ext3 Bug in 2.4.20

[kwall@kurtwerks.com]

On Tue, Dec 03, 2002 at 05:59:25PM -0800, Net Llama! wrote:
> On 12/03/02 16:31, Jerry McBride wrote:
> >
> >That said, I've had no problems using it on either laptops, desktops or 
> >servers.
> >The one big selling point is... you can always go back to ext2 when ever 
> >you
> >wish, with nothing more than a single command. Try doing that with reiser, 
> >xfs,
> >jffs, etc...
> 
> I think the question is why i'd want to go back to ext2?  I've never had 
> any need nor desire to revert to any other filesystem since using XFS. 
> the fact that you'd even have a need to do it with ext3 speaks volumes 
> about its lack of maturity.

Straw man. The capability to revert is a by-product of being able to
convert (almost) seamlessly from ext2 to ext3. Which, I might add, you
can't do with other filesystems. ext3 is a compatibility solution. To
the degree that it propagates ext2's shortcomings, ext3 shares whatever
immaturity ext2 had.

Oh, and speaking of data corruption problems, XFS has them: 1.2 
corrupts pages when the FS block size is less than the page size, 
because the buffer list would end up having multiple list heads.
This is not in itself a problem but if a page is scribbled on via
mmap() and via buffered writes from system calls (say, write()),
the pages would not be written out. 

Kurt
-- 
Fertility is hereditary.  If your parents didn't have any children,
neither will you.