From patchwork Fri Feb 8 10:46:54 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Luca Coelho X-Patchwork-Id: 10802693 X-Patchwork-Delegate: luca@coelho.fi Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 79A1A13BF for ; Fri, 8 Feb 2019 10:47:24 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 68B2E2D984 for ; Fri, 8 Feb 2019 10:47:24 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id 5C9AD2DF2F; Fri, 8 Feb 2019 10:47:24 +0000 (UTC) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on pdx-wl-mail.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-7.9 required=2.0 tests=BAYES_00,MAILING_LIST_MULTI, RCVD_IN_DNSWL_HI autolearn=ham version=3.3.1 Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id DAAD82D984 for ; Fri, 8 Feb 2019 10:47:23 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1727650AbfBHKrW (ORCPT ); Fri, 8 Feb 2019 05:47:22 -0500 Received: from paleale.coelho.fi ([176.9.41.70]:56946 "EHLO farmhouse.coelho.fi" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1727134AbfBHKrW (ORCPT ); Fri, 8 Feb 2019 05:47:22 -0500 Received: from 91-156-4-241.elisa-laajakaista.fi ([91.156.4.241] helo=redipa.ger.corp.intel.com) by farmhouse.coelho.fi with esmtpsa (TLS1.2:ECDHE_RSA_AES_128_GCM_SHA256:128) (Exim 4.91) (envelope-from ) id 1gs3gl-0002jg-Au; Fri, 08 Feb 2019 12:47:20 +0200 From: Luca Coelho To: kvalo@codeaurora.org Cc: linux-wireless@vger.kernel.org, Sara Sharon , Luca Coelho Date: Fri, 8 Feb 2019 12:46:54 +0200 Message-Id: <20190208104712.16210-3-luca@coelho.fi> X-Mailer: git-send-email 2.20.1 In-Reply-To: <20190208104712.16210-1-luca@coelho.fi> References: <20190208104712.16210-1-luca@coelho.fi> MIME-Version: 1.0 Subject: [PATCH 02/20] iwlwifi: pcie: fix TX while flushing Sender: linux-wireless-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-wireless@vger.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP From: Sara Sharon When flushing TX queues no new TX should go into the system. However, in the following scenario we get TX: 1. Queues are stopped and there are packets in overflow queue 2. Station is removed and flush begins 3. Flush empties space, and reclaim path TXes SKB from overflow queue. Note that the fact the queues are stopped during the process doesn't matter - the packet will be TXed since the TX path doesn't care if TX queues are stopped or not, just if there is space in the queue, which there is, since we just freed a packet. A fix here is rather complicated, since the flow is very racy. Change code not to warn if we are TXing from overflow TX. In case there is TX from both overflow TX and TX path we will miss a warning we optimally had, but we can live with that. Make sure we don't return before overflow queue is empty, otherwise we will think queues are empty, but they will be refilled, resulting with assert. Signed-off-by: Sara Sharon Fixes: 3955525d5d17 ("iwlwifi: pcie: buffer packets to avoid overflowing Tx queues") Signed-off-by: Luca Coelho --- .../wireless/intel/iwlwifi/pcie/internal.h | 2 ++ .../net/wireless/intel/iwlwifi/pcie/trans.c | 24 +++++++++++++++++-- drivers/net/wireless/intel/iwlwifi/pcie/tx.c | 10 ++++++++ 3 files changed, 34 insertions(+), 2 deletions(-) diff --git a/drivers/net/wireless/intel/iwlwifi/pcie/internal.h b/drivers/net/wireless/intel/iwlwifi/pcie/internal.h index 9e1bcafad786..0ecd90d050e6 100644 --- a/drivers/net/wireless/intel/iwlwifi/pcie/internal.h +++ b/drivers/net/wireless/intel/iwlwifi/pcie/internal.h @@ -400,6 +400,8 @@ struct iwl_txq { u32 id; int low_mark; int high_mark; + + bool overflow_tx; }; static inline dma_addr_t diff --git a/drivers/net/wireless/intel/iwlwifi/pcie/trans.c b/drivers/net/wireless/intel/iwlwifi/pcie/trans.c index 6c30c88fc41e..4b31b0cdbd09 100644 --- a/drivers/net/wireless/intel/iwlwifi/pcie/trans.c +++ b/drivers/net/wireless/intel/iwlwifi/pcie/trans.c @@ -2240,6 +2240,7 @@ static int iwl_trans_pcie_wait_txq_empty(struct iwl_trans *trans, int txq_idx) struct iwl_trans_pcie *trans_pcie = IWL_TRANS_GET_PCIE_TRANS(trans); struct iwl_txq *txq; unsigned long now = jiffies; + bool overflow_tx; u8 wr_ptr; /* Make sure the NIC is still alive in the bus */ @@ -2251,18 +2252,37 @@ static int iwl_trans_pcie_wait_txq_empty(struct iwl_trans *trans, int txq_idx) IWL_DEBUG_TX_QUEUES(trans, "Emptying queue %d...\n", txq_idx); txq = trans_pcie->txq[txq_idx]; + + spin_lock_bh(&txq->lock); + overflow_tx = txq->overflow_tx || + !skb_queue_empty(&txq->overflow_q); + spin_unlock_bh(&txq->lock); + wr_ptr = READ_ONCE(txq->write_ptr); - while (txq->read_ptr != READ_ONCE(txq->write_ptr) && + while ((txq->read_ptr != READ_ONCE(txq->write_ptr) || + overflow_tx) && !time_after(jiffies, now + msecs_to_jiffies(IWL_FLUSH_WAIT_MS))) { u8 write_ptr = READ_ONCE(txq->write_ptr); - if (WARN_ONCE(wr_ptr != write_ptr, + /* + * If write pointer moved during the wait, warn only + * if the TX came from op mode. In case TX came from + * trans layer (overflow TX) don't warn. + */ + if (WARN_ONCE(wr_ptr != write_ptr && !overflow_tx, "WR pointer moved while flushing %d -> %d\n", wr_ptr, write_ptr)) return -ETIMEDOUT; + wr_ptr = write_ptr; + usleep_range(1000, 2000); + + spin_lock_bh(&txq->lock); + overflow_tx = txq->overflow_tx || + !skb_queue_empty(&txq->overflow_q); + spin_unlock_bh(&txq->lock); } if (txq->read_ptr != txq->write_ptr) { diff --git a/drivers/net/wireless/intel/iwlwifi/pcie/tx.c b/drivers/net/wireless/intel/iwlwifi/pcie/tx.c index d8773e0a6062..9fbd37d23e85 100644 --- a/drivers/net/wireless/intel/iwlwifi/pcie/tx.c +++ b/drivers/net/wireless/intel/iwlwifi/pcie/tx.c @@ -1181,6 +1181,15 @@ void iwl_trans_pcie_reclaim(struct iwl_trans *trans, int txq_id, int ssn, __skb_queue_head_init(&overflow_skbs); skb_queue_splice_init(&txq->overflow_q, &overflow_skbs); + /* + * We are going to transmit from the overflow queue. + * Remember this state so that wait_for_txq_empty will know we + * are adding more packets to the TFD queue. It cannot rely on + * the state of &txq->overflow_q, as we just emptied it, but + * haven't TXed the content yet. + */ + txq->overflow_tx = true; + /* * This is tricky: we are in reclaim path which is non * re-entrant, so noone will try to take the access the @@ -1209,6 +1218,7 @@ void iwl_trans_pcie_reclaim(struct iwl_trans *trans, int txq_id, int ssn, iwl_wake_queue(trans, txq); spin_lock_bh(&txq->lock); + txq->overflow_tx = false; } if (txq->read_ptr == txq->write_ptr) {