From: Al Viro <viro@ZenIV.linux.org.uk>
To: linux-fsdevel@vger.kernel.org
Cc: Linus Torvalds <torvalds@linux-foundation.org>,
linux-kernel@vger.kernel.org, Stephen Tweedie <sct@redhat.com>,
Jeremy Eder <jeder@redhat.com>
Subject: [PATCH 4/5][RFC][CFT] reduce m_start() cost...
Date: Wed, 5 Mar 2014 03:51:47 +0000 [thread overview]
Message-ID: <20140305035147.GD26528@ZenIV.linux.org.uk> (raw)
In-Reply-To: <20140305034751.GW18016@ZenIV.linux.org.uk>
Don't rescan the mount list in m_start() if there hadn't been
any changes and we are right at the needed place of just before
it (that's the only occuring ones on sequential read() without
seeks).
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
fs/mount.h | 5 ++++-
fs/namespace.c | 21 ++++++++++++++++++---
fs/proc_namespace.c | 1 +
3 files changed, 23 insertions(+), 4 deletions(-)
diff --git a/fs/mount.h b/fs/mount.h
index a17458c..61e1c05 100644
--- a/fs/mount.h
+++ b/fs/mount.h
@@ -10,7 +10,7 @@ struct mnt_namespace {
struct user_namespace *user_ns;
u64 seq; /* Sequence number to prevent loops */
wait_queue_head_t poll;
- int event;
+ u64 event;
};
struct mnt_pcp {
@@ -104,6 +104,9 @@ struct proc_mounts {
struct mnt_namespace *ns;
struct path root;
int (*show)(struct seq_file *, struct vfsmount *);
+ void *cached_mount;
+ u64 cached_event;
+ loff_t cached_index;
};
#define proc_mounts(p) (container_of((p), struct proc_mounts, m))
diff --git a/fs/namespace.c b/fs/namespace.c
index f28c4fd..00f12af 100644
--- a/fs/namespace.c
+++ b/fs/namespace.c
@@ -29,7 +29,7 @@
#define HASH_SHIFT ilog2(PAGE_SIZE / sizeof(struct list_head))
#define HASH_SIZE (1UL << HASH_SHIFT)
-static int event;
+static u64 event;
static DEFINE_IDA(mnt_id_ida);
static DEFINE_IDA(mnt_group_ida);
static DEFINE_SPINLOCK(mnt_id_lock);
@@ -1064,14 +1064,29 @@ static void *m_start(struct seq_file *m, loff_t *pos)
struct proc_mounts *p = proc_mounts(m);
down_read(&namespace_sem);
- return seq_list_start(&p->ns->list, *pos);
+ if (p->cached_event == p->ns->event) {
+ void *v = p->cached_mount;
+ if (*pos == p->cached_index)
+ return v;
+ if (*pos == p->cached_index + 1) {
+ v = seq_list_next(v, &p->ns->list, &p->cached_index);
+ return p->cached_mount = v;
+ }
+ }
+
+ p->cached_event = p->ns->event;
+ p->cached_mount = seq_list_start(&p->ns->list, *pos);
+ p->cached_index = *pos;
+ return p->cached_mount;
}
static void *m_next(struct seq_file *m, void *v, loff_t *pos)
{
struct proc_mounts *p = proc_mounts(m);
- return seq_list_next(v, &p->ns->list, pos);
+ p->cached_mount = seq_list_next(v, &p->ns->list, pos);
+ p->cached_index = *pos;
+ return p->cached_mount;
}
static void m_stop(struct seq_file *m, void *v)
diff --git a/fs/proc_namespace.c b/fs/proc_namespace.c
index 7be26f0..1a81373 100644
--- a/fs/proc_namespace.c
+++ b/fs/proc_namespace.c
@@ -267,6 +267,7 @@ static int mounts_open_common(struct inode *inode, struct file *file,
p->root = root;
p->m.poll_event = ns->event;
p->show = show;
+ p->cached_event = ~0ULL;
return 0;
--
1.7.10.4
next prev parent reply other threads:[~2014-03-05 3:51 UTC|newest]
Thread overview: 18+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-03-05 3:47 [PATCHES][RFC][CFT] scalability fixes for shitloads of mounts Al Viro
2014-03-05 3:49 ` [PATCH 1/5][RFC][CFT] percpu fixes, part 1 Al Viro
2014-03-06 19:20 ` Tejun Heo
2014-03-06 20:30 ` Al Viro
2014-03-06 20:47 ` Tejun Heo
2014-03-07 2:52 ` Al Viro
2014-03-07 12:30 ` Tejun Heo
2014-03-14 18:45 ` Al Viro
2014-03-14 18:47 ` Tejun Heo
2014-03-14 18:53 ` Al Viro
2014-03-17 20:12 ` [PATCH percpu/for-3.15] percpu: allocation size should be even Tejun Heo
2014-03-05 3:50 ` [PATCH 2/5][RFC][CFT] fold pcpu_split_block() into the only caller Al Viro
2014-03-06 19:21 ` Tejun Heo
2014-03-05 3:51 ` [PATCH 3/5][RFC][CFT] smarter propagate_mnt() Al Viro
2014-03-05 3:51 ` Al Viro [this message]
2014-03-05 3:52 ` [PATCH 5/5][RFC][CFT] resizable namespace.c hashes Al Viro
2014-03-07 17:17 ` Andi Kleen
2014-03-07 18:38 ` Al Viro
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=20140305035147.GD26528@ZenIV.linux.org.uk \
--to=viro@zeniv.linux.org.uk \
--cc=jeder@redhat.com \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=sct@redhat.com \
--cc=torvalds@linux-foundation.org \
/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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).