Re: kldload won't load

On 12-Jul-2004 Robert Watson wrote:
> 
> On Mon, 12 Jul 2004, John Polstra wrote:
> 
>> No, it needs to be fixed.  It's printing a totally incorrect error
>> message, and nobody should have to use dmesg to find out what's really
>> happened. 
> 
> Well, the problem here is that the errno error-reporting mechanism can
> report but not describe errors.  We could add a new EKLDLINKER to point at
> a linker error (or the like), or a whole set of new errnos, but the
> mechanism even then couldn't report which symbols are missing, etc.  One
> or more linker-specific error values would probably be a useful start.

I fully understand the implementation difficulties, but we have to
be careful to observe the distinction between correct behavior and
easy-to-implement behavior.  The kldload(2) API doesn't support
correct behavior, so that's where the focus needs to be -- not on
making users feel stupid for failing to look in the dmesg output.

John