kernel/u-boot.git
17 years ago[PATCH] Change macro name for UartLite
Michal Simek [Sun, 23 Sep 2007 22:25:11 +0000 (00:25 +0200)]
[PATCH] Change macro name for UartLite
because PowerPC 405 can use UartLite as console

17 years ago[PATCH] Add support for design without interrupt controller
Michal Simek [Sun, 23 Sep 2007 22:21:19 +0000 (00:21 +0200)]
[PATCH] Add support for design without interrupt controller
Polling timer

17 years ago[FIX] resolve problem with cpu without barrel shifter
Michal Simek [Sun, 23 Sep 2007 22:19:48 +0000 (00:19 +0200)]
[FIX] resolve problem with cpu without barrel shifter

17 years ago[FIX] repair email address
Michal Simek [Sun, 23 Sep 2007 22:18:46 +0000 (00:18 +0200)]
[FIX] repair email address

17 years ago[FIX] repair MFSL commands
Michal Simek [Sun, 23 Sep 2007 22:17:42 +0000 (00:17 +0200)]
[FIX] repair MFSL commands

17 years agosynchronizition with mainline
Michal Simek [Sun, 23 Sep 2007 22:08:37 +0000 (00:08 +0200)]
synchronizition with mainline

17 years agoMerge ../u-boot
Michal Simek [Sun, 23 Sep 2007 22:04:22 +0000 (00:04 +0200)]
Merge ../u-boot

17 years agov1.3.0-rc2 v1.3.0-rc2
Wolfgang Denk [Wed, 19 Sep 2007 22:04:14 +0000 (00:04 +0200)]
v1.3.0-rc2

Signed-off-by: Wolfgang Denk <wd@denx.de>
17 years agoMerge with git+ssh://gemini_vpn/home/wd/git/u-boot/master
Wolfgang Denk [Tue, 18 Sep 2007 19:39:39 +0000 (21:39 +0200)]
Merge ... git+ssh://gemini_vpn/home/wd/git/u-boot/master

17 years agoAvoid compiler warning.
Wolfgang Denk [Tue, 18 Sep 2007 19:36:35 +0000 (21:36 +0200)]
Avoid compiler warning.

Signed-off-by: Wolfgang Denk <wd@denx.de>
17 years agoBugfix: remove embedded null (\0) from CFG_BOOTFILE macro in TQM8540_config
Grant Likely [Tue, 18 Sep 2007 18:24:57 +0000 (12:24 -0600)]
Bugfix: remove embedded null (\0) from CFG_BOOTFILE macro in TQM8540_config

/bin/bash and /bin/dash (which /bin/sh is linked to on ubuntu) handle embedded
nulls in a string differently.  For example, the following statement:
    echo "this is a string\0" > afile
Will produce the following with /bin/bash:
    "this is a string\0"
But with /bin/dash, will produce:
    "this is a string

Bug fixed by moving the embedded null out of the makefile and into the
config header.  Also renamed the macro to avoid usage colision with the same
macro used by other board ports.

Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
17 years agoMCC200: fix build warning
Wolfgang Denk [Tue, 18 Sep 2007 15:40:27 +0000 (17:40 +0200)]
MCC200: fix build warning

The MCC200 board config file includes version.h for some customer-
specific setting, which causes warnings with "make depend"; build
version.h before depend.

Signed-off-by: Wolfgang Denk <wd@denx.de>
17 years agoTQM8xx[LM]: Fix broken environment alignment.
Wolfgang Denk [Sun, 16 Sep 2007 15:10:04 +0000 (17:10 +0200)]
TQM8xx[LM]: Fix broken environment alignment.

With recent toolchains, the environment sectors were no longer aligned to
sector boundaries. The reason was a combination of two bugs:

1) common/environment.c assumed that CONFIG_TQM8xxL would be defined
   for all TQM8xxL and TQM8xxM boards. But "include/common.h", where
   this gets defined, is not included here (and cannot be included
   without causing lots of problems).

   Added a new #define CFG_USE_PPCENV for all boards which really
   want to put the environment is a ".ppcenv" section.

2) The linker scripts just include environment.o, silently assuming
   that the objects in that file are really in the order in which
   they are coded in the C file, i. e. "environment" first, then
   "redundand_environment", and "env_size" last. However, current
   toolchains (GCC-4.x) reorder the objects, causing the environment
   data not to start on a flash sector boundary:

   Instead of: we got:

40008000 T environment 40008000 T env_size
4000c000 T redundand_environment 40008004 T redundand_environment
40010000 T env_size 4000c004 T environment

   Note: this patch fixes just the first part, and cures the alignment
   problem by making sure that "env_size" gets placed correctly. However,
   we still have a potential issue because primary and redundant
   environment sectors are actually swapped, i. e. we have now:

40008000 T redundand_environment
4000c000 T environment
40010000 T env_size

   This shall be fixed in the next version.

