Bugzilla Bug 160333

title says it all. Attacching the patch. Not the best solution but it:
- works
- does not harm
- is very useful

doslowusb option is automagically triggered only when really needed.

Created an attachment (id=105583) [edit]
smarter "slowusb" triggering

No.

This breaks support on the IBM bladecenter, at least.

hey, it's just an enhancement, it does not break anything! doslowusb or
--slowusb is still usable as before.

btw, what a wonderful way of trying to improve this "slowusb" hack... why does
it break? is the IBM bladecenter the only problem? I think that it could be
fixed in the same way too... What a lazy answer :-(

as usual... an incomplete answer... I don't usually expect this kind of
feedback from a developer :S

http://sources.gentoo.org/viewcvs.py/genkernel/trunk/generic/initrd.scripts?rev=486&view=log

so, fixed instead?

No.  I've had some issues with the patch, especially on slower machines, as it
causes some serious delays, which I am trying to work out.

but it's in the SVN, so the resolution of this bug is wrong (WONTFIX).
maybe that the "serious" delay is caused by the "sleep 10" line...? That line
was always triggered, even in the past, appending "doslowusb" boot option.
I still can't understand the reasons of what you wrote here, because you added
that to the source code anyway... and it's there since _pre2 AFAIK.

Ok, so reopening.

This caused a huge number of complaints on the 2007.0 release, so I am
reverting this patch.

true, because your patch contained a small bug. But, that's your problem.