linux-kernel - Re: [PATCH 3/3] cpufreq: qcom: fix writes in read-only memory region

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:   Mon, 10 Oct 2022 11:30:34 +0530
From:   Viresh Kumar <viresh.kumar@...aro.org>
To:     Fabien Parent <fabien.parent@...aro.org>
Cc:     ilia.lin@...nel.org, agross@...nel.org, bjorn.andersson@...aro.org,
        rafael@...nel.org, linux-pm@...r.kernel.org,
        linux-arm-msm@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 3/3] cpufreq: qcom: fix writes in read-only memory region

On 01-10-22, 19:10, Fabien Parent wrote:
> This commit fixes a kernel oops because of a write in some read-only memory:
> 
> 	[    9.068287] Unable to handle kernel write to read-only memory at virtual address ffff800009240ad8
> 	..snip..
> 	[    9.138790] Internal error: Oops: 9600004f [#1] PREEMPT SMP
> 	..snip..
> 	[    9.269161] Call trace:
> 	[    9.276271]  __memcpy+0x5c/0x230
> 	[    9.278531]  snprintf+0x58/0x80
> 	[    9.282002]  qcom_cpufreq_msm8939_name_version+0xb4/0x190
> 	[    9.284869]  qcom_cpufreq_probe+0xc8/0x39c
> 	..snip..
> 
> The following line defines a pointer that point to a char buffer stored
> in read-only memory:
> 
> 	char *pvs_name = PVS_NAME;
> 
> This pointer is meant to hold a template "speedXX-pvsXX-vXX" where the
> XX values get overridden by the qcom_cpufreq_krait_name_version function. Since
> the template is actually stored in read-only memory, when the function
> executes the following call we get an oops:
> 
> 	snprintf(*pvs_name, sizeof(PVS_NAME), "speed%d-pvs%d-v%d",
> 		 speed, pvs, pvs_ver);
> 
> To fix this issue, we instead store the template name onto the stack by
> using the following syntax:
> 
> 	char pvs_name_buffer[] = PVS_NAME;
> 
> Because the `pvs_name` needs to be able to be assigned to NULL, the
> template buffer is stored in the pvs_name_template and not under the
> pvs_name variable.
> 
> Signed-off-by: Fabien Parent <fabien.parent@...aro.org>

No Fixes or Cc:Stable for this ? This looks like a bug fix and this
should be sent before patch 2/3 so it can be back ported properly.

> ---
>  drivers/cpufreq/qcom-cpufreq-nvmem.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/cpufreq/qcom-cpufreq-nvmem.c b/drivers/cpufreq/qcom-cpufreq-nvmem.c
> index 64ce077a4848..3e097262f612 100644
> --- a/drivers/cpufreq/qcom-cpufreq-nvmem.c
> +++ b/drivers/cpufreq/qcom-cpufreq-nvmem.c
> @@ -269,7 +269,8 @@ static int qcom_cpufreq_probe(struct platform_device *pdev)
>  	struct nvmem_cell *speedbin_nvmem;
>  	struct device_node *np;
>  	struct device *cpu_dev;
> -	char *pvs_name = PVS_NAME_TEMPLATE;
> +	char pvs_name_buffer[] = PVS_NAME_TEMPLATE;
> +	char *pvs_name = pvs_name_buffer;
>  	unsigned cpu;
>  	const struct of_device_id *match;
>  	int ret;
> -- 
> 2.37.2

-- 
viresh