Re: [PATCH] gpu: host1x: mipi: Disable the clock tegra114_mipi_power_down() enables
From: Svyatoslav Ryhel
Date: Sun Nov 22 2026 - 02:15:54 EST
нд, 22 лист. 2026 р. о 07:12 EMCP333 via B4 Relay
<devnull+emcp.enser.io@xxxxxxxxxx> пише:
>
> From: EMCP333 <emcp@xxxxxxxx>
>
> tegra114_mipi_power_down() enables the MIPI calibration clock to write
> the bias pad registers and never disables it. tegra114_mipi_power_up()
> pairs the same clk_enable() with a clk_disable(). Every DSI power-down
> (each DPMS off, each suspend) therefore leaves one more enable behind,
> and the clock never gates again.
>
> On a Tegra Note 7 (Tegra114) the mipi-cal enable count is 0 at boot and
> 4 after four DPMS-offs, with the clock still running while the display
> is off.
>
> No register access follows the new clk_disable(), and every other access
> in this driver takes the clock itself, so nothing depends on the leaked
> enable.
>
> Tested on that board: the mipi-cal enable count stays at 0 across three
> DSI power-down/power-up cycles.
>
> The code and this changelog were prepared with the help of an AI coding
> assistant. The submitter has reviewed both and takes responsibility for
> them.
>
> Fixes: 15372d4be7f0 ("gpu: host1x: mipi: Power down regulators when unused")
> Signed-off-by: EMCP333 <emcp@xxxxxxxx>
> ---
> drivers/gpu/host1x/tegra114-mipi.c | 2 ++
> 1 file changed, 2 insertions(+)
>
> diff --git a/drivers/gpu/host1x/tegra114-mipi.c b/drivers/gpu/host1x/tegra114-mipi.c
> index 49ed05d1a6..06130d453b 100644
> --- a/drivers/gpu/host1x/tegra114-mipi.c
> +++ b/drivers/gpu/host1x/tegra114-mipi.c
> @@ -197,6 +197,8 @@ static int tegra114_mipi_power_down(struct tegra_mipi *mipi)
> value |= MIPI_CAL_BIAS_PAD_PDVCLAMP;
> tegra_mipi_writel(mipi, value, MIPI_CAL_BIAS_PAD_CFG0);
>
> + clk_disable(mipi->clk);
> +
> return 0;
> }
>
>
> ---
> base-commit: 7475ea8b58616b33c04afb976cf0fbaad9dd63fb
> change-id: 20261122-b4-host1x-mipi-clk-leak-ebf2c34b1640
>
> Best regards,
> --
> EMCP333 <emcp@xxxxxxxx>
>
>
NAK
I wonder if this is AI slopped, or straight up stolen? Or maybe both?
Existing patch sent 25th of Oct 2026:
https://lore.kernel.org/lkml/20261025-upstream-host1x-mipi-clk-v1-1-6bc6e755b4cb@xxxxxxxxxxxxxxxxxx/
Reviewed by me and Mikko.