platform/kernel/u-boot.git
15 years agoMerge branch 'master' of git://git.denx.de/u-boot-i2c
Wolfgang Denk [Sun, 15 Feb 2009 21:21:17 +0000 (22:21 +0100)]
Merge branch 'master' of git://git.denx.de/u-boot-i2c

15 years agoi.MX31: Start the I2C clock on driver initialisation
Guennadi Liakhovetski [Fri, 13 Feb 2009 08:23:36 +0000 (09:23 +0100)]
i.MX31: Start the I2C clock on driver initialisation

i.MX31 powers on with most clocks running, so, after a power on this explicit
clock start up is not required. However, as Linux boots it disables most clocks
to save power. This includes the I2C clock. If we then soft reboot from Linux
the I2C clock stays off. This breaks the phycore, which has its environment in
I2C EEPROM. Fix the problem by explicitly starting the clock in I2C driver
initialisation routine.

Signed-off-by: Guennadi Liakhovetski <lg@denx.de>
Ack-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
15 years agoi2c.h: drop i2c_reg_{read, write} hack for Blackfin parts
Mike Frysinger [Thu, 12 Feb 2009 01:36:14 +0000 (20:36 -0500)]
i2c.h: drop i2c_reg_{read, write} hack for Blackfin parts

The Blackfin i2c driver has been rewritten thus the special ifdefs in the
common code are no longer needed.

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
15 years agoMerge branch 'master' of git://git.denx.de/u-boot-ppc4xx
Wolfgang Denk [Thu, 12 Feb 2009 07:36:52 +0000 (08:36 +0100)]
Merge branch 'master' of git://git.denx.de/u-boot-ppc4xx

15 years ago82xx, mgcoge: fix compile error
Heiko Schocher [Thu, 12 Feb 2009 07:08:54 +0000 (08:08 +0100)]
82xx, mgcoge: fix compile error

With actual u-boot compiling the mgcoge port fails, because
since commit ba705b5b1a97b47388ed48858bef6bf7b6bfcd56 it is
necessary to define CONFIG_NET_MULTI.

Seems to me the mgcoge port is the only actual existing 8260
port who uses CONFIG_ETHER_ON_SCC, so no other 8260 port needed
to be fixed.

Signed-off-by: Heiko Schocher <hs@denx.de>
15 years agoppc4xx: Add README entry for CONFIG_PCI_DISABLE_PCIE
Dirk Eibach [Mon, 9 Feb 2009 07:18:34 +0000 (08:18 +0100)]
ppc4xx: Add README entry for CONFIG_PCI_DISABLE_PCIE

Signed-off-by: Dirk Eibach <eibach@gdsys.de>
Signed-off-by: Stefan Roese <sr@denx.de>
15 years agoppc4xx: Fix initialization of the SDRAM_CODT register
Carolyn Smith [Thu, 12 Feb 2009 05:13:44 +0000 (06:13 +0100)]
ppc4xx: Fix initialization of the SDRAM_CODT register

This fixes the initialization of the SDRAM_CODT register in the ppc4xx DDR2
initialization code. It also removes use of the SDRAM_CODT_FEEDBACK_RCV_SINGLE_END
and SDRAM_CODT_FEEDBACK_DRV_SINGLE_END #define's since they are reserved bits.

Signed-off-by: Carolyn Smith <carolyn.smith@tektronix.com>
Signed-off-by: Stefan Roese <sr@denx.de>
15 years agoppc4xx: Fix problem with board_eth_init() vs cpu_eth_init() on AMCC boards
Stefan Roese [Wed, 11 Feb 2009 08:29:33 +0000 (09:29 +0100)]
ppc4xx: Fix problem with board_eth_init() vs cpu_eth_init() on AMCC boards

Some AMCC eval boards do have a board_eth_init() function calling
pci_eth_init(). These boards need to call cpu_eth_init() explicitly now
with the new eth_init rework.

Signed-off-by: Stefan Roese <sr@denx.de>
15 years agoppc4xx: Autocalibration can set RDCC to over aggressive value.
Adam Graham [Mon, 9 Feb 2009 21:18:12 +0000 (13:18 -0800)]
ppc4xx: Autocalibration can set RDCC to over aggressive value.

The criteria of the AMCC SDRAM Controller DDR autocalibration
U-Boot code is to pick the largest passing write/read/compare
window that also has the smallest SDRAM_RDCC.[RDSS] Read Sample
Cycle Select value.

On some Kilauea boards the DDR autocalibration algorithm can
find a large passing write/read/compare window with a small
SDRAM_RDCC.[RDSS] aggressive value of Read Sample Cycle Select
value "T1 Sample".

This SDRAM_RDCC.[RDSS] Read Sample Cycle Select value of
"T1 Sample" proves to be to aggressive when later on U-Boot
relocates into DDR memory and executes.

The memory traces on the Kilauea board are short so on some
Kilauea boards the SDRAM_RDCC.[RDSS] Read Sample Cycle Select
value of "T1 Sample" shows up as a potentially valid value for
the DDR autocalibratiion algorithm.

The fix is to define a weak default function which provides
the minimum SDRAM_RDCC.[RDSS] Read Sample Cycle Select value
to accept for DDR autocalibration.  The default will be the
"T2 Sample" value.  A board developer who has a well defined
board and chooses to be more aggressive can always provide
their own board specific string function with the more
aggressive "T1 Sample" value or stick with the default
minimum SDRAM_RDCC.[RDSS] value of "T2".

