From: Michal Hocko <mhocko@kernel.org>
To: Davidlohr Bueso <dave@stgolabs.net>
Cc: LKML <linux-kernel@vger.kernel.org>,
Peter Zijlstra <peterz@infradead.org>,
Ingo Molnar <mingo@redhat.com>,
Thomas Gleixner <tglx@linutronix.de>,
"H. Peter Anvin" <hpa@zytor.com>,
"David S. Miller" <davem@davemloft.net>,
Tony Luck <tony.luck@intel.com>,
Andrew Morton <akpm@linux-foundation.org>,
Chris Zankel <chris@zankel.net>,
Max Filippov <jcmvbkbc@gmail.com>,
x86@kernel.org, linux-alpha@vger.kernel.org,
linux-ia64@vger.kernel.org, linux-s390@vger.kernel.org,
linux-sh@vger.kernel.org, sparclinux@vger.kernel.org,
linux-xtensa@linux-xtensa.org, linux-arch@vger.kernel.org
Subject: Re: [PATCH 03/11] locking, rwsem: introduce basis for down_write_killable
Date: Mon, 4 Apr 2016 11:17:00 +0200 [thread overview]
Message-ID: <20160404091659.GA13463@dhcp22.suse.cz> (raw)
In-Reply-To: <20160402044125.GC5329@linux-uzut.site>
On Fri 01-04-16 21:41:25, Davidlohr Bueso wrote:
> On Fri, 01 Apr 2016, Michal Hocko wrote:
>
> >From: Michal Hocko <mhocko@suse.com>
> >
> >Introduce a generic implementation necessary for down_write_killable.
> >This is a trivial extension of the already existing down_write call
> >which can be interrupted by SIGKILL. This patch doesn't provide
> >down_write_killable yet because arches have to provide the necessary
> >pieces before.
> >
> >rwsem_down_write_failed which is a generic slow path for the
> >write lock is extended to allow a task state and renamed to
> >__rwsem_down_write_failed_state. The return value is either a valid
> >semaphore pointer or ERR_PTR(-EINTR).
> >
> >rwsem_down_write_failed_killable is exported as a new way to wait for
> >the lock and be killable.
> >
> >For rwsem-spinlock implementation the current __down_write it updated
> >in a similar way as __rwsem_down_write_failed_state except it doesn't
> >need new exports just visible __down_write_killable.
> >
> >Architectures which are not using the generic rwsem implementation are
> >supposed to provide their __down_write_killable implementation and
> >use rwsem_down_write_failed_killable for the slow path.
>
> So in a nutshell, this is supposed to be the (writer) rwsem counterpart of
> mutex_lock_killable() and down_killable(), right?
Yes.
> [...]
>
> >--- a/kernel/locking/rwsem-xadd.c
> >+++ b/kernel/locking/rwsem-xadd.c
> >@@ -433,12 +433,13 @@ static inline bool rwsem_has_spinner(struct rw_semaphore *sem)
> >/*
> > * Wait until we successfully acquire the write lock
> > */
> >-__visible
> >-struct rw_semaphore __sched *rwsem_down_write_failed(struct rw_semaphore *sem)
> >+static inline struct rw_semaphore *
> >+__rwsem_down_write_failed_state(struct rw_semaphore *sem, int state)
>
> fwiw I'm not a fan of the _state naming. While I understand why you chose it, I feel
> it does not really describe the purpose of the call itself. The state logic alone is
> really quite small and therefore should not govern the function name. Why not just apply
> kiss and label things _common, ie like mutexes do? This would also standardize names a
> bit.
I really do not care much about naming. So if _common sounds better I
can certainly rename.
>
> >{
> > long count;
> > bool waiting = true; /* any queued threads before us */
> > struct rwsem_waiter waiter;
> >+ struct rw_semaphore *ret = sem;
> >
> > /* undo write bias from down_write operation, stop active locking */
> > count = rwsem_atomic_update(-RWSEM_ACTIVE_WRITE_BIAS, sem);
> >@@ -478,7 +479,7 @@ struct rw_semaphore __sched *rwsem_down_write_failed(struct rw_semaphore *sem)
> > count = rwsem_atomic_update(RWSEM_WAITING_BIAS, sem);
> >
> > /* wait until we successfully acquire the lock */
> >- set_current_state(TASK_UNINTERRUPTIBLE);
> >+ set_current_state(state);
> > while (true) {
> > if (rwsem_try_write_lock(count, sem))
> > break;
> >@@ -486,21 +487,39 @@ struct rw_semaphore __sched *rwsem_down_write_failed(struct rw_semaphore *sem)
> >
> > /* Block until there are no active lockers. */
> > do {
> >+ if (signal_pending_state(state, current)) {
>
> ^^ unlikely()?
The generated code is identical after I've added unlikely. I haven't
tried more gcc versions (mine is 5.3.1) but is this worth it?
>
> >+ raw_spin_lock_irq(&sem->wait_lock);
>
> If the lock is highly contended + a bad workload for spin-on-owner, this could take a while :)
> Of course, this is a side effect of the wait until no active lockers optimization which avoids
> the wait_lock in the first place, so fortunately it somewhat mitigates the situation.
>
> >+ ret = ERR_PTR(-EINTR);
> >+ goto out;
> >+ }
> > schedule();
> >- set_current_state(TASK_UNINTERRUPTIBLE);
> >+ set_current_state(state);
> > } while ((count = sem->count) & RWSEM_ACTIVE_MASK);
> >
> > raw_spin_lock_irq(&sem->wait_lock);
> > }
> >+out:
> > __set_current_state(TASK_RUNNING);
> >-
>
> You certainly don't want this iff exiting due to TASK_KILLABLE situation.
