| 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
| ||
|
Message-ID: <efbc7027-5717-ba29-2b6c-e9b780558315@quicinc.com>
Date: Wed, 8 Mar 2023 11:22:27 +0530
From: Varadarajan Narayanan <quic_varada@...cinc.com>
To: Dmitry Baryshkov <dmitry.baryshkov@...aro.org>
CC: Andy Gross <agross@...nel.org>,
Bjorn Andersson <andersson@...nel.org>,
Konrad Dybcio <konrad.dybcio@...aro.org>,
Rob Herring <robh+dt@...nel.org>,
Krzysztof Kozlowski <krzysztof.kozlowski+dt@...aro.org>,
<linux-arm-msm@...r.kernel.org>, <devicetree@...r.kernel.org>,
<linux-kernel@...r.kernel.org>, Georgi Djakov <djakov@...nel.org>
Subject: Re: [PATCH 7/8] arm64: dts: qcom: ipq9574: Add USB related nodes
On 3/7/2023 5:19 PM, Dmitry Baryshkov wrote:
> On 07/03/2023 08:36, Varadarajan Narayanan wrote:
>>
>> On 3/6/2023 5:21 PM, Dmitry Baryshkov wrote:
>>> On 06/03/2023 13:26, Varadarajan Narayanan wrote:
>>>> Dmitry,
>>>>
>>>> On 3/2/2023 9:52 PM, Dmitry Baryshkov wrote:
>>>>> On Thu, 2 Mar 2023 at 11:57, Varadarajan Narayanan
>>>>> <quic_varada@...cinc.com> wrote:
>>>>>> Add USB phy and controller related nodes
>>>>>>
>>>>>> Signed-off-by: Varadarajan Narayanan <quic_varada@...cinc.com>
>>>>>> ---
>>>>>> arch/arm64/boot/dts/qcom/ipq9574.dtsi | 92
>>>>>> +++++++++++++++++++++++++++++++++++
>>>>>> 1 file changed, 92 insertions(+)
>>>>>>
>>>>>> diff --git a/arch/arm64/boot/dts/qcom/ipq9574.dtsi
>>>>>> b/arch/arm64/boot/dts/qcom/ipq9574.dtsi
>>>>>> index 2bb4053..319b5bd 100644
>>>>>> --- a/arch/arm64/boot/dts/qcom/ipq9574.dtsi
>>>>>> +++ b/arch/arm64/boot/dts/qcom/ipq9574.dtsi
>>>
>>> [skipped]
>>>
>>>
>>>>>> + usb3: usb3@...0000 {
>>>>> You know the drill. This node is in the wrong place.
>>>>>
>>>>>> + compatible = "qcom,dwc3";
>>>>>> + reg = <0x8AF8800 0x400>;
>>>>>> + #address-cells = <1>;
>>>>>> + #size-cells = <1>;
>>>>>> + ranges;
>>>>>> +
>>>>>> + clocks = <&gcc GCC_SNOC_USB_CLK>,
>>>>>> + <&gcc GCC_ANOC_USB_AXI_CLK>,
>>>>>> + <&gcc GCC_USB0_MASTER_CLK>,
>>>>>> + <&gcc GCC_USB0_SLEEP_CLK>,
>>>>>> + <&gcc GCC_USB0_MOCK_UTMI_CLK>;
>>>>>> +
>>>>>> + clock-names = "sys_noc_axi",
>>>>>> + "anoc_axi",
>>>>>> + "master",
>>>>>> + "sleep",
>>>>>> + "mock_utmi";
>>>>> Please fix the indentation of the lists.
>>>>>
>>>>>> +
>>>>>> + assigned-clocks = <&gcc GCC_SNOC_USB_CLK>,
>>>>>> + <&gcc
>>>>>> GCC_ANOC_USB_AXI_CLK>,
>>>>> Why do you assign clock rates to the NOC clocks? Should they be set
>>>>> using the interconnect instead?
>>>>
>>>> The SNOC and ANOC run at a fixed speed of 350MHz and 342MHz
>>>> respectively and are not scaled. These clocks are for the interface
>>>> between the USB block and the SNOC/ANOC. Do we still need to use
>>>> interconnect?
