Re: Re: Error reporting capabilities for libxc

[Gerd Hoffmann <kraxel@suse.de>]

Keir Fraser wrote:
> On 24/10/06 08:47, "Gerd Hoffmann" <kraxel@suse.de> wrote:
> 
>> What number space we are talking about btw?  errno?  something else?
> 
> I think a separate numbering space, where each err number would have some
> number of auxiliary arguments is being suggested, probably with an
> xc_{get,set}_errnum() interface (since the 'errnum' would not be a simple
> scalar so would be a bit of a pain to pass around).

I somehow feel this is a bit overdesigned.  I've just walked through my
domain builder rewrite, trying to assign useful error codes.  After all
there are not that many different cases:

 (1) internal errors (all sorts of sanity checks which normally
     never ever fail except when coding new bugs ;) ).
 (2) out of memory.
 (3) reading some file failed (kernel / initrd / hvmloader / whatelse).
 (4) invalid kernel image (broken elf headers, truncated file, ...).
 (5) incompatible kernel image (pae vs. nonpae, ppc kernel on x86, ...).
 (6) some invalid parameter (such as asking for a feature not supported
     by hypervisor or guest kernel, ...)

And the is only one where a fixed set of parameters specifying in detail
what went wrong somehow makes sense: for #3 this would be the filename
and maybe the reason (EPERM, ENOENT, ...).  For the other ones IMHO only
free text makes sense as detailed description ...

cheers,
  Gerd

-- 
Gerd Hoffmann <kraxel@suse.de>
http://www.suse.de/~kraxel/julika-dora.jpeg