From: Jason Baron <jbaron@redhat.com>
To: Frederic Weisbecker <fweisbec@gmail.com>
Cc: peterz@infradead.org, mathieu.desnoyers@polymtl.ca,
hpa@zytor.com, rostedt@goodmis.org, mingo@elte.hu,
tglx@linutronix.de, andi@firstfloor.org, roland@redhat.com,
rth@redhat.com, masami.hiramatsu.pt@hitachi.com, avi@redhat.com,
davem@davemloft.net, sam@ravnborg.org, ddaney@caviumnetworks.com,
michael@ellerman.id.au, linux-kernel@vger.kernel.org
Subject: Re: [PATCH 2/2] jump label: introduce static_branch()
Date: Wed, 5 Jan 2011 16:14:44 -0500 [thread overview]
Message-ID: <20110105211444.GD2896@redhat.com> (raw)
In-Reply-To: <20110105171516.GB1692@nowhere>
On Wed, Jan 05, 2011 at 06:15:18PM +0100, Frederic Weisbecker wrote:
> On Wed, Jan 05, 2011 at 10:43:12AM -0500, Jason Baron wrote:
> > diff --git a/include/linux/jump_label.h b/include/linux/jump_label.h
> > index 152f7de..0ad9c2e 100644
> > --- a/include/linux/jump_label.h
> > +++ b/include/linux/jump_label.h
> > @@ -22,6 +22,11 @@ struct module;
> >
> > #ifdef HAVE_JUMP_LABEL
> >
> > +static __always_inline bool static_branch(struct jump_label_key *key)
> > +{
> > + return __static_branch(key);
>
> Not very important, but __static_branch() would be more self-explained
> if it was called arch_static_branch().
>
> > +}
> > +
> > extern struct jump_entry __start___jump_table[];
> > extern struct jump_entry __stop___jump_table[];
> >
> > @@ -42,11 +47,12 @@ struct jump_label_key {
> > int state;
> > };
> >
> > -#define JUMP_LABEL(key, label) \
> > -do { \
> > - if (unlikely(((struct jump_label_key *)key)->state)) \
> > - goto label; \
> > -} while (0)
> > +static __always_inline bool static_branch(struct jump_label_key *key)
> > +{
> > + if (unlikely(key->state))
> > + return true;
> > + return false;
> > +}
> >
> > static inline int jump_label_enabled(struct jump_label_key *key)
> > {
> > @@ -78,14 +84,4 @@ static inline void jump_label_unlock(void) {}
> >
> > #endif
> >
> > -#define COND_STMT(key, stmt) \
> > -do { \
> > - __label__ jl_enabled; \
> > - JUMP_LABEL_ELSE_ATOMIC_READ(key, jl_enabled); \
> > - if (0) { \
> > -jl_enabled: \
> > - stmt; \
> > - } \
> > -} while (0)
> > -
> > #endif
> > diff --git a/include/linux/jump_label_ref.h b/include/linux/jump_label_ref.h
> > index 8a76e89..5178696 100644
> > --- a/include/linux/jump_label_ref.h
> > +++ b/include/linux/jump_label_ref.h
> > @@ -7,19 +7,23 @@
> > struct jump_label_key_counter {
> > atomic_t ref;
> > struct jump_label_key key;
> > -}
> > +};
> >
> > #ifdef HAVE_JUMP_LABEL
> >
> > -#define JUMP_LABEL_ELSE_ATOMIC_READ(key, label, counter) JUMP_LABEL(key, label)
> > +static __always_inline bool static_branch_else_atomic_read(struct jump_label_key *key, atomic_t *count)
> > +{
> > + return __static_branch(key);
> > +}
>
> How about having only static_branch() but the key would be handled only
> by ways of get()/put().
>
> Simple boolean key enablement would work in this scheme as well as branches
> based on refcount. So that the users could avoid maintaining both key and count,
> this would be transparently handled by the jump label API.
>
> Or am I missing something?
>
right. this is a good point. I had a 'jump_label_inc()',
'jump_label_dec()' essentially providing this. However, when jump labels
are disabled we didn't want to incur an atomic_read() everywhere.
Furthermore, the use of the atomic_t type within jump_label.h, causes
#include dependencies problems, since atomic.h ends up including
jump_label.h...
