* [PATCH] hugetlb: remove sysctl zero and infinity values
@ 2006-03-06 22:49 Dave Hansen
2006-03-07 0:30 ` Chen, Kenneth W
0 siblings, 1 reply; 2+ messages in thread
From: Dave Hansen @ 2006-03-06 22:49 UTC (permalink / raw)
To: wli; +Cc: linux-mm, Dave Hansen
These appear pretty redundant to me. The sysctl helper that
is used only deals with unsigned longs, so it doesn't make
much sense to try to deal with values less than zero here.
It is equally strange to try to impose a maximum of ~0UL.
There's also something a little bit fishy with putting
max_huge_pages in the sysctl table _and_ setting it manually
in the handler function. But, I'll leave that for another day.
---
work-dave/include/linux/hugetlb.h | 1 -
work-dave/kernel/sysctl.c | 2 --
work-dave/mm/hugetlb.c | 1 -
3 files changed, 4 deletions(-)
diff -puN kernel/sysctl.c~hugetlb-sysctl-remove-zero-infinity kernel/sysctl.c
--- work/kernel/sysctl.c~hugetlb-sysctl-remove-zero-infinity 2006-03-06 12:28:55.000000000 -0800
+++ work-dave/kernel/sysctl.c 2006-03-06 12:33:45.000000000 -0800
@@ -755,8 +755,6 @@ static ctl_table vm_table[] = {
.maxlen = sizeof(unsigned long),
.mode = 0644,
.proc_handler = &hugetlb_sysctl_handler,
- .extra1 = (void *)&hugetlb_zero,
- .extra2 = (void *)&hugetlb_infinity,
},
{
.ctl_name = VM_HUGETLB_GROUP,
diff -puN mm/hugetlb.c~hugetlb-sysctl-remove-zero-infinity mm/hugetlb.c
--- work/mm/hugetlb.c~hugetlb-sysctl-remove-zero-infinity 2006-03-06 12:28:55.000000000 -0800
+++ work-dave/mm/hugetlb.c 2006-03-06 12:28:55.000000000 -0800
@@ -19,7 +19,6 @@
#include <linux/hugetlb.h>
-const unsigned long hugetlb_zero = 0, hugetlb_infinity = ~0UL;
static unsigned long nr_huge_pages, free_huge_pages;
unsigned long max_huge_pages;
static struct list_head hugepage_freelists[MAX_NUMNODES];
diff -puN include/linux/hugetlb.h~hugetlb-sysctl-remove-zero-infinity include/linux/hugetlb.h
--- work/include/linux/hugetlb.h~hugetlb-sysctl-remove-zero-infinity 2006-03-06 12:28:55.000000000 -0800
+++ work-dave/include/linux/hugetlb.h 2006-03-06 12:28:55.000000000 -0800
@@ -28,7 +28,6 @@ int hugetlb_fault(struct mm_struct *mm,
unsigned long address, int write_access);
extern unsigned long max_huge_pages;
-extern const unsigned long hugetlb_zero, hugetlb_infinity;
extern int sysctl_hugetlb_shm_group;
/* arch callbacks */
_
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply [flat|nested] 2+ messages in thread* RE: [PATCH] hugetlb: remove sysctl zero and infinity values
2006-03-06 22:49 [PATCH] hugetlb: remove sysctl zero and infinity values Dave Hansen
@ 2006-03-07 0:30 ` Chen, Kenneth W
0 siblings, 0 replies; 2+ messages in thread
From: Chen, Kenneth W @ 2006-03-07 0:30 UTC (permalink / raw)
To: 'Dave Hansen', wli; +Cc: linux-mm
Dave Hansen wrote on Monday, March 06, 2006 2:50 PM
> There's also something a little bit fishy with putting
> max_huge_pages in the sysctl table _and_ setting it manually
> in the handler function. But, I'll leave that for another day.
max_huge_pages looks OK, maybe it has a bad name. Because that is
a variable used to pass desired hugetlb pool size by sys admin. It
is used only in the reservation path. Kernel pretty much needs at
least two variables: what is the desired target and what is current
reservation state (that tracked by nr_huge_pages).
- Ken
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2006-03-07 0:30 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2006-03-06 22:49 [PATCH] hugetlb: remove sysctl zero and infinity values Dave Hansen
2006-03-07 0:30 ` Chen, Kenneth W
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox