The new DDR2 SPD spec is backwards-compatible with the old one, but
there are DIMMs which are being produced which have new SPD version
fields that fail to work with U-boot.
To make the code a bit more readable, we add some human-readable
SPD_DIMM_TYPE_* constants.
Signed-off-by: Kyle Moffett
There are a couple preliminary patches to common code, but basically
everything else is limited to our own board-support code.
I look forward to your comments and criticisms!
Cheers,
Kyle Moffett
--
Overall diffstat below:
MAKEALL |2 +
Mak
, we disable the poll and error return on chips
whose EEPROMS indicate no management. As a protection against
misconfigured chipsets, we still delay for the entire management poll
timeout.
Signed-off-by: Kyle Moffett
---
drivers/net/e1000.c |7 +++
1 files changed, 7 insertions(+), 0
t code.
A GIT tree containing these patches may be found here:
git://opensource.exmeritus.com/hww-1u-1a/u-boot.git master
http://opensource.exmeritus.com/git/hww-1u-1a/u-boot.git/master
I look forward to your comments and criticisms.
Cheers,
Kyle Moffett
--
Overall diffstat below
, we disable the poll and error return on chips
whose EEPROMS indicate no management. As a protection against
misconfigured chipsets, we still delay for the entire management poll
timeout.
Signed-off-by: Kyle Moffett
---
Jeff Kirsher has already accepted a similar
patch to the Linux kernel into his
The new DDR2 SPD spec is backwards-compatible with the old one, but
there are DIMMs which are being produced which have new SPD version
fields that fail to work with U-boot.
To make the code a bit more readable, we add some human-readable
SPD_DIMM_TYPE_* constants.
Signed-off-by: Kyle Moffett
_MINI_UDIMM 0x20
>
> These aren't 8xxx-specific, so they should go somewhere common. It
> looks like many of them are already defined in include/ddr_spd.h; I'd
> recommend adding new types there.
Ok, I'll respin this using ddr_spd.h! Thanks for the comments!
Cheers,
bit.
>> +/* DDR2 SO-RDIMM SPD EEPROM is at I2C0-0x51 */
>> +#define CONFIG_SYS_SPD_BUS_NUM 0
>> +#define SPD_EEPROM_ADDRESS1 0x51
>> +#if 0
>> +#define CONFIG_CMD_EEPROM
>> +#endif
>
> Remove dead code. Same comment applies for future if 0s.
Same as ab
heers,
Kyle Moffett
___
U-Boot mailing list
U-Boot@lists.denx.de
http://lists.denx.de/mailman/listinfo/u-boot
To make it possible to use the "sspi" command with the e1000 firmware
EEPROM we add a small "generic SPI" driver wrapper around the existing
e1000 SPI backend.
Signed-off-by: Kyle Moffett
---
drivers/net/e1000.c | 92 ++-
To ease the implementation of other MPC85xx board ports, several common
GPIO helpers are added to .
Since each of these compiles to no more than 4-5 instructions it would
be very inefficient to call them out of line, therefore we put them
entirely in the header file.
Signed-off-by: Kyle Moffett
looks like this on the wire:
Time >>>
--
... [7, 6, 5, 4, 3, 2, 1, 0, 15, 14, 13, 12, 11, 10, 9, 8], ...
--
(MSB is 15, LSB is 0).
Signed-off-by: Kyle Moffet
Consolidate the test for a dual-port NIC to one location for easy
modification, then fix support for the dual-port 82571.
Signed-off-by: Kyle Moffett
---
drivers/net/e1000.c | 66 +-
drivers/net/e1000.h |6
2 files changed, 33
hopefully make such EEPROM errors much more
straightforward to handle correctly in boot scripts and the like.
It is also necessary for a followup patch which allows SPI programming
of an e1000 controller's EEPROM even if the checksum is invalid.