Also put in a autocalibration loop fix for case where current
write/read/compare passing window size is the same as a prior
window size, then in this case choose the write/read/compare
result that has the associated smallest RDCC T-Sample value.

Signed-off-by: Adam Graham <agraham@amcc.com>
Signed-off-by: Stefan Roese <sr@denx.de>
15 years agoppc4xx: Fix problem with CONFIG_MAX_MEM_MAPPED in include/asm-ppc/config.h
Stefan Roese [Wed, 11 Feb 2009 08:37:12 +0000 (09:37 +0100)]
ppc4xx: Fix problem with CONFIG_MAX_MEM_MAPPED in include/asm-ppc/config.h

CONFIG_SDRAM_PPC4xx_IBM_DDR2 is not set when include/asm-ppc/config.h is
included. So for katmai, CONFIG_MAX_MEM_MAPPED will get set to 256MB.

It makes perfect sense to set CONFIG_MAX_MEM_MAPPED to 2GB for all PPC4xx
boards right now.

Signed-off-by: Stefan Roese <sr@denx.de>
15 years agoCoding style cleanup; update CHANGELOG
Wolfgang Denk [Wed, 11 Feb 2009 23:08:39 +0000 (00:08 +0100)]
Coding style cleanup; update CHANGELOG

Signed-off-by: Wolfgang Denk <wd@denx.de>
15 years agoAdd feature-removal-schedule.txt
Peter Tyser [Fri, 30 Jan 2009 22:36:40 +0000 (16:36 -0600)]
Add feature-removal-schedule.txt

Signed-off-by: Peter Tyser <ptyser@xes-inc.com>
15 years ago8xx serial, smc: Coding-Style cleanup serial SMC driver
Heiko Schocher [Tue, 10 Feb 2009 08:32:38 +0000 (09:32 +0100)]
8xx serial, smc: Coding-Style cleanup serial SMC driver

Signed-off-by: Heiko Schocher <hs@denx.de>
15 years ago8xx serial, smc: add configurable SMC Rx buffer len
Heiko Schocher [Tue, 10 Feb 2009 08:31:47 +0000 (09:31 +0100)]
8xx serial, smc: add configurable SMC Rx buffer len

This patch adds the configuration option CONFIG_SYS_SMC_RXBUFLEN.
With this option it is possible to allow the receive
buffer for the SMC on 8xx to be greater then 1. In case
CONFIG_SYS_SMC_RXBUFLEN == 1 this driver works as the
old version.

When defining CONFIG_SYS_SMC_RXBUFLEN also
CONFIG_SYS_MAXIDLE must be defined to setup the maximum
idle timeout for the SMC.

Signed-off-by: Heiko Schocher <hs@denx.de>
15 years agocommon/{hush, kgdb, serial}.c: build by COBJS-$(...) in Makefile
Mike Frysinger [Fri, 6 Feb 2009 02:04:36 +0000 (21:04 -0500)]
common/{hush, kgdb, serial}.c: build by COBJS-$(...) in Makefile

Move global '#ifdef CONFIG_xxx .... #endif' out of the .c files and into
the COBJS-$(CONFIG_xxx) in the Makefile.  Also delete unused var in kgdb
code in the process.

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
15 years agobzip2: move ifdef handling to Makefile COBJS-$(...)
Mike Frysinger [Fri, 6 Feb 2009 02:04:50 +0000 (21:04 -0500)]
bzip2: move ifdef handling to Makefile COBJS-$(...)

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
15 years agoFix whitespace damage: double space changed to a tab
Jerry Van Baren [Fri, 6 Feb 2009 03:18:02 +0000 (22:18 -0500)]
Fix whitespace damage: double space changed to a tab

At some point an intentional double space at the end of the sentence
got changed into a tab in the GPL header line:
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the

This patch fixes the damage.

Signed-off-by: Gerald Van Baren <vanbaren@cideas.com>
15 years agoMerge branch 'master' of git://git.denx.de/u-boot-cfi-flash
Wolfgang Denk [Wed, 11 Feb 2009 21:24:51 +0000 (22:24 +0100)]
Merge branch 'master' of git://git.denx.de/u-boot-cfi-flash

15 years agocfi: make flash_get_info() non static
Heiko Schocher [Tue, 10 Feb 2009 08:53:29 +0000 (09:53 +0100)]
cfi: make flash_get_info() non static

If on your board is more than one flash, you must know
the size of every single flash, for example, for updating
the DTS before booting Linux. So make this function
flash_get_info() extern, and you can have all info
about your flashes.

Signed-off-by: Heiko Schocher <hs@denx.de>
Signed-off-by: Stefan Roese <sr@denx.de>
15 years agonet: removed board-specific CONFIGs from MPC5xxx FEC driver
Ben Warren [Fri, 6 Feb 2009 07:58:25 +0000 (23:58 -0800)]
net: removed board-specific CONFIGs from MPC5xxx FEC driver

Added new CONFIG options for the three type of MAC-PHY interconnect and
applied them all relevant board config files

Signed-off-by: Ben Warren <biggerbadderben@gmail.com>
15 years agonet/sntp.c: move ifdef into Makefile COBJS-$(...)
Mike Frysinger [Fri, 6 Feb 2009 02:04:47 +0000 (21:04 -0500)]
net/sntp.c: move ifdef into Makefile COBJS-$(...)

