FreeBSD supports of EXT4 ?

ralphbsz since that old zombie of oracle ./. Google still lurches around in some court room, could it be that using the on-disk format alone would be enough to sumon-monster(lawyer) on your doorstep? APIs seem to be cancerous after all.
I don't know. In general, reverse-engineering undocumented interfaces (such as APIs to binary libraries, network protocols to services, and on-disk formats for data) is less risky than outright reading source code, if the goal of the reverse-engineering is to create interoperable software. But given that there have been several long and expensive lawsuits about reverse engineering (plus the infamous Linus Torvalds v. Andy Tridgell v. Larry McVoy spat, which at least didn't lead to lawsuits, at least none that I know of) shows that reverse engineering at the interface level is not without risk.
 
I think that it is important to FreeBSD to have EXT4.

Ext4 is definitely better than ZFS for most type of harddisk/systems (at least for Raspberry).

Ext4 would offer user to move from / to Linux and FreeBSD operating system.
User can have EXT4 FS and do their work either on BSD or on Linux.
Anyhow there is mostly only two choices of Unix, the real one or the Linux one.

It is good because it would give freedom to users.

It reminds me Microsoft or Google, that propose just services to get more users. The more you lock the user, the better.

what about freedom ?
https://en.wikipedia.org/wiki/Open-source_software
 
Ext4 is definitely better than ZFS for most type of harddisk/systems (at least for Raspberry).
Better? :) Please ... Does ext4 support data integrity with checksums? Nope. Does it support compression with various algorithms? Nope. Can you create redundant array like RAID10/RAID5/RAID6/RAIDO50/RAID60 with ext4? Nope. Can You speed up reads with additional cache on SSD with ext4? Nope.

The ext4 filesystem is YEARS behind ZFS.

Ext4 would offer user to move from / to Linux and FreeBSD operating system.
User can have EXT4 FS and do their work either on BSD or on Linux.
Anyhow there is mostly only two choices of Unix, the real one or the Linux one.
Linux is free to implement full FreeBSD' UFS support so there will not be a problem.

You can use ZFS for that NOW without any modifications ...
 
Hmm, as a user I rarely give a hoot about the politics of licenses, etc. But I do care about being able to use technology to get things done. In my case I have multiple platforms and need to share data between them.
My workstation is running TrueOS with ZFS mirrors for O/S and data. I dual and sometimes to triple boot to get certain things on different *nix platforms to get around h/w and s/w compatibility. Sharing the ZFS data mirror is not friction-less but using a separate drive such as a USB stick is. However it is slow. So that means using a drive with some modern file system that is well enough supported to reliably transfer files.
My only decision is which FS to use that is well supported and can easily be shared by different O/S's?
Using FAT means all files ends up having the execute flag set, so is not optimum. Frankly I've not had any issues using ext4 (under Linux) since it was implemented, and for sharing files I don't care what it is. UFS seem to be well supported on Linux and can easily be remounted. That might be a good way to go.
 
My only decision is which FS to use that is well supported and can easily be shared by different O/S's?
Here's a good solution, which is reliable and doesn't have compatibility problems: Don't use a file system on disk with multiple implementations in multiple OSes. Instead, connect your disk and file system to a single OS, and use it as a storage server, which exports the data using a networking protocol or two (typical candidates include NFS and CIFS). Keep that server running at all times. The reboot your clients into different OSes.

The bad news: It will cause performance problems, because all data will have to go over a network (more on that below).

In theory, it also requires a second machine, and a physical network, which may be a bit of a hassle. But that can be easily taken care of by using virtual machines: Run the "file server" on the real host, using real disks. Then boot the various operating systems you want to us in virtual machine clients, using the VM technology of your choice. That also gets rid of the speed impact of using a physical network: the virtual network within a single host is much faster, and with modern machines can run at full disk speed.
 
A very late reaction, but even so...

Using FAT means all files ends up having the execute flag set, so is not optimum.
It doesn't have to, look into mount_msdosfs(8). With -m you can set a mask for files within the filesystem whereas -M specifies a mask which should be applied to directories.