Signed-off-by: Wolfgang Denk <wd@denx.de>
17 years agoTQM8xx/FPS8xx: adjust flash partitions for 2.6 ARCH=powerpc kernels
Wolfgang Denk [Sun, 16 Sep 2007 00:39:35 +0000 (02:39 +0200)]
TQM8xx/FPS8xx: adjust flash partitions for 2.6 ARCH=powerpc kernels

Signed-off-by: Wolfgang Denk <wd@denx.de>
17 years agoTypo fix in tsec.c
urwithsughosh@gmail.com [Mon, 10 Sep 2007 18:54:56 +0000 (14:54 -0400)]
Typo fix in tsec.c

Fixup for the break statement in wrong place.

[Patch by urwithsughosh@gmail.com]
Acked-by: Andy Fleming <afleming@freescale.com>
Signed-off-by: Wolfgang Denk <wd@denx.de>

17 years agoFix do_div() usage in nand process output
Matthias Fuchs [Tue, 11 Sep 2007 15:04:00 +0000 (17:04 +0200)]
Fix do_div() usage in nand process output

Fix usage of do_div() in nand erase|read|write process output.

The last patch to nand_util.c introduced do_div() instead of libgcc's
implementation. But do_div() returns the quotient in its first
macro parameter and not as result.

Signed-off-by: Matthias Fuchs <matthias.fuchs@esd-electronics.com>
17 years agoNAND: Add CFG_NAND_QUIET option
Matthias Fuchs [Wed, 12 Sep 2007 10:36:53 +0000 (12:36 +0200)]
NAND: Add CFG_NAND_QUIET option

This config option sets the default for the progress information
output behavior that can also be configured through the 'quiet'
environment variable.

The legacy NAND code does not print the current progress info
on the console. So this option is for backward compatibility for
units that are in the field and where setting the quiet variable
is not an option. With CFG_NAND_QUIET set to '1' the console
progress info is turned off. This can still be overwritten
through the environment variable.

Signed-off-by: Matthias Fuchs <matthias.fuchs@esd-electronics.com>
17 years agoColdFire: fix build error becasue of bad type of mii_init()
Liew Tsi Chung-r5aahp [Thu, 13 Sep 2007 23:06:05 +0000 (16:06 -0700)]
ColdFire: fix build error becasue of bad type of mii_init()

Signed-off-by: TsiChungLiew <Tsi-Chung.Liew@freescale.com>
17 years agoColdFire: Fix build error caused by pixis.c
Liew Tsi Chung-r5aahp [Thu, 13 Sep 2007 23:04:05 +0000 (16:04 -0700)]
ColdFire: Fix build error caused by pixis.c

Moved the #include <asm/cache.h> inside the #ifdef CONFIG_FSL_PIXIS.

Signed-off-by: TsiChungLiew <Tsi-Chung.Liew@freescale.com>
17 years agoUpdate MPC8349ITX*_config to place config.tmp in right place.
Sam Sparks [Fri, 14 Sep 2007 17:14:42 +0000 (11:14 -0600)]
Update MPC8349ITX*_config to place config.tmp in right place.

MPC834ITX*_config does not store config.tmp at the correct locatation,
causing MPC8349ITXGP to have the wrong TEXT_BASE.

Signed-off-by: Sam Sparks <SSparks@twacs.com>
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
17 years agoFix cases where DECLARE_GLOBAL_DATA_PTR was not declared as global
Wolfgang Denk [Sat, 15 Sep 2007 18:48:41 +0000 (20:48 +0200)]
Fix cases where DECLARE_GLOBAL_DATA_PTR was not declared as global

Signed-off-by: Wolfgang Denk <wd@denx.de>
17 years agoMake DECLARE_GLOBAL_DATA_PTR global for DaVinci
Dirk Behme [Sat, 15 Sep 2007 09:55:42 +0000 (11:55 +0200)]
Make DECLARE_GLOBAL_DATA_PTR global for DaVinci

As discussed in [1], DECLARE_GLOBAL_DATA_PTR has to be global and not
function local.

Signed-off-by: Dirk Behme <dirk.behme@gmail.com>
[1] http://article.gmane.org/gmane.comp.boot-loaders.u-boot/31805

17 years agocm5200: Fix a typo introduced by afaac86fe2948ac84cd9a12bbed883b3c683e7d9
Bartlomiej Sieka [Thu, 13 Sep 2007 16:21:48 +0000 (18:21 +0200)]
cm5200: Fix a typo introduced by afaac86fe2948ac84cd9a12bbed883b3c683e7d9

Signed-off-by: Marian Balakowicz <m8@semihalf.com>
17 years agoFix memory corruption problem on STX GP3 SSA Board.
Wolfgang Denk [Tue, 11 Sep 2007 22:48:57 +0000 (00:48 +0200)]
Fix memory corruption problem on STX GP3 SSA Board.

Signed-off-by: Wolfgang Denk <wd@denx.de>
17 years ago[GP3SSA] Add define CONFIG_MPC85XX_PCI2 in config file to allow u-boot to
Grzegorz Bernacki [Tue, 11 Sep 2007 13:42:11 +0000 (15:42 +0200)]
[GP3SSA] Add define CONFIG_MPC85XX_PCI2 in config file to allow u-boot to
scan on second pci bus.

