| Message ID | 20220429153431.308829-4-apatel@ventanamicro.com |
|---|---|
| State | New |
| Headers | show |
| Series | QEMU RISC-V priv spec version fixes | expand |
Hi Anup, If we want to limit the generated ISA string to/after a specific privilege spec version. Shouldn't we also check the privilege spec version when these extensions are enabled? Otherwise, it's possible that one extension is enabled, but the privilege spec version is smaller than the one in which the extension is supported. (This is possible if user specifies the privileged spec version through the command line.) The ISA string therefore won't include the enabled extension. Regards, Frank Chang On Fri, Apr 29, 2022 at 11:49 PM Anup Patel <apatel@ventanamicro.com> wrote: > Most of the multi-letter extensions (such as Svpbmt, Svnapot, Svinval, > etc) are only available after Priv spec v1.12 so ISA string generation > should check the minimum required priv spec version for all extensions. > > Fixes: a775398be2e ("target/riscv: Add isa extenstion strings to the > device tree") > Signed-off-by: Anup Patel <apatel@ventanamicro.com> > --- > target/riscv/cpu.c | 36 +++++++++++++++++++----------------- > 1 file changed, 19 insertions(+), 17 deletions(-) > > diff --git a/target/riscv/cpu.c b/target/riscv/cpu.c > index 02ee7d45d8..d8c88b96bc 100644 > --- a/target/riscv/cpu.c > +++ b/target/riscv/cpu.c > @@ -44,6 +44,7 @@ static const char riscv_single_letter_exts[] = > "IEMAFDQCPVH"; > struct isa_ext_data { > const char *name; > bool enabled; > + uint32_t min_priv_ver; > }; > > const char * const riscv_int_regnames[] = { > @@ -974,7 +975,7 @@ static void riscv_cpu_class_init(ObjectClass *c, void > *data) > device_class_set_props(dc, riscv_cpu_properties); > } > > -#define ISA_EDATA_ENTRY(name, prop) {#name, cpu->cfg.prop} > +#define ISA_EDATA_ENTRY(name, prop, priv) {#name, cpu->cfg.prop, priv} > > static void riscv_isa_string_ext(RISCVCPU *cpu, char **isa_str, int > max_str_len) > { > @@ -1000,25 +1001,26 @@ static void riscv_isa_string_ext(RISCVCPU *cpu, > char **isa_str, int max_str_len) > * extensions by an underscore. > */ > struct isa_ext_data isa_edata_arr[] = { > - ISA_EDATA_ENTRY(zfh, ext_zfh), > - ISA_EDATA_ENTRY(zfhmin, ext_zfhmin), > - ISA_EDATA_ENTRY(zfinx, ext_zfinx), > - ISA_EDATA_ENTRY(zhinx, ext_zhinx), > - ISA_EDATA_ENTRY(zhinxmin, ext_zhinxmin), > - ISA_EDATA_ENTRY(zdinx, ext_zdinx), > - ISA_EDATA_ENTRY(zba, ext_zba), > - ISA_EDATA_ENTRY(zbb, ext_zbb), > - ISA_EDATA_ENTRY(zbc, ext_zbc), > - ISA_EDATA_ENTRY(zbs, ext_zbs), > - ISA_EDATA_ENTRY(zve32f, ext_zve32f), > - ISA_EDATA_ENTRY(zve64f, ext_zve64f), > - ISA_EDATA_ENTRY(svinval, ext_svinval), > - ISA_EDATA_ENTRY(svnapot, ext_svnapot), > - ISA_EDATA_ENTRY(svpbmt, ext_svpbmt), > + ISA_EDATA_ENTRY(zfh, ext_zfh, PRIV_VERSION_1_12_0), > + ISA_EDATA_ENTRY(zfhmin, ext_zfhmin, PRIV_VERSION_1_12_0), > + ISA_EDATA_ENTRY(zfinx, ext_zfinx, PRIV_VERSION