linux-kernel - Re: [PATCH v2 2/3] soc: qcom: rpmhpd: Remove mx/cx relationship on sc7280

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: <YbEYyGHqaHGeRXsV@builder.lan>
Date:   Wed, 8 Dec 2021 14:42:48 -0600
From:   Bjorn Andersson <bjorn.andersson@...aro.org>
To:     Rajendra Nayak <quic_rjendra@...cinc.com>
Cc:     agross@...nel.org, linux-arm-msm@...r.kernel.org,
        linux-kernel@...r.kernel.org, swboyd@...omium.org, mka@...omium.org
Subject: Re: [PATCH v2 2/3] soc: qcom: rpmhpd: Remove mx/cx relationship on
 sc7280

On Tue 07 Dec 04:08 CST 2021, Rajendra Nayak wrote:

> While the requirement to specify the active + sleep and active-only MX
> power-domains as the parents of the corresponding CX power domains is
> applicable for most SoCs, we have some like the sc7280 where this 
> dependency is not applicable.
> Define new rpmhpd structs for cx and cx_ao without the mx as
> parent and use them for sc7280.
> 
> Signed-off-by: Rajendra Nayak <quic_rjendra@...cinc.com>
> ---
>  drivers/soc/qcom/rpmhpd.c | 18 ++++++++++++++++--
>  1 file changed, 16 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/soc/qcom/rpmhpd.c b/drivers/soc/qcom/rpmhpd.c
> index c71481d..4599efe 100644
> --- a/drivers/soc/qcom/rpmhpd.c
> +++ b/drivers/soc/qcom/rpmhpd.c
> @@ -120,6 +120,20 @@ static struct rpmhpd cx_ao = {
>  	.res_name = "cx.lvl",
>  };
>  
> +static struct rpmhpd cx_ao_no_parent;
> +static struct rpmhpd cx_no_parent = {

There are multiple variations of how each of these can be parented, but
only one way they can be without a parent. So how about we turn this the
other way around?

I.e. let's name this one "cx" and the existing one "cx_w_mx_parent".


This will be particularly useful when you look at mmcx, which on
8150/8180 has mx as parent and on 8450 has cx as parent.


PS. Unfortunately I had merged 8450 since you wrote this series, I tried
to just fix it up as I applied your patch, but noticed 8450_cx and
8450_mmcx and wanted to get your opinion on this first.

Regards,
Bjorn

> +	.pd = { .name = "cx", },
> +	.peer = &cx_ao_no_parent,
> +	.res_name = "cx.lvl",
> +};
> +
> +static struct rpmhpd cx_ao_no_parent = {
> +	.pd = { .name = "cx_ao", },
> +	.active_only = true,
> +	.peer = &cx_no_parent,
> +	.res_name = "cx.lvl",
> +};
> +
>  static struct rpmhpd mmcx_ao;
>  static struct rpmhpd mmcx = {
>  	.pd = { .name = "mmcx", },
> @@ -273,8 +287,8 @@ static const struct rpmhpd_desc sc7180_desc = {
>  
>  /* SC7280 RPMH powerdomains */
>  static struct rpmhpd *sc7280_rpmhpds[] = {
> -	[SC7280_CX] = &cx,
> -	[SC7280_CX_AO] = &cx_ao,
> +	[SC7280_CX] = &cx_no_parent,
> +	[SC7280_CX_AO] = &cx_ao_no_parent,
>  	[SC7280_EBI] = &ebi,
>  	[SC7280_GFX] = &gfx,
>  	[SC7280_MX] = &mx,
> -- 
> 2.7.4
>