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Ben Warren <biggerbadderben@gmail.com>
15 years agotsec: Fix a bug in soft-resetting
Andy Fleming [Wed, 4 Feb 2009 00:26:41 +0000 (18:26 -0600)]
tsec: Fix a bug in soft-resetting

SOFT_RESET must be asserted for at least 3 TX clocks.  Usually, that's about 30
clock cycles, so it's been mostly working.  But we had no guarantee, and at
slower bitrates, it's just over a microsecond (over 1000 clock cycles).  This
enforces a 2 microsecond gap between assertion and deassertion.

Signed-off-by: Andy Fleming <afleming@freescale.com>
Signed-off-by: Ben Warren <biggerbadderben@gmail.com>
15 years agoFix 100Mbs ethernet operation on sh7763 based boards
Simon Munton [Mon, 2 Feb 2009 09:44:08 +0000 (09:44 +0000)]
Fix 100Mbs ethernet operation on sh7763 based boards

100Mbs ethernet does not work on sh7763 chips due to the wrong value being
used in the GECMR register. Following diff fixes the problem

Signed-off-by: Simon Munton <simon@nidoran.m5data.com>
Signed-off-by: Ben Warren <biggerbadderben@gmail.com>
15 years agoFix MPC8260 with ethernet on SCC
ksi@koi8.net [Sat, 7 Feb 2009 00:27:55 +0000 (16:27 -0800)]
Fix MPC8260 with ethernet on SCC

This fixes MPC8260 compilation with ethernet on SCC. Probably was a
typo or something...

Signed-off-by: Sergey Kubushyn <ksi@koi8.net>
Signed-off-by: Ben Warren <biggerbadderben@gmail.com>
15 years ago82xx serial, smc: Coding-Style cleanup serial SMC driver
Heiko Schocher [Fri, 30 Jan 2009 11:56:15 +0000 (12:56 +0100)]
82xx serial, smc: Coding-Style cleanup serial SMC driver

Signed-off-by: Heiko Schocher <hs@denx.de>
15 years ago82xx serial, smc: add configurable SMC Rx buffer len
Heiko Schocher [Fri, 30 Jan 2009 11:55:38 +0000 (12:55 +0100)]
82xx serial, smc: add configurable SMC Rx buffer len

This patch adds the configuration option CONFIG_SYS_SMC_RXBUFLEN.
With this option it is possible to allow the receive
buffer for the SMC on 82xx to be greater then 1. In case
CONFIG_SYS_SMC_RXBUFLEN == 1 this driver works as the
old version.

When defining CONFIG_SYS_SMC_RXBUFLEN also
CONFIG_SYS_MAXIDLE must be defined to setup the maximum
idle timeout for the SMC.

Signed-off-by: Heiko Schocher <hs@denx.de>
15 years agoppc: Fix roll over bug in flush_cache()
Kumar Gala [Fri, 6 Feb 2009 14:08:06 +0000 (08:08 -0600)]
ppc: Fix roll over bug in flush_cache()

If we call flush_cache(0xfffff000, 0x1000) it would never
terminate the loop since end = 0xffffffff and we'd roll over
our counter from 0xfffffe0 to 0 (assuming a 32-byte cache line)

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
15 years agoppc: Move CONFIG_MAX_MEM_MAPPED to common config.h
Kumar Gala [Fri, 6 Feb 2009 02:40:58 +0000 (20:40 -0600)]
ppc: Move CONFIG_MAX_MEM_MAPPED to common config.h

Moved CONFIG_MAX_MEM_MAPPED to the asm/config.h so its kept consistent
between the two current users (lib_ppc/board.c, 44x SPD DDR2).

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Acked-by: Stefan Roese <sr@denx.de>
15 years agoAdd an architecture specific config.h for common defines
Kumar Gala [Fri, 6 Feb 2009 02:40:57 +0000 (20:40 -0600)]
Add an architecture specific config.h for common defines

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>
15 years agompc8641hpcn: Change PCI MEM pci bus address
Becky Bruce [Wed, 4 Feb 2009 00:10:56 +0000 (18:10 -0600)]
mpc8641hpcn: Change PCI MEM pci bus address

Now that the rest of u-boot can support it, change the PCI bus
address of the PCI MEM regions from 0x80000000 to 0xc0000000,
and use the same bus address for both PCI1 and PCI2.  This will
maximize the amount of PCI address space left over to map RAM
on systems with large amounts of memory.

Signed-off-by: Becky Bruce <beckyb@kernel.crashing.org>
15 years agodrivers/block/ahci: Fix pci mapping bug
Becky Bruce [Wed, 4 Feb 2009 00:10:55 +0000 (18:10 -0600)]
drivers/block/ahci: Fix pci mapping bug

The code assumes that the pci bus address and the virtual
address used to access a region are the same, but they might
not be.  Fix this assumption.

Signed-off-by: Becky Bruce <beckyb@kernel.crashing.org>
15 years agoMPC8641HPCN: Enable CONFIG_ADDR_MAP
Becky Bruce [Wed, 4 Feb 2009 00:10:54 +0000 (18:10 -0600)]
MPC8641HPCN: Enable CONFIG_ADDR_MAP

Signed-off-by: Becky Bruce <beckyb@kernel.crashing.org>
15 years agompc8641hpcn: Clean up PCI mapping concepts
Becky Bruce [Wed, 4 Feb 2009 00:10:53 +0000 (18:10 -0600)]
mpc8641hpcn: Clean up PCI mapping concepts

Clean up PCI mapping concepts in the 8641 config - rename _BASE
to _BUS, as it's actually a PCI bus address, separate virtual
and physical addresses into _VIRT and _PHYS, and use each
appopriately.

