Very puzzled, your entire mtd0 is bad. I'm looking for this section: [ 5.197563] Bad eraseblock 0 at 0x000000000000 [ 5.202061] Bad eraseblock 1 at 0x000000020000 [ 5.206560] Bad eraseblock 2 at 0x000000040000 [ 5.211114] Bad eraseblock 3 at 0x000000060000 [ 5.215611] Bad eraseblock 4 at 0x000000080000 [ 5.220125] Bad eraseblock 5 at 0x0000000a0000 [ 5.2246by bodhi - uBoot
sigma Wrote: ------------------------------------------------------- > I cant kwboot the uart.kwb, at 94% is always an > xmodem error, somehow i thought the watchdog is > disabled with this new uboot. Yes, the watchdog is disabled when u-boot starts. But UART booting is before that! so the binary must be loaded completely and then u-boot starts. That was the reason we've beeby bodhi - uBoot
InfoScav, Ah! I thought you want to use USB so that you can unplug it easily. However, you can do the same with HDD label using the mount command, instead of pmount. The reason I prefer partition label over the UUID: it is more flexible, you can even swap disks as long as the label is used for mounting, while UUID is fixed to that physical disk. And I don't put the entry in fstab becauby bodhi - Debian
@Davide, QuoteAs I understood from your reply to sigma, new Uboot has different location for its env and won't touch the stock one. But, since it's new uboot, I may as well create a new more powerful environment (scripting), correct? Also, did you include EFI partition support? Yes, it has all the good stuff: EFI, Hush shell scripting, boot with Ext4, >2TB HDD … I mentioby bodhi - Debian
InfoScav, The simplest solution is using pmount. Always label the HDD to something meaningful, e.g. MUSIC. pmount will take care of the mounting the disk to /media. So even if you don't want to automount (this is another subject), pmount is going to work well for it. That disk partition is always going to be /media/MUSIC regardless which Linux box you plug it in. pmount /dev/sda1 /mby bodhi - Debian
@jfharper, QuoteMy assumption is I don't have a filesystem on the GF base, so the red blinking occurs...is it looking fo boot files and cannot find them? No, the blinking red in this case means your envs are not define properly (due to bad blocks). The end result is, yes, u-boot could not find the boot files, but likely not because they are not there, it probably just searching in theby bodhi - uBoot
@jfharper, You're lucky the flashing did not mess up mtd0 too badly:) you got 4 bad blocks on mtd0 !!! There is no space left for envs. For now, you can boot into Arch, this is the best you can hope for. Don't try anything else, it might brick it for good :) Let me think about it a bit and get back to you.by bodhi - uBoot
sigma Wrote: ------------------------------------------------------- > i flashed uboot.2013.10-tld-1.nsa325.mtd0.kwb on > nsa325 with 3.16.0-tld-2. Now the new uboot is > coming up, but it dont read the envs from mtd1, > right? > can i copy the envs from mtd1 or have i to kwboot > with uart? Cool! right, the new u-boot uses diferent location: 0xC0000. So you can cut/paby bodhi - uBoot
@davidedg, Have you tried the test U-Boot? I'm running NAND version and it has been very stable. @WarheadsSE BTW, I have tried to see if I can chainload stock U-Boot from the new U-Boot. I could not get it to load, it just hang. It seems like stock U-boot is old, but not old enough to do this successfully. But there is a path to stock after you've installed new U-Boot: we can UARTby bodhi - Debian
jfharper, Have you rebooted? or still in Arch? > flash_erase /dev/mtd0 0 4 > nandwrite /dev/mtd0 > uboot-mtd0-orig-goflexhome.kwb If your above flashing of the original U-Boot was successful, then you're back to stock upon reboot. What the ouput of these comands look like? you can post the log here. Quote> Is the above uBoot the one I got from Jeffs > tutorial?by bodhi - uBoot
InfoScav Wrote: ------------------------------------------------------- > thanks a lot bodhi for that! clears it up well for > me. option no. 4 it is for me then! > > and i just took the risk & flashed the new uBoot > too successfully! :) Excellent!by bodhi - Debian
sigma, This bootcmd should be: bootcmd='run bootcmd_linux' And to_stock and to_linux should be: to_stock=setenv mainlineLinux no; setenv bootcmd \'run bootcmd_stock\'; saveenv; reset to_linux=setenv mainlineLinux yes; setenv bootcmd \'run bootcmd_linux\'; saveenv; reset BTW, you can post in this NSA325 thread to get help quicker: http://forum.doozan.comby bodhi - Debian
@sigma, Your envs look odd: to_stock=setenv mainlineLinux no; setenv bootcmd \'run bootcmd_stock'\; saveenv; reset to_linux=setenv mainlineLinux yes; setenv bootcmd \'run bootcmd_linux'\; saveenv; reset bootcmd='run bootcmd_linux\; saveenv; reset Have you tried to set it back to stock and back to linux again? in serial console: run to_stock run to_linuxby bodhi - Debian
