powerpc: Introduce device tree control and driver model

PATCH New Simon Glass 4 participants 33 messages Dec 15, 2014 – Jan 26, 2015 PowerPC

Share & tools

Simon Glass

author

2014-12-15 14:19 UTC #0 lore raw

cover powerpc: Introduce device tree control and driver model

This series does a small amount of tweaking to support device tree control (CONFIG_OF_CONTROL) on PowerPC platforms. It also adds support for driver model. In both cases the main effort is to set things up correctly before calling board_init_f().

A new generic function, board_init_f_mem() is introduced. This does the various memory calculations in C code, since they are messy in assembler and every architecture should in fact be the same. A later series will adjust ARM and x86 to use this function.

As an example, the Canyonlands boards are converted over to use device tree control and driver model for their serial console. It should be fairly straightforward to convert over other boards.

Simon Glass (18):
Introduce board_init_f_mem() to handle early memory layout
powerpc: Permit device tree control of U-Boot (CONFIG_OF_CONTROL)
powerpc: ppc4xx: canyonlands: config: Tidy up CONFIGs and config.mk
powerpc: ppc4xx: Move CANYONLANDS/GLACIER/ARCHES to Kconfig
powerpc: ppc4xx: Add ramboot config for glacier
powerpc: ppc4xx: canyonlands: Move to generic board
powerpc: ppc4xx: dts: Bring in canyonlands device tree files
powerpc: ppc4xx: Call board_init_f_mem() for generic board
powerpc: ppc4xx: Add a gpio.h header file
powerpc: ppc4xx: Allow the end of u-boot.bin to be found
powerpc: ppc4xx: Use CONFIG_OF_CONTROL for canyonlands boards
ppc: amcc: Omit unneeded ns16550 CONFIG if using driver model
powerpc: Add serial driver for driver model
dm: powerpc: ppc4xx: Move glacier to use driver model for serial
powerpc: Add linkage.h file
serial: Support an early UART for debugging
serial: ns16550: Support debug UART
powerpc: ppc4xx: Provide early debug UART defaults

Stefan Roese (1):
WIP: powerpc: ppc4xx: Somehow BSS is not cleared in RAMBOOT case

 arch/Kconfig                                |   1 +
 arch/powerpc/cpu/ppc4xx/44x_spd_ddr2.c      |   8 +
 arch/powerpc/cpu/ppc4xx/config.mk           |   5 +-
 arch/powerpc/cpu/ppc4xx/cpu_init.c          |   2 +
 arch/powerpc/cpu/ppc4xx/start.S             |  18 +-
 arch/powerpc/cpu/ppc4xx/u-boot.lds          |   8 +-
 arch/powerpc/dts/Makefile                   |  11 +
 arch/powerpc/dts/arches.dts                 | 339 ++++++++++++++++
 arch/powerpc/dts/canyonlands.dts            | 555 ++++++++++++++++++++++++++
 arch/powerpc/dts/glacier.dts                | 579 ++++++++++++++++++++++++++++
 arch/powerpc/include/asm/arch-ppc4xx/gpio.h |   7 +
 arch/powerpc/include/asm/linkage.h          |   7 +
 arch/powerpc/include/asm/ppc460ex_gt.h      |   2 +
 arch/powerpc/lib/board.c                    |   3 +
 board/amcc/canyonlands/Kconfig              |  35 ++
 board/amcc/canyonlands/MAINTAINERS          |   1 +
 board/amcc/canyonlands/config.mk            |   2 -
 board/amcc/canyonlands/u-boot-ram.lds       |  85 ++++
 common/board_f.c                            |  18 +
 common/board_r.c                            |   8 +-
 configs/arches_defconfig                    |   5 +-
 configs/canyonlands_defconfig               |   5 +-
 configs/glacier_defconfig                   |   5 +-
 configs/glacier_ramboot_defconfig           |   8 +
 drivers/serial/Kconfig                      |  59 +++
 drivers/serial/Makefile                     |   1 +
 drivers/serial/ns16550.c                    |  42 +-
 drivers/serial/serial_ppc.c                 |  40 ++
 include/configs/amcc-common.h               |   2 +
 include/configs/canyonlands.h               |  38 +-
 include/debug_uart.h                        | 139 +++++++
 31 files changed, 2006 insertions(+), 32 deletions(-)
 create mode 100644 arch/powerpc/dts/Makefile
 create mode 100644 arch/powerpc/dts/arches.dts
 create mode 100644 arch/powerpc/dts/canyonlands.dts
 create mode 100644 arch/powerpc/dts/glacier.dts
 create mode 100644 arch/powerpc/include/asm/arch-ppc4xx/gpio.h
 create mode 100644 arch/powerpc/include/asm/linkage.h
 create mode 100644 board/amcc/canyonlands/u-boot-ram.lds
 create mode 100644 configs/glacier_ramboot_defconfig
 create mode 100644 drivers/serial/serial_ppc.c
 create mode 100644 include/debug_uart.h

Signature · 1 line

2.2.0.rc0.207.ga3a616c

Simon Glass

author

2014-12-15 14:19 UTC #1 lore raw

PATCH 1/19 Introduce board_init_f_mem() to handle early memory layout

At present on some architectures we set up the following before calling board_init_f():

- global_data
- stack
- early malloc memory

Adding the code to support early malloc and global data setup to every arch's assembler start-up is a pain. Also this code is not actually architecture-specific. We can use common code for all architectures and with a bit of care we can write this code in C.

Add a new function to deal with this. It should be called after memory is available, with a pointer to the top of the area that should be used before relocation. The function will set things up and return the lowest memory address that it allocated/used. That can then be set as the top of the stack.

Note that on some archs this function will use the stack, so the stack pointer should be set to same value as is pased to board_init_f_mem(). A margin of 128 bytes will be left for this stack, so that it is not overwritten. This means that 128 bytes is wasted by this early call. This is not strictly necessary on several more modern archs, so we could remove this at the cost of some arch-dependent code.

With this function there is no-longer any need for the assembler code to zero global_data or set up the early malloc pointers.

Signed-off-by Simon Glass <[email protected]>

---

 common/board_f.c | 18 ++++++++++++++++++
 1 file changed, 18 insertions(+)

common/board_f.c +18 -0

index 98c9c72..9652e1d 100644
@@ -1064,4 +1064,22 @@ void board_init_f_r(void)
 	/* NOTREACHED - board_init_r() does not return */
 	hang();
 }
+#else
+ulong board_init_f_mem(ulong top)
+{
+	/* Leave space for the stack we are running with now */
+	top -= 0x80;
+
+	top -= sizeof(struct global_data);
+	top = ALIGN(top, 16);
+	gd = (struct global_data *)top;
+	memset((void *)gd, '\0', sizeof(*gd));
+
+#ifdef CONFIG_SYS_MALLOC_F_LEN
+	top -= CONFIG_SYS_MALLOC_F_LEN;
+	gd->malloc_base = top;
+#endif
+
+	return top;
+}
 #endif /* CONFIG_X86 */

Signature · 1 line

2.2.0.rc0.207.ga3a616c

Simon Glass

author

2014-12-15 14:19 UTC #2 lore raw

PATCH 2/19 powerpc: Permit device tree control of U-Boot (CONFIG_OF_CONTROL)

In reply to Simon Glass

Enable this in the Kconfig so that PowerPC boards can use device tree to configure U-Boot.

Signed-off-by Simon Glass <[email protected]>

---

 arch/Kconfig | 1 +
 1 file changed, 1 insertion(+)

arch/Kconfig +1 -0

index f63cc5a..67e99e9 100644
@@ -39,6 +39,7 @@ config OPENRISC
 config PPC
 	bool "PowerPC architecture"
 	select HAVE_PRIVATE_LIBGCC
+	select SUPPORT_OF_CONTROL
 
 config SANDBOX
 	bool "Sandbox"

Signature · 1 line

2.2.0.rc0.207.ga3a616c

Simon Glass

author

2014-12-15 14:19 UTC #3 lore raw

PATCH 3/19 powerpc: ppc4xx: canyonlands: config: Tidy up CONFIGs and config.mk

In reply to Simon Glass

Many CONFIG options have an unnecessary value of 1. CONFIG_440 is set in the various board config files. Also simplify the CONFIG_440 check in config.mk

