August 2008 - U-Boot

[U-Boot] [patch] arm920t fix constant error in start.S
by Andrew Dyer 26 Aug '08

26 Aug '08

Code in cpu/arm920t/start.S will die with a compilation error if CONFIG_STACKSIZE + CFG_MALLOC_LEN works out to an invalid constant for the ARM sub instruction. Change the code so that each is subtracted independently to avoid the error. Signed-off-by: Andrew Dyer <adyer(a)righthandtech.com> --- cpu/arm920t/start.S | 6 ++++-- 1 files changed, 4 insertions(+), 2 deletions(-) diff --git a/cpu/arm920t/start.S b/cpu/arm920t/start.S index 62231f8..5d0fec6 100644 --- a/cpu/arm920t/start.S +++ b/cpu/arm920t/start.S @@ -315,7 +315,8 @@ cpu_init_crit: sub sp, sp, #S_FRAME_SIZE stmia sp, {r0 - r12} @ Calling r0-r12 ldr r2, _armboot_start - sub r2, r2, #(CONFIG_STACKSIZE+CFG_MALLOC_LEN) + sub r2, r2, #(CONFIG_STACKSIZE) + sub r2, r2, #(CFG_MALLOC_LEN) sub r2, r2, #(CFG_GBL_DATA_SIZE+8) @ set base 2 words into abort stack ldmia r2, {r2 - r3} @ get pc, cpsr add r0, sp, #S_FRAME_SIZE @ restore sp_SVC @@ -348,7 +349,8 @@ cpu_init_crit: .macro get_bad_stack ldr r13, _armboot_start @ setup our mode stack - sub r13, r13, #(CONFIG_STACKSIZE+CFG_MALLOC_LEN) + sub r13, r13, #(CONFIG_STACKSIZE) + sub r13, r13, #(CFG_MALLOC_LEN) sub r13, r13, #(CFG_GBL_DATA_SIZE+8) @ reserved a couple spots in abort stack str lr, [r13] @ save caller lr / spsr -- 1.5.4.3

1 0

[U-Boot] [RFC] [PATCH 0/3] MPC8349EMDS: Using mpc83xx generic pci infrastructure for PCI Agent mode
by Ira W. Snyder 26 Aug '08

26 Aug '08

Hello u-boot-users, This is my attempt at adding CONFIG_PCISLAVE support to the MPC8349EMDS board. I noticed that there was some shared infrastructure for the PCI controller in the 83xx-series CPU's, so I removed all of the old code, and used the infrastructure instead. I have boot tested the board in host mode (using the 83XX_GENERIC_PCI infrastructure) but I do not have the IO Expander board, so I cannot test that PCI cards still work. Linux happily initializes the PCI interface, and lspci -vvv is identical before and after my patch. I copied the configuration bit unlocking routine from other boards support for CONFIG_PCISLAVE. I added it to the shared infrastructure so it would have access to the static pci_controller that is set up and used by the other infrastructure code. I then went ahead and added support for CONFIG_PCISLAVE to the MPC8349EMDS board. I have boot tested the board in a desktop pc. I have booted Linux on the board in this mode (with CONFIG_PCI=n so it does not try to re-initialize other PCI devices on the bus. Linux is only set up for Host-mode PCI), and have successfully developed both U-Boot and Linux drivers that implement a virtual serial port and virtual ethernet device (between the host pc and the MPC8349EMDS board) using only the PCI bus and the MPC8349EMDS message registers. I'm happy to post them if there is any interest. I'm looking for suggestions and improvements so that these can be accepted into the U-Boot source. Thanks, Ira

2 4

[U-Boot] [PATCH v2] RTC: Fix Makefile problem with COBJS-$(CONFIG_RTC_DS1307 || CONFIG_RTC_DS1338)
by Stefan Roese 25 Aug '08

