Re: [PATCH v4 09/13] hw/riscv/virt-acpi-build.c: Add MMU node in RHCT

[Sunil V L <sunilvl@ventanamicro.com>]

On Thu, Oct 26, 2023 at 10:31:51AM +0200, Andrew Jones wrote:
> On Thu, Oct 26, 2023 at 01:37:09AM +0530, Sunil V L wrote:
> > MMU type information is available via MMU node in RHCT. Add this node in
> > RHCT.
> > 
> > Signed-off-by: Sunil V L <sunilvl@ventanamicro.com>
> > Reviewed-by: Daniel Henrique Barboza <dbarboza@ventanamicro.com>
> > ---
> >  hw/riscv/virt-acpi-build.c | 37 ++++++++++++++++++++++++++++++++++++-
> >  1 file changed, 36 insertions(+), 1 deletion(-)
> > 
> > diff --git a/hw/riscv/virt-acpi-build.c b/hw/riscv/virt-acpi-build.c
> > index ebe7062b9b..dc7c0213f5 100644
> > --- a/hw/riscv/virt-acpi-build.c
> > +++ b/hw/riscv/virt-acpi-build.c
> > @@ -159,6 +159,8 @@ static void build_rhct(GArray *table_data,
> >      size_t len, aligned_len;
> >      uint32_t isa_offset, num_rhct_nodes, cmo_offset = 0;
> >      RISCVCPU *cpu = &s->soc[0].harts[0];
> > +    uint32_t mmu_offset = 0;
> > +    uint8_t satp_mode_max;
> >      char *isa;
> >  
> >      AcpiTable table = { .sig = "RHCT", .rev = 1, .oem_id = s->oem_id,
> > @@ -178,6 +180,10 @@ static void build_rhct(GArray *table_data,
> >          num_rhct_nodes++;
> >      }
> >  
> > +    if (cpu->cfg.satp_mode.supported != 0) {
> > +        num_rhct_nodes++;
> > +    }
> > +
> >      /* Number of RHCT nodes*/
> >      build_append_int_noprefix(table_data, num_rhct_nodes, 4);
> >  
> > @@ -233,6 +239,26 @@ static void build_rhct(GArray *table_data,
> >          }
> >      }
> >  
> > +    /* MMU node structure */
> > +    if (cpu->cfg.satp_mode.supported != 0) {
> > +        satp_mode_max = satp_mode_max_from_map(cpu->cfg.satp_mode.map);
> > +        mmu_offset = table_data->len - table.table_offset;
> > +        build_append_int_noprefix(table_data, 2, 2);    /* Type */
> > +        build_append_int_noprefix(table_data, 8, 2);    /* Length */
> > +        build_append_int_noprefix(table_data, 0x1, 2);  /* Revision */
> > +        build_append_int_noprefix(table_data, 0, 1);    /* Reserved */
> > +        /* MMU Type */
> > +        if (satp_mode_max == VM_1_10_SV57) {
> > +            build_append_int_noprefix(table_data, 2, 1);    /* Sv57 */
> > +        } else if (satp_mode_max == VM_1_10_SV48) {
> > +            build_append_int_noprefix(table_data, 1, 1);    /* Sv48 */
> > +        } else if (satp_mode_max == VM_1_10_SV39) {
> > +            build_append_int_noprefix(table_data, 0, 1);    /* Sv39 */
> > +        } else {
> > +            assert(1);
> > +        }
> > +    }
> > +
> >      /* Hart Info Node */
> >      for (int i = 0; i < arch_ids->len; i++) {
> >          len = 16;
> > @@ -245,17 +271,26 @@ static void build_rhct(GArray *table_data,
> >              num_offsets++;
> >          }
> >  
> > +        if (mmu_offset) {
> > +            len += 4;
> > +            num_offsets++;
> > +        }
> > +
> >          build_append_int_noprefix(table_data, len, 2);
> >          build_append_int_noprefix(table_data, 0x1, 2); /* Revision */
> >          /* Number of offsets */
> >          build_append_int_noprefix(table_data, num_offsets, 2);
> >          build_append_int_noprefix(table_data, i, 4);   /* ACPI Processor UID */
> > -
> >          /* Offsets */
> >          build_append_int_noprefix(table_data, isa_offset, 4);
> > +        if (mmu_offset) {
> > +            build_append_int_noprefix(table_data, mmu_offset, 4);
> > +        }
> > +
> 
> In the previous version of this patch the MMU node was getting generated
> above the CMO node, so its offset was less than those of the CMO node,
> and why I recommended moving it up here. But, in this version, the MMU
> node is now getting generated after the CMO node, so moving this up
> means the offsets are still not in ascending order.
> 
Yeah, after changing here I realized MMU node type is logically better
to be created after cmo. So, I changed the creation order but forgot
reorder here. Will update in the next revision.

Thanks,
Sunil