This is a read-only implementation of ext4 for FUSE. The main reason this exists is to be able to read linux partitions from OSX. However, it should work on top of any FUSE implementation. Linux and FreeBSD have been tested to some point and I've heard that OpenSolaris should also work.
Write support will only come if I get the time, knowledge, patience and nerve to support it. Most of them I lack, so it's a long shot. However, the fact that ext4fuse is read-only also means that it's completely safe to use.
##Installation
If you use OS X I suggest you rely on the homebrew project.
Once you have homebrew installed, simply type the following three commands:
$ brew tap homebrew/fuse
$ brew install Caskroom/cask/osxfuse
$ brew install ext4fuse
At least on Leopard, you need to add your user to the operator group so you can have readonly permissions to the disks. Use this:
$ sudo dscl . append /Groups/operator GroupMembership <your-user>
Also, you will need to know the name of your ext4 partition. Take a
look at the Mac Disk Utility. It should be something like /dev/disk0s5.
Simply install it through the ports tree:
$ cd /usr/ports/sysutils/fusefs-ext4fuse && make install clean
Remember that you need the fuse module loaded. In my experience it doesn't load automatically, but then again, I have nearly zero experience with FreeBSD.
If you prefer bleeding edge, get the source, untar it and compile using:
$ make
or in case you are on FreeBSD:
$ gmake
You need to have pkg-config for the compilation to work as well as the FUSE
kernel module. For OSX you should use fuse4x (notice that fuse4x is also
available via brew install).
You can mount a filesystem like this:
$ ext4fuse <device> <mountpoint>
If you compiled from source, and you haven't manually installed ext4fuse in
your $PATH, go to the directory where you did the compilation and run this
$ ./ext4fuse <device> <mountpoint>
The should be the partition device and the is the directory where you want to mount your partition.
If you notice a problem, please file a bug report.
If you have a reproducible problem the easiest for debugging is to share the filesystem. First of all, umount the partition, then you can create a backup like this:
$ dd if=<device> bs=64K | gzip -c > filesystem.backup.gz
Then, just upload the .gz file somewhere.
However, I understand that you generally do not want to do that. In that case you can also generate a log file. Notice that the log file still contains the directory listings.
To get a logfile, you can run ext4fuse like this:
$ ext4fuse <device> <mountpoint> -o logfile=/dev/stdout
If you do not want to share the logfile, another option is to provide a backtrace with gdb or a coredump (a coredump might contain file data).
Finally, you can always drop a mail:
- All code is religiously Little Endian only. If you don't know what this means, you are probably OK (ie, you are using an intel or amd cpu). The code should be better tested on x86-64, you should not be using anything else on modern hardware anyway.
- Block numbers over 32 bits aren't supported. You hit those when you reach around the terabyte, and I don't have any way to test that. It should be quite easy to fix, but I don't feel like spending time on something that neither has a use for me or can be proved to be correct. I don't have such big disks :P.