Commit 3c0659b "ARM: Avoid compiler optimization for readb, writeb
and friends." introduced I/O accessors with memory barriers.
Unfortunately the new write*() accessors introduced a bug:
The problem is that the argument "v" gets evaluated twice. This
breaks code like used here (from "drivers/net/dnet.c"):
for (i = 0; i < wrsz; i++)
writel(*bufp++, &dnet->regs->TX_DATA_FIFO);
Use auxiliary variables to avoid such problems.
Signed-off-by: Wolfgang Denk <wd@denx.de>
Cc: Albert Aribaud <albert.aribaud@free.fr>
Cc: Alexander Holler <holler@ahsoftware.de>
Cc: Dirk Behme <dirk.behme@googlemail.com>
gcc 4.5.1 seems to ignore (at least some) volatile definitions,
avoid that as done in the kernel.
Reading C99 6.7.3 8 and the comment 114) there, I think it is a bug of that
gcc version to ignore the volatile type qualifier used e.g. in __arch_getl().
Anyway, using a definition as in the kernel headers avoids such optimizations when
gcc 4.5.1 is used.
Maybe the headers as used in the current linux-kernel should be used,
but to avoid large changes, I've just added a small change to the current headers.
Signed-off-by: Alexander Holler <holler@ahsoftware.de>
Signed-off-by: Dirk Behme <dirk.behme@googlemail.com>
Signed-off-by: Wolfgang Denk <wd@denx.de>
Cc: Alessandro Rubini <rubini-list@gnudd.com>
Tested-by: Thomas Weber <weber@corscience.de>
Acked-by: Alexander Holler <holler@ahsoftware.de>
Tested-by: Alexander Holler <holler@ahsoftware.de>
These functions are undefined on ARM when using __io. These are the commonly
used versions and can be redefined.
Signed-off-by: Marek Vasut <marek.vasut@gmail.com>
This helps to clean up the include/ directory so that it only contains
non-architecture-specific headers and also matches Linux's directory
layout which many U-Boot developers are already familiar with.
Signed-off-by: Peter Tyser <ptyser@xes-inc.com>
Some Freescale's processors of different architecture
have the same peripheral (eSDHC controller in PowerPC
and i.MX51). This patch adds accessors for
the internal registers of the SOCs, as already
implemented in the PowerPC architecture.
Signed-off-by: Stefano Babic <sbabic@denx.de>
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>
Also add CONFIG_PHYS_64BIT on powerpc to deal with 32-bit ppc's
that have larger physical addresses like 44x, 85xx, and 86xx.
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
map_physmem() returns a virtual address which can be used to access a
given physical address without involving the cache. unmap_physmem()
should be called when the virtual address returned by map_physmem() is
no longer needed.
This patch adds a stub implementation which simply returns the
physical address cast to a uchar * for all architectures except AVR32,
which converts the physical address to an uncached virtual mapping.
unmap_physmem() is a no-op on all architectures, but if any
architecture needs to do such mappings through the TLB, this is the
hook where those TLB entries can be invalidated.
Signed-off-by: Haavard Skinnemoen <hskinnemoen@atmel.com>
Some CPUs like PPC, BLACKFIN need sync() to ensure cfi flash write command
is fully finished. The sync() is defined in each CPU's io.h file. For
those CPUs which do not need sync for now, a dummy sync() is defined in
their io.h as well.
Signed-off-by: Haiying Wang <Haiying.Wang@freescale.com>
add support for the TI OMAP2420 processor and its H4 reference
board
* Patch by Christian Pellegrin, 24 Sep 2004:
Added support for NE2000 compatible (DP8390, DP83902) NICs.
* Patch by Murray Jensen, 24 Jun 2003:
- make sure to use only U-boot provided header files
- fix problems with ".rodata.str1.4" section as used by GCC-3.x
the HSYNC/VSYNC. Requires new CPLD code (Version 101 for Rev. 100
boards, version 153 for Rev. 200 boards).
* Patch by Vladimir Gurevich, 12 Mar 2003:
Fix relocation problem of statically initialized string pointers
in common/cmd_pci.c
* Patch by Kai-Uwe Blöm, 12 Mar 2003:
Cleanup & bug fixes for JFFS2 code:
- the memory mangement was broken. It caused havoc on malloc by
writing beyond the block boundaries.
- the length calculation for files was wrong, sometimes resulting
in short file reads.
- data copying now optionally takes fragment version numbers into
account, to avoid copying from older data.
See doc/README.JFFS2 for details.
Wolfgang Denk
Alexander Holler
Marek Vasut
Terry Lv
Peter Tyser
Stefano Babic
Kumar Gala
Marcel Ziswiler
Haavard Skinnemoen
Haiying Wang
Wolfgang Denk
wdenk