linux-kernel - Re: [PATCHv3 5/5] clk: samsung: exynos542x: Move FSYS subsystem clocks to its sub-CMU

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]
Message-ID: <20200312113618.GA6206@pi3>
Date:   Thu, 12 Mar 2020 12:36:18 +0100
From:   Krzysztof Kozlowski <krzk@...nel.org>
To:     Anand Moon <linux.amoon@...il.com>
Cc:     Linux USB Mailing List <linux-usb@...r.kernel.org>,
        devicetree <devicetree@...r.kernel.org>,
        linux-arm-kernel <linux-arm-kernel@...ts.infradead.org>,
        linux-samsung-soc@...r.kernel.org,
        Linux Kernel <linux-kernel@...r.kernel.org>,
        "open list:COMMON CLK FRAMEWORK" <linux-clk@...r.kernel.org>,
        Rob Herring <robh+dt@...nel.org>,
        Kukjin Kim <kgene@...nel.org>,
        Marek Szyprowski <m.szyprowski@...sung.com>,
        Bartlomiej Zolnierkiewicz <b.zolnierkie@...sung.com>,
        Felipe Balbi <balbi@...nel.org>,
        Sylwester Nawrocki <s.nawrocki@...sung.com>,
        Tomasz Figa <tomasz.figa@...il.com>,
        Chanwoo Choi <cw00.choi@...sung.com>,
        Michael Turquette <mturquette@...libre.com>,
        Stephen Boyd <sboyd@...nel.org>
Subject: Re: [PATCHv3 5/5] clk: samsung: exynos542x: Move FSYS subsystem
 clocks to its sub-CMU