>>>
>>> Maybe I misunderstand something here. If the snoc and anoc speeds
>>> are at 350 MHz and 342 MHz, why do you assign clock-rates of 200 MHz?
>>>
>>> Is it enough to call clk_prepare_enable() for these clocks or the
>>> rate really needs to be set?
>>
>> The rate of 200MHz is not being set for the SNOC/ANOC. It is for the
>> NIU that connects the USB and SNOC/ANOC. The reason for setting the
>> rate to 200MHz is to configure the RCG parent for these interface
>> clocks. That said can we configure this RCG standalone in the driver
>> and enable these clocks?
>
> We discussed this separately with Georgi Djakov. Let me quote his IRC
> message: "it sounds like this is for USB port that connects to the
> NOC. if bandwidth scaling is not needed (or other interconnect
> configuration), then maybe this can go without interconnect provider
> driver."
>
> However as we discover more and more about this platform (e.g. PCIe
> using the aggre_noc region to setup some magic registers, see [1]),
> I'm more and more biased towards suggesting implementing the
> interconnect driver to setup all these tiny little things. With the DT
> tree being an ABI, it is much preferable to overestimate the needs
> rather than underestimating them (and having to cope with the
> backwards compatibility issues).
>
> Generally I think that PCIe/USB/whatever should not poke into NoC
> registers or NoC/NIU clocks directly (because this is a very
> platform-specific item). Rather than that it should tell the
> icc/opp/whatever subsystem, "please configure the SoC for me to work".
>
> [1]
> https://lore.kernel.org/linux-arm-msm/30cf9717-dcca-e984-c506-c71b7f8e32cd@quicinc.com/
Sure, will post a new revision that uses interconnect subsystem.
Thanks
Varada
>>>>>> + <&gcc GCC_USB0_MASTER_CLK>,
>>>>>> + <&gcc
>>>>>> GCC_USB0_MOCK_UTMI_CLK>;
>>>>>> + assigned-clock-rates = <200000000>,
>>>>>> + <200000000>,
>>>>>> + <200000000>,
>>>>>> + <24000000>;
>>>>>> +
>>>>>> + resets = <&gcc GCC_USB_BCR>;
>>>>>> + status = "disabled";
>>>>>> +
>>>>>> + dwc_0: dwc3@...0000 {
>>>>>> + compatible = "snps,dwc3";
>>>>>> + reg = <0x8A00000 0xcd00>;
>>>>>> + clock-names = "ref";
>>>>>> + clocks = <&gcc
>>>>>> GCC_USB0_MOCK_UTMI_CLK>;
>>>>> clocks before clock-names
>>>>>
>>>>>> + interrupts = <GIC_SPI 140 IRQ_TYPE_LEVEL_HIGH>;
>>>>>> + phys = <&qusb_phy_0>, <&usb0_ssphy>;
>>>>>> + phy-names = "usb2-phy", "usb3-phy";
>>>>>> + tx-fifo-resize;
>>>>>> + snps,dis_ep_cache_eviction;
>>>>>> + snps,is-utmi-l1-suspend;
>>>>>> + snps,hird-threshold = /bits/ 8
>>>>>> <0x0>;
>>>>>> + snps,dis_u2_susphy_quirk;
>>>>>> + snps,dis_u3_susphy_quirk;
>>>>>> + snps,quirk-frame-length-adjustment = <0x0A87F0A0>;
>>>>>> + dr_mode = "host";
>>>>>> + };
>>>>>> + };
>>>>>> +
>>>>>> pcie0_phy: phy@...00 {
>>>>>> compatible =
>>>>>> "qcom,ipq9574-qmp-gen3x1-pcie-phy";
>>>>>> reg = <0x00084000 0x1bc>; /* Serdes PLL */
>>>>>> --
>>>>>> 2.7.4
>>>>
>>>> Will address these and post a new revision.
>>>>
>>>> Thanks
>>>>
>>>> Varada
>>>>
>>>
>
Powered by blists - more mailing lists