_1_12_0), > + ISA_EDATA_ENTRY(zhinx, ext_zhinx, PRIV_VERSION_1_12_0), > + ISA_EDATA_ENTRY(zhinxmin, ext_zhinxmin, PRIV_VERSION_1_12_0), > + ISA_EDATA_ENTRY(zdinx, ext_zdinx, PRIV_VERSION_1_12_0), > + ISA_EDATA_ENTRY(zba, ext_zba, PRIV_VERSION_1_12_0), > + ISA_EDATA_ENTRY(zbb, ext_zbb, PRIV_VERSION_1_12_0), > + ISA_EDATA_ENTRY(zbc, ext_zbc, PRIV_VERSION_1_12_0), > + ISA_EDATA_ENTRY(zbs, ext_zbs, PRIV_VERSION_1_12_0), > + ISA_EDATA_ENTRY(zve32f, ext_zve32f, PRIV_VERSION_1_12_0), > + ISA_EDATA_ENTRY(zve64f, ext_zve64f, PRIV_VERSION_1_12_0), > + ISA_EDATA_ENTRY(svinval, ext_svinval, PRIV_VERSION_1_12_0), > + ISA_EDATA_ENTRY(svnapot, ext_svnapot, PRIV_VERSION_1_12_0), > + ISA_EDATA_ENTRY(svpbmt, ext_svpbmt, PRIV_VERSION_1_12_0), > }; > > for (i = 0; i < ARRAY_SIZE(isa_edata_arr); i++) { > - if (isa_edata_arr[i].enabled) { > + if (isa_edata_arr[i].enabled && > + cpu->env.priv_ver >= isa_edata_arr[i].min_priv_ver) { > new = g_strconcat(old, "_", isa_edata_arr[i].name, NULL); > g_free(old); > old = new; > -- > 2.34.1 > > >
On Sat, Apr 30, 2022 at 8:39 AM Frank Chang <frank.chang@sifive.com> wrote: > > Hi Anup, > > If we want to limit the generated ISA string to/after a specific privilege spec version. > Shouldn't we also check the privilege spec version when these extensions are enabled? > Otherwise, it's possible that one extension is enabled, > but the privilege spec version is smaller than the one in which the extension is supported. > (This is possible if user specifies the privileged spec version through the command line.) > The ISA string therefore won't include the enabled extension. This patch is only a temporary fix for the sifive_u machine where I am seeing some of these new extensions available in ISA string. We need a separate series to have the priv spec version influence individual extension enabling/disabling. Regards, Anup > > Regards, > Frank Chang > > > On Fri, Apr 29, 2022 at 11:49 PM Anup Patel <apatel@ventanamicro.com> wrote: >> >> Most of the multi-letter extensions (such as Svpbmt, Svnapot, Svinval, >> etc) are only available after Priv spec v1.12 so ISA string generation >> should check the minimum required priv spec version for all extensions. >> >> Fixes: a775398be2e ("target/riscv: Add isa extenstion strings to the >> device tree") >> Signed-off-by: Anup Patel <apatel@ventanamicro.com> >> --- >> target/riscv/cpu.c | 36 +++++++++++++++++++----------------- >> 1 file changed, 19 insertions(+), 17 deletions(-) >> >> diff --git a/target/riscv/cpu.c b/target/riscv/cpu.c >> index 02ee7d45d8..d8c88b96bc 100644 >> --- a/target/riscv/cpu.c >> +++ b/target/riscv/cpu.c >> @@ -44,6 +44,7 @@ static const char riscv_single_letter_exts[] = "IEMAFDQCPVH"; >> struct isa_ext_data { >> const char *name; >> bool enabled; >> + uint32_t min_priv_ver; >> }; >> >> const char * const riscv_int_regnames[] = { >> @@ -974,7 +975,7 @@ static void riscv_cpu_class_init(ObjectClass *c, void *data) >> device_class_set_props(dc, riscv_cpu_properties); >> } >> >> -#define ISA_EDATA_ENTRY(name, prop) {#name, cpu->cfg.prop} >> +#define ISA_EDATA_ENTRY(name, prop, priv) {#name, cpu->cfg.prop, priv} >> >> static void riscv_isa_string_ext(RISCVCPU *cpu, char **isa_str, int max_str_len) >> { >> @@ -1000,25 +1001,26 @@ static void riscv_isa_string_ext(RISCVCPU *cpu, char **isa_str, int max_str_len) >> * extensions by an underscore. >> */ >> struct isa_ext_data isa_edata_arr[] = { >> - ISA_EDATA_ENTRY(zfh, ext_zfh), >> - ISA_EDATA_ENTRY(zfhmin, ext_zfhmin), >> - ISA_EDATA_ENTRY(zfinx, ext_zfinx), >> - ISA_EDATA_ENTRY(zhinx, ext_zhinx), >> - ISA_EDATA_ENTRY(zhinxmin, ext_zhinxmin), >> - ISA_EDATA_ENTRY(zdinx, ext_zdinx), >> - ISA_EDATA_ENTRY(zba, ext_zba), >> - ISA_EDATA_ENTRY(zbb, ext_zbb), >> - ISA_EDATA_ENTRY(zbc, ext_zbc), >> - ISA_EDATA_ENTRY(zbs, ext_zbs), >> - ISA_EDATA_ENTRY(zve32f, ext_zve32f), >> - ISA_EDATA_ENTRY(zve64f, ext_zve64f), >> - ISA_EDATA_ENTRY(svinval, ext_svinval), >> - ISA_EDATA_ENTRY(svnapot, ext_svnapot), >> - ISA_EDATA_ENTRY(svpbmt, ext_svpbmt), >> + ISA_EDATA_ENTRY(zfh, ext_zfh, PRIV_VERSION_1_12_0), >> + ISA_EDATA_ENTRY(zfhmin, ext_zfhmin, PRIV_VERSION_1_12_0), >> + ISA_EDATA_ENTRY(zfinx, ext_zfinx, PRIV_VERSION_1_12_0), >> + ISA_EDATA_ENTRY(zhinx, ext_zhinx, PRIV_VERSION_1_12_0), >> + ISA_EDATA_ENTRY(zhinxmin, ext_zhinxmin, PRIV_VERSION_1_12_0), >> + ISA_EDATA_ENTRY(zdinx, ext_zdinx, PRIV_VERSION_1_12_0), >> + ISA_EDATA_ENTRY(zba, ext_zba, PRIV_VERSION_1_12_0), >> + ISA_EDATA_ENTRY(zbb, ext_zbb, PRIV_VERSION_1_12_0), >> + ISA_EDATA_ENTRY(zbc, ext_zbc, PRIV_VERSION_1_12_0), >> + ISA_EDATA_ENTRY(zbs, ext_zbs, PRIV_VERSION_1_12_0), >> + ISA_EDATA_ENTRY(zve32f, ext_zve32f, PRIV_VERSION_1_12_0), >> + ISA_EDATA_ENTRY(zve64f, ext_zve64f, PRIV_VERSION_1_12_0), >> + ISA_EDATA_ENTRY(svinval, ext_svinval, PRIV_VERSION_1_12_0), >> + ISA_EDATA_ENTRY(svnapot, ext_svnapot, PRIV_VERSION_1_12_0), >> + ISA_EDATA_ENTRY(svpbmt, ext_svpbmt, PRIV_VERSION_1_12_0), >> }; >> >> for (i = 0; i < ARRAY_SIZE(isa_edata_arr); i++) { >> - if (isa_edata_arr[i].enabled) { >> + if (isa_edata_arr[i].enabled && >> + cpu->env.priv_ver >= isa_edata_arr[i].min_priv_ver) { >> new = g_strconcat(old, "_", isa_edata_arr[i].name, NULL); >> g_free(old); >> old = new; >> -- >> 2.34.1 >> >>
On Sat, Apr 30, 2022 at 12:30 PM Anup Patel <anup@brainfault.org> wrote: > On Sat, Apr 30, 2022 at 8:39 AM Frank Chang <frank.chang@sifive.com> > wrote: > > > > Hi Anup, > > > > If we want to limit the generated ISA string to/after a specific > privilege spec version. > > Shouldn't we also check the privilege spec version when these extensions > are enabled? > > Otherwise, it's possible that one extension is enabled, > > but the privilege spec version is smaller than the one in which the > extension is supported. > > (This is possible if user specifies the privileged spec version