Re: Can bad NAND blocks cause USB's "device descriptor read/64, error -110" ?
Yhaw! It was enough today just to flash this: http://downloads.openmoko.org/releases/Om2008.8-update/Om2008.8-gta02-20080903.uImage.bin and cdc_ether immediately gave out usb0 (before to even get the usb `new full speed USB device using uhci_h' line, I had to suspend and then wake up. So, it was the kernel's problem (and for some reason the 20080826 did not work (bad image) on my FR DateCode: 20080725). Case close, into the new explorations! ;) and thanks for everyone's care, -- sledge Andy Green rašė: > -BEGIN PGP SIGNED MESSAGE- > Hash: SHA1 > > Somebody in the thread at some point said: > | Hello dear all, > | > | My new FR (DateCode 20080725) is suspected to having bad NAND blocks: > | It had been suffering kernel panics etc. after as soon as I reflashed it > | with Om2008.8 on 20080831. Subsequent reflashings rendered my FR less > | unstable. > > Copy of the panic backtrace would be interesting if you can get ahold of > it. Usually every panic is a story although there's no guarantee we can > figure it out. > > The kernel image itself is CRC-checked by U-Boot before it is run. So > it is doubtful we can say that kernel is corrupted by "bad NAND". > > | I'm still suffering with USB cdc_ether, I'm getting `device descriptor > | read/64, error -110' and no solution found out there on the net helps. > > This can be to do with the host PC USB device, cables, stack as well as > Freerunner. Are there any other exotic variations of that in your host > PC dmesg or is it only that? > > | Last time tried with: > | u-boot,rootfs (20080902) -- from Openmoko_update page > | kernel (20080902, too) -- from daily page. > | (!)With the 20080826 kernel from releases/_update page says wrong image > | format whatsoever.. (and still does if I try to revert to the 20080826 > one) > > That's a funny error to get, can it actually DFU anything reliably? > That is also using the dodgy USB connection of course... > > - -Andy > > -BEGIN PGP SIGNATURE- > Version: GnuPG v1.4.9 (GNU/Linux) > Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org > > iEYEARECAAYFAki98T4ACgkQOjLpvpq7dMpJHQCfW5H0o85lQdh8H1siLHev4Ml1 > Jd0An0sDcAqCg3O6Xxam5xmp27ADEztI > =Wnwd > -END PGP SIGNATURE- > > ___ > support mailing list > support@lists.openmoko.org > https://lists.openmoko.org/mailman/listinfo/support > > ___ support mailing list support@lists.openmoko.org https://lists.openmoko.org/mailman/listinfo/support
Re: Can bad NAND blocks cause USB's "device descriptor read/64, error -110" ?
Thanks for answering, Andy, Andy Green rašė: > Copy of the panic backtrace would be interesting if you can get ahold of > it. Usually every panic is a story although there's no guarantee we can > figure it out. > Kernel panic just came out once (when I first launched the `Locations') , then I reflashed the kernel (as was suggested in #openmoko because of maybe bad blocks) and no other panics I've experienced so far. I have no debug board. Are there any other ways [logs] of obtaining that late kernel panic? > | I'm still suffering with USB cdc_ether, I'm getting `device descriptor > | read/64, error -110' and no solution found out there on the net helps. > > This can be to do with the host PC USB device, cables, stack as well as > Freerunner. Are there any other exotic variations of that in your host > PC dmesg or is it only that? > Well, are you sure? Because I get cdc_usb without problems (/dev/ttyACM0 created OK). Is cdc_ether protocol so much different [complex] that it is being affected by cable or other hardware (or FR's hardware-based fault?), whilst I get a normal cdc_ether behaviour. Maybe it's the u-boot bug, killing off usb completely? -> http://n2.nabble.com/Om-2008.8%3A-usb%3A-unable-to-enumerate-USB-device-tp680918p719848.html But I find no other [older/testing?] u-boot to try out with. Could you guys recommend any flavours? Yesterday I have reflashed kernel from 20080901 to 20080902 and, after pressing around power (with buggy Om2008's suspend) got this: usb 6-1: device descriptor read/64, error -84 But that exotic variation was only once. After reboot it swayed back to -110. > | (!)With the 20080826 kernel from releases/_update page says wrong image > | format whatsoever.. (and still does if I try to revert to the 20080826 > one) > > That's a funny error to get, can it actually DFU anything reliably? > That is also using the dodgy USB connection of course... > The exact error is (with NAND and NOR u-booter): 2097152 bytes read:OK Wrong image format for bootm command ERROR: can't get kernel image! And it happens always (even yesterday) with http://downloads.openmoko.org/releases/Om2008.8-update/Om2008.8-gta02-20080826.uImage.bin So we cannot blame it on dfu (unless it is kernel-version sensitive :)) and dodgy connection, as any newer (testing) kernels and roofss get always flashed ok. Should I try older kernel? Let's make my FR happen, -- sledge ___ support mailing list support@lists.openmoko.org https://lists.openmoko.org/mailman/listinfo/support
Re: Can bad NAND blocks cause USB's "device descriptor read/64, error -110" ?
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Somebody in the thread at some point said: | Hello dear all, | | My new FR (DateCode 20080725) is suspected to having bad NAND blocks: | It had been suffering kernel panics etc. after as soon as I reflashed it | with Om2008.8 on 20080831. Subsequent reflashings rendered my FR less | unstable. Copy of the panic backtrace would be interesting if you can get ahold of it. Usually every panic is a story although there's no guarantee we can figure it out. The kernel image itself is CRC-checked by U-Boot before it is run. So it is doubtful we can say that kernel is corrupted by "bad NAND". | I'm still suffering with USB cdc_ether, I'm getting `device descriptor | read/64, error -110' and no solution found out there on the net helps. This can be to do with the host PC USB device, cables, stack as well as Freerunner. Are there any other exotic variations of that in your host PC dmesg or is it only that? | Last time tried with: | u-boot,rootfs (20080902) -- from Openmoko_update page | kernel (20080902, too) -- from daily page. | (!)With the 20080826 kernel from releases/_update page says wrong image | format whatsoever.. (and still does if I try to revert to the 20080826 one) That's a funny error to get, can it actually DFU anything reliably? That is also using the dodgy USB connection of course... - -Andy -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.9 (GNU/Linux) Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org iEYEARECAAYFAki98T4ACgkQOjLpvpq7dMpJHQCfW5H0o85lQdh8H1siLHev4Ml1 Jd0An0sDcAqCg3O6Xxam5xmp27ADEztI =Wnwd -END PGP SIGNATURE- ___ support mailing list support@lists.openmoko.org https://lists.openmoko.org/mailman/listinfo/support
Can bad NAND blocks cause USB's "device descriptor read/64, error -110" ?
Hello dear all, My new FR (DateCode 20080725) is suspected to having bad NAND blocks: It had been suffering kernel panics etc. after as soon as I reflashed it with Om2008.8 on 20080831. Subsequent reflashings rendered my FR less unstable. I'm still suffering with USB cdc_ether, I'm getting `device descriptor read/64, error -110' and no solution found out there on the net helps. Last time tried with: u-boot,rootfs (20080902) -- from Openmoko_update page kernel (20080902, too) -- from daily page. (!)With the 20080826 kernel from releases/_update page says wrong image format whatsoever.. (and still does if I try to revert to the 20080826 one) Tried everything I could google, no cure: * I have usbnet, cdc_ether compiled as modules. uhci_hcd, too, -- modprobe -r uhci_hcd (re-insert) does not help; * http://wiki.openmoko.org/wiki/Uboot#Troubleshooting - /dev/ttyACM0 gets created OK, from NAND and from NOR * I use the USB cable supplied in my packaging, even tried the other one. * Could not try out http://n2.nabble.com/Om-2008.8%3A-usb%3A-unable-to-enumerate-USB-device-tp680918p719848.html, because I don't know where to get the u-boot'er any more fresh than I have (1.3.2, Aug 26) Thanks for helping pushing this through towards an initially-usable FR; and keep up the great good work, lads! -- sledge -- View this message in context: http://n2.nabble.com/Can-bad-NAND-blocks-cause-USB%27s-%22device-descriptor-read-64%2C-error--110%22---tp835094p835094.html Sent from the Openmoko Support mailing list archive at Nabble.com. ___ support mailing list support@lists.openmoko.org https://lists.openmoko.org/mailman/listinfo/support