Add data cache support for ARM V7 systems. Used cache flush
functions from linux:arch/arm/mm/cache-v7.S developed from
Catalin Marinas.
Enable "cache" command on Beagle board and test performance.
Test 1: Loading 127 MB of data from NAND flash into RAM:
Instr. Cache off on on
Data Cache off off on
--------------------------------------------------
Beagle (Cortex A8) 116s 106s 30.3s = x 3.8
Test 2: uncompressing a gzipped image from RAM to RAM
(size compressed: 6.5 MiB, uncompressed: 35 MiB):
Instr. Cache off on on
Data Cache off off on
--------------------------------------------------
Beagle (Cortex A8) 1.84s 1.64s 0.12s = x 15.3
Portions of this work were supported by funding from
the CE Linux Forum.
Signed-off-by: Heiko Schocher <hs@denx.de>
Reviewed-by: Ben Gardiner<bengardiner@nanometrics.ca>
Upcoming Beagle and Overo revisions use POP memory with 256MB or 512MB
per bank. This patches uses the SDRC settings from x-load or the config
header to set up timing properly.
Signed-off-by: Steve Sakoman <steve@sakoman.com>
Signed-off-by: Sandeep Paulraj <s-paulraj@ti.com>
Using the reset command on OMAP36XX/37XX and OMAP4 caused a hang. This
patch uses the reset bit appropriate for each CPU architecture.
Signed-off-by: Steve Sakoman <steve@sakoman.com>
Signed-off-by: Sandeep Paulraj <s-paulraj@ti.com>
The workarounds for errata 621766 and 725233 should only be applied
on affected Cortex-A8 revisions. Recent chips use r3px cores where
these have been fixed.
Signed-off-by: Mans Rullgard <mans@mansr.com>
Signed-off-by: Steve Sakoman <steve@sakoman.com>
Signed-off-by: Sandeep Paulraj <s-paulraj@ti.com>
This function consists entirely of inline asm statements, so writing
it directly in a .S file is simpler. Additionally, the inline asm is
not safe as is, since registers are not guaranteed to be preserved
between asm() statements.
Signed-off-by: Mans Rullgard <mans@mansr.com>
Signed-off-by: Steve Sakoman <steve@sakoman.com>
Signed-off-by: Sandeep Paulraj <s-paulraj@ti.com>
On OMAP34xx ES1.0, the L2 enable bit can only be set in secure mode,
so an SMC call to the ROM monitor is required. On later versions,
and on newer devices, this bit is banked and we can set it directly.
The code checked only the ES revision of the chip, and hence incorrectly
used the ROM call on ES1.0 versions of other devices.
This patch adds a check for chip family as well as revision, and also
removes some code duplication between the enable and disable functions.
Signed-off-by: Mans Rullgard <mans@mansr.com>
Signed-off-by: Steve Sakoman <steve@sakoman.com>
Signed-off-by: Sandeep Paulraj <s-paulraj@ti.com>
This patch configures clocks properly when a 36XX/37XX
processor is detected.
Signed-off-by: Steve Sakoman <steve@sakoman.com>
Signed-off-by: Sandeep Paulraj <s-paulraj@ti.com>
TI has added new processors to the OMAP3 family. This patch enhances
the code in sysinfo.c to detect which family member is present.
Signed-off-by: Steve Sakoman <steve@sakoman.com>
Signed-off-by: Sandeep Paulraj <s-paulraj@ti.com>
The functions in syslib.c can be shared, so this patch moves it from
cpu/omap3 to cpu/omap-common
Signed-off-by: Steve Sakoman <steve@sakoman.com>
Signed-off-by: Sandeep Paulraj <s-paulraj@ti.com>
* Fix: return value of get_tbclk
* this fixes issue with prematurely restart/retry, if BOOT_RETRY_TIMEOUT is used
Signed-off-by: Jens Scharsig <js_at_ng@scharsoft.de>
Because of peripheral devices can select clock sources,
separate the peripheral clocks. (pwm, uart and so on)
It just return the pclk at s5pc1xx SoC,
but s5pc210 SoC must be calculated by own clock register setting.
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Fixed kw_winctrl_calcsize() off-by-1 bug which caused mapping
windows size to be cut by half.