Signed-off-by: Kyle Moffett
---
d
looks like this on the wire:
Time >>>
--
... [7, 6, 5, 4, 3, 2, 1, 0, 15, 14, 13, 12, 11, 10, 9, 8], ...
--
(MSB is 15, LSB is 0).
Signed-off-by: Kyle Moffet
To make it possible to use the "sspi" command with the e1000 firmware
EEPROM we add a small "generic SPI" driver wrapper around the existing
e1000 SPI backend.
Signed-off-by: Kyle Moffett
---
drivers/net/e1000.c | 92 ++-
As an aide to debugging, we should print out the expected value of the
EEPROM checksum in addition to just saying that it is wrong.
Signed-off-by: Kyle Moffett
---
drivers/net/e1000.c | 48 ++--
1 files changed, 30 insertions(+), 18 deletions
hardware misfeature. Proper
support for the hardware reset mechanism has been left for a later
patch series to address.
Signed-off-by: Kyle Moffett
Cc: Andy Fleming
Cc: Kumar Gala
--
I believe all review comments have been addressed, please let me know if
there is anything else I need to
xactly
identical, but it is very close to the version in Linux v3.2-rc4.
Signed-off-by: Kyle Moffett
---
tools/setlocalversion | 138 -
1 files changed, 113 insertions(+), 25 deletions(-)
diff --git a/tools/setlocalversion b/tools/setlocalvers
ion of tools/setlocalversion in the kernel correctly handles
those situations by using different plumbing commands.
The version from the kernel is not directly usable as it has code for
supporting CONFIG_LOCALVERSION from Kconfig, but the version that was
imported is very similar to the one in Lin
cted RDIMM VL493T5663D-E6M
2 GiB (DDR2, 64-bit, CL=4, ECC on)
DDR Chip-Select Interleaving Mode: CS0+CS1
After this patch, it is:
DRAM: Detected RDIMM VL493T5663D-E6M
2 GiB (DDR2, 64-bit, CL=4, ECC on)
DDR Chip-Select Interleaving Mode: CS0+CS1
Signed-off-by: Kyle M
but always uses an address of 0x00.
This fixes the logic for that particular case, hopefully without
breaking any other board configurations.
Signed-off-by: Kyle Moffett
Cc: Heiko Schocher
---
include/common.h |6 +++---
1 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/include/
code can be removed.
Signed-off-by: Kyle Moffett
Cc: Ben Warren
---
drivers/net/ne2000_base.c | 17 +---
drivers/usb/eth/usb_ether.c |4 +--
net/eth.c | 56 --
3 files changed, 35 insertions(+), 42 deletions(-)
diff
, fix an incorrect address for the kernel in FLASH memory.
Signed-off-by: Kyle Moffett
Cc: Andy Fleming
Cc: Kumar Gala
---
include/configs/HWW1U1A.h |7 +--
1 files changed, 5 insertions(+), 2 deletions(-)
diff --git a/include/configs/HWW1U1A.h b/include/configs/HWW1U1A.h
index 123fc
"run restore_eeprom && saveenv")
Signed-off-by: Kyle Moffett
Cc: Andy Fleming
Cc: Kumar Gala
---
include/configs/HWW1U1A.h | 29 +++--
1 files changed, 27 insertions(+), 2 deletions(-)
diff --git a/include/configs/HWW1U1A.h b/include/configs/HWW1U1A.h
index
This new #define is set in config_cmd_defaults.h (which is automatically
included on every board by "mkconfig"), but this allows boards to elect
to omit the "reset" command if necessary with "#undef CONFIG_CMD_RESET".
Signed-off-by: Kyle Moffett
Cc: Wolfgang Denk
C
). Other than a few minor changes
for Kconfig, the imported version is very similar to Linux v3.2-rc4.
Signed-off-by: Kyle Moffett
Cc: Wolfgang Denk
Cc: Mike Frysinger
--
Changelog:
v3: Fix "localversion-*" file support and minimize kernel delta.
