linux-kernel - Re: [PATCH v3 3/7] clk: qcom: Add TCSR clock driver for Glymur

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: <c2f39786-5780-4124-9e41-6971428aa267@oss.qualcomm.com>
Date: Mon, 4 Aug 2025 19:51:31 +0530
From: Taniya Das <taniya.das@....qualcomm.com>
To: Konrad Dybcio <konrad.dybcio@....qualcomm.com>,
        Abel Vesa <abel.vesa@...aro.org>
Cc: kernel@....qualcomm.com, Pankaj Patil <quic_pankpati@...cinc.com>,
        Bjorn Andersson <andersson@...nel.org>,
        Michael Turquette <mturquette@...libre.com>,
        Stephen Boyd
 <sboyd@...nel.org>, Rob Herring <robh@...nel.org>,
        Krzysztof Kozlowski <krzk+dt@...nel.org>,
        Conor Dooley
 <conor+dt@...nel.org>,
        Taniya Das <quic_tdas@...cinc.com>, linux-arm-msm@...r.kernel.org,
        linux-clk@...r.kernel.org, devicetree@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v3 3/7] clk: qcom: Add TCSR clock driver for Glymur



On 8/4/2025 6:40 PM, Konrad Dybcio wrote:
> On 8/4/25 11:00 AM, Taniya Das wrote:
>>
>>
>> On 8/1/2025 5:24 PM, Konrad Dybcio wrote:
>>> On 8/1/25 7:31 AM, Abel Vesa wrote:
>>>> On 25-08-01 10:02:15, Taniya Das wrote:
>>>>>
>>>>>
>>>>> On 7/30/2025 4:55 PM, Abel Vesa wrote:
>>>>>> On 25-07-29 11:12:37, Taniya Das wrote:
>>>>>>> Add a clock driver for the TCSR clock controller found on Glymur, which
>>>>>>> provides refclks for PCIE, USB, and UFS.
>>>>>>>
>>>>>>> Signed-off-by: Taniya Das <taniya.das@....qualcomm.com>
>>>>>>> ---
>>>>>>>  drivers/clk/qcom/Kconfig         |   8 ++
>>>>>>>  drivers/clk/qcom/Makefile        |   1 +
>>>>>>>  drivers/clk/qcom/tcsrcc-glymur.c | 257 +++++++++++++++++++++++++++++++++++++++
>>>>>>>  3 files changed, 266 insertions(+)
>>>>>>>
>>>>>>
>>>>>> [...]
>>>>>>
>>>>>>> +
>>>>>>> +static struct clk_branch tcsr_edp_clkref_en = {
>>>>>>> +	.halt_reg = 0x1c,
>>>>>>> +	.halt_check = BRANCH_HALT_DELAY,
>>>>>>> +	.clkr = {
>>>>>>> +		.enable_reg = 0x1c,
>>>>>>> +		.enable_mask = BIT(0),
>>>>>>> +		.hw.init = &(const struct clk_init_data) {
>>>>>>> +			.name = "tcsr_edp_clkref_en",
>>>>>>> +			.ops = &clk_branch2_ops,
>>>>>>
>>>>>> As discussed off-list, these clocks need to have the bi_tcxo as parent.
>>>>>>
>>>>>> Otherwise, as far as the CCF is concerned these clocks will have rate 0,
>>>>>> which is obviously not the case.
>>>>>>
>>>>>> Bringing this here since there is a disconnect between X Elite and
>>>>>> Glymur w.r.t this now.
>>>>>
>>>>>
>>>>> The ref clocks are not required to be have a parent of bi_tcxo as these
>>>>> ideally can be left enabled(as a subsystem requirement) even if HLOS
>>>>> (APSS) goes to suspend. With the bi_tcxo parent the ARC vote from
>>>>> HLOS/APSS will not allow APSS to collapse.
>>>>
>>>> Is there a scenario where the APSS is collapsed and still the ref clock
>>>> needs to stay enabled ? Sorry, this doesn't make sense to me.
>>>
>>> MDSS is capable of displaying things from a buffer when the CPU is off,
>>> AFAICU
>>>
>>> We can do CXO_AO instead to have it auto-collapse if it's just Linux
>>> requesting it to stay on, I think.
>>>
>>
>> Thanks Konrad for adding the display use case.
>> Abel, we earlier also had some PCIe, USB use cases where we had to leave
>> the ref clocks ON and APSS could collapse.
> 
> XO votes will prevent CX collapse, not APSS collapse. CX also powers
> USB and PCIe so that only makes sense.
> 
> I think it's fair to just stick XO as the parent of every refclock
> today and think about the what-ifs (such as the mdss case I mentioned
> above) later - especially since we have no infra to take full advantage
> of it today (non-APSS RSCs etc.)
> 

When ref clock have been part of GCC, then also they didn't have any xo
as the parent, similar design we kept when it was moved to TCSR as well.

-- 
Thanks,
Taniya Das