linux-kernel - Re: [PATCH 1/2] arm64: dts: mediatek: mt8192: Make sure MSDCPLL's rate is 400MHz

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 for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date:   Mon, 19 Jun 2023 09:33:00 +0200
From:   AngeloGioacchino Del Regno 
        <angelogioacchino.delregno@...labora.com>
To:     Matthias Brugger <matthias.bgg@...il.com>,
        TingHan Shen (沈廷翰) 
        <TingHan.Shen@...iatek.com>,
        Wenbin Mei (梅文彬) 
        <Wenbin.Mei@...iatek.com>
Cc:     "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "robh+dt@...nel.org" <robh+dt@...nel.org>,
        "linux-mediatek@...ts.infradead.org" 
        <linux-mediatek@...ts.infradead.org>,
        "devicetree@...r.kernel.org" <devicetree@...r.kernel.org>,
        "kernel@...labora.com" <kernel@...labora.com>,
        Seiya Wang (王迺君) 
        <seiya.wang@...iatek.com>,
        "conor+dt@...nel.org" <conor+dt@...nel.org>,
        Chun-Jie Chen (陳浚桀) 
        <Chun-Jie.Chen@...iatek.com>,
        "krzysztof.kozlowski+dt@...aro.org" 
        <krzysztof.kozlowski+dt@...aro.org>,
        Weiyi Lu (呂威儀) <Weiyi.Lu@...iatek.com>,
        "ikjn@...omium.org" <ikjn@...omium.org>,
        "linux-arm-kernel@...ts.infradead.org" 
        <linux-arm-kernel@...ts.infradead.org>
Subject: Re: [PATCH 1/2] arm64: dts: mediatek: mt8192: Make sure MSDCPLL's
 rate is 400MHz

Il 15/06/23 13:16, Matthias Brugger ha scritto:
> 
> 
> On 15/06/2023 11:51, TingHan Shen (沈廷翰) wrote:
>> Hi AngeloGioacchino,
>>
>> On Mon, 2023-05-22 at 11:30 +0200, AngeloGioacchino Del Regno wrote:
>>> External email : Please do not click links or open attachments until you have 
>>> verified the sender or the content.
>>>
>>>
>>> Some bootloaders will set MSDCPLL's rate lower than 400MHz: what I have
>>> seen is this clock being set at around 384MHz.
>>> This is a performance concern (and possibly a stability one, for picky
>>> eMMC/SD cards) as the MSDC controller's internal divier will choose a
>>> frequency that is lower than expected, in the end causing a difference
>>> in the expected mmc/sd device's timings.
>>>
>>> Make sure that the MSDCPLL frequency is always set to 400MHz to both
>>> improve performance and reliability of the sd/mmc storage.
>>>
>>> Fixes: 5d2b897bc6f5 ("arm64: dts: mediatek: Add mt8192 clock controllers")
>>> Signed-off-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@...labora.com>
>>> ---
>>>   arch/arm64/boot/dts/mediatek/mt8192.dtsi | 2 ++
>>>   1 file changed, 2 insertions(+)
>>>
>>> diff --git a/arch/arm64/boot/dts/mediatek/mt8192.dtsi 
>>> b/arch/arm64/boot/dts/mediatek/mt8192.dtsi
>>> index 5c30caf74026..6fc14004f6fd 100644
>>> --- a/arch/arm64/boot/dts/mediatek/mt8192.dtsi
>>> +++ b/arch/arm64/boot/dts/mediatek/mt8192.dtsi
>>> @@ -677,6 +677,8 @@ apmixedsys: syscon@...0c000 {
>>>                          compatible = "mediatek,mt8192-apmixedsys", "syscon";
>>>                          reg = <0 0x1000c000 0 0x1000>;
>>>                          #clock-cells = <1>;
>>> +                       assigned-clocks = <&apmixedsys CLK_APMIXED_MSDCPLL>;
>>> +                       assigned-clock-rates = <400000000>;
>>>                  };
>>>
>>>                  systimer: timer@...17000 {
>>> -- 
>>> 2.40.1
>>>
>>
>> Comment from mtk emmc owner,
>>
>> "As we all know, the clock has some jitter, when we set MSDCPLL to 400M,
>> but the actual measurement is not exactly 200M.
>> For eMMC, the spec stipulates that clock cannot exceed 200M.
>> If MSDCPLL is set to 400M, the actual measurement may exceed the spec.
>> So we set MSDCPLL to 384M in the bootloader stage to avoid exceeding the spec."
>>

Thanks for the comment,
I haven't seen any issue with this commit, if not a slight performance improvement,
on MT8192 and MT8195, but if there's risk to overclock the card, then it's not ok.

In any case, what is the expected jitter percentage?
eMMC/SD cards do expect jitter by spec anyway.

Thanks,
Angelo


> 
> Thanks for the feedback. Given that I'm not aware of any regressions that got fixed 
> by this commits I will drop this series for now. We can keep on the discussion and 
> if needed add them in a later stage.
> 
> Regards,
> Matthias
>