platform/kernel/u-boot.git
16 years agoppc4xx: Make USB working with CONFIG_4xx_DCACHE defined
Matthias Fuchs [Fri, 9 Nov 2007 14:37:23 +0000 (15:37 +0100)]
ppc4xx: Make USB working with CONFIG_4xx_DCACHE defined

This patch disables the 44x d-cache on 'usb start' and
reenables it on 'usb stop'. This should be seen as a
temporary fix until the generic usb-ohci driver can
life with d-cache enabled.

Signed-off-by: Matthias Fuchs <matthias.fuchs@esd-electronics.com>
Signed-off-by: Stefan Roese <sr@denx.de>
16 years agoppc4xx: Remove redundant code from 4xx network driver
Matthias Fuchs [Fri, 9 Nov 2007 14:36:44 +0000 (15:36 +0100)]
ppc4xx: Remove redundant code from 4xx network driver

This patch removes some redundant code and decrements the end
address of cache flush and invalidate by 1.

Signed-off-by: Matthias Fuchs <matthias.fuchs@esd-electronics.com>
Signed-off-by: Stefan Roese <sr@denx.de>
16 years agoppc4xx: Remove In:/Out:/Err: boot output for AMCC Kilauea
Stefan Roese [Fri, 9 Nov 2007 11:19:58 +0000 (12:19 +0100)]
ppc4xx: Remove In:/Out:/Err: boot output for AMCC Kilauea

Signed-off-by: Stefan Roese <sr@denx.de>
16 years agoppc4xx: Make output a little shorter on I2C bootrom detection
Stefan Roese [Fri, 9 Nov 2007 11:18:54 +0000 (12:18 +0100)]
ppc4xx: Make output a little shorter on I2C bootrom detection

Most 4xx PPC capable of using an I2C bootrom for bootstrap setting
already print a line with the information which I2C bootrom is
used for bootstrap configuration. So we don't need this extra line
with "I2C boot EEPROM en-/dis-abled".

This patch also has a little code cleanup integrated.

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Make output a little shorter on PCIe detection
Stefan Roese [Mon, 5 Nov 2007 06:43:05 +0000 (07:43 +0100)]
ppc4xx: Make output a little shorter on PCIe detection

Now not max 3 lines but 2 lines are printed per PCIe port.

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Add AMCC Kilauea/Haleakala NAND booting support
Stefan Roese [Sat, 3 Nov 2007 11:08:28 +0000 (12:08 +0100)]
ppc4xx: Add AMCC Kilauea/Haleakala NAND booting support

This patch adds NAND booting support for the AMCC 405EX(r) eval boards.
Again, only one image supports both targets.

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Fix acadia_nand build problem
Stefan Roese [Wed, 31 Oct 2007 19:58:34 +0000 (20:58 +0100)]
ppc4xx: Fix acadia_nand build problem

Since the cache handling functions were moved from start.S into cache.S
the acadia NAND booting Makfile needs to be adapted accordingly.

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Add CONFIG_4xx_DCACHE compile options to enable cached SDRAM
Stefan Roese [Wed, 31 Oct 2007 19:57:11 +0000 (20:57 +0100)]
ppc4xx: Add CONFIG_4xx_DCACHE compile options to enable cached SDRAM

This patch adds the CONFIG_4xx_DCACHE options to some SDRAM init files
and to the Sequoia TLB init code. Now the cache can be enabled on 44x
boards by defining CONFIG_4xx_DCACHE in the board config file. This
option will disappear, when more boards use is successfully and no
more known problems exist.

This is tested successfully on Sequoia and Katmai. The only problem that
needs to be fixed is, that USB is not working on Sequoia right now, since
it will need some cache handling code too, similar to the 4xx EMAC driver.

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Enable CPU POST test for 4xx with dcache enabled
Stefan Roese [Wed, 31 Oct 2007 19:51:10 +0000 (20:51 +0100)]
ppc4xx: Enable CPU POST test for 4xx with dcache enabled

Now with caches enabled (i- and d-cache) on 44x, we need a chance to
disable the cache for the CPU POST tests, since these tests consist
of self modifying code. This is done via the new change_tlb() function.

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Change 4xx POST ethernet test to handle cached memory too
Stefan Roese [Wed, 31 Oct 2007 19:47:26 +0000 (20:47 +0100)]
ppc4xx: Change 4xx POST ethernet test to handle cached memory too

This patch enables the 4xx EMAC POST driver to work too, when dcache is
enabled.

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Remove temporary TLB entry in POST cache test only for 440
Stefan Roese [Wed, 31 Oct 2007 19:45:53 +0000 (20:45 +0100)]
ppc4xx: Remove temporary TLB entry in POST cache test only for 440

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Change 4xx ethernet driver to handle cached memory too
Stefan Roese [Wed, 31 Oct 2007 17:01:24 +0000 (18:01 +0100)]
ppc4xx: Change 4xx ethernet driver to handle cached memory too

This patch enables the 4xx EMAC driver to work too, when dcache is
enabled.

