Fix for USB sticks not working on ARM while using GCC 4.x compilers
The GCC-compiler makes an optimisation error while optimising the routine
usb_set_maxpacket(). This should be fixed in the compiler in the first place,
but there lots of compilers out there that makes this error, that it is
probably wiser to workaround it in U-boot itself.
What happens is that the register r3 is used as loop-counter 'i', but gets
overwritten later on. From there it starts using register r3 for several other
things and the assembler code is becoming a big mess. This is clearly a compiler bug.
This error occurs on at least several versions of Code Sourcery Lite compilers
for ARM. Like the Edition 2008q1, and 2008q3, It has also been seen on other
compilers, while compiling for armv4t, or armv5te with Os, O1 and O2.
We work around it by splitting up this routine in 2 parts, and making sure that
the split out part is NOT inlined any longer. This will make GCC spit out assembler
that do not show this problem. Another possibility is to adapt the Makefile to stop
optimisation for the complete file. I think this solution is nicer.
Signed-off-by: Remy Bohmer <linux@bohmer.net> Signed-off-by: Markus Klotzbuecher <mk@denx.de>
This patch is an attempt to implement autoprobing for the Lime
presence on the bus.
Configure GPCM for Lime CS2 and try to access chip ID registers.
Second read atempt delivers register values if the chip is present.
powerpc: Fix bootm to boot up again with a Ramdisk
Commit 2a1a2cb6 didnt remove the dummy mem reservation in fdt_chosen,
and this stopped Linux from booting with a Ramdisk. This patch fixes
this, by deleting the useless dummy mem reservation.
When booting with a Ramdisk, a fix offset FDT_RAMDISK_OVERHEAD is now
added to of_size, so we dont need anymore a dummy mem reservation.
I measured the value of FDT_RAMDISK_OVERHEAD on a MPC8270 based
system (=0x44 bytes) and rounded it up to 0x80).
Signed-off-by: Heiko Schocher <hs@denx.de> Acked-by: Kumar Gala <galak@kernel.crashing.org>
Anton Vorontsov [Fri, 27 Jun 2008 19:04:20 +0000 (23:04 +0400)]
fsl_elbc_nand: ecclayout cleanups
This patch deletes oobavail assignments, they're calculated by the nand
core code in nand_scan_tail, plus current oobavail values are wrong for
the LP NANDs.
Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com> Signed-off-by: Scott Wood <scottwood@freescale.com>
Anton Vorontsov [Fri, 27 Jun 2008 19:04:13 +0000 (23:04 +0400)]
fsl_elbc_nand: implement support for flash-based BBT
This patch implements support for flash-based BBT for chips working
through ELBC NAND controller, so that NAND core will not have to re-scan
for bad blocks on every boot.
Because ELBC controller may provide HW-generated ECCs we should adjust
bbt pattern and bbt version positions in the OOB free area.
Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com> Signed-off-by: Scott Wood <scottwood@freescale.com>
Anton Vorontsov [Fri, 27 Jun 2008 19:04:04 +0000 (23:04 +0400)]
fsl_elbc_nand: fix OOB workability for large page NAND chips
For large page chips, nand_bbt is looking into OOB area, and checking
for "0xff 0xff" pattern at OOB offset 0. That is, two bytes should be
reserved for bbt means.
But ELBC driver is specifying ecclayout so that oobfree area starts at
offset 1, so only one byte left for the bbt purposes.
This causes problems with any OOB users, namely JFFS2: after first mount
JFFS2 will fill all OOBs with "erased marker", so OOBs will contain:
And on the next boot, NAND core will rescan for bad blocks, then will
see "0xff 0x19" pattern, and will mark all blocks as bad ones.
To fix the issue we should implement our own bad block pattern: just one
byte at OOB start. Though, this will work only for x8 chips. For x16
chips two bytes must be checked. Since ELBC driver does not support x16
NANDs (yet), we're safe for now.
Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com> Signed-off-by: Scott Wood <scottwood@freescale.com>
This patch prevents linker error on AT91RM9200 boards, if
CONFIG_CMD_I2_TREE is set.
It implements i2c_set_bus_speed and i2c_get_bus_speed as a dummy function.
Andrew Dyer [Fri, 12 Sep 2008 00:20:46 +0000 (02:20 +0200)]
arm920t fix constant error in start.S
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@righthandtech.com>
Magnus Lilja [Fri, 29 Aug 2008 08:36:18 +0000 (10:36 +0200)]
i.MX31: Make the SPI bus and chip select configurable for MC13783
The i.MX31 has three SPI buses and each bus has several chip selects
and the MC13783 chip can be connected to any of these. The current
RTC driver for MC13783 is hardcoded for CSPI2/SS2.
