The patch titled
kernel/sys.c: get rid of expensive divides in groups_sort()
has been added to the -mm tree. Its filename is
kernel-sysc-get-rid-of-expensive-divides-in-groups_sort.patch
*** Remember to use Documentation/SubmitChecklist when testing your code ***
See http://www.zip.com.au/~akpm/linux/patches/stuff/added-to-mm.txt to find
out what to do about this
------------------------------------------------------
Subject: kernel/sys.c: get rid of expensive divides in groups_sort()
From: Eric Dumazet <dada1@xxxxxxxxxxxxx>
groups_sort() can be quite long if user loads a large gid table.
This is because GROUP_AT(group_info, some_integer) uses an integer divide.
So having to do XXX thousand divides during one syscall can lead to very
high latencies. (NGROUPS_MAX=65536)
In the past (25 Mar 2006), an analog problem was found in groups_search()
(commit d74beb9f33a5f16d2965f11b275e401f225c949d ) and at that time I
changed some variables to unsigned int.
I believe that a more generic fix is to make sure NGROUPS_PER_BLOCK is
unsigned.
Signed-off-by: Eric Dumazet <dada1@xxxxxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
include/linux/sched.h | 2 +-
kernel/sys.c | 20 ++++++++++----------
2 files changed, 11 insertions(+), 11 deletions(-)
diff -puN include/linux/sched.h~kernel-sysc-get-rid-of-expensive-divides-in-groups_sort include/linux/sched.h
--- a/include/linux/sched.h~kernel-sysc-get-rid-of-expensive-divides-in-groups_sort
+++ a/include/linux/sched.h
@@ -803,7 +803,7 @@ static inline int above_background_load(
struct io_context; /* See blkdev.h */
#define NGROUPS_SMALL 32
-#define NGROUPS_PER_BLOCK ((int)(PAGE_SIZE / sizeof(gid_t)))
+#define NGROUPS_PER_BLOCK ((unsigned int)(PAGE_SIZE / sizeof(gid_t)))
struct group_info {
int ngroups;
atomic_t usage;
diff -puN kernel/sys.c~kernel-sysc-get-rid-of-expensive-divides-in-groups_sort kernel/sys.c
--- a/kernel/sys.c~kernel-sysc-get-rid-of-expensive-divides-in-groups_sort
+++ a/kernel/sys.c
@@ -1145,16 +1145,16 @@ static int groups_to_user(gid_t __user *
struct group_info *group_info)
{
int i;
- int count = group_info->ngroups;
+ unsigned int count = group_info->ngroups;
for (i = 0; i < group_info->nblocks; i++) {
- int cp_count = min(NGROUPS_PER_BLOCK, count);
- int off = i * NGROUPS_PER_BLOCK;
- int len = cp_count * sizeof(*grouplist);
+ unsigned int cp_count = min(NGROUPS_PER_BLOCK, count);
+ unsigned int len = cp_count * sizeof(*grouplist);
- if (copy_to_user(grouplist+off, group_info->blocks[i], len))
+ if (copy_to_user(grouplist, group_info->blocks[i], len))
return -EFAULT;
+ grouplist += NGROUPS_PER_BLOCK;
count -= cp_count;
}
return 0;
@@ -1165,16 +1165,16 @@ static int groups_from_user(struct group
gid_t __user *grouplist)
{
int i;
- int count = group_info->ngroups;
+ unsigned int count = group_info->ngroups;
for (i = 0; i < group_info->nblocks; i++) {
- int cp_count = min(NGROUPS_PER_BLOCK, count);
- int off = i * NGROUPS_PER_BLOCK;
- int len = cp_count * sizeof(*grouplist);
+ unsigned int cp_count = min(NGROUPS_PER_BLOCK, count);
+ unsigned int len = cp_count * sizeof(*grouplist);
- if (copy_from_user(group_info->blocks[i], grouplist+off, len))
+ if (copy_from_user(group_info->blocks[i], grouplist, len))
return -EFAULT;
+ grouplist += NGROUPS_PER_BLOCK;
count -= cp_count;
}
return 0;
_
Patches currently in -mm which might be from dada1@xxxxxxxxxxxxx are
origin.patch
parport-dev-timeslice-is-an-unsigned-long-not-an-int.patch
git-net.patch
git-sched.patch
percpu-__percpu_alloc_mask-can-dynamically-size-percpu_data.patch
get-rid-of-nr_open-and-introduce-a-sysctl_nr_open.patch
get-rid-of-nr_open-and-introduce-a-sysctl_nr_open-fix.patch
lib-extablec-removes-an-expensive-integer-divide-in-search_extable.patch
kernel-sysc-get-rid-of-expensive-divides-in-groups_sort.patch
-
To unsubscribe from this list: send the line "unsubscribe mm-commits" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
