On Mon, Jan 25, 2010 at 10:01:29AM +0100, Alexander Leidinger wrote: > Quoting Doug Barton <do...@freebsd.org> (from Sun, 24 Jan 2010 > 21:29:42 -0800 (PST)): > > >On Sat, 23 Jan 2010, Kostik Belousov wrote: > > > >>I do not see a need for such rudimentary ELF editor in the base at all. > > > >So, perhaps it's time to move brandelf out of the base? And if so, > >perhaps Alexander's contribution could be incorporated into a port > >for it? > > Personally I do not see a reason why his work can not go into the base > system. From a feature point of view the patch is giving brandelf a > little bit more freedom what it is allowed to change. When I look at > what I do/did with various tools in FreeBSD which where not intended > to be used like this but where useful in some cases, I do not think we > should enforce the policy to allow only stuff in brandelf which we are > able to emulate. > > >>After the work of dchagin@/bz@, brandelf is needed only for the corner > >>cases, if at all. > > > >Hmm, I was fooling around with some linux'y stuff the other day and > >needed to brandelf it (don't remember what, obviously wasn't that > >important). :) > >If this happens again in the future, is it worth reporting > >somewhere? (-emulation@ ?) > > If it was to brandelf a static linux executable so that the FreeBSD > system does not reboot when executing the static linux executable, > then I would say it does not need to be reported and we still need > brandelf in the base system. > > If someone says that exactly this case has been fixed recently: it > would be great to hear on emulation@ about cases where brandelf is > still needed.
If static linux binary contains .note.ABI-tag section, and I believe that relatively modern binaries do, then brand is autodetected.
pgpi7OgS0xmvQ.pgp
Description: PGP signature
