]> git.sur5r.net Git - u-boot/log
u-boot
15 years agoMisc ti_pci1410a fixups
Graeme Russ [Sun, 23 Aug 2009 02:59:52 +0000 (12:59 +1000)]
Misc ti_pci1410a fixups

Removed do_pinit() - now declared in cmd_pcmcia.c

Added #define CONFIG_CMD_PCMCIA around pcmcia_off() in line with other
PCMCIA drivers

signed/unsigned type fixups

Added semi-colon after default: label as required by newer gcc

The only board that appears to use this driver is the sc520_spunk which
is very old and very likely very broken anyway. I do not have one to test
whether this patch breaks anything functionaly, I have can only check
that it compiles without warning or error

Signed-off-by: Graeme Russ <graeme.russ@gmail.com>
15 years agoMisc SATA fixups
Graeme Russ [Sun, 23 Aug 2009 02:59:51 +0000 (12:59 +1000)]
Misc SATA fixups

Cast first parameter to sata_cpy()

In /drivers/block/ata_piix.h, ata_id_has_lba48(), ata_id_has_lba(),
ata_id_has_dma(), ata_id_u32(), ata_id_u64() are all defined in
include/libata.h which is included in ata.h which is included by all files
which include ata_piix.h (only ata_piix.c) so these definitions are
supurflous to (and conlict with) this in libata.h. Interestingly, my
compiler complains about ata_id_u64 already being defined, but not
ata_id_u32

ata_dump_id() is defined in include/libata.h and should not be static
(maybe should even use ata_dump_id() in libata.c

Signed-off-by: Graeme Russ <graeme.russ@gmail.com>
15 years agoi386: Misc PCI fixups
Graeme Russ [Sun, 23 Aug 2009 02:59:50 +0000 (12:59 +1000)]
i386: Misc PCI fixups

Change PCI_REGION_MEMORY to PCI_REGION_SYS_MEMORY (Originally done in
commit ff4e66e93c1a, regressed by commit 6d7f610b09f8)

Cast PCI_ROM_ADDRESS_MASK to u32

Wrap probe_pci_video() call inside #ifdef CONFIG_VIDEO

Change call to pci_find_class() to pci_find_devices(). This is based on a
patch submitted on 1st March 2007 (Patch that fixes the compilation errors
for sc520_cdp board) by mushtaq_k

This patch requires that PCI_VIDEO_VENDOR_ID and PCI_VIDEO_DEVICE_ID be
specified in the board config file.  Dummy values have been added for the
SC520 CDP board to enable compilation, but since I do not have one of these,
I do know what the values should be

Signed-off-by: Graeme Russ <graeme.russ@gmail.com>
15 years agoFix sc520 timer interrupt generation
Graeme Russ [Sun, 23 Aug 2009 02:59:49 +0000 (12:59 +1000)]
Fix sc520 timer interrupt generation

The current implementation has the timer being started before the interrupt
handler is installed. It the interrupt occurs before the handler is
installed, the timer interrupt is never reset and the timer stops

Signed-off-by: Graeme Russ <graeme.russ@gmail.com>
15 years agoFix environment configuration for eNET board
Graeme Russ [Sun, 23 Aug 2009 02:59:48 +0000 (12:59 +1000)]
Fix environment configuration for eNET board

The current configuration of the Environment has the redundant copy of the
environment in the Boot Flash - This was never the intent. The Environment
should instead be in the first two sectors of the first Strata Flash

Signed-off-by: Graeme Russ <graeme.russ@gmail.com>
15 years agoi386: Fix regression introduced by commit 8c63d47651f7
Graeme Russ [Sun, 23 Aug 2009 02:59:47 +0000 (12:59 +1000)]
i386: Fix regression introduced by commit 8c63d47651f7

A local variable was deleted that should not have been

Signed-off-by: Graeme Russ <graeme.russ@gmail.com>
15 years agoi386: Change inline asm global symbols to local
Graeme Russ [Sun, 23 Aug 2009 02:59:46 +0000 (12:59 +1000)]
i386: Change inline asm global symbols to local

gcc 4.3.2 optimiser creates multiple copies of inline asm (who knows why)
Remove use of global names for labels to prevent 'symbol already defined'
errors

Signed-off-by: Graeme Russ <graeme.russ@gmail.com>
15 years agoi386: Add errno.h
Graeme Russ [Sun, 23 Aug 2009 02:59:45 +0000 (12:59 +1000)]
i386: Add errno.h

Signed-off-by: Graeme Russ <graeme.russ@gmail.com>
15 years agoConsolidate arch-specific mem_malloc_init() implementations
Peter Tyser [Sat, 22 Aug 2009 04:05:21 +0000 (23:05 -0500)]
Consolidate arch-specific mem_malloc_init() implementations

Signed-off-by: Peter Tyser <ptyser@xes-inc.com>
15 years agoStandardize mem_malloc_init() implementation
Peter Tyser [Sat, 22 Aug 2009 04:05:20 +0000 (23:05 -0500)]
Standardize mem_malloc_init() implementation

This lays the groundwork to allow architectures to share a common
mem_malloc_init().

Note that the x86 implementation was not modified as it did not fit the
mold of all other architectures.

Signed-off-by: Peter Tyser <ptyser@xes-inc.com>
15 years agoConsolidate arch-specific sbrk() implementations
Peter Tyser [Sat, 22 Aug 2009 04:05:19 +0000 (23:05 -0500)]
Consolidate arch-specific sbrk() implementations

Signed-off-by: Peter Tyser <ptyser@xes-inc.com>
15 years agoatmel_df_pow2: standalone to convert dataflashes to pow2
Mike Frysinger [Thu, 23 Jul 2009 20:37:03 +0000 (16:37 -0400)]
atmel_df_pow2: standalone to convert dataflashes to pow2

Atmel DataFlashes by default operate with pages that are slightly bigger
than normal binary sizes (i.e. many are 1056 byte pages rather than 1024
bytes).  However, they also have a "power of 2" mode where the pages show
up with the normal binary size.  The latter mode is required in order to
boot with a Blackfin processor, so many people wish to convert their
DataFlashes on their development systems to this mode.  This standalone
application does just that.