Code:
/dev/da0s1      /media          msdosfs rw,noauto,-u=peter,-g=wheel,-m=644,-M=755 0 0
This is an entry in my /etc/fstab which I'm using to get easier access to my memory sticks. As you can see only the owner (me) gets write access to files (and directories) whereas the rest has only read access. But no execution bits get set (except for directories of course).
 
I think it might be a delay in updating the documentation (isn't it always) but I'm in progress in an upgrade to 12.0-RELEASE and I saw that notification that ext4 write support has been added to the ext2fs however the documentation https://www.freebsd.org/cgi/man.cgi?query=ext2fs&sektion=5&manpath=FreeBSD+12.0-RELEASE+and+Ports or should that be ext2fs() is still dated something in 2016 and says that ext4 is read-only.

With regard to the licencing I come to the BSD vs GPL with a different view. I've been using GNU/Linux for many years and IMVHO the GPL has been a mechanism to stop BigEvilCorp from unfairly benefiting from the work of volunteers who have donated their time and effort into a pool of resources on the understanding that they can freely make use of the others who have done the same. As such it is one way of doing things but it would be foolish to say that that is the only way.

As a lead contributor to a F.O.S.S. project however the GPL has one big advantage to me in that you just say the code is under that and then you do not have to worry about the terms of any other library or code that you make use of that is also under that license. My project does make use of other libraries and pieces of code with other Free licenses however I am getting frustrated with all the individual copies of their licenses I have to keep track of (that are identical except for the authors' and product names) that I am having to make sure can be displayed in the About <Application> dialogue!

Also, it has been suggested that the "one file system that in the hardware shall bind all the others" :D is ZFS but due to the fact that it's Open Source Licence is not compatible with the GPL 2.0 Linux Kernel pretty much all Linux distribution cannot distribute binaries containing kernel drivers for it. Ubuntu has stuck their neck out but they are on thin ice I think! As a systemd-less (well not as PID 1 anyhow) Debian Linux user I do not have access to ZFS (or UFS) file-systems on my triple boot main PC from Linux and getting access to files from the FreeBSD side is only possible via a spare ext2 partition I keep for this purpose. With the previous read-only limitation on the ext4 it was equally hard to send files to the Linux side from FreeBSD.

Anyhow I hope that this improvement is as good as it sounds, because it will make it a bit easier for people who want to migrate to (or at least investigate the benefits of) FreeBSD from Linux before they get subsumed by creature feep from the hydra that is systemd.
 
Who, where, and how, would come after me if I redistribute forked GPL software under BSD license?

Has this ever happened?

What if the author of original software, initially licensed under GPL, forks his own project and issues say a BSD3 license. Who would be against this? Is the GNu organization after such violations?
 
Who, where, and how, would come after me if I redistribute forked GPL software under BSD license?

Has this ever happened?

What if the author of original software, initially licensed under GPL, forks his own project and issues say a BSD3 license. Who would be against this? Is the GNu organization after such violations?

What I could find so far:

The Free Software Foundation (FSF) and Software Freedom Conservancy (Conservancy) today lead worldwide efforts to ensure compliance with the GPL family of licenses.


FWIW.
 
The only trouble with UFS/UFS2 on Linux is that, I think it was dropped from Debian a couple of major version numbers ago. If I am wrong I hope someone will correct me.
😉
 
What if the author of original software, initially licensed under GPL, forks his own project and issues say a BSD3 license. Who would be against this? Is the GNu organization after such violations?
JWM window manager used to be under GPL, then the author changed its license to MIT. Of course, the author only applied this license change to what he fully authored, and not anything under another authorship.

For anything else, that can't and shouldn't be done. I wouldn't want a program which was mislead on about the license.

When GPL eats up and muddies MIT, ISC and BSD code, and has adapted to eat up newer versions of licenses, it is a nuisance. The same can be said of commercial code, which is all ironic. Still, applying or misleading a more free license on a more restrictive one shouldn't be tried.
 
Ironically, L. Torvalds requirement (and with some many thousand of contributors now it is probably almost impossible to change it even if he wanted to) that the GNU/Linux kernel is under GPL 2 only is hurting another OS that I have recently started to experiment with - the GNU/HURD - the small band of interpid developers there would love to use some stuff from Linux but because the Hurd is under GPL 3 or later it is impossible to use element from the Linux kernel... 😬
 
Back
Top