25 Aug '08

This "||" doesn't seem to work. Now using the idea suggest by Scott Wood to combine both config options into one line. This even allows defining both options and not generating the target object twice. Signed-off-by: Stefan Roese <sr(a)denx.de> --- drivers/rtc/Makefile | 3 ++- 1 files changed, 2 insertions(+), 1 deletions(-) diff --git a/drivers/rtc/Makefile b/drivers/rtc/Makefile index 2af3ee5..86f855f 100644 --- a/drivers/rtc/Makefile +++ b/drivers/rtc/Makefile @@ -32,7 +32,8 @@ COBJS-y += date.o COBJS-$(CONFIG_RTC_DS12887) += ds12887.o COBJS-$(CONFIG_RTC_DS1302) += ds1302.o COBJS-$(CONFIG_RTC_DS1306) += ds1306.o -COBJS-$(CONFIG_RTC_DS1307 || CONFIG_RTC_DS1338) += ds1307.o +COBJS-$(CONFIG_RTC_DS1307)$(CONFIG_RTC_DS1338) += ds1307.o +COBJS-y += $(COBJS-yy) COBJS-$(CONFIG_RTC_DS1337) += ds1337.o COBJS-$(CONFIG_RTC_DS1374) += ds1374.o COBJS-$(CONFIG_RTC_DS1556) += ds1556.o -- 1.5.6.5

2 1

[U-Boot] Please pull: u-boot-fdt.git
by Jerry Van Baren 25 Aug '08

25 Aug '08

The following changes since commit afe3848b79a7ff351e9fbf3a7c162d2de002279b: Wolfgang Denk (1): Merge branch 'master' of git://git.denx.de/u-boot-ppc4xx are available in the git repository at: http://git.denx.de/u-boot-fdt.git master David Gibson (5): dtc: Enable and fix -Wpointer-arith warnings dtc: Enable and fix -Wcast-qual warnings libfdt: Increase namespace-pollution paranoia libfdt: Forgot one function when cleaning the namespace libfdt: Implement fdt_get_property_namelen() and fdt_getprop_namelen() Gerald Van Baren (1): Add note on dereferencing /aliases pointers Kumar Gala (1): libfdt: Add support for using aliases in fdt_path_offset() Wolfram Sang (1): libfdt: Improve documentation in libfdt.h common/cmd_fdt.c | 4 +- include/libfdt.h | 66 +++++++++++++----- include/libfdt_env.h | 12 ++++ libfdt/fdt.c | 10 ++-- libfdt/fdt_ro.c | 87 ++++++++++++++++++------- libfdt/fdt_rw.c | 166 +++++++++++++++++++++++---------------------- libfdt/fdt_strerror.c | 34 +++++----- libfdt/fdt_sw.c | 40 ++++++------ libfdt/fdt_wip.c | 9 ++- libfdt/libfdt_internal.h | 20 +++--- 10 files changed, 267 insertions(+), 181 deletions(-)

2 1

[U-Boot] [patch 1/2] Enable the use of an uncompressed kernel image for U-boot uImage
by Remy Bohmer 25 Aug '08

25 Aug '08

Normally the 'zImage' is used as base for uImage. Enable the capability to use the raw uncompressed 'Image' instead Signed-off-by: Remy Bohmer <linux(a)bohmer.net> --- arch/arm/Kconfig | 10 ++++++++++ arch/arm/boot/Makefile | 6 +++++- 2 files changed, 15 insertions(+), 1 deletion(-) Index: linux-2.6.27-rc4/arch/arm/Kconfig =================================================================== --- linux-2.6.27-rc4.orig/arch/arm/Kconfig 2008-08-25 20:58:33.000000000 +0200 +++ linux-2.6.27-rc4/arch/arm/Kconfig 2008-08-25 20:58:36.000000000 +0200 @@ -967,6 +967,16 @@ config XIP_KERNEL If unsure, say N. +config UNCOMPRESSED_UIMAGE + bool "Use uncompressed kernel image for uImage" + default n + help + Enable this option if you want uImage to be based on the kernel + 'Image' instead of 'zImage'. This will increase the kernel image + size, but it can improve kernel boot time in some situations. + + If unsure, say N. + config XIP_PHYS_ADDR hex "XIP Kernel Physical Location" depends on XIP_KERNEL Index: linux-2.6.27-rc4/arch/arm/boot/Makefile =================================================================== --- linux-2.6.27-rc4.orig/arch/arm/boot/Makefile 2008-08-25 20:58:33.000000000 +0200 +++ linux-2.6.27-rc4/arch/arm/boot/Makefile 2008-08-25 20:58:36.000000000 +0200 @@ -70,7 +70,11 @@ else $(obj)/uImage: LOADADDR=$(ZRELADDR) endif -$(obj)/uImage: $(obj)/zImage FORCE +ifeq ($(CONFIG_UNCOMPRESSED_UIMAGE),y) +$(obj)/uImage: $(obj)/Image FORCE +else +$(obj)/uImage: $(obj)/zImage FORCE +endif $(call if_changed,uimage) @echo ' Image $@ is ready' --