through > the command line.) > > The ISA string therefore won't include the enabled extension. > > This patch is only a temporary fix for the sifive_u machine where I am > seeing some > of these new extensions available in ISA string. > > We need a separate series to have the priv spec version influence > individual extension > enabling/disabling. > If that's the case, Reviewed-by: Frank Chang <frank.chang@sifive.com> > > Regards, > Anup > > > > > Regards, > > Frank Chang > > > > > > On Fri, Apr 29, 2022 at 11:49 PM Anup Patel <apatel@ventanamicro.com> > wrote: > >> > >> Most of the multi-letter extensions (such as Svpbmt, Svnapot, Svinval, > >> etc) are only available after Priv spec v1.12 so ISA string generation > >> should check the minimum required priv spec version for all extensions. > >> > >> Fixes: a775398be2e ("target/riscv: Add isa extenstion strings to the > >> device tree") > >> Signed-off-by: Anup Patel <apatel@ventanamicro.com> > >> --- > >> target/riscv/cpu.c | 36 +++++++++++++++++++----------------- > >> 1 file changed, 19 insertions(+), 17 deletions(-) > >> > >> diff --git a/target/riscv/cpu.c b/target/riscv/cpu.c > >> index 02ee7d45d8..d8c88b96bc 100644 > >> --- a/target/riscv/cpu.c > >> +++ b/target/riscv/cpu.c > >> @@ -44,6 +44,7 @@ static const char riscv_single_letter_exts[] = > "IEMAFDQCPVH"; > >> struct isa_ext_data { > >> const char *name; > >> bool enabled; > >> + uint32_t min_priv_ver; > >> }; > >> > >> const char * const riscv_int_regnames[] = { > >> @@ -974,7 +975,7 @@ static void riscv_cpu_class_init(ObjectClass *c, > void *data) > >> device_class_set_props(dc, riscv_cpu_properties); > >> } > >> > >> -#define ISA_EDATA_ENTRY(name, prop) {#name, cpu->cfg.prop} > >> +#define ISA_EDATA_ENTRY(name, prop, priv) {#name, cpu->cfg.prop, priv} > >> > >> static void riscv_isa_string_ext(RISCVCPU *cpu, char **isa_str, int > max_str_len) > >> { > >> @@ -1000,25 +1001,26 @@ static void riscv_isa_string_ext(RISCVCPU *cpu, > char **isa_str, int max_str_len) > >> * extensions by an underscore. > >> */ > >> struct isa_ext_data isa_edata_arr[] = { > >> - ISA_EDATA_ENTRY(zfh, ext_zfh), > >> - ISA_EDATA_ENTRY(zfhmin, ext_zfhmin), > >> - ISA_EDATA_ENTRY(zfinx, ext_zfinx), > >> - ISA_EDATA_ENTRY(zhinx, ext_zhinx), > >> - ISA_EDATA_ENTRY(zhinxmin, ext_zhinxmin), > >> - ISA_EDATA_ENTRY(zdinx, ext_zdinx), > >> - ISA_EDATA_ENTRY(zba, ext_zba), > >> - ISA_EDATA_ENTRY(zbb, ext_zbb), > >> - ISA_EDATA_ENTRY(zbc, ext_zbc), > >> - ISA_EDATA_ENTRY(zbs, ext_zbs), > >> - ISA_EDATA_ENTRY(zve32f, ext_zve32f), > >> - ISA_EDATA_ENTRY(zve64f, ext_zve64f), > >> - ISA_EDATA_ENTRY(svinval, ext_svinval), > >> - ISA_EDATA_ENTRY(svnapot, ext_svnapot), > >> - ISA_EDATA_ENTRY(svpbmt, ext_svpbmt), > >> + ISA_EDATA_ENTRY(zfh, ext_zfh, PRIV_VERSION_1_12_0), > >> + ISA_EDATA_ENTRY(zfhmin, ext_zfhmin, PRIV_VERSION_1_12_0), > >> + ISA_EDATA_ENTRY(zfinx, ext_zfinx, PRIV_VERSION_1_12_0), > >> + ISA_EDATA_ENTRY(zhinx, ext_zhinx, PRIV_VERSION_1_12_0), > >> + ISA_EDATA_ENTRY(zhinxmin, ext_zhinxmin, PRIV_VERSION_1_12_0), > >> + ISA_EDATA_ENTRY(zdinx, ext_zdinx, PRIV_VERSION_1_12_0), > >> + ISA_EDATA_ENTRY(zba, ext_zba, PRIV_VERSION_1_12_0), > >> + ISA_EDATA_ENTRY(zbb, ext_zbb, PRIV_VERSION_1_12_0), > >> + ISA_EDATA_ENTRY(zbc, ext_zbc, PRIV_VERSION_1_12_0), > >> + ISA_EDATA_ENTRY(zbs, ext_zbs, PRIV_VERSION_1_12_0), > >> + ISA_EDATA_ENTRY(zve32f, ext_zve32f, PRIV_VERSION_1_12_0), > >> + ISA_EDATA_ENTRY(zve64f, ext_zve64f, PRIV_VERSION_1_12_0), > >> + ISA_EDATA_ENTRY(svinval, ext_svinval, PRIV_VERSION_1_12_0), > >> + ISA_EDATA_ENTRY(svnapot, ext_svnapot, PRIV_VERSION_1_12_0), > >> + ISA_EDATA_ENTRY(svpbmt, ext_svpbmt, PRIV_VERSION_1_12_0), > >> }; > >> > >> for (i = 0; i < ARRAY_SIZE(isa_edata_arr); i++) { > >> - if (isa_edata_arr[i].enabled) { > >> + if (isa_edata_arr[i].enabled && > >> + cpu->env.priv_ver >= isa_edata_arr[i].min_priv_ver) { > >> new = g_strconcat(old, "_", isa_edata_arr[i].name, NULL); > >> g_free(old); > >> old = new; > >> -- > >> 2.34.1 > >> > >> >
On Sat, Apr 30, 2022 at 2:31 PM Anup Patel <anup@brainfault.org> wrote: > > On Sat, Apr 30, 2022 at 8:39 AM Frank Chang <frank.chang@sifive.com> wrote: > > > > Hi Anup, > > > > If we want to limit the generated ISA string to/after a specific privilege spec version. > > Shouldn't we also check the privilege spec version when these extensions are enabled? > > Otherwise, it's possible that one extension is enabled, > > but the privilege spec version is smaller than the one in which the extension is supported. > > (This is possible if user specifies the privileged spec version through the command line.) > > The ISA string therefore won't include the enabled extension. > > This patch is only a temporary fix for the sifive_u machine where I am > seeing some > of these new extensions available in ISA string. > > We need a separate series to have the priv spec version influence > individual extension > enabling/disabling. I agree with Frank, I think it makes more sense to just never enable the extension instead of not telling the guest it's enabled. Alistair > > Regards, > Anup > > > > > Regards, > > Frank Chang > > > > > > On Fri, Apr 29, 2022 at 11:49 PM Anup Patel <apatel@ventanamicro.com> wrote: > >> > >> Most of the multi-letter extensions (such as Svpbmt, Svnapot, Svinval, > >> etc) are only available after Priv spec v1.12 so ISA string generation > >> should check the minimum required priv spec version for all extensions. > >> > >> Fixes: a775398be2e ("target/riscv: Add isa extenstion strings to the > >> device tree") > >> Signed-off-by: Anup Patel <apatel@ventanamicro.com> > >> --- > >> target/riscv/cpu.c | 36 +++++++++++++++++++----------------- > >> 1 file changed, 19 insertions(+), 17 deletions(-) > >> > >> diff --git a/target/riscv/cpu.c b/target/riscv/cpu.c > >> index 02ee7d45d8..d8c88b96bc 100644 > >> --- a/target/riscv/cpu.c > >> +++ b/target/riscv/cpu.c > >> @@ -44,6 +44,7 @@ static const char riscv_single_letter_exts[] = "IEMAFDQCPVH"; > >> struct isa_ext_data { > >> const char *name; > >> bool enabled; > >> + uint32_t min_priv_ver; > >> }; > >> > >> const char * const riscv_int_regnames[] = { > >> @@ -974,7 +975,7 @@ static void riscv_cpu_class_init(ObjectClass *c, void *data) > >> device_class_set_props(dc, riscv_cpu_properties); > >> } > >> > >> -#define ISA_EDATA_ENTRY(name, prop) {#name, cpu->cfg.prop} > >> +#define ISA_EDATA_ENTRY(name, prop, priv) {#name, cpu->cfg.prop, priv} > >> > >> static void riscv_isa_string_ext(RISCVCPU *cpu, char **isa_str, int max_str_len) > >> { > >> @@ -1000,25 +1001,26 @@ static void riscv_isa_string_ext(RISCVCPU *cpu, char **isa_str, int max_str_len) > >> * extensions by an underscore. > >> */ > >> struct isa_ext_data isa_edata_arr[] = { > >> - ISA_EDATA_ENTRY(zfh, ext_zfh), > >> - ISA_EDATA_ENTRY(zfhmin, ext_zfhmin), > >> - ISA_EDATA_ENTRY(zfinx, ext_zfinx), > >> - ISA_EDATA_ENTRY(zhinx, ext_zhinx), > >> - ISA_EDATA_ENTRY(zhinxmin, ext_zhinxmin), > >> - ISA_EDATA_ENTRY(zdinx, ext_zdinx), > >> - ISA_EDATA_ENTRY(zba, ext_zba), > >> - ISA_EDATA_ENTRY(zbb, ext_zbb), > >> - ISA_EDATA_ENTRY(zbc, ext_zbc), > >> - ISA_EDATA_ENTRY(zbs, ext_zbs), > >> - ISA_EDATA_ENTRY(zve32f, ext_zve32f), > >> - ISA_EDATA_ENTRY(zve64f, ext_zve64f), > >> - ISA_EDATA_ENTRY(svinval, ext_svinval), > >> - ISA_EDATA_ENTRY(svnapot, ext_svnapot), > >> - ISA_EDATA_ENTRY(svpbmt, ext_svpbmt), > >> + ISA_EDATA_ENTRY(zfh, ext_zfh, PRIV_VERSION_1_12_0), > >> + ISA_EDATA_ENTRY(zfhmin, ext_zfhmin, PRIV_VERSION_1_12_0), > >> + ISA_EDATA_ENTRY(zfinx, ext_zfinx, PRIV_VERSION_1_12_0), > >> + ISA_EDATA_ENTRY(zhinx, ext_zhinx, PRIV_VERSION_1_12_0), > >> + ISA_EDATA_ENTRY(zhinxmin, ext_zhinxmin, PRIV_VERSION_1_12_0), > >> + ISA_EDATA_ENTRY(zdinx, ext_zdinx, PRIV_VERSION_1_12_0), > >> + ISA_EDATA_ENTRY(zba, ext_zba, PRIV_VERSION_1_12_0), > >> + ISA_EDATA_ENTRY(zbb, ext_zbb, PRIV_VERSION_1_12_0), > >> + ISA_EDATA_ENTRY(zbc, ext_zbc, PRIV_VERSION_1_12_0), > >> + ISA_EDATA_ENTRY(zbs, ext_zbs, PRIV_VERSION_1_12_0), > >> + ISA_EDATA_ENTRY(zve32f, ext_zve32f, PRIV_VERSION_1_12_0), > >> + ISA_EDATA_ENTRY(zve64f, ext_zve64f, PRIV_VERSION_1_12_0), > >> + ISA_EDATA_ENTRY(svinval, ext_svinval, PRIV_VERSION_1_12_0), > >> + ISA_EDATA_ENTRY(svnapot, ext_svnapot, PRIV_VERSION_1_12_0), > >> + ISA_EDATA_ENTRY(svpbmt, ext_svpbmt, PRIV_VERSION_1_12_0), > >> }; > >> > >> for (i = 0; i < ARRAY_SIZE(isa_edata_arr); i++) { > >> - if (isa_edata_arr[i].enabled) { > >> + if (isa_edata_arr[i].enabled && > >> + cpu->env.priv_ver >= isa_edata_arr[i].min_priv_ver) { > >> new = g_strconcat(old, "_", isa_edata_arr[i].name, NULL); > >> g_free(old); > >> old = new; > >> -- > >> 2.34.1 > >> > >> >
