Stefan Roese [Sat, 5 Jan 2008 08:12:41 +0000 (09:12 +0100)]
ppc4xx: Update Katmai/44x_spd_ddr2.c code for optimal DDR2 setup
On Katmai the complete auto-calibration somehow doesn't seem to
produce the best results, meaning optimal values for RQFD/RFFD.
This was discovered by GDA using a high bandwidth scope,
analyzing the DDR2 signals. GDA provided a fixed value for RQFD,
so now on Katmai "only" RFFD is auto-calibrated.
This patch also adds RDCC calibration as mentioned on page 7 of
the AMCC PowerPC440SP/SPe DDR2 application note:
"DDR1/DDR2 Initialization Sequence and Dynamic Tuning"
First, it adds function gpio_read_in_bit() to read the a bit from the
GPIO Input Register (GPIOx_IR) in the same way that function
gpio_read_out_bit() reads a bit from the GPIO Output Register
(GPIOx_OR).
Second, it modifies function gpio_set_chip_configuration() to provide
an additional option for configuring the GPIO from the
"CFG_4xx_GPIO_TABLE".
According to the 440EPx User's Manual, when an alternate output is used,
the three-state control is configured in one of two ways, depending on
the particular output. The first option is to select the corresponding
alternate three-state control in the GPIOx_TRSH/L registers. The second
option is to select the GPIO Three-State Control Register (GPIOx_TCR) in
the GPIOx_TRSH/L registers, and set the corresponding bit in the
GPIOx_TCR register to enable the output. For example, the Manual
specifies configuring the GPIO00 Alternate 1 Signal (PreAddr07) to use
the alternate three-state control (first option), and specifies
configuring the GPIO32 Alternate 1 Signal (USB2OM0) with the output
enabled in the GPIOx_TCR register (second option).
Currently, gpio_set_chip_configuration() configures all alternate signal
outputs to use the first option. This patch allow the second option to
be selected by setting the "out_val" element in the table entry to
"GPIO_OUT_1". The first option is used when the "out_val" element is
set to "GPIO_OUT_0". Because "out_val" is not currently used when an
alternate signal is selected, and because all current GPIO tables set
"out_val" to "GPIO_OUT_0" for all alternate signals, this patch should
not change any existing configurations.
Wolfgang Denk [Wed, 2 Jan 2008 14:54:45 +0000 (15:54 +0100)]
Make scripts and Makefiles POSIX compliant
The bash builtin versions of the "test" (resp. "[") command allow
using "==" for string comparisons, but POSIX compatible implemen-
tations (like /usr/bin/test) insist on using "=" only. On such systems
you will see:
$ /usr/bin/test a == a && echo OK
/usr/bin/test: ==: binary operator expected
This patch fixes Makefiles and scripts to use POSIX style.
Stefan Roese [Fri, 28 Dec 2007 15:08:08 +0000 (16:08 +0100)]
ppc4xx: Enable 405EP PCI arbiter per default on all boards
In an attmemt to clean up the 4xx start.S file, I removed the enabling
of the internal 405EP PCI arbiter. This is needed for multiple other
405EP platforms, like most of the esd 405EP. Now the internal PCI
arbiter is enabled again per default as it has been before.
Signed-off-by: Stefan Roese <sr@denx.de> Acked-by: Matthias Fuchs <matthias.fuchs@esd-electronics.com>
Larry Johnson [Thu, 27 Dec 2007 15:54:48 +0000 (10:54 -0500)]
Add 440EPx DDR2 SPD DIMM support
This patch adds SPD DDR2 support for the 440EPx ("Denali") SDRAM
controller. It should also work on the 440GRx. It is based on the DDR2
SPD code for the 440EP/440EPx, but makes no provision for DDR1 support.
This code has been tested on prototype Korat boards with three Kingston
DIMMS: 512 MiB ECC (one rank), 512 MiB non-ECC (one rank) and 1 GiB ECC
(two ranks). The Korat board has a single DIMM socket, but support has
been provided (though not tested) for boards with two DIMM sockets.
Larry Johnson [Sat, 22 Dec 2007 20:15:30 +0000 (15:15 -0500)]
Copy 440EPx/GRx SDRAM data-eye search to common directory
This patch creates a non-board-specific file for performing the SDRAM
data-eye search. It also adds ECC error checking to the test of valid
data on readback when ECC is enabled.
Larry Johnson [Sat, 22 Dec 2007 20:15:13 +0000 (15:15 -0500)]
Add definitions for 440EPx/GRx SDRAM controller to ppc440.h
This patch adds the Denali SDRAM controller definitions to "ppc440.h".
It also fixes two typos in the definitions, so the board-specific
"sdram.h" files containing these definitions are also fixed to avoid
compiler warnings.
Larry Johnson [Thu, 27 Dec 2007 15:50:55 +0000 (10:50 -0500)]
Add Ethernet 1000BASE-X support for PPC4xx
This patch adds a new switch: "CONFIG_PHY_DYNAMIC_ANEG". When this symbol
is defined, the PHY will advertise it's capabilities for autonegotiation
based on the capabilities shown in the PHY's status registers, including
1000BASE-X. When "CONFIG_PHY_DYNAMIC_ANEG" is not defined, the PHY will
advertise hard-coded capabilities, as before.
Larry Johnson [Sat, 22 Dec 2007 20:14:00 +0000 (15:14 -0500)]
Add driver for STMicroelectronics M41T60 RTC
This driver is based on the driver for the M41T11. In the intended
application, the RTC will be powered by a large capacitor, rather than a
battery. The driver therefore checks to see whether the RTC has lost
power. The chip's OUT bit is normally reset from its power-up state. If
the OUT bit is read as set, or if the date and time are not valid, then the
RTC is assumed to have lost power, and its date and time are reset to
1900-01-01 00:00:00.