2 1

Re: [U-Boot] Question about FIT images with uboot 1.3.4 (solved)
by Remy Bohmer 25 Aug '08

25 Aug '08

Hello All, > I know the image itself is correct. mkimage -l shows correct > information, but when copied to the target, the target detects a FIT > header but complains that it cannot find the description section. > Legacy images work properly. > Tomorrow I will hook up the debugger and get this problem tackled. Finally I got the problem tracked down. It was very stupid (as always) but the reason why FIT-images did not work here, was because the last bytes of the image were overwritten by downloading the ram drive image. The legacy uImage just fit in the 3MB memory space, and the FIT image just did not. (it is quite big, because I use an uncompressed kernel) So, moving up the ramdrive image to a higher address solved this problem... Keeps worth mentioning that the FIT error message (no description found) had put me completely in the wrong direction... For convenience, I also have a kernel patch here that allows the kernel to build a FIT image immediately while building the uImage. I will post that on LKML soon. Kind Regards, Remy

3 2

[U-Boot] Multiple USB controller support under Uboot
by mike zheng 25 Aug '08

25 Aug '08

Hello, I am using ISP156x USB PCI controller with Uboot. Right now, the code can find only 1 root hub when I use "usb inf" or "usb tree" command to check the USB tree. If I list the PCI device, it shows there are 2 OHCI and 1 EHCI host controller. Anyone knows why I have only one root hub? What shall the USB tree looks like for ISP156x USB PCI host controller? Thanks, Mike

1 0

[U-Boot] [patch 0/2] Several U-boot uImage building patches
by Remy Bohmer 25 Aug '08

25 Aug '08

Here follows a series of 2 patches that provide some additional functionality to building Linux kernel uImages for U-boot: * Make it possible to base uImage on the non-compressed kernel 'Image' instead of 'zImage'. * Add support for the new U-boot uImage file format, called FIT. This will allow a kernel to be build directly in this new format instead of the legacy format, IF the proper tools are available (dtc v1.2.0 and mkimage 1.3.3; dtc is checked by version-number, the proper mkimage is detected by cmd-line options, it must provide a '-f' option) If older, or improper tools are detected, then the Legacy style image is build after issuing a warning message. Signed-off-by: Remy Bohmer <linux(a)bohmer.net> --

1 0

[U-Boot] [PATCH] Add CONFIG_HDBOOTCOMMAND to the environment
by Kumar Gala 25 Aug '08

25 Aug '08

Provide a boot command for being able to boot from a hard drive Signed-off-by: Kumar Gala <galak(a)kernel.crashing.org> --- common/env_common.c | 3 +++ common/environment.c | 3 +++ tools/env/fw_env.c | 3 +++ 3 files changed, 9 insertions(+), 0 deletions(-) diff --git a/common/env_common.c b/common/env_common.c index d51c211..271e71f 100644 --- a/common/env_common.c +++ b/common/env_common.c @@ -64,6 +64,9 @@ uchar default_environment[] = { #ifdef CONFIG_BOOTCOMMAND "bootcmd=" CONFIG_BOOTCOMMAND "\0" #endif +#ifdef CONFIG_HDBOOTCOMMAND + "hdboot=" CONFIG_HDBOOTCOMMAND "\0" +#endif #ifdef CONFIG_RAMBOOTCOMMAND "ramboot=" CONFIG_RAMBOOTCOMMAND "\0" #endif diff --git a/common/environment.c b/common/environment.c index 3b9914f..17377a7 100644 --- a/common/environment.c +++ b/common/environment.c @@ -108,6 +108,9 @@ env_t environment __PPCENV__ = { #if defined(CONFIG_BOOTCOMMAND) "bootcmd=" CONFIG_BOOTCOMMAND "\0" #endif +#ifdef CONFIG_HDBOOTCOMMAND + "hdboot=" CONFIG_HDBOOTCOMMAND "\0" +#endif #if defined(CONFIG_RAMBOOTCOMMAND) "ramboot=" CONFIG_RAMBOOTCOMMAND "\0" #endif diff --git a/tools/env/fw_env.c b/tools/env/fw_env.c index b8bca91..6520c22 100644 --- a/tools/env/fw_env.c +++ b/tools/env/fw_env.c @@ -87,6 +87,9 @@ static char default_environment[] = { #if defined(CONFIG_BOOTCOMMAND) "bootcmd=" CONFIG_BOOTCOMMAND "\0" #endif +#ifdef CONFIG_HDBOOTCOMMAND + "hdboot=" CONFIG_HDBOOTCOMMAND "\0" +#endif #if defined(CONFIG_RAMBOOTCOMMAND) "ramboot=" CONFIG_RAMBOOTCOMMAND "\0" #endif -- 1.5.5.1

