linux-kernel - Re: [PATCH 3/4] phy: qcom-qmp: Add optional SW reset

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:   Fri, 20 Dec 2019 15:41:30 +0800
From:   Can Guo <cang@...eaurora.org>
To:     Vinod Koul <vkoul@...nel.org>
Cc:     Asutosh Das <asutoshd@...eaurora.org>,
        Kishon Vijay Abraham I <kishon@...com>,
        linux-arm-msm@...r.kernel.org,
        Bjorn Andersson <bjorn.andersson@...aro.org>,
        Andy Gross <agross@...nel.org>,
        Jeffrey Hugo <jeffrey.l.hugo@...il.com>,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH 3/4] phy: qcom-qmp: Add optional SW reset

On 2019-12-20 15:10, Vinod Koul wrote:
> On 20-12-19, 14:00, Can Guo wrote:
>> On 2019-12-20 12:24, Vinod Koul wrote:
>> > On 20-12-19, 08:49, cang@...eaurora.org wrote:
>> > > On 2019-12-20 08:22, cang@...eaurora.org wrote:
>> > > > On 2019-12-19 23:04, Vinod Koul wrote:
>> > > >
>> > > > >  	/* start SerDes and Phy-Coding-Sublayer */
>> > > > >  	qphy_setbits(pcs, cfg->regs[QPHY_START_CTRL], cfg->start_ctrl);
>> > >
>> > > I thought your change would be like this
>> > >
>> > > diff --git a/drivers/phy/qualcomm/phy-qcom-qmp.c
>> > > b/drivers/phy/qualcomm/phy-qcom-qmp.c
>> > > index 8e642a6..a4ab4b7 100755
>> > > --- a/drivers/phy/qualcomm/phy-qcom-qmp.c
>> > > +++ b/drivers/phy/qualcomm/phy-qcom-qmp.c
>> > > @@ -166,6 +166,7 @@ static const unsigned int
>> > > sdm845_ufsphy_regs_layout[] =
>> > > {
>> > >  };
>> > >
>> > >  static const unsigned int sm8150_ufsphy_regs_layout[] = {
>> > > +       [QPHY_SW_RESET]                 = 0x08,
>> > >         [QPHY_START_CTRL]               = 0x00,
>> > >         [QPHY_PCS_READY_STATUS]         = 0x180,
>> > >  };
>> > > @@ -1390,7 +1391,6 @@ static const struct qmp_phy_cfg
>> > > sm8150_ufsphy_cfg = {
>> > >         .pwrdn_ctrl             = SW_PWRDN,
>> > >
>> > >         .is_dual_lane_phy       = true,
>> > > -       .no_pcs_sw_reset        = true,
>> > >  };
>> > >
>> > >  static void qcom_qmp_phy_configure(void __iomem *base,
>> > > @@ -1475,6 +1475,9 @@ static int qcom_qmp_phy_com_init(struct
>> > > qmp_phy *qphy)
>> > >                              SW_USB3PHY_RESET_MUX | SW_USB3PHY_RESET);
>> > >         }
>> > >
>> > > +       if ((cfg->type == PHY_TYPE_UFS) && (!cfg->no_pcs_sw_reset))
>> > > +               qphy_setbits(pcs, cfg->regs[QPHY_SW_RESET], SW_RESET);
>> >
>> > Well am not sure if no_pcs_sw_reset would do this and side effect on
>> > other phys (somehow older ones dont seem to need this). That was the
>> > reason for a new flag and to be used for specific instances
>> >
>> > Thanks
>> 
>> Hi Vinod,
>> 
>> That is why I added the check as cfg->type == PHY_TYPE_UFS, meaning 
>> this
>> change will only apply to UFS.
>> FYI, start from 8150(include 8150), QPHY_SW_RESET is present in PHY's
>> PCS register. no_pcs_sw_reset = TRUE should only be given to 845 and 
>> older
>> targets, like 8998, because they don't have this QPHY_SW_RESET in 
>> PHY's
>> register per their design, that's why they leverage the reset control
>> provided by UFS controller.
> 
> I have removed no_pcs_sw_reset and tested.
> 
> Well as you said even with UFS we have variations between various 
> chips,
> so I thought leaving it separate might be better than creating a chance
> of regression on older platforms!
> 
> Moreover, are we sure that the reset wont be there for other qmp phy's
> in future other than UFS...
> 
> Thanks

Hi Vinod

We are just removing the no_pcs_sw_reset for 8150, right? Why is it
possibly impacting 845 or older paltforms?

In future, we will no longer need no_pcs_sw_reset for any newer QCOM UFS
PHY designs, as it is only for 845 and older platforms.

I am sure QPHY_SW_RESET will be within PHY's address space since 8150.
Otherwise, it will be a regression in UFS PHY design. We had a lot of
discussion about this on 845 years ago, then design team decided to add
it on later platforms, so I don't see a reason to remove it again. :)

I am not sure about the other qmp phys, but so long as UFS PHY needs the
reset, we need to keep it, as phy-qcom-qmp.c is a common driver. I am
not sure if I get your point here. Please correct me I am wrong.

Thanks,

Can Guo.