Kumar Gala [Wed, 27 Feb 2008 22:30:47 +0000 (16:30 -0600)]
85xx: Don't icbi when unlocking the cache
There is no reason to icbi when invalidating the temporary stack in
the d-cache. Its impossible on e500 to have the i-cache contain
any addresses in the temp stack and it can be problematic in generating
transactions on the bus to non-valid addresses.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Andy Fleming [Wed, 27 Feb 2008 20:29:58 +0000 (14:29 -0600)]
Invalidate INIT_RAM TLB mappings
Commit 0db37dc... (and some others) changed the INIT_RAM TLB
mappings to be unguarded. This collided with an existing "bug"
where the mappings for the INIT_RAM were being kept around.
This meant that speculative loads to those addresses were
succeeding in the TLB, and going out to the bus, where they
were causing an exception (there's nothing at that address). The
Flash code was coincidentally causing such a speculative load.
Rather than go back to mapping the INIT RAM as guarded, we fix
it so that the entries for the INIT_RAM are invalidated. Thus
the speculative loads will fail in the TLB, and have no effect.
Signed-off-by: Andy Fleming <afleming@freescale.com>
Remove the __STRICT_ANSI__ check from the __u64/__s64 declaration on 32bit targets.
----------------------------------------------------------------
Olaf Hering [Wed, 17 Oct 2007 06:27:13 +0000 (23:27 -0700)]
Remove the __STRICT_ANSI__ check from the __u64/__s64 declaration on
32bit targets.
GCC can be made to warn about usage of long long types with ISO C90
(-ansi), but only with -pedantic. You can write this in a way that even
then it doesn't cause warnings, namely by:
#ifdef __GNUC__
__extension__ typedef __signed__ long long __s64;
__extension__ typedef unsigned long long __u64;
#endif
The __extension__ keyword in front of this switches off any pedantic
warnings for this expression.
Signed-off-by: Olaf Hering <olh@suse.de> Cc: <linux-arch@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
----------------------------------------------------------------
Marcel Moolenaar [Fri, 22 Feb 2008 18:48:07 +0000 (10:48 -0800)]
make define2mk.sed work on FreeBSD
In the thread "[1.3.2-rc1] MPC8548CDS/MPC8555CDS configs fails to link",
the define2mk.sed script was identified as the source of the link
failure on FreeBSD. The problem is that sed(1) does not always support
the '+' operator. It isn't on FreeBSD. The attach patch implements the
equivalent, using the '*' operator instead and should work everywhere.
Larry Johnson [Thu, 21 Feb 2008 18:58:16 +0000 (13:58 -0500)]
LM75 bug fix for negative temperatures
When the LM75 temperature sensor measures a temperature below 0 C, the
current driver does not perform sign extension, so the result returned is
256 C too high. This patch fixes the problem.
Yuri Tikhonov [Thu, 21 Feb 2008 13:23:42 +0000 (14:23 +0100)]
lwmon5: enable hardware watchdog
Some boards (e.g. lwmon5) may use rather small watchdog intervals, so
causing it to reboot the board if U-Boot does a long busy-wait with
udelay(). Thus, for these boards we have to restart WD more
frequently.
This patch splits the busy-wait udelay() into smaller, predefined,
intervals, so that the watchdog timer may be resetted with the
configurable (CONFIG_WD_PERIOD) interval.
ppc4xx: Support for ATI Radeon 9200 card on sequoia
Adds configuration option for ATI Radeon 9200 card
support to sequoia config file. If CONFIG_VIDEO
is enabled, TEXT_BASE should be changed to 0xFFF80000.
Kumar Gala [Mon, 18 Feb 2008 14:09:37 +0000 (08:09 -0600)]
ppc: Allow boards to specify effective amount of memory
For historical reasons we limited the stack to 256M because some boards
could only map that much via BATS. However newer boards are capable of
mapping more memory (for example 85xx is capable of doing up to 2G).
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Mike Frysinger [Mon, 18 Feb 2008 10:10:07 +0000 (05:10 -0500)]
error check autoconf.mk generation
If any of the steps for generating autoconf.mk fail currently, they go
unnoticed. To fix, we can simply add 'set -e' to the long list of commands.
