[Skiboot] [PATCH] hdat: add support for system and type strings

[Oliver O'Halloran]

On Tue, Dec 20, 2016 at 2:05 PM, Stewart Smith
<stewart at linux.vnet.ibm.com> wrote:
> Oliver O'Halloran <oohall at gmail.com> writes:
>
>> In P9 the compatible string is supplied by hostboot through the HDAT.
>> This patch add support for using these strings to set the compatible
>> property in the device tree rather than using the machine ID number
>> scheme traditionally used in the HDAT.
>>
>> Cc: Vasant Hegde <hegdevasant at linux.vnet.ibm.com>
>> Signed-off-by: Oliver O'Halloran <oohall at gmail.com>
>> ---
>>  hdata/spira.c | 59 ++++++++++++++++++++++++++++++++++++-----------------------
>>  hdata/spira.h | 11 +++++++++++
>>  2 files changed, 47 insertions(+), 23 deletions(-)
>>
>> diff --git a/hdata/spira.c b/hdata/spira.c
>> index 592197e7685b..c83146ab0991 100644
>> --- a/hdata/spira.c
>> +++ b/hdata/spira.c
>> @@ -752,8 +752,6 @@ static void add_nx(void)
>>  static void add_iplparams_sys_params(const void *iplp, struct dt_node *node)
>>  {
>>       const struct iplparams_sysparams *p;
>> -     u32 sys_type;
>> -     const char *sys_family;
>>       const struct HDIF_common_hdr *hdif = iplp;
>>       u16 version = be16_to_cpu(hdif->version);
>>
>> @@ -772,29 +770,44 @@ static void add_iplparams_sys_params(const void *iplp, struct dt_node *node)
>>
>>       dt_add_property_nstr(node, "ibm,sys-model", p->sys_model, 4);
>>
>> -     /* Compatible is 2 entries: ibm,powernv and ibm,<platform>
>> +     /*
>> +      * Compatible has up to three entries:
>> +      *      "ibm,powernv", the system family and system type.
>> +      *
>> +      * On P9 and above the family and type strings come from the HDAT
>> +      * directly. On P8 we find it from the system ID numbers.
>>        */
>> -     sys_type = be32_to_cpu(p->system_type);
>> -     switch(sys_type >> 28) {
>> -     case 0:
>> -             sys_family = "ibm,squadrons";
>> -             break;
>> -     case 1:
>> -             sys_family = "ibm,eclipz";
>> -             break;
>> -     case 2:
>> -             sys_family = "ibm,apollo";
>> -             break;
>> -     case 3:
>> -             sys_family = "ibm,firenze";
>> -             break;
>> -     default:
>> -             sys_family = NULL;
>> -             prerror("IPLPARAMS: Unknown system family\n");
>> -             break;
>> +     if (version >= 0x70) {
>> +             dt_add_property_strings(dt_root, "compatible",
>> +                     "ibm,powernv", p->sys_family_str, p->sys_type_str);
>> +     } else {
>> +             u32 sys_type = be32_to_cpu(p->system_type);
>> +             const char *sys_family;
>> +
>> +             switch(sys_type >> 28) {
>> +             case 0:
>> +                     sys_family = "ibm,squadrons";
>> +                     break;
>> +             case 1:
>> +                     sys_family = "ibm,eclipz";
>> +                     break;
>> +             case 2:
>> +                     sys_family = "ibm,apollo";
>> +                     break;
>> +             case 3:
>> +             case 4: /* HACK: this is actually "ibm,p9", but there's no
>> +                      * useful destinction between the two right now */
>> +                     sys_family = "ibm,firenze";
>> +                     break;
>> +             default:
>> +                     sys_family = NULL;
>> +                     prerror("IPLPARAMS: Unknown system family\n");
>> +                     break;
>> +             }
>> +
>> +             dt_add_property_strings(dt_root, "compatible", "ibm,powernv",
>> +                                     sys_family);
>
> I wonder if we should assert that sys_type == the string we expect?
>
> or... could always construct a compatible string and add it as *well* as
> what we have in the hdat string...

The issue is that there's no meaningful system_type numbers for
non-IBM systems. We can only rely on the system_type when running on
an IBM system, but we have no way of knowing when we are since this is
the code that's supposed to figure that out.

>>       }
>> -     dt_add_property_strings(dt_root, "compatible", "ibm,powernv",
>> -                             sys_family);
>>
>>       /* Grab nest frequency when available */
>>       if (version >= 0x005b) {
>> diff --git a/hdata/spira.h b/hdata/spira.h
>> index eabf7f953ab7..b2863342f5a5 100644
>> --- a/hdata/spira.h
>> +++ b/hdata/spira.h
>> @@ -321,6 +321,17 @@ struct iplparams_sysparams {
>>       uint8_t         hw_page_table_size;     /* >= 0x59 */
>>       __be16          hv_disp_wheel;          /* >= 0x58 */
>>       __be32          nest_freq_mhz;          /* >= 0x5b */
>> +     uint8_t         split_core_mode;        /* >= 0x5c */
>> +     uint8_t         reserved1[3];
>> +     uint8_t         vendor_name[64];        /* >= 0x5f */
>> +     __be16          sys_sec_settings;       /* >= 0x60 */
>> +     __be16          tpm_config;
>> +     __be16          tmps_per_drawer;
>> +     __be16          reserved2;
>> +     uint8_t         hw_hash_keys;
>
> This should be uint6_t[64] according to the 10.3g spec?
>
> --
> Stewart Smith
> OPAL Architect, IBM.
>