Paul Mackerras

The invitation did go to spam, but I found it. Seems like coverity is complaining nearly every time we use warn() (or info(), fatal(), error(), etc.) because of the existence...

> it does not make any assumption. it just follows the build process So why does it think that the warn() called from various places in ipcp.c could ever be...

> there's non zero probabilty that "ppp" [configures itself](https://github.com/ppp-project/ppp/blob/master/configure.ac#L14) to use plugins from pppoe folder. Sure... and? The point is that coverity is putting together a use of a function...

I just saw a new batch of scan results via email. They all seem to be predicated on the assumption that read() can return absolutely any integer, whereas by definition...

Oh, and there were a couple of reports that would be valid if tdb->hash_size could ever be >= 2^31. But if you look at the code, it's only ever going...

I agree there's a problem - PPPOEInitDevice() is called from PPPoEDevnameHook(), which is called when the ethernet device name is seen while parsing the arguments. So if the pppoe-padi-timeout or...

I pushed a proposed fix to the pppoe-options branch in this repository. @ipaton1 please test it and confirm whether it fixes the problem. If it does I'll merge it.

@pali any comment? I guess that if you want to know about all concentrators, it is reasonable to wait a little while after the first response to see if any...

I got a reply from @pali via email, with a patch. @ipaton1 please test the pppoe-verbose-fix branch in this repository and tell me whether it fixes your problem.

Thanks for noting that.