diff --git a/target/riscv/cpu.c b/target/riscv/cpu.c index 02ee7d45d8..d8c88b96bc 100644 --- a/target/riscv/cpu.c +++ b/target/riscv/cpu.c @@ -44,6 +44,7 @@ static const char riscv_single_letter_exts[] = "IEMAFDQCPVH"; struct isa_ext_data { const char *name; bool enabled; + uint32_t min_priv_ver; }; const char * const riscv_int_regnames[] = { @@ -974,7 +975,7 @@ static void riscv_cpu_class_init(ObjectClass *c, void *data) device_class_set_props(dc, riscv_cpu_properties); } -#define ISA_EDATA_ENTRY(name, prop) {#name, cpu->cfg.prop} +#define ISA_EDATA_ENTRY(name, prop, priv) {#name, cpu->cfg.prop, priv} static void riscv_isa_string_ext(RISCVCPU *cpu, char **isa_str, int max_str_len) { @@ -1000,25 +1001,26 @@ static void riscv_isa_string_ext(RISCVCPU *cpu, char **isa_str, int max_str_len) * extensions by an underscore. */ struct isa_ext_data isa_edata_arr[] = { - ISA_EDATA_ENTRY(zfh, ext_zfh), - ISA_EDATA_ENTRY(zfhmin, ext_zfhmin), - ISA_EDATA_ENTRY(zfinx, ext_zfinx), - ISA_EDATA_ENTRY(zhinx, ext_zhinx), - ISA_EDATA_ENTRY(zhinxmin, ext_zhinxmin), - ISA_EDATA_ENTRY(zdinx, ext_zdinx), - ISA_EDATA_ENTRY(zba, ext_zba), - ISA_EDATA_ENTRY(zbb, ext_zbb), - ISA_EDATA_ENTRY(zbc, ext_zbc), - ISA_EDATA_ENTRY(zbs, ext_zbs), - ISA_EDATA_ENTRY(zve32f, ext_zve32f), - ISA_EDATA_ENTRY(zve64f, ext_zve64f), - ISA_EDATA_ENTRY(svinval, ext_svinval), - ISA_EDATA_ENTRY(svnapot, ext_svnapot), - ISA_EDATA_ENTRY(svpbmt, ext_svpbmt), + ISA_EDATA_ENTRY(zfh, ext_zfh, PRIV_VERSION_1_12_0), + ISA_EDATA_ENTRY(zfhmin, ext_zfhmin, PRIV_VERSION_1_12_0), + ISA_EDATA_ENTRY(zfinx, ext_zfinx, PRIV_VERSION_1_12_0), + ISA_EDATA_ENTRY(zhinx, ext_zhinx, PRIV_VERSION_1_12_0), + ISA_EDATA_ENTRY(zhinxmin, ext_zhinxmin, PRIV_VERSION_1_12_0), + ISA_EDATA_ENTRY(zdinx, ext_zdinx, PRIV_VERSION_1_12_0), + ISA_EDATA_ENTRY(zba, ext_zba, PRIV_VERSION_1_12_0), + ISA_EDATA_ENTRY(zbb, ext_zbb, PRIV_VERSION_1_12_0), + ISA_EDATA_ENTRY(zbc, ext_zbc, PRIV_VERSION_1_12_0), + ISA_EDATA_ENTRY(zbs, ext_zbs, PRIV_VERSION_1_12_0), + ISA_EDATA_ENTRY(zve32f, ext_zve32f, PRIV_VERSION_1_12_0), + ISA_EDATA_ENTRY(zve64f, ext_zve64f, PRIV_VERSION_1_12_0), + ISA_EDATA_ENTRY(svinval, ext_svinval, PRIV_VERSION_1_12_0), + ISA_EDATA_ENTRY(svnapot, ext_svnapot, PRIV_VERSION_1_12_0), + ISA_EDATA_ENTRY(svpbmt, ext_svpbmt, PRIV_VERSION_1_12_0), }; for (i = 0; i < ARRAY_SIZE(isa_edata_arr); i++) { - if (isa_edata_arr[i].enabled) { + if (isa_edata_arr[i].enabled && + cpu->env.priv_ver >= isa_edata_arr[i].min_priv_ver) { new = g_strconcat(old, "_", isa_edata_arr[i].name, NULL); g_free(old); old = new;
Most of the multi-letter extensions (such as Svpbmt, Svnapot, Svinval, etc) are only available after Priv spec v1.12 so ISA string generation should check the minimum required priv spec version for all extensions. Fixes: a775398be2e ("target/riscv: Add isa extenstion strings to the device tree") Signed-off-by: Anup Patel <apatel@ventanamicro.com> --- target/riscv/cpu.c | 36 +++++++++++++++++++----------------- 1 file changed, 19 insertions(+), 17 deletions(-)