yeah because PLi is using 3.3.6
i sent 3.3.8 changes , we should wait
Posted 21 May 2013 - 13:19
yeah because PLi is using 3.3.6
i sent 3.3.8 changes , we should wait
Open Vision sources: https://github.com/OpenVisionE2
Posted 21 May 2013 - 18:08
nfidump is now available since >3 years and since then there was always a working version available.
I didn't start the misuse of nfidump to support clones, so they are now stuck with old binaries lacking all new functionality like lzo compression or ubifs support and should not complain about the consequences.
And yes, I'm fully aware that is is a simple =m in the kernel config file to build the nandsim kernel module which means everybody with an up-to-date OE can build it themselves.
That's the reason why I'm a little bit dissapointed that we now have to collect modules from dubious sources instead of getting properly built standard module from the Feed of the image.
But when I'm dissapointed I stop complaining, I start acting
Edited by gutemine, 21 May 2013 - 18:10.
Posted 21 May 2013 - 23:09
I would now be ready, but we still miss lots of nandsim modules for various boxes:
>>>>>>>>>>>> nfidump 0.10.0 by gutemine <<<<<<<<<<<<
based on code from tmbinc, noggie and others
extracting UBI image ...
extracting root image ...
extracting root with ubifs via nandsim ....
>>>>>>>>>>>> nfidump 0.10.0 finished <<<<<<<<<<<<<<<
On my dm8000 it now works nicely to extract root_cfe_auto.bin via nandsim module, hence I'm pretty sure that it will work also on your boxes.
Get the modules, then you get the testkit.
Edited by gutemine, 21 May 2013 - 23:11.
Posted 22 May 2013 - 14:38
We now started preliminary NP tests on the VU boxes, anybody having ET/CT box and already found a matching nadsim module and wants to get the preliminary kit for early bird testing, then write PM.
But you have to hurry, because if everything works as expected I'll publically post the testkit late this evening.
And I can only support 2-3 users via PM, so don't be dissapointed if you volunteer but have to wait until then.
Testinstructions are simple - install nandsim ipk so that modprobe nandsim works, then install nfidump 0.10.0 ipk and then copy some DMM Images and some extracted ZIP images for your and other boxes to a directory of your harddisk like /hdd/backup
Then it is a simple nfidump /hdd/backup/imagename.nfi /hdd/backup/extractdirectory or /hdd/backup/xxx/root_cfe_auto.bin /hdd/backup/extractdirectory in telnet and we see if it works.
If it doesn't work I need the telnet output.
Edited by gutemine, 22 May 2013 - 14:43.
Posted 22 May 2013 - 21:42
You should not only download, you shout try out and report back if it works as it should
If it fails I need the telnet output and maybe the dmesg output.
And don't report any FATAL: nandsim not found errors , then you didn't install the nandsim module so the modprobe nandsim failed ...
Edited by gutemine, 22 May 2013 - 21:43.
Posted 23 May 2013 - 00:47
some nandsim modules
more soon
Open Vision sources: https://github.com/OpenVisionE2
Posted 23 May 2013 - 04:46
Edited by gutemine, 23 May 2013 - 04:47.
Posted 23 May 2013 - 15:17
Edited by greatred, 23 May 2013 - 15:20.
Posted 23 May 2013 - 17:32
Thanks, so it looks like Uno s not using 512 subpages from the 2048 byte blocks.
Remember, I have to develop/test this on an dm8000, because it also has 2048k Blocks in Flash, BUT it uses subpages.
Is any of your boxes using subpagest at all, because if not, then I will simply disable it ?
But we now have almost 40 downloads, and so far I only know that it works on sol2 to and not (yet) on uno, any further inputs ?
Coould you try to unpack a solo2 image on the uno with nfidump 0.10.0 so that we see if it then cashes also ?
Could you try also unpacking an DMM image on the uno ?
Edited by gutemine, 23 May 2013 - 17:36.
Posted 23 May 2013 - 23:38
root@vuuno:/# nfidump /media/hdd/backup/plisolo2/vuplus/solo2/root_cfe_auto.bin /media/hdd/backup/extractplisolo2 >>>>>>>>>>>> nfidump 0.10.0 by gutemine <<<<<<<<<<<< based on code from tmbinc, noggie and others extracting UBI image ... extracting root image ... extracting root with ubifs via nandsim ...Segmentation fault root@vuuno:/# nfidump /media/hdd/backup/Merlin-3_OE-2.0-dm8000-20130427.nfi /media/hdd/backup/Merlindm8000 >>>>>>>>>>>> nfidump 0.10.0 by gutemine <<<<<<<<<<<< based on code from tmbinc, noggie and others extracting dm8000 image ... extracting boot image ... extracting boot with jffs2 ... extracting root image ... extracting root with ubifs via nandsim ...
On Merlin extract the process has hanged - a forced reboot was the only solution to get rid of the process.
It freezes on every 2/3 nfidump extraction tries.
dmesg attached.
Posted 23 May 2013 - 23:47
Could you try to add -v 2048 to the nfidump command, eg.
nfidump -v 2048 /hdd/backup/xxxx/root_auto_cfe.bin /hdd/backup/imagedir
This enforces the blocksize at the ubiattach, maybe it helps, becuaset he crahes in nandsim driver are always caused by wrong allignment.
At least you now know why I prefere block2mtd and spent so much time to make it run.
And uno is not that great on the memorysize - I run a permanent swapfile on all my boxes, maybe I will have to add at least a temporary one to the nfidump binary too.
Does the hanging/crash stop if you mannually add a swapfile ?
Edited by gutemine, 23 May 2013 - 23:51.
Posted 24 May 2013 - 07:24
root@vuuno:/# nfidump -v 2048 /media/hdd/backup/plisolo2/vuplus/solo2/root_cfe_auto.bin /media/hdd/backup/extractplisolo 2 Segmentation fault root@vuuno:/# nfidump -v 2048 /media/hdd/backup/Merlin-3_OE-2.0-dm8000-20130427.nfi /media/hdd/backup/Merlindm8000 Segmentation fault root@vuuno:/#
Immediately getting the segmentation fault error when launching with "-v 2048" attribute.
Posted 24 May 2013 - 21:30
Sorry for letting you wait, but I had a busy week.
Attached is a Version 0.10.1 of nfidump
Now the -v option for the vid header offset option should also work.
But normally this should not be needed
Please continue your tests with the attached version!
gutemine
Edited by gutemine, 24 May 2013 - 21:31.
Posted 25 May 2013 - 10:06
The segmentation fualt is not the point, the point is the reason why - if it is again due to wrong ubiattach commands then dmesg should tell the full story.
And I didn#t fix the segmentation fault, I just fixed the -v option so that we can try out if it is avoidable and with which parameter.
But I'm patient, if people don't want to have a feature and don't help to get it I will not push them.
Edited by gutemine, 25 May 2013 - 10:07.
0 members, 0 guests, 0 anonymous users