[robbak]

This is an interesting issue.

For many years, it was considered to be unavoidable: After all, if an attached hard disk went awol mid-session, it seemed like panicing was the only sane thing for the OS to do.

Then, with removable USB devices, this started happening more often, but still most developers considered a panic to be the correct response, as it always had been in the past.

One of the developers of PC-BSD wondered why this should be, went digging for the actual cause(s), and found only relatively simple, fixable bugs in the pathways (seldom used in fixed servers!) that only occurred when a filesystem device was removed.

So now the FreeBSD developers have made similar fixes, I see. Nice to know!

But still, umount your filesystems before removing devices: Kernel Panics were only half the story - filesystem corruption is still as much of an issue as it ever was. Still, I would like to be able to put mount/umount code in devfs.rules, and have it just work(TM)!