Date: Thu, 19 Jun 2003 01:01:28 -0400 From: Alexander Kabaev <kabaev@mail.ru> To: "Matthew N. Dodd" <mdodd@FreeBSD.org> Cc: cvs-all@FreeBSD.org Subject: Re: cvs commit: src/libexec/rtld-elf debug.c rtld.1 rtld.c rtld.h Message-ID: <20030619010128.13160a72.kabaev@mail.ru> In-Reply-To: <20030619003055.G76384@sasami.jurai.net> References: <200306190355.h5J3tc0A060501@repoman.freebsd.org> <20030619002345.4f626848.kabaev@mail.ru> <20030619003055.G76384@sasami.jurai.net>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, 19 Jun 2003 00:40:04 -0400 (EDT) "Matthew N. Dodd" <mdodd@FreeBSD.org> wrote: > On Thu, 19 Jun 2003, Alexander Kabaev wrote: > > Just out of curiosity what does that buy you compared to > > objdump/readelf tools? Frankly I do not think this kind of logging > > output belongs to rtld and would like to see this patch backed out. > > It is also possible there are some advantages I am missing, in that > > case an explanation will be very much appreciated. > > This allows you to see what relocations were performed. > > No other tools in the tree provide this functionality. > > Anyone working with the MD bits has probably used similar code at one > point or another. > True, but these helper functions usually never hit the tree as they are considered to be very task specific and probably not interesting for a majority of our user base. I am withdrawing my objections with a note that ever growing amount of diagnostic functionality in rtld will have to be addressed some day. There is an option to offload most of it into a separate demand-loaded library in a manner similar to Solaris.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20030619010128.13160a72.kabaev>