Signed-off-by: Grzegorz Bernacki <gjb@semihalf.com>
17 years ago[ppc4xx] Individual handling of sdram.c for bamboo_nand build
Grzegorz Bernacki [Tue, 11 Sep 2007 10:57:52 +0000 (12:57 +0200)]
[ppc4xx] Individual handling of sdram.c for bamboo_nand build

Bamboo has a file sdram.c which needs special treatment when building in
separate directory. It has to be linked to build directory otherwise it is
not seen.

Signed-off-by: Grzegorz Bernacki <gjb@semihalf.com>
17 years ago[FIX] Microblaze ML401 - repare FLASH handling
Michal Simek [Mon, 10 Sep 2007 22:37:04 +0000 (00:37 +0200)]
[FIX] Microblaze ML401 - repare FLASH handling

17 years agoMerge git://www.denx.de/git/u-boot
Michal Simek [Mon, 10 Sep 2007 22:29:27 +0000 (00:29 +0200)]
Merge git://www.denx.de/git/u-boot

17 years agoMerge with /home/raj/git/u-boot#ads5121_fec_optimize
Wolfgang Denk [Mon, 10 Sep 2007 21:21:16 +0000 (23:21 +0200)]
Merge ... /home/raj/git/u-boot#ads5121_fec_optimize

17 years agoAllocate CPU Architecture Code for STMicroelectronics' ST200.
Sean MCGOOGAN [Mon, 10 Sep 2007 15:55:59 +0000 (16:55 +0100)]
Allocate CPU Architecture Code for STMicroelectronics' ST200.

Signed-off-by: Sean McGoogan <Sean.McGoogan@st.com>
---------------------------------------------------

17 years agoUpdate version to match current state.
Wolfgang Denk [Mon, 10 Sep 2007 18:42:31 +0000 (20:42 +0200)]
Update version to match current state.

Signed-off-by: Wolfgang Denk <wd@denx.de>
17 years ago[MPC512x] Streamline frame handling in the FEC driver
Grzegorz Bernacki [Mon, 10 Sep 2007 15:39:08 +0000 (17:39 +0200)]
[MPC512x] Streamline frame handling in the FEC driver

- convert frame size settings to be derived from a single base
- set frame size to the recommended default value

Signed-off-by: Grzegorz Bernacki <gjb@semihalf.com>
17 years agoRemove compiler warning: target CPU does not support interworking
Kyungmin Park [Mon, 10 Sep 2007 02:34:00 +0000 (11:34 +0900)]
Remove compiler warning: target CPU does not support interworking

Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
17 years agoFix compile error in spc1920 config.
Wolfgang Denk [Sun, 9 Sep 2007 19:21:33 +0000 (21:21 +0200)]
Fix compile error in spc1920 config.

Signed-off-by: Markus Klotzbücher <mk@denx.de>
Signed-off-by: Wolfgang Denk <wd@denx.de>
17 years agoBugfix: make bootm+libfdt compile on boards with no flash
Grant Likely [Fri, 7 Sep 2007 15:25:07 +0000 (09:25 -0600)]
Bugfix: make bootm+libfdt compile on boards with no flash

Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
17 years agoMerge with /home/raj/git/u-boot#440SPe_PCIe_fixes
Wolfgang Denk [Sat, 8 Sep 2007 18:52:57 +0000 (20:52 +0200)]
Merge with /home/raj/git/u-boot#440SPe_PCIe_fixes

17 years agoMerge with /home/raj/git/u-boot#ads5121_fixes
Wolfgang Denk [Sat, 8 Sep 2007 18:45:59 +0000 (20:45 +0200)]
Merge with /home/raj/git/u-boot#ads5121_fixes

17 years ago[PPC440SPe] PCIe environment settings for Katmai and Yucca
Grzegorz Bernacki [Fri, 7 Sep 2007 16:35:37 +0000 (18:35 +0200)]
[PPC440SPe] PCIe environment settings for Katmai and Yucca

- 'pciconfighost' is set by default in order to be able to scan bridges
behind the primary host/PCIe

- 'pciscandelay' env variable is recognized to allow for user-controlled
delay before the PCIe bus enumeration; some peripheral devices require a
significant delay before they can be scanned (e.g. LSI8408E); without the
delay they are not detected

Signed-off-by: Grzegorz Bernacki <gjb@semihalf.com>
17 years ago[PPC440SPe] Improve PCIe configuration space access
Grzegorz Bernacki [Fri, 7 Sep 2007 16:20:23 +0000 (18:20 +0200)]
[PPC440SPe] Improve PCIe configuration space access

- correct configuration space mapping
- correct bus numbering
- better access to config space

Prior to this patch, the 440SPe host/PCIe bridge was able to configure only the
first device on the first bus. We now allow to configure up to 16 buses;
also, scanning for devices behind the PCIe-PCIe bridge is supported, so
peripheral devices farther in hierarchy can be identified.