---
tools/setlocalversion | 185 +
The included series of 3 patches fixes several bugs in the FAT code
and the USB storage stack to make it possible to read from a FAT
filesystem image embedded as an "El-Torito boot image" in an ISO9660
volume (IE: a CD or DVD).
This has been tested on the eXMeritus HWW-1U-1A.
an ISO9660 boot
volume (El-Torito format) by reordering the filesystem checks and
reading the 0x55 0xAA "DOS boot signature" and FAT/FAT32 magic number
from the first sector of the partition instead of from sector 0.
Signed-off-by: Kyle Moffett
---
fs/fat/fa
ate an error but instead return 0 to
indicate the number of blocks successfully read (IE: None).
The FAT filesystem should defensively check to ensure that it got all of
the sectors that it asked for when reading.
Signed-off-by: Kyle Moffett
---
fs/fat/fat.c | 18 ++
1 files chan
D-ROM media has a 2048-byte blocksize, resulting in a
maximum transfer size of 40960 bytes, which does not fit.
Since the EHCI specification is impossibly obtuse and far beyond my
comprehension, I chose to dynamically compute the limit based on the
blocksize.
Signed-off-by: Kyle Moffett
---
c
w commentary on these patches and I believe
that they should be acceptable for merging into "next" now.
Cheers,
Kyle Moffett
--
Interested in my work on the Debian powerpcspe port?
I'm keeping a blog here: http://pureperl.blogspot.com/
an ISO9660 boot
volume (El-Torito format) by reordering the filesystem checks and
reading the 0x55 0xAA "DOS boot signature" and FAT/FAT32 magic number
from the first sector of the partition instead of from sector 0.
Signed-off-by: Kyle Moffett
--
v2:
I have removed the unnecess
ate an error but instead return 0 to
indicate the number of blocks successfully read (IE: None).
The FAT filesystem should defensively check to ensure that it got all of
the sectors that it asked for when reading.
Signed-off-by: Kyle Moffett
--
v2: No change
---
fs/fat/fat.c |
D-ROM media has a 2048-byte blocksize, resulting in a
maximum transfer size of 40960 bytes, which does not fit.
Since the EHCI specification is impossibly obtuse and far beyond my
comprehension, I chose to dynamically compute the limit based on the
blocksize.
Signed-off-by: Kyle Moffett
--
v2:
Signed-off-by: Kyle Moffett
---
arch/powerpc/cpu/mpc8xxx/ddr/util.c | 18 +-
1 files changed, 9 insertions(+), 9 deletions(-)
diff --git a/arch/powerpc/cpu/mpc8xxx/ddr/util.c
b/arch/powerpc/cpu/mpc8xxx/ddr/util.c
index 02908b4..104d360 100644
--- a/arch/powerpc/cpu/mpc8xxx
hopefully make such EEPROM errors much more
straightforward to handle correctly in boot scripts and the like.
It is also necessary for a followup patch which allows SPI programming
of an e1000 controller's EEPROM even if the checksum is invalid.
Signed-off-by: Kyle Moffett
---
d
looks like this on the wire:
Time >>>
-
... [7, 6, 5, 4, 3, 2, 1, 0, 15, 14, 13, 12, 11, 10, 9, 8], ...
-
(MSB is 15, LSB is 0).
Signed-off-by: Kyle Moffet
1 insertions(+), 115 deletions(-)
Most of the new code is hidden behind a couple of config options, so it won't
affect other users of e1000 chips unless they explicitly request it.
Cheers,
Kyle Moffett
___
U-Boot mailing list
U-Boot@lists.denx.de
htt
Consolidate the test for a dual-port NIC to one location for easy
modification, then fix support for the dual-port 82571.
Signed-off-by: Kyle Moffett
---
drivers/net/e1000.c | 66 +-
drivers/net/e1000.h |6
2 files changed, 33
To make it possible to use the "sspi" command with the e1000 firmware
EEPROM we add a small "generic SPI" driver wrapper around the existing
e1000 SPI backend.
