netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Hiroaki SHIMODA <shimoda.hiroaki@gmail.com>
To: Tom Herbert <therbert@google.com>
Cc: Denys Fedoryshchenko <denys@visp.net.lb>,
	netdev@vger.kernel.org, e1000-devel@lists.sourceforge.net,
	jeffrey.t.kirsher@intel.com, jesse.brandeburg@intel.com,
	eric.dumazet@gmail.com, davem@davemloft.net
Subject: Re: Strange latency spikes/TX network stalls on Sun Fire X4150(x86) and e1000e
Date: Wed, 30 May 2012 09:06:02 +0900	[thread overview]
Message-ID: <20120530090602.6204d857.shimoda.hiroaki@gmail.com> (raw)
In-Reply-To: <CA+mtBx_uYy9XcRvpD2E46FuMBFu38iQvCiwFHWqbhPBmY=JfOg@mail.gmail.com>

While reading the bql code, I have some questions.

1) dql_completed() and dql_queued() can be called concurrently,
   so dql->num_queued could change while processing
   dql_completed().
   Is it intentional to refer num_queued from "dql->" each time ?

2) From the comment in the code
   *   - The queue was over-limit in the previous interval and
   *     when enqueuing it was possible that all queued data
   *     had been consumed.

   and

   * Queue was not starved, check if the limit can be decreased.
   * A decrease is only considered if the queue has been busy in
   * the whole interval (the check above). 

   the calculation of all_prev_completed should take into account
   completed == dql->prev_num_queued case ?
   On current implementation, limit shrinks easily and some NIC
   hit TX stalls.
   To mitigate TX stalls, should we fix all_prev_completed rather
   than individual driver ?

3) limit calculation fails to consider integer wrap around in
   one place ?

Here is the patch what I meant.

diff --git a/lib/dynamic_queue_limits.c b/lib/dynamic_queue_limits.c
@@ -11,22 +11,27 @@
 #include <linux/dynamic_queue_limits.h>
 
 #define POSDIFF(A, B) ((A) > (B) ? (A) - (B) : 0)
+#define POSDIFFI(A, B) ((int)((A) - (B)) > 0 ? (A) - (B) : 0)
+#define AFTER_EQ(A, B) ((int)((A) - (B)) >= 0)
 
 /* Records completed count and recalculates the queue limit */
 void dql_completed(struct dql *dql, unsigned int count)
 {
 	unsigned int inprogress, prev_inprogress, limit;
-	unsigned int ovlimit, all_prev_completed, completed;
+	unsigned int ovlimit, completed, num_queued;
+	bool all_prev_completed;
+
+	num_queued = dql->num_queued;
 
 	/* Can't complete more than what's in queue */
-	BUG_ON(count > dql->num_queued - dql->num_completed);
+	BUG_ON(count > num_queued - dql->num_completed);
 
 	completed = dql->num_completed + count;
 	limit = dql->limit;
-	ovlimit = POSDIFF(dql->num_queued - dql->num_completed, limit);
-	inprogress = dql->num_queued - completed;
+	ovlimit = POSDIFF(num_queued - dql->num_completed, limit);
+	inprogress = num_queued - completed;
 	prev_inprogress = dql->prev_num_queued - dql->num_completed;
-	all_prev_completed = POSDIFF(completed, dql->prev_num_queued);
+	all_prev_completed = AFTER_EQ(completed, dql->prev_num_queued);
 
 	if ((ovlimit && !inprogress) ||
 	    (dql->prev_ovlimit && all_prev_completed)) {
@@ -45,7 +50,7 @@ void dql_completed(struct dql *dql, unsigned int count)
 		 *     of bytes both sent and completed in the last interval,
 		 *     plus any previous over-limit.
 		 */
-		limit += POSDIFF(completed, dql->prev_num_queued) +
+		limit += POSDIFFI(completed, dql->prev_num_queued) +
 		     dql->prev_ovlimit;
 		dql->slack_start_time = jiffies;
 		dql->lowest_slack = UINT_MAX;
@@ -104,7 +109,7 @@ void dql_completed(struct dql *dql, unsigned int count)
 	dql->prev_ovlimit = ovlimit;
 	dql->prev_last_obj_cnt = dql->last_obj_cnt;
 	dql->num_completed = completed;
-	dql->prev_num_queued = dql->num_queued;
+	dql->prev_num_queued = num_queued;
 }
 EXPORT_SYMBOL(dql_completed);

  parent reply	other threads:[~2012-05-30  0:06 UTC|newest]

