Probably you don't have /dev/console in your HDD copy.
Probably you missed some flags to the 'cp' command or your original
filesystem doesn't have /dev/console at all.
yang shaobo wrote:
I made a small filesystem by Devrocket ( follow this article :
Use ip=board_ipaddress:nfs_server_addess or ip=none if you want to
disable kernel IP autoconfiguration at all.
See Documentation/nfsroot.txt:
ip=client-ip:server-ip:gw-ip:netmask:hostname:device:autoconf
This parameter tells the kernel how to configure IP addresses of devices
and also how
No luck,shared libraries are also giving same issue.
is there any other compilation option which forces strict alignment?
If the sources you are trying to use have been written without alignment
in mind, I guess you won't be able to 'fix' things by only setting a
compilation switch.
Hi ashok,
See this page: http://www.nslu2-linux.org/wiki/Info/Alignment
Ivan
[EMAIL PROTECTED] wrote:
Hi,
i have some apllication which are running on x86 without any issue.
But if i cross compile those application for davinci board with MV
toolchain, it is getting segmentation fault.
Hi Robert,
However, vmlist is not exported, and I would expect that there is a
helper function somewhere for traversing vmlist, but I can't find
anything.
Probably you can consider splitting CMEM in a always-bult-in and
built-in-or-module part?
For user-to-phys address translation, I
Hi Brian,
I'm not a Codec Engine expert, but recently I ran into a similar problem.
For one of our projects we had one physical area in memory readable and
writable from:
Kernel
Userspace
EDMA
The mappings of Kernel and Userspace used different virtual addresses (and
thus occupied
Align your ramdisk size to a multiple of 512 (the page size of your NAND)
0x16cbcc -- 0x16cc00 ==
nand write 0x8070 0x104000 0x16CC00
Ivan
___
Davinci-linux-open-source mailing list
Davinci-linux-open-source@linux.davincidsp.com
Hi Ruiyu,
1. Your erase command is not correct.
2. You didn't enable MTD block device support in kernel config
JFFS2 needs the so-called cleanmarkers written to empty blocks. For
NAND, these are special markers written in spare area. Instead of
flash_eraseall, use flash_eraseall -j.
And for
Hi,
DaVinci RBL expects ECC to be placed in a non-standard format.
See sprue14a.pdf, page 131 for ECC format description.
You can also see code from TI UBL or one of its opensource replacements
(see [1]).
Ivan
[1]
http://wiki.davincidsp.com/index.php?title=RBL_UBL_and_host_program
Hi Lloyd,
Personally, I am more interested in SOURCE portability than executable
portability. My development tools are written in C or CPP - these are the
gold standard of Linux.
I have seen little that .NET offers that provides anything at this low a
level. Frankly, downloaders are trivial
Hi Thom,
Do you think this low cost DaVinci unit will be powerful enough to play
H.264 video?
BR,
Ivan
Maughan, Thomas wrote:
Dirk,
This is indeed a low cost DaVinci processor.
The DM355 DVEVM (board + dvsdk + 'demonstration version' of montavista
pro v4.0 + beta 2.6.10 lsp) cost is $495
Hi all,
We are already discussing UBLs and their flashing. I think that we can
extend this topic and start talking about complete firmware flashing and
upgrade. By complete firmware I mean UBL, Tertiary Bootloader, Kernel,
Userspace, FPGA Configuration and so on.
What do you think?
BR,
Ivan
Hi Lloyd,
U-Boot can load kernel images from JFFS2. Unfortunately JFFS2 mount (scan)
speed is VERY poor even with the read-only option turned on.
I can't give you concrete figures, but as far as I remember, on a 2 MB
uncompressed, read-only JFFS2 partition with only one file -- the kernel,
Hi Dirk,
I think we should take a two step approach here:
1) Have *one* functional, flexible, extensible, under public source
control (...) open source replacement for TI DaVinci bootloader.
2) Try to use this then golden bootloader to start a basic
discussion about the above topics on
Hi Daniel,
I am developing a DVFlasher/TI UBL replacement for our boards.
I need some help regarding RBL UART bootmode.
Is it possible, from NAND UBL, to jump to some address (0x4xxx) and have
RBL start UART boot (BOOTME BOOTME BOOTME BOOTME)?
I need this in case regular NAND UBL fails to
Do you have /dev/console in your rootfs?
