On Mon, 27 Feb 2017, Joonas Lahtinen <joonas.lahtinen@xxxxxxxxxxxxxxx> wrote:
> On pe, 2017-02-24 at 16:40 +0100, Arkadiusz Hiler wrote:
>> `guc_firmware_path` and `huc_firmware_path` module parameters are added.
>>
>> Using the parameter disabled version checks and loads desired firmware
>> instead of the default one.
>>
>> Cc: Chris Wilson <chris@xxxxxxxxxxxxxxxxxx>
>> Cc: Joonas Lahtinen <joonas.lahtinen@xxxxxxxxxxxxxxx>
>> Cc: Michal Winiarski <michal.winiarski@xxxxxxxxx>
>> Signed-off-by: Arkadiusz Hiler <arkadiusz.hiler@xxxxxxxxx>
>
> <SNIP>
>
>> @@ -230,6 +232,14 @@ module_param_named(guc_log_level, i915.guc_log_level, int, 0400);
>> MODULE_PARM_DESC(guc_log_level,
>> "GuC firmware logging level (-1:disabled (default), 0-3:enabled)");
>>
>> +module_param_named(guc_firmware_path, i915.guc_firmware_path, charp, 0400);
>
> I'm pretty sure this should be _unsafe, because it overrides the
> version checks. Cc'd Jani for this.
Yes, please. I replied the same thing to Chris' patches adding the same
thing.
BR,
Jani.
>
>> @@ -479,7 +479,11 @@ void intel_guc_select_fw(struct intel_guc *guc)
>> guc->fw.load_status = INTEL_UC_FIRMWARE_NONE;
>> guc->fw.fw = INTEL_UC_FW_TYPE_GUC;
>>
>> - if (IS_SKYLAKE(dev_priv)) {
>> + if (i915.guc_firmware_path) {
>> + guc->fw.path = i915.guc_firmware_path;
>> + guc->fw.major_ver_wanted = 0;
>> + guc->fw.minor_ver_wanted = 0;
>
> Or, we could keep the wanted version number, only replace the path, and
> spit out WARN/taint kernel if some other version was detected?
>
> But I guess the main purpose is to override version (not provide
> request_firmware workarounds), so my vote is to make the param _unsafe.
>
> Regards, Joonas
--
Jani Nikula, Intel Open Source Technology Center
_______________________________________________
Intel-gfx mailing list
Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