This corrected all windows address configuration
Signed-off-by: Prafulla Wadaskar <prafulla@marvell.com>
Fix orion5x_winctrl_calcsize() off-by-1 bug which caused mapping
windows to be cut by half. This afected all windows including NOR
flash (causing half the flash to be unaccessible) but DRAM was and
still is fine as its size is determined otherwise.
Signed-off-by: Albert Aribaud <albert.aribaud@free.fr>
Because of these are common files around s5p Socs, rename from s5pc1xx to s5p.
And getting cpu_id is SoC specific, so move to SoC's header file.
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
This patch adds basic support for s5pc210.
s5p-common will be used by all of s5p SoCs.
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
This patch is intended to prepare the other S5P SoC. (s5pc210)
If use SoC specific defines then can't share with other SoC.
So, make the accessor functions for access the base address by common way.
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Turn all ORION5X_DEF{ADR,SZ}_xxx macros into ORION5X_{ADR,SZ}_xxx
and allow defining them from board code to override defaults. This
is particularly useful for defining board-specific FLASH address
and size in board header file rather than having to tweak orion5x
code.
Signed-off-by: Albert Aribaud <albert.aribaud@free.fr>
get_timer_masked() should return current timestamp,
not current ticks from hardware register.
Tested on one custom board with NAND flash.
Without this patch, NAND write always TIMEOUT
because get_timer(0) return a big value.
This patch applies for u-boot-2010.06
Signed-off-by: Li Haibo <hbli@sinocastel.com>
The driver name does not need to be writable, so constify it.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Ben Warren <biggerbadderben@gmail.com>
The OMAP4 x-load code sets gptimer1 clock source to 32Khz. This isn't
acceptable for udelay. This patch changes from gptimer1 to gptimer2,
which uses sys_clk at 38.4 Mhz.
Signed-off-by: Steve Sakoman <steve@sakoman.com>
Signed-off-by: Sandeep Paulraj <s-paulraj@ti.com>
Add functional multiplexing support for OMAP4 pads.
Configure all the pads for the OMAP4430 SDP
and OMAP4 Panda boards
Signed-off-by: Steve Sakoman <steve@sakoman.com>
Signed-off-by: Aneesh V <aneesh@ti.com>
Signed-off-by: Sandeep Paulraj <s-paulraj@ti.com>
This patch adds a gpmc_init function for OMAP4 and adds calls to
gpmc_init for existing OMAP4 boards: panda and sdp4430
Signed-off-by: Steve Sakoman <steve@sakoman.com>
Signed-off-by: Sandeep Paulraj <s-paulraj@ti.com>
Add definitions and initialization in orion5x for mvgbe.
Add orion5x in mvgbe SoC includes.
Signed-off-by: Albert Aribaud <albert.aribaud@free.fr>
Acked-by: Prafulla Wadaskar <prafulla@marvell.com>
Signed-off-by: Ben Warren <biggerbadderben@gmail.com>
Rename all references to kirkwood in mvgbe symbols
throughout the whole codebase.
Signed-off-by: Albert Aribaud <albert.aribaud@free.fr>
Acked-by: Prafulla Wadaskar <prafulla@marvell.com>
Signed-off-by: Ben Warren <biggerbadderben@gmail.com>
This patch adds minimum support for OMAP4. Code which can be shared
between OMAP3 and OMAP4 is placed in arch/arm/cpu/armv7/omap-common
Signed-off-by: Aneesh V <aneesh@ti.com>
Signed-off-by: Steve Sakoman <steve@sakoman.com>
Signed-off-by: Sandeep Paulraj <s-paulraj@ti.com>
The purpose of this patch is to prepare for adding the OMAP4 architecture, which is Cortex A9
Cortex A8 and A9 both belong to the armv7 architecture, hence the name change.
The two architectures are similar enough that substantial code can be shared.
Signed-off-by: Aneesh V <aneesh@ti.com>
Signed-off-by: Steve Sakoman <steve@sakoman.com>
Signed-off-by: Sandeep Paulraj <s-paulraj@ti.com>
The hush shell dynamically allocates (and re-allocates) memory for the
argument strings in the "char *argv[]" argument vector passed to
commands. Any code that modifies these pointers will cause serious
corruption of the malloc data structures and crash U-Boot, so make
sure the compiler can check that no such modifications are being done
by changing the code into "char * const argv[]".