Signed-off-by: Grzegorz Bernacki <gjb@semihalf.com>
17 years ago[PPC440SPe] Convert machine check exceptions handling
Grzegorz Bernacki [Fri, 7 Sep 2007 15:46:18 +0000 (17:46 +0200)]
[PPC440SPe] Convert machine check exceptions handling

Convert using fixup mechanism to suppressing MCK for the duration of config
read/write transaction: while fixups work fine with the case of a precise
exception, we identified a major drawback with this approach when there's
an imprecise case. In this scenario there is the following race condition:
the fixup is (by design) set to catch the instruction following the one
actually causing the exception; if an interrupt (e.g. decrementer) happens
between those two instructions, the ISR code is executed before the fixup
handler the machine check is no longer protected by the fixup handler as it
appears as within the ISR code. In consequence the fixup approach is being
phased out and replaced with explicit suppressing of MCK during a PCIe
config read/write cycle.

Signed-off-by: Grzegorz Bernacki <gjb@semihalf.com>
17 years agoFix typo in MAKEALL script. v1.3.0-rc1
Wolfgang Denk [Fri, 7 Sep 2007 15:43:36 +0000 (17:43 +0200)]
Fix typo in MAKEALL script.

Signed-off-by: Wolfgang Denk <wd@denx.de>
17 years ago[MPC512x] Proper handling of larger frames in the FEC driver
Grzegorz Bernacki [Fri, 7 Sep 2007 15:09:21 +0000 (17:09 +0200)]
[MPC512x] Proper handling of larger frames in the FEC driver

When frame larger than local RX buffer is received, it is split and handled
by two buffer descriptors. Prior to this patch the FEC driver discarded
contents of a buffer descriptor without the 'LAST' bit set, so the first
part of the frame was lost in case of larger frames. This fix allows to
safely combine the two pieces into the whole frame.

Signed-off-by: Grzegorz Bernacki <gjb@semihalf.com>
17 years ago[MPC512x] Correct fixup relocation
Rafal Jaworowski [Fri, 7 Sep 2007 15:05:36 +0000 (17:05 +0200)]
[MPC512x] Correct fixup relocation

Signed-off-by: Rafal Jaworowski <raj@semihalf.com>
17 years agoUpdate CHANGELOG, minor coding style cleanup.
Wolfgang Denk [Thu, 6 Sep 2007 23:21:25 +0000 (01:21 +0200)]
Update CHANGELOG, minor coding style cleanup.

17 years agoPXA270: Added support for TrizepsIV board.
stefano babic [Thu, 30 Aug 2007 21:01:49 +0000 (23:01 +0200)]
PXA270: Added support for TrizepsIV board.

This patch add support for the Trizeps IV module (520Mhz).

Signed-off-by: Stefano Babic <sbabic@denx.de>
17 years agoPXA270: Add support for multiple serial ports.
stefano babic [Thu, 30 Aug 2007 20:57:04 +0000 (22:57 +0200)]
PXA270: Add support for multiple serial ports.

This patch adds support for multiple serial ports to the PXA target.
FFUART, BTUART and STUART are supported.

Signed-off-by: Stefano Babic <sbabic@denx.de>
17 years agoPXA270: fix compile issue (invalid lvalue)
stefano babic [Thu, 30 Aug 2007 20:48:47 +0000 (22:48 +0200)]
PXA270: fix compile issue (invalid lvalue)

Code is broken for PXA270 due to "invalid lvalue in assignment".

This patch fix it in pxa-regs.h

Signed-off-by: Stefano Babic <sbabic@denx.de>
17 years agoAdd BUILD_DIR support for bios emulator.
Jason Jin [Thu, 30 Aug 2007 10:19:05 +0000 (18:19 +0800)]
Add BUILD_DIR support for bios emulator.

Signed-off-by: Jason Jin <Jason.jin@freescale.com>
17 years ago[MIPS] Remove inline asm string functions
Shinya Kuribayashi [Fri, 31 Aug 2007 05:41:51 +0000 (14:41 +0900)]
[MIPS] Remove inline asm string functions

Stop using inline string functions on MIPS as other ARCHs do so,
since the optimized inline asm versions are not small.

This change is triggered by a following MIPS build error:
common/libcommon.a(exports.o)(.text+0xdc): In function `jumptable_init':
common/exports.c:32: undefined reference to `strcmp'
make: *** [u-boot] Error 1

Signed-off-by: Shinya Kuribayashi <shinya.kuribayashi@necel.com>
17 years ago[MIPS] Update asm string header
Shinya Kuribayashi [Fri, 31 Aug 2007 05:41:45 +0000 (14:41 +0900)]
[MIPS] Update asm string header

This patches contains several bugfixes and cleanups in the latest upstream:

 - Don't include linux/config.h
 - Remove buggy inline version of memscan.
 - Merge with Linux 2.6.11-rc3.
 - Fix undefined reference to strcpy in binfmt_misc caused by gcc 3.4.
 - Goodbye mips64.  31704 lines of code bite the dust.
 - Replace extern inline with static inline.
 - Fix return value of strncpy.
 - Remove a bunch more "$1" clobbers.

Signed-off-by: Shinya Kuribayashi <shinya.kuribayashi@necel.com>
17 years agoMerge with /home/wd/git/u-boot/custodian/u-boot-mpc85xx
Wolfgang Denk [Thu, 6 Sep 2007 22:15:04 +0000 (00:15 +0200)]
Merge ... /home/wd/git/u-boot/custodian/u-boot-mpc85xx

17 years agoFix do_bootm_linux() so that multi-file images with FDT blob boot.
Bartlomiej Sieka [Tue, 4 Sep 2007 15:31:22 +0000 (17:31 +0200)]
Fix do_bootm_linux() so that multi-file images with FDT blob boot.