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Add change_tlb function to modify I attribute of TLB(s)
Stefan Roese [Wed, 31 Oct 2007 16:59:22 +0000 (17:59 +0100)]
ppc4xx: Add change_tlb function to modify I attribute of TLB(s)

This function is used to either turn cache on or off in a specific
memory area.

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Remove cache definition from 4xx board config files
Stefan Roese [Wed, 31 Oct 2007 16:57:52 +0000 (17:57 +0100)]
ppc4xx: Remove cache definition from 4xx board config files

All 4xx board config files don't need the cache definitions anymore.
These are now defined in common headers.

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Rework 4xx cache support
Stefan Roese [Wed, 31 Oct 2007 16:55:58 +0000 (17:55 +0100)]
ppc4xx: Rework 4xx cache support

New cache handling functions added and all existing functions
moved from start.S into seperate cache.S.

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Remove compiler warning from previous commit
Stefan Roese [Tue, 23 Oct 2007 16:03:12 +0000 (18:03 +0200)]
ppc4xx: Remove compiler warning from previous commit

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Remove temporary TLB entry in POST cache test
Stefan Roese [Tue, 23 Oct 2007 12:40:30 +0000 (14:40 +0200)]
ppc4xx: Remove temporary TLB entry in POST cache test

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Change autonegotiation timeout from 4 to 5 seconds
Stefan Roese [Tue, 23 Oct 2007 12:05:08 +0000 (14:05 +0200)]
ppc4xx: Change autonegotiation timeout from 4 to 5 seconds

I lately noticed, that newer 4xx board with GBit support sometimes don't
finish link autonegotiation in 4 seconds. Changing this timeout to 5
seconds seems fine here.

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Change 4xx_enet & miiphy to use out_be32() and friends
Stefan Roese [Tue, 23 Oct 2007 12:03:17 +0000 (14:03 +0200)]
ppc4xx: Change 4xx_enet & miiphy to use out_be32() and friends

This patch changes all in32/out32 calls to use the recommended in_be32/
out_be32 macros instead.

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Fix POST ethernet test for Haleakala
Stefan Roese [Thu, 25 Oct 2007 10:24:59 +0000 (12:24 +0200)]
ppc4xx: Fix POST ethernet test for Haleakala

The POST ethernet test needed to be changed to dynamically determine
the count of ethernet devices. This code is cloned from the 4xx
ethernet driver.

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Correct UART input clock calculation and passing to fdt
Stefan Roese [Tue, 23 Oct 2007 09:31:05 +0000 (11:31 +0200)]
ppc4xx: Correct UART input clock calculation and passing to fdt

We now use a value in the gd (global data) structure for the UART input
frequency, since the PPC4xx_SYS_INFO struct is always rewritten completely
in get_sys_info().

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Add initial AMCC Haleakala PPC405EXr eval board support
Stefan Roese [Tue, 23 Oct 2007 08:10:08 +0000 (10:10 +0200)]
ppc4xx: Add initial AMCC Haleakala PPC405EXr eval board support

The Haleakala is nearly identical with the Kilauea eval board. The only
difference is that the 405EXr only supports one EMAC and one PCIe
interface. This patch adds support for the Haleakala board by using
the identical image for Kilauea and Haleakala. The distinction is done
by comparing the PVR.

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Fixed offset of refresh rate type for Bamboo on-board DDR SDRAM
Eugene O'Brien [Tue, 23 Oct 2007 06:29:10 +0000 (08:29 +0200)]
ppc4xx: Fixed offset of refresh rate type for Bamboo on-board DDR SDRAM

This patch also adds a note to the fixed DDR setup for Bamboo NAND booting:

Note:
As found out by Eugene O'Brien <eugene.obrien@advantechamt.com>, the fixed
DDR setup has problems (U-Boot crashes randomly upon TFTP), when the DIMM
modules are still plugged in. So it is recommended to remove the DIMM
modules while using the NAND booting code with the fixed SDRAM setup!

Signed-off-by: Eugene O'Brien <eugene.obrien@advantechamt.com>
Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Add SNTP support to AMCC Katmai, Kilauea & Makalu boards
Stefan Roese [Mon, 22 Oct 2007 14:24:44 +0000 (16:24 +0200)]
ppc4xx: Add SNTP support to AMCC Katmai, Kilauea & Makalu boards

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Rework of 4xx serial driver (4)
Stefan Roese [Mon, 22 Oct 2007 14:22:40 +0000 (16:22 +0200)]
ppc4xx: Rework of 4xx serial driver (4)

Change 4xx_uart.c:

- Use in_8/out_8 macros instead of in8/out8
- No need for UART_BASE marco anymore, now really handled via function
  parameter
- serial_init_common() introduced
- Further coding style cleanup

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Rework of 4xx serial driver (3)
Stefan Roese [Mon, 22 Oct 2007 13:45:49 +0000 (15:45 +0200)]
ppc4xx: Rework of 4xx serial driver (3)

Change all linker scripts to reference the changed driver name iop480_uart.o.

