[PATCH v6 08/11] drm/xe/configfs: Keep default device config settings together

[Lucas De Marchi]

On Tue, Aug 05, 2025 at 05:09:52PM +0200, Michal Wajdeczko wrote:
>
>
>On 8/5/2025 3:28 PM, Lucas De Marchi wrote:
>> On Thu, Jul 31, 2025 at 09:33:36PM +0200, Michal Wajdeczko wrote:
>>> For easier maintenance add a placeholder where we can keep all
>>> default device configuration settings in one place.
>>>
>>> Signed-off-by: Michal Wajdeczko <michal.wajdeczko at intel.com>
>>> Cc: Lucas De Marchi <lucas.demarchi at intel.com>
>>> Reviewed-by: John Harrison <John.C.Harrison at Intel.com>
>>> ---
>>> drivers/gpu/drm/xe/xe_configfs.c | 17 +++++++++++++----
>>> 1 file changed, 13 insertions(+), 4 deletions(-)
>>>
>>> diff --git a/drivers/gpu/drm/xe/xe_configfs.c b/drivers/gpu/drm/xe/xe_configfs.c
>>> index 7ad9fc65e21d..150e7f2becc8 100644
>>> --- a/drivers/gpu/drm/xe/xe_configfs.c
>>> +++ b/drivers/gpu/drm/xe/xe_configfs.c
>>> @@ -97,6 +97,16 @@ struct xe_config_group_device {
>>>     struct mutex lock;
>>> };
>>>
>>> +static struct xe_config_device device_defaults = {
>>
>> any reason why this would not be const?
>>
>
>while today we are extending our configfs support only
>with completely new attributes (like survivability mode
>or allowed engines) I was assuming that shortly we will
>also move to configfs some of our existing modparams,
>which are device specific, rather than module specific,
>like:
> - GuC/HuC firmwares
> - max VFs
> - ...
>
>so to keep the idea of 'device_defaults' as the only
>place where default values are stored, I assumed, maybe
>wrong, that we will need this to be non-const (either as
>new location of some modparams or cached values of the
>modparam attributes)

I don't think we will have non-const defaults. If it needs to be
different per platform, then we'd rather copy the "module defaults"
(since it's a static) to the device struct and work from there.

Anyway, as I said in my other reply, I added the const and pushed. If we
need in future we can remove it.

thanks
Lucas De Marchi

>
>in [1] I was trying to make first step with max_vfs,
>but patch was dropped for a while until this series will
>be completed and we will have some better understanding
>how to proceed
>
>Michal
>
>[1] https://patchwork.freedesktop.org/patch/664176/?series=151660&rev=1
>