This modification is the result of debugging a strange crash caused
after adding a new command, which used the following argument
processing code which has been working perfectly fine in all Unix
systems since version 6 - but not so in U-Boot:
int main (int argc, char **argv)
{
while (--argc > 0 && **++argv == '-') {
/* ====> */ while (*++*argv) {
switch (**argv) {
case 'd':
debug++;
break;
...
default:
usage ();
}
}
}
...
}
The line marked "====>" will corrupt the malloc data structures and
usually cause U-Boot to crash when the next command gets executed by
the shell. With the modification, the compiler will prevent this with
an
error: increment of read-only location '*argv'
N.B.: The code above can be trivially rewritten like this:
while (--argc > 0 && **++argv == '-') {
char *arg = *argv;
while (*++arg) {
switch (*arg) {
...
Signed-off-by: Wolfgang Denk <wd@denx.de>
Acked-by: Mike Frysinger <vapier@gentoo.org>
The ARM ABI requires that the stack be aligned to 8 bytes as it is noted
in Procedure Call Standard for the ARM Architecture:
http://infocenter.arm.com/help/topic/com.arm.doc.ihi0042d/index.html
Unaligned SP also causes the problem with variable-length arrays
allocation when VLA address becomes less than stack pointer during
aligning of this address, so the next 'push' in the stack overwrites
first 4 bytes of VLA.
Signed-off-by: Vitaly Kuzmichev <vkuzmichev@mvista.com>
Tested on tx25(mx25), imx27lite(mx27), qong(mx31) and trab(s3c2400)
Tested-by: Wolfgang Denk <wd@denx.de>
The push / pop instructions used in this file are available only with
more recent tool chains:
cache.S: Assembler messages:
cache.S:133: Error: bad instruction `push {r0,r1,r2,lr}'
cache.S:160: Error: bad instruction `pop {r1,r2,r3,pc}'
cache.S:164: Error: bad instruction `push {r0,r1,r2,lr}'
cache.S:191: Error: bad instruction `pop {r1,r2,r3,pc}'
Change push/pop into stmfd/ldmfd instructions to support older
versions of binutils as well.
I verified that the modified source code generates exactly the same
binary code.
Signed-off-by: Wolfgang Denk <wd@denx.de>
Cc: Sandeep Paulraj <s-paulraj@ti.com>
Cc: Tom Rix <tom@bumblecow.com>
This patch adds support for the Marvell Orion5x SoC.
It has no use alone, and must be followed by a patch
to add Orion5x support for serial, then support for
the ED Mini V2, an Orion5x-based product from LaCie.
Signed-off-by: Albert Aribaud <albert.aribaud@free.fr>
Part of this patch is by: Mikhail Kshevetskiy.
Stack must be aligned to 8 bytes on PXA (possibly all armv5te) for LDRD/STRD
instructions. In case LDRD/STRD is issued on an unaligned address, the behaviour
is undefined.
The issue was observed when working with the NAND code, which was rendered
disfunctional. Also, the vsprintf() function had serious problems with printing
64bit wide long longs. After aligning the stack, this wrong behaviour is no
longer present.
Tested on:
Marvell Littleton PXA310 board
Toradex Colibri PXA320 board
Aeronix Zipit Z2 PXA270 handheld
Voipac PXA270 board
Signed-off-by: Marek Vasut <marek.vasut@gmail.com>
This patch adds support for the EMIF4 interface
available in the AM35x processors.
Signed-off-by: Vaibhav Hiremath <hvaibhav@ti.com>
Signed-off-by: Sanjeev Premi <premi@ti.com>
Signed-off-by: Sandeep Paulraj <s-paulraj@ti.com>
Consolidated SDRC related functions into one file - sdrc.c
And also replaced sdrc_init with generic memory init
function (mem_init), this generalization of omap memory setup
is necessary to support the new emif4 interface introduced in AM3517.
