From: Andrew Morton <akpm@linux-foundation.org>
To: Tom Zanussi <zanussi@comcast.net>
Cc: a.p.zijlstra@chello.nl, prasad@linux.vnet.ibm.com,
mbligh@google.com, linux-kernel@vger.kernel.org,
torvalds@linux-foundation.org, tglx@linutronix.de,
compudj@krystal.dyndns.org, rostedt@goodmis.org, od@novell.com,
fche@redhat.com, hch@lst.de, dwilder@us.ibm.com
Subject: Re: [PATCH 1/3] relay - clean up subbuf switch
Date: Tue, 23 Sep 2008 13:15:48 -0700 [thread overview]
Message-ID: <20080923131548.e3ac1533.akpm@linux-foundation.org> (raw)
In-Reply-To: <1222147622.6875.136.camel@charm-linux>
On Tue, 23 Sep 2008 00:27:02 -0500
Tom Zanussi <zanussi@comcast.net> wrote:
> Clean up relay_switch_subbuf() and make waking up consumers optional.
>
> Over time, relay_switch_subbuf() has accumulated some cruft - this
> patch cleans it up and at the same time makes available some of it
> available as common functions that any subbuf-switch implementor would
> need (this is partially in preparation for the next patch, which makes
> the subbuf-switch function completely replaceable). It also removes
> the hard-coded reader wakeup and moves it into a replaceable callback
> called notify_consumers(); this allows any given tracer to implement
> consumer notification as it sees fit.
>
> Signed-off-by: Tom Zanussi <tzanussi@gmail.com>
>
> diff --git a/include/linux/relay.h b/include/linux/relay.h
> index 953fc05..17f0515 100644
> --- a/include/linux/relay.h
> +++ b/include/linux/relay.h
> @@ -159,6 +159,15 @@ struct rchan_callbacks
> * The callback should return 0 if successful, negative if not.
> */
> int (*remove_buf_file)(struct dentry *dentry);
> +
> + /*
> + * notify_consumers - new sub-buffer available, let consumers know
> + * @buf: the channel buffer
> + *
> + * Called during sub-buffer switch. Applications which don't
> + * want to notify anyone should implement an empty version.
> + */
> + void (*notify_consumers)(struct rchan_buf *buf);
> };
Does this comment format and placement get properly processed by the
kerneldoc tools?
> /*
> @@ -186,6 +195,48 @@ extern size_t relay_switch_subbuf(struct rchan_buf *buf,
> size_t length);
>
> /**
> + * relay_event_toobig - is event too big to fit in a sub-buffer?
> + * @buf: relay channel buffer
> + * @length: length of event
> + *
> + * Returns 1 if too big, 0 otherwise.
> + *
> + * switch_subbuf() helper function.
> + */
> +static inline int relay_event_toobig(struct rchan_buf *buf, size_t length)
> +{
> + return length > buf->chan->subbuf_size;
> +}
> +
> +/**
> + * relay_update_filesize - increase relay file i_size by length
> + * @buf: relay channel buffer
> + * @length: length to add
> + *
> + * switch_subbuf() helper function.
> + */
> +static inline void relay_update_filesize(struct rchan_buf *buf, size_t length)
> +{
> + if (buf->dentry)
> + buf->dentry->d_inode->i_size += length;
> + else
> + buf->early_bytes += length;
> +
> + smp_mb();
> +}
What locking protects the non-atomic modification of the 64-bit i_size
here?
> +/**
> + * relay_inc_produced - increase number of sub-buffers produced by 1
> + * @buf: relay channel buffer
> + *
> + * switch_subbuf() helper function.
> + */
> +static inline void relay_inc_produced(struct rchan_buf *buf)
> +{
> + buf->subbufs_produced++;
> +}
This also needs caller-provided locking. That's part of the function's
interface and should be documented,
> +/**
> * relay_write - write data into the channel
> * @chan: relay channel
> * @data: data to be written
> diff --git a/kernel/relay.c b/kernel/relay.c
> index 8d13a78..53652f1 100644
> --- a/kernel/relay.c
> +++ b/kernel/relay.c
> @@ -324,6 +324,21 @@ static int remove_buf_file_default_callback(struct dentry *dentry)
> return -EINVAL;
> }
>
> +/*
> + * notify_consumers() default callback.
