From mboxrd@z Thu Jan  1 00:00:00 1970
From: Ben Greear <greearb@candelatech.com>
Date: Fri, 07 Jan 2011 22:12:54 -0800
Subject: [ath9k-devel] Dealing with packets out of BAW (tx progress blocked)
Message-ID: <4D280066.50409@candelatech.com>
List-Id: <ath9k-devel.lists.ath9k.org>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
To: ath9k-devel@lists.ath9k.org

Still working on the tx-logic hang.  I am seeing cases where
we cannot form aggr packets for transmit because
the fi->seqno is a bit less than tid->seq_start.

Should we just drop these packets...or maybe there is
no valid way we should ever get in this state??

Code check that fails is below:

static enum ATH_AGGR_STATUS ath_tx_form_aggr(struct ath_softc *sc,
					     struct ath_txq *txq,
					     struct ath_atx_tid *tid,
					     struct list_head *bf_q,
					     int *aggr_len)
{
#define PADBYTES(_len) ((4 - ((_len) % 4)) % 4)
	struct ath_buf *bf, *bf_first, *bf_prev = NULL;
	int rl = 0, nframes = 0, ndelim, prev_al = 0;
	u16 aggr_limit = 0, al = 0, bpad = 0,
		al_delta, h_baw = tid->baw_size / 2;
	enum ATH_AGGR_STATUS status = ATH_AGGR_DONE;
	struct ieee80211_tx_info *tx_info;
	struct ath_frame_info *fi;

	bf_first = list_first_entry(&tid->buf_q, struct ath_buf, list);

	do {
		bf = list_first_entry(&tid->buf_q, struct ath_buf, list);
		fi = get_frame_info(bf->bf_mpdu);

		/* do not step over block-ack window */
		if (!BAW_WITHIN(tid->seq_start, tid->baw_size, fi->seqno)) {
			if (dbg_hack)
				printk("  tid->seq_start: %hu baw_size: %hu"
				       " baw-head: %i  baw-tail: %i seqno: %i (BAW_CLOSED)\n",
				       tid->seq_start, tid->baw_size,
				       tid->baw_head, tid->baw_tail, fi->seqno);


Printout of one of the cases where we hit this debug code:

  tid->seq_start: 4 baw_size: 64 baw-head: 3  baw-tail: 3 seqno: 0 (BAW_CLOSED)


-- 
Ben Greear <greearb@candelatech.com>
Candela Technologies Inc  http://www.candelatech.com