Thus, what I've proposed here, is to the have the very simple
jump_label_enable()/disable(), and leave reference counting to the
caller.
thanks,
-Jason
next prev parent reply other threads:[~2011-01-05 21:15 UTC|newest]
Thread overview: 113+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-01-05 15:43 [PATCH 0/2] jump label: 2.6.38 updates Jason Baron
2011-01-05 15:43 ` [PATCH 1/2] jump label: make enable/disable o(1) Jason Baron
2011-01-05 17:31 ` Steven Rostedt
2011-01-05 21:19 ` Jason Baron
2011-01-05 15:43 ` [PATCH 2/2] jump label: introduce static_branch() Jason Baron
2011-01-05 17:15 ` Frederic Weisbecker
2011-01-05 17:46 ` Steven Rostedt
2011-01-05 18:52 ` H. Peter Anvin
2011-01-05 21:19 ` Jason Baron
2011-01-05 21:14 ` Jason Baron [this message]
2011-01-05 17:32 ` David Daney
2011-01-05 17:43 ` Steven Rostedt
2011-01-05 18:44 ` David Miller
2011-01-05 20:04 ` Steven Rostedt
2011-01-05 18:56 ` H. Peter Anvin
2011-01-05 19:14 ` Ingo Molnar
2011-01-05 19:32 ` David Daney
2011-01-05 19:50 ` Ingo Molnar
2011-01-05 20:07 ` David Daney
2011-01-05 20:08 ` H. Peter Anvin
2011-01-05 20:18 ` Ingo Molnar
2011-01-05 21:16 ` Jason Baron
2011-01-05 17:41 ` Steven Rostedt
2011-01-09 18:48 ` Mathieu Desnoyers
2011-02-11 19:25 ` [PATCH 0/2] jump label: 2.6.38 updates Peter Zijlstra
2011-02-11 21:13 ` Mathieu Desnoyers
[not found] ` <BLU0-SMTP101B686C32E10BA346B15F896EF0@phx.gbl>
2011-02-11 21:38 ` Peter Zijlstra
2011-02-11 22:15 ` Jason Baron
2011-02-11 22:19 ` H. Peter Anvin
2011-02-11 22:30 ` Mathieu Desnoyers
2011-02-11 22:20 ` Mathieu Desnoyers
[not found] ` <BLU0-SMTP8562BA758CF8AAE5323AE296EF0@phx.gbl>
2011-02-11 22:27 ` Jason Baron
2011-02-11 22:32 ` Mathieu Desnoyers
2011-02-12 18:47 ` Peter Zijlstra
2011-02-14 12:27 ` Ingo Molnar
2011-02-14 15:51 ` Jason Baron
2011-02-14 15:57 ` Peter Zijlstra
2011-02-14 16:04 ` Jason Baron
2011-02-14 16:14 ` Mathieu Desnoyers
[not found] ` <BLU0-SMTP4069A1A89F06CDFF9B28F896D00@phx.gbl>
2011-02-14 16:25 ` Peter Zijlstra
2011-02-14 16:29 ` Jason Baron
2011-02-14 16:37 ` Peter Zijlstra
2011-02-14 16:43 ` Mathieu Desnoyers
2011-02-14 16:46 ` Steven Rostedt
2011-02-14 16:53 ` Peter Zijlstra
2011-02-14 17:18 ` Steven Rostedt
2011-02-14 17:23 ` Mike Frysinger
2011-02-14 17:27 ` Peter Zijlstra
2011-02-14 17:29 ` Mike Frysinger
2011-02-14 17:38 ` Peter Zijlstra
2011-02-14 17:45 ` Mike Frysinger
2011-02-14 17:38 ` Will Newton
2011-02-14 17:43 ` Peter Zijlstra
2011-02-14 17:50 ` Will Newton
2011-02-14 18:04 ` Peter Zijlstra
2011-02-14 18:24 ` Peter Zijlstra
2011-02-14 18:53 ` Mathieu Desnoyers
2011-02-14 21:29 ` Steven Rostedt
2011-02-14 21:39 ` Steven Rostedt
2011-02-14 21:46 ` David Miller
2011-02-14 22:20 ` Steven Rostedt
2011-02-14 22:21 ` Steven Rostedt
2011-02-14 22:21 ` H. Peter Anvin
2011-02-14 22:29 ` Mathieu Desnoyers
[not found] ` <BLU0-SMTP98BFCC52FD41661DD9CC1E96D00@phx.gbl>
