Linux kernel development

Hillf Danton | 1 Jul 2012 15:27

[RFC patch] BFS: 421-1

With 15 patches collected, the 421-1 is ready with iso untouched.
Note diff is based not on 420 but on 421.

--- a/kernel/sched/bfs.c	Sun Jul  1 20:39:30 2012
+++ b/kernel/sched/bfs.c	Fri Jun 15 20:00:52 2012
 <at>  <at>  -113,7 +113,6  <at>  <at> 
 #define USER_PRIO(p)		((p) - MAX_RT_PRIO)
 #define TASK_USER_PRIO(p)	USER_PRIO((p)->static_prio)
 #define MAX_USER_PRIO		(USER_PRIO(MAX_PRIO))
-#define SCHED_PRIO(p)		((p) + MAX_RT_PRIO)
 #define STOP_PRIO		(MAX_RT_PRIO - 1)

 /*
 <at>  <at>  -150,6 +149,19  <at>  <at>  int rr_interval __read_mostly = 6;
  */
 int sched_iso_cpu __read_mostly = 70;

+#ifdef CONFIG_SMP
+enum {
+	EDL_CK,		//default
+	EDL_MS,		//map cache distance to milliseconds
+	EDL_NONE,	//strict edl
+};
+int edl_mode = EDL_CK;
+
+unsigned long	grab_rq_lock = 0,
+		wait_rq_lock = 0,
+		tsk_csw = 0,
+		cpu_csw = 0;
+#endif

(Continue reading)

headers

Hillf Danton | 1 Jul 2012 15:29

Re: [RFC patch] BFS: 421-1

On Sun, Jul 1, 2012 at 9:27 PM, Hillf Danton <dhillf <at> gmail.com> wrote:
> With 15 patches collected, the 421-1 is ready with iso untouched.
> Note diff is based not on 420 but on 421.

Here is the diff on 420.

--- a/kernel/sched/bfs.c	Sun Jul  1 20:44:00 2012
+++ b/kernel/sched/bfs.c	Fri Jun 15 20:00:52 2012
 <at>  <at>  -113,7 +113,6  <at>  <at> 
 #define USER_PRIO(p)		((p) - MAX_RT_PRIO)
 #define TASK_USER_PRIO(p)	USER_PRIO((p)->static_prio)
 #define MAX_USER_PRIO		(USER_PRIO(MAX_PRIO))
-#define SCHED_PRIO(p)		((p) + MAX_RT_PRIO)
 #define STOP_PRIO		(MAX_RT_PRIO - 1)

 /*
 <at>  <at>  -133,7 +132,7  <at>  <at> 

 void print_scheduler_version(void)
 {
-	printk(KERN_INFO "BFS CPU scheduler v0.420 by Con Kolivas.\n");
+	printk(KERN_INFO "BFS 421-1 based on v0.420 by Con Kolivas\n");
 }

 /*
 <at>  <at>  -150,6 +149,19  <at>  <at>  int rr_interval __read_mostly = 6;
  */
 int sched_iso_cpu __read_mostly = 70;

+#ifdef CONFIG_SMP

(Continue reading)

headers

Heinz Diehl | 1 Jul 2012 16:58

Re: [RFC patch] BFS: 421-1

On 01.07.2012, Hillf Danton wrote: 

> With 15 patches collected, the 421-1 is ready with iso untouched.
> Note diff is based not on 420 but on 421.
[....]

There is no v.421 on ck.kolivas.org, and it seems it never has
been, at least there is a patch which updates v.420 directly to v.422
which indicates that there isn't any v.421. So what version is your 
patchset actually based on?

Btw: Con Kolivas released BFS v.422 already some weeks ago, you maybe
should think on rebasing your patchset on that one?

Confused..

headers

Emmanuel Benisty | 1 Jul 2012 18:13

Re: [RFC patch] BFS: 421-1

On Sun, Jul 1, 2012 at 9:58 PM, Heinz Diehl <htd <at> fancy-poultry.org> wrote:
> On 01.07.2012, Hillf Danton wrote:
>
>> With 15 patches collected, the 421-1 is ready with iso untouched.
>> Note diff is based not on 420 but on 421.
> [....]
>
> There is no v.421 on ck.kolivas.org, and it seems it never has
> been, at least there is a patch which updates v.420 directly to v.422
> which indicates that there isn't any v.421. So what version is your
> patchset actually based on?

Con skipped 421 because Hillf did release a 421 BFS, he wanted to
avoid confusion. There are two patches in this thread, the first one
is based on Hillf's 421 (you can find the patch on lkml) and a second
one based on Con's 420 BFS. Hope it's a bit more clear now.

 <at> Hillf, 421-1 builds and boots fine. Too early for me to give any
feedback on improvements and/or regressions this patch brings.