Signed-off-by: Kyle Moffett
---
drivers/net/e1000.c | 92 ++-
As an aide to debugging, we should print out the expected value of the
EEPROM checksum in addition to just saying that it is wrong.
Signed-off-by: Kyle Moffett
---
drivers/net/e1000.c | 48 ++--
1 files changed, 30 insertions(+), 18 deletions
() command function instead of
directly poking at the MPC85xx "RSTCR" (reset control) register.
Signed-off-by: Kyle Moffett
---
arch/powerpc/cpu/mpc85xx/cpu.c |7 +--
1 files changed, 5 insertions(+), 2 deletions(-)
diff --git a/arch/powerpc/cpu/mpc85xx/cpu.c b/arch/powerpc/cpu/mpc
Hello again everyone!
After a relatively long hiatus working on other projects, I'm resubmitting
updated versions of our board-support patches for review and inclusion.
This patch series is based off the latest master branch as of Feb 11th,
specifically, commit d1a79b71f7c5fd9e277e0feb35f049289df
Use #define constants to enhance readability of DDR2/3 SPD parsing code.
Also add the DDR2 type for an SO-RDIMM module to the switch statement.
Signed-off-by: Kyle Moffett
---
arch/powerpc/cpu/mpc8xxx/ddr/ddr2_dimm_params.c | 34 +++--
arch/powerpc/cpu/mpc8xxx/ddr
the Linux kernel seems
to require a full 64-bit divide on any 32-bit PowerPC.
Build-and-boot-tested on the HWW-1U-1A board using DDR2 SPD detection.
Signed-off-by: Kyle Moffett
---
Author's note: This patch really needs a bunch more review and testing, but
I only have access to a very limit
issue.
Specifically, the routines are: _ashldi3(), _ashrdi3(), and _lshrdi3().
The Makefile framework was copied from the U-Boot ARM port.
Signed-off-by: Kyle Moffett
---
arch/powerpc/lib/Makefile | 28 -
arch/powerpc/lib/_ashldi3.S | 45
ned-off-by: Kyle Moffett
---
common/fdt_support.c |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/common/fdt_support.c b/common/fdt_support.c
index 6c98e5b..edcf04a 100644
--- a/common/fdt_support.c
+++ b/common/fdt_support.c
@@ -394,7 +394,7 @@ int fdt_fixup_memory_banks(
be enabled with CONFIG_MPC85XX_GENERIC_GPIO
Signed-off-by: Kyle Moffett
---
arch/powerpc/include/asm/mpc85xx_gpio.h | 124 +++
1 files changed, 124 insertions(+), 0 deletions(-)
create mode 100644 arch/powerpc/include/asm/mpc85xx_gpio.h
diff --git a/arch/powerpc
seems to work, but I don't
yet have a usable 36-bit kernel or DTB, so it's mostly untested.
* CPU reset is a little quirky due to hardware misfeature, see the
extensive comments in the board_reset_r() function in hww-1u-1a.c
Signed-off-by: Kyle Moffett
---
M
Consolidate the test for a dual-port NIC to one location for easy
modification, then fix support for the dual-port 82571.
Signed-off-by: Kyle Moffett
Cc: Ben Warren
---
No changes since v1
drivers/net/e1000.c | 66 +-
drivers/net/e1000.h
As an aide to debugging, we should print out the expected value of the
EEPROM checksum in addition to just saying that it is wrong.
Signed-off-by: Kyle Moffett
Cc: Ben Warren
---
No changes since v1
drivers/net/e1000.c | 48 ++--
1 files changed
hopefully make such EEPROM errors much more
straightforward to handle correctly in boot scripts and the like.
It is also necessary for a followup patch which allows SPI programming
of an e1000 controller's EEPROM even if the checksum is invalid.
