Skip to contentSkip to conversation

Introduce generic pre-relocation board_f.c

PATCH New Simon Glass 3 participants 7 messages Mar 11, 2013 – Mar 30, 2013 1 message not in the archive

Subsystems (4)
lore
Share & tools
lore
#0 lore raw

PATCH v9 9/31 Introduce generic pre-relocation board_f.c

This file handles common pre-relocation init for boards which use the generic framework.

It starts up the console, DRAM, performs relocation and then jumps to post-relocation init.

---
Changes in v9:
- Rebase x86 global_data patch on top of mainline

Changes in v8: None
Changes in v7: None
Changes in v6:
- Tidy up stack reservation code to separate out ARM
- Fix up ordering of dram init (specifically to get ARM SPL boards to boot)
- Add debug() to show RAM top
- Adjust the way the top of memory is calculated

Changes in v5:
- Add fdtdec header file to board_f.c
- Remove setup_global_data_ptr()
- Tidy up stack init
- Put global data on stack in board_init_f()
- Add fdt relocation
- Remove fdt relocation field from x86 arch_global_data
- Deal with change of board_init_f() semantics on ARM
- Save boot_flags in global_data in board_init_f()
- Add print_cpuinfo() prototype to include/common.h

Changes in v4:
- Use asm/sections.h instead of asm-generic/sections.h

Changes in v3:
- Cast away the volatile on gd for memcpy()

Changes in v2: None

 arch/x86/lib/board.c              |   2 +-
 arch/x86/lib/init_helpers.c       |   2 +-
 arch/x86/lib/relocate.c           |   8 +-
 common/Makefile                   |   3 +
 common/board_f.c                  | 579 ++++++++++++++++++++++++++++++++++++++
 include/asm-generic/global_data.h |   2 +
 include/common.h                  |   1 +
 7 files changed, 591 insertions(+), 6 deletions(-)
 create mode 100644 common/board_f.c
arch/x86/lib/board.c +1 -1
index 2441a66..555301a 100644
@@ -219,7 +219,7 @@ static void do_init_loop(init_fnc_t **init_fnc_ptr)
 
 void board_init_f(ulong boot_flags)
 {
-	gd->fdt_blob = gd->arch.new_fdt = NULL;
+	gd->fdt_blob = gd->new_fdt = NULL;
 	gd->flags = boot_flags;
 
 	do_init_loop(init_sequence_f);
arch/x86/lib/init_helpers.c +1 -1
index 414fdcc..7df9536 100644
@@ -111,7 +111,7 @@ int calculate_relocation_address(void)
 	 */
 	if (gd->fdt_blob) {
 		dest_addr -= fdt_size;
-		gd->arch.new_fdt = (void *)dest_addr;
+		gd->new_fdt = (void *)dest_addr;
 		dest_addr &= ~15;
 	}
 #endif
arch/x86/lib/relocate.c +4 -4
index 3e370f2..e893c2b 100644
@@ -49,15 +49,15 @@ int copy_uboot_to_ram(void)
 
 int copy_fdt_to_ram(void)
 {
-	if (gd->arch.new_fdt) {
+	if (gd->new_fdt) {
 		ulong fdt_size;
 
 		fdt_size = ALIGN(fdt_totalsize(gd->fdt_blob) + 0x1000, 32);
 
-		memcpy(gd->arch.new_fdt, gd->fdt_blob, fdt_size);
+		memcpy(gd->new_fdt, gd->fdt_blob, fdt_size);
 		debug("Relocated fdt from %p to %p, size %lx\n",
-		       gd->fdt_blob, gd->arch.new_fdt, fdt_size);
-		gd->fdt_blob = gd->arch.new_fdt;
+		       gd->fdt_blob, gd->new_fdt, fdt_size);
+		gd->fdt_blob = gd->new_fdt;
 	}
 
 	return 0;
common/Makefile +3 -0
index 719fc23..a5f5b1a 100644
@@ -36,6 +36,9 @@ COBJS-y += s_record.o
 COBJS-y += xyzModem.o
 COBJS-y += cmd_disk.o
 
