Re: [PATCH v3 6/8] clk: sunxi-ng: div: add read-only operation support

[sashiko-bot@kernel.org]

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

Pre-existing issues:
- [High] The existing ccu_div_determine_rate_helper incorrectly asks the hardware parent for a non-predivided rate when CLK_SET_RATE_PARENT is set, resulting in unvalidated parent rate requests.
--

commit 20e3c89e1fcae37e3eaf3a9be6d0c4e28094d07d
Author: Jerome Brunet <jbrunet@baylibre.com>

clk: sunxi-ng: div: add read-only operation support

This commit adds support for sunxi-ng read-only dividers to be used
by the a733 oscillator detection logic.

> diff --git a/drivers/clk/sunxi-ng/ccu_div.c b/drivers/clk/sunxi-ng/ccu_div.c
> index 62d680ccb524b..d1c8c7baa12dc 100644
> --- a/drivers/clk/sunxi-ng/ccu_div.c
> +++ b/drivers/clk/sunxi-ng/ccu_div.c
> @@ -84,6 +84,36 @@ static int ccu_div_determine_rate(struct clk_hw *hw,
>  					     req, ccu_div_determine_rate_helper, cd);
>  }

[Severity: High]
This is a pre-existing issue, but does ccu_div_determine_rate_helper
incorrectly ask the hardware parent for a non-predivided rate when
CLK_SET_RATE_PARENT is set?

When ccu_div_determine_rate is called on a clock with a pre-divider, it
delegates to ccu_mux_helper_determine_rate, which calls
ccu_div_determine_rate_helper. That helper then directly invokes
divider_determine_rate() which queries req->best_parent_hw for
req->rate * div.

Could this cause the clock core to query the upstream parent for a target
rate without accounting for the pre-divider? If the core subsequently sets
the parent rate to an unvalidated higher rate, it seems the parent PLL may
fail or clamp because it was asked for an achievable rate that it cannot
actually sustain when multiplied by the pre-divider.

[ ... ]

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260702-a733-rtc-v3-0-eb2580374de6@baylibre.com?part=6