This is simpler and more robust than placing '|| exit $$?' after every line.
Stefano Babic [Mon, 18 Feb 2008 07:03:51 +0000 (08:03 +0100)]
Fix bug in dependency checking
By adding VERSION_FILE to the PHONY targets the script
/tools/setlocalversion is always called and version_autogenerated.h
is replaced only if the script find a modified source file.
Mike Frysinger [Sat, 16 Feb 2008 07:40:18 +0000 (02:40 -0500)]
easylogo: clean up some more and add -r (rgb) support
Michael Hennerich added support for outputting an image in RGB format rather
than forcing YUYV all the time. This makes obvious sense if the display you
have takes RGB input rather than YUYV.
Rather than hack in support for options, I've converted it to use getopt and
cleaned up the argument parsing in the process.
Signed-off-by: Michael Hennerich <michael.hennerich@analog.com> Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Mike Frysinger [Sat, 16 Feb 2008 07:12:37 +0000 (02:12 -0500)]
Makefile: add target for $(LDSCRIPT)
If the $(LDSCRIPT) does not exist (normally it's board/$(BOARD)/u-boot.lds),
then change into the board directory and try and create it. This allows you
to generate the linker script on the fly based upon board defines (like the
Blackfin boards do).
There should be no regressions due to this change as the normal case is to
already have a u-boot.lds file. If that's the case, then there's nothing to
generate, and so make will always exit. The fix here is that if the linker
script does not exist, the implicit rules take over and attempt to guess how
to generate the file.
Do not use uninitialized cmd_reset; issue both AMD and Intel reset
commands instead
From a short test, it looks like AMD-style flash roms treat *any* unknown
command write as a reset, at least when in CFI Query mode, so issuing the
Intel reset command to AMD-style flashs seems safe (from the small sample I
have), plus the 3-cycle magic sequence should kick the state machine into
the right state even without a reset command. Since the AMD-style flashs
require the unlock sequence for real operation, I chose to try the AMD reset
command first, so that Intel flashs do no see an invalid command prior to
the CFI query.
I have tested the patch on AM29LV320-style flashs from Fujitsu and Macronix,
plus Intel StrataFlash.
Signed-off-by: Michael Schwingen <michael@schwingen.org> Signed-off-by: Stefan Roese <sr@denx.de>
Yuri Tikhonov [Thu, 21 Feb 2008 10:06:07 +0000 (11:06 +0100)]
Fix CPU POST test failure
The CPU POST test code (run from cpu_post_exec_31()) doesn't follow the
ABI carefully, at least the CR3, CR4, and CR5 fields of CR are clobbered
by it. The gcc-4.2 with its more aggressive optimization exposes this fact.
This patch just saves the CR value before running the test code, so allowing
it to do anything it wants with CR.
Jon Loeliger [Wed, 20 Feb 2008 20:22:26 +0000 (14:22 -0600)]
86xx: Fix GUR PCI config registers properly.
Back in commit 975a083a5ef785c414b35f9c5b8ae25b26b41524 where
I tried to "8610HPCD: Fix typos in two PCI setup registers", I
botched it due to not realizing that 8610 and 8641 had different
Global Utility Register defintions, one of which was like 85xx,
and the other wasn't. Correct this problem by introducing two
symbols, one for each 86xx SoC, but neither of which is named
anything like 85xx.
My bad. Lovely Wednesday with git bisect. You know.
Jon Loeliger [Wed, 20 Feb 2008 18:24:11 +0000 (12:24 -0600)]
8610HPCD: Don't use VIDEO/CFB_CONSOLE by default.
Without an actual supported video card hooked up, enabling
the CONFIG_VIDEO by default just makes it look broken by
routing all console output to the video card. Don't.
Jon Loeliger [Tue, 19 Feb 2008 18:31:08 +0000 (12:31 -0600)]
8610HPCD: Fix typos in two PCI setup registers.
The two symbols MPC86xx_PORDEVSR_IO_SEL and MPC86xx_PORBMSR_HA
were erroneously present as 85xx names and values, leftover from
the clone wars. Fix this by removing the 85xx cruft from the
86xx codebase.