Signed-off-by: Kyle Moffett
Cc: Ben Warren
---
looks like this on the wire:
Time >>>
--
... [7, 6, 5, 4, 3, 2, 1, 0, 15, 14, 13, 12, 11, 10, 9, 8], ...
--
(MSB is 15, LSB is 0).
Signed-off-by: Kyle Moff
To make it possible to use the "sspi" command with the e1000 firmware
EEPROM we add a small "generic SPI" driver wrapper around the existing
e1000 SPI backend.
Signed-off-by: Kyle Moffett
Cc: Ben Warren
---
No changes since v1
driver
ned-off-by: Kyle Moffett
Cc: Jerry Van Baren
---
Changelog:
v2: Resubmitted separately from the other HWW-1U-1A changes.
common/fdt_support.c |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/common/fdt_support.c b/common/fdt_support.c
index 6c98e5b..edcf04a 100644
--
the Linux kernel seems
to require a full 64-bit divide on any 32-bit PowerPC.
Build-and-boot-tested on the HWW-1U-1A board using DDR2 SPD detection.
Signed-off-by: Kyle Moffett
Cc: Andy Fleming
Cc: Kumar Gala
Cc: Wolfgang Denk
Cc: Kim Phillips
---
Ok, so this patch touches a rather sensitive p
06706d6e4fbeb0216213, and are GPLv2+.
The Makefile framework was copied from the U-Boot ARM port.
Signed-off-by: Kyle Moffett
Cc: Wolfgang Denk
Cc: Kim Phillips
Cc: Andy Fleming
Cc: Kumar Gala
Cc: Stefan Roese
---
Changelog:
v2: Resubmitted separately from the other HWW-1U-1A patches
arch/
be enabled with CONFIG_MPC85XX_GENERIC_GPIO
Signed-off-by: Kyle Moffett
Cc: Andy Fleming
Cc: Kumar Gala
Cc: Peter Tyser
---
Changelog:
v2: Moved the inline functions to a non-board-specific header
v3: Added generic Linux-standard GPIO wrappers
v4: Improved comments and fixed minor bugs in
switch statement.
Signed-off-by: Kyle Moffett
Cc: Andy Fleming
Cc: Kumar Gala
Cc: Kim Phillips
Cc: Wolfgang Denk
---
Changelog:
v2: Moved the constants to include/ddr_spd.h and also fixed DDR3.
v3: No changes
v4: Fixed up excessively long lines
arch/powerpc/cpu/mpc8xxx/ddr
seems to work, but I don't
yet have a usable 36-bit kernel or DTB, so it's mostly untested.
* CPU reset is a little quirky due to hardware misfeature, see the
extensive comments in the board_reset() function in hww1u1a.c
Signed-off-by: Kyle Moffett
Cc: Andy Fleming
Cc:
th of them.
A new generic system_reset() function was introduced which simply calls
board_reset() and cpu_reset() in that order, and all explicit callers of
do_reset() were changed to use system_reset() instead.
Signed-off-by: Kyle Moffett
Cc: Wolfgang Denk
Cc: Albert Aribaud
Cc: Mike Frysinge
fically, the MIPS do_reset() function used to look like this:
_machine_restart();
printf("*** restart failed ***\n");
When those platforms are fixed then it should be safe to remove the weak
fallback __arch_restart() function.
Cheers,
Kyle Moffett
_
art() hook. This works for all
hardware platforms which have a properly defined hardware reset
capability.
Signed-off-by: Kyle Moffett
Cc: Wolfgang Denk
---
api/api.c |3 +-
common/cmd_boot.c | 128 -
common/cmd_bootm.c
do_reset() functions.
Signed-off-by: Kyle Moffett
Cc: Wolfgang Denk
---
board/dave/PPChameleonEVB/PPChameleonEVB.c |4 ++--
board/esd/apc405/apc405.c |4 ++--
board/esd/ar405/ar405.c|2 +-
board/esd/ash405/ash405.c |4 ++--
board
Divide-by-zero errors should be controlled by the CONFIG_PANIC_HANG
configuration variable just like any other kind of trappable U-Boot bug.
