Re: VIA Firewire still broken in 2.6.5 (also broken in 2.6.4, worked in 2.6.3)

[Alexandre Oliva <aoliva@redhat.com>]

On Apr 15, 2004, Felix von Leitner <felix-kernel@fefe.de> wrote:

> In 2.6.4, the kernel would detect my cold plugged Maxtor disk, but then
> freeze and eventually time out trying to mount the file system on it in
> the boot process.

> In 2.6.5, largely the same, but the kernel works when the disk is hot
> plugged in after the boot process.  Having it plugged in during the boot
> process still fails, and unplugging and replugging it after the boot
> process still fails.

> After a few write accesses on the file system, however, 2.6.5 panics.

> Is this a known problem?  Anyone working on it?

Yup.  There seems to have been some progress in this regard in the
linux1394 mailing lists and SVN repository, but it's still a bit
unstable.

Personally, I've been using the 2.6.3 ieee1394 sources in newer kernel
releases, and it's been working very well.  Perhaps we should revert
drivers/ieee1394 to the 2.6.3 state until the problems are addressed
in the linux1394 repository, to only then have it merged here.
What we have now was a merge of a non-functional state of the tree,
which servers nobody.

I can easily provide a patch to restore the working state, if people
think this would be a good idea.  Just say the word.

-- 
Alexandre Oliva             http://www.ic.unicamp.br/~oliva/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}