From patchwork Fri Oct 14 05:00:47 2016 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: Vignesh Raghavendra X-Patchwork-Id: 9376113 Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork.web.codeaurora.org (Postfix) with ESMTP id 68A4060779 for ; Fri, 14 Oct 2016 05:01:39 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 5C5D62A2C6 for ; Fri, 14 Oct 2016 05:01:39 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id 4F9E92A2CA; Fri, 14 Oct 2016 05:01:39 +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=-6.9 required=2.0 tests=BAYES_00,RCVD_IN_DNSWL_HI autolearn=unavailable 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 8DC6C2A2C6 for ; Fri, 14 Oct 2016 05:01:38 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751645AbcJNFBR (ORCPT ); Fri, 14 Oct 2016 01:01:17 -0400 Received: from bear.ext.ti.com ([198.47.19.11]:49552 "EHLO bear.ext.ti.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751364AbcJNFBQ (ORCPT ); Fri, 14 Oct 2016 01:01:16 -0400 Received: from dflxv15.itg.ti.com ([128.247.5.124]) by bear.ext.ti.com (8.13.7/8.13.7) with ESMTP id u9E50sHn011536; Fri, 14 Oct 2016 00:00:54 -0500 Received: from DFLE73.ent.ti.com (dfle73.ent.ti.com [128.247.5.110]) by dflxv15.itg.ti.com (8.14.3/8.13.8) with ESMTP id u9E50riI004171; Fri, 14 Oct 2016 00:00:53 -0500 Received: from dflp33.itg.ti.com (10.64.6.16) by DFLE73.ent.ti.com (128.247.5.110) with Microsoft SMTP Server id 14.3.294.0; Fri, 14 Oct 2016 00:00:53 -0500 Received: from a0132425.india.ti.com (ileax41-snat.itg.ti.com [10.172.224.153]) by dflp33.itg.ti.com (8.14.3/8.13.8) with ESMTP id u9E50o1w012275; Fri, 14 Oct 2016 00:00:51 -0500 From: Vignesh R To: Vinod Koul CC: Sebastian Andrzej Siewior , Peter Ujfalusi , Russell King - ARM Linux , , , , Vignesh R Subject: [PATCH v4] dmaengine: omap-dma: add support for pause of non-cyclic transfers Date: Fri, 14 Oct 2016 10:30:47 +0530 Message-ID: <20161014050047.28447-1-vigneshr@ti.com> X-Mailer: git-send-email 2.10.1 In-Reply-To: <20161010120755.26153-1-vigneshr@ti.com> References: <20161010120755.26153-1-vigneshr@ti.com> MIME-Version: 1.0 Sender: linux-omap-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-omap@vger.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP From: Sebastian Andrzej Siewior This DMA driver is used by 8250-omap on DRA7-evm. There is one requirement that is to pause a transfer. This is currently used on the RX side. It is possible that the UART HW aborted the RX (UART's RX-timeout) but the DMA controller starts the transfer shortly after. Before we can manually purge the FIFO we need to pause the transfer, check how many bytes it already received and terminate the transfer without it making any progress. From testing on the TX side it seems that it is possible that we invoke pause once the transfer has completed which is indicated by the missing CCR_ENABLE bit but before the interrupt has been noticed. In that case the interrupt will come even after disabling it. The AM572x manual says that we have to wait for the CCR_RD_ACTIVE & CCR_WR_ACTIVE bits to be gone before programming it again here is the drain loop. Also it looks like without the drain the TX-transfer makes sometimes progress. One note: The pause + resume combo is broken because after resume the the complete transfer will be programmed again. That means the already transferred bytes (until the pause event) will be sent again. This is currently not important for my UART user because it does only pause + terminate. v3…v4: - update subject line. v2…v3: - rephrase the comment based on Russell's information / feedback. v1…v2: - move the drain loop into