From: Alan Stern <stern@rowland.harvard.edu>
To: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Oleg Nesterov <oleg@tv-sign.ru>,
Kernel development list <linux-kernel@vger.kernel.org>
Subject: Re: [patch] cpufreq: mark cpufreq_tsc() as core_initcall_sync
Date: Tue, 21 Nov 2006 12:56:21 -0500 (EST) [thread overview]
Message-ID: <Pine.LNX.4.44L0.0611211244200.6140-100000@iolanthe.rowland.org> (raw)
In-Reply-To: <20061119214315.GI4427@us.ibm.com>
Here's another potential problem with the fast path approach. It's not
very serious, but you might want to keep it in mind.
The idea is that a reader can start up on one CPU and finish on another,
and a writer might see the finish event but not the start event. For
example:
Reader A enters the critical section on CPU 0 and starts
accessing the old data area.
Writer B updates the data pointer and starts executing
srcu_readers_active_idx() to check if the fast path can be
used. It sees per_cpu_ptr(0)->c[idx] == 1 because of
Reader A.
Reader C runs srcu_read_lock() on CPU 0, setting
per_cpu_ptr[0]->c[idx] to 2.
Reader C migrates to CPU 1 and leaves the critical section;
srcu_read_unlock() sets per_cpu_ptr(1)->c[idx] to -1.
Writer B finishes the cpu loop in srcu_readers_active_idx(),
seeing per_cpu_ptr(1)->c[idx] == -1. It computes sum =
1 + -1 == 0, takes the fast path, and exits immediately
from synchronize_srcu().
Writer B deallocates the old data area while Reader A is still
using it.
This requires two context switches to take place while the cpu loop in
srcu_readers_active_idx() runs, so perhaps it isn't realistic. Is it
worth worrying about?
Alan
next prev parent reply other threads:[~2006-11-21 17:56 UTC|newest]
Thread overview: 108+ messages / expand[flat|nested] mbox.gz Atom feed top
2006-11-16 20:00 BUG: cpufreq notification broken Thomas Gleixner
2006-11-16 20:15 ` [patch] cpufreq: mark cpufreq_tsc() as core_initcall_sync Ingo Molnar
2006-11-16 20:45 ` Thomas Gleixner
2006-11-16 21:47 ` Linus Torvalds
2006-11-16 22:03 ` Alan Stern
2006-11-16 22:21 ` Linus Torvalds
2006-11-17 3:06 ` Alan Stern
2006-11-17 6:51 ` Paul E. McKenney
2006-11-17 9:29 ` Jens Axboe
2006-11-17 18:39 ` Oleg Nesterov
2006-11-18 0:28 ` Paul E. McKenney
2006-11-18 16:15 ` Alan Stern
2006-11-18 17:14 ` Paul E. McKenney
2006-11-18 19:34 ` Oleg Nesterov
2006-11-19 21:26 ` Paul E. McKenney
2006-11-18 21:00 ` Alan Stern
2006-11-18 21:25 ` Oleg Nesterov
2006-11-18 22:13 ` Alan Stern
2006-11-18 22:46 ` Oleg Nesterov
2006-11-19 20:12 ` Alan Stern
2006-11-19 21:43 ` Paul E. McKenney
2006-11-20 17:19 ` Alan Stern
2006-11-20 17:58 ` Jens Axboe
2006-11-20 19:39 ` Alan Stern
2006-11-20 20:13 ` Jens Axboe
2006-11-20 21:39 ` Alan Stern
2006-11-21 7:39 ` Jens Axboe
2006-11-20 18:57 ` Oleg Nesterov
2006-11-20 20:01 ` Alan Stern
2006-11-20 20:51 ` Paul E. McKenney
2006-11-21 20:04 ` Oleg Nesterov
2006-11-21 20:54 ` Alan Stern