> + */
> +static void notify_consumers_default_callback(struct rchan_buf *buf)
> +{
> + if (waitqueue_active(&buf->read_wait))
> + /*
> + * Calling wake_up_interruptible() from here
> + * will deadlock if we happen to be logging
> + * from the scheduler (trying to re-grab
> + * rq->lock), so defer it.
> + */
> + __mod_timer(&buf->timer, jiffies + 1);
> +}
> +
> /* relay channel default callbacks */
> static struct rchan_callbacks default_channel_callbacks = {
> .subbuf_start = subbuf_start_default_callback,
> @@ -331,6 +346,7 @@ static struct rchan_callbacks default_channel_callbacks = {
> .buf_unmapped = buf_unmapped_default_callback,
> .create_buf_file = create_buf_file_default_callback,
> .remove_buf_file = remove_buf_file_default_callback,
> + .notify_consumers = notify_consumers_default_callback,
> };
>
> /**
> @@ -508,6 +524,8 @@ static void setup_callbacks(struct rchan *chan,
> cb->create_buf_file = create_buf_file_default_callback;
> if (!cb->remove_buf_file)
> cb->remove_buf_file = remove_buf_file_default_callback;
> + if (!cb->notify_consumers)
> + cb->notify_consumers = notify_consumers_default_callback;
> chan->cb = cb;
> }
>
> @@ -726,30 +744,17 @@ size_t relay_switch_subbuf(struct rchan_buf *buf, size_t length)
> void *old, *new;
> size_t old_subbuf, new_subbuf;
>
> - if (unlikely(length > buf->chan->subbuf_size))
> + if (unlikely(relay_event_toobig(buf, length)))
> goto toobig;
>
> if (buf->offset != buf->chan->subbuf_size + 1) {
> buf->prev_padding = buf->chan->subbuf_size - buf->offset;
> old_subbuf = buf->subbufs_produced % buf->chan->n_subbufs;
> buf->padding[old_subbuf] = buf->prev_padding;
> - buf->subbufs_produced++;
> - if (buf->dentry)
> - buf->dentry->d_inode->i_size +=
> - buf->chan->subbuf_size -
> - buf->padding[old_subbuf];
> - else
> - buf->early_bytes += buf->chan->subbuf_size -
> - buf->padding[old_subbuf];
> - smp_mb();
> - if (waitqueue_active(&buf->read_wait))
> - /*
> - * Calling wake_up_interruptible() from here
> - * will deadlock if we happen to be logging
> - * from the scheduler (trying to re-grab
> - * rq->lock), so defer it.
> - */
> - __mod_timer(&buf->timer, jiffies + 1);
> + relay_inc_produced(buf);
> + relay_update_filesize(buf, buf->chan->subbuf_size -
> + buf->padding[old_subbuf]);
> + buf->chan->cb->notify_consumers(buf);
> }
>
> old = buf->data;
> @@ -763,7 +768,7 @@ size_t relay_switch_subbuf(struct rchan_buf *buf, size_t length)
> buf->data = new;
> buf->padding[new_subbuf] = 0;
>
> - if (unlikely(length + buf->offset > buf->chan->subbuf_size))
> + if (unlikely(relay_event_toobig(buf, length + buf->offset)))
> goto toobig;
I think you can put the unlikely() into relay_event_toobig() and gcc
will dtrt. If that has any value.
