Messages in this thread | | | Date | Sun, 20 Jul 2008 12:51:39 +0400 | From | Oleg Nesterov <> | Subject | Re: [PATCH] coredump: kill mm->core_done |
| |
On 07/19, Roland McGrath wrote: > > It is indeed ugly. I think the place to start is with moving all the > coredump code out of exit_mm() into a new subfunction that is just passed > the mm pointer, and doing the same with coredump_finish() as you've done, > but leaving the completion logic inside those functions the same.
The next patch does this.
This patch goes first because it kills the last field in mm_struct which is needed for the coredump code, so new new function (exit_coredump) doesn't need the mm pointer at all.
Oleg.
| |