2006-11-21 22:07 ` Paul E. McKenney
2006-11-20 20:38 ` Paul E. McKenney
2006-11-21 16:44 ` Oleg Nesterov
2006-11-21 19:56 ` Paul E. McKenney
2006-11-21 20:26 ` Alan Stern
2006-11-21 23:03 ` Paul E. McKenney
2006-11-22 2:17 ` Alan Stern
2006-11-22 17:01 ` Paul E. McKenney
2006-11-26 22:25 ` Oleg Nesterov
2006-11-27 21:10 ` Alan Stern
2006-11-28 1:47 ` Paul E. McKenney
2006-11-20 19:17 ` Paul E. McKenney
2006-11-20 20:22 ` Alan Stern
2006-11-21 17:55 ` Paul E. McKenney
2006-11-21 17:56 ` Alan Stern [this message]
2006-11-21 19:13 ` Paul E. McKenney
2006-11-21 20:40 ` Alan Stern
2006-11-22 18:08 ` Paul E. McKenney
2006-11-21 21:01 ` Oleg Nesterov
2006-11-22 0:51 ` Paul E. McKenney
2006-11-18 18:46 ` Oleg Nesterov
2006-11-19 21:07 ` Paul E. McKenney
2006-11-20 7:15 ` Jens Axboe
2006-11-20 16:59 ` Paul E. McKenney
2006-11-20 17:55 ` Jens Axboe
2006-11-20 20:09 ` Paul E. McKenney
2006-11-20 20:21 ` Jens Axboe
2006-11-18 19:02 ` Oleg Nesterov
2006-11-19 21:23 ` Paul E. McKenney
2006-11-17 19:15 ` Paul E. McKenney
2006-11-17 19:27 ` Alan Stern
2006-11-18 0:38 ` Paul E. McKenney
2006-11-18 4:33 ` Alan Stern
2006-11-18 4:51 ` Andrew Morton
2006-11-18 5:57 ` Paul E. McKenney
2006-11-19 19:00 ` Oleg Nesterov
2006-11-19 20:21 ` Alan Stern
2006-11-19 20:55 ` Oleg Nesterov
2006-11-19 21:09 ` Alan Stern
2006-11-19 21:17 ` Oleg Nesterov
2006-11-19 21:54 ` Paul E. McKenney
2006-11-19 22:28 ` Oleg Nesterov
2006-11-20 5:47 ` Paul E. McKenney
2006-11-19 21:20 ` Paul E. McKenney
2006-11-19 21:50 ` Oleg Nesterov
2006-11-19 22:04 ` Paul E. McKenney
2006-11-23 14:59 ` Oleg Nesterov
2006-11-23 20:40 ` Paul E. McKenney
2006-11-23 21:34 ` Oleg Nesterov
2006-11-23 21:49 ` Oleg Nesterov
2006-11-27 4:33 ` Paul E. McKenney
2006-11-24 18:21 ` Oleg Nesterov
2006-11-24 20:04 ` Jens Axboe
2006-11-24 20:47 ` Alan Stern
2006-11-24 21:13 ` Oleg Nesterov
2006-11-25 3:24 ` Alan Stern
2006-11-25 17:14 ` Oleg Nesterov
2006-11-25 22:06 ` Alan Stern
2006-11-26 21:34 ` Oleg Nesterov
2006-11-27 5:02 ` Paul E. McKenney
2006-11-27 16:11 ` Oleg Nesterov
2006-11-27 16:56 ` Oleg Nesterov
2006-11-29 19:29 ` Paul E. McKenney
2006-11-29 20:16 ` Oleg Nesterov
2006-11-29 23:08 ` Paul E. McKenney
2006-11-30 0:01 ` Oleg Nesterov
2006-11-17 2:33 ` Paul E. McKenney
2006-11-16 20:52 ` Peter Zijlstra
2006-11-16 21:20 ` Andrew Morton
2006-11-16 21:27 ` Thomas Gleixner
2006-11-20 19:57 ` Linus Torvalds
2006-11-16 20:27 ` BUG: cpufreq notification broken Alan Stern
2006-11-16 21:09 ` Thomas Gleixner
2006-11-16 21:26 ` Alan Stern
2006-11-16 21:36 ` Thomas Gleixner
2006-11-16 21:56 ` Alan Stern
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=Pine.LNX.4.44L0.0611211244200.6140-100000@iolanthe.rowland.org \
--to=stern@rowland.harvard.edu \
--cc=linux-kernel@vger.kernel.org \
--cc=oleg@tv-sign.ru \
--cc=paulmck@linux.vnet.ibm.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.