2011-02-14 22:33 ` David Miller
2011-02-14 22:33 ` David Miller
2011-02-14 22:37 ` Matt Fleming
2011-02-14 23:03 ` Mathieu Desnoyers
[not found] ` <BLU0-SMTP166A8555C791786059B0FF96D00@phx.gbl>
2011-02-14 23:09 ` Paul E. McKenney
2011-02-14 23:29 ` Mathieu Desnoyers
[not found] ` <BLU0-SMTP4599FAAD7330498472B87396D00@phx.gbl>
2011-02-15 0:19 ` Segher Boessenkool
2011-02-15 0:48 ` Mathieu Desnoyers
2011-02-15 1:29 ` Steven Rostedt
[not found] ` <BLU0-SMTP984E876DBDFBC13F4C86F896D00@phx.gbl>
2011-02-15 0:42 ` Paul E. McKenney
2011-02-15 0:51 ` Mathieu Desnoyers
2011-02-15 11:53 ` Will Newton
2011-02-18 19:03 ` Paul E. McKenney
2011-02-14 23:19 ` H. Peter Anvin
2011-02-15 11:01 ` Will Newton
2011-02-15 13:31 ` H. Peter Anvin
2011-02-15 13:49 ` Steven Rostedt
2011-02-15 14:04 ` Will Newton
2011-02-15 21:11 ` Will Simoneau
2011-02-15 21:27 ` David Miller
2011-02-15 21:56 ` Will Simoneau
2011-02-16 10:15 ` Will Newton
2011-02-16 12:18 ` Steven Rostedt
2011-02-16 12:41 ` Will Newton
2011-02-16 13:24 ` Mathieu Desnoyers
2011-02-16 22:51 ` Will Simoneau
2011-02-17 0:53 ` Please watch your cc lists Andi Kleen
2011-02-17 0:56 ` David Miller
2011-02-17 1:04 ` Michael Witten
2011-02-17 10:55 ` [PATCH 0/2] jump label: 2.6.38 updates Will Newton
[not found] ` <BLU0-SMTP80F56386E7E060A3B2020B96D20@phx.gbl>
2011-02-17 1:55 ` Masami Hiramatsu
2011-02-17 3:19 ` H. Peter Anvin
2011-02-17 16:03 ` Mathieu Desnoyers
[not found] ` <BLU0-SMTP71BCB155CBAE79997EE08D96D20@phx.gbl>
2011-02-17 3:36 ` Steven Rostedt
2011-02-17 16:13 ` Mathieu Desnoyers
[not found] ` <BLU0-SMTP51D40A5B1DACA8883D6AB596D50@phx.gbl>
2011-02-17 20:09 ` Steven Rostedt
2011-02-15 22:20 ` Benjamin Herrenschmidt
2011-02-16 8:35 ` Ingo Molnar
2011-02-17 1:04 ` H. Peter Anvin
2011-02-17 12:51 ` Ingo Molnar
[not found] ` <BLU0-SMTP637B2E9372CFBF3A0B5B0996D00@phx.gbl>
2011-02-14 23:25 ` David Miller
2011-02-14 23:34 ` Mathieu Desnoyers
[not found] ` <20110214233405.GC17432@Krystal>
2011-02-14 23:52 ` Mathieu Desnoyers
2011-02-14 22:15 ` Matt Fleming
2011-02-15 15:20 ` Heiko Carstens
[not found] ` <BLU0-SMTP64371A838030ED92A7CCB696D00@phx.gbl>
2011-02-14 18:54 ` Jason Baron
2011-02-14 19:20 ` Peter Zijlstra
2011-02-14 19:48 ` Mathieu Desnoyers
2011-02-14 16:11 ` Mathieu Desnoyers
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=20110105211444.GD2896@redhat.com \
--to=jbaron@redhat.com \
--cc=andi@firstfloor.org \
--cc=avi@redhat.com \
--cc=davem@davemloft.net \
--cc=ddaney@caviumnetworks.com \
--cc=fweisbec@gmail.com \
--cc=hpa@zytor.com \
--cc=linux-kernel@vger.kernel.org \
--cc=masami.hiramatsu.pt@hitachi.com \
--cc=mathieu.desnoyers@polymtl.ca \
--cc=michael@ellerman.id.au \
--cc=mingo@elte.hu \
--cc=peterz@infradead.org \
--cc=roland@redhat.com \
--cc=rostedt@goodmis.org \
--cc=rth@redhat.com \
--cc=sam@ravnborg.org \
--cc=tglx@linutronix.de \
/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.