[U-Boot-Users] at91sam9260-ek intermittenly fails at u-boot init
I am struggling with this problem for a week now and not making any progress. So any suggestions, ideas or pointer will be greatly appreciated.
The board works with u-boot 1.3.2 and at91bootstrap(1.5 I only have the binary) booting from NAND. We have changes the SDRAM from 32bit databus to 16bit and NAND to 32M. I am using U-boot 1.3.2 and AT91bootStrap 1.8.
I changed the SDRAMC and PIOC init per Atmel's application notes in at91bootstrap.
U-boot fails intermittently at the start. When it fails, it didn't make it to my first printf statement. When I use JTAG to debug it, it always works.
If the board has been let without power for 1+ hour, it will also boot up when I apply power to it(something to do with the battery perhaps). Once it booted up,it is very stable. I can run Mtest etc.
We tried replacing NAND, SDRAM, etc. Along the way we are battling bugs with SAM-BA as well. We think that we have observed the same behavior on another board. We also experimented with different bootstraps and u-boot binaries.
The only thing we can think of is Reset cycle or some h/ware init code in u-boot that might not be compatible with at91bootstrap's init code.
We had at91bootstrap 1.5 working well in 32-bit mode with u-boot 1.3.2. But I don't have the source. I wonder what is the current stable version of u-boot and at91bootstrap?
Thanks in advance for your help.
Norman
Norman Cheung wrote:
U-boot fails intermittently at the start. When it fails, it didn't make it to my first printf statement. When I use JTAG to debug it, it always works.
If the board has been let without power for 1+ hour, it will also boot up when I apply power to it(something to do with the battery perhaps). Once it booted up,it is very stable. I can run Mtest etc.
...
The only thing we can think of is Reset cycle or some h/ware init code in u-boot that might not be compatible with at91bootstrap's init code.
We had *exactly* the same problem with our AT91RM9200 based custom board recently. After battling lots of days trying to find a bug within software we realized that problem actually hides in hardware. Boards reset line is controlled via small supervisor AVR. After tuning that watchdog properly according to datasheets and leaving small reserve, finally board is working as supposed. Trying to fix this I believe we read half of the internet topics related to that :), and I have found similar problem description in details (don't remember where, though). There was one guy claiming that there should be correct reset procedure during cold reset, or AT91 CPU's memory controller will fail to initialize. Based on that, we had reviewed our watchdog timings, have separated NRST and NTRST (!) giving each personal delay, and now it rocks.
I sincerely believe that french engineers of Atmel's EK/DK boards are too competent to make such a mistake, but probably it is worth trying to test this.
Kind regards, Andrejs Cainikovs.
Le mercredi 25 juin 2008 à 01:36 +0000, Norman Cheung a écrit :
I am struggling with this problem for a week now and not making any progress. So any suggestions, ideas or pointer will be greatly appreciated.
[...]
The board works with u-boot 1.3.2 and at91bootstrap(1.5 I only have the binary) booting from NAND.
The source (for the latest version) is freely available on the Atmel site: http://www.atmel.com/dyn/products/tools_card.asp?tool_id=4093
If you really need the source for an older version, I'm sure the Atmel support will send it to you upon request.
U-boot fails intermittently at the start. When it fails, it didn't make it to my first printf statement. When I use JTAG to debug it, it always works.
If the board has been let without power for 1+ hour, it will also boot up when I apply power to it(something to do with the battery perhaps).
Did you try disconnecting the battery ? Or replacing the battery with a new one ?
[...]
We had at91bootstrap 1.5 working well in 32-bit mode with u-boot 1.3.2. But I don't have the source.
See above.
I wonder what is the current stable version of u-boot and at91bootstrap?
AT91Bootstrap seems to be at v1.9.
As for U-Boot, try the latest git (which will soon be u-boot v1.3.4).
(BTW, if Wolfgang is listening, it would be nice to have proper tags in git as well as releases on the ftp server for the -pre version. Since the merge window is closed, I guess we're up to v1.3.4-pre1 now, but this version tag cannot be found anywhere...)
Stelian.
(BTW, if Wolfgang is listening, it would be nice to have proper tags in git as well as releases on the ftp server for the -pre version. Since the merge window is closed, I guess we're up to v1.3.4-pre1 now, but this version tag cannot be found anywhere...)
good idea
Best Regards, J.
Many thanks Stelian for the replies. My comments inline.
Norman
Stelian Pop wrote:
Le mercredi 25 juin 2008 à 01:36 +0000, Norman Cheung a écrit :
I am struggling with this problem for a week now and not making any progress. So any suggestions, ideas or pointer will be greatly appreciated.
[...]
Since the post earlier, I have narrowed the problem to the bootstrap during the copy from NAND to SDRAM, memories are corrupted. I think I have a marginal NAND chip. I played around the SMC cycle timing in bootstrap, I have a setting that I can boot u-boot with power cycle 6 out of 10 times. So I am convinced that it is not an u-boot issue, Thanks!
The board works with u-boot 1.3.2 and at91bootstrap(1.5 I only have the binary) booting from NAND.
The source (for the latest version) is freely available on the Atmel site: http://www.atmel.com/dyn/products/tools_card.asp?tool_id=4093
If you really need the source for an older version, I'm sure the Atmel support will send it to you upon request.
U-boot fails intermittently at the start. When it fails, it didn't make it to my first printf statement. When I use JTAG to debug it, it always works.
If the board has been let without power for 1+ hour, it will also boot up when I apply power to it(something to do with the battery perhaps).
Did you try disconnecting the battery ? Or replacing the battery with a new one ?
[...]
