At 2004-10-22T18:48:34+1300, Volker Kuhlmann wrote:
> So running umount is safe then, as I initially said. Which Joe User
> gives a *(&@*# whether it's because of the umount program or the umount
> system call?
Sure--it's safe with the potential exception of the write-back caching
issue. I never said that unmounting the filesystem was not safe--it was
you who made bogus claims about the safety of the unmounting filesystems
using the Linux FAT driver.
What I was doing was explaining how things work for people who are
interested.
> I almost certainly would have given it a few seconds. As I said, under
> the same circumstances Linux filesystems have never shown that problem.
You didn't answer my question. Clean shut down, or hard power off?
Waiting "a few seconds" is meaningless--the drive can potentially defer
flushing writes in the cache for as long as it likes if it's doesn't
have a reason to flush. Plus, if it's a drive the kernel is unable to
flush the cache for, there's not much you can do other than pray.
I should know better than to respond to the rest of your post, but I'm
bored...
> Yes I know you think I don't have a clue, however I've had
Not true. You're quite capable, but that doesn't reduce the fact that
you tend to rant and make incorrect claims about topics that you are not
an expert in.
> acknowledgement from a kernel developer that the bugs you refer to "in
> other areas" do in fact exist.
What are you referring to? The handling of DMA on IDE devices? There's
a difference between an acknowledgement that some random bugs exist, and
some random "kernel developer" confirming that The World According To
Volker is correct.
> I trust my observations more than your opinion of my skills.
Your assumptions about the safety of unmounting FAT filesystems under
Linux, based on your observations, are wrong. You've been corrected.
It's not a big deal.
> I also seem to recall that the smartmontools author thought what I
> said about disk SMART was essentially correct.
So you might have been right about something, maybe? His message was
fairly vague, unfortunately, and he wasn't able to clarify any of the
technical questions I posed. Besides, the fact that you might be right
about that (which I don't believe that you totally are) doesn't make you
right all the time. Should others on the list, myself included, avoid
posting explanations about how they understand things work, or how
things actually work (after having read the source to understand it) to
avoid your wrath?
I don't see what your statement adds to this discussion, other than a
bit of puffery.
> Perhaps it's you who has his nose a bit too high up?
If you get come down off of your high chair, I won't have to tilt my
head so far back to see you.
Cheers,
-mjg
--
Matthew Gregan |/
/| [EMAIL PROTECTED]