Not sure I got your point here.
--
Michal Hocko
SUSE Labs
next prev parent reply other threads:[~2016-04-04 9:17 UTC|newest]
Thread overview: 55+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-04-01 11:04 [PATCH 0/11] introduce down_write_killable for rw_semaphore v2 Michal Hocko
2016-04-01 11:04 ` [PATCH 01/11] locking, rwsem: get rid of __down_write_nested Michal Hocko
2016-04-02 0:28 ` Davidlohr Bueso
2016-04-01 11:04 ` [PATCH 02/11] locking, rwsem: drop explicit memory barriers Michal Hocko
2016-04-02 1:17 ` Davidlohr Bueso
2016-04-04 9:03 ` Michal Hocko
2016-04-04 9:06 ` [PATCH 1/2] xtensa, rwsem: drop superfluous arch specific implementation Michal Hocko
2016-04-04 9:06 ` [PATCH 2/2] sh, " Michal Hocko
2016-04-06 9:26 ` Peter Zijlstra
2016-04-06 9:50 ` Geert Uytterhoeven
2016-04-06 10:27 ` Peter Zijlstra
2016-04-04 10:23 ` [PATCH 1/2] xtensa, " Max Filippov
2016-04-06 9:06 ` [PATCH] sparc, " Michal Hocko
2016-04-01 11:04 ` [PATCH 03/11] locking, rwsem: introduce basis for down_write_killable Michal Hocko
2016-04-02 4:41 ` Davidlohr Bueso
2016-04-04 9:17 ` Michal Hocko [this message]
2016-04-04 9:21 ` Peter Zijlstra
2016-04-07 6:58 ` Davidlohr Bueso
2016-04-07 7:38 ` Michal Hocko
2016-05-10 10:43 ` Tetsuo Handa
2016-05-10 11:53 ` Michal Hocko
2016-05-10 12:38 ` Peter Zijlstra
2016-05-10 13:57 ` Tetsuo Handa
2016-05-11 7:23 ` Michal Hocko
2016-05-11 8:28 ` Michal Hocko
2016-05-11 8:44 ` Peter Zijlstra
2016-05-11 9:04 ` Michal Hocko
2016-05-11 9:17 ` Peter Zijlstra
2016-05-11 9:31 ` Michal Hocko
2016-05-11 9:41 ` Peter Zijlstra
2016-05-11 13:59 ` Michal Hocko
2016-05-11 18:03 ` Michal Hocko
2016-05-12 11:57 ` [PATCH] locking, rwsem: Fix down_write_killable() Peter Zijlstra
2016-05-12 12:15 ` [tip:locking/rwsem] locking/rwsem: " tip-bot for Peter Zijlstra
2016-05-12 16:59 ` [PATCH] locking, rwsem: " Michal Hocko
2016-05-15 20:57 ` [tip:locking/rwsem] locking/rwsem: " tip-bot for Peter Zijlstra
2016-05-12 12:12 ` [PATCH 03/11] locking, rwsem: introduce basis for down_write_killable Peter Zijlstra
2016-05-12 12:19 ` Michal Hocko
2016-05-12 13:58 ` Peter Zijlstra
2016-05-12 19:42 ` Waiman Long
2016-05-11 8:35 ` Peter Zijlstra
2016-05-11 9:02 ` Michal Hocko
2016-04-01 11:04 ` [PATCH 04/11] alpha, rwsem: provide __down_write_killable Michal Hocko
2016-04-01 11:04 ` [PATCH 05/11] ia64, " Michal Hocko
2016-04-01 11:04 ` [PATCH 06/11] s390, " Michal Hocko
2016-04-01 11:04 ` [PATCH 07/11] sh, " Michal Hocko
2016-04-01 11:04 ` [PATCH 08/11] sparc, " Michal Hocko
2016-04-01 11:04 ` [PATCH 09/11] xtensa, " Michal Hocko
2016-04-01 11:04 ` [PATCH 10/11] x86, " Michal Hocko
2016-04-06 18:31 ` Peter Zijlstra
2016-04-01 11:04 ` [PATCH 11/11] locking, rwsem: provide down_write_killable Michal Hocko
-- strict thread matches above, loose matches on Subject: below --
2016-02-29 12:58 [PATCH 0/11] introduce down_write_killable for rw_semaphore Michal Hocko
2016-02-29 12:58 ` [PATCH 03/11] locking, rwsem: introduce basis for down_write_killable Michal Hocko
2016-03-30 13:25 ` Peter Zijlstra
2016-03-31 8:33 ` Michal Hocko
2016-03-31 8:44 ` Peter Zijlstra
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=20160404091659.GA13463@dhcp22.suse.cz \
--to=mhocko@kernel.org \
--cc=akpm@linux-foundation.org \
--cc=chris@zankel.net \
--cc=dave@stgolabs.net \
--cc=davem@davemloft.net \
--cc=hpa@zytor.com \
--cc=jcmvbkbc@gmail.com \
--cc=linux-alpha@vger.kernel.org \
--cc=linux-arch@vger.kernel.org \
--cc=linux-ia64@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-s390@vger.kernel.org \
--cc=linux-sh@vger.kernel.org \
--cc=linux-xtensa@linux-xtensa.org \
--cc=mingo@redhat.com \
--cc=peterz@infradead.org \
--cc=sparclinux@vger.kernel.org \
--cc=tglx@linutronix.de \
--cc=tony.luck@intel.com \
--cc=x86@kernel.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).