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
15 years agoBlackfin: cm-bf548: fix device->stdio_dev fallout
Mike Frysinger [Wed, 2 Sep 2009 09:52:37 +0000 (05:52 -0400)]
Blackfin: cm-bf548: fix device->stdio_dev fallout

The recent 52cb4d4fb348 commit which renamed device to stdio_dev missed the
cm-bf548's video board.

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
15 years agoBlackfin: enable 64bit printf for nand
Mike Frysinger [Wed, 2 Sep 2009 08:21:16 +0000 (04:21 -0400)]
Blackfin: enable 64bit printf for nand

Since the NAND code now uses 64bit code, make sure we enable support for
ADI Blackfin boards in printf to avoid the warning:
nand_util.c:45:2: warning: #warning Please define CONFIG_SYS_64BIT_VSPRINTF for correct output!

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
15 years agoBlackfin: use scratch pad for exception stack
Mike Frysinger [Tue, 25 Aug 2009 00:48:04 +0000 (20:48 -0400)]
Blackfin: use scratch pad for exception stack

If the memory layout pushes the stack out of the default DCPLB coverage,
the exception handler may trigger a double fault by trying to push onto
the uncovered stack.  So handle the exception stack similar to the kernel
by using the top of the scratch pad SRAM.

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
15 years agoBlackfin: increase default console size
Mike Frysinger [Tue, 25 Aug 2009 00:36:25 +0000 (20:36 -0400)]
Blackfin: increase default console size

The default console size indirectly applies to length of env vars, so a
smaller length makes it hard to pass longer command lines to kernels.

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
15 years agoBlackfin: fix debug printf modifiers
Mike Frysinger [Mon, 24 Aug 2009 23:03:18 +0000 (19:03 -0400)]
Blackfin: fix debug printf modifiers

The display_global_data() function generated warnings with pretty much
every variable.

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
15 years agoBlackfin: cm-bf537u: new board port
Harald Krapfenbauer [Thu, 20 Aug 2009 23:20:41 +0000 (19:20 -0400)]
Blackfin: cm-bf537u: new board port

The CM-BF537U is similar to the CM-BF537E module, but enough to need its
own board port.

Signed-off-by: Harald Krapfenbauer <Harald.Krapfenbauer@bluetechnix.at>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
15 years agoBlackfin: change global data register from P5 to P3
Robin Getz [Mon, 17 Aug 2009 15:23:02 +0000 (15:23 +0000)]
Blackfin: change global data register from P5 to P3

Since the Blackfin ABI favors higher scratch registers by default, use the
last scratch register (P3) for global data rather than the first (P5).
This allows the compiler's register allocator to use higher number scratch
P registers, which in turn better matches the Blackfin instruction set,
which reduces the size of U-Boot by more than 1024 bytes...

Signed-off-by: Robin Getz <robin.getz@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
15 years agoBlackfin: enable more network commands for ADI dev boards
Robin Getz [Tue, 11 Aug 2009 14:20:13 +0000 (14:20 +0000)]
Blackfin: enable more network commands for ADI dev boards

Add dns and ntp to default networking commands, and ask for more dhcp
options to better configure the network environment.

Signed-off-by: Robin Getz <robin.getz@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
15 years agoBlackfin: bf537-stamp: comment CF-Flash Card Support better
Michael Hennerich [Thu, 18 Jun 2009 09:12:50 +0000 (09:12 +0000)]
Blackfin: bf537-stamp: comment CF-Flash Card Support better

Signed-off-by: Michael Hennerich <michael.hennerich@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
15 years agoBlackfin: use +(filesize) to make sure we are only doing what is necessary
Robin Getz [Fri, 10 Jul 2009 18:37:15 +0000 (18:37 +0000)]
Blackfin: use +(filesize) to make sure we are only doing what is necessary

Signed-off-by: Robin Getz <robin.getz@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
15 years agoTI DaVinci: DM646x: Initial Support for DM646x SOC
Sandeep Paulraj [Tue, 8 Sep 2009 15:37:39 +0000 (11:37 -0400)]
TI DaVinci: DM646x: Initial Support for DM646x SOC

DM646x is an SOC from TI which has both an ARM and a DSP.
There are multiple variants of the SOC mainly dealing with different
core speeds.
This patch adds the initial framework for the DM646x SOC.

Signed-off-by: Sandeep Paulraj <s-paulraj@ti.com>
15 years agoTI DaVinci: DM6446: Fix Compilation error in NAND mode
Sandeep Paulraj [Wed, 9 Sep 2009 19:26:00 +0000 (15:26 -0400)]
TI DaVinci: DM6446: Fix Compilation error in NAND mode

The Default mode that is built for the Davinci DVEVM happens
to be the NOR mode.
When we want to build for the NAND mode, we get a compilation
error. This is overcome by defining the CONFIG_MTD_DEVICE
flag in the NAND mode.
The image built for NAND mode was successfully tested on the
DaVinci DM6446 EVM.

Signed-off-by: Sandeep Paulraj <s-paulraj@ti.com>
15 years agoOMAP3 Move cache routine to cache.S
Tom Rix [Thu, 10 Sep 2009 19:27:57 +0000 (15:27 -0400)]
OMAP3 Move cache routine to cache.S

v7_flush_dcache_all, because it depends on omap ROM code is not
generic.  Rename the function to 'invalidate_dcache' and move it
to the omap cpu directory.

Collect the other omap cache routines l2_cache_enable and
l2_cache_disable with invalide_dcache into cache.S.  This
means removing the old cache.c file that contained l2_cache_enable
and l2_cache_disable.

The conversion from cache.c to cache.S was done most through
disassembling the uboot binary.  The only significant change was
to change the comparision for the return of get_cpu_rev from

   cmp r0, #0
   beq earlier_than_label

Which was lost information to

   cmp r0, #CPU_3XX_ES20
   blt earlier_than_label

The paths through the enable routine were verified by
adding an infinite loop and seeing the hang.  Then
removing the infinite loop and seeing it continue.

The disable routine is similar enough that it was not
tested with this method.

Run tested by cold booting from nand on beagle and zoom1.
Compile tested on MAKEALL arm.

