linux-kernel - Re: [PATCH] mmc: mmci_sdmmc: fix power on issue due to pwr

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:   Wed, 22 Apr 2020 15:40:05 +0200
From:   Ludovic BARRE <ludovic.barre@...com>
To:     Ulf Hansson <ulf.hansson@...aro.org>
CC:     Rob Herring <robh+dt@...nel.org>,
        Srini Kandagatla <srinivas.kandagatla@...aro.org>,
        Maxime Coquelin <mcoquelin.stm32@...il.com>,
        Alexandre Torgue <alexandre.torgue@...com>,
        Linux ARM <linux-arm-kernel@...ts.infradead.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        DTML <devicetree@...r.kernel.org>,
        "linux-mmc@...r.kernel.org" <linux-mmc@...r.kernel.org>,
        <linux-stm32@...md-mailman.stormreply.com>
Subject: Re: [PATCH] mmc: mmci_sdmmc: fix power on issue due to pwr_reg
 initialization

hi Ulf

Le 4/21/20 à 11:38 AM, Ulf Hansson a écrit :
> On Tue, 21 Apr 2020 at 11:25, Ulf Hansson <ulf.hansson@...aro.org> wrote:
>>
>> On Mon, 20 Apr 2020 at 18:18, Ludovic Barre <ludovic.barre@...com> wrote:
>>>
>>> This patch fix a power-on issue, and avoid to retry the power sequence.
>>>
>>> In power off sequence: sdmmc must set pwr_reg in "power-cycle" state
>>> (value 0x2), to prevent the card from being supplied through the signal
>>> lines (all the lines are driven low).
>>>
>>> In power on sequence: when the power is stable, sdmmc must set pwr_reg
>>> in "power-off" state (value 0x0) to drive all signal to high before to
>>> set "power-on".
>>
>> Just a question to gain further understanding.
>>
>> Let's assume that the controller is a power-on state, because it's
>> been initialized by the boot loader. When the mmc core then starts the
>> power-on sequence (not doing a power-off first), would $subject patch
>> then cause the
>> MMCIPOWER to remain as is, or is it going to be overwritten?

On sdmmc controller, the PWRCTRL[1:0] field of MMCIPOWER register allow
to manage sd lines and has a specific bahavior.

PWRCTRL value:
  - 0x0: After reset, Reset: the SDMMC is disabled and the clock to the
         Card is stopped, SDMMC_D[7:0], and SDMMC_CMD are HiZ and
         SDMMC_CK is driven low.
         When written 00, power-off: the SDMMC is disabled and the clock
         to the card is stopped, SDMMC_D[7:0], SDMMC_CMD and SDMMC_CK
         are driven high.

  - 0x2: Power-cycle, the SDMMC is disabled and the clock to the card is
         stopped, SDMMC_D[7:0], SDMMC_CMD and SDMMC_CK are driven low.

  - 0x3: Power-on: the card is clocked, The first 74 SDMMC_CK cycles the
         SDMMC is still disabled. After the 74 cycles the SDMMC is
         enabled and the SDMMC_D[7:0], SDMMC_CMD and SDMMC_CK are
         controlled according the SDMMC operation.
         **Any further write will be ignored, PWRCTRL value
         will keep 0x3**. when the SDMMC is ON (0x3) only a reset could
         change pwrctrl value and the state of sdmmc lines.

So if the lines are already "ON", the power-on sequence (decribed in
commit message) not overwrite the pwctrl field and not disturb the sdmmc 
lines.

>>
>> I am a little worried that we may start to rely on boot loader
>> conditions, which isn't really what we want either...
>>

We not depend of boot loader conditions.

This patch simply allows to drive high the sd lines before to set
"power-on" value (no effect if already power ON).

>>>
>>> To avoid writing the same value to the power register several times, this
>>> register is cached by the pwr_reg variable. At probe pwr_reg is initialized
>>> to 0 by kzalloc of mmc_alloc_host.
>>>
>>> Like pwr_reg value is 0 at probing, the power on sequence fail because
>>> the "power-off" state is not writes (value 0x0) and the lines
>>> remain drive to low.
>>>
>>> This patch initializes "pwr_reg" variable with power register value.
>>> This it done in sdmmc variant init to not disturb default mmci behavior.
>>>
>>> Signed-off-by: Ludovic Barre <ludovic.barre@...com>
>>
>> Besides the comment, the code and the approach seems reasonable to me.
> 
> Another related question. I just realized why you probably haven't set
> .pwrreg_nopower for the variant_stm32_sdmmc and variant_stm32_sdmmcv2.
> 
> I guess it's because you need a slightly different way to restore the
> context of MMCIPOWER register at ->runtime_resume(), rather than just
> re-writing it with the saved register values. Is this something that
> you are looking into as well?

Yes exactly, the sequence is slightly different. I can't write 0 on 
mmci_runtime_suspend, and can't just re-writing the saved register.

Regards
Ludo

> 
> [...]
> 
> Kind regards
> Uffe
>