[Nut-upsuser] usbhid-ups looping and unresponsive after update

Charles Lepple clepple at gmail.com
Thu Feb 7 04:03:13 UTC 2008


On Feb 6, 2008 9:55 PM, John Darrah <xyllyx at gmail.com> wrote:
> On Wed, Feb 06, 2008 at 09:29:47AM +0100, Arjen de Korte wrote:
> > > After updating from 2.2.0 to 2.2.1, the following command
> > > seems to go into loop and becomes unresponsive except for a
> > > "kill -9":
> > >
> > > /lib/nut/usbhid-ups -u root -DDD -a trippy
> >
> > Running the driver in debug mode has always prevented it from
> > backgrounding
>
> I understand this quite well.
>
> >, so nut-2.2.0 would also require a 'kill -9' to stop it,
>
> The previous driver would stop by typing a CTRL-C when
> running interactively in debug mode. The new driver will not
> respond to a CTRL-C. It does not respond to a "kill -1",
> "kill -2" or a "kill -15" either (Which may or may not be
> OK).
>
> I just thought it strange that the new version would require
> a SIGKILL to terminate rather than a SIGINT or SIGTERM. I
> guess it not a problem, just an observation of differing
> behavior between versions.

Is there a chance that either the kernel or libusb was upgraded in
between 2.2.0 and 2.2.1?

I can't think of any changes that would cause that sort of problem,
unless we somehow increased the polling rate to the point that we
tripped across some kernel problem. As it stands, we don't install
special handlers for SIGINT or SIGTERM, so they should kill the
process.

Also, when the data is printed at high speed, does it look relatively
correct, or does it look like there are errors for most of the
printouts?

-- 
- Charles Lepple



More information about the Nut-upsuser mailing list