Date: Tue, 1 Oct 1996 14:51:34 -0700 (MST) From: Terry Lambert <terry@lambert.org> To: jdp@polstra.com (John Polstra) Cc: current@FreeBSD.ORG Subject: Re: WARNING: botched ld.so commit! :-( Message-ID: <199610012151.OAA02594@phaeton.artisoft.com> In-Reply-To: <199610011615.JAA03350@austin.polstra.com> from "John Polstra" at Oct 1, 96 09:15:40 am
next in thread | previous in thread | raw e-mail | index | archive | help
> > Even I don't touch rtld :-). Scary thought if I worked on it :-). > > Just to reinforce that sentiment, I worked on rtld for a while, and it's > nerve-wracking! It's like trying to replace a rope that you and the > rest of the world are hanging from. (By the neck, if you screw up.) > > I think we all know Peter well enough to forego giving him too much > flack over this. (Same goes for John Dyson and the occasional > hiccups that come with the great things he's doing for us.) For what it's worth: a correct ELF implemented is expected to have the kernel execution class loader map the ld.so into the process address space instead of relying on the crt0.o to do it for you. That's the reason for the "large enough" offset at the start of the SVR4 EABI specification prior to the image mapping location. Going to something like this would allow the kernel to automatically establish correct versioning, and you could overrride using userconfig in case you *really* screwed up. Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199610012151.OAA02594>