Signed-off-by: Kyle Moffett
Cc: Albert Aribaud
---
arch/arm/lib/div0.c |6 +++---
1 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/arch
: Kyle Moffett
Cc: Albert Aribaud
---
arch/arm/lib/interrupts.c | 21 +++--
1 files changed, 7 insertions(+), 14 deletions(-)
diff --git a/arch/arm/lib/interrupts.c b/arch/arm/lib/interrupts.c
index 74ff5ce..156a23c 100644
--- a/arch/arm/lib/interrupts.c
+++ b/arch/arm/lib
The exported function table already assigns XF_do_reset properly, so
there should be no need for this processor to manually reassign it.
This change is required before ARM is converted away from using the
legacy do_reset() handler.
Signed-off-by: Kyle Moffett
Cc: Albert Aribaud
Cc: Reinhard
().
Signed-off-by: Kyle Moffett
Cc: Albert Aribaud
Cc: Reinhard Meyer
---
arch/arm/cpu/arm920t/at91/reset.c |4 ++--
arch/arm/cpu/arm920t/at91rm9200/reset.c |6 +++---
board/atmel/at91rm9200dk/at91rm9200dk.c |2 +-
3 files changed, 6 insertions(+), 6 deletions(-)
diff --git a/arch
reset code will probably work even when the CPU is in a bad state,
so no separate __arch_emergency_restart() function is required.
Signed-off-by: Kyle Moffett
Cc: Albert Aribaud
---
arch/arm/lib/reset.c |6 +-
1 files changed, 1 insertions(+), 5 deletions(-)
diff --git a/arch/arm/lib
The bfin_reset_or_hang function unnecessarily duplicates the panic()
logic based on CONFIG_PANIC_HANG.
This patch deletes 20 lines of code and just calls panic() instead.
This also makes the following generic-restart conversion patch simpler.
Signed-off-by: Kyle Moffett
Cc: Mike Frysinger
-off-by: Kyle Moffett
Cc: Reinhard Meyer
---
arch/avr32/cpu/cpu.c |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/arch/avr32/cpu/cpu.c b/arch/avr32/cpu/cpu.c
index e4489bb..ca2226b 100644
--- a/arch/avr32/cpu/cpu.c
+++ b/arch/avr32/cpu/cpu.c
@@ -76,7 +76,7 @@ void
reset code will probably work even when the CPU is in a bad state,
so no separate __arch_emergency_restart() function is required.
Signed-off-by: Kyle Moffett
Cc: Graeme Russ
---
arch/i386/cpu/cpu.c |4 +---
1 files changed, 1 insertions(+), 3 deletions(-)
diff --git a/arch/i386/cpu/cpu.c b
CPU may be in
a bad or undefined state, but it *appears* to be OK.
As a result no separate __arch_emergency_restart() function should be
needed.
Signed-off-by: Kyle Moffett
Cc: Mike Frysinger
---
arch/blackfin/cpu/cpu.h |1 -
arch/blackfin/cpu/reset.c |4 +---
board/bf537
back to their FLASH reset
vector, therefore they have no-op __board_emergency_restart() functions.
(If the CPU is in an invalid state then that probably won't work).
Signed-off-by: Kyle Moffett
Cc: Shinya Kuribayashi
---
arch/mips/cpu/cpu.c| 13 -
arch/mips/includ
an invalid state then jump-to-FLASH probably won't
work.
Signed-off-by: Kyle Moffett
Cc: Scott McNutt
---
arch/nios2/cpu/cpu.c | 12 +++-
1 files changed, 11 insertions(+), 1 deletions(-)
diff --git a/arch/nios2/cpu/cpu.c b/arch/nios2/cpu/cpu.c
index ef360ee..9f40188 100644
-off-by: Kyle Moffett
Cc: Nobuhiro Iwamatsu
---
arch/sh/lib/board.c |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/arch/sh/lib/board.c b/arch/sh/lib/board.c
index 07361b4..9206311 100644
--- a/arch/sh/lib/board.c
+++ b/arch/sh/lib/board.c
@@ -204,7 +204,7 @@ void
y_restart() function.