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Rework of 4xx serial driver (2)
Stefan Roese [Mon, 22 Oct 2007 13:44:39 +0000 (15:44 +0200)]
ppc4xx: Rework of 4xx serial driver (2)

Change all linker scripts to reference the changed driver name 4xx_uart.o.

Note: In most cased all these explicit referencing of these object files
in the linker scripts is not neccessary. Only for manually embedded
environment into the U-Boot image, which is not done is most cases.

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Rework of 4xx serial driver (1)
Stefan Roese [Mon, 22 Oct 2007 13:09:59 +0000 (15:09 +0200)]
ppc4xx: Rework of 4xx serial driver (1)

This patch starts the rework of the PPC4xx serial driver. First we split
the file into two seperate files, one 4xx_uart.c with the 405/440 UART
handling code and the other one iop480_uart.c with the UART code for the
PLX-Tech IOP480 PPC (PPC403 based).

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Correct UART input clock calculation and passing to fdt
Stefan Roese [Mon, 22 Oct 2007 08:30:38 +0000 (10:30 +0200)]
ppc4xx: Correct UART input clock calculation and passing to fdt

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Add initial AMCC Makalu 405EX support
Stefan Roese [Mon, 22 Oct 2007 05:34:34 +0000 (07:34 +0200)]
ppc4xx: Add initial AMCC Makalu 405EX support

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Add freqUART to CPU speed detection
Stefan Roese [Mon, 22 Oct 2007 05:33:52 +0000 (07:33 +0200)]
ppc4xx: Add freqUART to CPU speed detection

This value is needed later for the device tree configuration of
the uart clock.

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc: Small Kilauea cleanup of config file
Stefan Roese [Sun, 21 Oct 2007 12:26:29 +0000 (14:26 +0200)]
ppc: Small Kilauea cleanup of config file

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agortc: Add Xicor/Intersil X1205 RTC support
Stefan Roese [Sun, 21 Oct 2007 06:16:12 +0000 (08:16 +0200)]
rtc: Add Xicor/Intersil X1205 RTC support

This patch adds support for the Xicor/Intersil X1205 RTC used on the
AMCC Makalu eval board. This driver is basically cloned from the Linux
driver version (2.6.23).

This patch also introduces the Linux bcd.h header for the BCD2BIN/
BIN2BCD conversions. In the future some of the other U-Boot RTC driver
should be converted to also use this header instead of implementing
their own local copy of these functions/macros.

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Consolidate some of the 405 and 440 macros/structs into 4xx
Stefan Roese [Sun, 21 Oct 2007 06:12:41 +0000 (08:12 +0200)]
ppc4xx: Consolidate some of the 405 and 440 macros/structs into 4xx

This patch moves some common 4xx macros and the PPC405_SYS_INFO/
PPC440_SYS_INFO structure into the common ppc4xx.h header.

Lot's of other macros are good candidates to be consolidated this way
in the future.

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Fix size setup in Kilauea DDR2 init routine
Stefan Roese [Sun, 21 Oct 2007 06:05:18 +0000 (08:05 +0200)]
ppc4xx: Fix size setup in Kilauea DDR2 init routine

The size was initilized wrong. Instead of 256MB, the DDR2 controller
was setup to 512MB. Now the correct values is used.

This patch also does a little cleanup and adds a comment here.

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Define CONFIG_BOOKE for all PPC440 based processors
Eugene O'Brien [Thu, 18 Oct 2007 15:29:04 +0000 (17:29 +0200)]
ppc4xx: Define CONFIG_BOOKE for all PPC440 based processors

CONFIG_BOOKE must be defined for PPC440 processors so that the proper SPR
number is used to access system registers.

Signed-off-by: Eugene O'Brien <eugene.obrien@advantechamt.com>
Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Change inbound PCIe location for endpoint tests on Katmai
Stefan Roese [Thu, 18 Oct 2007 05:42:27 +0000 (07:42 +0200)]
ppc4xx: Change inbound PCIe location for endpoint tests on Katmai

On Yucca & Katmai, the inbound memory map pointed to 0x4.0000.0000, which
is the internal SRAM. Since I now ported and tested this endpoint mode
on Kilauea successfully to map to 0 (SDRAM), I also changed this for
Katmai.

Yucca will stay at internal SRAM for now. Not sure if somebody relies on
this setup.

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Add PCIe endpoint support on Kilauea (405EX)
Stefan Roese [Thu, 18 Oct 2007 05:39:38 +0000 (07:39 +0200)]
ppc4xx: Add PCIe endpoint support on Kilauea (405EX)

This patch adds endpoint support for the AMCC Kilauea eval board. It can
be tested by connecting a reworked PCIe cable (only 1x lane singles
connected) to another root-complex.

In this test setup, a 64MB inbound window is configured at BAR0 which maps
to 0 on the PLB side. So accessing this BAR0 from the root-complex will
access the first 64MB of the SDRAM on the PPC side.

