Replicant

Before, EFS backups used the following as part of the procedure:

adb shell "cat /dev/block/platform/*/by-name/EFS > /EFS.img" 

At some point, this started to create corrupted images.

There might be a way to fix the images.

In BackupResearch revision 15, doak complex tested the following as part of research to use adb pull in one command:

Be aware of the @head --bytes=-2@ which is needed to get rid of the nasty @0x0d 0x0a@ line ending returned by @adb@.
(Tested on Recovery of Replicant 6.0 0003.)

So the images that were backuped with the bad command have might just have the two extra bytes mentioned above.

• The two bytes are appended, so we could still try to parse the filesystem but that would be error prone.
• The last 2 bytes of a non corrupted backup might be the same than the ones we know are appended. Removing them would corrupt good backups.
• We don't know if the EFS partition size differ across various versions (16G, 32G) of a same device variant (GT-I9300). However we could try to find that with adb. If the backup is bigger than the partition it's supposed to be restored to, and that the two bytes are 0x0d and 0x0a, we could create a <file>.fixed for instance.