[PATCH v2 2/3] drm: Add MIPI support function and macros
Doug Anderson
dianders at chromium.org
Tue Jul 29 21:34:24 UTC 2025
Hi,
On Mon, Jul 28, 2025 at 10:44 PM Brigham Campbell
<me at brighamcampbell.com> wrote:
>
> Create mipi_dsi_dcs_read_multi(), which accepts a mipi_dsi_multi_context
> struct for improved error handling and cleaner panel driver code.
>
> Create mipi_dsi_dcs_write_var_seq_multi() and
> mipi_dsi_generic_write_var_seq_multi() macros which allow MIPI panel
> drivers to write non-static data to display controllers.
>
> Signed-off-by: Brigham Campbell <me at brighamcampbell.com>
> ---
>
> I looked into using __builtin_constant_p() to extend
> mipi_dsi_{generic,dcs}_write_seq_multi() to accept both static and
> non-static sequences of bytes and store them accordingly, it looked
> promising at first, but I found no such solution ultimately. At the very
> least, if we find a solution at some point, my hope is that cocinelle
> could be used to replace each of the new var_seq_multi() usages among
> drivers with an improved seq_multi().
>
>
> drivers/gpu/drm/drm_mipi_dsi.c | 37 ++++++++++++++++++++++++++++++++++
> include/drm/drm_mipi_dsi.h | 35 ++++++++++++++++++++++++++++++++
> 2 files changed, 72 insertions(+)
nit: maybe make the subject a bit more specific, like:
drm: Add MIPI _multi reader func and two new write variants
> @@ -415,6 +417,22 @@ void mipi_dsi_dcs_set_tear_off_multi(struct mipi_dsi_multi_context *ctx);
> mipi_dsi_generic_write_multi(ctx, d, ARRAY_SIZE(d)); \
> } while (0)
>
> +/**
> + * mipi_dsi_generic_write_var_seq_multi - transmit non-static data using a
> + * generic write packet
nit: "non-constant", not "non-static"
>From the caller's point of view the difference is that the data is
compile-time constant in one case and not compile-time constant in the
other case. It happens that means you can _store_ it in a "static
const" in one case and not in the other case, but that doesn't make
the parameters "static".
Other than nits:
Reviewed-by: Douglas Anderson <dianders at chromium.org>
More information about the dri-devel
mailing list