From patchwork Fri Feb 5 09:06:40 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Luca Coelho X-Patchwork-Id: 12069701 X-Patchwork-Delegate: luca@coelho.fi Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-16.7 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS,URIBL_BLOCKED,USER_AGENT_GIT autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id D9E5EC43381 for ; Fri, 5 Feb 2021 09:33:13 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by mail.kernel.org (Postfix) with ESMTP id 9C7D364F9E for ; Fri, 5 Feb 2021 09:33:13 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S230150AbhBEJdA (ORCPT ); Fri, 5 Feb 2021 04:33:00 -0500 Received: from paleale.coelho.fi ([176.9.41.70]:43820 "EHLO farmhouse.coelho.fi" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S230267AbhBEJVs (ORCPT ); Fri, 5 Feb 2021 04:21:48 -0500 Received: from 91-156-6-193.elisa-laajakaista.fi ([91.156.6.193] helo=redipa.ger.corp.intel.com) by farmhouse.coelho.fi with esmtpsa (TLS1.3) tls TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384 (Exim 4.93) (envelope-from ) id 1l7x4n-0044i1-MY; Fri, 05 Feb 2021 11:06:54 +0200 From: Luca Coelho To: kvalo@codeaurora.org Cc: linux-wireless@vger.kernel.org Date: Fri, 5 Feb 2021 11:06:40 +0200 Message-Id: X-Mailer: git-send-email 2.30.0 In-Reply-To: <20210205090642.1553849-1-luca@coelho.fi> References: <20210205090642.1553849-1-luca@coelho.fi> MIME-Version: 1.0 Subject: [RESEND PATCH 10/12] iwlwifi: pcie: Disable softirqs during Rx queue init Precedence: bulk List-ID: X-Mailing-List: linux-wireless@vger.kernel.org From: Ilan Peer When Rx queues are configured during module init, NAPI is enabled while the Rx queue lock is held. However, since softirqs are not disabled, it is possible that and IRQ would fire and call iwl_pcie_rx_handle() which would also try to acquire the Rx lock. Prevent this by disabling softirqs during Rx queue configuration, as part of module init flow. Signed-off-by: Ilan Peer Signed-off-by: Luca Coelho --- drivers/net/wireless/intel/iwlwifi/pcie/rx.c | 36 ++++++++++---------- 1 file changed, 18 insertions(+), 18 deletions(-) diff --git a/drivers/net/wireless/intel/iwlwifi/pcie/rx.c b/drivers/net/wireless/intel/iwlwifi/pcie/rx.c index e3e53419b526..407809c7e958 100644 --- a/drivers/net/wireless/intel/iwlwifi/pcie/rx.c +++ b/drivers/net/wireless/intel/iwlwifi/pcie/rx.c @@ -301,7 +301,7 @@ static void iwl_pcie_rxsq_restock(struct iwl_trans *trans, if (!test_bit(STATUS_DEVICE_ENABLED, &trans->status)) return; - spin_lock(&rxq->lock); + spin_lock_bh(&rxq->lock); while ((iwl_rxq_space(rxq) > 0) && (rxq->free_count)) { __le32 *bd = (__le32 *)rxq->bd; /* The overwritten rxb must be a used one */ @@ -320,14 +320,14 @@ static void iwl_pcie_rxsq_restock(struct iwl_trans *trans, rxq->write = (rxq->write + 1) & RX_QUEUE_MASK; rxq->free_count--; } - spin_unlock(&rxq->lock); + spin_unlock_bh(&rxq->lock); /* If we've added more space for the firmware to place data, tell it. * Increment device's write pointer in multiples of 8. */ if (rxq->write_actual != (rxq->write & ~0x7)) { - spin_lock(&rxq->lock); + spin_lock_bh(&rxq->lock); iwl_pcie_rxq_inc_wr_ptr(trans, rxq); - spin_unlock(&rxq->lock); + spin_unlock_bh(&rxq->lock); } } @@ -433,28 +433,28 @@ void iwl_pcie_rxq_alloc_rbs(struct iwl_trans *trans, gfp_t priority, while (1) { unsigned int offset; - spin_lock(&rxq->lock); + spin_lock_bh(&rxq->lock); if (list_empty(&rxq->rx_used)) { - spin_unlock(&rxq->lock); + spin_unlock_bh(&rxq->lock); return; } - spin_unlock(&rxq->lock); + spin_unlock_bh(&rxq->lock); page = iwl_pcie_rx_alloc_page(trans, &offset, priority); if (!page) return; - spin_lock(&rxq->lock); + spin_lock_bh(&rxq->lock); if (list_empty(&rxq->rx_used)) { - spin_unlock(&rxq->lock); + spin_unlock_bh(&rxq->lock); __free_pages(page, trans_pcie->rx_page_order); return; } rxb = list_first_entry(&rxq->rx_used, struct iwl_rx_mem_buffer, list); list_del(&rxb->list); - spin_unlock(&rxq->lock); + spin_unlock_bh(&rxq->lock); BUG_ON(rxb->page); rxb->page = page; @@ -466,19 +466,19 @@ void iwl_pcie_rxq_alloc_rbs(struct iwl_trans *trans, gfp_t priority, DMA_FROM_DEVICE); if (dma_mapping_error(trans->dev, rxb->page_dma)) { rxb->page = NULL; - spin_lock(&rxq->lock); + spin_lock_bh(&rxq->lock); list_add(&rxb->list, &rxq->rx_used); - spin_unlock(&rxq->lock); + spin_unlock_bh(&rxq->lock); __free_pages(page, trans_pcie->rx_page_order); return; } - spin_lock(&rxq->lock); + spin_lock_bh(&rxq->lock); list_add_tail(&rxb->list, &rxq->rx_free); rxq->free_count++; - spin_unlock(&rxq->lock); + spin_unlock_bh(&rxq->lock); } } @@ -1112,7 +1112,7 @@ static int _iwl_pcie_rx_init(struct iwl_trans *trans) for (i = 0; i < trans->num_rx_queues; i++) { struct iwl_rxq *rxq = &trans_pcie->rxq[i]; - spin_lock(&rxq->lock); + spin_lock_bh(&rxq->lock); /* * Set read write pointer to reflect that we have processed * and used all buffers, but have not restocked the Rx queue @@ -1148,7 +1148,7 @@ static int _iwl_pcie_rx_init(struct iwl_trans *trans) napi_enable(&rxq->napi); } - spin_unlock(&rxq->lock); + spin_unlock_bh(&rxq->lock); } /* move the pool to the default queue and allocator ownerships */ @@ -1190,9 +1190,9 @@ int iwl_pcie_rx_init(struct iwl_trans *trans) iwl_pcie_rxq_restock(trans, trans_pcie->rxq); - spin_lock(&trans_pcie->rxq->lock); + spin_lock_bh(&trans_pcie->rxq->lock); iwl_pcie_rxq_inc_wr_ptr(trans, trans_pcie->rxq); - spin_unlock(&trans_pcie->rxq->lock); + spin_unlock_bh(&trans_pcie->rxq->lock); return 0; }