On Sun, Nov 02, 2025 at 05:02:37PM +0100, Marek Vasut wrote:
>On 10/28/25 10:08 AM, Shawn Guo wrote:
>> On Tue, Oct 28, 2025 at 04:21:05PM +0800, Rain Yang wrote:
>> > On Mon, Oct 27, 2025 at 08:57:08AM +0800, Shawn Guo wrote:
>> > > On Sat, Oct 11, 2025 at 12:53:29PM +0200, Marek Vasut wrote:
>> > > > On 9/25/25 10:38 PM, Marek Vasut wrote:
>> > > > > The instance of the GPU populated in i.MX95 is the G310, describe
>> > > > > this
>> > > > > GPU in the DT. Include dummy GPU voltage regulator and OPP tables.
>> > > > Is there still anything that should be changed with this patchset, or
>> > > > can it
>> > > > be applied ?
>> > >
>> > > I'm waiting binding change to be applied first. Or is it already done?
>> > >
>> > > Shawn
>> > >
>> > Hi Shawn,
>> > It might be better for Marek to remove the always-on GPUAPB clock from the
>> > DTS, given that
>> > it's an SCMI protocol clock and the kernel expects it to be available
>> > during operations.
>> >
>> > Also, the gpu_fix_reg appears unnecessary for the driver and could be
>> > removed[1].
>> >
>> > [1]
>> > https://cgit.freedesktop.org/drm/drm-misc/commit/?id=a8cb5ca53690aa809f4f65e14192753073e61a71,
>> > https://cgit.freedesktop.org/drm/drm-misc/commit/?id=02df3543f3e0ea572e2c739605ebd6c20e1149c4
>>
>> Ah, sorry! I did not know we haven't reached agreement on the change.
>The feedback above is new to me, so I am somewhat surprised too.
>
>I'll just send a rebased V4 now.
Hi Marek,
thanks for your patch to make i.MX95 GPU upstreamed.
Maybe I didn't express it clearly in the previous e-mails.
the always-on GPUAPB clock must be removed, otherwise, it will result in
redundant and unnecessary
function call to M33.
