From mboxrd@z Thu Jan 1 00:00:00 1970 From: Paul Moore To: David Howells Subject: Re: [PATCH] Fix a potentially uninitialised variable in SELinux hooks Date: Wed, 9 Jul 2008 21:23:40 -0400 Cc: sds@tycho.nsa.gov, jmorris@namei.org, selinux@tycho.nsa.gov References: <200807091847.14347.paul.moore@hp.com> <8198.1215643142@redhat.com> <8883.1215645130@redhat.com> In-Reply-To: <8883.1215645130@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Message-Id: <200807092123.40667.paul.moore@hp.com> Sender: owner-selinux@tycho.nsa.gov List-Id: selinux@tycho.nsa.gov On Wednesday 09 July 2008 7:12:10 pm David Howells wrote: > Paul Moore wrote: > > ret = selinux_parse_skb_ipv4(...); > > if (ret != 0) > > addrp = (src ? ...); > > else > > addrp = NULL; > > I guess you mean "*addrp = ..." in that case, otherwise you haven't > eliminated anything. Yes, I did. Too much pseudo in my pseudo-code I guess. > Personally, I prefer to add extra variables if it makes things > clearer, and I prefer to use gotos for error handling. It eliminates > the else-statements that you would otherwise introduce, and the > goto-label can be used as documentation of a sort too. Furthermore, > it moves the error handling clearly out of the main route through the > function. Having said that, I should reorder my patch to put the > parse_error segment last - then I can ditch the okay label and use > break instead of goto. > > It comes down to personal preference, I guess. Doesn't it always :) I don't think it makes a huge difference either way, and since you are the one fixing the problem I'll let you decide. -- paul moore linux @ hp -- This message was distributed to subscribers of the selinux mailing list. If you no longer wish to subscribe, send mail to majordomo@tycho.nsa.gov with the words "unsubscribe selinux" without quotes as the message.