Signed-off-by: Tom Rix <Tom.Rix@windriver.com>
15 years agoTI DaVinci: Remove references to SZ_xx
Sandeep Paulraj [Tue, 8 Sep 2009 21:09:52 +0000 (17:09 -0400)]
TI DaVinci: Remove references to SZ_xx

This patch removes the asm/sizes.h header file from being
included in the DaVinci SOC configs.
References to SZ_xx have been replaced by appropriate
bit shifted values.

Signed-off-by: Sandeep Paulraj <s-paulraj@ti.com>
Acked-by: Wolfgang Denk <wd@denx.de>
15 years agoARM: DaVinci: Adding Support for DaVinci DM365 EVM
Sandeep Paulraj [Sat, 15 Aug 2009 15:20:58 +0000 (11:20 -0400)]
ARM: DaVinci: Adding Support for DaVinci DM365 EVM

This patch adds support for the DM365 EVM.
It has been tested on a DM365 EVM.

Signed-off-by: Sandeep Paulraj <s-paulraj@ti.com>
15 years agoARM: DaVinci: DaVinci DM365 SOC specific code
Sandeep Paulraj [Sat, 15 Aug 2009 15:20:44 +0000 (11:20 -0400)]
ARM: DaVinci: DaVinci DM365 SOC specific code

This patch adds support for DaVinci DM365 SOC.

Signed-off-by: Sandeep Paulraj <s-paulraj@ti.com>
15 years agoDM9000 init for pm9261
Ilko Iliev [Sat, 5 Sep 2009 00:51:34 +0000 (02:51 +0200)]
DM9000 init for pm9261

Signed-off-by: Ilko Iliev <iliev@ronetix.at>
15 years agoAdd support for the DevKit8000 board
Frederik Kriewitz [Sun, 23 Aug 2009 10:56:42 +0000 (12:56 +0200)]
Add support for the DevKit8000 board

This patch adds support for the DevKit8000 board.

Signed-off-by: Frederik Kriewitz <frederik@kriewitz.eu>
15 years agoomap3: move the other boards to board/
Jean-Christophe PLAGNIOL-VILLARD [Sun, 23 Aug 2009 14:32:40 +0000 (16:32 +0200)]
omap3: move the other boards to board/

Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
15 years agoarm: move Logicpd's boards to board/logicpd/
Jean-Christophe PLAGNIOL-VILLARD [Sun, 23 Aug 2009 14:32:39 +0000 (16:32 +0200)]
arm: move Logicpd's boards to board/logicpd/

Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
15 years agoomap: move TI's boards to board/ti/
Jean-Christophe PLAGNIOL-VILLARD [Sun, 23 Aug 2009 14:32:38 +0000 (16:32 +0200)]
omap: move TI's boards to board/ti/

Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
15 years agoSupport for the Calao TNY-A9260/TNY-A9G20 boards
Albin Tonnerre [Thu, 20 Aug 2009 14:04:49 +0000 (16:04 +0200)]
Support for the Calao TNY-A9260/TNY-A9G20 boards

The Calao TNY-A9260 and TNY-9G20 are boards manufactured and sold by
Calao Systems <http://www.calao-systems.com>. Their components are very
similar to the AT91SAM9260EK board, so their configuration is based on
the configuration of this board. There are however some differences:
different clocks, no LCD, no ethernet. They also can use SPI EEPROM to
store the environment.

Signed-off-by: Albin Tonnerre <albin.tonnerre@free-electrons.com>
Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
15 years agoarm: Kirkwood: add SYSRSTn Duration Counter Support
Prafulla Wadaskar [Thu, 20 Aug 2009 15:29:28 +0000 (20:59 +0530)]
arm: Kirkwood: add SYSRSTn Duration Counter Support

This feature can be used to trigger special command "sysrstcmd" using
reset key long press event and environment variable "sysrstdelay" is set
(useful for reset to factory or manufacturing mode execution)

Kirkwood SoC implements a hardware-based SYSRSTn duration counter.
When SYSRSTn is asserted low, a SYSRSTn duration counter is running.
The counter value is stored in the SYSRSTn Length Counter Register
The counter is based on the 25-MHz reference clock (40ns)
It is a 29-bit counter, yielding a maximum counting duration of
2^29/25 MHz (21.4 seconds). When the counter reach its maximum value,
it remains at this value until counter reset is triggered by setting
bit 31 of KW_REG_SYSRST_CNT

Implementation:
Upon long reset assertion (> ${sysrstdelay} in secs) sysrstcmd will be
executed if pre-defined in environment variables.
This feature will be disabled if "sysrstdelay" variable is unset.

for-ex.
setenv sysrst_cmd "echo starting factory reset;
   nand erase 0xa0000 0x20000;
   echo finish ed sysrst command;"
will erase particular nand sector if triggered by this event

Signed-off-by: Prafulla Wadaskar <prafulla@marvell.com>
15 years agoAdd support for the Calao SBC35-A9G20 board
Albin Tonnerre [Mon, 24 Aug 2009 16:03:26 +0000 (18:03 +0200)]
Add support for the Calao SBC35-A9G20 board

The Calao SBC35-A9G20 board is manufactured and sold by Calao Systems
<http://www.calao-systems.com>. It is built around an AT91SAM9G20 ARM SoC
running at 400MHz. It features an Ethernet port, an SPI RTC backed by an onboard
battery , an SD/MMC slot, a CompactFlash slot, 64Mo of SDRAM, 256Mo of NAND
flash, two USB host ports, and an USB device port. More informations can be
found at <http://www.calao-systems.com/articles.php?lng=en&pg=5936>

Signed-off-by: Albin Tonnerre <albin.tonnerre@free-electrons.com>
15 years agoimx27lite: add support for imx27lite board from LogicPD
Ilya Yanok [Mon, 10 Aug 2009 22:32:09 +0000 (02:32 +0400)]
imx27lite: add support for imx27lite board from LogicPD

This patch adds support for i.MX27-LITEKIT development board from
LogicPD. This board uses i.MX27 SoC and has 2MB NOR flash, 64MB NAND
flash, FEC ethernet controller integrated into i.MX27.

