linux-kernel - Re: [PATCH 1/3] drivers: remoteproc: Make mdt

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:   Wed, 26 Jul 2017 11:11:08 -0700
From:   Bjorn Andersson <bjorn.andersson@...aro.org>
To:     Sricharan R <sricharan@...eaurora.org>
Cc:     ohad@...ery.com, robh+dt@...nel.org, mark.rutland@....com,
        andy.gross@...aro.org, david.brown@...aro.org,
        linux-remoteproc@...r.kernel.org, devicetree@...r.kernel.org,
        linux-kernel@...r.kernel.org, linux-arm-msm@...r.kernel.org,
        linux-soc@...r.kernel.org
Subject: Re: [PATCH 1/3] drivers: remoteproc: Make mdt_loader firmware
 authentication optional

On Thu 29 Jun 07:17 PDT 2017, Sricharan R wrote:

> qcom_mdt_load function loads the mdt type firmware and
> authenticates it as well. Make the authentication only
> when requested by the caller, so that the function can be used
> by self-authenticating remoteproc as well.
> 

This is good, we should be able to save some duplication the current
MSA PIL as well by this.

> Signed-off-by: Sricharan R <sricharan@...eaurora.org>
> ---
>  drivers/remoteproc/qcom_adsp_pil.c  |  3 ++-
>  drivers/remoteproc/qcom_wcnss.c     |  3 ++-
>  drivers/soc/qcom/mdt_loader.c       | 24 ++++++++++++++----------
>  include/linux/soc/qcom/mdt_loader.h |  2 +-

We have two additional callers being merged, so changing the prototype
of qcom_mdt_load() will cause issues.

I think the best approach is to leave all callers untouched, make the
current implementation of qcom_mdt_load() internal and provide two
exported wrapper functions: qcom_mdt_load() and qcom_mdt_load_no_init().

These wrappers would call the internal function with the appropriate
value of the boolean.

[..]
> diff --git a/drivers/soc/qcom/mdt_loader.c b/drivers/soc/qcom/mdt_loader.c
[..]
>  int qcom_mdt_load(struct device *dev, const struct firmware *fw,
>  		  const char *firmware, int pas_id, void *mem_region,
> -		  phys_addr_t mem_phys, size_t mem_size)
> +		  phys_addr_t mem_phys, size_t mem_size, bool auth)

We're not authenticating even with @auth=true, so please name this
"pas_init".

>  {
>  	const struct elf32_phdr *phdrs;
>  	const struct elf32_phdr *phdr;
[..]
> @@ -142,12 +144,14 @@ int qcom_mdt_load(struct device *dev, const struct firmware *fw,
>  	}
>  
>  	if (relocate) {
> -		ret = qcom_scm_pas_mem_setup(pas_id, mem_phys, max_addr - min_addr);
> -		if (ret) {
> -			dev_err(dev, "unable to setup relocation\n");
> -			goto out;
> +		if (auth) {
> +			ret = qcom_scm_pas_mem_setup(pas_id, mem_phys,
> +						     max_addr - min_addr);
> +			if (ret) {
> +				dev_err(dev, "unable to setup relocation\n");
> +				goto out;
> +			}
>  		}
> -

I like this empty line, please let me have it.

>  		/*
>  		 * The image is relocatable, so offset each segment based on
>  		 * the lowest segment address.

Regards,
Bjorn