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Dynamic configuration of 4xx PCIe mode as root or endpoint mode
Stefan Roese [Sat, 13 Oct 2007 14:43:23 +0000 (16:43 +0200)]
ppc4xx: Dynamic configuration of 4xx PCIe mode as root or endpoint mode

This patch adds support for dynamic configuration of PCIe ports for the
AMCC PPC4xx boards equipped with PCIe interfaces. These are the PPC440SPe
boards Yucca & Katmai and the 405EX board Kilauea.

This dynamic configuration is done via the "pcie_mode" environement
variable. This variable can be set to "EP" or "RP" for endpoint or
rootpoint mode. Multiple values can be joined via the ":" delimiter.
Here an example:

pcie_mode=RP:EP:EP

This way, PCIe port 0 will be configured as rootpoint, PCIe port 1 and 2
as endpoint.

Per default Yucca will be configured as:
pcie_mode=RP:EP:EP

Per default Katmai will be configured as:
pcie_mode=RP:RP:REP

Per default Kilauea will be configured as:
pcie_mode=RP:RP

Signed-off-by: Tirumala R Marri <tmarri@amcc.com>
Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Enable device tree support (fdt) on Kilauea per default
Stefan Roese [Thu, 11 Oct 2007 09:15:59 +0000 (11:15 +0200)]
ppc4xx: Enable device tree support (fdt) on Kilauea per default

This patch enables the fdt support on the AMCC Kilauea eval board.
Additionally now EBC ranges fdt fixup is included to support NOR
FLASH mapping via the Linux physmap_of driver.

This Kilauea port now support booting arch/ppc and arch/powerpc
Linux kernels. The default environment "net_nfs" is for arch/ppc
and "net_nfs_fdt" is for arch/powerpc. In the long run, arch/ppc
support will be removed.

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Add additional debug info to 4xx fdt support
Stefan Roese [Thu, 11 Oct 2007 09:11:45 +0000 (11:11 +0200)]
ppc4xx: Add additional debug info to 4xx fdt support

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Fix small merge problems with CPCI440 and Acadia boards
Stefan Roese [Fri, 5 Oct 2007 19:28:58 +0000 (21:28 +0200)]
ppc4xx: Fix small merge problems with CPCI440 and Acadia boards

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Fix small merge problem in 4xx_enet.c
Stefan Roese [Fri, 5 Oct 2007 15:35:10 +0000 (17:35 +0200)]
ppc4xx: Fix small merge problem in 4xx_enet.c

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Add initial AMCC Kilauea 405EX support
Stefan Roese [Fri, 5 Oct 2007 15:11:30 +0000 (17:11 +0200)]
ppc4xx: Add initial AMCC Kilauea 405EX support

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Add PPC405EX support
Stefan Roese [Fri, 5 Oct 2007 15:10:59 +0000 (17:10 +0200)]
ppc4xx: Add PPC405EX support

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Cleanup of 4xx PCI and PCIe support (renaming)
Stefan Roese [Fri, 5 Oct 2007 15:09:36 +0000 (17:09 +0200)]
ppc4xx: Cleanup of 4xx PCI and PCIe support (renaming)

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Add initial fdt support to 4xx (first needed on 405EX)
Stefan Roese [Fri, 5 Oct 2007 15:07:50 +0000 (17:07 +0200)]
ppc4xx: Add initial fdt support to 4xx (first needed on 405EX)

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoPOST: Add 405EX support to 4xx UART POST test
Stefan Roese [Fri, 5 Oct 2007 15:04:57 +0000 (17:04 +0200)]
POST: Add 405EX support to 4xx UART POST test

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoDTT: Prepare DS1775 driver for use of different I2C addresses
Stefan Roese [Fri, 5 Oct 2007 13:10:02 +0000 (15:10 +0200)]
DTT: Prepare DS1775 driver for use of different I2C addresses

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: 4xx_pcie: Change PCIe status output to match common style
Stefan Roese [Fri, 5 Oct 2007 12:23:43 +0000 (14:23 +0200)]
ppc4xx: 4xx_pcie: Change PCIe status output to match common style

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: 4xx_pcie: Disable debug output as default
Stefan Roese [Fri, 5 Oct 2007 07:22:33 +0000 (09:22 +0200)]
ppc4xx: 4xx_pcie: Disable debug output as default

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: 4xx_pcie: More general cleanup and 405EX PCIe support added
Stefan Roese [Fri, 5 Oct 2007 07:18:23 +0000 (09:18 +0200)]
ppc4xx: 4xx_pcie: More general cleanup and 405EX PCIe support added

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: 4xx_pcie: Change CFG_PCIE_MEMSIZE to 128MB on Yucca & Katmai
Stefan Roese [Fri, 5 Oct 2007 05:57:20 +0000 (07:57 +0200)]
ppc4xx: 4xx_pcie: Change CFG_PCIE_MEMSIZE to 128MB on Yucca & Katmai

128MB seems to be the smallest possible value for the memory size
for on PCIe port. With this change now the BAR's of the PCIe cards
are accessible under U-Boot.