+# boards
+COBJS-$(CONFIG_SYS_GENERIC_BOARD) += board_f.o
+
 # core command
 COBJS-y += cmd_boot.o
 COBJS-$(CONFIG_CMD_BOOTM) += cmd_bootm.o
common/board_f.c new file +579 -0

579 lines not on this page · raw message

include/asm-generic/global_data.h +2 -0
index b8ac024..ebfe2a3 100644
@@ -81,6 +81,8 @@ typedef struct global_data {
 	unsigned long reloc_off;
 	struct global_data *new_gd;	/* relocated global data */
 	const void *fdt_blob;	/* Our device tree, NULL if none */
+	void *new_fdt;		/* Relocated FDT */
+	unsigned long fdt_size;	/* Space reserved for relocated FDT */
 	void **jt;		/* jump table */
 	char env_buf[32];	/* buffer for getenv() before reloc. */
 	struct arch_global_data arch;	/* architecture-specific data */
include/common.h +1 -0
index 6d52924..2ff8db1 100644
@@ -311,6 +311,7 @@ extern ulong monitor_flash_len;
 int mac_read_from_eeprom(void);
 extern u8 _binary_dt_dtb_start[];	/* embedded device tree blob */
 int set_cpu_clk_info(void);
+int print_cpuinfo(void);
 
 /**
  * Show the DRAM size in a board-specific way
Signature · 1 line
1.8.1.3
#1 lore raw

PATCH v9 13/31 arm: Enable generic board support

This enables generic board support so that ARM boards can define CONFIG_SYS_GENERIC_BOARD.

---
Changes in v9:
- Use !CONFIG instead of nCONFIG

Changes in v8:
- Define __HAVE_ARCH_GENERIC_BOARD in ARM's config.mk

Changes in v7: None
Changes in v6: None
Changes in v5: None
Changes in v4: None
Changes in v3: None
Changes in v2: None

 arch/arm/config.mk            | 3 +++
 arch/arm/include/asm/u-boot.h | 9 +++++++++
 arch/arm/lib/Makefile         | 3 +++
 3 files changed, 15 insertions(+)
arch/arm/config.mk +3 -0
index 24b9d7c..a0c89b7 100644
@@ -31,6 +31,9 @@ CONFIG_STANDALONE_LOAD_ADDR = 0xc100000
 endif
 endif
 
+# Support generic board on ARM
+__HAVE_ARCH_GENERIC_BOARD := y
+
 PLATFORM_CPPFLAGS += -DCONFIG_ARM -D__ARM__
 
 # Choose between ARM/Thumb instruction sets
arch/arm/include/asm/u-boot.h +9 -0
index 2ba98bc..a33fefa 100644
@@ -36,6 +36,12 @@
 #ifndef _U_BOOT_H_
 #define _U_BOOT_H_	1
 
+#ifdef CONFIG_SYS_GENERIC_BOARD
+/* Use the generic board which requires a unified bd_info */
+#include <asm-generic/u-boot.h>
+#else
+
+#ifndef __ASSEMBLY__
 typedef struct bd_info {
 	unsigned int	bi_baudrate;	/* serial console baudrate */
     ulong	        bi_arch_number;	/* unique id for this board */
@@ -49,6 +55,9 @@ typedef struct bd_info {
 	ulong size;
     }			bi_dram[CONFIG_NR_DRAM_BANKS];
 } bd_t;
+#endif
+
+#endif /* !CONFIG_SYS_GENERIC_BOARD */
 
