From: akpm@linux-foundation.org
To: adobriyan@gmail.com, ebiederm@xmission.com,
keescook@chromium.org, oleg@redhat.com,
mm-commits@vger.kernel.org
Subject: + pid-delete-struct-pidmap-nr_free.patch added to -mm tree
Date: Mon, 02 Oct 2017 16:13:31 -0700 [thread overview]
Message-ID: <59d2c81b.sN3iiDcgxwFHtFRg%akpm@linux-foundation.org> (raw)
The patch titled
Subject: pid: delete struct pidmap::nr_free
has been added to the -mm tree. Its filename is
pid-delete-struct-pidmap-nr_free.patch
This patch should soon appear at
http://ozlabs.org/~akpm/mmots/broken-out/pid-delete-struct-pidmap-nr_free.patch
and later at
http://ozlabs.org/~akpm/mmotm/broken-out/pid-delete-struct-pidmap-nr_free.patch
Before you just go and hit "reply", please:
a) Consider who else should be cc'ed
b) Prefer to cc a suitable mailing list as well
c) Ideally: find the original patch on the mailing list and do a
reply-to-all to that, adding suitable additional cc's
*** Remember to use Documentation/SubmitChecklist when testing your code ***
The -mm tree is included into linux-next and is updated
there every 3-4 working days
------------------------------------------------------
From: Alexey Dobriyan <adobriyan@gmail.com>
Subject: pid: delete struct pidmap::nr_free
There is a check in pid allocation code to skip a full page:
if (likely(atomic_read(&map->nr_free))) {
...
In practice it doesn't do anything. To skip a pidmap page one has to have
32K consecutive pids allocated at the same time which doesn't happen.
Currently the price is _every_ fork/exit on every system being slower than
necessary.
Link: http://lkml.kernel.org/r/20170909203649.GB4791@avx2
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
---
include/linux/pid_namespace.h | 1 -
kernel/pid.c | 28 ++++++++++------------------
kernel/pid_namespace.c | 6 ------
3 files changed, 10 insertions(+), 25 deletions(-)
diff -puN include/linux/pid_namespace.h~pid-delete-struct-pidmap-nr_free include/linux/pid_namespace.h
--- a/include/linux/pid_namespace.h~pid-delete-struct-pidmap-nr_free
+++ a/include/linux/pid_namespace.h
@@ -11,7 +11,6 @@
#include <linux/ns_common.h>
struct pidmap {
- atomic_t nr_free;
void *page;
};
diff -puN kernel/pid.c~pid-delete-struct-pidmap-nr_free kernel/pid.c
--- a/kernel/pid.c~pid-delete-struct-pidmap-nr_free
+++ a/kernel/pid.c
@@ -68,9 +68,6 @@ static inline int mk_pid(struct pid_name
*/
struct pid_namespace init_pid_ns = {
.kref = KREF_INIT(2),
- .pidmap = {
- [ 0 ... PIDMAP_ENTRIES-1] = { ATOMIC_INIT(BITS_PER_PAGE), NULL }
- },
.last_pid = 0,
.nr_hashed = PIDNS_HASH_ADDING,
.level = 0,
@@ -106,7 +103,6 @@ static void free_pidmap(struct upid *upi
int offset = nr & BITS_PER_PAGE_MASK;
clear_bit(offset, map->page);
- atomic_inc(&map->nr_free);
}
/*
@@ -181,20 +177,17 @@ static int alloc_pidmap(struct pid_names
if (unlikely(!map->page))
return -ENOMEM;
}
- if (likely(atomic_read(&map->nr_free))) {
- for ( ; ; ) {
- if (!test_and_set_bit(offset, map->page)) {
- atomic_dec(&map->nr_free);
- set_last_pid(pid_ns, last, pid);
- return pid;
- }
- offset = find_next_offset(map, offset);
- if (offset >= BITS_PER_PAGE)
- break;
- pid = mk_pid(pid_ns, map, offset);
- if (pid >= pid_max)
- break;
+ for (;;) {
+ if (!test_and_set_bit(offset, map->page)) {
+ set_last_pid(pid_ns, last, pid);
+ return pid;
}
+ offset = find_next_offset(map, offset);
+ if (offset >= BITS_PER_PAGE)
+ break;
+ pid = mk_pid(pid_ns, map, offset);
+ if (pid >= pid_max)
+ break;
}
if (map < &pid_ns->pidmap[(pid_max-1)/BITS_PER_PAGE]) {
++map;
@@ -591,7 +584,6 @@ void __init pidmap_init(void)
init_pid_ns.pidmap[0].page = kzalloc(PAGE_SIZE, GFP_KERNEL);
/* Reserve PID 0. We never call free_pidmap(0) */
set_bit(0, init_pid_ns.pidmap[0].page);
- atomic_dec(&init_pid_ns.pidmap[0].nr_free);
init_pid_ns.pid_cachep = KMEM_CACHE(pid,
SLAB_HWCACHE_ALIGN | SLAB_PANIC | SLAB_ACCOUNT);
diff -puN kernel/pid_namespace.c~pid-delete-struct-pidmap-nr_free kernel/pid_namespace.c
--- a/kernel/pid_namespace.c~pid-delete-struct-pidmap-nr_free
+++ a/kernel/pid_namespace.c
@@ -98,7 +98,6 @@ static struct pid_namespace *create_pid_
struct pid_namespace *ns;
unsigned int level = parent_pid_ns->level + 1;
struct ucounts *ucounts;
- int i;
int err;
err = -EINVAL;
@@ -139,11 +138,6 @@ static struct pid_namespace *create_pid_
INIT_WORK(&ns->proc_work, proc_cleanup_work);
set_bit(0, ns->pidmap[0].page);
- atomic_set(&ns->pidmap[0].nr_free, BITS_PER_PAGE - 1);
-
- for (i = 1; i < PIDMAP_ENTRIES; i++)
- atomic_set(&ns->pidmap[i].nr_free, BITS_PER_PAGE);
next reply other threads:[~2017-10-02 23:13 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-10-02 23:13 akpm [this message]
2017-10-03 16:15 ` + pid-delete-struct-pidmap-nr_free.patch added to -mm tree Oleg Nesterov
2017-10-03 17:39 ` Gargi Sharma
2017-10-03 21:56 ` Alexey Dobriyan
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=59d2c81b.sN3iiDcgxwFHtFRg%akpm@linux-foundation.org \
--to=akpm@linux-foundation.org \
--cc=adobriyan@gmail.com \
--cc=ebiederm@xmission.com \
--cc=keescook@chromium.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mm-commits@vger.kernel.org \
--cc=oleg@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.