Thread overview: 52+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-05-15 14:15 Strange latency spikes/TX network stalls on Sun Fire X4150(x86) and e1000e Denys Fedoryshchenko
2012-05-17 13:42 ` Denys Fedoryshchenko
2012-05-17 16:54   ` Denys Fedoryshchenko
2012-05-18 14:04     ` Denys Fedoryshchenko
2012-05-19  2:07       ` Tom Herbert
2012-05-19  2:29         ` Denys Fedoryshchenko
2012-05-19  2:40         ` Denys Fedoryshchenko
2012-05-20 17:40           ` Tom Herbert
2012-05-20 18:53             ` Denys Fedoryshchenko
2012-05-20 19:07               ` Eric Dumazet
2012-05-20 19:18                 ` Denys Fedoryshchenko
2012-05-21  3:56                   ` Eric Dumazet
2012-05-21  8:06                     ` Denys Fedoryshchenko
2012-05-21  8:30                       ` Eric Dumazet
2012-05-21  8:40                         ` Eric Dumazet
2012-05-21  9:22                           ` Denys Fedoryshchenko
2012-05-22 17:11                           ` Denys Fedoryshchenko
2012-05-22 17:24                             ` Eric Dumazet
2012-05-25  6:01                               ` Tom Herbert
2012-05-25  6:22                                 ` Eric Dumazet
2012-05-25 16:59                                   ` Tom Herbert
2012-05-25 17:18                                     ` Eric Dumazet
2012-05-29 14:25             ` Hiroaki SHIMODA
2012-05-29 14:54               ` Tom Herbert
2012-05-29 15:11                 ` Eric Dumazet
2012-05-29 19:52                 ` Denys Fedoryshchenko
2012-05-30  0:06                 ` Hiroaki SHIMODA [this message]
2012-05-30  8:40                   ` Eric Dumazet
2012-05-30 10:43                     ` Hiroaki SHIMODA
2012-05-30 11:08                       ` Eric Dumazet
2012-05-30 11:20                         ` Joe Perches
2012-05-30 11:59                           ` Eric Dumazet
2012-05-30 14:09                             ` Joe Perches
2012-05-30 14:42                               ` Eric Dumazet
2012-05-30 14:50                                 ` Joe Perches
2012-05-30 11:29                         ` Hiroaki SHIMODA
2012-05-30 14:49                           ` Eric Dumazet
2012-05-30 15:00                             ` dave taht
2012-05-30 22:19                             ` Hiroaki SHIMODA
2012-05-30 10:52                     ` David Laight
2012-05-30 11:04                       ` Eric Dumazet
2012-05-30 11:12                         ` Eric Dumazet
2012-06-06  5:10               ` Eric Dumazet
2012-06-06  8:43                 ` Hiroaki SHIMODA
2012-06-06 16:26                   ` Jesse Brandeburg
2012-06-06 17:05                     ` Eric Dumazet
2012-06-06 17:26                       ` David Miller
2012-06-06 17:19                     ` Hiroaki SHIMODA
2012-06-06 18:21                       ` Tom Herbert
2012-06-06 18:23                         ` David Miller
2012-06-06 18:46                           ` Stephen Hemminger
2012-05-20 18:13       ` Eric Dumazet

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=20120530090602.6204d857.shimoda.hiroaki@gmail.com \
    --to=shimoda.hiroaki@gmail.com \
    --cc=davem@davemloft.net \
    --cc=denys@visp.net.lb \
    --cc=e1000-devel@lists.sourceforge.net \
    --cc=eric.dumazet@gmail.com \
    --cc=jeffrey.t.kirsher@intel.com \
    --cc=jesse.brandeburg@intel.com \
    --cc=netdev@vger.kernel.org \
    --cc=therbert@google.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 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).