Fix incorrect blob address calculation in do_bootm_linux() that prevents
booting the kernel from a multi-file image (kernel + initrd + blob).

Also, make minor updates to the U-Boot's output and to the coding style.

Signed-off-by: Bartlomiej Sieka <tur@semihalf.com>
17 years agoMerge with /home/wd/git/u-boot/custodian/u-boot-mpc5xxx
Wolfgang Denk [Thu, 6 Sep 2007 22:13:11 +0000 (00:13 +0200)]
Merge ... /home/wd/git/u-boot/custodian/u-boot-mpc5xxx

17 years agoAdd support for Sil680 IDE controller.
Gary Jennejohn [Fri, 31 Aug 2007 12:29:04 +0000 (14:29 +0200)]
Add support for Sil680 IDE controller.

o add drivers/sil680.c to support the Sil680 IDE-controller.
o drivers/Makefile: add sil680.o.

Signed-off-by: Gary Jennejohn <garyj@denx.de>
17 years agobootm/fdt: Only process the fdt if an fdt address was provided
Grant Likely [Thu, 6 Sep 2007 15:47:40 +0000 (09:47 -0600)]
bootm/fdt: Only process the fdt if an fdt address was provided

Boards with CONFIG_OF_LIBFDT enabled are not able to boot old-style
kernels using the board info structure (instead of passing a device tree)
This change allows the old style booting to be used if the fdt argument
was not passed to 'bootm'.

Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Acked-by: Kim Phillips <kim.phillips@freescale.com>
17 years agoMigrate 5xxx boards from CONFIG_OF_FLAT_TREE to CONFIG_OF_LIBFDT
Grant Likely [Thu, 6 Sep 2007 15:46:23 +0000 (09:46 -0600)]
Migrate 5xxx boards from CONFIG_OF_FLAT_TREE to CONFIG_OF_LIBFDT

Affects boards: icecube (lite5200), jupiter, motionpro, tqm5200

Tested on: lite5200b

Note: the fixup functions have not been moved to a common place.  This
patch is targeted for immediate merging as in solves a build issue, but
the final name/location of the fixups is still subject to debate.  I
propose to merge this now, and move the fixups in the next merge window
to be usable by all targets.

Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
17 years agolibfdt: add convenience function fdt_find_and_setprop()
Grant Likely [Thu, 6 Sep 2007 15:46:17 +0000 (09:46 -0600)]
libfdt: add convenience function fdt_find_and_setprop()

Given the path to a node, fdt_find_and_setprop() allows a property value
to be set directly.

Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
17 years agoFix ULI RTC support on MPC8544 DS
Kumar Gala [Thu, 30 Aug 2007 21:18:18 +0000 (16:18 -0500)]
Fix ULI RTC support on MPC8544 DS

The RTC on the M1575 ULI chipset requires a dummy read before
we are able to talk to the RTC.  We accomplish this by adding a
second memory region to the PHB the ULI is on and read from it.

The second region is added to maintain compatiabilty with Linux's
view of the PCI memory map.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
17 years agoft_board_setup update 85xx/86xx of pci/pcie bus-range property.
Ed Swarthout [Thu, 30 Aug 2007 06:58:48 +0000 (01:58 -0500)]
ft_board_setup update 85xx/86xx of pci/pcie bus-range property.

pcie is now differentiated from pci.  Add 8641 bus-range updates.

Signed-off-by: Ed Swarthout <Ed.Swarthout@freescale.com>
17 years agoMerge with git://www.denx.de/git/u-boot.git
Stefan Roese [Sun, 2 Sep 2007 12:02:19 +0000 (14:02 +0200)]
Merge with git://www.denx.de/git/u-boot.git

17 years agoppc4xx: (Re-)Enable CONFIG_PCI_PNP on AMCC 440EPx Sequoia
Gary Jennejohn [Fri, 31 Aug 2007 13:21:46 +0000 (15:21 +0200)]
ppc4xx: (Re-)Enable CONFIG_PCI_PNP on AMCC 440EPx Sequoia

The 440EPx has a problem when the PCI_CACHE_LINE_SIZE register is
set to non-zero, because it doesn't support MRM (memory-read-
multiple) correctly. We now added the possibility to configure
this register in the board config file, so that the default value
of 8 can be overridden.

Here the details of this patch:

o drivers_pci_auto.c: introduce CFG_PCI_CACHE_LINE_SIZE to allow
  board-specific settings. As an example the sequoia board requires 0.
  Idea from Stefan Roese <sr@denx.de>.
o board/amcc/sequoia/init.S: add a TLB mapping at 0xE8000000 for the
  PCI IO-space. Obtained from Stefan Roese <sr@denx.de>.
o include/configs/sequoia.h: turn CONFIG_PCI_PNP back on and set
  CFG_PCI_CACHE_LINE_SIZE to 0.

Signed-off-by: Gary Jennejohn <garyj@denx.de>
Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoFix TFTP OACK code for short packets.
Wolfgang Denk [Fri, 31 Aug 2007 08:01:51 +0000 (10:01 +0200)]
Fix TFTP OACK code for short packets.

