Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Mon, 21 Sep 2020 19:36:06 -0500
From:   Bjorn Andersson <bjorn.andersson@linaro.org>
To:     nguyenb@codeaurora.org
Cc:     Rob Herring <robh@kernel.org>, Can Guo <cang@codeaurora.org>,
        Asutosh Das <asutoshd@codeaurora.org>,
        "Martin K. Petersen" <martin.petersen@oracle.com>,
        SCSI <linux-scsi@vger.kernel.org>,
        linux-arm-msm <linux-arm-msm@vger.kernel.org>,
        Avri Altman <Avri.Altman@wdc.com>,
        Vinod Koul <vkoul@kernel.org>,
        "open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS" 
        <devicetree@vger.kernel.org>,
        open list <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH v1 1/2] scsi: dt-bindings: ufs: Add vcc-voltage-level for
 UFS
Message-ID: <20200922003606.GA40811@yoga>
References: <cover.1598939393.git.nguyenb@codeaurora.org>
 <0a9d395dc38433501f9652a9236856d0ac840b77.1598939393.git.nguyenb@codeaurora.org>
 <20200914183505.GA357@bogus>
 <d332e61cea4fef237507f1404efa724a@codeaurora.org>
 <CAL_Jsq+YV-GjAhVVHtgNz6xFR=bEgSwWKY+QGixRQJ5Ov75pag@mail.gmail.com>
 <e489cee219d48e9f5e48dc30518f445b@codeaurora.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <e489cee219d48e9f5e48dc30518f445b@codeaurora.org>
Precedence: bulk

On Mon 21 Sep 19:22 CDT 2020, nguyenb@codeaurora.org wrote:

> On 2020-09-18 12:01, Rob Herring wrote:
> > On Tue, Sep 15, 2020 at 2:10 AM <nguyenb@codeaurora.org> wrote:
> > > 
> > > On 2020-09-14 11:35, Rob Herring wrote:
> > > > On Mon, Aug 31, 2020 at 11:00:47PM -0700, Bao D. Nguyen wrote:
> > > >> UFS's specifications supports a range of Vcc operating
> > > >> voltage levels. Add documentation for the UFS's Vcc voltage
> > > >> levels setting.
> > > >>
> > > >> Signed-off-by: Can Guo <cang@codeaurora.org>
> > > >> Signed-off-by: Asutosh Das <asutoshd@codeaurora.org>
> > > >> Signed-off-by: Bao D. Nguyen <nguyenb@codeaurora.org>
> > > >> ---
> > > >>  Documentation/devicetree/bindings/ufs/ufshcd-pltfrm.txt | 2 ++
> > > >>  1 file changed, 2 insertions(+)
> > > >>
> > > >> diff --git a/Documentation/devicetree/bindings/ufs/ufshcd-pltfrm.txt
> > > >> b/Documentation/devicetree/bindings/ufs/ufshcd-pltfrm.txt
> > > >> index 415ccdd..7257b32 100644
> > > >> --- a/Documentation/devicetree/bindings/ufs/ufshcd-pltfrm.txt
> > > >> +++ b/Documentation/devicetree/bindings/ufs/ufshcd-pltfrm.txt
> > > >> @@ -23,6 +23,8 @@ Optional properties:
> > > >>                            with "phys" attribute, provides phandle to
> > > >> UFS PHY node
> > > >>  - vdd-hba-supply        : phandle to UFS host controller supply
> > > >> regulator node
> > > >>  - vcc-supply            : phandle to VCC supply regulator node
> > > >> +- vcc-voltage-level     : specifies voltage levels for VCC supply.
> > > >> +                          Should be specified in pairs (min, max),
> > > >> units uV.
> > > >
> > > > The expectation is the regulator pointed to by 'vcc-supply' has the
> > > > voltage constraints. Those constraints are supposed to be the board
> > > > constraints, not the regulator operating design constraints. If that
> > > > doesn't work for your case, then it should be addressed in a common way
> > > > for the regulator binding.
> > > The UFS regulator has a min_uV and max_uV limits. Currently, the min
> > > and
> > > max are hardcoded
> > > to UFS2.1 Spec allowed values of 2.7V and 3.6V respectively.
> > > With this change, I am trying to fix a couple issues:
> > > 1. The 2.7V min value only applies to UFS2.1 devices. with UFS3.0+
> > > devices, the VCC min should be 2.4V.
> > > Hardcoding the min_uV to 2.7V does not work for UFS3.0+ devices.
> > 
> > Don't you know the device version attached and can adjust the voltage
> > based on that? Or you have to set the voltage first?
> Yes it is one of the solutions. Once detect the UFS device is version 3.0+,
> you can lower
> the voltage to 2.5V from the hardcoded value used by the driver. However, to
> change the
> Vcc voltage, the host needs to follow a sequence to ensure safe operations
> after Vcc change
> (device has to be in sleep mode, Vcc needs to go down to 0 then up to 2.5V.)
> Also same sequence is repeated for every host initialization which is
> inconvenient.
> 

It sounds like you're suggesting that we detect the UFS device using
some voltage, then depending on version we might lower it to 2.5V.

I'm afraid I don't see any of this either documented or implemented in
these patches.

What is this initial detection voltage and how to you configure it?

Regards,
Bjorn

> > 
> > > 2. Allow users to select a different Vcc voltage within the allowed
> > > range.
> > > Using the min value, the UFS device is operating at marginal Vcc
> > > voltage.
> > > In addition the PMIC and the board designs may add some variables
> > > especially at extreme
> > > temperatures. We observe stability issues when using the min Vcc
> > > voltage.
> > 
> > Again, we have standard regulator properties for this already that you
> > can tune per board.
> Thank you for the suggestion.
> 
> > 
> > Rob