2 2

[U-Boot] [PATCH] Resubmit: Correct ARM Versatile Timer Initialization
by Gururaja Hebbar K R 25 Aug '08

25 Aug '08

- According to ARM Dual-Timer Module (SP804) TRM (ARM DDI0271), -- Timer Value Register @ TIMER Base + 4 is Read-only. -- Prescale Value (Bits 3-2 of TIMER Control register) can only be one of 00,01,10. 11 is undefined. -- CFG_HZ for Versatile board is set to #define CFG_HZ (1000000 / 256) So Prescale bits is set to indicate - 8 Stages of Prescale, Clock divided by 256 - The Timer Control Register has one Undefined/Shouldn't Use Bit So we should do read/modify/write Operation Signed-off-by: Gururaja Hebbar <gururajakr(a)sanyo.co.in> --- cpu/arm926ejs/versatile/timer.c | 37 +++++++++++++++++++++++++++++++++---- 1 files changed, 33 insertions(+), 4 deletions(-) diff --git a/cpu/arm926ejs/versatile/timer.c b/cpu/arm926ejs/versatile/timer.c index 32872d2..17ba387 100755 --- a/cpu/arm926ejs/versatile/timer.c +++ b/cpu/arm926ejs/versatile/timer.c @@ -46,12 +46,41 @@ static ulong timestamp; static ulong lastdec; -/* nothing really to do with interrupts, just starts up a counter. */ +#define TIMER_ENABLE (1 << 7) +#define TIMER_MODE_MSK (1 << 6) +#define TIMER_MODE_FR (0 << 6) +#define TIMER_MODE_PD (1 << 6) + +#define TIMER_INT_EN (1 << 5) +#define TIMER_PRS_MSK (3 << 2) +#define TIMER_PRS_8S (1 << 3) +#define TIMER_SIZE_MSK (1 << 2) +#define TIMER_ONE_SHT (1 << 0) + int timer_init (void) { - *(volatile ulong *)(CFG_TIMERBASE + 0) = CFG_TIMER_RELOAD; /* TimerLoad */ - *(volatile ulong *)(CFG_TIMERBASE + 4) = CFG_TIMER_RELOAD; /* TimerValue */ - *(volatile ulong *)(CFG_TIMERBASE + 8) = 0x8C; + ulong TmrCntrlVal; + + /*1st disable the Timer*/ + TmrCntrlVal = *(volatile ulong *)(CFG_TIMERBASE + 8); + TmrCntrlVal &= ~TIMER_ENABLE; + *(volatile ulong *)(CFG_TIMERBASE + 8) = TmrCntrlVal; + + /* The Timer Control Register has one Undefined/Shouldn't Use Bit + * So we should do read/modify/write Operation + */ + + /* Timer Mode : Free Running + * Interrupt : Disabled + * Prescale : 8 Stage, Clk/256 + * Tmr Siz : 16 Bit Counter + * Tmr in Wrapping Mode + */ + TmrCntrlVal = *(volatile ulong *)(CFG_TIMERBASE + 8); + TmrCntrlVal &= ~(TIMER_MODE_MSK | TIMER_INT_EN | TIMER_PRS_MSK | TIMER_SIZE_MSK | TIMER_ONE_SHT ); + TmrCntrlVal |= (TIMER_ENABLE | TIMER_PRS_8S); + + *(volatile ulong *)(CFG_TIMERBASE + 8) = TmrCntrlVal; /* init the timestamp and lastdec value */ reset_timer_masked(); -- 1.5.6.4

2 2