The old code had a loop limit overflow bug which caused a semi-
infinite loop for small packets, because in "i<len-8", "i" was signed,
but "len" was unsigned, and "len-8" became a huge number for small
values of "len".

This is a workaround which replaces broken commit 8f1bc284.

Signed-off-by: Wolfgang Denk <wd@denx.de>
17 years agoBackout commit 8f1bc284 as it causes TFTP to fail.
Wolfgang Denk [Thu, 30 Aug 2007 12:42:15 +0000 (14:42 +0200)]
Backout commit 8f1bc284 as it causes TFTP to fail.

Signed-off-by: Wolfgang Denk <wd@denx.de>
17 years agoRevert "Fix MPC8544DS PCIe3 scsi."
Ed Swarthout [Thu, 30 Aug 2007 07:26:17 +0000 (02:26 -0500)]
Revert "Fix MPC8544DS PCIe3 scsi."

This reverts commit 9468e680.
Commit 16e23c3f5da removing allocation of PCSRBAR is sufficient.

Signed-off-by: Ed Swarthout <Ed.Swarthout@freescale.com>
17 years agotftp: don't implicity trust the format of recevied packets
Grant Likely [Thu, 30 Aug 2007 00:26:24 +0000 (18:26 -0600)]
tftp: don't implicity trust the format of recevied packets

The TFTP OACK code trusts that the incoming packet is formated as
ASCII text and can be processed by string functions. It also has a
loop limit overflow bug where if the packet length is less than 8, it
ends up looping over *all* of memory to find the 'blksize' string.

This patch solves the problem by forcing the packet to be null
terminated and using strstr() to search for the sub string.

Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
17 years agoMerge with git+ssh://gemini/home/wd/git/u-boot/master
Wolfgang Denk [Wed, 29 Aug 2007 14:22:54 +0000 (16:22 +0200)]
Merge ... git+ssh://gemini/home/wd/git/u-boot/master

17 years agosbc8641: remove unused OF_FLAT_TREE_MAX_SIZE
Kim Phillips [Wed, 29 Aug 2007 14:06:05 +0000 (09:06 -0500)]
sbc8641: remove unused OF_FLAT_TREE_MAX_SIZE

this had slipped through the cracks, since the sbc board was added
after I wrote the original patch to remove all these symbols, and
before it was merged.

Signed-off-by: Kim Phillips <kim.phillips@freescale.com>
17 years agoAdd mii_init() prototype
Wolfgang Denk [Wed, 29 Aug 2007 12:05:30 +0000 (14:05 +0200)]
Add mii_init() prototype

to get rid of a *lot* of compiler warnings.

Signed-off-by: Wolfgang Denk <wd@denx.de>
17 years agoDisable network support on cmi_mpc5xx board
Wolfgang Denk [Wed, 29 Aug 2007 11:35:03 +0000 (13:35 +0200)]
Disable network support on cmi_mpc5xx board

..because it caused compiler errors and there seems to be no
board maintainer to take care of this.

Signed-off-by: Wolfgang Denk <wd@denx.de>
17 years agoFix MPC8544DS PCIe3 scsi.
Kumar Gala [Mon, 20 Aug 2007 14:44:00 +0000 (09:44 -0500)]
Fix MPC8544DS PCIe3 scsi.

<ed.swarthout@freescale.com>

The problem is pciauto_setup_device() getting called from fsl_pci_init.c
is allocating memory space it doesn't need.

Signed-off-by: Ed Swarthout <ed.swarthout@freescale.com>
Signed-off-by: Andy Fleming <afleming@freescale.com>
17 years ago8548cds fixes
Ed Swarthout [Tue, 21 Aug 2007 14:38:59 +0000 (09:38 -0500)]
8548cds fixes

Restore CONFIG_EXTRA_ENV_SETTINGS definition which contains the
correct consoledev needed for linux boot.
Standardize on fdt{file,addr} var to hold dtb file name.

Set PCI inbound memory region from CFG_MEMORY_{BUS,PHYS}.

Signed-off-by: Ed Swarthout <Ed.Swarthout@freescale.com>
17 years agoEnable L2 cache for MPC8568MDS board
Haiying Wang [Thu, 23 Aug 2007 19:20:54 +0000 (15:20 -0400)]
Enable L2 cache for MPC8568MDS board

The L2 cache size is 512KB for 8568, print out the correct informaiton.

Signed-off-by: Haiying Wang <Haiying.Wang@freescale.com>
17 years agoRemove the bios emulator binary files from MAI board
Jason Jin [Wed, 22 Aug 2007 09:54:49 +0000 (17:54 +0800)]
Remove the bios emulator binary files from MAI board

Signed-off-by: Jason Jin <Jason.jin@freescale.com>
17 years agosupport board vendor-common makefiles
Kim Phillips [Tue, 21 Aug 2007 22:00:17 +0000 (17:00 -0500)]
support board vendor-common makefiles

if a board/$(VENDOR)/common/Makefile exists, build it.

also add the first such case, board/freescale/common/Makefile, to
handle building board-shared EEPROM, PIXIS, and MDS-PIB code, as
dictated by board configuration.

thusly get rid of alternate build dir errors such as:

FATAL: can't create /work/wd/tmp/u-boot-ppc/board/freescale/mpc8360emds/../common/pq-mds-pib.o: No such file or directory

by putting the common/ mkdir command in its proper place (the common
Makefile). Common bits from existing individual board Makefiles have
been removed.

