Sender: quic_rjendra=quicinc.com@mg.codeaurora.org
DMARC-Filter: OpenDMARC Filter v1.4.1 smtp.codeaurora.org 75C9AC4360D
Message-ID: <7e9f6cbe-17c9-53ff-95a8-2f2bbdeb1d23@quicinc.com>
Date:   Thu, 9 Dec 2021 12:29:12 +0530
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:91.0) Gecko/20100101
 Thunderbird/91.3.2
Subject: Re: [PATCH v2 2/3] soc: qcom: rpmhpd: Remove mx/cx relationship on
 sc7280
Content-Language: en-US
To:     Bjorn Andersson <bjorn.andersson@linaro.org>
Cc:     agross@kernel.org, linux-arm-msm@vger.kernel.org,
        linux-kernel@vger.kernel.org, swboyd@chromium.org, mka@chromium.org
References: <1638871712-18636-1-git-send-email-quic_rjendra@quicinc.com>
 <1638871712-18636-3-git-send-email-quic_rjendra@quicinc.com>
 <YbEYyGHqaHGeRXsV@builder.lan>
From:   Rajendra Nayak <quic_rjendra@quicinc.com>
In-Reply-To: <YbEYyGHqaHGeRXsV@builder.lan>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk


On 12/9/2021 2:12 AM, Bjorn Andersson wrote:
> 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@quicinc.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.

I agree that sounds like a reasonable thing to do, I hadn't looked at 8450
so did not notice it, I will rebase my patches on top and repost.

> 
> 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
>>