One big note: This only works for PCIe port 0 & 1. For port 2 this
currently doesn't work, since the base address is now 0xc0000000
(0xb0000000 + 2 * 0x08000000), and this is already occupied by
CFG_PCIE0_CFGBASE. But solving this issue for port 2 would mean
to change the base addresses completely and this change would have
too much impact right now.

This patch adds debug output to the 4xx pcie driver too.

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: 4xx_pcie: Fix problem with SDRN access using port number as idx
Stefan Roese [Wed, 3 Oct 2007 19:16:32 +0000 (21:16 +0200)]
ppc4xx: 4xx_pcie: Fix problem with SDRN access using port number as idx

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Rename 405gp_pci to 4xx_pci since its used on all 4xx platforms
Stefan Roese [Wed, 3 Oct 2007 13:01:02 +0000 (15:01 +0200)]
ppc4xx: Rename 405gp_pci to 4xx_pci since its used on all 4xx platforms

These files were introduced with the IBM 405GP but are currently used on all
4xx PPC platforms. So the name doesn't match the content anymore. This patch
renames the files to 4xx_pci.c/h.

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Add a comment for 405EX PCIe endpoint configuration
Stefan Roese [Wed, 3 Oct 2007 12:14:58 +0000 (14:14 +0200)]
ppc4xx: Add a comment for 405EX PCIe endpoint configuration

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Make 440SPe PCIe code more generic to use on different 4xx PPCs (3)
Stefan Roese [Wed, 3 Oct 2007 08:38:09 +0000 (10:38 +0200)]
ppc4xx: Make 440SPe PCIe code more generic to use on different 4xx PPCs (3)

(3) This patch introduces macros like SDRN_PESDR_DLPSET(port) to access
    the SDR registers of the PCIe ports. This makes the overall design
    clearer, since it removed a lot of switch statements which are not
    needed anymore.

    Also, the functions ppc4xx_init_pcie_rootport() and
    ppc4xx_init_pcie_entport() are merged into a single function
    ppc4xx_init_pcie_port(), since most of the code was duplicated.
    This makes maintainance and porting to other 4xx platforms
    easier.

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Make 440SPe PCIe code more generic to use on different 4xx PPCs (2)
Stefan Roese [Wed, 3 Oct 2007 05:48:09 +0000 (07:48 +0200)]
ppc4xx: Make 440SPe PCIe code more generic to use on different 4xx PPCs (2)

This patch is the first patch of a series to make the 440SPe PCIe code
usable on different 4xx PPC platforms. In preperation for the new 405EX
which is also equipped with PCIe interfaces.

(2) This patch renames the functions from 440spe_ to 4xx_ with a
    little additional cleanup

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Make 440SPe PCIe code more generic to use on different 4xx PPCs (1)
Stefan Roese [Wed, 3 Oct 2007 05:34:10 +0000 (07:34 +0200)]
ppc4xx: Make 440SPe PCIe code more generic to use on different 4xx PPCs (1)

This patch is the first patch of a series to make the 440SPe PCIe code
usable on different 4xx PPC platforms. In preperation for the new 405EX
which is also equipped with PCIe interfaces.

(1) This patch renames the files from 440spe_pcie to 4xx_pcie

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoMerge git://www.denx.de/git/u-boot
Stefan Roese [Sat, 27 Oct 2007 11:43:40 +0000 (13:43 +0200)]
Merge git://www.denx.de/git/u-boot

17 years agoTQM5200: increase kernel_addr_r and fdt_addr_r (hinted by Wolfgang Denk).
Bartlomiej Sieka [Thu, 25 Oct 2007 15:20:01 +0000 (17:20 +0200)]
TQM5200: increase kernel_addr_r and fdt_addr_r (hinted by Wolfgang Denk).

Signed-off-by: Bartlomiej Sieka <tur@semihalf.com>
17 years agoMerge branch 'motionpro_ng' of /home/tur/git/u-boot
Wolfgang Denk [Wed, 24 Oct 2007 09:05:28 +0000 (11:05 +0200)]
Merge branch 'motionpro_ng' of /home/tur/git/u-boot

17 years agoTQM5200: fix spurious characters on second serial interface
Martin Krause [Mon, 22 Oct 2007 14:45:53 +0000 (16:45 +0200)]
TQM5200: fix spurious characters on second serial interface

With this patch PSC3 is configured as UART. This is done, because if
the pins of PSC3 are not configured at all (-> all pins are GPI),
due to crosstalk, spurious characters may be send over the RX232_2_TXD
signal line.

Signed-off-by: Martin Krause <martin.krause@tqs.de>
17 years agoTQM5200S: fix commands for STK52xx base board because of missing SM501 grafic controller
Martin Krause [Mon, 22 Oct 2007 14:40:06 +0000 (16:40 +0200)]
TQM5200S: fix commands for STK52xx base board because of missing SM501 grafic controller

Some commands for the STK52xx base board try to access the SM501 grafic
controller. But the TQM5200S has no grafic controller (only the TQM5200
and the TQM5200B have). This patch deactivates the commands accessing
the SM501 for the TQM5200S.