Signed-off-by Simon Glass <[email protected]>

---

 arch/powerpc/cpu/ppc4xx/config.mk |  5 +----
 board/amcc/canyonlands/config.mk  |  2 --
 include/configs/canyonlands.h     | 34 +++++++++++++++++-----------------
 3 files changed, 18 insertions(+), 23 deletions(-)

arch/powerpc/cpu/ppc4xx/config.mk +1 -4

index f87c9dc..9cb41bb 100644
@@ -7,10 +7,7 @@
 
 PLATFORM_CPPFLAGS += -mstring -msoft-float
 
-cfg=$(srctree)/include/configs/$(CONFIG_SYS_CONFIG_NAME:"%"=%).h
-is440:=$(shell grep CONFIG_440 $(cfg))
-
-ifneq (,$(findstring CONFIG_440,$(is440)))
+ifneq (,$(CONFIG_440))
 PLATFORM_CPPFLAGS += -Wa,-m440 -mcpu=440
 else
 PLATFORM_CPPFLAGS += -Wa,-m405 -mcpu=405

board/amcc/canyonlands/config.mk +0 -2

index 63b8973..5cc90d2 100644
@@ -8,8 +8,6 @@
 # AMCC 460EX/460GT Evaluation Board (Canyonlands) board
 #
 
-PLATFORM_CPPFLAGS += -DCONFIG_440=1
-
 ifeq ($(debug),1)
 PLATFORM_CPPFLAGS += -DDEBUG
 endif

include/configs/canyonlands.h +17 ~~-17~~

index 8eeb15c..7b1f368 100644
@@ -19,10 +19,10 @@
  * and Arches dual (460GT)
  */
 #ifdef CONFIG_CANYONLANDS
-#define CONFIG_460EX		1	/* Specific PPC460EX		*/
+#define CONFIG_460EX			/* Specific PPC460EX		*/
 #define CONFIG_HOSTNAME		canyonlands
 #else
-#define CONFIG_460GT		1	/* Specific PPC460GT		*/
+#define CONFIG_460GT			/* Specific PPC460GT		*/
 #ifdef CONFIG_GLACIER
 #define CONFIG_HOSTNAME		glacier
 #else

@@ -32,7 +32,7 @@
 #endif
 #endif
 
-#define CONFIG_440		1
+#define CONFIG_440
 
 #ifndef CONFIG_SYS_TEXT_BASE
 #define CONFIG_SYS_TEXT_BASE	0xFFF80000

@@ -45,10 +45,10 @@
 
 #define CONFIG_SYS_CLK_FREQ	66666667	/* external freq to pll	*/
 
-#define CONFIG_BOARD_EARLY_INIT_F	1	/* Call board_early_init_f */
-#define CONFIG_BOARD_EARLY_INIT_R	1	/* Call board_early_init_r */
-#define CONFIG_MISC_INIT_R		1	/* Call misc_init_r */
-#define CONFIG_BOARD_TYPES		1	/* support board types */
+#define CONFIG_BOARD_EARLY_INIT_F		/* Call board_early_init_f */
+#define CONFIG_BOARD_EARLY_INIT_R		/* Call board_early_init_r */
+#define CONFIG_MISC_INIT_R			/* Call misc_init_r */
+#define CONFIG_BOARD_TYPES			/* support board types */
 
 /*-----------------------------------------------------------------------
  * Base addresses -- Note these are effective addresses where the

@@ -134,7 +134,7 @@
  *----------------------------------------------------------------------*/
 #define CONFIG_SYS_FLASH_CFI			/* The flash is CFI compatible	*/
 #define CONFIG_FLASH_CFI_DRIVER		/* Use common CFI driver	*/
-#define CONFIG_SYS_FLASH_CFI_AMD_RESET	1	/* Use AMD (Spansion) reset cmd */
+#define CONFIG_SYS_FLASH_CFI_AMD_RESET	/* Use AMD (Spansion) reset cmd */
 
 #define CONFIG_SYS_FLASH_BANKS_LIST    {CONFIG_SYS_FLASH_BASE}
 #define CONFIG_SYS_MAX_FLASH_BANKS	1	/* max number of memory banks		*/

@@ -172,9 +172,9 @@
  * I2C SPD DIMM autodetection/calibration doesn't fit into the 4k of boot
  * code.
  */
-#define CONFIG_SPD_EEPROM	1	/* Use SPD EEPROM for setup	*/
+#define CONFIG_SPD_EEPROM		/* Use SPD EEPROM for setup	*/
 #define SPD_EEPROM_ADDRESS	{0x50, 0x51}	/* SPD i2c spd addresses*/
-#define CONFIG_DDR_ECC		1	/* with ECC support		*/
+#define CONFIG_DDR_ECC			/* with ECC support		*/
 #define CONFIG_DDR_RQDC_FIXED	0x80000038 /* fixed value for RQDC	*/
 
 #else /* defined(CONFIG_ARCHES) */

@@ -262,8 +262,8 @@
 #define CONFIG_4xx_CONFIG_BLOCKSIZE		16
 
 /* I2C SYSMON (LM75, AD7414 is almost compatible)			*/
-#define CONFIG_DTT_LM75		1		/* ON Semi's LM75	*/
-#define CONFIG_DTT_AD7414	1		/* use AD7414		*/
+#define CONFIG_DTT_LM75				/* ON Semi's LM75	*/
+#define CONFIG_DTT_AD7414			/* use AD7414		*/
 #define CONFIG_DTT_SENSORS	{0}		/* Sensor addresses	*/
 #define CONFIG_SYS_DTT_MAX_TEMP	70
 #define CONFIG_SYS_DTT_LOW_TEMP	-30

@@ -275,14 +275,14 @@
 
 #if !defined(CONFIG_ARCHES)
 /* RTC configuration */
-#define CONFIG_RTC_M41T62	1
+#define CONFIG_RTC_M41T62
 #define CONFIG_SYS_I2C_RTC_ADDR	0x68
 #endif
 
 /*-----------------------------------------------------------------------
  * Ethernet
  *----------------------------------------------------------------------*/
-#define CONFIG_IBM_EMAC4_V4	1
+#define CONFIG_IBM_EMAC4_V4
 
 #define CONFIG_HAS_ETH0
 #define CONFIG_HAS_ETH1

@@ -322,9 +322,9 @@
 #define CONFIG_GPCS_PHY2_ADDR   0xC
 #endif	/* !defined(CONFIG_ARCHES) */
 
-#define CONFIG_PHY_RESET	1	/* reset phy upon startup	*/
-#define CONFIG_PHY_GIGE		1	/* Include GbE speed/duplex detection */
-#define CONFIG_PHY_DYNAMIC_ANEG	1
+#define CONFIG_PHY_RESET		/* reset phy upon startup	*/
+#define CONFIG_PHY_GIGE			/* Include GbE speed/duplex detection */
+#define CONFIG_PHY_DYNAMIC_ANEG
 
 /*-----------------------------------------------------------------------
  * USB-OHCI

Signature · 1 line

2.2.0.rc0.207.ga3a616c

Simon Glass

author

2014-12-15 14:19 UTC #4 lore raw

PATCH 4/19 powerpc: ppc4xx: Move CANYONLANDS/GLACIER/ARCHES to Kconfig

In reply to Simon Glass

Move these options to Kconfig and remove them from the CONFIG files.

Signed-off-by Simon Glass <[email protected]>

---

 board/amcc/canyonlands/Kconfig | 20 ++++++++++++++++++++
 configs/arches_defconfig       |  2 +-
 configs/canyonlands_defconfig  |  2 +-
 configs/glacier_defconfig      |  2 +-
 include/configs/canyonlands.h  |  2 ++
 5 files changed, 25 insertions(+), 3 deletions(-)

board/amcc/canyonlands/Kconfig +20 -0

index 530a6ef..0fc6877 100644
@@ -9,4 +9,24 @@ config SYS_VENDOR
 config SYS_CONFIG_NAME
 	default "canyonlands"
 
+choice BOARD_TYPE
+	prompt "Select which board to build for"
+
+config CANYONLANDS
+	bool "Glacier"
+	help
+	  Select this to build for the Canyonlands 460EX board.
+
+config GLACIER
+	bool "Glacier"
+	help
+	  Select this to build for the Glacier 460GT board.
+
+config ARCHES
+	bool "Arches"
+	help
+	  Select this to build for the Arches dual 460GT board.
+
+endchoice
+
 endif

configs/arches_defconfig +1 -1

index 18d0a14..60e6ef9 100644
@@ -1,4 +1,4 @@
-CONFIG_SYS_EXTRA_OPTIONS="ARCHES"
 CONFIG_PPC=y
 CONFIG_4xx=y
 CONFIG_TARGET_CANYONLANDS=y
+CONFIG_ARCHES=y

configs/canyonlands_defconfig +1 -1

index 09b9ab9..37a2b4d 100644
@@ -1,4 +1,4 @@
-CONFIG_SYS_EXTRA_OPTIONS="CANYONLANDS"
 CONFIG_PPC=y
 CONFIG_4xx=y
 CONFIG_TARGET_CANYONLANDS=y
+CONFIG_CANYONLANDS=y

configs/glacier_defconfig +1 -1

index 2a66bfb..436b9f8 100644
@@ -1,4 +1,4 @@
-CONFIG_SYS_EXTRA_OPTIONS="GLACIER"
 CONFIG_PPC=y
 CONFIG_4xx=y
 CONFIG_TARGET_CANYONLANDS=y
+CONFIG_GLACIER=y

include/configs/canyonlands.h +2 -0

index 7b1f368..ed790cc 100644
@@ -11,6 +11,8 @@
 #ifndef __CONFIG_H
 #define __CONFIG_H
 
