Messages in this thread | | | From | Namhyung Kim <> | Subject | Re: [RFC PATCH 14/43] ptrace, frv: change signature of arch_ptrace() | Date | Fri, 27 Aug 2010 21:58:26 +0900 |
| |
David Howells <dhowells@redhat.com> writes: > Namhyung Kim <namhyung@gmail.com> wrote: >> I just wanted to let you know it depends on that. > > The patch being part of the series is probably sufficient, though a note of > the subject line of the previous patch would be useful. >
I see. Will do that hereafter.
>> What is the proper way to handle this? > > A summary of the changes being made is good: > > ptrace: Fix up the arguments arch_ptrace() in arch FRV > > Fix up the arguments to arch_ptrace() to take account of the fact that > addr and data are now unsigned long rather than long as of a preceding > patch in this series. > > Signed-off-by: ... >
Thanks. I will use this on the next round. :-)
> Note, however, that if the earlier patch breaks the compilation and then this > patch fixes it up, you should roll this patch into the earlier patch, and the > earlier patch is not complete without it. > > Think what happens if patch 3/43 breaks an arch, and then patch 43/43, say, > mends that arch, and then bisection lands on patch 3 during its progress. You > may end up having to 'git bisect skip' all the patches between 3 and 43 one at > a time. >
In this series, 3/43 changes the prototype of arch_ptrace() in include/linux/ptrace.h and the following patches change it for each arch in arch/xxx/kernel/ptrace.c. Do you mean all of arch change patches should be combinded into a patch?
-- Regards, Namhyung Kim
| |