Signed-off-by: Ilya Yanok <yanok@emcraft.com>
Acked-by: Wolfgang Denk <wd@denx.de>
15 years agoat91sam9260/afeb9260: Fix SPI initialization
Albin Tonnerre [Tue, 1 Sep 2009 09:26:20 +0000 (11:26 +0200)]
at91sam9260/afeb9260: Fix SPI initialization

Commit 7ebafb7ec1a0285af8380623c009576f92583b98 introduced a mistake in the spi
init function call for those boards. This patch fixes this.

Signed-off-by: Albin Tonnerre <albin.tonnerre@free-electrons.com>
15 years agoRemove duplicate set_cr
Simon Kagstrom [Tue, 18 Aug 2009 09:13:44 +0000 (11:13 +0200)]
Remove duplicate set_cr

Remove duplicate set_cr

set_cr is defined in both asm-arm/proc-armv/system.h and
include/asm-arm/system.h. This patch removes it (and some duplicate
defines) from the former.

Signed-off-by: Simon Kagstrom <simon.kagstrom@netinsight.net>
15 years agoMerge branch 'next' of ../next
Wolfgang Denk [Mon, 31 Aug 2009 20:21:47 +0000 (22:21 +0200)]
Merge branch 'next' of ../next

15 years agoPrepare 2009.08 v2009.08
Wolfgang Denk [Mon, 31 Aug 2009 17:57:42 +0000 (19:57 +0200)]
Prepare 2009.08

Update CHANGELOG

Signed-off-by: Wolfgang Denk <wd@denx.de>
15 years agoAdd common code dir for Matrix Vision boards.
Andre Schwarz [Mon, 31 Aug 2009 14:18:24 +0000 (16:18 +0200)]
Add common code dir for Matrix Vision boards.

This fixes current build failure.

Signed-off-by: Andre Schwarz <andre.schwarz@matrix-vision.de>
mvblm7.c: fix warning: implicit declaration of function
'mv_reset_environment'

Signed-off-by: Wolfgang Denk <wd@denx.de>
15 years agobugfix CONFIG_SYS_CONSOLE_INFO_QUIET
David Brownell [Sun, 30 Aug 2009 18:05:29 +0000 (11:05 -0700)]
bugfix CONFIG_SYS_CONSOLE_INFO_QUIET

The "console: unify printing current devices" patch goofed:
CONFIG_SYS_CONSOLE_INFO_QUIET is supposed to *REMOVE* boot
time noise, not add it.  Said patch changed the #ifndefs
to #ifdef; this one restores them to the proper sense.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
15 years agofsl: simplify the "mac id" command, improve boot-time informational message
Timur Tabi [Fri, 28 Aug 2009 21:56:45 +0000 (16:56 -0500)]
fsl: simplify the "mac id" command, improve boot-time informational message

The "mac id" command took a 4-character parameter as the identifier string.
However, for any given board, only one kind of identifier is acceptable, so it
makes no sense to ask the user to type it in.  Instead, if the user enters
"mac id", the identifier (and also the version, if it's NXID) will
automatically be set to the correct value.

Improve the message that is displayed when EEPROM is read during boot.  It now
displays "EEPROM:" and then either an error message or the EEPROM identifier
if successful.

If the identifier in EEPROM is valid, then always reject a bad CRC, even if the
CRC field has not been initialized.

Don't force the MAC address count to MAX_NUM_PORTS or less.  Forcing the value
to be changed resulting in an in-memory copy that does not match what's in
hardware, even though the user did not request that change.

Finally, always update the CRC value in the in-memory copy after any field
is changed, so that the CRC is always correct.

Signed-off-by: Timur Tabi <timur@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
15 years ago85xx: Added PCIe support for P1 P2 RDB
Poonam Aggrwal [Fri, 21 Aug 2009 01:59:58 +0000 (07:29 +0530)]
85xx: Added PCIe support for P1 P2 RDB

Call fsl_pci_init_port() to initialize all the PCIe ports on the board.

Signed-off-by: Poonam Aggrwal <poonam.aggrwal@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
15 years agodriver/fsl_pci: Add fsl_pci_init_port function to initialize a PCI controller
Poonam Aggrwal [Fri, 21 Aug 2009 01:59:42 +0000 (07:29 +0530)]
driver/fsl_pci: Add fsl_pci_init_port function to initialize a PCI controller

fsl_pci_init_port can be called from board specific PCI initialization
routines to setup the PCI (or PCIe) controller.  This will reduce code
redundancy in most of the 85xx/86xx FSL board ports that setup PCI.

Signed-off-by: Poonam Aggrwal <poonam.aggrwal@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
15 years ago85xx: Improve MPIC initialization
Timur Tabi [Thu, 20 Aug 2009 22:41:11 +0000 (17:41 -0500)]
85xx: Improve MPIC initialization

The MPIC initialization code for Freescale e500 CPUs was not using I/O
accessors, and it was not issuing a read-back to the MPIC after setting
mixed mode.  This may be the cause of a spurious interrupt on some systems.

Signed-off-by: Timur Tabi <timur@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
15 years ago85xx: Added support for P1011RDB and P2010RDB
Poonam Aggrwal [Thu, 20 Aug 2009 13:29:18 +0000 (18:59 +0530)]
85xx: Added support for P1011RDB and P2010RDB

P1011 and P2010 are single core variants of P1010 and P2020 respectively.
The board(RDB) will be same.

Signed-off-by: Poonam Aggrwal <poonam.aggrwal@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
15 years ago85xx: Added single core members of FSL P1xx/P2xx processors series
Poonam Aggrwal [Thu, 20 Aug 2009 13:27:45 +0000 (18:57 +0530)]
85xx: Added single core members of FSL P1xx/P2xx processors series

P1011 - Single core variant of P1020
P2010 - Single core variant of P2020

Signed-off-by: Poonam Aggrwal <poonam.aggrwal@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
15 years ago85xx: P1020RDB Support Added
Poonam Aggrwal [Thu, 20 Aug 2009 13:27:02 +0000 (18:57 +0530)]
85xx: P1020RDB Support Added

Signed-off-by: Poonam Aggrwal <poonam.aggrwal@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
15 years ago85xx: Added CONFIG_MAX_CPUS for P1020
Poonam Aggrwal [Thu, 20 Aug 2009 13:25:35 +0000 (18:55 +0530)]
85xx: Added CONFIG_MAX_CPUS for P1020

Signed-off-by: Poonam Aggrwal <poonam.aggrwal@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
15 years ago85xx: Add L2SRAM Register's macro definition
Mingkai Hu [Tue, 18 Aug 2009 07:37:15 +0000 (15:37 +0800)]
85xx: Add L2SRAM Register's macro definition

Signed-off-by: Mingkai Hu <Mingkai.hu@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
15 years ago85xx: Fix memory test range on MPC8536DS
Felix Radensky [Sat, 15 Aug 2009 12:08:37 +0000 (15:08 +0300)]
85xx: Fix memory test range on MPC8536DS

With current values of CONFIG_SYS_MEMTEST_START and CONFIG_SYS_MEMTEST_END
memory test hangs if run without arguments. Set them to sane values, so
that all available 512MB of RAM excluding exception vectors at the bottom
and u-boot code and stack at the top can be tested.

Signed-off-by: Felix Radensky <felix@embedded-sol.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
15 years ago85xx: Removed BEDBUG support on P1_P2_RDB
Kumar Gala [Wed, 12 Aug 2009 05:10:44 +0000 (00:10 -0500)]
85xx: Removed BEDBUG support on P1_P2_RDB

To match all other 85xx platforms we are removing BEDBUG support.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
15 years ago85xx: Init pci ethernet cards if we enable any on MPC8572DS
Kumar Gala [Sat, 8 Aug 2009 15:42:30 +0000 (10:42 -0500)]
85xx: Init pci ethernet cards if we enable any on MPC8572DS

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
15 years agoxes: Use proper IO access functions
Peter Tyser [Fri, 7 Aug 2009 18:16:34 +0000 (13:16 -0500)]
xes: Use proper IO access functions

Also fix some minor whitespace oddities while we're cleaning up

Signed-off-by: Peter Tyser <ptyser@xes-inc.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
15 years ago85xx: Move to a common linker script
Kumar Gala [Fri, 7 Aug 2009 18:00:55 +0000 (13:00 -0500)]
85xx: Move to a common linker script

There are really no differences between all the 85xx linker scripts so
we can just move to a single common one.  Board code is still able to
override the common one if need be.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
15 years ago85xx: Added P1020 Processor Support.
Poonam Aggrwal [Fri, 31 Jul 2009 06:38:27 +0000 (12:08 +0530)]
85xx: Added P1020 Processor Support.

P1020 is another member of QorIQ series of processors which falls in ULE
category. It is an e500 based dual core SOC.

Being a scaled down version of P2020 it has following differences:
- 533MHz - 800MHz core frequency.
- 256Kbyte L2 cache
- Ethernet controllers with classification capabilities.
Also the SOC is pin compatible with P2020

Signed-off-by: Poonam Aggrwal <poonam.aggrwal@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
15 years ago85xx: Add support for P2020RDB board
Poonam Aggrwal [Wed, 5 Aug 2009 07:59:24 +0000 (13:29 +0530)]
85xx: Add support for P2020RDB board

The code base adds P1 & P2 RDB platforms support.
The folder and file names can cater to future SOCs of P1/P2 family.
P1 & P2 processors are 85xx platforms, part of Freescale QorIQ series.

Tested following on P2020RDB:
1. eTSECs
2. DDR, NAND, NOR, I2C.

Signed-off-by: Poonam Aggrwal <poonam.aggrwal@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
15 years ago8xxx: Removed CONFIG_NUM_CPUS from 85xx/86xx
Poonam Aggrwal [Fri, 31 Jul 2009 06:38:14 +0000 (12:08 +0530)]
8xxx: Removed CONFIG_NUM_CPUS from 85xx/86xx

The number of CPUs are getting detected dynamically by checking the
processor SVR value.  Also removed CONFIG_NUM_CPUS references from all
the platforms with 85xx/86xx processors.

This can help to use the same u-boot image across the platforms.

Also revamped and corrected few Freescale Copyright messages.

Signed-off-by: Poonam Aggrwal <poonam.aggrwal@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
15 years ago8xxx: Refactored common cpu specific code for 85xx/86xx into one file.
Poonam Aggrwal [Fri, 31 Jul 2009 06:37:45 +0000 (12:07 +0530)]
8xxx: Refactored common cpu specific code for 85xx/86xx into one file.

Removed same code pieces from cpu/mpc85xx/cpu.c and cpu/mpc86xx/cpu.c
and moved to cpu/mpc8xxx/cpu.c(new file)

Signed-off-by: Poonam Aggrwal <poonam.aggrwal@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
15 years agostx: create common vendor/board hierarchy for STx boards
Alex Dubov [Fri, 7 Aug 2009 05:28:32 +0000 (15:28 +1000)]
stx: create common vendor/board hierarchy for STx boards

Move files belonging to the STx boards into common vendor directory and
update the Makefile to reflect this.

Signed-off-by: Alex Dubov <oakad@yahoo.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
15 years ago85xx: Remove unused CONFIG_CLEAR_LAW0 defines
Peter Tyser [Tue, 4 Aug 2009 22:38:00 +0000 (17:38 -0500)]
85xx: Remove unused CONFIG_CLEAR_LAW0 defines

Signed-off-by: Peter Tyser <ptyser@xes-inc.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
15 years ago86xx: Remove redudant PLATFORM_CPPFLAGS
Kumar Gala [Thu, 6 Aug 2009 23:38:43 +0000 (18:38 -0500)]
86xx: Remove redudant PLATFORM_CPPFLAGS

For historic reasons we had defined some additional PLATFORM_CPPFLAGS like:

PLATFORM_CPPFLAGS += -DCONFIG_MPC86xx=1
PLATFORM_CPPFLAGS += -DCONFIG_MPC8641=1

However these are all captured in the config.h and thus redudant.  Also
moved common 86xx flags into cpu/mpc86xx/config.mk.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
15 years ago85xx: Remove redudant PLATFORM_CPPFLAGS
Kumar Gala [Thu, 6 Aug 2009 23:28:34 +0000 (18:28 -0500)]
85xx: Remove redudant PLATFORM_CPPFLAGS

