From: Jack Burton <jack@saosce.com.au>
Subject: Re: [patch 1 of 3] iscsid(8): fix memory leak
To: Claudio Jeker <cjeker@diehard.n-r-g.com>
Cc: tech@openbsd.org
Date: Fri, 17 Jan 2025 18:23:58 +1030

On Thu, 16 Jan 2025 21:24:58 +0100
Claudio Jeker <cjeker@diehard.n-r-g.com> wrote:
> > > I will commit this diff with the above mentioned changes
> > > tomorrow.  
> > 
> > Thanks; much appreciated.  
> 
> Diff is in. I guess we could really always allocate value and always
> use a const char for the key (I see no reason to support or enforce
> allocation there).

Agreed.  Assuming we follow the "always get in first with proposals"
approach of my patch #3, the only genuine need to support allocated keys
would be if we had to respond to any proposals that could only be made
by the target ... and looking at section 13 of RFC 7143 all such
proposals are declarative, so we don't need to respond to them.
Perfect.

Once the replacement for my original patch #2 is done & dusted, I'll
redo my patch #3 to suit that *and* to avoid allocating any keys.