From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from smtp.kernel.org (aws-us-west-2-korg-mail-alma10-1.taild15c8.ts.net [100.103.45.18]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id CA9074F4030; Tue, 13 Oct 2026 11:49:39 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=100.103.45.18 ARC-Seal:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1791892181; cv=none; b=n6qTg/x32TIAl9syqcqQxDxnreDW9Rs1vhuienOXphHliDjm4vWP/Ak7bc5V7rdSYPwtcvuoDwERCleF21B+xQ9dlXnDBiJDLUMMuvwcw24m+DkLXJXAQVEIX3Z19Rg0g4Y0DPa3Qc45DVDo0OL0Dnf/JkhZKPYoctZMMEPlM6g= ARC-Message-Signature:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1791892181; c=relaxed/simple; bh=k8iiltljkaH+v5tO61qd7aYGzsmzzT9b1toFsCvLlgc=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=J+vnywrggRrSEkCPuPCxxQaIgCnZBsgJUTD5Ue0fvHgkcd4bAgSApTGGqba/B3q/0b2wae5zEkBOKbPoUpRrzFZKfaIZBAYFEbWxOJAk/4lcY6HybTLiGVYCXFmI4CHqh3PCQfgKPJBsmGk+KzShceODv4+6tP3G7WWc0MtxHqM= ARC-Authentication-Results:i=1; smtp.subspace.kernel.org; dkim=pass (1024-bit key) header.d=linuxfoundation.org header.i=@linuxfoundation.org header.b=kRHxzAzO; arc=none smtp.client-ip=100.103.45.18 Authentication-Results: smtp.subspace.kernel.org; dkim=pass (1024-bit key) header.d=linuxfoundation.org header.i=@linuxfoundation.org header.b="kRHxzAzO" Received: by smtp.kernel.org (Postfix) with ESMTPSA id 1EBDE1F000FF; Tue, 13 Oct 2026 11:49:38 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=linuxfoundation.org; s=korg; t=1791892179; bh=5z1nkH+99aAT01aSwogDfCMeiHRCI4qGURf0RHuvj/M=; h=From:To:Cc:Subject:Date:In-Reply-To:References; b=kRHxzAzOBiYxN4Jy2gqIstvR4ICtWNimZEemD0OTIIuqup37dnrmn3JfOfUM9Jj+t x+X4dlR6yMuVo0KaNRAqxCdK8uH3eSPtnDfiIzUMy0JCK116CoBpMS0FQqSFmeqM43 7GB2DAK0stea2O47sFizC+SgNWOT6y4SmP3j9mJU= From: Greg Kroah-Hartman To: stable@vger.kernel.org Cc: Greg Kroah-Hartman , patches@lists.linux.dev, Suraj Gupta , Jakub Kicinski Subject: [PATCH 7.2 088/281] net: xilinx: axienet: Free outstanding DMA buffers on dmaengine stop Date: Tue, 13 Oct 2026 13:36:44 +0200 Message-ID: <20261013112857.270286557@linuxfoundation.org> X-Mailer: git-send-email 2.56.0 In-Reply-To: <20261013112855.045212108@linuxfoundation.org> References: <20261013112855.045212108@linuxfoundation.org> User-Agent: quilt/0.69 X-stable: review X-Patchwork-Hint: ignore Precedence: bulk X-Mailing-List: patches@lists.linux.dev List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: 8bit 7.2-stable review patch. If anyone has any objections, please let me know. ------------------ From: Suraj Gupta commit bcf2573a68a808a5b54cadafa72b6672e662c2bf upstream. In the dmaengine path the driver pre-submits RX buffers and holds in-flight TX buffers whose SKBs are DMA-mapped by the driver and freed only in the completion callbacks. On ndo_stop() dmaengine_terminate_sync() aborts these descriptors without running their callbacks, and the driver then frees only the ring shells, leaking every SKB still owned by the engine and its DMA mapping on each ifdown. With 128 RX buffers pre-posted per channel, the mapping leak can eventually exhaust a limited IOMMU aperture. Clear the slot's skb in the TX and RX callbacks so a non-NULL skb marks a slot that still owns a live, DMA-mapped buffer, and on stop unmap and free every such buffer. axienet_dma_rx_cb() runs from the DMA tasklet and re-arms the RX ring on each completion, so it can race axienet_stop(): a completion may submit a fresh buffer after dmaengine_terminate_sync() has returned, leaving the channel armed with a buffer the teardown then frees while the engine may still write into it (dma_release_channel() does not stop it either). Add a lock that axienet_dma_rx_cb() holds across the @stopping check and the resubmit, and axienet_stop() holds to set @stopping before terminating. Once @stopping is set no callback can arm a new buffer, and any armed just before is aborted by the terminate, so teardown only frees buffers the engine no longer owns. Fixes: 6a91b846af85 ("net: axienet: Introduce dmaengine support") Cc: stable@vger.kernel.org Signed-off-by: Suraj Gupta Link: https://patch.msgid.link/20260928184207.2361931-1-suraj.gupta2@amd.com Signed-off-by: Jakub Kicinski Signed-off-by: Greg Kroah-Hartman --- drivers/net/ethernet/xilinx/xilinx_axienet.h | 7 ++- drivers/net/ethernet/xilinx/xilinx_axienet_main.c | 51 ++++++++++++++++++---- 2 files changed, 47 insertions(+), 11 deletions(-) --- a/drivers/net/ethernet/xilinx/xilinx_axienet.h +++ b/drivers/net/ethernet/xilinx/xilinx_axienet.h @@ -523,8 +523,9 @@ struct skbuf_dma_descriptor { * @stats_work: Work for reading the hardware statistics counters often enough * to catch overflows. * @dma_err_task: Work structure to process Axi DMA errors - * @stopping: Set when @dma_err_task shouldn't do anything because we are - * about to stop the device. + * @stopping: Set when we are about to stop the device: makes @dma_err_task + * a no-op (legacy DMA path) and fences RX descriptor + * resubmission in axienet_dma_rx_cb() (dmaengine path). * @tx_irq: Axidma TX IRQ number * @rx_irq: Axidma RX IRQ number * @eth_irq: Ethernet core IRQ number @@ -545,6 +546,7 @@ struct skbuf_dma_descriptor { * @tx_ring_tail: TX skb ring buffer tail index. * @rx_ring_head: RX skb ring buffer head index. * @rx_ring_tail: RX skb ring buffer tail index. + * @rx_submit_lock: Protects RX ring resubmission vs teardown in dmaengine path. */ struct axienet_local { struct net_device *ndev; @@ -626,6 +628,7 @@ struct axienet_local { int tx_ring_tail; int rx_ring_head; int rx_ring_tail; + spinlock_t rx_submit_lock; }; /** --- a/drivers/net/ethernet/xilinx/xilinx_axienet_main.c +++ b/drivers/net/ethernet/xilinx/xilinx_axienet_main.c @@ -881,6 +881,7 @@ static void axienet_dma_tx_cb(void *data u64_stats_update_end(&lp->tx_stat_sync); dma_unmap_sg(lp->dev, skbuf_dma->sgl, skbuf_dma->sg_len, DMA_TO_DEVICE); dev_consume_skb_any(skbuf_dma->skb); + skbuf_dma->skb = NULL; netif_txq_completed_wake(txq, 1, len, CIRC_SPACE(lp->tx_ring_head, lp->tx_ring_tail, TX_BD_NUM_MAX), 2); @@ -1171,6 +1172,7 @@ static void axienet_dma_rx_cb(void *data &meta_max_len); dma_unmap_single(lp->dev, skbuf_dma->dma_address, lp->max_frm_size, DMA_FROM_DEVICE); + skbuf_dma->skb = NULL; if (IS_ERR(app_metadata)) { if (net_ratelimit()) @@ -1193,10 +1195,17 @@ static void axienet_dma_rx_cb(void *data u64_stats_update_end(&lp->rx_stat_sync); rx_submit: + spin_lock(&lp->rx_submit_lock); + if (lp->stopping) { + spin_unlock(&lp->rx_submit_lock); + return; + } + for (i = 0; i < CIRC_SPACE(lp->rx_ring_head, lp->rx_ring_tail, RX_BUF_NUM_DEFAULT); i++) axienet_rx_submit_desc(lp->ndev); dma_async_issue_pending(lp->rx_chan); + spin_unlock(&lp->rx_submit_lock); } /** @@ -1541,6 +1550,7 @@ static int axienet_init_dmaengine(struct lp->tx_ring_head = 0; lp->rx_ring_tail = 0; lp->rx_ring_head = 0; + lp->stopping = false; lp->tx_skb_ring = kzalloc_objs(*lp->tx_skb_ring, TX_BD_NUM_MAX); if (!lp->tx_skb_ring) { ret = -ENOMEM; @@ -1752,20 +1762,42 @@ static int axienet_stop(struct net_devic free_irq(lp->rx_irq, ndev); axienet_dma_bd_release(ndev); } else { + struct skbuf_dma_descriptor *skbuf_dma; + + spin_lock_bh(&lp->rx_submit_lock); + lp->stopping = true; + spin_unlock_bh(&lp->rx_submit_lock); + dmaengine_terminate_sync(lp->tx_chan); - dmaengine_synchronize(lp->tx_chan); dmaengine_terminate_sync(lp->rx_chan); - dmaengine_synchronize(lp->rx_chan); - - for (i = 0; i < TX_BD_NUM_MAX; i++) - kfree(lp->tx_skb_ring[i]); - kfree(lp->tx_skb_ring); - for (i = 0; i < RX_BUF_NUM_DEFAULT; i++) - kfree(lp->rx_skb_ring[i]); - kfree(lp->rx_skb_ring); dma_release_channel(lp->rx_chan); dma_release_channel(lp->tx_chan); + + /* Unmap and free any buffer the terminate did not reclaim, so it + * is not leaked; a non-NULL skb marks such a slot. + */ + for (i = 0; i < TX_BD_NUM_MAX; i++) { + skbuf_dma = lp->tx_skb_ring[i]; + if (skbuf_dma && skbuf_dma->skb) { + dma_unmap_sg(lp->dev, skbuf_dma->sgl, + skbuf_dma->sg_len, DMA_TO_DEVICE); + dev_kfree_skb_any(skbuf_dma->skb); + } + kfree(skbuf_dma); + } + kfree(lp->tx_skb_ring); + + for (i = 0; i < RX_BUF_NUM_DEFAULT; i++) { + skbuf_dma = lp->rx_skb_ring[i]; + if (skbuf_dma && skbuf_dma->skb) { + dma_unmap_single(lp->dev, skbuf_dma->dma_address, + lp->max_frm_size, DMA_FROM_DEVICE); + dev_kfree_skb_any(skbuf_dma->skb); + } + kfree(skbuf_dma); + } + kfree(lp->rx_skb_ring); } netdev_reset_queue(ndev); @@ -3062,6 +3094,7 @@ static int axienet_probe(struct platform spin_lock_init(&lp->rx_cr_lock); spin_lock_init(&lp->tx_cr_lock); + spin_lock_init(&lp->rx_submit_lock); INIT_WORK(&lp->rx_dim.work, axienet_rx_dim_work); lp->rx_dim_enabled = true; lp->rx_dim.profile_ix = 1;