Re: [PATCH 4/4] clk: qcom: gcc-glymur: use shared ops for GCC UFS RCGs

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:

Pre-existing issues:
- [High] The patch fixes a UFS GDSC hang for `gcc-glymur.c` by using `clk_rcg2_shared_ops` to park UFS RCGs at XO during init. However, identical UFS RCGs in sibling drivers (`gcc-eliza.c`, `gcc-kaanapali.c`, and others) still use `clk_rcg2_shared_no_init_park_ops` and remain vulnerable to the same GDSC wedge issue.
--

commit ea0c46ac321975e91a07a59d21b631f1f46de735
Author: Jagadeesh Kona <jagadeesh.kona@oss.qualcomm.com>

clk: qcom: gcc-glymur: use shared ops for GCC UFS RCGs

This patch switches GCC UFS RCGs in the gcc-glymur driver to use
clk_rcg2_shared_ops instead of clk_rcg2_shared_no_init_park_ops. This ensures
the clocks are parked at XO during registration, which is required for the UFS
GDSC power-on sequence.

> diff --git a/drivers/clk/qcom/gcc-glymur.c b/drivers/clk/qcom/gcc-glymur.c
> index a47a2c6b961ba..1957c847568af 100644
> --- a/drivers/clk/qcom/gcc-glymur.c
> +++ b/drivers/clk/qcom/gcc-glymur.c
> @@ -2327,7 +2327,7 @@ static struct clk_rcg2 gcc_ufs_phy_axi_clk_src = {
>  		.parent_data = gcc_parent_data_6,
>  		.num_parents = ARRAY_SIZE(gcc_parent_data_6),
>  		.flags = CLK_SET_RATE_PARENT,
> -		.ops = &clk_rcg2_shared_no_init_park_ops,
> +		.ops = &clk_rcg2_shared_ops,

[Severity: High]
This is a pre-existing issue, but this commit appears to only fix the UFS GDSC
hang for gcc-glymur.c, leaving identical UFS RCGs in sibling drivers vulnerable
to the same problem.

For example, in drivers/clk/qcom/gcc-eliza.c:

static struct clk_rcg2 gcc_ufs_phy_axi_clk_src = {
	...
		.ops = &clk_rcg2_shared_no_init_park_ops,
};

And similarly in drivers/clk/qcom/gcc-kaanapali.c:

static struct clk_rcg2 gcc_ufs_phy_axi_clk_src = {
	...
		.ops = &clk_rcg2_shared_no_init_park_ops,
};

Since these drivers share the same hardware IP design and define the exact same
gcc_ufs_phy_gdsc and RCGs, won't they suffer from the identical wedge failure
during UFS controller probe if the bootloader leaves the clocks in an unparked,
off state?

Could this fix be extended to the sibling SoC drivers to prevent UFS
initialization failures across the subsystem?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260905-glymur-ufs-clk-changes-v1-0-cdff2c67a8c2@oss.qualcomm.com?part=4