For historic reasons we had defined some additional PLATFORM_CPPFLAGS
like:

PLATFORM_CPPFLAGS += -DCONFIG_E500=1
PLATFORM_CPPFLAGS += -DCONFIG_MPC85xx=1
PLATFORM_CPPFLAGS += -DCONFIG_MPC8548=1

However these are all captured in the config.h and thus redudant.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
15 years ago85xx: Add a 36-bit physical configuration for MPC8536DS
Kumar Gala [Thu, 30 Jul 2009 20:54:07 +0000 (15:54 -0500)]
85xx: Add a 36-bit physical configuration for MPC8536DS

We move all IO addressed (CCSR, localbus, PCI) above the 4G boundary
to allow for larger memory sizes.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Acked-by: Wolfgang Denk <wd@denx.de>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
15 years ago85xx: Cleanup whitespace in mpc8536ds.c
Kumar Gala [Tue, 4 Aug 2009 14:10:03 +0000 (09:10 -0500)]
85xx: Cleanup whitespace in mpc8536ds.c

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
15 years agopci/fsl_pci_init: Rework PCI ATMU setup to handle >4G of memory
Kumar Gala [Wed, 5 Aug 2009 12:59:35 +0000 (07:59 -0500)]
pci/fsl_pci_init: Rework PCI ATMU setup to handle >4G of memory

The old PCI ATMU setup code would just mimic the PCI regions into the
ATMU registers.  For simple memory maps in which all memory, MMIO, etc
space fit into 4G this works ok.  However there are issues with we have
>4G of memory as we know can't access all of memory and we need to
ensure that PCICSRBAR (PEXCSRBAR on PCIe) isn't overlapping with
anything since we can't turn it off.

We first setup outbound windows based on what the board code setup
in the pci regions for MMIO and IO access.  Next we place PCICSRBAR
below the MMIO window.  After which we try to setup the inbound windows
to map as much of memory as possible.

On PCIe based controllers we are able to overmap the ATMU setup since
RX & TX links are separate but report the proper amount of inbound
address space to the region tracking to ensure there is no overlap.

On PCI based controllers we use as many inbound windows as available to
map as much of the memory as possible.

Additionally we changed all the CCSR register access to use proper IO
accessor functions.  Also had to add CONFIG_SYS_CCSRBAR_PHYS to some
86xx platforms that didn't have it defined.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
15 years agopci/fsl_pci_init: Use PCIe capability to determine if controller is PCIe
Kumar Gala [Wed, 5 Aug 2009 12:49:27 +0000 (07:49 -0500)]
pci/fsl_pci_init: Use PCIe capability to determine if controller is PCIe

Change the code to use the PCIe capabilities register to determine if we
are a PCIe controller or not.  Additionally cleaned up some white space
and formatting in the file.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
15 years agopci/fsl_pci_init: Fold fsl_pci_setup_inbound_windows into fsl_pci_init
Kumar Gala [Tue, 4 Aug 2009 02:02:02 +0000 (21:02 -0500)]
pci/fsl_pci_init: Fold fsl_pci_setup_inbound_windows into fsl_pci_init

Every platform that calls fsl_pci_init calls fsl_pci_setup_inbound_windows
before it calls fsl_pci_init.  There isn't any reason to just call it
from fsl_pci_init and simplify things a bit.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
15 years agopci/fsl_pci_init: Fold pci_setup_indirect into fsl_pci_init
Kumar Gala [Tue, 4 Aug 2009 01:44:55 +0000 (20:44 -0500)]
pci/fsl_pci_init: Fold pci_setup_indirect into fsl_pci_init

Every platform that calls fsl_pci_init calls pci_setup_indirect before
it calls fsl_pci_init.  There isn't any reason to just call it from
fsl_pci_init and simplify things a bit.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
15 years agoUse common code for Matrix Vision boards
André Schwarz [Thu, 27 Aug 2009 12:48:35 +0000 (14:48 +0200)]
Use common code for Matrix Vision boards

Clean up existing boards (mvBC-P/MPC5200 and mvBL-M7/MPC8343) by
using common code.

Signed-off-by: André Schwarz <andre.schwarz@matrix-vision.de>
15 years agoMerge branch 'next' of git://git.denx.de/u-boot-nand-flash into next
Wolfgang Denk [Thu, 27 Aug 2009 22:17:41 +0000 (00:17 +0200)]
Merge branch 'next' of git://git.denx.de/u-boot-nand-flash into next

15 years agompc83xx: mpc8349itx - accommodate larger kernel sizes & unzero the bootdelay
Kim Phillips [Thu, 27 Aug 2009 02:27:37 +0000 (21:27 -0500)]
mpc83xx: mpc8349itx - accommodate larger kernel sizes & unzero the bootdelay

apparently the ITX was missed last round.

Also make bootdelay consistent with other boards, so as to give on the
opportunity to fix mistakenly set bootcmd without having checked for an
bootdelay zero setting first.

Signed-off-by: Kim Phillips <kim.phillips@freescale.com>
15 years agompc83xx: match dtb filename references to their dts equivalents in the linux kernel
Kim Phillips [Thu, 27 Aug 2009 02:25:46 +0000 (21:25 -0500)]
mpc83xx: match dtb filename references to their dts equivalents in the linux kernel

Signed-off-by: Kim Phillips <kim.phillips@freescale.com>
15 years agoNAND: DaVinci: V2 Adding 4 BIT ECC support
Sandeep Paulraj [Tue, 18 Aug 2009 14:10:42 +0000 (10:10 -0400)]
NAND: DaVinci: V2 Adding 4 BIT ECC support

This patch adds 4 BIT ECC support in the DaVinci NAND
driver. Tested on both the DM355 and DM365.

Signed-off-by: Sandeep Paulraj <s-paulraj@ti.com>
Signed-off-by: Scott Wood <scottwood@freescale.com>
15 years agoMTD:NAND: ADD new ECC mode NAND_ECC_HW_OOB_FIRST
Sandeep Paulraj [Mon, 10 Aug 2009 17:27:56 +0000 (13:27 -0400)]
MTD:NAND: ADD new ECC mode NAND_ECC_HW_OOB_FIRST

This patch adds the new mode NAND_ECC_HW_OOB_FIRST in the nand code to
support 4-bit ECC on TI DaVinci devices with large page (up to 2K) NAND
chips.  This ECC mode is similar to NAND_ECC_HW, with the exception of
read_page API that first reads the OOB area, reads the data in chunks,
feeds the ECC from OOB area to the ECC hw engine and perform any
correction on the data as per the ECC status reported by the engine.

