Hi, Shawn
Anson Huang
Best Regards!
> -----Original Message-----
> From: Shawn Guo <shawnguo@xxxxxxxxxx>
> Sent: Friday, September 28, 2018 4:45 PM
> To: Anson Huang <anson.huang@xxxxxxx>
> Cc: robh+dt@xxxxxxxxxx; mark.rutland@xxxxxxx; s.hauer@xxxxxxxxxxxxxx;
> kernel@xxxxxxxxxxxxxx; Fabio Estevam <fabio.estevam@xxxxxxx>;
> linux@xxxxxxxxxxxxxxx; mturquette@xxxxxxxxxxxx; sboyd@xxxxxxxxxx; Jacky
> Bai <ping.bai@xxxxxxx>; A.s. Dong <aisheng.dong@xxxxxxx>;
> devicetree@xxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx;
> linux-arm-kernel@xxxxxxxxxxxxxxxxxxx; linux-clk@xxxxxxxxxxxxxxx; dl-linux-imx
> <linux-imx@xxxxxxx>
> Subject: Re: [PATCH V2 1/4] ARM: imx: add i.mx6ulz msl support
>
> On Wed, Sep 19, 2018 at 02:04:45PM +0800, Anson Huang wrote:
> > The i.MX 6ULZ processor is a high-performance, ultra cost-efficient
> > consumer Linux processor featuring an advanced implementation of a
> > single Arm(r) Cortex(r)-A7 core, which operates at speeds up to 900 MHz.
> >
> > This patch adds basic MSL support for i.MX6ULZ, the i.MX6ULZ has same
> > soc_id as i.MX6ULL, and SRC_SBMR2 bit[6] is to differentiate i.MX6ULZ
> > from i.MX6ULL, 1'b1 means i.MX6ULZ and 1'b0 means i.MX6ULL.
> >
> > Signed-off-by: Anson Huang <Anson.Huang@xxxxxxx>
> > ---
> > arch/arm/mach-imx/anatop.c | 20 ++++++++++++++++++++
> > arch/arm/mach-imx/cpu.c | 3 +++
> > arch/arm/mach-imx/mach-imx6ul.c | 1 +
> > arch/arm/mach-imx/mxc.h | 7 +++++++
> > arch/arm/mach-imx/pm-imx6.c | 4 ++--
> > 5 files changed, 33 insertions(+), 2 deletions(-)
> >
> > diff --git a/arch/arm/mach-imx/anatop.c b/arch/arm/mach-imx/anatop.c
> > index 61f3d94..45d618a 100644
> > --- a/arch/arm/mach-imx/anatop.c
> > +++ b/arch/arm/mach-imx/anatop.c
> > @@ -31,6 +31,8 @@
> > #define ANADIG_DIGPROG_IMX6SL 0x280
> > #define ANADIG_DIGPROG_IMX7D 0x800
> >
> > +#define SRC_SBMR2 0x1c
> > +
> > #define BM_ANADIG_REG_2P5_ENABLE_WEAK_LINREG 0x40000
> > #define BM_ANADIG_REG_2P5_ENABLE_PULLDOWN 0x8
> > #define BM_ANADIG_REG_CORE_FET_ODRIVE 0x20000000
> > @@ -148,6 +150,24 @@ void __init imx_init_revision_from_anatop(void)
> > major_part = (digprog >> 8) & 0xf;
> > minor_part = digprog & 0xf;
> > revision = ((major_part + 1) << 4) | minor_part;
> > +
> > + if ((digprog >> 16) == MXC_CPU_IMX6ULL) {
> > + void __iomem *src_base;
> > + u32 sbmr2;
> > +
> > + np = of_find_compatible_node(NULL, NULL,
> > + "fsl,imx6ul-src");
> > + src_base = of_iomap(np, 0);
> > + WARN_ON(!src_base);
> > + sbmr2 = readl_relaxed(src_base + SRC_SBMR2);
> > + iounmap(src_base);
> > +
> > + /* src_sbmr2 bit 6 is to identify if it is i.MX6ULZ */
> > + if (sbmr2 & (1 << 6)) {
> > + digprog &= ~(0xff << 16);
> > + digprog |= (MXC_CPU_IMX6ULZ << 16);
> > + }
> > + }
> > }
> >
> > mxc_set_cpu_type(digprog >> 16 & 0xff); diff --git
> > a/arch/arm/mach-imx/cpu.c b/arch/arm/mach-imx/cpu.c index
> > c6b1bf9..c73593e 100644
> > --- a/arch/arm/mach-imx/cpu.c
> > +++ b/arch/arm/mach-imx/cpu.c
> > @@ -136,6 +136,9 @@ struct device * __init imx_soc_device_init(void)
> > case MXC_CPU_IMX6ULL:
> > soc_id = "i.MX6ULL";
> > break;
> > + case MXC_CPU_IMX6ULZ:
> > + soc_id = "i.MX6ULZ";
> > + break;
> > case MXC_CPU_IMX6SLL:
> > soc_id = "i.MX6SLL";
> > break;
> > diff --git a/arch/arm/mach-imx/mach-imx6ul.c
> > b/arch/arm/mach-imx/mach-imx6ul.c index 6cb8a22..4ffe3c8 100644
> > --- a/arch/arm/mach-imx/mach-imx6ul.c
> > +++ b/arch/arm/mach-imx/mach-imx6ul.c
> > @@ -90,6 +90,7 @@ static void __init imx6ul_init_late(void) static
> > const char * const imx6ul_dt_compat[] __initconst = {
> > "fsl,imx6ul",
> > "fsl,imx6ull",
> > + "fsl,imx6ulz",
>
> Can we have "fsl,imx6ull" on the DT compatible, so that we can save the
> changes on kernel side, like this and the clock driver update (patch #2)?
