All PLX based chips - Important Upgrade Note [updated]

[slycat]

I was fearful of reflashing nand so I commented those lines out, I only meant to reiterate that i used uImage.pci, curious if that could've given me a problem.

Im not sure I understand what you mean by subing /dev/sdX1 for NAND. When in chroot jail, I am running from said /dev/sdX1 which is considered root (but I do understand the "mount -o bind /dev dev/" call prior to entering chroot). Should I change the ##Flash Kernel calls to use /dev/sdX1 instead of /dev/mtd1? It doesn't seem correct.

Also, zeroing the drive would erase my backup. If I were to do this, I would just start from scratch, but that is what I want to avoid.

[Geoff]

$this->bbcode_second_pass_quote('slycat', 'S')hould I change the ##Flash Kernel calls to use /dev/sdX1 instead of /dev/mtd1? It doesn't seem correct.

[slycat]

Thanks for clearing that up, Geoff. It was partly my mistake to not say what the setup was. I didn't do the whole sata boot with a kernel on the drive, just boot from PPro NAND, catches first partition of sata as rootfs, and runs from there.

By having trouble booting from the older spare rootfs, I am positive the nand flash was handled correctly even from the chroot jail. I checked what pogoman said about there being random, tty*, mtd* and sdX* folders in /dev/ and they all were there. I'm guessing there is something else going on in my rootfs that's not allowing it to boot.
It seems I am the one that failed and not the update script (though I never had any real doubt in WHSE's unix-fu). There must be some other baddies I updated to. Gonna check the usual suspects of linux, mkinitcpio, and filesystem.

[WarheadsSE]

You need to be sure you have the lib tree in place, to match the kernel in nand. They are not interchangeable.

[slycat]

After going in circles I got the plug to *begin* booting.
Then from everything.log
$this->bbcode_second_pass_code('', 'Jan 1 00:00:27 AtomAnt dhcpcd[675]: version 5.5.6 starting
Jan 1 00:00:27 AtomAnt dhcpcd[675]: eth0: interface not found or invalid')
So used rc.conf configuration for network from latest rootfs. No eth0 errror, but no dhcpcd start up either (which IS called in the latest rootfs).
Transferred latest set-oxnas-mac, redone update-oxnas (for depmod purposes), doubled network: still no connectivity, but solid green led, no dhcpcd call. It does continue to boot as I notice it calls to add the swap drive (which I added to rc.local). So I added "dhcpcd eth0" and still the same outcome and still no dhcpcd.

Should anyone have any ideas what to try next please let me know.

EDIT:
With all the time I spent trying to cover my old rootfs, I was able to rebuild from the latest rootfs transferring all my special bits within a single evening. Word of advice to anyone else who gets fubar'd like me: just start with what works.

[WarheadsSE]

jmandawg: I split the thread!

[icewewe]

Gah, I see I missed the module request period. The new kernel is missing the XFS filesystem module

[rampage]

I saw modules and firmwares from Linux 3.1 under lib folder, are they in use after apply this upgrade fix?

[WarheadsSE]

I have to answer this so often..


NO.

[Cybertimber2009]

If I haven't done a pacman -Syu in awhile (according to logs, since 2012-05-19), do I need to do that first? Or run this script and then update?

Edit: ok... based on the third post, part of the updater script perform a pacman update. I'm thinking it would be ok to run now, but i'd appreciate a confirmation. Sorry for the stupid question warhead.