Signed-off-by: Becky Bruce <beckyb@kernel.crashing.org>
15 years agompc86xx: Add support to populate addr map based on BATs
Becky Bruce [Wed, 4 Feb 2009 00:10:52 +0000 (18:10 -0600)]
mpc86xx: Add support to populate addr map based on BATs

If CONFIG_ADDR_MAP is enabled, update the address map
whenever we write a bat.

Signed-off-by: Becky Bruce <beckyb@kernel.crashing.org>
15 years agopowerpc: Move duplicated BAT defines to mmu.h
Becky Bruce [Wed, 4 Feb 2009 00:10:51 +0000 (18:10 -0600)]
powerpc: Move duplicated BAT defines to mmu.h

The BAT fields are architected; there's no need for these to be in
cpu-specific files.  Drop the duplication and move these to
include/asm-ppc/mmu.h.  Also, remove the BL_xxx defines that were only
used by the alaska board, and switch to using the BATU_BL_xxx defines
used by all the other boards.  The BL_ defines previously in use
had to be shifted into the proper position for use, which was inefficient.

Signed-off-by: Becky Bruce <beckyb@kernel.crashing.org>
15 years agodrivers/pci: Create pci_map_bar function
Becky Bruce [Wed, 4 Feb 2009 00:10:50 +0000 (18:10 -0600)]
drivers/pci: Create pci_map_bar function

It is no longer always true that the pci bus address can be
used as the virtual address for pci accesses.  pci_map_bar()
is created to return the virtual address for a pci region.

Signed-off-by: Becky Bruce <beckyb@kernel.crashing.org>
15 years agompc8641hpcn: Set up outbound pci windows before inbound
Becky Bruce [Wed, 4 Feb 2009 00:10:49 +0000 (18:10 -0600)]
mpc8641hpcn: Set up outbound pci windows before inbound

Because the inbound pci windows are mapped generously, set up
the more specific outbound windows first.  This way, when we
search the pci regions for something, we will hit on the more
specific region.  This can actually be a problem on systems
with large amounts of RAM.

Signed-off-by: Becky Bruce <beckyb@kernel.crashing.org>
15 years agompc8641hpcn: Use physical address in flash banks defintion
Becky Bruce [Mon, 2 Feb 2009 22:34:52 +0000 (16:34 -0600)]
mpc8641hpcn: Use physical address in flash banks defintion

If the VA and PA of the flash aren't the same, the banks list
should be initialized to hold the physical address.  Correct this.

Signed-off-by: Becky Bruce <beckyb@kernel.crashing.org>
15 years agoMerge branch 'master' of ssh://gemini/home/wd/git/u-boot/master
Wolfgang Denk [Sat, 7 Feb 2009 22:51:52 +0000 (23:51 +0100)]
Merge branch 'master' of ssh://gemini/home/wd/git/u-boot/master

15 years agopci: give preference to non-PCI_REGION_SYS_MEMORY regions when matching
Kumar Gala [Fri, 6 Feb 2009 15:49:32 +0000 (09:49 -0600)]
pci: give preference to non-PCI_REGION_SYS_MEMORY regions when matching

When we search for an address match in pci_hose_{phys_to_bus,bus_to_phys}
we should give preference to memory regions that aren't system memory.

Its possible that we have over mapped system memory in the regions and
we want to avoid depending on the order of the regions.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
15 years agopci: Rename PCI_REGION_MEMORY to PCI_REGION_SYS_MEMORY for clarity
Kumar Gala [Fri, 6 Feb 2009 15:49:31 +0000 (09:49 -0600)]
pci: Rename PCI_REGION_MEMORY to PCI_REGION_SYS_MEMORY for clarity

The PCI_REGION_MEMORY and PCI_REGION_MEM are a bit to similar and
can be confusing when reading the code.

Rename PCI_REGION_MEMORY to PCI_REGION_SYS_MEMORY to clarify its used
for system memory mapping purposes.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
15 years agoMerge branch 'master' of git://git.denx.de/u-boot-nand-flash
Wolfgang Denk [Sat, 7 Feb 2009 22:37:10 +0000 (23:37 +0100)]
Merge branch 'master' of git://git.denx.de/u-boot-nand-flash

15 years agoMerge branch 'master' of git://git.denx.de/u-boot-coldfire
Wolfgang Denk [Sat, 7 Feb 2009 22:24:38 +0000 (23:24 +0100)]
Merge branch 'master' of git://git.denx.de/u-boot-coldfire

15 years agoMerge branch 'master' of git://git.denx.de/u-boot-blackfin
Wolfgang Denk [Sat, 7 Feb 2009 21:53:45 +0000 (22:53 +0100)]
Merge branch 'master' of git://git.denx.de/u-boot-blackfin

15 years agoMerge branch 'master' of git://git.denx.de/u-boot-cfi-flash
Wolfgang Denk [Sat, 7 Feb 2009 21:17:44 +0000 (22:17 +0100)]
Merge branch 'master' of git://git.denx.de/u-boot-cfi-flash

15 years agoMerge branch 'master' of git://git.denx.de/u-boot-ppc4xx
Wolfgang Denk [Sat, 7 Feb 2009 21:17:19 +0000 (22:17 +0100)]
Merge branch 'master' of git://git.denx.de/u-boot-ppc4xx

15 years agoMerge branch 'master' of git://git.denx.de/u-boot-mpc5xxx
Wolfgang Denk [Sat, 7 Feb 2009 21:08:53 +0000 (22:08 +0100)]
Merge branch 'master' of git://git.denx.de/u-boot-mpc5xxx

15 years agoNAND: Add timeout for reset command
Peter Tyser [Wed, 4 Feb 2009 19:47:22 +0000 (13:47 -0600)]
NAND: Add timeout for reset command

Without the timeout present an infinite loop can occur if the
NAND device is broken or not present.

