirtState *s, int socket,
>      int cpu;
>      uint32_t cpu_phandle;
>      MachineState *ms = MACHINE(s);
> -    char *name, *cpu_name, *core_name, *intc_name, *sv_name;
> +    char *cpu_name, *core_name, *intc_name, *sv_name;
>      bool is_32_bit = riscv_is_32bit(&s->soc[0]);
>      uint8_t satp_mode_max;
>
> @@ -236,9 +236,7 @@ static void create_fdt_socket_cpus(RISCVVirtState *s, int socket,
>              g_free(sv_name);
>          }
>
> -        name = riscv_isa_string(cpu_ptr);
> -        qemu_fdt_setprop_string(ms->fdt, cpu_name, "riscv,isa", name);
> -        g_free(name);
> +        riscv_isa_write_fdt(cpu_ptr, ms->fdt, cpu_name);
>
>          if (cpu_ptr->cfg.ext_zicbom) {
>              qemu_fdt_setprop_cell(ms->fdt, cpu_name, "riscv,cbom-block-size",
> diff --git a/target/riscv/cpu.c b/target/riscv/cpu.c
> index 5b5da970f2..1c8c81ca4c 100644
> --- a/target/riscv/cpu.c
> +++ b/target/riscv/cpu.c
> @@ -31,6 +31,7 @@
>  #include "hw/qdev-properties.h"
>  #include "migration/vmstate.h"
>  #include "fpu/softfloat-helpers.h"
> +#include "sysemu/device_tree.h"
>  #include "sysemu/kvm.h"
>  #include "sysemu/tcg.h"
>  #include "kvm/kvm_riscv.h"
> @@ -1875,6 +1876,58 @@ char *riscv_isa_string(RISCVCPU *cpu)
>      return isa_str;
>  }
>
> +#ifndef CONFIG_USER_ONLY
> +static char **riscv_isa_extensions_list(RISCVCPU *cpu, int *count)
> +{
> +    int maxlen = ARRAY_SIZE(riscv_single_letter_exts) + ARRAY_SIZE(isa_edata_arr);
> +    char **extensions = g_new(char *, maxlen);
> +
> +    for (int i = 0; i < sizeof(riscv_single_letter_exts) - 1; i++) {
> +        if (cpu->env.misa_ext & RV(riscv_single_letter_exts[i])) {
> +            extensions[*count] = g_new(char, 2);
> +            snprintf(extensions[*count], 2, "%c",
> +                     qemu_tolower(riscv_single_letter_exts[i]));
> +            (*count)++;
> +        }
> +    }
> +
> +    for (const RISCVIsaExtData *edata = isa_edata_arr; edata->name; edata++) {
> +        if (isa_ext_is_enabled(cpu, edata->ext_enable_offset)) {
> +            extensions[*count] = g_strdup(edata->name);
> +            (*count)++;
> +        }
> +    }
> +
> +    return extensions;
> +}
> +
> +void riscv_isa_write_fdt(RISCVCPU *cpu, void *fdt, char *nodename)
> +{
> +    const size_t maxlen = sizeof("rv128i");
> +    g_autofree char *isa_base = g_new(char, maxlen);
> +    g_autofree char *riscv_isa;
> +    char **isa_extensions;
> +    int count = 0;
> +    int xlen = 16 << cpu->env.misa_mxl_max;
> +
> +    riscv_isa = riscv_isa_string(cpu);
> +    qemu_fdt_setprop_string(fdt, nodename, "riscv,isa", riscv_isa);
> +
> +    snprintf(isa_base, maxlen, "rv%di", xlen);
> +    qemu_fdt_setprop_string(fdt, nodename, "riscv,isa-base", isa_base);
> +
> +    isa_extensions = riscv_isa_extensions_list(cpu, &count);
> +    qemu_fdt_setprop_string_array(fdt, nodename, "riscv,isa-extensions",
> +                                  isa_extensions, count);
> +
> +    for (int i = 0; i < count; i++) {
> +        g_free(isa_extensions[i]);
> +    }
> +
> +    g_free(isa_extensions);
> +}
> +#endif
> +
>  #define DEFINE_CPU(type_name, initfn)      \
>      {                                      \
>          .name = type_name,                 \
> diff --git a/target/riscv/cpu.h b/target/riscv/cpu.h
> index 5f3955c38d..192d0c2d31 100644
> --- a/target/riscv/cpu.h
> +++ b/target/riscv/cpu.h
> @@ -510,6 +510,7 @@ char *riscv_isa_string(RISCVCPU *cpu);
>  #define cpu_mmu_index riscv_cpu_mmu_index
>
>  #ifndef CONFIG_USER_ONLY
> +void riscv_isa_write_fdt(RISCVCPU *cpu, void *fdt, char *nodename);
>  void riscv_cpu_do_transaction_failed(CPUState *cs, hwaddr physaddr,
>                                       vaddr addr, unsigned size,
>                                       MMUAccessType access_type,
> --
> 2.39.2
>
>