Support for adjusting the speed of the clock to improve accuracy is
provided through an environment variable.
Stefan Roese [Sat, 22 Dec 2007 11:18:26 +0000 (12:18 +0100)]
ppc4xx: Fix problem in 44x cache POST routine
As repoted by Larry Johnson, running "diag run cache" caused a crash
in U-Boot. This problem was introduced by a patch that removed the
TLB entry for the cache test after the test has completed. Since this
TLB was only setup once, a 2nd attempt to run this cache test
failed with a crash. Now this TLB entry is created every time the
routine is called.
Stefan Roese [Tue, 18 Dec 2007 07:44:51 +0000 (08:44 +0100)]
ppc4xx: fdt: Cleanup setup of cpu node setup
Now the cpu node setup ("timebase-frequency" and "clock-frequency") is
without using the absolute path to the cpu node. This makes it possible
to use this U-Boot version with both versions of cpu-node naming
"cpu@0" and the former "PowerPC,440EPx@0".
ppc4xx: Add CONFIG_BOOTP_SUBNETMASK to Sequoia board config
When using dhcp/bootp the "netmask" environment variable is not
set because CONFIG_BOOTP_SUBNETMASK is not defined. But usually this is
desireable, so the following patch adds this this option to the board
config.
Signed-off-by: Markus Klotzbuecher <mk@denx.de> Signed-off-by: Stefan Roese <sr@denx.de>
Larry Johnson [Sat, 27 Oct 2007 16:48:15 +0000 (12:48 -0400)]
Fix/enhance ECC POST for 440EPx/GRx
This patch allows the ECC POST to be used for different boards with the
PPC440 Denali SDRAM controller. Modifications include skipping the test
if ECC is not enabled (as for non-ECC DIMMs) and adding synchronization
to prevent timing errors.
Stefan Roese [Thu, 27 Dec 2007 06:50:54 +0000 (07:50 +0100)]
cfi_flash: Add missing check for erased dest to flash_write_cfibuffer()
The check for an sufficiently erased destination was missing in the
buffered write function of the cfi flash driver (when
CFG_FLASH_USE_BUFFER_WRITE is defined). This patch adds this check to that
writing to such a region will fail with the currect error message.
TQM885D: Exchanged SDRAM timing by a more relaxed timing.
CAS-Latency=2, Write Recovery Time tWR=2
The max. supported bus frequency is 66 MHz. Therefore, changed
threshold to switch from 1:1 mode to 2:1 from 80 MHz to 66 MHz.
Signed-off-by: Martin Krause <martin.krause@tqs.de>
Martin Krause [Wed, 26 Sep 2007 15:55:56 +0000 (17:55 +0200)]
TQM885D: use calculated cpuclk instead of measuring it
On the TQM885D the measurement of cpuclk with the PIT reference
timer ist not necessary. Since all module variants use the same
external 10 MHz oscillator, the cpuclk only depends on the PLL
configuration - which is readable by software.
Signed-off-by: Martin Krause <martin.krause@tqs.de>
At 133 MHz the current SDRAM refresh rate is too fast
(measured 4 * 1.17 us).
CFG_MAMR_PTA changes from 39 to 128. This result
in a refresh rate of 4 * 7.8 us at the default clock
66 MHz. At 133 MHz the value will be then 4 * 3.8 us.
This is a compromise until a new method is found to
adjust the refresh rate.
Signed-off-by: Martin Krause <martin.krause@tqs.de>
Run fixups based on the JEDEC manufacturer ID independent of the
command set ID.
This changes current behaviour: Previously, geometry reversal for AMD
chips were done based on the command set ID, while they are now done
based on the JEDEC manufacturer and device ID.
Also add fixup for top-boot Atmel chips. A fixup is needed for
AT49BV6416(T) too, but since u-boot currently only reads the low byte
of the device ID, there's no way to tell it apart from AT49BV642D,
which should not have this fixup. Since AT49BV642D support is
necessary to get ATNGW100 board support into mainline, I've commented
out the fixup for now.
Move things like reading JEDEC IDs and fixing up geometry reversal
into separate functions. The geometry reversal fixup is now performed
by altering the qry structure directly, which makes the sector init
code slightly cleaner.
Read out the whole CFI Standard Query structure after successful cfi
identification. This allows subsequent code to access this information
directly without having to go through flash_read_uchar() and friends.
The timer code depends on the timer interrupt to keep track of the
upper 32 bits of the cycle counter. This obviously doesn't work when
interrupts are disabled the whole time.
Add all the ATSTK1000 daughterboards to MAINTAINERS along with their
"mother". Also update the entry for ATSTK1000 to be not only about the
AP7000 CPU; it's intended to handle all CPUs in the AT32AP family.
The SoC-specific code for all the AT32AP700x CPUs is practically
identical; the only difference is that some chips have less features
than others. By doing this rename, we can add support for the AP7000
derivatives simply by making some features conditional.
Use map_physmem() and unmap_physmem() to convert from physical to
virtual addresses. This gives the arch a chance to provide an uncached
mapping for flash accesses.
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.
Introduce flash_read{8,16,32,64) and flash_write{8,16,32,64} and use
them to access the flash memory. This makes it clearer when the flash
is actually being accessed; merely dereferencing a volatile pointer
looks just like any other kind of access.
Implement __raw_{read,write}[bwl] on all architectures
This adds implementations of __raw_read[bwl] and __raw_write[bwl] to
m68k, ppc, nios and nios2. The m68k and ppc implementations were taken
from Linux.
projects / u-boot / log