Messages in this thread | | | From | (Alan Cox) | Subject | Re: [PATCH] [SECURITY] suid procs exec'd with bad 0,1,2 fds | Date | Tue, 4 Aug 1998 22:30:11 +0100 (BST) |
| |
> To quote Linus - "Ostrich" 8) - Sun are doing the right thing. > > If they are doing it for all old sparc32 binaries, no they are not > doing the right thing. They are breaking a defined interface.
Configurably - exactly as we should be doing. In fact Sun actually tell you quite specifically that if it doesn't work tough. You have(had?) to get the Solaris kit, you had to enable it via the sysctl like stuff. Sun are quite obviously aware of the issues
> Not only does it break nearly all existing elf ABI's I know of, it is > a dirty and unclean implementation of the solution to top it off. Go
Indeed - the Solar stack patch is a hack - its a beautiful trick - but its not something I'd advocate being in generic 2.0 either - intel doesnt do non exec stacks right, full stop. Suns can do it
> more so down Sun's path, put bits in the ELF header which indicate > this "extension" and then we'll be more open to such a solution > hitting the main kernel sources...
It should be two things:
Elf bits for "non exec stack" and "exec stack" Maybe a sysctl for
0 Honour all requests, default to ABI specified 1 Honour all requests, default to nonexec
(I say maybe - I don't see why ld.so can't handle all of this via LD_PRELOAD)
Thats all I'd advocate on stack handling.
Alan
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.altern.org/andrebalsa/doc/lkml-faq.html
| |