Signed-off-by: Peter Tyser <ptyser@xes-inc.com>
Signed-off-by: Scott Wood <scottwood@freescale.com>
15 years agoNAND: Silence warning when CONFIG_SYS_NAND_QUIET_TEST
Peter Tyser [Wed, 4 Feb 2009 19:39:40 +0000 (13:39 -0600)]
NAND: Silence warning when CONFIG_SYS_NAND_QUIET_TEST

Commit cfa460adfdefcc30d104e1a9ee44994ee349bb7b removed support
for disabling the "No NAND device found!!!" warning when
CONFIG_SYS_NAND_QUIET_TEST was defined.  This re-adds support
for silencing the warning.

Signed-off-by: Peter Tyser <ptyser@xes-inc.com>
Signed-off-by: Scott Wood <scottwood@freescale.com>
15 years agoNAND: Fixed invalid pointers to static relocated chip names
Valeriy Glushkov [Mon, 19 Jan 2009 14:32:59 +0000 (16:32 +0200)]
NAND: Fixed invalid pointers to static relocated chip names

Dear Wolfgang,

You are right, the patch was ugly.
The new one seems to be better.

Signed-off-by: Valeriy Glushkov <gvv@lstec.com>
Signed-off-by: Scott Wood <scottwood@freescale.com>
15 years agoenv_nand: fix env memory release
derek@siconix.com [Mon, 26 Jan 2009 21:08:17 +0000 (14:08 -0700)]
env_nand: fix env memory release

This fixes a bug that tmp environment memory not being released.

Signed-off-by: Derek Ou <dou@siconix.com>
Signed-off-by: Scott Wood <scottwood@freescale.com>
15 years agoColdfire: M527x: Add missing GPIO register address defines
Richard Retanubun [Thu, 5 Feb 2009 14:33:50 +0000 (09:33 -0500)]
Coldfire: M527x: Add missing GPIO register address defines

Add missing GPIO registers address definition for Coldfire M5271.

Signed-off-by: Richard Retanubun <RichardRetanubun@RuggedCom.com>
15 years agoColdfire: mcfmii: Allow non-autonegotiating PHYs to use mii command
Richard Retanubun [Fri, 23 Jan 2009 19:42:58 +0000 (14:42 -0500)]
Coldfire: mcfmii: Allow non-autonegotiating PHYs to use mii command

Modified mii_init to support boards with PHYs that are not set to
autonegotiate, but still want to use u-boot's mii commands to probe
the smi bus. Such PHYs will not set the Autonegotiate-done bit.

Signed-off-by: Richard Retanubun <RichardRetanubun@RuggedCom.com>
15 years agoColdfire: Applied baudrate formula of serial_init to serial_setbrg
Richard Retanubun [Fri, 23 Jan 2009 16:44:30 +0000 (11:44 -0500)]
Coldfire: Applied baudrate formula of serial_init to serial_setbrg

Applied the patch for baudrate divider value truncation for
serial_init to serial_setbrg as well.

Signed-off-by: Richard Retanubun <RichardRetanubun@RuggedCom.com>
15 years agoColdfire: M5271EVB: Board header update (dependencies)
Richard Retanubun [Fri, 23 Jan 2009 19:07:05 +0000 (14:07 -0500)]
Coldfire: M5271EVB: Board header update (dependencies)

Cleanup for M5271EVB:
Added clarification on the use of CONFIG_SYS_CLOCK.
Modified to use u-boot's HUSH parser.
Cleanup on environment settings.
Removed compiler warning by defining CONFIG_SYS_CS0_*

Dependencies:
Added the use of CONFIG_SYS_MCF_SYNCR for clock multiplier.
This depends on a patch to include/asm-m68k/m5271.h
that defines the multiplier and divider ratios.

Removed the definition of CONFIG_SYS_FECI2C.
This depends on a patch that removes the use of it in
cpu/mcf52x2/cpu_init.c

Signed-off-by: Richard Retanubun <RichardRetanubun@RuggedCom.com>
15 years agoColdfire: M5271: Allow board header file to specify clock multiplier
Richard Retanubun [Thu, 29 Jan 2009 19:36:06 +0000 (14:36 -0500)]
Coldfire: M5271: Allow board header file to specify clock multiplier

M5271 dynamic clock multiplier. It is currently fixed at 100MHz.

Allow the board header file to set their own multiplier and divider.
Added the #define for the multiplier and divider to the cpu header file.

Signed-off-by: Richard Retanubun <RichardRetanubun@RuggedCom.com>
15 years agoColdfire: M5271EVB: Remove usage of CONFIG_SYS_FECI2C
Richard Retanubun [Fri, 23 Jan 2009 15:47:13 +0000 (10:47 -0500)]
Coldfire: M5271EVB: Remove usage of CONFIG_SYS_FECI2C

Discontinue the use of CONFIG_SYS_FECI2C (only used by M5271EVB).
Use read-modify-write to activate the FEC pins without disabling I2C.

