Bug#774657: linux-image-3.16.0-4-amd64: firmware user helper floods log on missing firmware

[Alexander Greiner-Bär <alexander.greiner-baer@web.de>]

Hi Ben,

On Monday 05 January 2015 20:41:31
Ben Hutchings <ben@decadent.org.uk> wrote:
[...]
> Well that's a stupid thing for network-manager to do!

You are right. Because downgrading to 3.14 fixed the problem I made
the wrong assumption that a change in 3.15 breaks userland behaviour.
After thinking about it (see below) the problem I called a kernel bug
just reveals a bug in userland.
So you can close this bug. And thanks for cloning it as #774663.

[...]
> Why does the upstream change not do that?  I don't want to apply a fix
> that isn't upstream.

I have discussed this with Shuah Khan who made the last change in
this area. Because file system is not involved there the right
error is returning ENOMEM or EAGAIN. After also looking at the rt73usb
driver I come to the conclusion that it was also with 3.14 and before
possible that network-manager may receive an error when initialising
the interface. So only checking for ENOENT and stucking in a loop
otherwise is a network-manager bug.

Regards,
Alexander