> return length;
>
next prev parent reply other threads:[~2008-09-23 20:17 UTC|newest]
Thread overview: 122+ messages / expand[flat|nested] mbox.gz Atom feed top
2008-09-19 21:33 Unified tracing buffer Martin Bligh
2008-09-19 21:42 ` Randy Dunlap
2008-09-19 21:57 ` Martin Bligh
2008-09-19 22:41 ` Olaf Dabrunz
2008-09-19 22:19 ` Martin Bligh
2008-09-20 8:10 ` Olaf Dabrunz
2008-09-20 8:29 ` Steven Rostedt
2008-09-20 11:40 ` Mathieu Desnoyers
2008-09-20 8:26 ` Steven Rostedt
2008-09-20 11:44 ` Mathieu Desnoyers
2008-09-19 22:28 ` Olaf Dabrunz
2008-09-19 22:09 ` Martin Bligh
2008-09-19 23:18 ` Frank Ch. Eigler
2008-09-20 8:50 ` Steven Rostedt
2008-09-20 13:37 ` Mathieu Desnoyers
2008-09-20 13:51 ` Steven Rostedt
2008-09-20 14:54 ` Steven Rostedt
2008-09-22 18:45 ` Mathieu Desnoyers
2008-09-22 21:39 ` Steven Rostedt
2008-09-23 3:27 ` Mathieu Desnoyers
2008-09-20 0:07 ` Peter Zijlstra
2008-09-22 14:07 ` K.Prasad
2008-09-22 14:45 ` Peter Zijlstra
2008-09-22 16:29 ` Martin Bligh
2008-09-22 16:36 ` Peter Zijlstra
2008-09-22 20:50 ` Masami Hiramatsu
2008-09-23 3:05 ` Mathieu Desnoyers
2008-09-23 2:49 ` Mathieu Desnoyers
2008-09-23 5:25 ` Tom Zanussi
2008-09-23 9:31 ` Peter Zijlstra
2008-09-23 18:13 ` Mathieu Desnoyers
2008-09-23 18:33 ` Christoph Lameter
2008-09-23 18:56 ` Linus Torvalds
2008-09-23 13:50 ` Mathieu Desnoyers
2008-09-23 14:00 ` Martin Bligh
2008-09-23 17:55 ` K.Prasad
2008-09-23 18:27 ` Martin Bligh
2008-09-24 3:50 ` Tom Zanussi
2008-09-24 5:42 ` K.Prasad
2008-09-25 6:07 ` [RFC PATCH 0/8] current relay cleanup patchset Tom Zanussi
2008-09-25 6:07 ` [RFC PATCH 1/8] relay - Clean up relay_switch_subbuf() and make waking up consumers optional Tom Zanussi
2008-09-25 6:07 ` [RFC PATCH 2/8] relay - Make the relay sub-buffer switch code replaceable Tom Zanussi
2008-09-25 6:07 ` [RFC PATCH 3/8] relay - Add channel flags to relay, remove global callback param Tom Zanussi
2008-09-25 6:07 ` [RFC PATCH 4/8] relay - Add reserved param to switch-subbuf, in preparation for non-pad write/reserve Tom Zanussi
2008-09-25 6:07 ` [RFC PATCH 5/8] relay - Map the first sub-buffer at the end of the buffer, for temporary convenience Tom Zanussi
2008-09-25 6:07 ` [RFC PATCH 6/8] relay - Replace relay_reserve/relay_write with non-padded versions Tom Zanussi
2008-09-25 6:07 ` [RFC PATCH 7/8] relay - Remove padding-related code from relay_read()/relay_splice_read() et al Tom Zanussi
2008-09-25 6:08 ` [RFC PATCH 8/8] relay - Clean up remaining padding-related junk Tom Zanussi
2008-09-23 5:27 ` [PATCH 1/3] relay - clean up subbuf switch Tom Zanussi
2008-09-23 20:15 ` Andrew Morton [this message]
2008-09-23 5:27 ` [PATCH 2/3] relay - make subbuf switch replaceable Tom Zanussi
2008-09-23 20:17 ` Andrew Morton
2008-09-23 5:27 ` [PATCH 3/3] relay - add channel flags Tom Zanussi
2008-09-23 20:20 ` Andrew Morton
2008-09-24 3:57 ` Tom Zanussi