This patch has been accepted by Andrew Morton and can be found at

http://userweb.kernel.org/~akpm/mmotm/broken-out/mtd-nand-add-new-ecc-mode-ecc_hw_oob_first.patch

Signed-off-by: Sandeep Paulraj <s-paulraj@ti.com>
Signed-off-by: Sneha Narnakaje <nsnehaprabha@ti.com>
Signed-off-by: Scott Wood <scottwood@freescale.com>
15 years agomxc_nand: add nand driver for MX2/MX3
Ilya Yanok [Mon, 10 Aug 2009 22:32:54 +0000 (02:32 +0400)]
mxc_nand: add nand driver for MX2/MX3

Driver for NFC NAND controller found on Freescale's MX2 and MX3
processors. Ported from Linux. Tested only with i.MX27 but should
works with other MX2 and MX3 processors too.

Signed-off-by: Ilya Yanok <yanok@emcraft.com>
Signed-off-by: Scott Wood <scottwood@freescale.com>
15 years agoNAND: ADD page Parameter to all read_page/read_page_raw API's
Sandeep Paulraj [Mon, 10 Aug 2009 17:27:46 +0000 (13:27 -0400)]
NAND: ADD page Parameter to all read_page/read_page_raw API's

This patch adds a new "page" parameter to all NAND read_page/read_page_raw
APIs.  The read_page API for the new mode ECC_HW_OOB_FIRST requires the
page information to send the READOOB command and read the OOB area before
the data area.

This patch has been accepted by Andrew Morton and can be found at
http://userweb.kernel.org/~akpm/mmotm/broken-out/mtd-nand-add-page-parameter-to-all-read_page-read_page_raw-apis.patch

WE would like this to become part of the u-boot GIT as well

Signed-off-by: Sandeep Paulraj <s-paulraj@ti.com>
Signed-off-by: Sneha Narnakaje <nsnehaprabha@ti.com>
Signed-off-by: Scott Wood <scottwood@freescale.com>
15 years ago83xx, kmeter1: added NAND support
Heiko Schocher [Tue, 21 Jul 2009 15:13:40 +0000 (17:13 +0200)]
83xx, kmeter1: added NAND support

Signed-off-by: Heiko Schocher <hs@denx.de>
Signed-off-by: Scott Wood <scottwood@freescale.com>
15 years agoOneNAND: Remove unused read_spareram
Kyungmin Park [Tue, 21 Jul 2009 02:58:04 +0000 (11:58 +0900)]
OneNAND: Remove unused read_spareram

Remove unused read_spareram and add unlock_all as kernel does

Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Scott Wood <scottwood@freescale.com>
15 years agoKB9202: Add NAND support
Matthias Kaehlcke [Thu, 16 Jul 2009 19:19:29 +0000 (21:19 +0200)]
KB9202: Add NAND support

Add KB9202 NAND driver

Signed-off-by: Matthias Kaehlcke <matthias@kaehlcke.net>
Signed-off-by: Scott Wood <scottwood@freescale.com>
15 years agoOneNAND: Remove unused read_spareram
Kyungmin Park [Tue, 21 Jul 2009 02:58:04 +0000 (11:58 +0900)]
OneNAND: Remove unused read_spareram

Remove unused read_spareram and add unlock_all as kernel does

Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
15 years agogalaxy5200: Cleanup typo and trailing whitespace
Eric Millbrandt [Tue, 25 Aug 2009 15:30:26 +0000 (10:30 -0500)]
galaxy5200: Cleanup typo and trailing whitespace

Signed-off-by: Eric Millbrandt <emillbrandt@dekaresearch.com>
15 years agoMerge branch 'next' of git://www.denx.de/git/u-boot-coldfire into next
Wolfgang Denk [Wed, 26 Aug 2009 19:29:32 +0000 (21:29 +0200)]
Merge branch 'next' of git://www.denx.de/git/u-boot-coldfire into next

15 years agoColdFire: Fix compile warning messages
TsiChung Liew [Wed, 22 Jul 2009 18:42:45 +0000 (18:42 +0000)]
ColdFire: Fix compile warning messages

Change %08lX to %08X in board.c. Remove unused variable
'oscillator' in mcf5227x/cpu_init.c and 'scm2' in
mcf532x/cpu_init.c. Provide argument type cast in
drivers/dma/MCD_dmaApi.c.

Signed-off-by: TsiChung Liew <tsicliew@gmail.com>
15 years agoColdFire: Fix missing _IO_BASE which caused compile error
TsiChung Liew [Wed, 22 Jul 2009 16:32:39 +0000 (16:32 +0000)]
ColdFire: Fix missing _IO_BASE which caused compile error

The compile error was caused by a recent patch. Affected platforms -
M5253DEMO.h, M5253EVBE.h, and M54455EVB.h. Adding the _IO_BASE
automatically defined to 0 in asm-m68k/io.h if it isn't set in
platform configuration file.

Signed-off-by: TsiChung Liew <tsicliew@gmail.com>
15 years agoSupport up to 7 banks for ids as specified in JEDEC JEP106Z
Niklaus Giger [Wed, 22 Jul 2009 15:13:24 +0000 (17:13 +0200)]
Support up to 7 banks for ids as specified in JEDEC JEP106Z

see http://www.jedec.org/download/search/jep106Z.pdf
Add some second source legacy flash chips 256x8.