Signed-off-by: Vaibhav Hiremath <hvaibhav@ti.com>
Signed-off-by: Sandeep Paulraj <s-paulraj@ti.com>
initialized for CS1
From: Vaibhav Hiremath <hvaibhav@ti.com>
The patch makes sure that size for SDRC CS1 gets calculated
only when the CS1 SDRC is initialized.
Signed-off-by: Vaibhav Hiremath <hvaibhav@ti.com>
Signed-off-by: Sandeep Paulraj <s-paulraj@ti.com>
TNETV107X is a Texas Instruments SoC based on an ARM1176 core, and with a
bunch on on-chip integrated peripherals. This is an initial commit with
basic functionality, more commits with drivers, etc. to follow.
Signed-off-by: Cyril Chemparathy <cyril@ti.com>
Signed-off-by: Sandeep Paulraj <s-paulraj@ti.com>
The current ARM1176 CPU specific code is too specific to the SMDK6400
architecture. The following changes were necessary prerequisites for the
addition of other SoCs based on ARM1176.
Existing board's (SMDK6400) configuration has been modified to keep behavior
unchanged despite these changes.
1. Peripheral port remap configurability
The earlier code had hardcoded remap values specific to s3c64xx in start.S.
This change makes the peripheral port remap addresses and sizes configurable.
2. U-Boot code relocation support
Most architectures allow u-boot code to run initially at a different
address (possibly in NOR) and then get relocated to its final resting place
in RAM. Added support for this capability in ARM1176 architecture.
3. Disable TCM if necessary
If a ROM based bootloader happened to have initialized TCM, we disable it here
to keep things sane.
4. Remove unnecessary SoC specific includes
ARM1176 code does not really need this SoC specific include. The presence
of this include prevents builds on other ARM1176 archs.
5. Modified virt-to-phys conversion during MMU disable
The original MMU disable code masks out too many bits from the load address
when it tries to figure out the physical address of the jump target label.
Consequently, it ends up branching to the wrong address after disabling the
MMU.
Signed-off-by: Cyril Chemparathy <cyril@ti.com>
Signed-off-by: Sandeep Paulraj <s-paulraj@ti.com>
The ARM1136 cache_flush() function uses the "mcr p15, 0, rn, c7, c7, 0"
instruction which means "Invalidate Both Caches" when in fact the intent
is to clean and invalidate all caches. So add an "mcr p15, 0, %0, c7,
c10, 0" instruction to "Clean Entire Data Cache" prior to the "Invalidate
Both Caches" instruction to insure that memory is consistent with any
dirty cache lines.
Also fix a couple of "flush v*" comments in ARM1136 cpu_init_crit() so
that they correctly describe the actual ARM1136 CP15 C7 Cache Operations
used.
Signed-off-by: George G. Davis <gdavis@mvista.com>
Fix the following compiler problems:
arch/arm/cpu/arm920t/a320/liba320.a(timer.o): In function `udelay':
/home/wd/git/u-boot/work/arch/arm/cpu/arm920t/a320/timer.c:160: multiple definition of `udelay'
lib/libgeneric.a(time.o):/home/wd/git/u-boot/work/lib/time.c:34: first defined here
lib/libgeneric.a(time.o): In function `udelay':
time.c:(.text+0x1c): undefined reference to `__udelay'
Signed-off-by: Wolfgang Denk <wd@denx.de>
Fix MX51 CPU detect message.
Original string was:
CPU: Freescale i.MX51 family 3.0V at 800 MHz
which can be misinterpreted as 3.0 Volts instead of the silicon revision.
,change it to:
CPU: Freescale i.MX51 family rev3.0 at 800 MHz
Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Replace call to imx_get_mpllclk with imx_get_armclk
to show frequency of ARM core instead of mpll internal
bus in print_cpuinfo.
Signed-off-by: John Rigby <jcrigby@gmail.com>
CC: Stefano Babic <sbabic@denx.de>
The lowlevel_init file contained some hard-coded values
to setup the RAM. These board related values are moved into
the board configuration file.
Signed-off-by: Stefano Babic <sbabic@denx.de>
Because of s5pc1xx gpio is same as s5p seires SoC,
move gpio functions to drvier/gpio/
and modify structure's name from s5pc1xx_ to s5p_.
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
