linux-kernel - Re: [PATCH v6 5/6] clk: qcom: lpassaudiocc-sc7280: Merge lpasscc into lpass

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:   Tue, 31 Jan 2023 12:26:26 -0800
From:   Stephen Boyd <swboyd@...omium.org>
To:     Srinivasa Rao Mandadapu <quic_srivasam@...cinc.com>,
        agross@...nel.org, andersson@...nel.org, broonie@...nel.org,
        konrad.dybcio@...ainline.org, krzysztof.kozlowski+dt@...aro.org,
        linux-arm-msm@...r.kernel.org, linux-clk@...r.kernel.org,
        linux-kernel@...r.kernel.org, mturquette@...libre.com,
        quic_plai@...cinc.com, quic_rohkumar@...cinc.com,
        robh+dt@...nel.org
Subject: Re: [PATCH v6 5/6] clk: qcom: lpassaudiocc-sc7280: Merge lpasscc into lpass_aon

Quoting Srinivasa Rao Mandadapu (2023-01-31 01:29:16)
>
> On 1/31/2023 6:34 AM, Stephen Boyd wrote:
> Thanks for your Time Stephen!!!
> > Quoting Srinivasa Rao Mandadapu (2023-01-26 02:14:24)
> >> Merge lpasscc clocks into lpass_aon clk_regmap structure as they
> >> are using same register space.
> >> Add conditional check for doing lpasscc clock registration only
> >> if regname specified in device tree node.
> >> In existing implementation, lpasscc clocks and lpass_aon clocks are
> >> being registered exclusively and overlapping if both of them are
> >> to be used.
> >> This is required to avoid such overlapping and to register
> >> lpasscc clocks and lpass_aon clocks simultaneously.
> > Can you describe the register ranges that are overlapping?
> Okay. Will add register ranges in description.

Thanks!

> >
> > Here's what I see in DT right now:
> >
> >                  lpasscc: lpasscc@...0000 {
> >                          compatible = "qcom,sc7280-lpasscc";
> >                          reg = <0 0x03000000 0 0x40>,
> >                                <0 0x03c04000 0 0x4>;
> >                          ...
> >                  };
> >
> >                  lpass_audiocc: clock-controller@...0000 {
> >                          compatible = "qcom,sc7280-lpassaudiocc";
> >                          reg = <0 0x03300000 0 0x30000>,
> >                                <0 0x032a9000 0 0x1000>;
> >                          ...
> >                  };
> >
> >                  lpass_aon: clock-controller@...0000 {
> >                          compatible = "qcom,sc7280-lpassaoncc";
> >                          reg = <0 0x03380000 0 0x30000>;
> >                          ...
> >                  };
> >
> >                  lpass_core: clock-controller@...0000 {
> >                          compatible = "qcom,sc7280-lpasscorecc";
> >                          reg = <0 0x03900000 0 0x50000>;
> >                          ...
> >                  };
> >
> > Presumably lpascc is really supposed to be a node named
> > 'clock-controller' and is the node that is overlapping with lpass_aon?
>
> Okay. As it's been coming previous patches, didn't change the name.
>
> May be we need to do it as separate patch.

Sure, another patch to rename lpasscc to clock-controller would be
appreciated.

>
> Yes. It's overlapping with lpass_aon ( <0 0x03380000 0 0x30000>).
>
> CC clocks range is <0 0x03389000 0 0x24>;

Is that a new register range for lpasscc? Why do we have that node at
all? Can we add different properties to the existing lpass_audiocc,
lpass_aon, or lpass_core nodes to indicate what clks should or shouldn't
be registered or provided to the kernel?

>
> >
> >> Fixes: 4ab43d171181 ("clk: qcom: Add lpass clock controller driver for SC7280")
> >> Signed-off-by: Srinivasa Rao Mandadapu <quic_srivasam@...cinc.com>
> >> Tested-by: Mohammad Rafi Shaik <quic_mohs@...cinc.com>
> >> ---
> >>   drivers/clk/qcom/lpassaudiocc-sc7280.c | 13 +++++++++----
> >>   1 file changed, 9 insertions(+), 4 deletions(-)
> >>
> >> diff --git a/drivers/clk/qcom/lpassaudiocc-sc7280.c b/drivers/clk/qcom/lpassaudiocc-sc7280.c
> >> index 1339f92..8e2f433 100644
> >> --- a/drivers/clk/qcom/lpassaudiocc-sc7280.c
> >> +++ b/drivers/clk/qcom/lpassaudiocc-sc7280.c
> >> @@ -826,10 +829,12 @@ static int lpass_aon_cc_sc7280_probe(struct platform_device *pdev)
> >>                  return ret;
> >>
> >>          if (of_property_read_bool(pdev->dev.of_node, "qcom,adsp-pil-mode")) {
> >> -               lpass_audio_cc_sc7280_regmap_config.name = "cc";
> >> -               desc = &lpass_cc_sc7280_desc;
> >> -               ret = qcom_cc_probe(pdev, desc);
> >> -               goto exit;
> >> +               res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "cc");
> > We shouldn't need to check for reg-name property. Instead, the index
> > should be the only thing that matters.
>
> As qcom_cc_probe() function is mapping the zero index reg property, and
>
> in next implementation qcom_cc_really_probe() is also probing zero index
> reg property,
>
> unable to map the same region twice.

Use qcom_cc_probe_by_index()?

>
> Hence all I want here is to skip this cc clock probing by keeping some
> check.
>
> If we remove, it may cause ABI break.
>

I'm not sure what you mean here about ABI break, but hopefully just
using qcom_cc_probe_by_index() works!