On Thu, Sep 24, 2026 at 01:51:13PM -0400, Nicolas Dufresne wrote:
> Hi Conor,
> 
> Le jeudi 24 septembre 2026 à 17:55 +0100, Conor Dooley a écrit :
> > On Thu, Sep 24, 2026 at 03:04:52PM +0200, Benjamin Gaignard wrote:
> > > nxp,imx8mq-vpu has been tagged has deprecated since a while.
> > > After driver clean up it is time to remove it.
> > 
> > No users in linux doesn't mean no users, but I am guessing that there's
> > genuinely no users elsewhere and anyone using out of tree devicetrees
> > should have updated due to the warning at runtime?
> 
> I've checked the vendor DT before they moved to the mainline driver, and they
> never used this binding (actually they never made bindings). 
> Here's what the DT was:
> 
> 		vpu: vpu@38300000 {
> 			compatible = "nxp,imx8mq-hantro";
> 			reg = <0x38300000 0x200000>;
> 			reg-names = "regs_hantro";
> 			interrupts = <GIC_SPI 7 IRQ_TYPE_LEVEL_HIGH>, <GIC_SPI 8 IRQ_TYPE_LEVEL_HIGH>;
> 			interrupt-names = "irq_hantro_g1", "irq_hantro_g2";
> 			clocks = <&clk IMX8MQ_CLK_VPU_G1_ROOT>, <&clk IMX8MQ_CLK_VPU_G2_ROOT>, <&clk IMX8MQ_CLK_VPU_DEC_ROOT>;
> 			clock-names = "clk_hantro_g1", "clk_hantro_g2", "clk_hantro_bus";
> 			assigned-clocks = <&clk IMX8MQ_CLK_VPU_G1>, <&clk IMX8MQ_CLK_VPU_G2>, <&clk IMX8MQ_CLK_VPU_BUS>;
> 			assigned-clock-parents = <&clk IMX8MQ_VPU_PLL_OUT>, <&clk IMX8MQ_VPU_PLL_OUT>, <&clk IMX8MQ_SYS1_PLL_800M>;
> 			assigned-clock-rates = <600000000>, <600000000>, <800000000>;
> 			power-domains = <&pgc_vpu>;
> 			status = "disabled";
> 		};
> 
> I've not aware of any other OS or booloader that offers media capabilities using
> Linux mainline DT. The now deprecated nxp,imx8mq-vpu compatible had a short live
> in the Linux tree if my memory is correct. Hope this is enough justification.

Yeah, and this kind of detail (plus the fact that the driver emitted a
warning) are very helpful in the commit message for patches like this.
Hint hint Benjamin!