+#include <linux/kconfig.h>
+
 /*-----------------------------------------------------------------------
  * High Level Configuration Options
  *----------------------------------------------------------------------*/

Signature · 1 line

2.2.0.rc0.207.ga3a616c

Masahiro Yamada

2015-01-06 07:29 UTC #5 lore raw

PATCH 4/19 powerpc: ppc4xx: Move CANYONLANDS/GLACIER/ARCHES to Kconfig

Hi Simon,

include/configs/canyonlands.h line 15

Quoted 10 lines · On Mon, 15 Dec 2014 07:19:39 -0700 Simon Glass <[email protected]> wrote:

include/configs/canyonlands.h +2 -0

index 7b1f368..ed790cc 100644
@@ -11,6 +11,8 @@
 #ifndef __CONFIG_H
 #define __CONFIG_H
 
+#include <linux/kconfig.h>
+

Please do not explicitely include <linux/kconfig.h> from C sources. It is automatically included.

See Makefile

UBOOTINCLUDE    := \
                -Iinclude \
                $(if $(KBUILD_SRC), -I$(srctree)/include) \
                -I$(srctree)/arch/$(ARCH)/include \
                -include $(srctree)/include/linux/kconfig.h

Best Regards
Masahiro Yamada

Simon Glass

author

2014-12-15 14:19 UTC #6 lore raw

PATCH 5/19 powerpc: ppc4xx: Add ramboot config for glacier

In reply to Simon Glass

Add a new ramboot config for glacier so that it is possible to test U-Boot loaded over Ethernet instead of using JTAG.

Signed-off-by Simon Glass <[email protected]>

---

 arch/powerpc/cpu/ppc4xx/44x_spd_ddr2.c |  8 ++++
 board/amcc/canyonlands/MAINTAINERS     |  1 +
 board/amcc/canyonlands/u-boot-ram.lds  | 79 ++++++++++++++++++++++++++++++++++
 configs/glacier_ramboot_defconfig      |  5 +++
 4 files changed, 93 insertions(+)
 create mode 100644 board/amcc/canyonlands/u-boot-ram.lds
 create mode 100644 configs/glacier_ramboot_defconfig

arch/powerpc/cpu/ppc4xx/44x_spd_ddr2.c +8 -0

index 71bb9d7..7202c3f 100644
@@ -424,6 +424,14 @@ phys_size_t initdram(int board_type)
 	int write_recovery;
 	phys_size_t dram_size = 0;
 
+	if (IS_ENABLED(CONFIG_SYS_RAMBOOT)) {
+		/*
+		 * Reduce RAM size to avoid overwriting memory used by
+		 * current stack? Not sure what is happening.
+		 */
+		return sdram_memsize() / 2;
+	}
+
 	num_dimm_banks = sizeof(iic0_dimm_addr);
 
 	/*------------------------------------------------------------------

board/amcc/canyonlands/MAINTAINERS +1 -0

index 52bf004..8be8a52 100644
@@ -6,3 +6,4 @@ F:	include/configs/canyonlands.h
 F:	configs/arches_defconfig
 F:	configs/canyonlands_defconfig
 F:	configs/glacier_defconfig
+F:	configs/glacier_ramboot_defconfig

board/amcc/canyonlands/u-boot-ram.lds new file +79 -0

79 lines not on this page · raw message

configs/glacier_ramboot_defconfig new file +5 -0

new file mode 100644
index 0000000..4fc2303
@@ -0,0 +1,5 @@
+CONFIG_SYS_EXTRA_OPTIONS="SYS_RAMBOOT,SYS_TEXT_BASE=0x01000000,SYS_LDSCRIPT=board/amcc/canyonlands/u-boot-ram.lds"
+CONFIG_PPC=y
+CONFIG_4xx=y
+CONFIG_TARGET_CANYONLANDS=y
+CONFIG_GLACIER=y

Signature · 1 line

2.2.0.rc0.207.ga3a616c

Simon Glass

author

2014-12-15 14:19 UTC #7 lore raw

PATCH 6/19 WIP: powerpc: ppc4xx: Somehow BSS is not cleared in RAMBOOT case

In reply to Simon Glass

From: Stefan Roese <[email protected]>

(NOT TO APPLY)

I'm really not sure why this doesn't work in the RAMBOOT case. But BSS is not cleared and because of this booting crashed / hangs at some stage later. Something with the GOT calculation / handling is incorrect.

To get this going for now, just clear the BSS again in the C code.

Signed-off-by Stefan Roese <[email protected]>

Signed-off-by Simon Glass <[email protected]>

---

 arch/powerpc/lib/board.c | 3 +++
 common/board_r.c         | 8 +++++++-
 2 files changed, 10 insertions(+), 1 deletion(-)

arch/powerpc/lib/board.c +3 -0

index e6d5355..2455e45 100644
@@ -108,6 +108,7 @@ DECLARE_GLOBAL_DATA_PTR;
 extern ulong __init_end;
 extern ulong __bss_end;
 ulong monitor_flash_len;
+extern ulong __bss_start;
 
 #if defined(CONFIG_CMD_BEDBUG)
 #include <bedbug/type.h>

@@ -611,6 +612,8 @@ void board_init_r(gd_t *id, ulong dest_addr)
 	/* The Malloc area is immediately below the monitor copy in DRAM */
 	malloc_start = dest_addr - TOTAL_MALLOC_LEN;
 
+	memset(&__bss_start, 0, &__bss_end - &__bss_start);
+
 #if defined(CONFIG_MPC85xx) || defined(CONFIG_MPC86xx)
 	/*
 	 * The gd->arch.cpu pointer is set to an address in flash before

common/board_r.c +7 -1

index a301cc2..2f80425 100644
@@ -1,4 +1,4 @@
-/*
+	/*
  * Copyright (c) 2011 The Chromium OS Authors.
  * (C) Copyright 2002-2006
  * Wolfgang Denk, DENX Software Engineering, wd at denx.de.

@@ -886,6 +886,12 @@ void board_init_r(gd_t *new_gd, ulong dest_addr)
 	gd = new_gd;
 #endif
 
+#ifdef CONFIG_GLACIER
+	ulong bss_start = (ulong)&__bss_start;
+	ulong bss_end = (ulong)&__bss_end;
+	memset((void *)bss_start, 0, bss_end - bss_start);
+#endif
+
 #ifdef CONFIG_NEEDS_MANUAL_RELOC
 	for (i = 0; i < ARRAY_SIZE(init_sequence_r); i++)
 		init_sequence_r[i] += gd->reloc_off;

Signature · 1 line

2.2.0.rc0.207.ga3a616c

Simon Glass

author

2014-12-15 14:19 UTC #8 lore raw

PATCH 7/19 powerpc: ppc4xx: canyonlands: Move to generic board

In reply to Simon Glass

Switch to generic board so that this board will not be broken/removed.

Signed-off-by Simon Glass <[email protected]>

---

 arch/powerpc/cpu/ppc4xx/cpu_init.c | 2 ++
 include/configs/canyonlands.h      | 2 ++
 2 files changed, 4 insertions(+)

arch/powerpc/cpu/ppc4xx/cpu_init.c +2 -0

index e5a0e21..5f5c720 100644
@@ -450,10 +450,12 @@ cpu_init_f (void)
 	      PLB4Ax_ACR_RDP_4DEEP);
 #endif /* CONFIG_440SP/SPE || CONFIG_460EX/GT || CONFIG_405EX */
 
+#ifndef CONFIG_SYS_GENERIC_BOARD
 	gd = (gd_t *)(CONFIG_SYS_INIT_RAM_ADDR + CONFIG_SYS_GBL_DATA_OFFSET);
 
 	/* Clear initial global data */
 	memset((void *)gd, 0, sizeof(gd_t));
+#endif
 }
 
 /*

include/configs/canyonlands.h +2 -0

index ed790cc..7a1499d 100644
@@ -13,6 +13,8 @@
 
 #include <linux/kconfig.h>
 
+#define CONFIG_SYS_GENERIC_BOARD
+
 /*-----------------------------------------------------------------------
  * High Level Configuration Options
  *----------------------------------------------------------------------*/

Signature · 1 line

2.2.0.rc0.207.ga3a616c

Simon Glass

author

2014-12-15 14:19 UTC #9 lore raw

PATCH 8/19 powerpc: ppc4xx: dts: Bring in canyonlands device tree files

In reply to Simon Glass

The canyonlands.h config file works with canyonlands, glacier and arches boards. Bring in the device tree files for these from Linux 3.17.