2008-09-20 0:26 ` Unified tracing buffer Marcel Holtmann
2008-09-20 9:03 ` Steven Rostedt
2008-09-20 13:55 ` Mathieu Desnoyers
2008-09-20 14:12 ` Arjan van de Ven
2008-09-22 18:52 ` Mathieu Desnoyers
2008-10-02 15:28 ` Jason Baron
2008-10-03 16:11 ` Mathieu Desnoyers
2008-10-03 18:37 ` Jason Baron
2008-10-03 19:10 ` Mathieu Desnoyers
2008-10-03 19:25 ` Jason Baron
2008-10-03 19:56 ` Mathieu Desnoyers
2008-10-03 20:25 ` Jason Baron
2008-10-03 21:52 ` Frank Ch. Eigler
2008-09-22 3:09 ` KOSAKI Motohiro
2008-09-22 9:57 ` Peter Zijlstra
2008-09-23 2:36 ` Mathieu Desnoyers
2008-09-22 13:57 ` K.Prasad
2008-09-22 19:45 ` Masami Hiramatsu
2008-09-22 20:13 ` Martin Bligh
2008-09-22 22:25 ` Masami Hiramatsu
2008-09-22 23:11 ` Darren Hart
2008-09-23 0:04 ` Masami Hiramatsu
2008-09-22 23:16 ` Martin Bligh
2008-09-23 0:05 ` Masami Hiramatsu
2008-09-23 0:12 ` Martin Bligh
2008-09-23 14:49 ` Masami Hiramatsu
2008-09-23 15:04 ` Mathieu Desnoyers
2008-09-23 15:30 ` Masami Hiramatsu
2008-09-23 16:01 ` Linus Torvalds
2008-09-23 17:04 ` Masami Hiramatsu
2008-09-23 17:30 ` Thomas Gleixner
2008-09-23 18:59 ` Masami Hiramatsu
2008-09-23 19:36 ` Thomas Gleixner
2008-09-23 19:38 ` Martin Bligh
2008-09-23 19:41 ` Thomas Gleixner
2008-09-23 19:50 ` Martin Bligh
2008-09-23 20:03 ` Thomas Gleixner
2008-09-23 21:02 ` Martin Bligh
2008-09-23 20:03 ` Masami Hiramatsu
2008-09-23 20:08 ` Thomas Gleixner
2008-09-23 15:46 ` Linus Torvalds
2008-09-23 0:39 ` Linus Torvalds
2008-09-23 1:26 ` Roland Dreier
2008-09-23 1:39 ` Steven Rostedt
2008-09-23 2:02 ` Mathieu Desnoyers
2008-09-23 2:26 ` Darren Hart
2008-09-23 2:31 ` Mathieu Desnoyers
2008-09-23 3:26 ` Linus Torvalds
2008-09-23 3:36 ` Mathieu Desnoyers
2008-09-23 4:05 ` Linus Torvalds
2008-09-23 3:43 ` Steven Rostedt
2008-09-23 4:10 ` Masami Hiramatsu
2008-09-23 4:17 ` Martin Bligh
2008-09-23 15:23 ` Masami Hiramatsu
2008-09-23 10:53 ` Steven Rostedt
2008-09-23 4:19 ` Linus Torvalds
2008-09-23 14:12 ` Mathieu Desnoyers
2008-09-23 2:30 ` Mathieu Desnoyers
2008-09-23 3:06 ` Masami Hiramatsu
2008-09-23 14:36 ` KOSAKI Motohiro
2008-09-23 15:02 ` Frank Ch. Eigler
2008-09-23 15:21 ` Masami Hiramatsu
2008-09-23 17:59 ` KOSAKI Motohiro
2008-09-23 18:28 ` Martin Bligh
2008-09-23 3:33 ` Andi Kleen
2008-09-23 3:47 ` Martin Bligh
2008-09-23 5:04 ` Andi Kleen
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=20080923131548.e3ac1533.akpm@linux-foundation.org \
--to=akpm@linux-foundation.org \
--cc=a.p.zijlstra@chello.nl \
--cc=compudj@krystal.dyndns.org \
--cc=dwilder@us.ibm.com \
--cc=fche@redhat.com \
--cc=hch@lst.de \
--cc=linux-kernel@vger.kernel.org \
--cc=mbligh@google.com \
--cc=od@novell.com \
--cc=prasad@linux.vnet.ibm.com \
--cc=rostedt@goodmis.org \
--cc=tglx@linutronix.de \
--cc=torvalds@linux-foundation.org \
--cc=zanussi@comcast.net \
/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
Powered by JetHome