linux-wireless.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Johannes Berg <johannes@sipsolutions.net>
To: Thomas Pedersen <thomas@cozybit.com>
Cc: linux-wireless@vger.kernel.org,
	Pedro Larbig <pedro.larbig@carhs.de>,
	linville@tuxdriver.com
Subject: Re: [PATCH 2/9] mac80211: Limit amount of HWMP frames and forwarded data packets in queues on mesh interfaces
Date: Wed, 24 Aug 2011 22:08:20 -0700	[thread overview]
Message-ID: <1314248900.5054.45.camel@jlt3.sipsolutions.net> (raw)
In-Reply-To: <1314236452-7226-3-git-send-email-thomas@cozybit.com> (sfid-20110825_034140_614123_A977E37D)

On Wed, 2011-08-24 at 18:40 -0700, Thomas Pedersen wrote:
> From: Pedro Larbig <pedro.larbig@carhs.de>
> 
> To avoid contention problems in a mesh network's high load areas, this patch
> adds a 2-stage packet dropping mechanism:
> * If the transmit queue for HWMP frames is filled with 256 or more packets,
> additional HWMP frames will be dropped
> * If the transmit queue for forwarded packets is at 384 or more, drop those,
> too

> +	/* Frames going through ieee80211_tx_skb will be on the voice queue,
> +	 * Therefor we need to check only IEEE80211_AC_VO */
> +	if (unlikely(skb_queue_len(&local->pending[IEEE80211_AC_VO]) >=
> +	   MESH_MGMT_QUEUE_LEN)) {
> +		kfree_skb(skb);
> +		I802_DEBUG_INC(local->tx_handlers_drop_mesh_mgmt);
> +		return;
> +	}

I still don't think this should be done. If the HW queue is full, we
will get a stop_queue from the driver, this could also tell mesh to stop
forwarding or so.

johannes


  reply	other threads:[~2011-08-25  5:08 UTC|newest]

Thread overview: 31+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-08-25  1:40 [PATCH 0/9] mesh fixes Thomas Pedersen
2011-08-25  1:40 ` [PATCH 1/9] mac80211: Fix RCU pointer dereference in mesh_path_discard_frame() Thomas Pedersen
2011-08-25  2:08   ` Johannes Berg
2011-08-25 18:16     ` Javier Cardona
2011-08-25 18:21       ` Johannes Berg
2011-08-25 18:45         ` Javier Cardona
2011-08-25 18:48           ` Johannes Berg
2011-08-25 19:04             ` Javier Cardona
2011-08-25  1:40 ` [PATCH 2/9] mac80211: Limit amount of HWMP frames and forwarded data packets in queues on mesh interfaces Thomas Pedersen
2011-08-25  5:08   ` Johannes Berg [this message]
2011-08-25 17:46     ` Javier Cardona
2011-08-25  1:40 ` [PATCH 3/9] mac80211: Remove mesh paths when an interface is removed Thomas Pedersen
2011-08-25  1:40 ` [PATCH 4/9] mac80211: Improve mpath state locking Thomas Pedersen
2011-08-25  1:40 ` [PATCH 5/9] mac80211: Remove redundant mesh path expiration checks Thomas Pedersen
2011-08-25  1:40 ` [PATCH 6/9] mac80211: Don't iterate twice over all mpaths when once in sufficient Thomas Pedersen
2011-08-25  1:40 ` [PATCH 7/9] mac80211: Consolidate {mesh,mpp}_path_flush into one function Thomas Pedersen
2011-08-25  1:40 ` [PATCH 8/9] mac80211: Don't take the mesh path resize lock when deleting an mpath Thomas Pedersen
2011-08-25  1:40 ` [PATCH 9/9] mac80211: Consolidate mesh path duplicated functions Thomas Pedersen
2011-08-27  0:18 ` [PATCH v2 0/8] mesh fixes Javier Cardona
2011-08-27  0:18   ` [PATCH v2 1/8] mac80211: Fix RCU pointer dereference in mesh_path_discard_frame() Javier Cardona
2011-08-27  0:18   ` [PATCH v2 2/8] mac80211: Remove mesh paths when an interface is removed Javier Cardona
2011-08-27  0:18   ` [PATCH v2 3/8] mac80211: Improve mpath state locking Javier Cardona
2011-08-27  0:18   ` [PATCH v2 4/8] mac80211: Remove redundant mesh path expiration checks Javier Cardona
2011-08-27  0:18   ` [PATCH v2 5/8] mac80211: Don't iterate twice over all mpaths when once in sufficient Javier Cardona
2011-08-27  0:18   ` [PATCH v2 6/8] mac80211: Consolidate {mesh,mpp}_path_flush into one function Javier Cardona
2011-08-29 13:49     ` Johannes Berg
2011-08-29 18:36       ` Javier Cardona
2011-08-29 18:38         ` Johannes Berg
2011-08-27  0:18   ` [PATCH v2 7/8] mac80211: Don't take the mesh path resize lock when deleting an mpath Javier Cardona
2011-08-29 13:49     ` Johannes Berg
2011-08-27  0:18   ` [PATCH v2 8/8] mac80211: Consolidate mesh path duplicated functions Javier Cardona

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=1314248900.5054.45.camel@jlt3.sipsolutions.net \
    --to=johannes@sipsolutions.net \
    --cc=linux-wireless@vger.kernel.org \
    --cc=linville@tuxdriver.com \
    --cc=pedro.larbig@carhs.de \
    --cc=thomas@cozybit.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).