From nobody Mon Sep 29 21:09:35 2025 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id D44D1C25B0F for ; Tue, 16 Aug 2022 04:42:39 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S231909AbiHPEmg (ORCPT ); Tue, 16 Aug 2022 00:42:36 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:42486 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S230474AbiHPEkI (ORCPT ); Tue, 16 Aug 2022 00:40:08 -0400 Received: from dfw.source.kernel.org (dfw.source.kernel.org [IPv6:2604:1380:4641:c500::1]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 9F51B4F66E; Mon, 15 Aug 2022 13:31:29 -0700 (PDT) Received: from smtp.kernel.org (relay.kernel.org [52.25.139.140]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by dfw.source.kernel.org (Postfix) with ESMTPS id 3F02460EE9; Mon, 15 Aug 2022 20:31:29 +0000 (UTC) Received: by smtp.kernel.org (Postfix) with ESMTPSA id 2BC03C433D6; Mon, 15 Aug 2022 20:31:27 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=linuxfoundation.org; s=korg; t=1660595488; bh=Wrm3g5UT3Db63BpJWdAvHlY44yFmCZwOM483tLdfX34=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=FAsfNfz9aRxZONY0lvaqu+RHdoUirF2EMeYxsEUcFnDiZ4OeW7qi7g5LINudqIH/4 Lix0f3K+D7EDpJmhV/aWsrfC4Ev3QGdSo3sWuwhfZBGXsh3sdihKHJGMfLNGW2qWt4 ZsVmM33Z+p04b4XENc/uF674mBPnyD2SgbPNxiFk= From: Greg Kroah-Hartman To: linux-kernel@vger.kernel.org Cc: Greg Kroah-Hartman , stable@vger.kernel.org, Viacheslav Mitrofanov , Vinod Koul , Sasha Levin Subject: [PATCH 5.19 0780/1157] dmaengine: sf-pdma: Add multithread support for a DMA channel Date: Mon, 15 Aug 2022 20:02:16 +0200 Message-Id: <20220815180510.706154414@linuxfoundation.org> X-Mailer: git-send-email 2.37.2 In-Reply-To: <20220815180439.416659447@linuxfoundation.org> References: <20220815180439.416659447@linuxfoundation.org> User-Agent: quilt/0.67 MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Type: text/plain; charset="utf-8" From: Viacheslav Mitrofanov [ Upstream commit b2cc5c465c2cb8ab697c3fd6583c614e3f6cfbcc ] When we get a DMA channel and try to use it in multiple threads it will cause oops and hanging the system. % echo 64 > /sys/module/dmatest/parameters/threads_per_chan % echo 10000 > /sys/module/dmatest/parameters/iterations % echo 1 > /sys/module/dmatest/parameters/run [ 89.480664] Unable to handle kernel NULL pointer dereference at virtual address 00000000000000a0 [ 89.488725] Oops [#1] [ 89.494708] CPU: 2 PID: 1008 Comm: dma0chan0-copy0 Not tainted 5.17.0-rc5 [ 89.509385] epc : vchan_find_desc+0x32/0x46 [ 89.513553] ra : sf_pdma_tx_status+0xca/0xd6 This happens because of data race. Each thread rewrite channels's descriptor as soon as device_prep_dma_memcpy() is called. It leads to the situation when the driver thinks that it uses right descriptor that actually is freed or substituted for other one. With current fixes a descriptor changes its value only when it has been used. A new descriptor is acquired from vc->desc_issued queue that is already filled with descriptors that are ready to be sent. Threads have no direct access to DMA channel descriptor. Now it is just possible to queue a descriptor for further processing. Fixes: 6973886ad58e ("dmaengine: sf-pdma: add platform DMA support for HiFi= ve Unleashed A00") Signed-off-by: Viacheslav Mitrofanov Link: https://lore.kernel.org/r/20220701082942.12835-1-v.v.mitrofanov@yadro= .com Signed-off-by: Vinod Koul Signed-off-by: Sasha Levin --- drivers/dma/sf-pdma/sf-pdma.c | 44 ++++++++++++++++++++++++----------- 1 file changed, 30 insertions(+), 14 deletions(-) diff --git a/drivers/dma/sf-pdma/sf-pdma.c b/drivers/dma/sf-pdma/sf-pdma.c index db5a4ef76077..4f8b8498c5c6 100644 --- a/drivers/dma/sf-pdma/sf-pdma.c +++ b/drivers/dma/sf-pdma/sf-pdma.c @@ -52,16 +52,6 @@ static inline struct sf_pdma_desc *to_sf_pdma_desc(struc= t virt_dma_desc *vd) static struct sf_pdma_desc *sf_pdma_alloc_desc(struct sf_pdma_chan *chan) { struct sf_pdma_desc *desc; - unsigned long flags; - - spin_lock_irqsave(&chan->lock, flags); - - if (chan->desc && !chan->desc->in_use) { - spin_unlock_irqrestore(&chan->lock, flags); - return chan->desc; - } - - spin_unlock_irqrestore(&chan->lock, flags); =20 desc =3D kzalloc(sizeof(*desc), GFP_NOWAIT); if (!desc) @@ -111,7 +101,6 @@ sf_pdma_prep_dma_memcpy(struct dma_chan *dchan, dma_add= r_t dest, dma_addr_t src, desc->async_tx =3D vchan_tx_prep(&chan->vchan, &desc->vdesc, flags); =20 spin_lock_irqsave(&chan->vchan.lock, iflags); - chan->desc =3D desc; sf_pdma_fill_desc(desc, dest, src, len); spin_unlock_irqrestore(&chan->vchan.lock, iflags); =20 @@ -170,11 +159,17 @@ static size_t sf_pdma_desc_residue(struct sf_pdma_cha= n *chan, unsigned long flags; u64 residue =3D 0; struct sf_pdma_desc *desc; - struct dma_async_tx_descriptor *tx; + struct dma_async_tx_descriptor *tx =3D NULL; =20 spin_lock_irqsave(&chan->vchan.lock, flags); =20 - tx =3D &chan->desc->vdesc.tx; + list_for_each_entry(vd, &chan->vchan.desc_submitted, node) + if (vd->tx.cookie =3D=3D cookie) + tx =3D &vd->tx; + + if (!tx) + goto out; + if (cookie =3D=3D tx->chan->completed_cookie) goto out; =20 @@ -241,6 +236,19 @@ static void sf_pdma_enable_request(struct sf_pdma_chan= *chan) writel(v, regs->ctrl); } =20 +static struct sf_pdma_desc *sf_pdma_get_first_pending_desc(struct sf_pdma_= chan *chan) +{ + struct virt_dma_chan *vchan =3D &chan->vchan; + struct virt_dma_desc *vdesc; + + if (list_empty(&vchan->desc_issued)) + return NULL; + + vdesc =3D list_first_entry(&vchan->desc_issued, struct virt_dma_desc, nod= e); + + return container_of(vdesc, struct sf_pdma_desc, vdesc); +} + static void sf_pdma_xfer_desc(struct sf_pdma_chan *chan) { struct sf_pdma_desc *desc =3D chan->desc; @@ -268,8 +276,11 @@ static void sf_pdma_issue_pending(struct dma_chan *dch= an) =20 spin_lock_irqsave(&chan->vchan.lock, flags); =20 - if (vchan_issue_pending(&chan->vchan) && chan->desc) + if (!chan->desc && vchan_issue_pending(&chan->vchan)) { + /* vchan_issue_pending has made a check that desc in not NULL */ + chan->desc =3D sf_pdma_get_first_pending_desc(chan); sf_pdma_xfer_desc(chan); + } =20 spin_unlock_irqrestore(&chan->vchan.lock, flags); } @@ -298,6 +309,11 @@ static void sf_pdma_donebh_tasklet(struct tasklet_stru= ct *t) spin_lock_irqsave(&chan->vchan.lock, flags); list_del(&chan->desc->vdesc.node); vchan_cookie_complete(&chan->desc->vdesc); + + chan->desc =3D sf_pdma_get_first_pending_desc(chan); + if (chan->desc) + sf_pdma_xfer_desc(chan); + spin_unlock_irqrestore(&chan->vchan.lock, flags); } =20 --=20 2.35.1