- Original Message -
From: Kumar Gala ga...@kernel.crashing.org
On Apr 12, 2011, at 12:04 AM, Alex Dubov wrote:
From: Alex Dubov oa...@yahoo.com
AMC8548 is a RapidIO development board in AMC form factor, featuring
MPC8548E
Patch needs some minor updates to apply build
Dear Alex Dubov,
In message 1318576261.97640.yahoomail...@web121013.mail.ne1.yahoo.com
you wrote:
I have looked into it previously but unfortunately, I could not think of a
reasonable solution.
Well, the ultimate cure to that problem is the removal of this board.
Boards
From: Kumar Gala ga...@kernel.crashing.org
When trying to build it we get:
Hi,
I have looked into it previously but unfortunately, I could not think of a
reasonable solution.
Some revisions of this board have only 16Mb of flash, so I thought it will
be wise to squeeze everything into one
From: Alex Dubov oa...@yahoo.com
AMC8548 is a RapidIO development board in AMC form factor, featuring MPC8548E
processor, DDR2 SO-DIMM slot, 16MB of hardwired NAND flash memory, real time
clock and additional serial EEPROM on i2c bus (enabled). USB controller is
available, but not presently
--- On Mon, 4/4/11, Wolfgang Denk w...@denx.de wrote:
In my blob I still got 8k free, and I don't think
there's too much use
for NFS support, which is enabled by default.
I just point out that this board does not build in our
build tests -
as the only PPC board, btw.
The error says
--- On Fri, 1/4/11, Wolfgang Denk w...@denx.de wrote:
In message 791834.81297...@web37605.mail.mud.yahoo.com
you wrote:
I pulled the git and recompiled, yet I don't see any
errors,
everything is
just fine. My compiler version is:
...
gcc version 4.5.1 (GCC)
Seems you are on
--- On Fri, 1/4/11, Andy Fleming aflem...@gmail.com wrote:
I ran into this, too, but it went
away when I updated my compiler. It
looks like the beginning was set a little too close to the
end?
I pulled the git and recompiled, yet I don't see any errors, everything is
just fine. My
is available at:
http://www.silicontkx.com/AMC8548.htm
Signed-off-by: Alex Dubov oa...@yahoo.com
---
This is really a patch repost from Fall, 2009. The board is old and never was
that good, but I still got those running, so I propped up an u-boot config as
well. Nothing fancy here, just a fairly
applied
- k
Thank you.
I have another board config for your consideration.
___
U-Boot mailing list
U-Boot@lists.denx.de
http://lists.denx.de/mailman/listinfo/u-boot
). This is achieved by means of
custom ld script.
Signed-off-by: Alex Dubov oa...@yahoo.com
---
Changes for v8:
- Move CONFIG_SYS_LDSCRIPT to board config file
- Explicitly refer to env_embedded.o in ld script, as this object is no
longer linked into libcommon.o
Changes for v7:
- Add
Please keep the boards.cfg clean from configuration
settings. Move
the definition of CONFIG_SYS_LDSCRIPT into your board
config header
file.
[Kumar also reported that this does not even build. I
confirm this.]
That's because of the relatively recent change to how the env_embedded.o
+/*
+ * For booting Linux, the board info and command line
data
+ * have to be in the first 16 MB of memory, since
this is
+ * the maximum mapped by the Linux kernel during
initialization.
+ */
+#define CONFIG_SYS_BOOTMAPSZ (16 20) /*
Initial Memory map for Linux*/
+
You
). This is achieved by means of
custom ld script.
Signed-off-by: Alex Dubov oa...@yahoo.com
---
Cahnges for v7:
- Add CONFIG_SYS_BOOTM_LEN define
Changes for v6:
- Remove 1s from feature defines and fix a SRIO ifdef in tlb.c
Changes for v5:
- Use new common SRIO configuration definitions
). This is achieved by means of
custom ld script.
Signed-off-by: Alex Dubov oa...@yahoo.com
---
Changes for v6:
- Remove 1s from feature defines and fix a SRIO ifdef in tlb.c
Changes for v5:
- Use new common SRIO configuration definitions.
- Replace initdram() with platform required
Environment is configured to precede the actual u-boot
image so that it's
located at the beginning of flash erase block (made
necessary by the recent
changes to the embedded environment handling).
I assume customer linker script is because of where you
have the environment. If
). This is achieved by means of
custom ld script.
Signed-off-by: Alex Dubov oa...@yahoo.com
---
Changes for v5:
- Use new common SRIO configuration definitions.
- Replace initdram() with platform required fixed_sdram().
- Don't use get_ram_size() in fixed_sdram() as TLBs are set after
).
Signed-off-by: Alex Dubov oa...@yahoo.com
---
Changes for v4:
- Replace config.mk supplied linker flags with custom linker script.
- Fix checkpatch errors.
Changes for v3:
- Use io accessor functions for all mmio accesses.
- Add configuration options for RTC and EEPROM on I2C buses
It seems, something had broke the embedded environment compilation in the
recent git head.
During DO_DEPS_ONLY phase, environment.h is not considered, thus
CONFIG_ENV_IS_EMBEDDED is never set with the obvious consequences.