Signed-off-by Simon Glass <[email protected]>

---

 arch/powerpc/dts/Makefile        |  11 +
 arch/powerpc/dts/arches.dts      | 339 +++++++++++++++++++++++
 arch/powerpc/dts/canyonlands.dts | 555 +++++++++++++++++++++++++++++++++++++
 arch/powerpc/dts/glacier.dts     | 579 +++++++++++++++++++++++++++++++++++++++
 4 files changed, 1484 insertions(+)
 create mode 100644 arch/powerpc/dts/Makefile
 create mode 100644 arch/powerpc/dts/arches.dts
 create mode 100644 arch/powerpc/dts/canyonlands.dts
 create mode 100644 arch/powerpc/dts/glacier.dts

arch/powerpc/dts/Makefile new file +11 -0

new file mode 100644
index 0000000..ad104b9
@@ -0,0 +1,11 @@
+dtb-$(CONFIG_TARGET_CANYONLANDS) += arches.dtb canyonlands.dtb glacier.dtb
+
+targets += $(dtb-y)
+
+DTC_FLAGS += -R 4 -p 0x1000
+
+PHONY += dtbs
+dtbs: $(addprefix $(obj)/, $(dtb-y))
+	@:
+
+clean-files := *.dtb

arch/powerpc/dts/arches.dts new file +339 -0

339 lines not on this page · raw message

arch/powerpc/dts/canyonlands.dts new file +555 -0

555 lines not on this page · raw message

arch/powerpc/dts/glacier.dts new file +579 -0

579 lines not on this page · raw message

Signature · 1 line

2.2.0.rc0.207.ga3a616c

Simon Glass

author

2014-12-15 14:19 UTC #10 lore raw

PATCH 9/19 powerpc: ppc4xx: Call board_init_f_mem() for generic board

In reply to Simon Glass

Call this function to set up our early memory.

Signed-off-by Simon Glass <[email protected]>

---

 arch/powerpc/cpu/ppc4xx/start.S | 18 +++++++++++++++++-
 1 file changed, 17 insertions(+), 1 deletion(-)

arch/powerpc/cpu/ppc4xx/start.S +17 -1

index 09a02d7..7a0f0d2 100644
@@ -760,6 +760,15 @@ _start:
 #endif
 
 	bl	cpu_init_f	/* run low-level CPU init code	   (from Flash) */
+#ifdef CONFIG_SYS_GENERIC_BOARD
+	mr	r3, r1
+	bl	board_init_f_mem
+	mr	r1, r3
+	li	r0,0
+	stwu	r0, -4(r1)
+	stwu	r0, -4(r1)
+#endif
+	li	r3, 0
 	bl	board_init_f
 	/* NOTREACHED - board_init_f() does not return */

@@ -1027,7 +1036,14 @@ _start:
 	GET_GOT			/* initialize GOT access			*/
 
 	bl	cpu_init_f	/* run low-level CPU init code	   (from Flash) */
-
+#ifdef CONFIG_SYS_GENERIC_BOARD
+	mr	r3, r1
+	bl	board_init_f_mem
+	mr	r1, r3
+	stwu	r0, -4(r1)
+	stwu	r0, -4(r1)
+#endif
+	li	r3, 0
 	bl	board_init_f	/* run first part of init code (from Flash)	*/
 	/* NOTREACHED - board_init_f() does not return */

Signature · 1 line

2.2.0.rc0.207.ga3a616c

Simon Glass

author

2014-12-15 14:19 UTC #11 lore raw

PATCH 10/19 powerpc: ppc4xx: Add a gpio.h header file

In reply to Simon Glass

This is required at present for device tree control. The ppc4xx does support GPIOs but does not seem to have a proper driver. So this file is empty.

Signed-off-by Simon Glass <[email protected]>

---

 arch/powerpc/include/asm/arch-ppc4xx/gpio.h | 7 +++++++
 1 file changed, 7 insertions(+)
 create mode 100644 arch/powerpc/include/asm/arch-ppc4xx/gpio.h

arch/powerpc/include/asm/arch-ppc4xx/gpio.h new file +7 -0

new file mode 100644
index 0000000..3d960c3
@@ -0,0 +1,7 @@
+/*
+ * (C) Copyright 2014 Google, Inc
+ *
+ * SPDX-License-Identifier:	GPL-2.0+
+ */
+
+/* This is empty for now as we don't support the generic GPIO interface */

Signature · 1 line

2.2.0.rc0.207.ga3a616c

Simon Glass

author

2014-12-15 14:19 UTC #12 lore raw

PATCH 11/19 powerpc: ppc4xx: Allow the end of u-boot.bin to be found

In reply to Simon Glass

Define an _end symbol indicating the end of u-boot.bin. Also add some dummy words into the link script to ensure that u-boot.bin will always extend that far. There may be a better way of doing this.

Signed-off-by Simon Glass <[email protected]>

---

 arch/powerpc/cpu/ppc4xx/u-boot.lds | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

arch/powerpc/cpu/ppc4xx/u-boot.lds +6 -2

index 8773178..1980508 100644
@@ -76,9 +76,13 @@ SECTIONS
   . = ALIGN(256);
   __init_begin = .;
   .text.init : { *(.text.init) }
-  .data.init : { *(.data.init) }
-  . = ALIGN(256);
+  .data.init : {
+	*(.data.init)
+	. = ALIGN(256);
+	LONG(0) LONG(0)		/* Extend u-boot.bin to here */
+  }
   __init_end = .;
+  _end = .;
 
 #ifndef CONFIG_SPL
 #ifdef CONFIG_440

Signature · 1 line

2.2.0.rc0.207.ga3a616c

Simon Glass

author

2014-12-15 14:19 UTC #13 lore raw

PATCH 12/19 powerpc: ppc4xx: Use CONFIG_OF_CONTROL for canyonlands boards

In reply to Simon Glass

Enable CONFIG_OF_CONTROL so that U-Boot on these three boards uses a device tree for its configuration.

Signed-off-by Simon Glass <[email protected]>

---

 board/amcc/canyonlands/u-boot-ram.lds | 10 ++++++++--
 configs/arches_defconfig              |  3 +++
 configs/canyonlands_defconfig         |  3 +++
 configs/glacier_defconfig             |  3 +++
 configs/glacier_ramboot_defconfig     |  3 +++
 5 files changed, 20 insertions(+), 2 deletions(-)

board/amcc/canyonlands/u-boot-ram.lds +8 -2

index 6765256..1750c74 100644
@@ -12,6 +12,7 @@ SECTIONS
   . = + SIZEOF_HEADERS;
   .text      :
   {
+    _image_copy_start = .;
     arch/powerpc/cpu/ppc4xx/start.o	(.text*)
     board/amcc/canyonlands/init.o	(.text*)

@@ -61,9 +62,14 @@ SECTIONS
   . = ALIGN(256);
   __init_begin = .;
   .text.init : { *(.text.init) }
-  .data.init : { *(.data.init) }
-  . = ALIGN(256);
+  .data.init : {
+	*(.data.init)
+	. = ALIGN(256);
+	LONG(0) LONG(0)		/* Extend u-boot.bin to here */
+  }
   __init_end = .;
+  _end = .;
+  _image_binary_end = .;
 
   __bss_start = .;
   .bss (NOLOAD)       :

configs/arches_defconfig +3 -0

index 60e6ef9..30c6932 100644
@@ -2,3 +2,6 @@ CONFIG_PPC=y
 CONFIG_4xx=y
 CONFIG_TARGET_CANYONLANDS=y
 CONFIG_ARCHES=y
+CONFIG_DEFAULT_DEVICE_TREE="arches"
+CONFIG_OF_CONTROL=y
+CONFIG_OF_SEPARATE=y

configs/canyonlands_defconfig +3 -0

index 37a2b4d..6f6cf14 100644
@@ -2,3 +2,6 @@ CONFIG_PPC=y
 CONFIG_4xx=y
 CONFIG_TARGET_CANYONLANDS=y
 CONFIG_CANYONLANDS=y
+CONFIG_DEFAULT_DEVICE_TREE="canyonlands"
+CONFIG_OF_CONTROL=y
+CONFIG_OF_SEPARATE=y

configs/glacier_defconfig +3 -0

index 436b9f8..e67fa32 100644
@@ -2,3 +2,6 @@ CONFIG_PPC=y
 CONFIG_4xx=y
 CONFIG_TARGET_CANYONLANDS=y
 CONFIG_GLACIER=y
+CONFIG_DEFAULT_DEVICE_TREE="glacier"
+CONFIG_OF_CONTROL=y
+CONFIG_OF_SEPARATE=y

configs/glacier_ramboot_defconfig +3 -0

index 4fc2303..e5b402e 100644
@@ -3,3 +3,6 @@ CONFIG_PPC=y
 CONFIG_4xx=y
 CONFIG_TARGET_CANYONLANDS=y
 CONFIG_GLACIER=y
+CONFIG_DEFAULT_DEVICE_TREE="glacier"
+CONFIG_OF_CONTROL=y
+CONFIG_OF_SEPARATE=y