Re: [PATCH v3 0/2] riscv: support new isa extension detection devicetree properties

[Alistair Francis]

On Wed, Jan 10, 2024 at 8:27 PM Conor Dooley <conor@kernel.org> wrote:
>
> From: Conor Dooley <conor.dooley@microchip.com>
>
> Making it a series to keep the standalone change to riscv_isa_string()
> that Drew reported separate.
>
> Changes in v3:
> - g_free() isa_extensions too
> - use misa_mxl_max rather than the compile target for the base isa
> - add a new patch changing riscv_isa_string() to do the same
> - drop a null check that cannot be null
> - rebased on top of Alistair's next branch

Do you mind rebasing on
https://github.com/alistair23/qemu/tree/riscv-to-apply.next again?
There was a big re-org recently so lots of rebasing is required

Alistair

>
> Changes in v2:
> - use g_strdup() for multiletter extension string copying
> - wrap stuff in #ifndef to prevent breaking the user mode build
> - rename riscv_isa_set_props() -> riscv_isa_write_fdt()
>
> CC: Alistair Francis <Alistair.Francis@wdc.com>
> CC: Bin Meng <bin.meng@windriver.com>
> CC: Palmer Dabbelt <palmer@dabbelt.com>
> CC: Weiwei Li <liwei1518@gmail.com>
> CC: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
> CC: Andrew Jones <ajones@ventanamicro.com>
> CC: Liu Zhiwei <zhiwei_liu@linux.alibaba.com>
> CC: qemu-riscv@nongnu.org
> CC: qemu-devel@nongnu.org
>
> Conor Dooley (2):
>   target/riscv: use misa_mxl_max to populate isa string rather than
>     TARGET_LONG_BITS
>   target/riscv: support new isa extension detection devicetree
>     properties
>
>  hw/riscv/sifive_u.c |  7 ++----
>  hw/riscv/spike.c    |  6 ++---
>  hw/riscv/virt.c     |  6 ++---
>  target/riscv/cpu.c  | 57 ++++++++++++++++++++++++++++++++++++++++++++-
>  target/riscv/cpu.h  |  1 +
>  5 files changed, 63 insertions(+), 14 deletions(-)
>
> --
> 2.39.2
>
>

Re: [PATCH v3 0/2] riscv: support new isa extension detection devicetree properties

[Conor Dooley]

[-- Attachment #1: Type: text/plain, Size: 947 bytes --]

On Mon, Jan 22, 2024 at 03:24:19PM +1000, Alistair Francis wrote:
> On Wed, Jan 10, 2024 at 8:27 PM Conor Dooley <conor@kernel.org> wrote:
> >
> > From: Conor Dooley <conor.dooley@microchip.com>
> >
> > Making it a series to keep the standalone change to riscv_isa_string()
> > that Drew reported separate.
> >
> > Changes in v3:
> > - g_free() isa_extensions too
> > - use misa_mxl_max rather than the compile target for the base isa
> > - add a new patch changing riscv_isa_string() to do the same
> > - drop a null check that cannot be null
> > - rebased on top of Alistair's next branch
> 
> Do you mind rebasing on
> https://github.com/alistair23/qemu/tree/riscv-to-apply.next again?
> There was a big re-org recently so lots of rebasing is required

I can, sure. Do you want me to introduce the macro that I mentioned in
the first patch as a helper for misa_mxl_max -> width conversions when I
do?

Thanks,
Conor.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]

Re: [PATCH v3 0/2] riscv: support new isa extension detection devicetree properties

[Alistair Francis]

On Mon, Jan 22, 2024 at 10:25 PM Conor Dooley <conor@kernel.org> wrote:
>
> On Mon, Jan 22, 2024 at 03:24:19PM +1000, Alistair Francis wrote:
> > On Wed, Jan 10, 2024 at 8:27 PM Conor Dooley <conor@kernel.org> wrote:
> > >
> > > From: Conor Dooley <conor.dooley@microchip.com>
> > >
> > > Making it a series to keep the standalone change to riscv_isa_string()
> > > that Drew reported separate.
> > >
> > > Changes in v3:
> > > - g_free() isa_extensions too
> > > - use misa_mxl_max rather than the compile target for the base isa
> > > - add a new patch changing riscv_isa_string() to do the same
> > > - drop a null check that cannot be null
> > > - rebased on top of Alistair's next branch
> >
> > Do you mind rebasing on
> > https://github.com/alistair23/qemu/tree/riscv-to-apply.next again?
> > There was a big re-org recently so lots of rebasing is required
>
> I can, sure. Do you want me to introduce the macro that I mentioned in
> the first patch as a helper for misa_mxl_max -> width conversions when I
> do?

Yes please!

Alistair

>
> Thanks,
> Conor.