given fs dies if SMP, if PREEMPT, if open two or more

given fs dies if SMP, if PREEMPT, if open two or more

[Pat LaVarre]

You may have noticed, I recently arrived here in linux-fsdevel,
committed to trying to help make linux udf.ko ever more perfect.

Can anyone here advise me specifically what to try next?

So far we've demoed some trouble in the linux-fsdevel thread "zeroes
read back more often than appended".  That trouble occurs if CONFIG_SMP,
if CONFIG_PREEMPT, or if I append nonzero bytes to multiple files in a
round robin sequence.

That says to me something in udf.ko doesn't know how to reenter itself,
right?  So my next step is to try reading while writing, to see if that
breaks also, right?  No?

What do you think?

Once my random guesses of specific kinds of trivial stress cease to
cause trouble, I figure I will then try the more complex:

http://www.codemonkey.org.uk/cruft/
http://www.zip.com.au/~akpm/linux/patches/stuff/
http://ltp.sourceforge.net/
http://www.iozone.org
http://www.google.com/search?q=postmark+linux
http://www.connectathon.org/
http://www.google.com/search?q=bonnie+linux

I mean to be asking what do you think I should try before risking the
possibly less discovered troubles per hour of that alternative.

What do you think?

Pat LaVarre

Re: given fs dies if SMP, if PREEMPT, if open two or more

[Pat LaVarre]

> So far we've demoed some trouble in the
> linux-fsdevel thread "zeroes read back more
> often than appended".  That trouble occurs if
> CONFIG_SMP, if CONFIG_PREEMPT, or if I append
> nonzero bytes to multiple files in a round
> robin sequence.
>
> That says to me something in udf.ko doesn't
> know how to reenter itself, right?  So my next
> step is to try reading while writing, to see
> if that breaks also, right?

I remember eventually I did try that, but in fact writing while reading
did Not break as readily as writing while writing.

Pat LaVarre