Re: [RFC PATCH 5/7] riscv: Use pgtable_l4_enabled to output mmu type in cpuinfo

[Alex Ghiti <alex@ghiti.fr>]

On 4/3/20 11:53 AM, Palmer Dabbelt wrote:
> On Sun, 22 Mar 2020 04:00:26 PDT (-0700), alex@ghiti.fr wrote:
>> Now that the mmu type is determined at runtime using SATP
>> characteristic, use the global variable pgtable_l4_enabled to output
>> mmu type of the processor through /proc/cpuinfo instead of relying on
>> device tree infos.
>>
>> Signed-off-by: Alexandre Ghiti <alex@ghiti.fr>
>> ---
>>  arch/riscv/boot/dts/sifive/fu540-c000.dtsi |  4 ----
>>  arch/riscv/kernel/cpu.c                    | 24 ++++++++++++----------
>>  2 files changed, 13 insertions(+), 15 deletions(-)
>>
>> diff --git a/arch/riscv/boot/dts/sifive/fu540-c000.dtsi 
>> b/arch/riscv/boot/dts/sifive/fu540-c000.dtsi
>> index 7db861053483..6138590a2229 100644
>> --- a/arch/riscv/boot/dts/sifive/fu540-c000.dtsi
>> +++ b/arch/riscv/boot/dts/sifive/fu540-c000.dtsi
>> @@ -50,7 +50,6 @@
>>              i-cache-size = <32768>;
>>              i-tlb-sets = <1>;
>>              i-tlb-size = <32>;
>> -            mmu-type = "riscv,sv39";
>>              reg = <1>;
>>              riscv,isa = "rv64imafdc";
>>              tlb-split;
>> @@ -74,7 +73,6 @@
>>              i-cache-size = <32768>;
>>              i-tlb-sets = <1>;
>>              i-tlb-size = <32>;
>> -            mmu-type = "riscv,sv39";
>>              reg = <2>;
>>              riscv,isa = "rv64imafdc";
>>              tlb-split;
>> @@ -98,7 +96,6 @@
>>              i-cache-size = <32768>;
>>              i-tlb-sets = <1>;
>>              i-tlb-size = <32>;
>> -            mmu-type = "riscv,sv39";
>>              reg = <3>;
>>              riscv,isa = "rv64imafdc";
>>              tlb-split;
>> @@ -122,7 +119,6 @@
>>              i-cache-size = <32768>;
>>              i-tlb-sets = <1>;
>>              i-tlb-size = <32>;
>> -            mmu-type = "riscv,sv39";
>>              reg = <4>;
>>              riscv,isa = "rv64imafdc";
>>              tlb-split;
>> diff --git a/arch/riscv/kernel/cpu.c b/arch/riscv/kernel/cpu.c
>> index 40a3c442ac5f..38a699b997a8 100644
>> --- a/arch/riscv/kernel/cpu.c
>> +++ b/arch/riscv/kernel/cpu.c
>> @@ -8,6 +8,8 @@
>>  #include <linux/of.h>
>>  #include <asm/smp.h>
>>
>> +extern bool pgtable_l4_enabled;
>> +
>>  /*
>>   * Returns the hart ID of the given device tree node, or -ENODEV if 
>> the node
>>   * isn't an enabled and valid RISC-V hart node.
>> @@ -54,18 +56,19 @@ static void print_isa(struct seq_file *f, const 
>> char *isa)
>>      seq_puts(f, "\n");
>>  }
>>
>> -static void print_mmu(struct seq_file *f, const char *mmu_type)
>> +static void print_mmu(struct seq_file *f)
>>  {
>> +    char sv_type[16];
>> +
>>  #if defined(CONFIG_32BIT)
>> -    if (strcmp(mmu_type, "riscv,sv32") != 0)
>> -        return;
>> +    strncpy(sv_type, "sv32", 5);
>>  #elif defined(CONFIG_64BIT)
>> -    if (strcmp(mmu_type, "riscv,sv39") != 0 &&
>> -        strcmp(mmu_type, "riscv,sv48") != 0)
>> -        return;
>> +    if (pgtable_l4_enabled)
>> +        strncpy(sv_type, "sv48", 5);
>> +    else
>> +        strncpy(sv_type, "sv39", 5);
>>  #endif
>> -
>> -    seq_printf(f, "mmu\t\t: %s\n", mmu_type+6);
>> +    seq_printf(f, "mmu\t\t: %s\n", sv_type);
>>  }
>>
>>  static void *c_start(struct seq_file *m, loff_t *pos)
>> @@ -90,14 +93,13 @@ static int c_show(struct seq_file *m, void *v)
>>  {
>>      unsigned long cpu_id = (unsigned long)v - 1;
>>      struct device_node *node = of_get_cpu_node(cpu_id, NULL);
>> -    const char *compat, *isa, *mmu;
>> +    const char *compat, *isa;
>>
>>      seq_printf(m, "processor\t: %lu\n", cpu_id);
>>      seq_printf(m, "hart\t\t: %lu\n", cpuid_to_hartid_map(cpu_id));
>>      if (!of_property_read_string(node, "riscv,isa", &isa))
>>          print_isa(m, isa);
>> -    if (!of_property_read_string(node, "mmu-type", &mmu))
>> -        print_mmu(m, mmu);
>> +    print_mmu(m);
>>      if (!of_property_read_string(node, "compatible", &compat)
>>          && strcmp(compat, "riscv"))
>>          seq_printf(m, "uarch\t\t: %s\n", compat);
> 
> Reviewed-by: Palmer Dabbelt <palmerdabbelt@google.com>

Thanks,

Alex