 /* For image.h:image_check_target_arch() */
 #define IH_ARCH_DEFAULT IH_ARCH_ARM
arch/arm/lib/Makefile +3 -0
index 57111af..24c7e7a 100644
@@ -39,7 +39,10 @@ GLCOBJS	+= div0.o
 SOBJS-y += crt0.o
 
 ifndef CONFIG_SPL_BUILD
+ifndef CONFIG_SYS_GENERIC_BOARD
 COBJS-y	+= board.o
+endif
+
 COBJS-y	+= bootm.o
 COBJS-$(CONFIG_SYS_L2_PL310) += cache-pl310.o
 SOBJS-$(CONFIG_USE_ARCH_MEMSET) += memset.o
Signature · 1 line
1.8.1.3
#2 lore raw

PATCH v9 17/31 ppc: Enable generic board support

In reply to Simon Glass

This enables generic board support so that ppc boards can define CONFIG_SYS_GENERIC_BOARD.

---
Changes in v9:
- Use !CONFIG instead of nCONFIG

Changes in v8:
- Define __HAVE_ARCH_GENERIC_BOARD in PPC's config.mk

Changes in v7: None
Changes in v6: None
Changes in v5: None
Changes in v4: None
Changes in v3: None
Changes in v2: None

 arch/powerpc/config.mk            | 3 +++
 arch/powerpc/include/asm/u-boot.h | 7 +++++++
 arch/powerpc/lib/Makefile         | 2 ++
 3 files changed, 12 insertions(+)
arch/powerpc/config.mk +3 -0
index b706281..e32d2bf 100644
@@ -29,6 +29,9 @@ PLATFORM_RELFLAGS += -fpic -mrelocatable -ffunction-sections -fdata-sections
 PLATFORM_CPPFLAGS += -DCONFIG_PPC -D__powerpc__
 PLATFORM_LDFLAGS  += -n
 
+# Support generic board on PPC
+__HAVE_ARCH_GENERIC_BOARD := y
+
 #
 # When cross-compiling on NetBSD, we have to define __PPC__ or else we
 # will pick up a va_list declaration that is incompatible with the
arch/powerpc/include/asm/u-boot.h +7 -0
index 7229a98..cf972d2 100644
@@ -34,6 +34,11 @@
  * include/asm-ppc/u-boot.h
  */
 
+#ifdef CONFIG_SYS_GENERIC_BOARD
+/* Use the generic board which requires a unified bd_info */
+#include <asm-generic/u-boot.h>
+#else
+
 #ifndef __ASSEMBLY__
 
 typedef struct bd_info {
@@ -144,6 +149,8 @@ typedef struct bd_info {
 
 #endif /* __ASSEMBLY__ */
 
+#endif /* !CONFIG_SYS_GENERIC_BOARD */
+
 /* For image.h:image_check_target_arch() */
 #define IH_ARCH_DEFAULT IH_ARCH_PPC
 
arch/powerpc/lib/Makefile +2 -0
index 86cf02a..59c723b 100644
@@ -59,8 +59,10 @@ SOBJS-y	+= reloc.o
 
 COBJS-$(CONFIG_BAT_RW) += bat_rw.o
 ifndef CONFIG_SPL_BUILD
+ifndef CONFIG_SYS_GENERIC_BOARD
 COBJS-y	+= board.o
 endif
+endif
 COBJS-y	+= bootm.o
 COBJS-y	+= cache.o
 COBJS-y	+= extable.o
Signature · 1 line
1.8.1.3
#3 lore raw

PATCH v9 21/31 x86: Enable generic board support

In reply to Simon Glass

This enables generic board support so that x86 boards can define CONFIG_SYS_GENERIC_BOARD.