Signed-off-by: Martin Krause <martin.krause@tqs.de>
17 years agoMips: Fix string functions differ prototype declaration
Jean-Christophe PLAGNIOL-VILLARD [Sun, 21 Oct 2007 07:14:28 +0000 (09:14 +0200)]
Mips: Fix string functions differ prototype declaration

Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
17 years agofsl_pci_init enable COMMAND_MEMORY if inbound window
Ed Swarthout [Fri, 19 Oct 2007 22:51:40 +0000 (17:51 -0500)]
fsl_pci_init enable COMMAND_MEMORY if inbound window

Patch 16e23c3f removed PCSRBAR allocation.  But passing zero windows
to pciauto_setup_device has the side effect of not getting
COMMAND_MEMORY set.

Signed-off-by: Ed Swarthout <Ed.Swarthout@freescale.com>
17 years agodelta: Fix OHCI_REGS_BASE undeclared and wait_ms implicit declaration
Jean-Christophe PLAGNIOL-VILLARD [Fri, 19 Oct 2007 08:55:24 +0000 (10:55 +0200)]
delta: Fix OHCI_REGS_BASE undeclared and wait_ms implicit declaration

Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
17 years agofix warning: no return statement in function returning non-void
Jean-Christophe PLAGNIOL-VILLARD [Fri, 19 Oct 2007 06:10:15 +0000 (08:10 +0200)]
fix warning: no return statement in function returning non-void

Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
17 years agoxsengine: Fix no partition type specified, use DOS as default
Jean-Christophe PLAGNIOL-VILLARD [Fri, 19 Oct 2007 04:33:45 +0000 (06:33 +0200)]
xsengine: Fix no partition type specified, use DOS as default

Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
17 years agolubbock: Fix no partition type specified, use DOS as default
Jean-Christophe PLAGNIOL-VILLARD [Thu, 18 Oct 2007 22:24:59 +0000 (00:24 +0200)]
lubbock: Fix no partition type specified, use DOS as default

Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
17 years agoCoding style: keep lists sorted; update CHANGELOG
Wolfgang Denk [Tue, 23 Oct 2007 14:50:03 +0000 (16:50 +0200)]
Coding style: keep lists sorted; update CHANGELOG

Signed-off-by: Wolfgang Denk <wd@denx.de>
17 years agoFix missing drivers makefile entries ds1722.c mw_eeprom.c
Jean-Christophe PLAGNIOL-VILLARD [Thu, 18 Oct 2007 22:09:05 +0000 (00:09 +0200)]
Fix missing drivers makefile entries ds1722.c mw_eeprom.c

Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
17 years agoFix warning differ in signedness in board/innokom/innokom.c
Jean-Christophe PLAGNIOL-VILLARD [Thu, 18 Oct 2007 22:07:39 +0000 (00:07 +0200)]
Fix warning differ in signedness in board/innokom/innokom.c

Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
17 years agofix pxa255_idp board
Marcel Ziswiler [Thu, 18 Oct 2007 22:25:33 +0000 (00:25 +0200)]
fix pxa255_idp board

The pxa255_idp being an old unmaintained board showed several issues:
1. CONFIG_INIT_CRITICAL was still defined.
2. Neither CONFIG_MAC_PARTITION nor CONFIG_DOS_PARTITION was defined.
3. Symbol flash_addr was undeclared.
4. The boards lowlevel_init function was still called memsetup.
5. The TEXT_BASE was still 0xa3000000 rather than 0xa3080000.
6. Using -march=armv5 instead of -march=armv5te resulted in lots of
'target CPU does not support interworking' warnings on recent compilers.
7. The PXA's serial driver redefined FFUART, BTUART and STUART used as
indexes rather than the register definitions from the pxa-regs header
file. Renamed them to FFUART_INDEX, BTUART_INDEX and STUART_INDEX to
avoid any ambiguities.
8. There were several redefinition warnings concerning ICMR, OSMR3,
OSCR, OWER, OIER, RCSR and CCCR in the PXA's assembly start file.
9. The board configuration file was rather outdated.
10. The part header file defined the vendor, product and revision arrays
as unsigned chars instead of just chars in the block_dev_desc_t
structure.

Signed-off-by: Marcel Ziswiler <marcel@ziswiler.com>
17 years agoMake MPC8266ADS command selection more robust
Rune Torgersen [Wed, 17 Oct 2007 16:56:31 +0000 (11:56 -0500)]
Make MPC8266ADS command selection more robust

Fix MPC8266 command line definition so it won't break when new commands
are added to u-boot.
Signed-off-by Rune Torgersen <runet@innovsys.com>

17 years agoMotion-PRO: Update configuration to accomodate next generation board.
Bartlomiej Sieka [Tue, 23 Oct 2007 11:14:10 +0000 (13:14 +0200)]
Motion-PRO: Update configuration to accomodate next generation board.

