According to the PPC reference implementation the udelay() function is
responsible for resetting the watchdog timer as frequently as needed.
Most other architectures do not meet that requirement, so long-running
operations might result in a watchdog reset.
This patch adds a generic udelay() function which takes care of
resetting the watchdog before calling an architecture-specific
__udelay().
Signed-off-by: Ingo van Lil <inguin@gmx.de>
Commits
02f99901ed52d61227b6
introduced a regression where platform-specific ffs/fls implementations
were defined away. This patch corrects that by using PLATFORM_xxx
instead of the name itself.
Signed-off-by: Simon Kagstrom <simon.kagstrom@netinsight.net>
Acked-by: Kumar Gala <galak@kernel.crashing.org>
Acked-by: Stefan Roese <sr@denx.de>
__set_bit and __clear_bit are defined in ubifs.h as well as in
asm/include/bitops.h for some architectures. This patch moves
the generic implementation to include/linux/bitops.h and uses
that unless it's defined by the architecture.
Signed-off-by: Simon Kagstrom <simon.kagstrom@netinsight.net>
If the memory layout pushes the stack out of the default DCPLB coverage,
the exception handler may trigger a double fault by trying to push onto
the uncovered stack. So handle the exception stack similar to the kernel
by using the top of the scratch pad SRAM.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
The default console size indirectly applies to length of env vars, so a
smaller length makes it hard to pass longer command lines to kernels.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Since the Blackfin ABI favors higher scratch registers by default, use the
last scratch register (P3) for global data rather than the first (P5).
This allows the compiler's register allocator to use higher number scratch
P registers, which in turn better matches the Blackfin instruction set,
which reduces the size of U-Boot by more than 1024 bytes...
Signed-off-by: Robin Getz <robin.getz@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Part of the mini Blackfin ABI with operating systems is that they can use
0x4f0-0x4f8 to pass log buffers to/from bootloaders. So add support to
U-Boot for reading the log buffer.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
This patch use blackfin errno.h implementation which
correspond Linux kernel one.
MIPS implemetation is different that's why I keep it.
I removed ppc_error_no.h from Marvell boards which
was the same too.
I have got ack from ppc40x, blackfin, arm, coldfire and avr custodians.
Acked-by: Stefan Roese <sr@denx.de>
Signed-off-by: Michal Simek <monstr@monstr.eu>
The Blackfin port was using asm/blackfin-config-{pre,post}.h to setup
common Blackfin board defines. The common method now is to use config.h,
so convert blackfin-config-post.h to that. Rename the still Blackfin
specific blackfin-config-pre.h to config-pre.h so the naming conventions
at least line up.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Newer Blackfin parts can an on-chip ROM that can boot LDRs over SPI flashes,
so add a new 'spibootldr' command to take advantage of it.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
A missing set of parenthesis caused the silicon revision to apply only to
the BF533 and not the BF531/BF532 variants.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
When dropping jump block support, the assumption was that all bootroms
supported entry point redirection via the EVT1 register. Unfortunately,
this turned out to be incorrect for the oldest Blackfin parts (BF533-0.2
and older and BF561). No one really noticed earlier because these parts
usually are booted by bypassing the bootrom entirely, and older BF533
parts are not supported at all (too many anomalies).
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Since the Blackfin timer code requires HZ to be 1000, barf on any board
that tries to use a different value.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Use the common net eth functions to setup the env/global data with the MAC
address, and properly handle the case where CONFIG_SYS_NO_FLASH is defined.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
The SPORT_TX registers cannot be read (the hardware will trigger an error),
so drop the read helper macros.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
We have common defines that we duplicate in various ways. Having an
arch specific config.h gives us a common location for those defines.
Eventually we should be able to replace this when we have proper
Kconfig support.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Move the default SPI CS that we boot from into common code so that it can
be used in other SPI drivers and environment settings.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
This is a port of the Linux Blackfin on-chip SDH driver to U-Boot.
Signed-off-by: Cliff Cai <cliff.cai@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Knowing the booting source of the part is useful, especially when the part
can switch dynamically between sources.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Set the default CONFIG_ENV_SPI_CS value to match the SPI CS that is used by
the Blackfin on-chip bootrom to boot out of SPI flash.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
This is a port of the Linux Blackfin on-chip ATAPI driver to U-Boot.
Signed-off-by: Sonic Zhang <Sonic.Zhang@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
The performance difference from doing an 8 bit DMA memcpy vs an optimized
core memcpy can be pretty big when you add in the overhead of setting up the
MDMA registers, cache flushes, etc... So only use dma_memcpy() when we
actually require it.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Take the cache flush functions from the kernel as they use hardware loops in
order to get optimal performance.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Calculating the clocks requires a bit of calls to gcc math functions, so
cache the values after the first run since they'll most likely never
change once U-Boot is up and running.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Our dcache invalidate function doesn't just invalidate, it also flushes.
So rename the function accordingly and fix the dma_memcpy() function so it
doesn't inadvertently corrupt the data destination.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
virt_to_phys() returns the physical address given a virtual. In most
cases this will be just the input value as the vast majority of
systems run in a 1:1 mode.
However in systems that are not running this way it should report the
physical address or ~0 if no mapping exists for the given virtual
address.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