Andrew Armstrong wrote:
I have managed to set the framebuffer as a virtual terminal output, so I
can see all my kernel messages, however I get an unable to open an
initial console message.
In this particular example I have disabled the serial port as I
Hi Anrew,
mknod /dev/console c 5 1
BR,
Ivan
___
Davinci-linux-open-source mailing list
Davinci-linux-open-source@linux.davincidsp.com
http://linux.davincidsp.com/mailman/listinfo/davinci-linux-open-source
Cool :)
By the way, did you patch framebuffer driver to get the right FB console
output? (By right I mean -- with correct colors)
BR,
Ivan
Andrew Armstrong wrote:
Ivan,
That does the trick nicely!!
Regards,
Andrew
On Thu, 2007-08-30 at 15:57 +0300, Ivan Tonchev wrote:
Hi Anrew
Hi Andrew,
I did not, is that why I am missing the first few characters of the
screen as well?
If you are talking about the black messages on black background - yes.
can you point me in the right direction of the patch!?
I am using OSD0 in 8 bit mode with OSD1 disabled. Colors are
Dirk Behme wrote:
Dirk Behme wrote:
Ivans mail: Ivan Tonchev wrote:
Hi Dirk,
Thanks for posting your sources. Unfortunately I am very pressed on time
and I can't look at them now.
Very happy to hear about yours and Sergey's work!
What I have is a somewhat complete flashing toolchain
Meher,
ARM9 vectors may be located either at addresses 0x (low) or
0x (high).
I guess your vector setting is high and your NOR flash is also visible
at 0x.
AFAIK, vector table location is controlled by
1. CP15 (ARM core coprocessor 15)
2. ARM core HIVECS pin
I
Where is this mapping information available?
I don't know. Ask TI people about it.
Or prove/unprove my suggestion with CCS yourself!
Connect with JTAG and compare memory regions 0x and 0x0200,
size 0x1
Good luck,
Ivan
___
Hi Meher,
I am not sure how to intrepet this.. Also why should the exceptions go
to the 0x0200 location instead of 0x?
They don't go to 0x0200 -- this is impossible with standard ARM9 cores.
Do you mean 0x is mapped to NOR Flash at 0x0200?
I guess the first
Hi Vitaly,
I have faced the situation here. We are doing the custom design on the
base of Davinci EVM. Previous revision had the Samsung K9F1208 nand
and the configuration worked just fine with our current DVFlasher, UBL
and u-boot SW. Another revision based on STM NAND512R3A flash (which
Hi Vitaly,
Don't worry about bad blocks after 0x03FF. SW-managed BBT (bad block
table) is stored there. They are marked bad, because MTD guys don't
want us to erase BBT by mistake.
Good luck with TI people.
The delay you are talking about is huge. I guess RBL is doing something
Hi Vitaly,
Talking about delay that is weird if RBL tries to fix the ECC errors
because the flash is empty and filled with 0xFF as I can see it from
the u-boot.
I misunderstood you. I thought that on the STM board, when the flash was
empty, you saw no delay. But obviously you were talking
Hi Vitaly,
Chasing this nasty problem starts to get real fun :)
As another test I run the new board connected with JTAG to the CSS and
tried to debug the RBL. And got wierd results:
if I do step-by-step execution for the code in ARM memory from address
0x4000 to address 0x403C allowing about
Hi Eric,
Is anyone working on adding SD card support to U-Boot for the 644x?
I have ported the SD card project from DaVinci CCS examples. Unfortunately
this project is not GPL and I cannot disclose my work.
Speaking with TI about this was on my list, but noone showed interest for
the SD card
Hi Scott,
Dirk and Ivan,
Being a newbie to linux and u-boot, and having discovered, actually
been directed to this list, belatedly, I am sure I have not done things
in the most efficient method.
I started with u-boot-1.1.3. After a lot of thrashing, I am now able to
write the UBL and
Hi Scott
Is there any other way to write this fs from within the kernel without
access to nandwrite? Any other formatting I need to do on the nand?
You have to erase nand and write the so-called cleanmarkers in block
spare areas.
Under Linux, flash_eraseall -j will do the job.
Under
Thanks!
NP :)
And one thing I missed:
You can write your image with u-boot's nand write, as long as you have
nand erase clean-ed the desired area in NAND. And of course your ECC
implementations in u-boot and kernel are compatible.
If you use u-boot for flashing, you will not need
Hi Dirk,
Thank you for contributing!