---
Changes in v9:
- Use !CONFIG instead of nCONFIG

Changes in v8: None
Changes in v7: None
Changes in v6: None
Changes in v5:
- Avoid setting up gd on x86 as it is already done

Changes in v4: None
Changes in v3: None
Changes in v2: None

 arch/x86/include/asm/u-boot.h | 11 +++++++++++
 arch/x86/lib/Makefile         |  3 +++
 common/board_r.c              |  2 ++
 3 files changed, 16 insertions(+)
arch/x86/include/asm/u-boot.h +11 -0
index 2f45c7b..df759fa 100644
@@ -39,6 +39,13 @@
 #include <config.h>
 #include <compiler.h>
 
+#ifdef CONFIG_SYS_GENERIC_BOARD
+/* Use the generic board which requires a unified bd_info */
+#include <asm-generic/u-boot.h>
+#else
+
+#ifndef __ASSEMBLY__
+
 typedef struct bd_info {
 	unsigned long	bi_memstart;	/* start of DRAM memory */
 	phys_size_t	bi_memsize;	/* size	 of DRAM memory in bytes */
@@ -60,6 +67,10 @@ typedef struct bd_info {
 	}bi_dram[CONFIG_NR_DRAM_BANKS];
 } bd_t;
 
+#endif /* __ASSEMBLY__ */
+
+#endif /* !CONFIG_SYS_GENERIC_BOARD */
+
 /* For image.h:image_check_target_arch() */
 #define IH_ARCH_DEFAULT IH_ARCH_I386
 
arch/x86/lib/Makefile +3 -0
index 9b24dc5..ee89354 100644
@@ -25,7 +25,10 @@ include $(TOPDIR)/config.mk
 
 LIB	= $(obj)lib$(ARCH).o
 
+ifeq ($(CONFIG_SYS_GENERIC_BOARD),)
 COBJS-y	+= board.o
+endif
+
 COBJS-y	+= bootm.o
 COBJS-y	+= cmd_boot.o
 COBJS-y	+= gcc.o
common/board_r.c +2 -0
index 29eccdf..230887d 100644
@@ -509,11 +509,13 @@ static int show_model_r(void)
 #endif
 
 /* enable exceptions */
+#ifdef CONFIG_ARM
 static int initr_enable_interrupts(void)
 {
 	enable_interrupts();
 	return 0;
 }
+#endif
 
 #ifdef CONFIG_CMD_NET
 static int initr_ethaddr(void)
Signature · 1 line
1.8.1.3

A message that is not in the archive (sent off-list or not yet received): [email protected]

#5 lore raw

v9 cover Create generic board init for ARM, x86, PPC

Dear Simon Glass,

Quoted 4 lines · In message <[email protected]> you wrote:

This series creates a generic board init implementation which contains the essential functions of the major arch/xxx/lib/board.c files. It is split into two parts: board_f.c for pre-relocation and board_r.c for post-relocation.

...

Quoted 7 lines

ARM is a relatively large board.c file and one which I can test, therefore I think it is a good target for this series. On the other hand, x86 is relatively small and simple, but different enough that it introduces a few issues to be solved. So I have chosen both ARM and x86 for this series. After a suggestion from Wolfgang I have added PPC also. This is the largest and most feature-full board, so hopefully we have all bases covered in this series. Other archs are mostly a subset of these.

I did some testing on Power Architecture systems - to be precise:

CPU Board
--------------------------
MPC860T TQM860L
MPC5200 TQM5200S
PPC405EX Kilauea
PPC440EP Yosemite
PPC440EX Sequoia
PPC440SPe Katmai
PPC460EX Canyonlands

On all boards I verified at least environment access (printenv, saveenv) and network operation (TFTP download); where available also PCI. Everything appears to work fine, no problems noted.

A well done job, thanks a lot!

For the whole series:

Best regards,

Wolfgang Denk

Signature · 10 lines
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
Certain old men prefer to rise at dawn, taking a cold bath and a long
walk with an empty stomach and otherwise mortifying the  flesh.  They
then point with pride to these practices as the cause of their sturdy
health  and ripe years; the truth being that they are hearty and old,
not because of their habits, but in spite of them. The reason we find
only robust persons doing this thing is that it has  killed  all  the
others who have tried it.  - Ambrose Bierce, "The Devil's Dictionary"
#6 lore raw

v9 cover Create generic board init for ARM, x86, PPC

Quoted 32 lines · Wolfgang Denk

Dear Simon Glass,

Quoted 4 lines · In message <[email protected]> you wrote:

This series creates a generic board init implementation which contains the essential functions of the major arch/xxx/lib/board.c files. It is split into two parts: board_f.c for pre-relocation and board_r.c for post-relocation.