Jon Loeliger [Mon, 18 Feb 2008 20:01:56 +0000 (14:01 -0600)]
86xx: Convert sbc8641d to use libfdt.
This is the proper fix for a missing closing brace in the function
ft_cpu_setup() noticed by joe.hamman <at> embeddedspecialties.com.
The ft_cpu_setup() function in mpc8641hpcn.c should have been
removed earlier as it was under the obsolete CONFIG_OF_FLAT_TREE,
but was missed. Only, the sbc8641d was nominally still using it.
It all got ripped out, and the funcality that was in ft_board_setup()
was refactored to remove the CPU portions into the new file
cpu/mpc86xx/fdt.c instead. Make sbc8641d use this now.
Based loosely on an original patch from joe.hamman@embeddedspecialties.com
Mike Nuss [Wed, 6 Feb 2008 16:10:11 +0000 (11:10 -0500)]
PPC440EPx: Optionally enable second I2C bus
The option CONFIG_I2C_MULTI_BUS does not have any effect on Sequoia, the
PPC440EPx reference platform, because IIC1 is never enabled. Add Sequoia board
code to turn on IIC1 if CONFIG_I2C_MULTI_BUS is selected.
Signed-off-by: Mike Nuss <mike@terascala.com> Cc: Stefan Roese <sr@denx.de>
Niklaus Giger [Tue, 5 Feb 2008 10:31:28 +0000 (11:31 +0100)]
ppc4xx: HCU4/5. Cleanups
- Fix some coding style violations.
- Use in/out_u16/32 where appropriate.
- Use register names from ppc405.h.
- Fix trace useage for Lauterbach.
- Remove obsolete generation HCU2.
- Renamed fixed_hcu4_sdram to init_ppc405_sdram.
If the length of the memory address range passed to the "mtest" command is
not of the form 2^x - 1, not all address lines are tested. This bug is
inherited from the original software at
http://www.netrino.com/Embedded-Systems/How-To/Memory-Test-Suite-C. Fix
this.
Mike Frysinger [Mon, 4 Feb 2008 22:44:23 +0000 (17:44 -0500)]
only update version header as needed
Constantly rebuilding the version header will force useless relinking, so we
simply need to compare the new header with the existing one before updating
it.
Wolfgang Denk [Thu, 14 Feb 2008 21:43:22 +0000 (22:43 +0100)]
PPC: Use r2 instead of r29 as global data pointer
R29 was an unlucky choice as with recent toolchains (gcc-4.2.x) gcc
will refuse to use load/store multiple insns; instead, it issues a
list of simple load/store instructions upon function entry and exit,
resulting in bigger code size, which in turn makes the build for a
few boards fail.
common/miiphyutil.c: Cleanup MII_DEBUG and debug()
Current MII_DEBUG is confusing in two ways. One is useless define-then-
undef at the top of the file. The other is there is only one debug() in
this file, and that doesn't seem worthwhile to bother having MII_DEBUG.
While there are many useful printf()/puts() debug codes, but they are for
DEBUG, not for MII_DEBUG.
This patch tries to put them all together into MII_DEBUG and debug().
Kyungmin Park [Thu, 17 Jan 2008 07:43:25 +0000 (16:43 +0900)]
OneNAND Initial Program Loader (IPL) support
This patch enables the OneNAND boot within U-Boot.
Before this work, we used another OneNAND IPL called X-Loader based
on open source. With this work, we can build the oneboot.bin image
without other program.
The build sequence is simple.
First, it compiles the u-boot.bin
Second, it compiles OneNAND IPL
Finally, it becomes the oneboot.bin from OneNAND IPL and u-boot.bin
The mechanism is similar with NAND boot except it boots from itself.
Another thing is that you can only use the OneNAND IPL only to work
other bootloader such as RedBoot and so on.
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Andy Fleming [Wed, 16 Jan 2008 19:06:59 +0000 (13:06 -0600)]
Fix CONFIG_MMC usage in fat code
A #if statement in fat.c depended on CONFIG_MMC, instead of
defined(CONFIG_MMC). This meant CONFIG_MMC needed to be defined
as "1" rather than just defined. Now it's better.
Signed-off-by: Andy Fleming <afleming@freescale.com>
projects / u-boot / log