[answers inline.]On Oct 19, 2010, at 12:49 AM, Petr Štetiar wrote:
Bridges Seth <seth%40sethbridges.com> [2010-10-18 23:52:05]:
> * MTD partitions look like they came from a 7200...standard spew looks like this:
>
> Searching for NAND flash...
> NAND device: Manufacturer ID: 0xec, Chip ID: 0x75 (Samsung NAND 32MiB 3,3V 8-bit)
> Scanning device for bad blocks
> Hardware ECC detected
> Using MBR partition definition
> Creating 4 MTD partitions on "NAND 32MiB 3,3V 8-bit":
> 0x00000000-0x02000000 : "whole chip"
> 0x00004000-0x00204000 : "kernel"
> 0x00204000-0x00404000 : "initrd"
> 0x00404000-0x02000000 : "rootfs"
> yaffs: dev is 7939 name is "1f:03"
> yaffs: Attempting MTD mount on 31.3, "1f:03"
This is from different kernel, right?
Yes, I was just including the MTD startup spew from my 2.4.x kernel for reference. (more about this below)
> * serial port ttyTS0 doesn't show up...the standard spew from the
> 2.4.26-ts11 kernel looks like this:
I've built the kernel as modular as possible and added just the necessary
stuff as built-in into the kernel. For now, you would need to unpack the
module from modules-2.6.34-r3-ts72xx.tgz and to autoload it at startup just do
something like this:
$ echo 8250_ts_ser1 > /etc/modutils/8250_ts_ser1
$ update-modules
$ reboot
cool, I'll have to try that.
In the future it should be possible to just:
$ opkg update
$ opkg install kernel-module-8250_ts_ser1
> I'm not sure where to go from here (except to bed), but I wanted to let you
> know the results of what I'd tested so far.
I would add some debugging into the pppd/chat so you can see all the commands
etc. How do you have the modem connected? It's via ts_ser1?
I had the chat spew turned up, and it got to the point where it tried to send AT to the modem, but nothing ever responded. Just previously, I had been talking to the modem with microcom over the same port...
> <begin output spew from kernel boot>
Thanks, that's what I needed. Looks like it works fine :-) If find some spare
time, please can you try to debug the ntpclient issue? If it hangs in the
ntpclient binary I can prepare some verbose version for you.
I'll poke at the ntpclient binary more later....I did try running it from the command line with -h time.apple.com and it just sat there. I didn't do any more than that.
> Configuring leds:
> platform:grled: nand-disk
> platform:rdled: heartbeat
BTW, is there red led on the ts7400? If so, started blinking after this?
There is a red LED, but I didn't notice it.
Can you pls also try to flash the UBIFS image and try to boot from it? Thanks.
My problem so far is that I've never been able to boot to a NAND root with a kernel I've built. Your comment about the UART driver not being in the kernel finally made me think that the NAND flash driver was only in as a module.
As a test, I grabbed the 2.4.26-ts11 source, touched up the config to include the MTD_NAND_TS7250 driver in the kernel and booted it up. For the first time ever, it actually recognized the partitions correctly and looked like it was going to boot.
Then it started puking all kinds of ecc errors like this:
page 91 in gc has no object >>mtd ecc error fix performed on chunk 60:0 >>mtd ecc error unfixed on chunk 60:1 >>Block 1 marked for retirement page 92 in gc has no object >>mtd ecc error unfixed on chunk 61:0 >>mtd ecc error unfixed on chunk 61:1 >>Block 1 marked for retirement
After the ECC spew stopped, I got a tty login screen, but nothing would respond. Now my board won't boot at all. :( I think my failure was that I was supposed to have included the MTD_NAND_TS7400.
I guess it is a good thing that I ordered the TS-9441 boot recovery board yesterday. Too bad it won't be delivered until tomorrow.
At any rate, when I get actually figure out how to get a kernel that boots from NAND, I'm happy to try flashing the UBI image down to my board.
Thanks,
-sb
__._,_.___ ![]()
__,_._,___
|