Signature · 1 line

2.2.0.rc0.207.ga3a616c

Stefan Roese

2015-01-26 13:17 UTC #14 lore raw

PATCH 12/19 powerpc: ppc4xx: Use CONFIG_OF_CONTROL for canyonlands boards

Quoted 4 lines · Simon Glass

Enable CONFIG_OF_CONTROL so that U-Boot on these three boards uses a device tree for its configuration.
Signed-off-by Simon Glass <[email protected]>

This need further work on the ppc4xx platforms to get it booting from flash. As now the dtb is appended to the "normal" U-Boot image. And this image is fine-tuned to fit into currently 512KiB. With the last word (reset vector of the ppc4xx) located at 0xffff.fffc. So we need to include this dtb into the image this way:

0xfff8.0000 U-Boot code / text etc
...
dtb
0xffff.fffc reset vector

Looking for the current options shows that perhaps CONFIG_OF_EMBED seems to be exactly what we are need. I'll got it working on Canyonlands after a bit of debugging. I'll post the necessary small patches soon...

Thanks,
Stefan

Simon Glass

author

2015-01-26 13:31 UTC #15 lore raw

PATCH 12/19 powerpc: ppc4xx: Use CONFIG_OF_CONTROL for canyonlands boards

Hi Stefan,

Quoted 15 lines · Stefan Roese

Quoted 6 lines · Simon Glass
Enable CONFIG_OF_CONTROL so that U-Boot on these three boards uses a device
tree for its configuration.
Signed-off-by Simon Glass <[email protected]>
This need further work on the ppc4xx platforms to get it booting from flash. As now the dtb is appended to the "normal" U-Boot image. And this image is fine-tuned to fit into currently 512KiB. With the last word (reset vector of the ppc4xx) located at 0xffff.fffc. So we need to include this dtb into the image this way:

Quoted 8 lines

0xfff8.0000 U-Boot code / text etc
...
dtb
0xffff.fffc reset vector
Looking for the current options shows that perhaps CONFIG_OF_EMBED seems to be exactly what we are need. I'll got it working on Canyonlands after a bit of debugging. I'll post the necessary small patches soon...

Sounds good, thanks!

It almost feels like you should have a mkimage image for this, to create a 512KB blob from u-boot.bin and u-boot.dtb. There must be linker magic to make this all hold together. But yes the embed option will build it in, so avoiding that.

Regards,
Simon

Bin Meng

2015-01-26 13:39 UTC #16 lore raw

PATCH 12/19 powerpc: ppc4xx: Use CONFIG_OF_CONTROL for canyonlands boards

Hi Simon,

Quoted 23 lines · Simon Glass

Hi Stefan,
Quoted 20 lines · Stefan Roese
Quoted 6 lines · Simon Glass
Enable CONFIG_OF_CONTROL so that U-Boot on these three boards uses a device
tree for its configuration.
Signed-off-by Simon Glass <[email protected]>
This need further work on the ppc4xx platforms to get it booting from flash. As now the dtb is appended to the "normal" U-Boot image. And this image is fine-tuned to fit into currently 512KiB. With the last word (reset vector of the ppc4xx) located at 0xffff.fffc. So we need to include this dtb into the image this way:
0xfff8.0000 U-Boot code / text etc
...
dtb
0xffff.fffc reset vector

Quoted 6 lines

Quoted 3 lines
Looking for the current options shows that perhaps CONFIG_OF_EMBED seems to be exactly what we are need. I'll got it working on Canyonlands after a bit of debugging. I'll post the necessary small patches soon...
Sounds good, thanks!

We need do the same for Freescale BookE powerpc series too.

Quoted 5 lines

It almost feels like you should have a mkimage image for this, to create a 512KB blob from u-boot.bin and u-boot.dtb. There must be linker magic to make this all hold together. But yes the embed option will build it in, so avoiding that.

Or maybe we can extend x86 solution (like ifdtool) to non-x86 architectures like BookE powerpc?

Regards,
Bin

Simon Glass

author

2014-12-15 14:19 UTC #17 lore raw

PATCH 13/19 ppc: amcc: Omit unneeded ns16550 CONFIG if using driver model

In reply to Simon Glass

This comes from the device tree or a call to get_uart_clock().

Signed-off-by Simon Glass <[email protected]>

---

 arch/powerpc/include/asm/ppc460ex_gt.h | 2 ++
 include/configs/amcc-common.h          | 2 ++
 2 files changed, 4 insertions(+)

arch/powerpc/include/asm/ppc460ex_gt.h +2 -0

index f41df0d..ea019aa 100644
@@ -19,10 +19,12 @@
 /* Memory mapped registers */
 #define CONFIG_SYS_PERIPHERAL_BASE	0xef600000 /* Internal Peripherals */
 
+#ifndef CONFIG_DM_SERIAL
 #define CONFIG_SYS_NS16550_COM1	(CONFIG_SYS_PERIPHERAL_BASE + 0x0300)
 #define CONFIG_SYS_NS16550_COM2	(CONFIG_SYS_PERIPHERAL_BASE + 0x0400)
 #define CONFIG_SYS_NS16550_COM3	(CONFIG_SYS_PERIPHERAL_BASE + 0x0500)
 #define CONFIG_SYS_NS16550_COM4	(CONFIG_SYS_PERIPHERAL_BASE + 0x0600)
+#endif
 
 #define GPIO0_BASE		(CONFIG_SYS_PERIPHERAL_BASE + 0x0b00)
 #define GPIO1_BASE		(CONFIG_SYS_PERIPHERAL_BASE + 0x0c00)

include/configs/amcc-common.h +2 -0

index 2aea899..73e1b0a 100644
@@ -20,8 +20,10 @@
  */
 #define CONFIG_SYS_NS16550
 #define CONFIG_SYS_NS16550_SERIAL
+#ifndef CONFIG_DM_SERIAL
 #define CONFIG_SYS_NS16550_REG_SIZE	1
 #define CONFIG_SYS_NS16550_CLK		get_serial_clock()
+#endif
 #define CONFIG_BAUDRATE		115200
 #define CONFIG_SYS_BAUDRATE_TABLE  \
     {300, 600, 1200, 2400, 4800, 9600, 19200, 38400, 57600, 115200, 230400}

Signature · 1 line

2.2.0.rc0.207.ga3a616c

Simon Glass

author

2014-12-15 14:19 UTC #18 lore raw

PATCH 14/19 powerpc: Add serial driver for driver model

In reply to Simon Glass

This uses the ns16550 driver but sets up the clock at run-time. It does not seem to be available in the device tree.

Signed-off-by Simon Glass <[email protected]>

---

 drivers/serial/Makefile     |  1 +
 drivers/serial/serial_ppc.c | 40 ++++++++++++++++++++++++++++++++++++++++
 2 files changed, 41 insertions(+)
 create mode 100644 drivers/serial/serial_ppc.c

drivers/serial/Makefile +1 -0

index 8c84942..dd3b1cb 100644
@@ -8,6 +8,7 @@
 ifdef CONFIG_DM_SERIAL
 obj-y += serial-uclass.o
 obj-$(CONFIG_PL01X_SERIAL) += serial_pl01x.o
+obj-$(CONFIG_PPC) += serial_ppc.o
 else
 obj-y += serial.o
 obj-$(CONFIG_PL010_SERIAL) += serial_pl01x.o

drivers/serial/serial_ppc.c new file +40 -0

new file mode 100644
index 0000000..47141c6
@@ -0,0 +1,40 @@
+/*
+ * Copyright (c) 2014 Google, Inc
+ *
+ * SPDX-License-Identifier:	GPL-2.0+
+ */
+
+#include <common.h>
+#include <dm.h>
+#include <ns16550.h>
+#include <serial.h>
+
+static const struct udevice_id ppc_serial_ids[] = {
+	{ .compatible = "ns16550" },
+	{ }
+};
+
+static int ppc_serial_ofdata_to_platdata(struct udevice *dev)
+{
+	struct ns16550_platdata *plat = dev_get_platdata(dev);
+	int ret;
+
+	ret = ns16550_serial_ofdata_to_platdata(dev);
+	if (ret)
+		return ret;
+	plat->clock = get_serial_clock();
+
+	return 0;
+}
+
+U_BOOT_DRIVER(serial_ns16550) = {
+	.name	= "serial_ppc",
+	.id	= UCLASS_SERIAL,
+	.of_match = ppc_serial_ids,
+	.ofdata_to_platdata = ppc_serial_ofdata_to_platdata,
+	.platdata_auto_alloc_size = sizeof(struct ns16550_platdata),
+	.priv_auto_alloc_size = sizeof(struct NS16550),
+	.probe = ns16550_serial_probe,
+	.ops	= &ns16550_serial_ops,
+	.flags	= DM_FLAG_PRE_RELOC,
+};

Signature · 1 line

2.2.0.rc0.207.ga3a616c

Simon Glass

author

2014-12-15 14:19 UTC #19 lore raw

PATCH 15/19 dm: powerpc: ppc4xx: Move glacier to use driver model for serial

In reply to Simon Glass

Adjust Kconfig to default to driver model for glacier, canyonlands and arches.

