Comments below: On 13/04/05, Blaisorblade wrote: > On Monday 11 April 2005 01:19, Ian McDonald wrote: > > With this patch applied to 2.6.11.7 it is my first kernel that > > compiles straight off :-) > Ok, that's the -bs tree, not the one I announced here (the -devel one) but > it's ok anyway. > > It does not apply one patch because it is already applied which is fine. > Yes, I merged that, it's the va_copy fix. Sorry about that... > > I did a quick test against 2.6.12-rc2 and it didn't like that at all > > (many rejected patches). That is to be expected though as that is not > > the target but it does mean a reasonable amount of work ahead for > > Paolo :-( > No; luckily, I'm mainly working on -rc2 and I've backported the needed stuff > to 2.6.11. Found the development snapshot and found the following problems: One patch did not apply. Documented in patch.out which is attached. Problem with arch/um/include/mode_kern.h as below In file included from arch/um/kernel/exec_kern.c:23: arch/um/include/mode_kern.h:12: unterminated string or character constant It doesn't like the #error on gcc 2.95.4 - if I do a remove of this line it compiles... will need a better fix than this though.... Problem with arch/um/kernel/init_task.c as below arch/um/kernel/init_task.c:30: duplicate initializer arch/um/kernel/init_task.c:30: (near initialization for `init_task.thread.arch.debugregs') This is probably a duplication in initialisation of struct just like the problem we had with enum in syscalls and again due to gcc 2.95.4. I haven't tracked this one down fully yet but can if needed buy may take a while... I then hacked half that line out and got further compile errors in arch/um/kernel/process_kern.c as per attached file err.out. I then stopped as this looked a little more complicated... If you need any further information please tell me... I have also checked that 2.6.12-rc2 without the patch does compile and it does with the addition of the sed patch. If you want me to help I can apply the patches one by one but this will take some time... Regards, Ian