Signed-off-by: Richard Retanubun <RichardRetanubun@RuggedCom.com>
15 years agoColdfire: cmd_bdinfo cleanup
Richard Retanubun [Fri, 23 Jan 2009 14:45:34 +0000 (09:45 -0500)]
Coldfire: cmd_bdinfo cleanup

CONFIG_M68K bdinfo cleanup:

Fixed compiler warning about baudrate printing.
format '%d' expects type 'int', but argument 2 has type 'long unsigned int'.

Added printing of "cpufreq"

Signed-off-by: Richard Retanubun <RichardRetanubun@RuggedCom.com>
15 years agoColdfire: Fix half-baud UART by adding M5271 to Coldfire v2 core list
Richard Retanubun [Fri, 23 Jan 2009 14:27:00 +0000 (09:27 -0500)]
Coldfire: Fix half-baud UART by adding M5271 to Coldfire v2 core list

Added the CONFIG_M5271 to the list of Coldfire V2 processor. This
was causing the bus clock (not CPU clock) to be declared twice as
fast as it actually is. This causes UARTS to operate at half the
specified baudrate.

Signed-off-by: Richard Retanubun <RichardRetanubun@RuggedCom.com>
15 years agoppc4xx: Make PCIE support selectable
Dirk Eibach [Tue, 3 Feb 2009 14:15:21 +0000 (15:15 +0100)]
ppc4xx: Make PCIE support selectable

On some platforms PCIE support is not required, but would be included
because the cpu supports it. To reduce fooprint it is now configurable
via CONFIG_PCI_DISABLE_PCIE.

Signed-off-by: Dirk Eibach <eibach@gdsys.de>
Signed-off-by: Stefan Roese <sr@denx.de>
15 years agoppc4xx: Only fixup opb attached UARTs
Matthias Fuchs [Tue, 3 Feb 2009 21:13:16 +0000 (22:13 +0100)]
ppc4xx: Only fixup opb attached UARTs

This patch updates the fdt UART clock fixup code to
only touch CPU internal UARTs on 4xx systems.
Only these UARTs are definitely clocked by gd->uart_clk.

Signed-off-by: Matthias Fuchs <matthias.fuchs@esd-electronics.com>
Signed-off-by: Stefan Roese <sr@denx.de>
15 years agoBlackfin: move default boot SPI CS to common code
Mike Frysinger [Tue, 13 Jan 2009 16:00:29 +0000 (11:00 -0500)]
Blackfin: move default boot SPI CS to common code

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>
15 years agoBlackfin: dynamically update UART speed when initializing
Mike Frysinger [Wed, 10 Dec 2008 17:33:54 +0000 (12:33 -0500)]
Blackfin: dynamically update UART speed when initializing

Previously, booting over the UART required the baud rate to be known ahead
of time.  Using a bit of tricky simple math, we can calculate the new board
rate based on the old divisors.

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Robin Getz <rgetz@blackfin.uclinux.org>
15 years agoBlackfin: add support for fast SPI reads with Boot ROM
Mike Frysinger [Tue, 9 Dec 2008 22:21:08 +0000 (17:21 -0500)]
Blackfin: add support for fast SPI reads with Boot ROM

Newer Blackfin boot roms support using the fast SPI read command rather than
just the slow one.  If the functionality is available, then use it.

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
15 years agoBlackfin: check for reserved settings in DDR MMRs
Mike Frysinger [Sun, 12 Oct 2008 01:46:52 +0000 (21:46 -0400)]
Blackfin: check for reserved settings in DDR MMRs

Some bits of the DDR MMRs should not be set.  If they do, bad things may
happen (like random failures or hardware destruction).

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
15 years agoBlackfin: set default voltage levels for BF538/BF539 parts
Mike Frysinger [Sun, 12 Oct 2008 01:54:00 +0000 (21:54 -0400)]
Blackfin: set default voltage levels for BF538/BF539 parts

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
15 years agoBlackfin: use on-chip syscontrol() rom function when available
Mike Frysinger [Sun, 1 Jun 2008 05:29:57 +0000 (01:29 -0400)]
Blackfin: use on-chip syscontrol() rom function when available

Newer Blackfin's have an on-chip rom with a syscontrol() function that needs
to be used to properly program the memory and voltage settings as it will
include (possibly critical) factory tested bias values.

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
15 years agocfi_flash: Fix typo in cfi_flash.c
Stefan Roese [Thu, 5 Feb 2009 10:44:52 +0000 (11:44 +0100)]
cfi_flash: Fix typo in cfi_flash.c

Patch "flash/cfi_flash: Use virtual sector start address, not phys"
introduced a small typo and compilation warning for systems with CFI
legacy support (e.g. hcu4). This patch fixes it.

Signed-off-by: Stefan Roese <sr@denx.de>
15 years agojedec_flash: Only use manufacturer defines from common flash.h
Stefan Roese [Thu, 29 Jan 2009 10:21:38 +0000 (11:21 +0100)]
jedec_flash: Only use manufacturer defines from common flash.h

This patch removes the double defined manufacturer defines from
jedec_flash.c. Since the common defines in flash.h are 32bit
we now need the (16) cast. This patch also removes the compilation
warning (e.g. seen on hcu5):

./MAKEALL hcu5
Configuring for hcu5 board...
jedec_flash.c:219: warning: large integer implicitly truncated to unsigned type

Signed-off-by: Stefan Roese <sr@denx.de>
15 years agocfi_flash: Silence compilation warning
Stefan Roese [Thu, 5 Feb 2009 10:25:57 +0000 (11:25 +0100)]
cfi_flash: Silence compilation warning

Patch "flash/cfi_flash: Use virtual sector start address, not phys"
introduced a small compilation warning. This patch fixes it.

Signed-off-by: Stefan Roese <sr@denx.de>
15 years agoflash/cfi_flash: Use virtual sector start address, not phys
Becky Bruce [Mon, 2 Feb 2009 22:34:51 +0000 (16:34 -0600)]
flash/cfi_flash: Use virtual sector start address, not phys

include/flash.h was commented to say that the address in
flash_info->start was a physical address.  However, from u-boot's
point of view, and looking at most flash code, it makes more
sense for this to be a virtual address.  So I corrected the
comment to indicate that this was a virtual address.

The only flash driver that was actually treating the address
as physical was the mtd/cfi_flash driver.  However, this code
was using it inconsistently as it actually directly dereferenced
the "start" element, while it used map_physmem to get a
virtual address in other places.  I changed this driver so
that the code which initializes the info->start field calls
map_physmem to get a virtual address, eliminating the need for
further map_physmem calls.  The code is now consistent.

The *only* place a physical address should be used is when defining the
flash banks list that is used to initialize the flash_info struct,
usually found in the board config file.

Signed-off-by: Becky Bruce <beckyb@kernel.crashing.org>
Signed-off-by: Stefan Roese <sr@denx.de>
15 years agoFix compiler warning
Wolfgang Denk [Wed, 4 Feb 2009 08:42:20 +0000 (09:42 +0100)]
Fix compiler warning

(shows up only when DEBUG is enabled)

Signed-off-by: Wolfgang Denk <wd@denx.de>
15 years agoBlackfin: update anomaly lists
Mike Frysinger [Mon, 6 Oct 2008 07:45:55 +0000 (03:45 -0400)]
Blackfin: update anomaly lists

Update the anomaly lists to match latest anomaly sheets.

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
15 years agoADS5121 Add PATA support
Ralph Kondziella [Mon, 26 Jan 2009 19:34:36 +0000 (12:34 -0700)]
ADS5121 Add PATA support

Original patch from Ralph Kondziella
plus clean up by Wolfgang Denk
plus changes by John Rigby
    use ips clock not lpc
    port forward to current u-boot release

Signed-off-by: Ralph Kondziella <rk@argos-messtechnik.de>
Signed-off-by: Wolfgang Denk <wd@denx.de>
Signed-off-by: John Rigby <jrigby@freescale.com>
15 years agoADS5121 Add IC Ident Module (IIM) support
Martha Marx [Mon, 26 Jan 2009 17:45:07 +0000 (10:45 -0700)]
ADS5121 Add IC Ident Module (IIM) support

IIM (IC Identification Module) is the fusebox for the mpc5121.
Use #define CONFIG_IIM to turn on the clock for this module
use #define CONFIG_CMD_FUSE to add fusebox commands.
Fusebox commands include the ability to read
the status, read the register cache, override the register cache,
program the fuses and sense them.

Signed-off-by: Martha Marx <mmarx@silicontkx.com>
Signed-off-by: John Rigby <jrigby@freescale.com>
15 years agoMerge branch 'master' of git://git.denx.de/u-boot-usb
Wolfgang Denk [Tue, 3 Feb 2009 22:16:15 +0000 (23:16 +0100)]
Merge branch 'master' of git://git.denx.de/u-boot-usb

15 years agoADS5121 Fix rev2 silicon pci iopad config
John Rigby [Fri, 23 Jan 2009 17:33:15 +0000 (10:33 -0700)]
ADS5121 Fix rev2 silicon pci iopad config

Reset config is not correct

Signed-off-by: John Rigby <jrigby@freescale.com>
15 years agoADS5121 DIU Add diu_bmp_addr env
John Rigby [Wed, 19 Nov 2008 20:57:34 +0000 (13:57 -0700)]
ADS5121 DIU Add diu_bmp_addr env

Add support for using a bmp other than
FSL_Logo_BMP for the DIU splash screen.

Can now set the env var "diu_bmp_addr" to
the address of a BMP in flash to use instead
of the default FSL_Logo_BMP.

Signed-off-by: Martha Marx <mmarx@silicontkx.com>
Signed-off-by: John Rigby <jrigby@freescale.com>
15 years agoADS5121 DIU Make inclusion of FSL logo optional
John Rigby [Thu, 30 Oct 2008 22:39:35 +0000 (16:39 -0600)]
ADS5121 DIU Make inclusion of FSL logo optional

Make inclusion of FSL logo optional and
turn it off by default.

Signed-off-by: John Rigby <jrigby@freescale.com>
15 years agoCompile warning fix in onenand_uboot.h
Remy Bohmer [Sun, 1 Feb 2009 11:27:53 +0000 (12:27 +0100)]
Compile warning fix in onenand_uboot.h

Regression since merge window after 2009.01

Signed-off-by: Remy Bohmer <linux@bohmer.net>
15 years agoUSB: Add EHCI support for VCT EHCI controller (really with driver now)
Stefan Roese [Thu, 29 Jan 2009 05:33:55 +0000 (06:33 +0100)]
USB: Add EHCI support for VCT EHCI controller (really with driver now)

Somehow I missed the real driver part in my last patch version. This patch
now adds the driver.