If the CPU is in an invalid state then jump-to-RAM probably won't work.
All of the rest of the reset code will probably work even when the CPU
is in a bad state, so they should be fine with the defaults.
Signed-off-by: Kyle Moffett
Cc: Wolfgang Denk
Cc: Stefan Roese
C
ther or not the code properly supports emergency
restart. A new __board_emergency_restart() may need to be implemented.
Signed-off-by: Kyle Moffett
Cc: Michal Simek
---
.../xilinx/microblaze-generic/microblaze-generic.c |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/boa
appear to either use
a CPU hardware feature or an onboard watchdog unit to perform the system
reset. As a hardware reset that will probably work even when the CPU is
in a bad state, so no separate __arch_emergency_restart() is required.
Signed-off-by: Kyle Moffett
Cc: Jason Jin
---
arch/m68k/cpu
All of the users of the legacy do_reset() function have been converted
to __arch_restart() or __board_restart() as appropriate, so the
compatibility calls to do_reset() may be removed.
In addition the do_generic_reset() function is renamed to the now-unused
name do_reset().
Signed-off-by: Kyle
together and moving
them into common code.
Signed-off-by: Kyle Moffett
Cc: Daniel Hellstrom
---
arch/sparc/cpu/leon2/cpu.c | 18 --
arch/sparc/cpu/leon3/cpu.c | 19 ---
arch/sparc/lib/board.c | 11 +++
3 files changed, 11 insertions(+), 37
invalid state then jump-to-FLASH probably won't
work.
Signed-off-by: Kyle Moffett
Cc: Daniel Hellstrom
---
arch/sparc/lib/board.c | 12 +++-
1 files changed, 11 insertions(+), 1 deletions(-)
diff --git a/arch/sparc/lib/board.c b/arch/sparc/lib/board.c
index 128ece7..e250274 1
together and moving
them into common code.
Signed-off-by: Kyle Moffett
Cc: Nobuhiro Iwamatsu
---
arch/sh/cpu/sh2/cpu.c |7 ---
arch/sh/cpu/sh2/watchdog.c |9 -
arch/sh/cpu/sh3/cpu.c |7 ---
arch/sh/cpu/sh3/watchdog.c |9 -
arch/sh/cpu/sh4/cpu.c
ardware without those patches.
It will not cause a build failure however, as this board port simply defines
an otherwise-unused __board_restart() function.
I believe these 3 patches at least are ready to be merged, as all concerns
presented so far have been resolved.
Cheer
functions. This can be enabled with the config option
CONFIG_MPC85XX_GENERIC_GPIO.
Signed-off-by: Kyle Moffett
Cc: Andy Fleming
Cc: Kumar Gala
Cc: Peter Tyser
---
arch/powerpc/include/asm/mpc85xx_gpio.h | 120 +++
1 files changed, 120 insertions(+), 0 deletions
switch statement.
Signed-off-by: Kyle Moffett
Cc: Andy Fleming
Cc: Kumar Gala
Cc: Kim Phillips
---
arch/powerpc/cpu/mpc8xxx/ddr/ddr2_dimm_params.c | 23 +++-
arch/powerpc/cpu/mpc8xxx/ddr/ddr3_dimm_params.c | 43 ++
common/ddr_spd.c
seems to work, but I don't
yet have a usable 36-bit kernel or DTB, so it's mostly untested.
* CPU reset is a little quirky due to hardware misfeature, see the
extensive comments in the __board_restart() function in hww1u1a.c
Signed-off-by: Kyle Moffett
Cc: Andy Fleming
Cc:
to require a full 64-bit divide on my 32-bit PowerPC.