Signed-off-by: Kim Phillips <kim.phillips@freescale.com>
17 years agoFix: TFTP is not working on little endian systems
stefano babic [Tue, 21 Aug 2007 13:52:33 +0000 (15:52 +0200)]
Fix: TFTP is not working on little endian systems

TFTP does not work anymore after multicast tftp
patch was applied on little endian systems.
This patch fix it.

Signed-off-by: Stefano Babic <sbabic@denx.de>
17 years agoFix MAC address setting in DM9000 driver.
stefano babic [Tue, 21 Aug 2007 13:50:33 +0000 (15:50 +0200)]
Fix MAC address setting in DM9000 driver.

The logic to check if there is a correct MAC address in the DM9000
EEPROM, added in the last patch, is wrong. Now the MAC address is
always taken from the environment, even if a suitable MAC is present
in the EEPROM.

Signed-off-by: Stefano Babic <sbabic@denx.de>
17 years agoMPC5xxx: fix some compiler warnings in USB code
Martin Krause [Tue, 21 Aug 2007 10:40:34 +0000 (12:40 +0200)]
MPC5xxx: fix some compiler warnings in USB code

Fix the following warnings:
- usb.c:xx: warning: function declaration isn't a prototype
- usb_ohci.c:xxx: warning: passing argument 1 of '__fswab32' makes integer
  from pointer wihtout a cast

Signed-off-by: Martin Krause <martin.krase@tqs.de>
17 years agofsl_pci_init - Remove self PCSRBAR allocation
Ed Swarthout [Tue, 21 Aug 2007 04:55:33 +0000 (23:55 -0500)]
fsl_pci_init - Remove self PCSRBAR allocation

CPU physical address space was being wasted by allocating a
PCSRBAR PCI inbound region to it's memory space.

As a rule, PCSRBAR should be left alone since it does not affect
transactions from self and other masters may have changed it.

Signed-off-by: Ed Swarthout <ed.swarthout@freescale.com>
17 years agoFix compiler warning in include/s3c2410.h
Martin Krause [Mon, 20 Aug 2007 11:56:47 +0000 (13:56 +0200)]
Fix compiler warning in include/s3c2410.h

This patch fixes the "type qualifiers ignored on fuction return tpye"
warning for include/s3c2410.h

Signed-off-by: Martin Krause <martin.krause@tqs.de>
17 years agoFix compilation error for omap2420h4_config.
Dirk Behme [Mon, 20 Aug 2007 05:09:05 +0000 (07:09 +0200)]
Fix compilation error for omap2420h4_config.

omap2420h4 switched to cfi, so remove old (already disabled) flash.c
and flash_probe() calls in env_flash.c.

Signed-off-by: Dirk Behme <dirk.behme@gmail.com>
17 years agofdt: remove unused OF_FLAT_TREE_MAX_SIZE references
Kim Phillips [Fri, 10 Aug 2007 19:34:14 +0000 (14:34 -0500)]
fdt: remove unused OF_FLAT_TREE_MAX_SIZE references

and make some minor corrections to the FDT part of the README.

Signed-off-by: Kim Phillips <kim.phillips@freescale.com>
17 years agoMinor coding style cleanup.
Wolfgang Denk [Tue, 28 Aug 2007 23:32:05 +0000 (01:32 +0200)]
Minor coding style cleanup.

Signed-off-by: Wolfgang Denk <wd@denx.de>
17 years agoMerge with /home/wd/git/u-boot/custodian/u-boot-ppc4xx
Wolfgang Denk [Tue, 28 Aug 2007 22:53:51 +0000 (00:53 +0200)]
Merge ... /home/wd/git/u-boot/custodian/u-boot-ppc4xx

17 years ago[UC101] Fix: if no CF in the board, U-Boot resets sometimes.
Heiko Schocher [Tue, 28 Aug 2007 15:40:33 +0000 (17:40 +0200)]
[UC101] Fix: if no CF in the board, U-Boot resets sometimes.

Signed-off-by: Heiko Schocher <hs@denx.de>
17 years agoIDE: - make ide_inb () and ide_outb () "weak", so boards can
Heiko Schocher [Tue, 28 Aug 2007 15:39:14 +0000 (17:39 +0200)]
IDE: - make ide_inb () and ide_outb () "weak", so boards can
          define there own I/O functions.
          (Needed for the pcs440ep board).
        - The default I/O Functions are again 8 Bit accesses.
        - Added CONFIG_CMD_IDE for the pcs440ep Board.

Signed-off-by: Heiko Schocher <hs@denx.de>
17 years agoPOST: limit memory test area to not touch global data anymore
Yuri Tikhonov [Sat, 25 Aug 2007 03:07:16 +0000 (05:07 +0200)]
POST: limit memory test area to not touch global data anymore

As experienced on lwmon5, on some boards the POST memory test can
corrupt the global data buffer (bd). This patch fixes this issue
by checking and limiting this area.