Signed-off-by: Stefan Roese <sr@denx.de>
Signed-off-by: Remy Bohmer <linux@bohmer.net>
15 years agoBlackfin: add driver for on-chip MMC/SD controller
Cliff Cai [Sat, 29 Nov 2008 23:22:38 +0000 (18:22 -0500)]
Blackfin: add driver for on-chip MMC/SD controller

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>
15 years agoBlackfin: add port muxing for BF51x SPI
Mike Frysinger [Sun, 12 Oct 2008 02:47:34 +0000 (22:47 -0400)]
Blackfin: add port muxing for BF51x SPI

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
15 years agoBlackfin: output booting source when booting
Mike Frysinger [Tue, 6 Jan 2009 11:16:19 +0000 (06:16 -0500)]
Blackfin: output booting source when booting

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>
15 years agoBlackfin: set default CONFIG_ENV_SPI_CS based on bootrom
Mike Frysinger [Thu, 11 Dec 2008 11:30:46 +0000 (06:30 -0500)]
Blackfin: set default CONFIG_ENV_SPI_CS based on bootrom

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>
15 years agoBlackfin: update asm-blackfin/posix_types.h to latest Linux version
Mike Frysinger [Thu, 11 Dec 2008 09:06:26 +0000 (04:06 -0500)]
Blackfin: update asm-blackfin/posix_types.h to latest Linux version

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
15 years agoBlackfin: add port I bits
Mike Frysinger [Sat, 6 Dec 2008 07:54:52 +0000 (02:54 -0500)]
Blackfin: add port I bits

Some people need to access port I, so make sure the pins are defined.

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
15 years agoBlackfin: add driver for on-chip ATAPI controller
Sonic Zhang [Thu, 27 Nov 2008 03:16:45 +0000 (22:16 -0500)]
Blackfin: add driver for on-chip ATAPI controller

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>
15 years agoBlackfin: add driver for on-chip NAND controller
Mike Frysinger [Thu, 27 Nov 2008 02:43:06 +0000 (21:43 -0500)]
Blackfin: add driver for on-chip NAND controller

This is a port of the Linux Blackfin on-chip NFC driver to U-Boot.

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
15 years agoBlackfin: build with -mno-fdpic
Mike Frysinger [Wed, 12 Nov 2008 12:18:15 +0000 (07:18 -0500)]
Blackfin: build with -mno-fdpic

Use the -mno-fdpic flag so that any Blackfin toolchain can be used to build
up u-boot, including ones that output FDPIC ELF by default.

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
15 years agoBlackfin: fix up EBIU defines
Mike Frysinger [Tue, 11 Nov 2008 10:43:57 +0000 (05:43 -0500)]
Blackfin: fix up EBIU defines

The EBIU defines for EBSZ 256/512 were incorrect.

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
15 years agoBlackfin: use 8/16/32 bit transfer widths in dma_memcpy()
Mike Frysinger [Wed, 5 Nov 2008 17:45:24 +0000 (12:45 -0500)]
Blackfin: use 8/16/32 bit transfer widths in dma_memcpy()

Rather than using 8bit transfers for everything, use 8/16/32 bit transfers
as usable with the source/destination addresses and the count size.

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
15 years agoBlackfin: only flag L1 instruction for DMA memcpy
Mike Frysinger [Wed, 5 Nov 2008 13:50:23 +0000 (08:50 -0500)]
Blackfin: only flag L1 instruction for DMA memcpy

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>
15 years agoBlackfin: dma_memcpy(): fix random failures
Mike Frysinger [Wed, 5 Nov 2008 12:20:37 +0000 (07:20 -0500)]
Blackfin: dma_memcpy(): fix random failures

We have to make sure the DMA channel is actually disabled in hardware before
attempting to reprogram it.  Otherwise the new settings are ignored and we
end up with random hangs/failures.

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
15 years agoBlackfin: rewrite cache handling functions
Mike Frysinger [Tue, 4 Nov 2008 05:04:03 +0000 (00:04 -0500)]
Blackfin: rewrite cache handling functions

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>
15 years agoBlackfin: setup bi_enetaddr for single nets
Mike Frysinger [Tue, 4 Nov 2008 03:30:05 +0000 (22:30 -0500)]
Blackfin: setup bi_enetaddr for single nets

For systems with CONFIG_NET_MULTI disabled, bi_enetaddr does not get setup
based on $ethaddr, so set it up.

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
15 years agoBlackfin: cache core/system clock values
Mike Frysinger [Sat, 25 Oct 2008 02:48:47 +0000 (22:48 -0400)]
Blackfin: cache core/system clock values

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>
15 years agoBlackfin: enable --gc-sections
Mike Frysinger [Fri, 24 Oct 2008 22:18:16 +0000 (18:18 -0400)]
Blackfin: enable --gc-sections

Start building all Blackfin boards with -ffunction-sections/-fdata-sections
and linking with --gc-sections.

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
15 years agoBlackfin: dont check baud if it wont actually get used
Mike Frysinger [Tue, 21 Oct 2008 01:08:54 +0000 (21:08 -0400)]
Blackfin: dont check baud if it wont actually get used

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
15 years agoBlackfin: add driver for on-chip SPI controller
Mike Frysinger [Tue, 14 Oct 2008 11:54:09 +0000 (07:54 -0400)]
Blackfin: add driver for on-chip SPI controller

This fills out the SPI backend for the Blackfin on-chip SPI peripheral.

Signed-off-by: Mike Frysinger <vapier@gentoo.org>