On Thu, Mar 12, 2020 at 04:04:57PM +0530, Anand Moon wrote:
> Hi Krzysztof,
> 
> Thanks for your review comments.
> 
> On Wed, 11 Mar 2020 at 20:12, Krzysztof Kozlowski <krzk@...nel.org> wrote:
> >
> > On Tue, Mar 10, 2020 at 07:48:54PM +0000, Anand Moon wrote:
> > > FSYS power domain support usbdrd3, pdma and usb2 power gaiting,
> > > hence move FSYS clk setting to sub-CMU block to support power domain
> > > on/off sequences for device nodes.
> > >
> > > Signed-off-by: Anand Moon <linux.amoon@...il.com>
> > > ---
> > > New patch in the series
> > > ---
> > >  drivers/clk/samsung/clk-exynos5420.c | 45 +++++++++++++++++++++-------
> > >  1 file changed, 34 insertions(+), 11 deletions(-)
> > >
> > > diff --git a/drivers/clk/samsung/clk-exynos5420.c b/drivers/clk/samsung/clk-exynos5420.c
> > > index c9e5a1fb6653..6c4c47dfcdce 100644
> > > --- a/drivers/clk/samsung/clk-exynos5420.c
> > > +++ b/drivers/clk/samsung/clk-exynos5420.c
> > > @@ -859,12 +859,6 @@ static const struct samsung_div_clock exynos5x_div_clks[] __initconst = {
> > >       DIV(0, "dout_maudio0", "mout_maudio0", DIV_MAU, 20, 4),
> > >       DIV(0, "dout_maupcm0", "dout_maudio0", DIV_MAU, 24, 8),
> > >
> > > -     /* USB3.0 */
> > > -     DIV(0, "dout_usbphy301", "mout_usbd301", DIV_FSYS0, 12, 4),
> > > -     DIV(0, "dout_usbphy300", "mout_usbd300", DIV_FSYS0, 16, 4),
> > > -     DIV(0, "dout_usbd301", "mout_usbd301", DIV_FSYS0, 20, 4),
> > > -     DIV(0, "dout_usbd300", "mout_usbd300", DIV_FSYS0, 24, 4),
> >
> > According to clock diagram these are still in CMU TOP, not FSYS.
> >
> > > -
> > >       /* MMC */
> > >       DIV(0, "dout_mmc0", "mout_mmc0", DIV_FSYS1, 0, 10),
> > >       DIV(0, "dout_mmc1", "mout_mmc1", DIV_FSYS1, 10, 10),
> > > @@ -1031,8 +1025,6 @@ static const struct samsung_gate_clock exynos5x_gate_clks[] __initconst = {
> > />
> > >       /* FSYS Block */
> > >       GATE(CLK_TSI, "tsi", "aclk200_fsys", GATE_BUS_FSYS0, 0, 0, 0),
> > > -     GATE(CLK_PDMA0, "pdma0", "aclk200_fsys", GATE_BUS_FSYS0, 1, 0, 0),
> > > -     GATE(CLK_PDMA1, "pdma1", "aclk200_fsys", GATE_BUS_FSYS0, 2, 0, 0),
> > >       GATE(CLK_UFS, "ufs", "aclk200_fsys2", GATE_BUS_FSYS0, 3, 0, 0),
> > >       GATE(CLK_RTIC, "rtic", "aclk200_fsys", GATE_IP_FSYS, 9, 0, 0),
> > >       GATE(CLK_MMC0, "mmc0", "aclk200_fsys2", GATE_IP_FSYS, 12, 0, 0),
> > > @@ -1040,9 +1032,6 @@ static const struct samsung_gate_clock exynos5x_gate_clks[] __initconst = {
> > >       GATE(CLK_MMC2, "mmc2", "aclk200_fsys2", GATE_IP_FSYS, 14, 0, 0),
> > >       GATE(CLK_SROMC, "sromc", "aclk200_fsys2",
> > >                       GATE_IP_FSYS, 17, CLK_IGNORE_UNUSED, 0),
> > > -     GATE(CLK_USBH20, "usbh20", "aclk200_fsys", GATE_IP_FSYS, 18, 0, 0),
> > > -     GATE(CLK_USBD300, "usbd300", "aclk200_fsys", GATE_IP_FSYS, 19, 0, 0),
> > > -     GATE(CLK_USBD301, "usbd301", "aclk200_fsys", GATE_IP_FSYS, 20, 0, 0),
> > >       GATE(CLK_SCLK_UNIPRO, "sclk_unipro", "dout_unipro",
> > >                       SRC_MASK_FSYS, 24, CLK_SET_RATE_PARENT, 0),
> > >
> > > @@ -1258,6 +1247,28 @@ static struct exynos5_subcmu_reg_dump exynos5x_gsc_suspend_regs[] = {
> > >       { DIV2_RATIO0, 0, 0x30 },       /* DIV dout_gscl_blk_300 */
> > >  };
> > >
> > > +/* USB3.0 */
> > > +static const struct samsung_div_clock exynos5x_fsys_div_clks[] __initconst = {
> > > +     DIV(0, "dout_usbphy301", "mout_usbd301", DIV_FSYS0, 12, 4),
> > > +     DIV(0, "dout_usbphy300", "mout_usbd300", DIV_FSYS0, 16, 4),
> > > +     DIV(0, "dout_usbd301", "mout_usbd301", DIV_FSYS0, 20, 4),
> > > +     DIV(0, "dout_usbd300", "mout_usbd300", DIV_FSYS0, 24, 4),
> > > +};
> > > +
> > > +static const struct samsung_gate_clock exynos5x_fsys_gate_clks[] __initconst = {
> > > +     GATE(CLK_PDMA0, "pdma0", "aclk200_fsys", GATE_BUS_FSYS0, 1, 0, 0),
> > > +     GATE(CLK_PDMA1, "pdma1", "aclk200_fsys", GATE_BUS_FSYS0, 2, 0, 0),
> > > +     GATE(CLK_USBH20, "usbh20", "aclk200_fsys", GATE_IP_FSYS, 18, 0, 0),
> > > +     GATE(CLK_USBD300, "usbd300", "aclk200_fsys", GATE_IP_FSYS, 19, 0, 0),
> > > +     GATE(CLK_USBD301, "usbd301", "aclk200_fsys", GATE_IP_FSYS, 20, 0, 0),
> > > +};
> > > +
> > > +static struct exynos5_subcmu_reg_dump exynos5x_fsys_suspend_regs[] = {
> > > +     { GATE_IP_FSYS, 0xffffffff, 0xffffffff }, /* FSYS gates */
> >
> > This looks wrong. GATE_IP_FSYS has fields also for FSYS2 clocks which
> > you are not suspending. They do not belong to this CMU.
> >
> 
> Ok. I change the from GATE_IP_FSYS to GATE_BUS_FSYS0 in the above
> exynos5x_fsys_gate_clks to make this consistent to used GATE_BUS_FSYS0 for CMU,
> with this change it works as per previously.

Wait, you should set here proper registers with proper mask.
> 
> > Don't you need to save also parts of GATE_BUS_FSYS0?
> 
> GATE_BUS_FSYS0 and GATE_IP_FSYS are already part of list
> of control register which are saved and restored during suspend and resume
> so no point in adding this here, I should drop the GATE_IP_FSYS reg
> dump over here.

Since registers are used in separate sub CMU devices, each should
save/restore its part.

Best regards,
Krzysztof