On 03/08/2016 12:02 PM, Tetsuo Handa wrote:
> Since we assigned mm = victim->mm before pr_err(),
> we don't need to dereference victim->mm again at pr_err().
> This saves a few instructions.
That sounds obvious, right. Yet once in a while I try to test these for
fun, and there can be indeed surprises :)
./scripts/bloat-o-meter says:
add/remove: 0/0 grow/shrink: 1/0 up/down: 1/0 (1)
function old new delta
oom_kill_process 1085 1086 +1
a naive asmdiff is too complicated to follow but it seems from the
number of lines that indeed there are less instructions in your case,
but still the code is a bit larger.
Just a reminder that compilers can be quite counter-intuitive. Anyway, a
liquid-helium-path like this probably doesn't need such
microoptimisations as an extra patch, given the ongoing churn in the oom
area?
> Signed-off-by: Tetsuo Handa <penguin-kernel@xxxxxxxxxxxxxxxxxxx>
> ---
> mm/oom_kill.c | 8 ++++----
> 1 file changed, 4 insertions(+), 4 deletions(-)
>
> diff --git a/mm/oom_kill.c b/mm/oom_kill.c
> index c84e784..1808db32 100644
> --- a/mm/oom_kill.c
> +++ b/mm/oom_kill.c
> @@ -756,10 +756,10 @@ void oom_kill_process(struct oom_control *oc, struct task_struct *p,
> do_send_sig_info(SIGKILL, SEND_SIG_FORCED, victim, true);
> mark_oom_victim(victim);
> pr_err("Killed process %d (%s) total-vm:%lukB, anon-rss:%lukB, file-rss:%lukB, shmem-rss:%lukB\n",
> - task_pid_nr(victim), victim->comm, K(victim->mm->total_vm),
> - K(get_mm_counter(victim->mm, MM_ANONPAGES)),
> - K(get_mm_counter(victim->mm, MM_FILEPAGES)),
> - K(get_mm_counter(victim->mm, MM_SHMEMPAGES)));
> + task_pid_nr(victim), victim->comm, K(mm->total_vm),
> + K(get_mm_counter(mm, MM_ANONPAGES)),
> + K(get_mm_counter(mm, MM_FILEPAGES)),
> + K(get_mm_counter(mm, MM_SHMEMPAGES)));
> task_unlock(victim);
>
> /*
>
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@xxxxxxxxx. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>