>
> compatible = "fsl,imx6ull", "fsl,imx6ulz";
>
> I'm not sure if there is any problem with this approach. But you can think
> about it.
>
> Shawn
Using this approach will save the changes in clk-imx6ul.c and mach-imx6ul.c,
but other changes will be still needed, since it is defined as a new SoC other
than a i.MX6ULL with different fuse settings. I can do the changes you suggested
to save those 2 files changes if you prefer this way, but current implementation
should also make sense if think about it from a new SoC perspective? What do
you prefer?
Thanks.
Anson.
>
> > NULL,
> > };
> >
> > diff --git a/arch/arm/mach-imx/mxc.h b/arch/arm/mach-imx/mxc.h index
> > 026e2ca..b130a53 100644
> > --- a/arch/arm/mach-imx/mxc.h
> > +++ b/arch/arm/mach-imx/mxc.h
> > @@ -40,6 +40,8 @@
> > #define MXC_CPU_IMX6Q 0x63
> > #define MXC_CPU_IMX6UL 0x64
> > #define MXC_CPU_IMX6ULL 0x65
> > +/* virtual cpu id for i.mx6ulz */
> > +#define MXC_CPU_IMX6ULZ 0x6b
> > #define MXC_CPU_IMX6SLL 0x67
> > #define MXC_CPU_IMX7D 0x72
> >
> > @@ -80,6 +82,11 @@ static inline bool cpu_is_imx6ull(void)
> > return __mxc_cpu_type == MXC_CPU_IMX6ULL; }
> >
> > +static inline bool cpu_is_imx6ulz(void) {
> > + return __mxc_cpu_type == MXC_CPU_IMX6ULZ; }
> > +
> > static inline bool cpu_is_imx6sll(void) {
> > return __mxc_cpu_type == MXC_CPU_IMX6SLL; diff --git
> > a/arch/arm/mach-imx/pm-imx6.c b/arch/arm/mach-imx/pm-imx6.c index
> > 529f4b5..87f45b9 100644
> > --- a/arch/arm/mach-imx/pm-imx6.c
> > +++ b/arch/arm/mach-imx/pm-imx6.c
> > @@ -313,7 +313,7 @@ int imx6_set_lpm(enum mxc_cpu_pwr_mode mode)
> > if (cpu_is_imx6sl())
> > val |= BM_CLPCR_BYPASS_PMIC_READY;
> > if (cpu_is_imx6sl() || cpu_is_imx6sx() || cpu_is_imx6ul() ||
> > - cpu_is_imx6ull() || cpu_is_imx6sll())
> > + cpu_is_imx6ull() || cpu_is_imx6sll() || cpu_is_imx6ulz())
> > val |= BM_CLPCR_BYP_MMDC_CH0_LPM_HS;
> > else
> > val |= BM_CLPCR_BYP_MMDC_CH1_LPM_HS; @@ -331,7
> +331,7 @@ int
> > imx6_set_lpm(enum mxc_cpu_pwr_mode mode)
> > if (cpu_is_imx6sl() || cpu_is_imx6sx())
> > val |= BM_CLPCR_BYPASS_PMIC_READY;
> > if (cpu_is_imx6sl() || cpu_is_imx6sx() || cpu_is_imx6ul() ||
> > - cpu_is_imx6ull() || cpu_is_imx6sll())
> > + cpu_is_imx6ull() || cpu_is_imx6sll() || cpu_is_imx6ulz())
> > val |= BM_CLPCR_BYP_MMDC_CH0_LPM_HS;
> > else
> > val |= BM_CLPCR_BYP_MMDC_CH1_LPM_HS;
> > --
> > 2.7.4
> >
