linux-kernel - Re: [PATCH] clk: Don't write error code into divider register

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]

Date:   Tue, 25 Jul 2017 17:32:08 -0700
From:   Stephen Boyd <sboyd@...eaurora.org>
To:     Peter De Schrijver <pdeschrijver@...dia.com>
Cc:     Michael Turquette <mturquette@...libre.com>,
        linux-clk@...r.kernel.org, linux-kernel@...r.kernel.org,
        Alex Frid <afrid@...dia.com>
Subject: Re: [PATCH] clk: Don't write error code into divider register

On 07/25, Peter De Schrijver wrote:
> From: Alex Frid <afrid@...dia.com>
> 
> Add a check for error returned by divider value calculation to avoid
> writing error code into hw register.
> 
> Signed-off-by: Alex Frid <afrid@...dia.com>
> Reviewed-by: Peter De Schrijver <pdeschrijver@...dia.com>
> Reviewed-by: Jon Mayo <jmayo@...dia.com>
> ---

Fixes: bca9690b9426 ("clk: divider: Make generic for usage elsewhere")

perhaps?

Also, curious how this got triggered? Presumably round_rate would
have failed before a set_rate call with something invalid came
through so that's why nobody has reported anything so far.

>  drivers/clk/clk-divider.c | 6 ++++--
>  1 file changed, 4 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/clk/clk-divider.c b/drivers/clk/clk-divider.c
> index 9bb472c..4ed516c 100644
> --- a/drivers/clk/clk-divider.c
> +++ b/drivers/clk/clk-divider.c
> @@ -385,12 +385,14 @@ static int clk_divider_set_rate(struct clk_hw *hw, unsigned long rate,
>  				unsigned long parent_rate)
>  {
>  	struct clk_divider *divider = to_clk_divider(hw);
> -	unsigned int value;
> +	int value;
>  	unsigned long flags = 0;
>  	u32 val;
>  
>  	value = divider_get_val(rate, parent_rate, divider->table,
>  				divider->width, divider->flags);
> +	if (value < 0)

Perhaps value should be an s32 if we're doing a writel with it
shifted so much. And then, we could declare it as a u32 and test
it with

	if ((s32)value < 0)

to see if there was an error. Sort of annoying that we've limited
the available space of divider_get_val() by combining the value
with the error code number space. We may want to not do that in
case people have huge dividers.

-- 
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum,
a Linux Foundation Collaborative Project