...

Quoted 7 lines

ARM is a relatively large board.c file and one which I can test, therefore I think it is a good target for this series. On the other hand, x86 is relatively small and simple, but different enough that it introduces a few issues to be solved. So I have chosen both ARM and x86 for this series. After a suggestion from Wolfgang I have added PPC also. This is the largest and most feature-full board, so hopefully we have all bases covered in this series. Other archs are mostly a subset of these.

I did some testing on Power Architecture systems - to be precise:

CPU Board
--------------------------
MPC860T TQM860L
MPC5200 TQM5200S
PPC405EX Kilauea
PPC440EP Yosemite
PPC440EX Sequoia
PPC440SPe Katmai
PPC460EX Canyonlands

On all boards I verified at least environment access (printenv, saveenv) and network operation (TFTP download); where available also PCI. Everything appears to work fine, no problems noted.

Quoted 6 lines

A well done job, thanks a lot!

For the whole series:

OK, I've taken http://patchwork.ozlabs.org/bundle/sjg/us-board/ but with
http://patchwork.ozlabs.org/patch/228080/ for 2/31 and applied this to u-boot/master now, thanks again!

Signature · 8 lines
Tom
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.denx.de/pipermail/u-boot/attachments/20130315/da1bf0d1/attachment.pgp>
#7 lore raw

v9 cover Create generic board init for ARM, x86, PPC

In reply to Wolfgang Denk

Hi Wolfgang,

Quoted 30 lines · Wolfgang Denk

Dear Simon Glass,

Quoted 4 lines · In message <[email protected]> you wrote:

This series creates a generic board init implementation which contains the essential functions of the major arch/xxx/lib/board.c files. It is split into two parts: board_f.c for pre-relocation and board_r.c for post-relocation.

...

Quoted 1 line

ARM is a relatively large board.c file and one which I can test,

therefore

Quoted 3 lines

I think it is a good target for this series. On the other hand, x86 is relatively small and simple, but different enough that it introduces a few issues to be solved. So I have chosen both ARM and x86 for this

series.

Quoted 3 lines

After a suggestion from Wolfgang I have added PPC also. This is the largest and most feature-full board, so hopefully we have all bases covered in this series. Other archs are mostly a subset of these.

I did some testing on Power Architecture systems - to be precise:

CPU Board
--------------------------

MPC860T         TQM860L
MPC5200         TQM5200S
PPC405EX        Kilauea
PPC440EP        Yosemite
PPC440EX        Sequoia
PPC440SPe       Katmai
PPC460EX        Canyonlands
Quoted 6 lines

On all boards I verified at least environment access (printenv, saveenv) and network operation (TFTP download); where available also PCI. Everything appears to work fine, no problems noted.

A well done job, thanks a lot!

Thanks for your comments which are much appreciated. There is still work to do, in cleaning up the implementation to reduce the differences between arch code (e.g. all the #ifdefs in board_f/r.c). But this is a great first step and I am pleased to get this in.

Quoted 11 lines

For the whole series:

Best regards,

Wolfgang Denk

Quoted 12 lines

--
DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de Certain old men prefer to rise at dawn, taking a cold bath and a long walk with an empty stomach and otherwise mortifying the flesh. They then point with pride to these practices as the cause of their sturdy health and ripe years; the truth being that they are hearty and old, not because of their habits, but in spite of them. The reason we find only robust persons doing this thing is that it has killed all the others who have tried it. - Ambrose Bierce, "The Devil's Dictionary"

But I wonder if this is trying to tell me to stop these tricky refactors :-)

Regards,
Simon