- It does the above stuff (1 2) and therefore makes your patch more
generic. Have a look to board/davinci/nand.c part of patch in
attachment. There, it introduces an additional #define
ENABLE_DAVINCI_NAND_HW_ECC which can be used to switch between
OK, I'll check with Roger and answer your thread in the u-boot mailing list.
BR,
Ivan
Philip Balister wrote:
Ivan Tonchev wrote:
The patch is against git. I am very interested in structural comments,
like the best way to support similar, but different boards.
My patch is against u-boot 1.2
Hi Lewis,
I think that ZD1211 based devices should work. We've used such a USB
device with another TI SoC running ARM9 Linux.
BR,
Ivan
Lewis Liu wrote:
Hi, Friends,
I'm planning to use the USB wireless card on the Davinci board. Does
anyone know if there is any special model that works?
Hi Yogesh,
I haven't tested this on GIT tree, but on 2.6.10 it works. At least for
SD :)
BR,
Ivan
Yogesh Kumar M wrote:
Hi all,
I use Git kernel 2.6.20 version,when compile the mmc/sd driver, error likes:
error: 'struct mmc_data' has no member named 'req'
error: 'struct mmc_data' has no
Hi all,
Do you know whether DaVinci MMC/SD supports MMC?
I have tested with with SD and MMC cards and it works fine with the SDs,
but doesn't detect the MMCs.
Do I have to change something to make it work with MMC?
BR,
Ivan
___
Hi Huy,
What's the status of the Wiki/FAQ proposal.
Over the past few days I collected valuable information about DaVinci
RBL ECC/Bad block (non)handling and I'd like to share it.
Is there an official place for such content already?
BR,
Ivan
Pham, Huy wrote:
Jean,
We're looking at the
Hi Philip,
I am trying to get a patch together to add support for the TI Davinci
processor into u-boot. I have started with one of the patches floating
around the internet and am in the process of modifying to support the
Lyrtech SFF SDR board. Currently, this patch will boot on the board and
Hi Uthappa,
We were working on an SPI EDMA driver, but our buffers were less than a
kilobyte.
Unfortunately we found that when programmed with EDMA, DaVinci SPI
controller does not always send the first byte. Sometimes it sent it,
sometimes it started from the second byte...
Did you
Hi Andy,
Thanks, but how do I set the range uncacheable? I don't see such an
option for mmap. Please advise. Thanks!
Have you tried with pgprot_noncached()?
I haven't used it myself, but found it in some framebuffer mmap handling
code. For example see linux/drivers/video/sa1100fb.c
Hi Suba,
You can find a u-boot port for DM320 here:
http://svn.neurostechnology.com/svn/neuros-bsp/trunk/bootloader/
The interesting file is board/dm320/nand.c. One thing that needs
mentioning is that this source code uses software ECC, which is not
compatible with the DM320 RBL.
I don't know
My u-boot compilation seems to go well. But during
ld, there are errors like this:
_udivSi3.oS: warning: duplicate section
.note.gnu.arm.ident has different contents.
And the map file is created, but no .bin or .srec.
Any ideas?
Which version of u-boot are you compiling and with
Hi Suba,
Why would you want the gel file?
I just wanted to compare it with the GEL files I use. I see that you
modified PLL, DDR2 and AEMIF initializations.
When using this gel
file and the JTAG, there are no ECC errors and the UBL
and u-boot are flashed correctly using that.
[mailto:[EMAIL PROTECTED]
Sent: Friday, April 27, 2007 7:40 AM
To: Ivan Tonchev; Allred, Daniel
Cc: Linux DaVinci; Subashini Ravindran
Subject: Re: UBL on custom board
Hi All,
We implemented the patch provided by Ivon
Tonchev. But the problem still persists. There is no
change
How much RAM do you have on your board?
Is it 256M?
ketan deshpande wrote:
Hi,
We are trying the DVFlasher_1_12 with the custom designed board
with NAND Flash.
The NAND flash chip we are using is a big block device.When we try to
flash the UBL we get the following messages.
Hi Andy, Constantine,
We're interested in this item too.
Is this Kconfig change sufficient for what we want? I'm not sure, but I
think there is more to be changed. It seems to me that Host/Peripheral
mode is selected by Kconfig and hardcoded in kernel image. See
, no software flow control
2. Then I read serial port with cat /dev/ttyS0
3. And I write to it with echo something /dev/ttyS0
You can achieve 1) with other tools, of which stty is a possible example...
