Date: Sun, 22 Jan 2006 07:37:23 +1100 From: Peter Jeremy <PeterJeremy@optushome.com.au> To: Tom Rhodes <trhodes@freebsd.org> Cc: cvs-src@freebsd.org, src-committers@freebsd.org, cvs-all@freebsd.org Subject: Re: cvs commit: src/sys/nfsserver nfs_serv.c Message-ID: <20060121203723.GT25397@cirb503493.alcatel.com.au> In-Reply-To: <200601211210.k0LCAXYl069896@repoman.freebsd.org> References: <200601211210.k0LCAXYl069896@repoman.freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sat, 2006-Jan-21 12:10:33 +0000, Tom Rhodes wrote: >trhodes 2006-01-21 12:10:33 UTC > > FreeBSD src repository > > Modified files: > sys/nfsserver nfs_serv.c > Log: > Remove some dead code. > > Found with: Coverity Prevent(tm) I'll put my $0.02 in and agree with Stefan Farfeleder. (Luckily, in this case, the notorious NFS macros are not involved). The comments on nfs_namei() state that dirp can be returned not-NULL even if an error occurs and a check of the code paths in nfs_namei() indicates that this is correct. Can you please re-evaluate your change. If (as I suspect), this is actually an incorrect report from Coverity, we should probably report it back to them to investigate. IMHO, as a general rule, I would suggest that dead code that does not trivially resolve to "if (0)" should be replaced by a KASSERT(9) so that if the code path isn't totally dead, we have some chance of finding out about it. Not to pick on Tom, but I've noticed a lot of Coverity fixes being applied. I would expect that Coverity reports would be validated in the same way any any other bug reports (even if that doesn mean wading through NFS's maze of twisty large macros). -- Peter Jeremy
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20060121203723.GT25397>