Signed-off-by Simon Glass <[email protected]>

---

 board/amcc/canyonlands/Kconfig | 9 +++++++++
 1 file changed, 9 insertions(+)

board/amcc/canyonlands/Kconfig +9 -0

index 0fc6877..cbc5ff9 100644
@@ -29,4 +29,13 @@ config ARCHES
 
 endchoice
 
+config DM
+	default y
+
+config DM_SERIAL
+	default y
+
+config SYS_MALLOC_F_LEN
+	default 0x400
+
 endif

Signature · 1 line

2.2.0.rc0.207.ga3a616c

Stefan Roese

2015-01-26 09:23 UTC #20 lore raw

PATCH 15/19 dm: powerpc: ppc4xx: Move glacier to use driver model for serial

Hi Simon,

finally I'm testing your patchset. Again sorry for the big delay here.

Quoted 16 lines · Simon Glass

Adjust Kconfig to default to driver model for glacier, canyonlands and arches.
Signed-off-by Simon Glass <[email protected]>
---
  board/amcc/canyonlands/Kconfig | 9 +++++++++
  1 file changed, 9 insertions(+)
board/amcc/canyonlands/Kconfig +0 -0
index 0fc6877..cbc5ff9 100644
@@ -29,4 +29,13 @@ config ARCHES
 
  endchoice

board/amcc/canyonlands/Kconfig config ARCHES · line 40

Quoted 10 lines

board/amcc/canyonlands/Kconfig +9 -0

 
+config DM
+	default y
+
+config DM_SERIAL
+	default y
+
+config SYS_MALLOC_F_LEN
+	default 0x400
+

Hmmm. This does not seem to work. With your patch series applied on current top-of-tree I get this error (for glacier_ramboot or canyonlands):

$ make -s -j10
board/amcc/canyonlands/Kconfig:38:warning: config symbol defined without type
drivers/serial/serial-uclass.c:29:2: error: #error "Serial is required before relocation - define CONFIG_SYS_MALLOC_F_LEN to make this work"
make[1]: *** [drivers/serial/serial-uclass.o] Error 1

Somehow the CONFIG_SYS_MALLOC_F_LEN define is not available in the .config used. Do you have any idea why it might be missing?

Thanks,
Stefan

Stefan Roese

2015-01-26 09:35 UTC #21 lore raw

PATCH 15/19 dm: powerpc: ppc4xx: Move glacier to use driver model for serial

Quoted 22 lines · Stefan Roese

Quoted 7 lines
board/amcc/canyonlands/Kconfig +0 -0
index 0fc6877..cbc5ff9 100644
@@ -29,4 +29,13 @@ config ARCHES
 
   endchoice
board/amcc/canyonlands/Kconfig config ARCHES · line 40
Quoted 10 lines
board/amcc/canyonlands/Kconfig +9 -0
 
+config DM
+	default y
+
+config DM_SERIAL
+	default y
+
+config SYS_MALLOC_F_LEN
+	default 0x400
+
Hmmm. This does not seem to work. With your patch series applied on current top-of-tree I get this error (for glacier_ramboot or canyonlands):

Quoted 7 lines

$ make -s -j10
board/amcc/canyonlands/Kconfig:38:warning: config symbol defined without type
drivers/serial/serial-uclass.c:29:2: error: #error "Serial is required before relocation - define CONFIG_SYS_MALLOC_F_LEN to make this work"
make[1]: *** [drivers/serial/serial-uclass.o] Error 1
Somehow the CONFIG_SYS_MALLOC_F_LEN define is not available in the .config used. Do you have any idea why it might be missing?

Okay. Found it. Its the missing type here. I'll send some follow-up patches once I have all tested on canyonlands here.

Thanks,
Stefan

Simon Glass

author

2015-01-26 13:33 UTC #22 lore raw

PATCH 15/19 dm: powerpc: ppc4xx: Move glacier to use driver model for serial

Hi Stefan,

Quoted 33 lines · Stefan Roese

Quoted 32 lines · Stefan Roese
Quoted 9 lines
board/amcc/canyonlands/Kconfig +0 -0
index 0fc6877..cbc5ff9 100644
@@ -29,4 +29,13 @@ config ARCHES
 
   endchoice
board/amcc/canyonlands/Kconfig config ARCHES · line 40
Quoted 10 lines
board/amcc/canyonlands/Kconfig +9 -0
 
+config DM
+       default y
+
+config DM_SERIAL
+       default y
+
+config SYS_MALLOC_F_LEN
+       default 0x400
+
Hmmm. This does not seem to work. With your patch series applied on current top-of-tree I get this error (for glacier_ramboot or canyonlands):
$ make -s -j10
board/amcc/canyonlands/Kconfig:38:warning: config symbol defined without
type
drivers/serial/serial-uclass.c:29:2: error: #error "Serial is required
before relocation - define CONFIG_SYS_MALLOC_F_LEN to make this work"
make[1]: *** [drivers/serial/serial-uclass.o] Error 1

Quoted 6 lines

Quoted 2 lines
Somehow the CONFIG_SYS_MALLOC_F_LEN define is not available in the .config used. Do you have any idea why it might be missing?
Okay. Found it. Its the missing type here. I'll send some follow-up patches once I have all tested on canyonlands here.

OK good. Also see u-boot-dm branch ppc-working if you need more context.

Regards,
Simon

Stefan Roese

2015-01-26 13:38 UTC #23 lore raw

PATCH 15/19 dm: powerpc: ppc4xx: Move glacier to use driver model for serial

Hi Simon,

Quoted 11 lines · Simon Glass

Quoted 11 lines
Quoted 11 lines
Hmmm. This does not seem to work. With your patch series applied on current top-of-tree I get this error (for glacier_ramboot or canyonlands):
$ make -s -j10
board/amcc/canyonlands/Kconfig:38:warning: config symbol defined without
type
drivers/serial/serial-uclass.c:29:2: error: #error "Serial is required
before relocation - define CONFIG_SYS_MALLOC_F_LEN to make this work"
make[1]: *** [drivers/serial/serial-uclass.o] Error 1

Quoted 8 lines

Quoted 6 lines
Quoted 2 lines
Somehow the CONFIG_SYS_MALLOC_F_LEN define is not available in the .config used. Do you have any idea why it might be missing?
Okay. Found it. Its the missing type here. I'll send some follow-up patches once I have all tested on canyonlands here.
OK good. Also see u-boot-dm branch ppc-working if you need more context.

I just posted 4 patches that are needed on top of your patchset from mid December. Please either include them in your next patchset or fold them into your patches. With these patches on top I can boot Canyonlands from NOR flash. So for the complete patch series:

Tested-by Stefan Roese <[email protected]>

Thanks,
Stefan

Simon Glass

author

2014-12-15 14:19 UTC #24 lore raw

PATCH 16/19 powerpc: Add linkage.h file

In reply to Simon Glass

This permits us to use linux/linkage.h on PowerPC machines.

Signed-off-by Simon Glass <[email protected]>

---

 arch/powerpc/include/asm/linkage.h | 7 +++++++
 1 file changed, 7 insertions(+)
 create mode 100644 arch/powerpc/include/asm/linkage.h

arch/powerpc/include/asm/linkage.h new file +7 -0

new file mode 100644
index 0000000..559b42e
@@ -0,0 +1,7 @@
+/*
+ * Copyright (c) 2014 Google, Inc
+ *
+ * SPDX-License-Identifier:	GPL-2.0+
+ */
+
+/* We don't need anything here at present */

Signature · 1 line

2.2.0.rc0.207.ga3a616c

Simon Glass

author

2014-12-15 14:19 UTC #25 lore raw

PATCH 17/19 serial: Support an early UART for debugging

In reply to Simon Glass

This came up in a discussion on the mailing list here:

https://patchwork.ozlabs.org/patch/384613/

My concerns at the time were:
- it doesn't need to be written in assembler
- it doesn't need to be ARM-specific

This patch provides a possible alternative. It works by allowing any serial driver to export one init function and provide a putc() function. These can be used to output debug data before the real serial driver is available.

