Re: CVS commit: src/sys/dev/usb

[Hans Rosenfeld <hans%netbsd.org@localhost>]

On Tue, Mar 25, 2025 at 06:42:38PM +1100, matthew green wrote:
> hm.  does this really change anything?

You're right. I missed that the ums.c code is a bit different from
uts.c and doesn't suffer from the same panic-inducing problem.

> i've had a couple of crashes near here recently and i haven't been
> able to understand them (or reproduce easily, with USBHIST), so i
> was thinking this might help me, but unless i've missed something
> above, this doesn't actually change behaviour.  the issue in uts.c
> does appear to have been a real problem, perhaps because it didn't
> have the 2nd if() that ums_disable() currently does.

Yes, uts_disable() was causing kern/59206 because of this. A friend of
mine wanted to try 10.1 on his slightly outdated Thinkpad an ran into
this immediately, which is why I tried to find the issue and fix it for
them.

> (note: the second if() could be removed now, since it will always
> be true if the code gets to this point.  oh, which makes it like
> the uts.c version.)

Will do.

Hans


-- 
%SYSTEM-F-ANARCHISM, The operating system has been overthrown