Signed-off-by: Yuri Tikhonov <yur@emcraft.com>
Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Add RTC POST test to lwmon5 board configuration
Stefan Roese [Fri, 24 Aug 2007 13:41:42 +0000 (15:41 +0200)]
ppc4xx: Add RTC POST test to lwmon5 board configuration

Since this RTC POST test is taking quite a while to complete
it's only initiated upon special keypress same as the complete
memory POST.

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Change GPIO signal for watchdog triggering on lwmon5
Stefan Roese [Fri, 24 Aug 2007 13:19:10 +0000 (15:19 +0200)]
ppc4xx: Change GPIO signal for watchdog triggering on lwmon5

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Add support for 2nd I2C EEPROM on lwmon5 board
Stefan Roese [Thu, 23 Aug 2007 09:02:37 +0000 (11:02 +0200)]
ppc4xx: Add support for 2nd I2C EEPROM on lwmon5 board

This patch adds support for the 2nd EEPROM (AT24C128) on the lwmon5
board. Now the "eeprom" command can be used to read/write from/to this
device. Additionally a new command was added "eepromwp" to en-/disable
the write-protect of this 2nd EEPROM.

The 1st EEPROM is not affected by this write-protect command.

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Remove unused option CFG_INIT_RAM_OCM
Stefan Roese [Wed, 22 Aug 2007 06:56:09 +0000 (08:56 +0200)]
ppc4xx: Remove unused option CFG_INIT_RAM_OCM

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoMerge with /home/stefan/git/u-boot/u-boot-ppc4xx
Stefan Roese [Tue, 21 Aug 2007 14:33:33 +0000 (16:33 +0200)]
Merge with /home/stefan/git/u-boot/u-boot-ppc4xx

17 years agoppc4xx: Add matrix kbd support to lwmon5 board (440EPx based)
Stefan Roese [Tue, 21 Aug 2007 14:27:57 +0000 (16:27 +0200)]
ppc4xx: Add matrix kbd support to lwmon5 board (440EPx based)

This patch adds support for the matrix keyboard on the lwmon5 board.
Since the implementation in the dsPCI is kind of compatible with the
"old" lwmon board, most of the code is copied from the lwmon
board directory.

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoFix some build errors.
Wolfgang Denk [Sun, 19 Aug 2007 08:27:34 +0000 (10:27 +0200)]
Fix some build errors.

Signed-off-by: Wolfgang Denk <wd@denx.de>
17 years agoUpdate CHANGELOG.
Wolfgang Denk [Sat, 18 Aug 2007 20:00:38 +0000 (22:00 +0200)]
Update CHANGELOG.

17 years agoMerge with /home/wd/git/u-boot/custodian/u-boot-coldfire
Wolfgang Denk [Sat, 18 Aug 2007 19:56:57 +0000 (21:56 +0200)]
Merge ... /home/wd/git/u-boot/custodian/u-boot-coldfire

17 years agoMerge with /home/wd/git/u-boot/custodian/u-boot-mpc85xx
Wolfgang Denk [Sat, 18 Aug 2007 19:50:01 +0000 (21:50 +0200)]
Merge ... /home/wd/git/u-boot/custodian/u-boot-mpc85xx

17 years agoMerge with /home/wd/git/u-boot/custodian/u-boot-mpc83xx
Wolfgang Denk [Sat, 18 Aug 2007 19:47:33 +0000 (21:47 +0200)]
Merge ... /home/wd/git/u-boot/custodian/u-boot-mpc83xx

17 years agoMerge with git+ssh://gemini_vpn/home/wd/git/u-boot/master
Wolfgang Denk [Sat, 18 Aug 2007 19:42:30 +0000 (21:42 +0200)]
Merge ... git+ssh://gemini_vpn/home/wd/git/u-boot/master

17 years agolib_ppc: make board_add_ram_info weak
Kim Phillips [Fri, 17 Aug 2007 03:52:39 +0000 (22:52 -0500)]
lib_ppc: make board_add_ram_info weak

platforms wishing to display RAM diagnostics in addition to size,
can do so, on one line, in their own board_add_ram_info()
implementation.

this consequently eliminates CONFIG_ADD_RAM_INFO.

Thanks to Stefan for the hint.

Signed-off-by: Kim Phillips <kim.phillips@freescale.com>
17 years agoPCI_READ_VIA_DWORD_OP: Fix *val uninitialized bug
Shinya Kuribayashi [Fri, 17 Aug 2007 03:43:44 +0000 (12:43 +0900)]
PCI_READ_VIA_DWORD_OP: Fix *val uninitialized bug

This patch has been sent on:
- 6 Jun 2007

Many users of PCI config read routines tend to ignore the function
ret value, and are only concerned about the contents of *val. Based
on this, pci_hose_read_config_{byte,word}_via_dword should initialize
the *val on dword read error.

Without this fix, for example, we'll go on scanning bus with vendor or
header_type uninitialized. This brings many unnecessary config trials.

Signed-off-by: Shinya Kuribayashi <shinya.kuribayashi@necel.com>