This implementation does not depend on driver model, and it is possible for it to operate without a stack on some architectures (e.g. PowerPC, ARM). It provides the same features as the ARM-specific debug.S but with more UART and architecture support.

Signed-off-by Simon Glass <[email protected]>

---

 drivers/serial/Kconfig |  46 ++++++++++++++++
 include/debug_uart.h   | 139 +++++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 185 insertions(+)
 create mode 100644 include/debug_uart.h

drivers/serial/Kconfig +46 -0

index a0b6e02..268bb42 100644
@@ -5,6 +5,52 @@ config DM_SERIAL
 	  If you want to use driver model for serial drivers, say Y.
 	  To use legacy serial drivers, say N.
 
+config DEBUG_UART
+	bool "Enable an early debug UART for debugging"
+	help
+	  The debug UART is intended for use very early in U-Boot to debug
+	  problems when an ICE or other debug mechanism is not available.
+
+	  To use it you should:
+	  - Make sure your UART supports this interface
+	  - Enable CONFIG_DEBUG_UART
+	  - Enable the CONFIG for your UART to tell it to provide this interface
+	        (e.g. CONFIG_DEBUG_UART_NS16550)
+	  - Define the required settings as needed (see below)
+	  - Call debug_uart_init() before use
+	  - Call debug_uart_putc() to output a character
+
+	  Depending on your platform it may be possible to use this UART before
+	  a stack is available.
+
+	  If your UART does not support this interface you can probably add
+	  support quite easily. Remember that you cannot use driver model and
+	  it is preferred to use no stack.
+
+	  You must not use this UART once driver model is working and the
+	  serial drivers are up and running (done in serial_init()). Otherwise
+	  the drivers may conflict and you will get strange output.
+
+config DEBUG_UART_BASE
+	hex "Base address of UART"
+	depends on DEBUG_UART
+	help
+	  This is the base address of your UART for memory-mapped UARTs.
+
+	  A default should be provided by your board, but if not you will need
+	  to use the correct value here.
+
+config DEBUG_UART_CLOCK
+	int "UART input clock"
+	depends on DEBUG_UART
+	help
+	  The UART input clock determines the speed of the internal UART
+	  circuitry. The baud rate is derived from this by dividing the input
+	  clock down.
+
+	  A default should be provided by your board, but if not you will need
+	  to use the correct value here.
+
 config UNIPHIER_SERIAL
 	bool "UniPhier on-chip UART support"
 	depends on ARCH_UNIPHIER && DM_SERIAL

include/debug_uart.h new file +139 -0

139 lines not on this page · raw message

Signature · 1 line

2.2.0.rc0.207.ga3a616c

Masahiro Yamada

2014-12-16 04:38 UTC #26 lore raw

PATCH 17/19 serial: Support an early UART for debugging

Hi Simon,

Quoted 6 lines · On Mon, 15 Dec 2014 07:19:52 -0700 Simon Glass <[email protected]> wrote:

This came up in a discussion on the mailing list here:
https://patchwork.ozlabs.org/patch/384613/
My concerns at the time were:
- it doesn't need to be written in assembler

Assembler can guarantee that stack is not used.

Quoted 1 line

- it doesn't need to be ARM-specific

OK.

Quoted 5 lines

This patch provides a possible alternative. It works by allowing any serial driver to export one init function and provide a putc() function. These can be used to output debug data before the real serial driver is available.
This implementation does not depend on driver model,

OK.

Quoted 2 lines

and it is possible for
it to operate without a stack on some architectures (e.g. PowerPC, ARM).

It depends on the optimization option -O2 / -Os.

In 18/19, you add "static inline" to calc_divisor().

"static inline" does not force-inline the code, but still leaves it to the compiler decision. (Marek Vasut also mentioned this in commit 4425e62856)

Best Regards
Masahiro Yamada

Simon Glass

author

2014-12-17 04:46 UTC #27 lore raw

PATCH 17/19 serial: Support an early UART for debugging

Hi Masahiro,

Quoted 7 lines · Masahiro Yamada

Hi Simon,
On Mon, 15 Dec 2014 07:19:52 -0700
Simon Glass <[email protected]> wrote:

Quoted 8 lines

Quoted 6 lines
This came up in a discussion on the mailing list here:
https://patchwork.ozlabs.org/patch/384613/
My concerns at the time were:
- it doesn't need to be written in assembler
Assembler can guarantee that stack is not used.

So can C if written correctly, at least on PowerPC, ARM and a few other archs.

Quoted 10 lines

Quoted 3 lines
- it doesn't need to be ARM-specific
OK.
Quoted 4 lines
This patch provides a possible alternative. It works by allowing any serial driver to export one init function and provide a putc() function. These can be used to output debug data before the real serial driver is available.

Quoted 8 lines

Quoted 1 line
This implementation does not depend on driver model,
OK.
Quoted 2 lines
and it is possible for
it to operate without a stack on some architectures (e.g. PowerPC, ARM).
It depends on the optimization option -O2 / -Os.

OK, fair enough.

Quoted 6 lines

In 18/19, you add "static inline" to calc_divisor().
"static inline" does not force-inline the code, but still leaves it to the compiler decision. (Marek Vasut also mentioned this in commit 4425e62856)

Sure, I understand.

For me this implementation works on ARM and PowerPC, and can be applied for any UART in U-Boot. We don't need to write assembler in each UART for each architecture we support. That's my point.

I'm not suggesting we drop the assembler version, just that we have another option. Of course if a particular UART driver wants to include assembler it can.

Regards,
Simon

Simon Glass

author

2014-12-15 14:19 UTC #28 lore raw

PATCH 18/19 serial: ns16550: Support debug UART

In reply to Simon Glass

Add debug UART functions to permit ns16550 to provide an early debug UART. Try to avoid using the stack so that this can be called from assembler before a stack is set up (at least on ARM and PowerPC).

Signed-off-by Simon Glass <[email protected]>

---

 drivers/serial/Kconfig   | 13 +++++++++++++
 drivers/serial/ns16550.c | 42 ++++++++++++++++++++++++++++++++++++++++--
 2 files changed, 53 insertions(+), 2 deletions(-)

drivers/serial/Kconfig +13 -0

index 268bb42..9d0d4c3 100644
@@ -31,6 +31,19 @@ config DEBUG_UART
 	  serial drivers are up and running (done in serial_init()). Otherwise
 	  the drivers may conflict and you will get strange output.
 
+choice
+	prompt "Select which UART will provide the debug UART"
+	depends on DEBUG_UART
+
+config DEBUG_UART_NS16550
+	bool "ns16550"
+	help
+	  Select this to enable a debug UART using the ns16550 driver. You
+	  will need to provide parameters to make this work. The driver will
+	  be available until the real driver model serial is running.
+
+endchoice
+
 config DEBUG_UART_BASE
 	hex "Base address of UART"
 	depends on DEBUG_UART

drivers/serial/ns16550.c +40 -2

69 lines not on this page · raw message

Signature · 1 line

2.2.0.rc0.207.ga3a616c

Masahiro Yamada

2014-12-16 04:39 UTC #29 lore raw

PATCH 18/19 serial: ns16550: Support debug UART

Hi Simon,

Quoted 8 lines · On Mon, 15 Dec 2014 07:19:53 -0700 Simon Glass <[email protected]> wrote:

drivers/serial/ns16550.c +8 -0

+#ifdef CONFIG_DEBUG_UART_NS16550
+void debug_uart_init(void)
+{
+	struct NS16550 *com_port = (struct NS16550 *)CONFIG_DEBUG_UART_BASE;
+	int baud_divisor;
+
+	baud_divisor = calc_divisor(com_port, CONFIG_DEBUG_UART_CLOCK,
+				    CONFIG_BAUDRATE);

drivers/serial/ns16550.c int NS16550_tstc(NS16550_t com_port) · line 245

Quoted 10 lines

drivers/serial/ns16550.c +10 -0

+
+	writeb(CONFIG_SYS_NS16550_IER, &com_port->ier);
+	writeb(UART_MCRVAL, &com_port->mcr);
+	writeb(UART_FCRVAL, &com_port->fcr);
+
+	writeb(UART_LCR_BKSE | UART_LCRVAL, &com_port->lcr);
+	writeb(baud_divisor & 0xff, &com_port->dll);
+	writeb((baud_divisor >> 8) & 0xff, &com_port->dlm);
+	writeb(UART_LCRVAL, &com_port->lcr);
+}

This function is duplicating NS16550_init() and NS16550_setbrg().

drivers/serial/ns16550.c int NS16550_tstc(NS16550_t com_port) · line 254

Quoted 8 lines

drivers/serial/ns16550.c +8 -0

+static inline void _debug_uart_putc(int ch)
+{
+	struct NS16550 *com_port = (struct NS16550 *)CONFIG_DEBUG_UART_BASE;
+
+	while (!(readb(&com_port->lsr) & UART_LSR_THRE))
+		;
+	writeb(ch, &com_port->thr);
+}

This is duplicating ns16550_serial_putc().