sigma, Try this: fw_setenv machid 118fby bodhi - Debian
InfoScav, If the main use is general purpose and perhaps involved with compiling sources code then I would use the HDD rootfs. OTOH, I would choose option 4 for media-storage & occasional downloads. All my media NAS servers boot from USB flash, this has many advantages over a single HDD drive: easier to backup rootfs image, portable HDD drive that you can unplug and carry. And I have a sinby bodhi - Debian
sigma Wrote: ------------------------------------------------------- > i got this error with 3.16.0-kirkwood-tld-2: > > Error: unrecognized/unsupported machine ID (r1 = > 0x0000020f). > ... > ... > Please check your kernel config and/or > bootloader. > > installed the new kernel over > 3.16.0-kirkwood-tld-1 Which box did you install this kernel onby bodhi - Debian
Both 3.16.0-kirkwood-tld-2 kernel and rootfs have been uploaded. Please see 1st post for download links.by bodhi - Debian
Me neither. I just thought about it because I came across the mcc module in stock kernel while browsing around. This module reads the temperature (pull down and pull up those 3 GPIO pins).by bodhi - Debian
InfoScav, > that! it makes sense that perhaps may be that's > why the HDD is not showing up. don't know. Not likely the reason. The GFNet u-boot works fine for GFHome. It's only different in other minor thing specific to GFHome such as LEDs. > > i tried interrupting the boot-process when it > gives the prompt 'hit any key to stop autoboot', &gby bodhi - Debian
InfoScav, Yeah, I forgot that you have the GoFlex Home! only one available SATA port (there are 2 supported by the SoC). But it should work with the U-Boot that you are running without problem. You are actually running the GoFlex Net version, not the GFHome version (work the same for booting, but not exactly support GH Home specifics in Debian kernel). You could interrupt netconsole and tryby bodhi - Debian
InfoScav, > UPDATE: attaching the boot flash-drive to > the USB-hub not working. the hub is anyway a > non-powered one. can't fathom why attaching the > HDD to the SATA port is not working. USB flash boot drive should work with the hub (it does not matter power or not). Look at netconsole output for the USB drive detection, see if anything strange during boot. [ 2by bodhi - Debian
@WarheadsSE, Do you have any plan to implement the NSA325 MCU status readout in the kernel? perhaps a module?by bodhi - Debian
mxhdrm, Did you keep the log of your attempt to flash U-boot? It's always good to include that, too. Your bad block is right within the first 1M of NAND, and to be precise it is in first 512K, which is where U-Boot is. The instruction in the blog mentioned above, with both the flash_erase and nandwrite commands must have failed, resulted in a bricked Pogoplug. So don't shutdown orby bodhi - uBoot
davidedg, It is defined in mtd nands part: cat arch/arm/mach-kirkwood/nsa325-setup.c.orig | grep -5 MTD static struct mtd_partition nsa325_nand_parts[] = { { .name = "uboot", .offset = 0, .size = SZ_1M, .mask_flags = MTD_WRITEABLE }, { This write flag was masked, and must be removed to allow wby bodhi - uBoot
sunfire, > When will 3.16.0-tld-2 be released? Not sure when exactly, could be a next few days. > I need a new u-boot, because the watchdog does not > support a cold start with debian. Boot with a fast USB thumb drive would help a little. But if the HDD is needed to be fsck then it would be too slow to survive the watchdog. > I hope i can fix it with a new submarine :)by bodhi - uBoot
# By default this script does nothing. ethtool -K eth0 tso off # turn on LEDby bodhi - Debian
jst818 Wrote: ------------------------------------------------------- > Do i need to enter this command every time debian > arm device is restarted ? Yes. So you can add it to /etc/rc.local so that it will executed during kernel start up.by bodhi - Debian
InfoScav, > nc -lup 6666 192.168.1.50 6666 > nc: cannot assign requested address nc has many variations on different platforms, so check the man page on Ubuntu to see if the command options are the same as on Debian wheezy. You might need to change the command above.by bodhi - Debian
sunfire, > root@nsa325:/home# dmesg | grep -i 'bad' > [ 21.769430] Scanning device for bad blocks > [ 21.779256] Bad eraseblock 100 at > 0x000000c80000 > [ 21.789546] Bad eraseblock 200 at > 0x000001900000 > These bad blocks are OK. They are far in other regions, not in first 1M. > 2. flash_erase u-boot in debian as root > > root@by bodhi - uBoot
zzbao, > are these bad blocks in the first 1M? Can I flash > the new uboot? Block 67 and the rest, they are way up in other area in NAND. Please be extremely careful, because the Pogo E02 does not have UART booting. If you have a typo in your commands, you will have a brick :) and the only way to recover is JTAG.by bodhi - uBoot