lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <87af7b7e-9c2f-41e1-af97-01d3f29f5970@linaro.org>
Date: Tue, 26 Mar 2024 21:51:24 +0100
From: Konrad Dybcio <konrad.dybcio@...aro.org>
To: Gabor Juhos <j4g8y7@...il.com>, Bjorn Andersson <andersson@...nel.org>,
 Michael Turquette <mturquette@...libre.com>, Stephen Boyd
 <sboyd@...nel.org>, Varadarajan Narayanan <quic_varada@...cinc.com>,
 Sricharan R <quic_srichara@...cinc.com>,
 Kathiravan T <quic_kathirav@...cinc.com>
Cc: linux-arm-msm@...r.kernel.org, linux-clk@...r.kernel.org,
 linux-kernel@...r.kernel.org, stable@...r.kernel.org
Subject: Re: [PATCH v2] clk: qcom: clk-alpha-pll: fix rate setting for Stromer
 PLLs

On 26.03.2024 1:15 PM, Gabor Juhos wrote:
> The clk_alpha_pll_stromer_set_rate() function writes inproper
> values into the ALPHA_VAL{,_U} registers which results in wrong
> clock rates when the alpha value is used.
> 
> The broken behaviour can be seen on IPQ5018 for example, when
> dynamic scaling sets the CPU frequency to 800000 KHz. In this
> case the CPU cores are running only at 792031 KHz:
> 
>   # cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_cur_freq
>   800000
>   # cat /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_cur_freq
>   792031
> 
> This happens because the function ignores the fact that the alpha
> value calculated by the alpha_pll_round_rate() function is only
> 32 bits wide which must be extended to 40 bits if it is used on
> a hardware which supports 40 bits wide values.
> 
> Extend the clk_alpha_pll_stromer_set_rate() function to convert
> the alpha value to 40 bits before wrinting that into the registers
> in order to ensure that the hardware really uses the requested rate.
> 
> After the change the CPU frequency is correct:
> 
>   # cat /sys/devices/system/cpu/cpu0/cpufreq/scaling_cur_freq
>   800000
>   # cat /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_cur_freq
>   800000
> 
> Cc: stable@...r.kernel.org
> Fixes: e47a4f55f240 ("clk: qcom: clk-alpha-pll: Add support for Stromer PLLs")
> Signed-off-by: Gabor Juhos <j4g8y7@...il.com>
> ---
> Changes in v2:
>   - fix subject prefix
>   - rebase on v6.9-rc1
>   - Link to v1: https://lore.kernel.org/r/20240324-alpha-pll-fix-stromer-set-rate-v1-1-335b0b157219@gmail.com
> 
> Depends on the following patch:
>   https://lore.kernel.org/r/20240315-apss-ipq-pll-ipq5018-hang-v2-1-6fe30ada2009@gmail.com
> ---
>  drivers/clk/qcom/clk-alpha-pll.c | 4 ++++
>  1 file changed, 4 insertions(+)
> 
> diff --git a/drivers/clk/qcom/clk-alpha-pll.c b/drivers/clk/qcom/clk-alpha-pll.c
> index 8a412ef47e163..8e98198d4b4b6 100644
> --- a/drivers/clk/qcom/clk-alpha-pll.c
> +++ b/drivers/clk/qcom/clk-alpha-pll.c
> @@ -2490,6 +2490,10 @@ static int clk_alpha_pll_stromer_set_rate(struct clk_hw *hw, unsigned long rate,
>  	rate = alpha_pll_round_rate(rate, prate, &l, &a, ALPHA_REG_BITWIDTH);
>  
>  	regmap_write(pll->clkr.regmap, PLL_L_VAL(pll), l);
> +
> +	if (ALPHA_REG_BITWIDTH > ALPHA_BITWIDTH)
> +		a <<= ALPHA_REG_BITWIDTH - ALPHA_BITWIDTH;

Uh.. that's not right, this is comparing two constants

Did you mean to use pll_alpha_width()?

Konrad

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