New board has faster oscillator and a different Flash chip. This affects:
- CFG_MPC5XXX_CLKIN
- SDRAM timings
- Flash CS configuration (timings)
- Flash sector size, and thus MTD partition layout
- malloc() arena size (due to bigger Flash sectors)
- smaller memory test range (due to bigger malloc() arena)

This patch also enables more extensive memory testing via "mtest".

Signed-off-by: Bartlomiej Sieka <tur@semihalf.com>
17 years agoMotion-PRO: Add setting of SDelay reg. to SDRAM controller configuration.
Bartlomiej Sieka [Tue, 23 Oct 2007 09:36:07 +0000 (11:36 +0200)]
Motion-PRO: Add setting of SDelay reg. to SDRAM controller configuration.

Per AN3221 (MPC5200B SDRAM Initialization and Configuration), the SDelay
register must be written a value of 0x00000004 as the first step of the
SDRAM contorller configuration.

Signed-off-by: Bartlomiej Sieka <tur@semihalf.com>
17 years agoppc4xx: lwmon5: Some further GPIO config changes
Stefan Roese [Tue, 23 Oct 2007 08:17:42 +0000 (10:17 +0200)]
ppc4xx: lwmon5: Some further GPIO config changes

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoMinor coding style cleanup; update CHANGELOG
Wolfgang Denk [Sat, 20 Oct 2007 23:01:17 +0000 (01:01 +0200)]
Minor coding style cleanup; update CHANGELOG

Signed-off-by: Wolfgang Denk <wd@denx.de>
17 years agoFix NE2000 driver:
Vlad Lungu [Thu, 4 Oct 2007 17:47:10 +0000 (20:47 +0300)]
Fix NE2000 driver:

Fixed typo in ne2000.h, thinko re n2k_inb() usage, don't try
to do anything in eth_stop() if eth_init() was not called.
Simplified RX path in order to avoid timeouts on really really
fast NE2000 cards (read: qemu with internal tftp), NetLoop() is
clever enough to cope with 1 packet per eth_rx().

Signed-off-by: Vlad Lungu <vlad@comsys.ro>
17 years agoMerge branch 'master' of git://www.denx.de/git/u-boot-tq-group
Wolfgang Denk [Sat, 20 Oct 2007 22:29:32 +0000 (00:29 +0200)]
Merge branch 'master' of git://denx.de/git/u-boot-tq-group

17 years agoMerge branch 'master' of /home/git/u-boot/
Wolfgang Denk [Sat, 20 Oct 2007 22:10:34 +0000 (00:10 +0200)]
Merge branch 'master' of /home/git/u-boot/

17 years agoFix two typos.
Detlev Zundel [Fri, 19 Oct 2007 14:47:26 +0000 (16:47 +0200)]
Fix two typos.

Signed-off-by: Detlev Zundel <dzu@denx.de>
17 years agoMerge branch 'master' of git://www.denx.de/git/u-boot-ppc4xx
Wolfgang Denk [Thu, 18 Oct 2007 20:02:35 +0000 (22:02 +0200)]
Merge branch 'master' of git://denx.de/git/u-boot-ppc4xx

17 years agompc83xx: Add configure entry for MPC83xx ATM support
Tony Li [Thu, 18 Oct 2007 09:47:19 +0000 (17:47 +0800)]
mpc83xx: Add configure entry for MPC83xx ATM support

Add MPC8360EMDS_ATM_config and MPC832XEMDS_ATM_config into
Makfile and MAKEALL

Signed-off-by: Tony Li <tony.li@freescale.com>
Signed-off-by: Kim Phillips <kim.phillips@freescale.com>
17 years agompc83xx: pq-mds-pib.c typo error
Tony Li [Thu, 18 Oct 2007 09:44:38 +0000 (17:44 +0800)]
mpc83xx: pq-mds-pib.c typo error

Correct to val8 from val.

Signed-off-by: Tony Li <tony.li@freescale.com>
Signed-off-by: Kim Phillips <kim.phillips@freescale.com>
17 years agoMerge git://www.denx.de/git/u-boot
Kim Phillips [Thu, 18 Oct 2007 15:02:16 +0000 (10:02 -0500)]
Merge git://www.denx.de/git/u-boot

17 years agoppc4xx: Add 667/133 (CPU/PLB) frequency setup to Sequoia bootstrap command
Stefan Roese [Wed, 17 Oct 2007 13:40:19 +0000 (15:40 +0200)]
ppc4xx: Add 667/133 (CPU/PLB) frequency setup to Sequoia bootstrap command

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoMake MPC8266ADS board compile again.
runet@innovsys.com [Tue, 16 Oct 2007 19:50:40 +0000 (14:50 -0500)]
Make MPC8266ADS board compile again.

Signed-off-by: Runet Torgersen <runet@innovsys.com>
17 years agoMerge branch 'master' of git+ssh://gemini_vpn/home/wd/git/u-boot/master
Wolfgang Denk [Tue, 16 Oct 2007 14:45:20 +0000 (16:45 +0200)]
Merge branch 'master' of git+ssh://gemini_vpn/home/wd/git/u-boot/master

17 years ago86xx: Allow for fewer DDR slots per memory controller.
Jon Loeliger [Mon, 27 Aug 2007 17:41:03 +0000 (12:41 -0500)]
86xx: Allow for fewer DDR slots per memory controller.