This patch makes make MC13783 SPI bus and chip select configurable
via CONFIG_MC13783_SPI_BUS and CONFIG_MC13783_SPI_CS.
Signed-off-by: Magnus Lilja <lilja.magnus@gmail.com>
Magnus Lilja [Fri, 29 Aug 2008 08:36:17 +0000 (10:36 +0200)]
i.MX31: Add reset_timer() and modify get_timer_masked().
This patch adds the reset_timer() function (needed by nand_base.c) and
modifies the get_timer_masked() to work in the same way as the omap24xx
function.
Signed-off-by: Magnus Lilja <lilja.magnus@gmail.com>
Hugo Villeneuve [Sat, 30 Aug 2008 21:06:55 +0000 (17:06 -0400)]
ARM DaVinci: Fix broken HW ECC for large page NAND.
Based on original patch by Bernard Blackham <bernard@largestprime.net>
U-boot's HW ECC support for large page NAND on Davinci is completely
broken. Some kernels, such as the 2.6.10 one supported by
MontaVista for DaVinci, rely upon this broken behaviour as they
share the same code for ECCs. In the existing scheme, error
detection *might* work on large page, but error correction
definitely does not. Small page ECC correction works, but the
format is not compatible with the mainline git kernel.
This patch adds ECC code that matches what is currently in the
Davinci git repository (since NAND support was added in 2.6.24).
This makes the ECC and OOB layout written by u-boot compatible with
Linux for both small page and large page devices and fixes ECC
correction for large page devices.
The old behaviour can be restored by defining the macro
CFG_DAVINCI_BROKEN_ECC, which is undefined by default.
Signed-off-by: Hugo Villeneuve <hugo.villeneuve@lyrtech.com> Acked-by: Sergey Kubushyn <ksi@koi8.net> Signed-off-by: Scott Wood <scottwood@freescale.com>
Petri Lehtinen [Wed, 10 Sep 2008 06:43:49 +0000 (09:43 +0300)]
autoscr: Fix one-character lines and non-newline terminated scripts
When not using hush, the autoscr command now executes lines that are
only one character long. It also runs the last line of scripts even if
it does not end in a newline.
Signed-off-by: Petri Lehtinen <petri.lehtinen@inoi.fi>
Andrew Klossner [Thu, 21 Aug 2008 14:12:26 +0000 (07:12 -0700)]
Fix printf errors under -DDEBUG
Fix printf format-string/arg mismatches under -DDEBUG.
These warnings occur with DEBUG defined for a platform using
cpu/mpc85xx. Users of other architectures can unearth similar
problems by adding the line "CFLAGS += -DDEBUG=1" in config.mk right
after "CFLAGS += $(call cc-option,-fno-stack-protector)".
Signed-off-by: Andrew Klossner <andrew@cesa.opbu.xerox.com> Signed-off-by: Andy Fleming <afleming@freescale.com>
Remy Bohmer [Wed, 20 Aug 2008 09:22:02 +0000 (11:22 +0200)]
Make usb-stop() safe to call multiple times in a row.
A recent commit (936897d4d1365452bbbdf8430db5e7769ef08d38)
enabled the usb_stop() command in common/cmd_bootm.c which was
not enabled for some time, because no board did actually set the
CFG_CMD_USB flag. So, now the usb_stop() is executed before
loading the linux kernel.
However, the usb_ohci driver hangs up (at least on AT91SAM) if the
driver is stopped twice (e.g. the peripheral clock is stopped on AT91).
If some other piece of code calls usb_stop() before the bootm command,
this command will hangup the system during boot.
(usb start and stop is typically used while booting from usb memory stick)
But, stopping the usb stack twice is useless anyway, and a flag already
existed that kept track on the usb_init()/usb_stop() calls.
So, we now check if the usb stack is really started before we stop it.
This problem is now fixed in both the upper as low-level layer.
Signed-off-by: Remy Bohmer <linux@bohmer.net> Acked-by: Markus Klotzbuecher <mk@denx.de>
FIT: add ability to check hashes of all images in FIT, improve output
- add function fit_all_image_check_hashes() that verifies if all
hashes of all images in the FIT are valid
- improve output of fit_image_check_hashes() when the hash check fails
Wolfgang Denk [Tue, 9 Sep 2008 13:45:18 +0000 (15:45 +0200)]
Makefile: compile and link each module just once
Several source files need to be compiled and linked when one or more
config options are selected. To allow for easy selection in the
Makefiles yet to avoild multiple compilation (which costs build time)
and especially multiple linking (which causes errors), we use
"COBJS = $(sort COBJS-y)" which eliminates duplicates.
By courtesy of Detlev Zundel who suggested this approach.
projects / u-boot / log