What does this patch save?

It looks like the driver code is just getting dirtier.

Best Regards
Masahiro Yamada

Simon Glass

author

2014-12-17 04:47 UTC #30 lore raw

PATCH 18/19 serial: ns16550: Support debug UART

Hi Masahiro,

Quoted 18 lines · Masahiro Yamada

Hi Simon,

Quoted 12 lines · On Mon, 15 Dec 2014 07:19:53 -0700 Simon Glass <[email protected]> wrote:

drivers/serial/ns16550.c +12 -0

+#ifdef CONFIG_DEBUG_UART_NS16550
+void debug_uart_init(void)
+{
+     struct NS16550 *com_port = (struct NS16550 *)CONFIG_DEBUG_UART_BASE;
+     int baud_divisor;
+
+     baud_divisor = calc_divisor(com_port, CONFIG_DEBUG_UART_CLOCK,
+                                 CONFIG_BAUDRATE);
+
+     writeb(CONFIG_SYS_NS16550_IER, &com_port->ier);
+     writeb(UART_MCRVAL, &com_port->mcr);
+     writeb(UART_FCRVAL, &com_port->fcr);

Quoted 8 lines

drivers/serial/ns16550.c int NS16550_tstc(NS16550_t com_port) · line 245
Quoted 6 lines
drivers/serial/ns16550.c +6 -0
+
+     writeb(UART_LCR_BKSE | UART_LCRVAL, &com_port->lcr);
+     writeb(baud_divisor & 0xff, &com_port->dll);
+     writeb((baud_divisor >> 8) & 0xff, &com_port->dlm);
+     writeb(UART_LCRVAL, &com_port->lcr);
+}
This function is duplicating NS16550_init() and NS16550_setbrg().

Yes I should fix that.

Quoted 13 lines

drivers/serial/ns16550.c int NS16550_tstc(NS16550_t com_port) · line 254
Quoted 11 lines
drivers/serial/ns16550.c +8 -0
+static inline void _debug_uart_putc(int ch)
+{
+     struct NS16550 *com_port = (struct NS16550 *)CONFIG_DEBUG_UART_BASE;
+
+     while (!(readb(&com_port->lsr) & UART_LSR_THRE))
+             ;
+     writeb(ch, &com_port->thr);
+}
This is duplicating ns16550_serial_putc().

Ditto.

Quoted 6 lines

What does this patch save?
It looks like the driver code is just getting dirtier.

See my other patch for an explanation. We could add this feature to any serial driver.

Regards,
Simon

Simon Glass

author

2014-12-15 14:19 UTC #31 lore raw

PATCH 19/19 powerpc: ppc4xx: Provide early debug UART defaults

In reply to Simon Glass

Provide defaults so that the early debug UART can be enabled.

Signed-off-by Simon Glass <[email protected]>

---

 board/amcc/canyonlands/Kconfig | 6 ++++++
 1 file changed, 6 insertions(+)

board/amcc/canyonlands/Kconfig +6 -0

index cbc5ff9..bfdf17a 100644
@@ -38,4 +38,10 @@ config DM_SERIAL
 config SYS_MALLOC_F_LEN
 	default 0x400
 
+config DEBUG_UART_BASE
+	default 0xef600300
+
+config DEBUG_UART_CLOCK
+	default 7372800
+
 endif

Signature · 1 line

2.2.0.rc0.207.ga3a616c

Simon Glass

author

2015-01-24 04:06 UTC #32 lore raw

cover powerpc: Introduce device tree control and driver model

In reply to Simon Glass

+Tom

Hi,

Quoted 35 lines · Simon Glass

This series does a small amount of tweaking to support device tree control (CONFIG_OF_CONTROL) on PowerPC platforms. It also adds support for driver model. In both cases the main effort is to set things up correctly before calling board_init_f().
A new generic function, board_init_f_mem() is introduced. This does the various memory calculations in C code, since they are messy in assembler and every architecture should in fact be the same. A later series will adjust ARM and x86 to use this function.
As an example, the Canyonlands boards are converted over to use device tree control and driver model for their serial console. It should be fairly straightforward to convert over other boards.
Simon Glass (18):
Introduce board_init_f_mem() to handle early memory layout
powerpc: Permit device tree control of U-Boot (CONFIG_OF_CONTROL)
powerpc: ppc4xx: canyonlands: config: Tidy up CONFIGs and config.mk
powerpc: ppc4xx: Move CANYONLANDS/GLACIER/ARCHES to Kconfig
powerpc: ppc4xx: Add ramboot config for glacier
powerpc: ppc4xx: canyonlands: Move to generic board
powerpc: ppc4xx: dts: Bring in canyonlands device tree files
powerpc: ppc4xx: Call board_init_f_mem() for generic board
powerpc: ppc4xx: Add a gpio.h header file
powerpc: ppc4xx: Allow the end of u-boot.bin to be found
powerpc: ppc4xx: Use CONFIG_OF_CONTROL for canyonlands boards
ppc: amcc: Omit unneeded ns16550 CONFIG if using driver model
powerpc: Add serial driver for driver model
dm: powerpc: ppc4xx: Move glacier to use driver model for serial
powerpc: Add linkage.h file
serial: Support an early UART for debugging
serial: ns16550: Support debug UART
powerpc: ppc4xx: Provide early debug UART defaults

Quoted 3 lines

Stefan Roese (1):
WIP: powerpc: ppc4xx: Somehow BSS is not cleared in RAMBOOT case

Are there any comments on this series? How should I go about getting it applied?

Quoted 43 lines

 arch/Kconfig                                |   1 +
 arch/powerpc/cpu/ppc4xx/44x_spd_ddr2.c      |   8 +
 arch/powerpc/cpu/ppc4xx/config.mk           |   5 +-
 arch/powerpc/cpu/ppc4xx/cpu_init.c          |   2 +
 arch/powerpc/cpu/ppc4xx/start.S             |  18 +-
 arch/powerpc/cpu/ppc4xx/u-boot.lds          |   8 +-
 arch/powerpc/dts/Makefile                   |  11 +
 arch/powerpc/dts/arches.dts                 | 339 ++++++++++++++++
 arch/powerpc/dts/canyonlands.dts            | 555 ++++++++++++++++++++++++++
 arch/powerpc/dts/glacier.dts                | 579 ++++++++++++++++++++++++++++
 arch/powerpc/include/asm/arch-ppc4xx/gpio.h |   7 +
 arch/powerpc/include/asm/linkage.h          |   7 +
 arch/powerpc/include/asm/ppc460ex_gt.h      |   2 +
 arch/powerpc/lib/board.c                    |   3 +
 board/amcc/canyonlands/Kconfig              |  35 ++
 board/amcc/canyonlands/MAINTAINERS          |   1 +
 board/amcc/canyonlands/config.mk            |   2 -
 board/amcc/canyonlands/u-boot-ram.lds       |  85 ++++
 common/board_f.c                            |  18 +
 common/board_r.c                            |   8 +-
 configs/arches_defconfig                    |   5 +-
 configs/canyonlands_defconfig               |   5 +-
 configs/glacier_defconfig                   |   5 +-
 configs/glacier_ramboot_defconfig           |   8 +
 drivers/serial/Kconfig                      |  59 +++
 drivers/serial/Makefile                     |   1 +
 drivers/serial/ns16550.c                    |  42 +-
 drivers/serial/serial_ppc.c                 |  40 ++
 include/configs/amcc-common.h               |   2 +
 include/configs/canyonlands.h               |  38 +-
 include/debug_uart.h                        | 139 +++++++
 31 files changed, 2006 insertions(+), 32 deletions(-)
 create mode 100644 arch/powerpc/dts/Makefile
 create mode 100644 arch/powerpc/dts/arches.dts
 create mode 100644 arch/powerpc/dts/canyonlands.dts
 create mode 100644 arch/powerpc/dts/glacier.dts
 create mode 100644 arch/powerpc/include/asm/arch-ppc4xx/gpio.h
 create mode 100644 arch/powerpc/include/asm/linkage.h
 create mode 100644 board/amcc/canyonlands/u-boot-ram.lds
 create mode 100644 configs/glacier_ramboot_defconfig
 create mode 100644 drivers/serial/serial_ppc.c
 create mode 100644 include/debug_uart.h

Quoted 3 lines

--
2.2.0.rc0.207.ga3a616c

Regards,
Simon