As a direct correlation exists between DDR DIMM slots
and SPD EEPROM addresses used to configure them, use
the individually defined SPD_EEPROM_ADDRESS* values to
determine if a DDR DIMM slot should have its SPD
configuration read or not.

Effectively, this now allows for 1 or 2 DIMM slots
per memory controller.

Signed-off-by: Jon Loeliger <jdl@freescale.com>
17 years agoMerge branch 'master' of /home/wd/git/u-boot/custodians
Wolfgang Denk [Mon, 15 Oct 2007 18:56:12 +0000 (20:56 +0200)]
Merge branch 'master' of /home/wd/git/u-boot/custodians

17 years agoMerge branch 'master' of git://www.denx.de/git/u-boot-usb
Wolfgang Denk [Mon, 15 Oct 2007 18:55:51 +0000 (20:55 +0200)]
Merge branch 'master' of git://denx.de/git/u-boot-usb

17 years agoMerge branch 'master' of git+ssh://gemini_vpn/home/wd/git/u-boot/master
Wolfgang Denk [Mon, 15 Oct 2007 10:59:05 +0000 (12:59 +0200)]
Merge branch 'master' of git+ssh://gemini_vpn/home/wd/git/u-boot/master

17 years agoPXA USB OHCI: "usb stop" implementation.
Rodolfo Giometti [Mon, 15 Oct 2007 09:59:17 +0000 (11:59 +0200)]
PXA USB OHCI: "usb stop" implementation.

Some USB keys need to be switched off before loading the kernel
otherwise they can remain in an undefined status which prevents them
to be correctly recognized by the kernel.

Signed-off-by: Rodolfo Giometti <giometti@linux.it>
17 years agoppc4xx: Fix bug in I2C bootstrap values for Sequoia/Rainier
Stefan Roese [Mon, 15 Oct 2007 09:39:00 +0000 (11:39 +0200)]
ppc4xx: Fix bug in I2C bootstrap values for Sequoia/Rainier

The I2C bootstrap values that can be setup via the "bootstrap" command,
were setup incorrect regarding the generation of the internal sync PCI
clock. The values for PLB clock == 133MHz were slighly incorrect and the
values for PLB clock == 166MHz were totally incorrect. This could
lead to a hangup upon booting while PCI configuration scan.

This patch fixes this issue and configures valid PCI divisor values
for the sync PCI clock, with respect to the provided external async
PCI frequency.

Here the values of the formula in the chapter 14.2 "PCI clocking"
from the 440EPx users manual:

AsyncPCICLK - 1MHz <= SyncPCIClk <= (2 * AsyncPCIClk) - 1MHz

33MHz async PCI frequency:
PLB = 133:
=>      32 <= 44.3 <= 65        (div = 3)

PLB = 166:
=>      32 <= 55.3 <= 65        (div = 3)

66MHz async PCI frequency:
PLB = 133:
=>      65 <= 66.5 <= 132       (div = 2)

PLB = 166:
=>      65 <= 83 <= 132         (div = 2)

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoppc4xx: Fix incorrect 33/66MHz PCI clock log-message on Sequoia & Yosemite
Stefan Roese [Mon, 15 Oct 2007 09:29:33 +0000 (11:29 +0200)]
ppc4xx: Fix incorrect 33/66MHz PCI clock log-message on Sequoia & Yosemite

The BCSR status bit for the 66MHz PCI operation was correctly
addressed (MSB/LSB problem). Now the correct currently setup
PCI frequency is displayed upon bootup.

This patch also fixes this problem on Rainier & Yellowstone, since these
boards use the same souce code as Sequoia & Yosemite do.

Signed-off-by: Stefan Roese <sr@denx.de>
17 years agoTQM860M: adjust for doubled flash sector size.
Martin Krause [Wed, 26 Sep 2007 15:55:56 +0000 (17:55 +0200)]
TQM860M: adjust for doubled flash sector size.

Adjust flash map to support the new S29GLxxN (N-Type) Flashes with
doubled sector size.

Signed-off-by: Martin Krause <martin.krause@tqs.de>
17 years agoTQM8xx: Fix CAN timing.
Jens Gehrlein [Wed, 26 Sep 2007 15:55:54 +0000 (17:55 +0200)]
TQM8xx: Fix CAN timing.

Signed-off-by: Martin Krause <martin.krause@tqs.de>
17 years agoTQM866M: fix SDRAM refresh
Martin Krause [Thu, 27 Sep 2007 12:54:36 +0000 (14:54 +0200)]
TQM866M: fix SDRAM refresh

At 133 MHz the current SDRAM refresh rate is too fast
(measured 4 * 1.17 us).
CFG_MAMR_PTA changes from 39 to 97. This result
in a refresh rate of 4 * 7.8 us at the default clock
50 MHz. At 133 MHz the value will be then 4 * 2.9 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>