Ivan Tonchev wrote:
And if you're OK with implementing your application in bash, you can
use
It'd be also great if you make the UBL more straightforward to port for
boards with less than 256 MB memory.
For current implementation I had to change:
.ddrram2 starting address in ubl_davinci.lds,
MAX_IMAGE_SIZE in ubl.h and
RAM_END_ADDR in ubl.h
For my memory map (128 mb), I set
Daniel J. Allred
Software Applications
Catalog DSP / Emerging End Equipment
*From:* Ivan Tonchev [mailto:[EMAIL PROTECTED]
*Sent:* Thursday, March 15, 2007 11:04 AM
*To:* Allred, Daniel
*Cc:* Monk, Roger; Linux DaVinci
*Subject
:* Ivan Tonchev [mailto:[EMAIL PROTECTED]
*Sent:* Thursday, March 15, 2007 11:53 AM
*To:* Allred, Daniel
*Cc:* Monk, Roger; Linux DaVinci
*Subject:* Re: UBL for 128MB DDR
Great to hear that you'd be dropping the s-rec format. It's bloats
images quite a bit and makes downloads over the UART more
Hi all,
I want to access my HDD in UDMA4 mode (ATA 66). Unfortunately it seems
that with current davinci driver, this isn't possible. Am I wrong?
hdparm tells me that I'm in udma2 and when I try to set udma4, I receive
this message: ide0: Speed warnings UDMA 3/4/5 is not functional.
If I'm
Hi DaVinci experts,
Has anyone of you used FIQs with DaVinci? Is this possible with the
current MV codebases (GIT and stable)? If not does anyone has pointers
to what modifications have to be applied? Any kind of examples would be
greatly appreciated!
Please help me. I'm a noob :)
Thanks,
Hi Daniel,
Link is up. I peeked at UBL source code and it seems that it has the
soft reset fix integrated. I haven't tested it yet though.
Ivan
Allred, Daniel wrote:
The link on the DVEVM update site has been fixed. Sorry for that!
Daniel
Daniel J. Allred
Software Applications
Hi Andy,
You can load u-boot to your board via JTAG. You'd need a JTAG and a
working GEL file for your board which you probably already have.
Code Composer Studio can load elf files, so assuming that your DDR is
initialized correctly (by the GEL file), you can load a compiled u-boot
*From:* Ivan Tonchev [mailto:[EMAIL PROTECTED]
*Sent:* Monday, March 12, 2007 5:22 AM
*To:* Allred, Daniel
*Cc:* Arie Muijnck; Carlos Ojea; Linux DaVinci
*Subject:* Re: New UBL source and DVFlasher App on DVEVM update site
Hi Daniel,
Link is up. I peeked at UBL source code and it seems that it has
*From:* Ivan Tonchev [mailto:[EMAIL PROTECTED]
*Sent:* Monday, March 12, 2007 5:22 AM
*To:* Allred, Daniel
*Cc:* Arie Muijnck; Carlos Ojea; Linux DaVinci
*Subject:* Re: New UBL source and DVFlasher App on DVEVM update site
Hi Daniel,
Link is up. I peeked at UBL source code
Hi Peter,
Can you share your u-boot-1.1.6 with us? AFAIK it has more flexible
support for NAND devices. E.g. you don't need to hardcode your u-boot for
particular NAND geometry, but rather it detects and handles this runtime.
It also supports a great deal of Manufacturer IDs and Device IDs (code
Hi Steve,
I reproduced your soft reset problem on our boards and on evm.
In the mean time TI provided us UBL source code, so I did some patching.
I implemented workarounds for both soft reset and STM NAND flash. Code
works OK at least for me. I submitted patches to Roger.
Ivan
Steve Yeh
this part.
Steve
2007/3/5, Ivan Tonchev [EMAIL PROTECTED]:
Hi Steve,
I reproduced your soft reset problem on our boards and on evm.
In the mean time TI provided us UBL source code, so I did some patching.
I
implemented workarounds for both soft reset and STM NAND flash. Code
works
OK
Hm, DaVinci soft reset is implemented as a watchdog timer reset (see
/include/asm/arch/system.h and arch/arm/mach-davinci/time.c:
davinci_watchdog_reset())
I don't know whether this watchdog reset resets MMU and if it doesn't,
whether the TI UBL takes steps to disable MMU before trying to
60 matches
Mail list logo