___
U-Boot mailing
1. I want to fit an environment and the bootloader
into a single flash
sector.
This is a very bad idea as it will open a window brick your
system at
each and every saveenv command.
I strongly recommend not to do that.
This issue is well understood.
Yet, we are talking an
atum8548
mpc8540
pm854/6
sbc8560
sbc8641d
socrates
mpc8536ds
mpc8540ads
mpc8560ads
mpc8572ds
mpc8610hpdc
mpc8641hpcn
p2020ds
tqm85xx
Could you please be a bit more specific, i. e. like quoting
file names
and line numbers?
These were produced by
grep -nr
).
Signed-off-by: Alex Dubov oa...@yahoo.com
---
MAINTAINERS|3 +
board/mercury/mpq101/Makefile | 53 +
board/mercury/mpq101/config.mk |6 +
board/mercury/mpq101/law.c | 55 ++
board/mercury/mpq101/mpq101.c | 146 ++
board/mercury/mpq101
/ensemble_mpq101_rapidio_powerquicc_iii.aspx
Environment is configured to precede the actual u-boot image so that it's
located at the beginning of flash erase block (made necessary by the recent
changes to the embedded environment handling).
Signed-off-by: Alex Dubov oa...@yahoo.com
---
MAINTAINERS
Greetings.
In the old times it was customary to tweak the linux kernel for each and
every board, creating named configurations, so naturally dts files ended
up in the kernel source tree. Yet, kernel build system does nothing with
those files. Moreover, during initial board development, u-boot
+phys_size_t
+initdram (int board_type)
+{
...
+ dram_size = setup_ddr_tlbs(1ull
(CONFIG_SYS_SDRAM_SIZE_LOG - 20));
+
+ puts( DDR: );
+
+ return dram_size 20;
You should use get_ram_size().
I still need to call setup_ddr_tlbs because get_ram_size won't do it for
+CONFIG_SYS_TEXT_BASE = 0xfffc0800
+LDFLAGS +=
--section-start=.ppcenv=$(CONFIG_ENV_ADDR)
Please mode defines to board config file and get rid of
config.mk
And another question: how am I supposed to specify linker flags in the
board config file (I suppose, you mean boards.cfg). There's
/ensemble_mpq101_rapidio_powerquicc_iii.aspx
Signed-off-by: Alex Dubov oa...@yahoo.com
---
board/mercury/mpq101/Makefile | 53 ++
board/mercury/mpq101/config.mk |7 +
board/mercury/mpq101/law.c | 54 ++
board/mercury/mpq101/mpq101.c | 167 ++
board/mercury/mpq101
Until now, custom environment settings were just defined in the board config
file. All that had to be done is to edit the file, recompile the u-boot and
flash it to the target.
Now, however, this approach is not longer acceptable. This means, that
environment must be either set manually on the
Just a general observation: if you are not sure that
tlb/law files can be
safely factored out and most of the ddr.c files are
actually board specific
overrides (common part being less than 10 lines) why
had you requested me
to create a common board hierarchy for these STX
boards in
-off-by: Alex Dubov oa...@yahoo.com
---
Makefile |3 +
board/stx/stxamc8548/Makefile | 51 ++
board/stx/stxamc8548/config.mk| 32
board/stx/stxamc8548/stxamc8548.c | 168 +++
board/stx/stxamc8548/u-boot.lds | 143
Put environment into .ppcenv section aligned on a
smaller boot eraseblock
boundary near flash end.
Hm... if you change to code to using a smaller boot
sector...
#define CONFIG_ENV_ADDR
(CONFIG_SYS_MONITOR_BASE + 0x38000)
#define
CONFIG_ENV_SECT_SIZE 0x4000 /*
Move board definition files for STx XTC, GP3 and SSA
boards into
common subdirectory and factor out common code.
-mno-spe flag common to all MPC85xx configurations
does not work
so change it to -mspe=no which does (GCC bug
37759).
Signed-off-by: Alex Dubov oa...@yahoo.com
Just a general observation: if you are not sure that tlb/law files can be
safely factored out and most of the ddr.c files are actually board specific
overrides (common part being less than 10 lines) why had you requested me
to create a common board hierarchy for these STX boards in the first
-off-by: Alex Dubov oa...@yahoo.com
---
board/stxamc8548/Makefile | 54 ++
board/stxamc8548/config.mk| 32
board/stxamc8548/ddr.c| 87 +
board/stxamc8548/law.c| 54 ++
board/stxamc8548/stxamc8548.c | 260 +++
board
I'm trying to setup a new u-boot on an old STx AMC8548 board. I discovered the
reason for instability I reported earlier, namely DDR's twoT_en option was not
set. However, now I have a different problem.
Attempt to access the network always fails with Got error 4 message. I can
see ARP
I've got an old STx AMC8548 board. It came with u-boot-1.1.4 which works.
I upgraded it to stock u-boot-1.3.4 which also works. However, nothing
newer will, despite a lot of effort I spent in a last couple of days.
The crux of the problem: cpu gets a weird program error exception
out of nowhere
35 matches
Mail list logo