On Tue, Nov 12, 2019 at 08:59:56PM +0800, Icenowy Zheng wrote:
>
>
> 于 2019年11月11日 GMT+08:00 下午8:39:36, Maxime Ripard <mripard@kernel.org> 写到:
> >Hi,
> >
> >Thanks for your patch
> >
> >On Sat, Nov 09, 2019 at 03:19:09PM +0000, Tian Yunhao wrote:
> >> The hws field of sun8i_v3s_hw_clks has only 74
> >> members. However, the number specified by CLK_NUMBER
> >> is 77 (= CLK_I2S0 + 1). This leads to runtime segmentation
> >> fault that is not always reproducible.
> >>
> >> This patch adds a protective field [CLK_NUMBER] which ensures
> >> ARRAY_SIZE(.hws) is always greater than .num, thus eliminates
> >> this error.
> >>
> >> Signed-off-by: Yunhao Tian <t123yh@outlook.com>
> >> ---
> >>  drivers/clk/sunxi-ng/ccu-sun8i-v3s.c | 2 ++
> >>  1 file changed, 2 insertions(+)
> >>
> >> diff --git a/drivers/clk/sunxi-ng/ccu-sun8i-v3s.c
> >b/drivers/clk/sunxi-ng/ccu-sun8i-v3s.c
> >> index 5c779eec454b..de7fce7f32e6 100644
> >> --- a/drivers/clk/sunxi-ng/ccu-sun8i-v3s.c
> >> +++ b/drivers/clk/sunxi-ng/ccu-sun8i-v3s.c
> >> @@ -617,6 +617,7 @@ static struct clk_hw_onecell_data
> >sun8i_v3s_hw_clks = {
> >>  		[CLK_AVS]		= &avs_clk.common.hw,
> >>  		[CLK_MBUS]		= &mbus_clk.common.hw,
> >>  		[CLK_MIPI_CSI]		= &mipi_csi_clk.common.hw,
> >> +		[CLK_NUMBER]    = NULL,
> >>  	},
> >>  	.num	= CLK_NUMBER,
> >>  };
> >> @@ -699,6 +700,7 @@ static struct clk_hw_onecell_data
> >sun8i_v3_hw_clks = {
> >>  		[CLK_AVS]		= &avs_clk.common.hw,
> >>  		[CLK_MBUS]		= &mbus_clk.common.hw,
> >>  		[CLK_MIPI_CSI]		= &mipi_csi_clk.common.hw,
> >> +		[CLK_NUMBER]    = NULL,
> >>  	},
> >>  	.num	= CLK_NUMBER,
> >
> >I'd rather have the number of clocks (.num) being properly set.
>
> However the maximum clock indices number is different on V3s and V3, because
> on V3s the last clock is missing.
>
> Should we define CLK_NUMBER_V3S here?

That, or we can just reference the last clock, we're not using
CLK_NUMBER anywhere else.

Maxime