[Skiboot] [PATCH 27/61] platforms/astbmc: Add ast2600

[Vasant Hegde]

On 7/29/21 1:54 PM, Joel Stanley wrote:
> On Mon, 19 Jul 2021 at 13:23, Vasant Hegde
> <hegdevasant at linux.vnet.ibm.com> wrote:
>>
>> From: Reza Arbab <arbab at linux.ibm.com>
>>
>> Signed-off-by: Reza Arbab <arbab at linux.ibm.com>
>> Signed-off-by: Vasant Hegde <hegdevasant at linux.vnet.ibm.com>
>> ---
>>   platforms/astbmc/astbmc.h |  2 ++
>>   platforms/astbmc/common.c | 19 +++++++++++++++++--
>>   2 files changed, 19 insertions(+), 2 deletions(-)
>>
>> diff --git a/platforms/astbmc/astbmc.h b/platforms/astbmc/astbmc.h
>> index 86631bc4e..00f221230 100644
>> --- a/platforms/astbmc/astbmc.h
>> +++ b/platforms/astbmc/astbmc.h
>> @@ -87,9 +87,11 @@ static struct slot_table_entry st_name[] = \
>>
>>   extern const struct bmc_hw_config bmc_hw_ast2400;
>>   extern const struct bmc_hw_config bmc_hw_ast2500;
>> +extern const struct bmc_hw_config bmc_hw_ast2600;
>>   extern const struct bmc_platform bmc_plat_ast2400_ami;
>>   extern const struct bmc_platform bmc_plat_ast2500_ami;
>>   extern const struct bmc_platform bmc_plat_ast2500_openbmc;
>> +extern const struct bmc_platform bmc_plat_ast2600_openbmc;
>>
>>   extern void astbmc_early_init(void);
>>   extern int64_t astbmc_ipmi_reboot(void);
>> diff --git a/platforms/astbmc/common.c b/platforms/astbmc/common.c
>> index d96e070e5..83ef70ad3 100644
>> --- a/platforms/astbmc/common.c
>> +++ b/platforms/astbmc/common.c
>> @@ -266,8 +266,9 @@ static void astbmc_fixup_dt_mbox(struct dt_node *lpc)
>>           * can indicate they support mbox using the scratch register, or ipmi
>>           * by configuring the hiomap ipmi command. If neither are configured
>>           * for P8 then skiboot will drive the flash controller directly.
>> +        * XXX P10
>>           */
>> -       if (proc_gen != proc_gen_p9 && !ast_scratch_reg_is_mbox())
>> +       if (proc_gen == proc_gen_p8 && !ast_scratch_reg_is_mbox())
>>                  return;
>>
>>          /* First check if the mbox interface is already there */
>> @@ -478,7 +479,7 @@ void astbmc_early_init(void)
>>                   * never MBOX. Thus only populate the MBOX node on P9 to allow
>>                   * fallback.
>>                   */
>> -               if (proc_gen == proc_gen_p9) {
>> +               if (proc_gen >= proc_gen_p9) {
>>                          astbmc_fixup_dt_mbox(dt_find_primary_lpc());
>>                          ast_setup_sio_mbox(MBOX_IO_BASE, MBOX_LPC_IRQ);
>>                  }
> 
> This part looks okay, the rainier BMC will talk without issue.
> 
>> @@ -530,6 +531,14 @@ const struct bmc_hw_config bmc_hw_ast2500 = {
>>          .mcr_scu_strap = 0x00000000,
>>   };
>>
>> +/* XXX P10: Update with Rainier values */
>> +const struct bmc_hw_config bmc_hw_ast2600 = {
>> +       .scu_revision_id = 0x05000303,
>> +       .mcr_configuration = 0x11200756,
>> +       .mcr_scu_mpll = 0x1008405F,
>> +       .mcr_scu_strap = 0x000030E0,
>> +};
> 
> This one is a bit suspect. It won't cause any issue, only because the
> PCI id has changed, so the quirk code won't call quirk_astbmc_vga to
> populate the device tree.
> 
> Probably best to omit it for now.

Sure. For now I will set this to ast2500?

-Vasant