[U-Boot] [PATCH 2/2] MSCC: Remove reset.c
Horatiu Vultur
horatiu.vultur at microchip.com
Wed Jan 16 15:01:59 UTC 2019
Hi Daniel,
The 01/16/2019 15:35, Daniel Schwierzeck wrote:
>
>
> Am 15.01.19 um 17:33 schrieb Horatiu Vultur:
> > Remove reset.c because it is not used anymore. It is superseed by
> > sysreset driver.
> >
> > Signed-off-by: Horatiu Vultur <horatiu.vultur at microchip.com>
> > ---
> > arch/mips/mach-mscc/Makefile | 2 +-
> > arch/mips/mach-mscc/include/mach/ddr.h | 18 ++++++++++++-
> > arch/mips/mach-mscc/reset.c | 47 ----------------------------------
> > 3 files changed, 18 insertions(+), 49 deletions(-)
> > delete mode 100644 arch/mips/mach-mscc/reset.c
> >
> > diff --git a/arch/mips/mach-mscc/Makefile b/arch/mips/mach-mscc/Makefile
> > index f5b6968..349e889 100644
> > --- a/arch/mips/mach-mscc/Makefile
> > +++ b/arch/mips/mach-mscc/Makefile
> > @@ -2,6 +2,6 @@
> >
> > CFLAGS_cpu.o += -finline-limit=64000
> >
> > -obj-y += cpu.o dram.o reset.o phy.o lowlevel_init.o
> > +obj-y += cpu.o dram.o phy.o lowlevel_init.o
> > obj-$(CONFIG_SOC_LUTON) += lowlevel_init_luton.o gpio.o
> > obj-$(CONFIG_SOC_OCELOT) += gpio.o
> > diff --git a/arch/mips/mach-mscc/include/mach/ddr.h b/arch/mips/mach-mscc/include/mach/ddr.h
> > index 7552acb..a2d20db 100644
> > --- a/arch/mips/mach-mscc/include/mach/ddr.h
> > +++ b/arch/mips/mach-mscc/include/mach/ddr.h
> > @@ -415,6 +415,22 @@ static inline void hal_vcoreiii_ddr_reset_release(void)
> > sleep_100ns(10000);
> > }
> >
> > +static void ocelot_restart(void)
> > +{
> > + register u32 resetbits = PERF_SOFT_RST_SOFT_CHIP_RST;
> > + (void)readl(BASE_DEVCPU_GCB + PERF_SOFT_RST);
> > +
> > + /* Make sure VCore is NOT protected from reset */
> > + clrbits_le32(BASE_CFG + ICPU_RESET, ICPU_RESET_CORE_RST_PROTECT);
> > +
> > + /* Change to SPI bitbang for SPI reset workaround... */
> > + writel(ICPU_SW_MODE_SW_SPI_CS_OE(1) | ICPU_SW_MODE_SW_SPI_CS(1) |
> > + ICPU_SW_MODE_SW_PIN_CTRL_MODE, BASE_CFG + ICPU_SW_MODE);
> > +
> > + /* Do the global reset */
> > + writel(resetbits, BASE_DEVCPU_GCB + PERF_SOFT_RST);
> > +}
> > +
> > /*
> > * DDR memory sanity checking failed, tally and do hard reset
> > *
> > @@ -443,7 +459,7 @@ static inline void hal_vcoreiii_ddr_failed(void)
> > * MIPS CPU (and the cache), and the CPU will start executing
> > * from the reset vector.
> > */
> > - reset = KSEG0ADDR(_machine_restart);
> > + reset = KSEG0ADDR(ocelot_restart);
>
> ah I didn't thought about that when suggesting a sysreset driver.Then
> it's maybe not a good idea at all. Or you export a function declaration
> like mscc_cpu_reset(void) and implement it for each MSCC SoC. You could
> call this still in hal_vcoreiii_ddr_failed() context. The sysreset
> driver then simply wraps this function. But then it's not really worth
> the overhead. Sorry for the troubles ;)
No problem, thanks for all the feedback :)
One thingy is that after the patch[1] is accepted, I would like to send
a new patch series for a new SoC. And this new SoC has a different reset
mechanism which means that I need to put another #ifdef inside
_machine_restart.
For me it is perfectly fine not to accept this patch series and after I
send the new patch series for new SoC to have another look at the
function _machine_restart.
[1] https://lists.denx.de/pipermail/u-boot/2019-January/354812.html
>
> > icache_lock((void *)reset, 128);
> > asm volatile ("jr %0"::"r" (reset));
> >
> > diff --git a/arch/mips/mach-mscc/reset.c b/arch/mips/mach-mscc/reset.c
> > deleted file mode 100644
> > index e0e610a..0000000
> > --- a/arch/mips/mach-mscc/reset.c
> > +++ /dev/null
> > @@ -1,47 +0,0 @@
> > -// SPDX-License-Identifier: (GPL-2.0+ OR MIT)
> > -/*
> > - * Copyright (c) 2018 Microsemi Corporation
> > - */
> > -
> > -#include <common.h>
> > -
> > -#include <asm/sections.h>
> > -#include <asm/io.h>
> > -
> > -#include <asm/reboot.h>
> > -
> > -void _machine_restart(void)
> > -{
> > -#if defined(CONFIG_SOC_JR2)
> > - register u32 reg = readl(BASE_CFG + ICPU_GENERAL_CTRL);
> > - /* Set owner */
> > - reg &= ~ICPU_GENERAL_CTRL_IF_SI_OWNER_M;
> > - reg |= ICPU_GENERAL_CTRL_IF_SI_OWNER(1);
> > - /* Set boot mode */
> > - reg |= ICPU_GENERAL_CTRL_BOOT_MODE_ENA;
> > - writel(reg, BASE_CFG + ICPU_GENERAL_CTRL);
> > - /* Read back in order to make BOOT mode setting active */
> > - reg = readl(BASE_CFG + ICPU_GENERAL_CTRL);
> > - /* Reset CPU only - still executing _here_. but from cache */
> > - writel(readl(BASE_CFG + ICPU_RESET) |
> > - ICPU_RESET_CORE_RST_CPU_ONLY |
> > - ICPU_RESET_CORE_RST_FORCE,
> > - BASE_CFG + ICPU_RESET);
> > -#else
> > - register u32 resetbits = PERF_SOFT_RST_SOFT_CHIP_RST;
> > - (void)readl(BASE_DEVCPU_GCB + PERF_SOFT_RST);
> > -
> > - /* Make sure VCore is NOT protected from reset */
> > - clrbits_le32(BASE_CFG + ICPU_RESET, ICPU_RESET_CORE_RST_PROTECT);
> > -
> > - /* Change to SPI bitbang for SPI reset workaround... */
> > - writel(ICPU_SW_MODE_SW_SPI_CS_OE(1) | ICPU_SW_MODE_SW_SPI_CS(1) |
> > - ICPU_SW_MODE_SW_PIN_CTRL_MODE, BASE_CFG + ICPU_SW_MODE);
> > -
> > - /* Do the global reset */
> > - writel(resetbits, BASE_DEVCPU_GCB + PERF_SOFT_RST);
> > -#endif
> > -
> > - while (1)
> > - ; /* NOP */
> > -}
> >
>
> --
> - Daniel
--
/Horatiu
More information about the U-Boot
mailing list