I tried disconnecting the battery and with the new SMC timing changes, I can boot up only one time after disconnecting the battery. I will go and get a new battery.
We had at91bootstrap 1.5 working well in 32-bit mode with u-boot 1.3.2. But I don't have the source.
See above.
I wonder what is the current stable version of u-boot and at91bootstrap?
AT91Bootstrap seems to be at v1.9.
As for U-Boot, try the latest git (which will soon be u-boot v1.3.4).
(BTW, if Wolfgang is listening, it would be nice to have proper tags in git as well as releases on the ftp server for the -pre version. Since the merge window is closed, I guess we're up to v1.3.4-pre1 now, but this version tag cannot be found anywhere...)
Stelian.
Dear all,
I am still struggling with this problem, but getting to reproduce this problem repeatably. Hoping folks in the forum can give me some pointers. Our Atmel FAE is clueless also.
This is repeatable, on two boards (both 9260 Rev A and Rev B): - bootstrap, u-boot, kernel, RFS all downloaded to NAND with SAM-BA 2.6 - Without recycle power, just push reset button; u-boot will come up. And I can continue to boot Kernel (2.6.25) - Push reset second time(and subsequent times), bootstrap will run and fail in the first few instruction of u-boot - Sometimes power cycle, u-boot will come up again
- When it runs, if I has u-boot boot kernel, linux will run. - Push reset button in Linux board will reboot and u-boot will run again (all the times) - If I use linux reboot, the board will stop after running bootstrap and (I believe) branched to u-boot.
Our h/ware guy has checked all the timing NAND, SDRAM etc and we have experimented with all the settings.
Thanks in advance for your help, Norman
Norman Cheung wrote:
Many thanks Stelian for the replies. My comments inline.
Norman
Stelian Pop wrote:
Le mercredi 25 juin 2008 à 01:36 +0000, Norman Cheung a écrit :
I am struggling with this problem for a week now and not making any progress. So any suggestions, ideas or pointer will be greatly appreciated.
[...]
Since the post earlier, I have narrowed the problem to the bootstrap during the copy from NAND to SDRAM, memories are corrupted. I think I have a marginal NAND chip. I played around the SMC cycle timing in bootstrap, I have a setting that I can boot u-boot with power cycle 6 out of 10 times. So I am convinced that it is not an u-boot issue, Thanks!
The board works with u-boot 1.3.2 and at91bootstrap(1.5 I only have the binary) booting from NAND.
The source (for the latest version) is freely available on the Atmel site: http://www.atmel.com/dyn/products/tools_card.asp?tool_id=4093
If you really need the source for an older version, I'm sure the Atmel support will send it to you upon request.
U-boot fails intermittently at the start. When it fails, it didn't make it to my first printf statement. When I use JTAG to debug it, it always works. If the board has been let without power for 1+ hour, it will also boot up when I apply power to it(something to do with the battery perhaps).
Did you try disconnecting the battery ? Or replacing the battery with a new one ?
[...]
I tried disconnecting the battery and with the new SMC timing changes, I can boot up only one time after disconnecting the battery. I will go and get a new battery.
We had at91bootstrap 1.5 working well in 32-bit mode with u-boot 1.3.2. But I don't have the source.
See above.
I wonder what is the current stable version of u-boot and at91bootstrap?
AT91Bootstrap seems to be at v1.9.
As for U-Boot, try the latest git (which will soon be u-boot v1.3.4).
(BTW, if Wolfgang is listening, it would be nice to have proper tags in git as well as releases on the ftp server for the -pre version. Since the merge window is closed, I guess we're up to v1.3.4-pre1 now, but this version tag cannot be found anywhere...)
Stelian.
Le vendredi 27 juin 2008 à 14:31 -0700, Norman Cheung a écrit :
Dear all,
Hi Norman,
I am still struggling with this problem, but getting to reproduce this problem repeatably. Hoping folks in the forum can give me some pointers. Our Atmel FAE is clueless also.
This is repeatable, on two boards (both 9260 Rev A and Rev B):
- bootstrap, u-boot, kernel, RFS all downloaded to NAND with SAM-BA 2.6
- Without recycle power, just push reset button; u-boot will come up.
And I can continue to boot Kernel (2.6.25)
- Push reset second time(and subsequent times), bootstrap will run and
fail in the first few instruction of u-boot
Sometimes power cycle, u-boot will come up again
When it runs, if I has u-boot boot kernel, linux will run.
Push reset button in Linux board will reboot and u-boot will run again
(all the times)
- If I use linux reboot, the board will stop after running bootstrap
and (I believe) branched to u-boot.
Our h/ware guy has checked all the timing NAND, SDRAM etc and we have experimented with all the settings.
It looks like you have a SDRAM initialisation problem. To confirm this (since you said that you cannot reproduce this problem with a JTAG), you need to modify the source code for AT91Bootstrap, and add a 'verify' pass (using a checksum for example) after the copy from NAND to SDRAM. This way you'll find out if the reading from NAND is going wrong, or the writing to SDRAM (I suspect the latter).
Next, you may encounter an issue with the reset sequence. I know that on some newer Atmel boards (AT91CAP9 and AT572D940HF) there is a problem when an external reset is performed while a SDRAM burst access is occuring. The workaround for this issue is to catch the external interrupt generated by the reset switch, and do the reset by software (see the AT572D940HF patches posted recently by Antonio Costa for the full code). I'm not sure the 9260 exhibits the same bug, but you may query your FAE for that.
Hope this gives some clues.
Stelian.
participants (5)
-
Andrejs Cainikovs -
Jean-Christophe PLAGNIOL-VILLARD -
Norman Cheung -
Norman Cheung -
Stelian Pop