From: Jeff Dike <jdike@ka...> - 2001-04-14 22:27:07
> I'm working on TUX: I took 2.4.2, Ingo's tux2-full-X patch and then
> applied the UML 2.4.3 patch.
Applying the UML patch to the corresponding kernel would have been a good idea.
> - I changed some pte,pmd calls in the UML file mm/memory.c -- a
> little worrying
There were some mm changes and I reorganized my pgtable.h to match. I don't
know if this has anything to do with what you did in memory.c, though.
> - arch/um/kernel/signal_kern.c has a different prototype for binfmt->
> vmlinux.o: In function `bust_spinlocks': /usr/src/kernel-tux/lib/
> bust_spinlocks.c:25: undefined reference to `unblank_screen'
These weren't problems in the Linus 2.4.2. They're in Alan's 2.4.3 tree, and
I sent him a patch (which he hasn't merged yet) to fix these. See
http://www.geocrawler.com/lists/3/SourceForge/709/0/5527690/ for the patch.
> - fs.h has a long sys_dup, but unistd.h has int sys_dup
> vmlinux.o: In function `inet_sock_destruct': /usr/src/kernel-tux/net/
> ipv4/af_inet.c:154: undefined reference to `show_stack'
These I haven't seen, so I don't know what to do about them.
> When starting UML with 2.4.2 and Tux patches I get a segfault in
> mounting /proc. Can someone tell me whatis wrong or have I hacked my
> way in too fast?
Why is this a problem? That's supposed to happen. It happens on native
This is just copy_from_user relying on the fault handler to tell it when to
stop copying user data.