Build-and-boot-tested on the HWW-1U-1A board using DDR2 SPD detection.
Signed-off-by: Kyle Moffett
Acked-by: York Sun
Cc: Andy Fleming
Cc: Kumar Gala
--
Changelog:
v2: Resubmitted separately from the other HWW-1U-1A patches
v3: Rebased on
ies.
I believe these 4 patches are ready to be merged, as all concerns presented
so far have been addressed.
Cheers,
Kyle Moffett
___
U-Boot mailing list
U-Boot@lists.denx.de
http://lists.denx.de/mailman/listinfo/u-boot
be enabled with CONFIG_MPC85XX_GENERIC_GPIO
Signed-off-by: Kyle Moffett
Cc: Andy Fleming
Cc: Kumar Gala
Cc: Peter Tyser
---
Changelog:
v2: Moved the inline functions to a non-board-specific header
v3: Added generic Linux-standard GPIO wrappers
v4: Improved comments and fixed minor bugs in the
switch statement.
Signed-off-by: Kyle Moffett
Cc: Andy Fleming
Cc: Kumar Gala
Cc: Kim Phillips
---
Changelog:
v2: Moved the constants to include/ddr_spd.h and also fixed DDR3
v3: No changes
v4: Fixed up excessively long lines
v5: No changes
v6: Rebased onto the 'next' bra
The HWW-1U-1A board needs to be able to override the "reset" command due
to hardware design limitations.
Signed-off-by: Kyle Moffett
Cc: Andy Fleming
Cc: Kumar Gala
---
Changelog:
v2: Removed in favor of more involved reset rework
v6: Resurrected again (the more involved rework
seems to work, but I don't
yet have a usable 36-bit kernel or DTB, so it's mostly untested.
* CPU reset is a little quirky due to hardware misfeature, see the
extensive comments in the board_reset() function in hww1u1a.c
Signed-off-by: Kyle Moffett
Cc: Andy Fleming
Cc:
eady to be merged.
Cheers,
Kyle Moffett
___
U-Boot mailing list
U-Boot@lists.denx.de
http://lists.denx.de/mailman/listinfo/u-boot
switch statement.
Signed-off-by: Kyle Moffett
Cc: Andy Fleming
Cc: Kumar Gala
Cc: Kim Phillips
--
Changelog:
v2: Moved the constants to include/ddr_spd.h and also fixed DDR3
v4: Fixed up excessively long lines
v6: Rebased onto the 'next' branch of git://git.denx.de/u-boot-mpc85xx.
be enabled with CONFIG_MPC85XX_GENERIC_GPIO.
Signed-off-by: Kyle Moffett
Cc: Andy Fleming
Cc: Kumar Gala
Cc: Peter Tyser
--
Changelog:
v2: Moved the inline functions to a non-board-specific header
v3: Added generic Linux-standard GPIO wrappers
v4: Improved comments and fixed minor bugs in the
The HWW-1U-1A board needs to be able to override the "reset" command due
to hardware design limitations.
Signed-off-by: Kyle Moffett
Cc: Andy Fleming
Cc: Kumar Gala
--
Changelog:
v2: Removed in favor of more involved reset rework
v6: Resurrected again (the more involved rework was
seems to work, but I don't
yet have a usable 36-bit kernel or DTB, so it's mostly untested.
* CPU reset is a little quirky due to hardware misfeature, see the
extensive comments in the board_reset() function in hww1u1a.c
Signed-off-by: Kyle Moffett
Cc: Andy Fleming
Cc:
, "e1000_spi.c", and it calls a few exported
functions in the main e1000.c file to coordinate EEPROM access.
Comments, critiques, and compliments are highly appreciated.
Cheers,
Kyle Moffett
--
Curious about my work on the Debian powerpcspe port?
I'm keeping a blog here: http://purepe
1 - 100 of 113 matches
Mail list logo