Signed-off-by: Niklaus Giger <niklaus.giger@member.fsf.org>
Signed-off-by: Stefan Roese <sr@denx.de>
15 years agoMerge branch 'next' of git://git.denx.de/u-boot-net into next
Wolfgang Denk [Tue, 25 Aug 2009 21:03:22 +0000 (23:03 +0200)]
Merge branch 'next' of git://git.denx.de/u-boot-net into next

15 years agoMerge branch 'master' into next
Wolfgang Denk [Tue, 25 Aug 2009 20:57:10 +0000 (22:57 +0200)]
Merge branch 'master' into next

15 years agoMerge branch 'master' of git://git.denx.de/u-boot-ppc4xx
Wolfgang Denk [Tue, 25 Aug 2009 20:54:01 +0000 (22:54 +0200)]
Merge branch 'master' of git://git.denx.de/u-boot-ppc4xx

15 years agoMerge branch 'master' of git://git.denx.de/u-boot-blackfin
Wolfgang Denk [Tue, 25 Aug 2009 20:51:40 +0000 (22:51 +0200)]
Merge branch 'master' of git://git.denx.de/u-boot-blackfin

15 years agonet: tsec - handle user interrupt while waiting for PHY auto negotiation to complete
Kim Phillips [Mon, 24 Aug 2009 19:32:26 +0000 (14:32 -0500)]
net: tsec - handle user interrupt while waiting for PHY auto negotiation to complete

if you don't have firmware installed for the PHY to come to life, this
wait can be painful - let's give the option to avoid it if we want.

Signed-off-by: Kim Phillips <kim.phillips@freescale.com>
Acked-by: Andy Fleming <afleming@freescale.com>
Signed-off-by: Ben Warren <biggerbadderben@gmail.com>
15 years agoAdd Transfer Size Option to tftp
Robin Getz [Thu, 20 Aug 2009 14:50:20 +0000 (10:50 -0400)]
Add Transfer Size Option to tftp

Optionally add RFC 2349 "Transfer Size Option", so we can minimize the
time spent sending data over the UART (now print a single line during a
tftp transfer).

 - If turned on (CONFIG_TFTP_TSIZE), U-Boot asks for the size of the file.
 - if receives the file size, a single line (50 chars) are printed.
     one hash mark == 2% of the file downloaded.
 - if it doesn't receive the file size (the server doesn't support RFC
     2349, prints standard hash marks (one mark for each UDP frame).

Signed-off-by: Robin Getz <rgetz@blackfin.uclinux.org>
Signed-off-by: Ben Warren <biggerbadderben@gmail.com>
15 years agoAdd debug message for Blackfin Ethernet Rx function.
Robin Getz [Mon, 24 Aug 2009 14:33:39 +0000 (10:33 -0400)]
Add debug message for Blackfin Ethernet Rx function.

Add a simple print for the Blackfin's Ethernet Rx function,
so we can debug incomming Ethernet functions easier.

Signed-off-by: Robin Getz <rgetz@blackfin.uclinux.org>
Signed-off-by: Ben Warren <biggerbadderben@gmail.com>
15 years agoConvert CS8900 Ethernet driver to CONFIG_NET_MULTI API
Ben Warren [Tue, 25 Aug 2009 20:09:37 +0000 (13:09 -0700)]
Convert CS8900 Ethernet driver to CONFIG_NET_MULTI API

All in-tree boards that use this controller have CONFIG_NET_MULTI added
Also:
  - changed CONFIG_DRIVER_CS8900 to CONFIG_CS8900
  - changed CS8900_BASE to CONFIG_CS8900_BASE
  - changed CS8900_BUS?? to CONFIG_CS8900_BUS??
  - cleaned up line lengths
  - modified VCMA9 command function that accesses the device
  - removed MAC address initialization from lib_arm/board.c

Signed-off-by: Ben Warren <biggerbadderben@gmail.com>
Tested-by: Wolfgang Denk <wd@denx.de>
Acked-by: Wolfgang Denk <wd@denx.de>
15 years agoarm nomadik: activate defrag choose 4k transfer block size
Alessandro Rubini [Fri, 7 Aug 2009 11:59:26 +0000 (13:59 +0200)]
arm nomadik: activate defrag choose 4k transfer block size

This chooses 4kB data size for both TFTP and NFS, as an example
about how to use support for IP fragments.

Signed-off-by: Alessandro Rubini <rubini@gnudd.com>
Signed-off-by: Ben Warren <biggerbadderben@gmail.com>
15 years agonfs: accept CONFIG_NFS_READ_SIZE from config file
Alessandro Rubini [Fri, 7 Aug 2009 11:59:16 +0000 (13:59 +0200)]
nfs: accept CONFIG_NFS_READ_SIZE from config file

To take advantage of defragmented packets, the config file
can define CONFIG_NFS_READ_SIZE to override the 1kB default.
No support is there for an environment variable by now.

Signed-off-by: Alessandro Rubini <rubini@gnudd.com>
Signed-off-by: Ben Warren <biggerbadderben@gmail.com>
15 years agotftp: get the tftp block size from config file and from the environment
Alessandro Rubini [Fri, 7 Aug 2009 11:59:06 +0000 (13:59 +0200)]
tftp: get the tftp block size from config file and from the environment

Increasing the block size is useful if CONFIG_IP_DEFRAG is
used. Howerver, the last fragments in a burst may overflow the
receiving ethernet, so the default is left at 1468, with thre new
CONFIG_TFTP_BLOCKSIZE for config files. Further, "tftpblocksize"
can be set in the environment.

Signed-off-by: Alessandro Rubini <rubini@gnudd.com>
Signed-off-by: Ben Warren <biggerbadderben@gmail.com>
15 years agonet: defragment IP packets
Alessandro Rubini [Fri, 7 Aug 2009 11:58:56 +0000 (13:58 +0200)]
net: defragment IP packets

The defragmenting code is enabled by CONFIG_IP_DEFRAG; the code is
useful for TFTP and NFS transfers.  The user can specify the maximum
defragmented payload as CONFIG_NET_MAXDEFRAG (default 16k).
Since NFS has a bigger per-packet overhead than TFTP, the static
reassembly buffer can hold CONFIG_NET_MAXDEFRAG + the NFS overhead.

The packet buffer is used as an array of "hole" structures, acting as
a double-linked list. Each new fragment can split a hole in two,
reduce a hole or fill a hole. No support is there for a fragment
overlapping two diffrent holes (i.e., thre new fragment is across an
already-received fragment).

Signed-off-by: Alessandro Rubini <rubini@gnudd.com>
Signed-off-by: Ben Warren <biggerbadderben@gmail.com>