Re: Set of 16 potential bugs

[christos%astron.com@localhost (Christos Zoulas)]

In article <kqiqja$ck4$1%ger.gmane.org@localhost>,
Christos Zoulas <christos%astron.com@localhost> wrote:
>In article <20130628014532.GA27908@marx.bitnet>,
>Jukka Ruohonen  <jruohonen%iki.fi@localhost> wrote:
>>kieeeOn Thu, Jun 27, 2013 at 05:34:22PM +0200, Maxime Villard wrote:
>>> Hi,
>>> I've been making a code scanner for a while, and I recently launched it
>>> on NetBSD's source tree to test some new rules. It found 16 potential
>>> bugs that I've summed up here:
>>
>>It is far from being a surprise that you've located bugs. But what is more
>>interesting, is that you've *only* located sixteen bugs from the code base
>>that dates to the 1980s.
>
>Yes, and the majority of them so far have been correct! Nice job.

The wrong ones are:
        - the assignment after a setjmp() == 0 will always happen, but the
          checker could not have known that. I made it clear now. (01-0x01)
        - the dnode uninitialized access cannot happen because v == -1 unless
          dnode is initialized. Not fixed. (04-0x01)
        - the madness of nfs macros; mb is actually initiliazed in the
          nfsm_reply macro. Not fixed. (05-0x01)

Thanks for the bug reports!

christos