omap_dma_drain_chan() instead of having it twice. - allow pause only for DMA_DEV_TO_MEM transfers if non-cyclic. Add a comment why DMA_MEM_TO_DEV not allowed. - clear pause on terminate_all. Otherwise pause() + terminate_all() will keep the pause bit set and we can't pause the following transfer. Signed-off-by: Sebastian Andrzej Siewior [vigneshr@ti.com: drain channel only when buffering is on, rebase to v4.8] Signed-off-by: Vignesh R Acked-by: Peter Ujfalusi --- drivers/dma/omap-dma.c | 124 +++++++++++++++++++++++++++++++++++-------------- 1 file changed, 89 insertions(+), 35 deletions(-) diff --git a/drivers/dma/omap-dma.c b/drivers/dma/omap-dma.c index 7ca27d4b1c54..fd6b9e6834ad 100644 --- a/drivers/dma/omap-dma.c +++ b/drivers/dma/omap-dma.c @@ -422,7 +422,30 @@ static void omap_dma_start(struct omap_chan *c, struct omap_desc *d) c->running = true; } -static void omap_dma_stop(struct omap_chan *c) +static void omap_dma_drain_chan(struct omap_chan *c) +{ + int i; + u32 val; + + /* Wait for sDMA FIFO to drain */ + for (i = 0; ; i++) { + val = omap_dma_chan_read(c, CCR); + if (!(val & (CCR_RD_ACTIVE | CCR_WR_ACTIVE))) + break; + + if (i > 100) + break; + + udelay(5); + } + + if (val & (CCR_RD_ACTIVE | CCR_WR_ACTIVE)) + dev_err(c->vc.chan.device->dev, + "DMA drain did not complete on lch %d\n", + c->dma_ch); +} + +static int omap_dma_stop(struct omap_chan *c) { struct omap_dmadev *od = to_omap_dma_dev(c->vc.chan.device); uint32_t val; @@ -435,7 +458,6 @@ static void omap_dma_stop(struct omap_chan *c) val = omap_dma_chan_read(c, CCR); if (od->plat->errata & DMA_ERRATA_i541 && val & CCR_TRIGGER_SRC) { uint32_t sysconfig; - unsigned i; sysconfig = omap_dma_glbl_read(od, OCP_SYSCONFIG); val = sysconfig & ~DMA_SYSCONFIG_MIDLEMODE_MASK; @@ -446,27 +468,19 @@ static void omap_dma_stop(struct omap_chan *c) val &= ~CCR_ENABLE; omap_dma_chan_write(c, CCR, val); - /* Wait for sDMA FIFO to drain */ - for (i = 0; ; i++) { - val = omap_dma_chan_read(c, CCR); - if (!(val & (CCR_RD_ACTIVE | CCR_WR_ACTIVE))) - break; - - if (i > 100) - break; - - udelay(5); - } - - if (val & (CCR_RD_ACTIVE | CCR_WR_ACTIVE)) - dev_err(c->vc.chan.device->dev, - "DMA drain did not complete on lch %d\n", - c->dma_ch); + if (!(c->ccr & CCR_BUFFERING_DISABLE)) + omap_dma_drain_chan(c); omap_dma_glbl_write(od, OCP_SYSCONFIG, sysconfig); } else { + if (!(val & CCR_ENABLE)) + return -EINVAL; + val &= ~CCR_ENABLE; omap_dma_chan_write(c, CCR, val); + + if (!(c->ccr & CCR_BUFFERING_DISABLE)) + omap_dma_drain_chan(c); } mb(); @@ -481,8 +495,8 @@ static void omap_dma_stop(struct omap_chan *c) omap_dma_chan_write(c, CLNK_CTRL, val); } - c->running = false; + return 0; } static void omap_dma_start_sg(struct omap_chan *c, struct omap_desc *d) @@ -836,6 +850,8 @@ static enum dma_status omap_dma_tx_status(struct dma_chan *chan, } else { txstate->residue = 0; } + if (ret == DMA_IN_PROGRESS && c->paused) + ret = DMA_PAUSED; spin_unlock_irqrestore(&c->vc.lock, flags); return ret; @@ -1247,10 +1263,8 @@ static int omap_dma_terminate_all(struct dma_chan *chan) omap_dma_stop(c); } - if (c->cyclic) { - c->cyclic = false; - c->paused = false; - } + c->cyclic = false; + c->paused = false; vchan_get_all_descriptors(&c->vc, &head); spin_unlock_irqrestore(&c->vc.lock, flags); @@ -1269,28 +1283,66 @@ static void omap_dma_synchronize(struct dma_chan *chan) static int omap_dma_pause(struct dma_chan *chan) { struct omap_chan *c = to_omap_dma_chan(chan); + struct omap_dmadev *od = to_omap_dma_dev(chan->device); + unsigned long flags; + int ret = -EINVAL; + bool can_pause; - /* Pause/Resume only allowed with cyclic mode */ - if (!c->cyclic) - return -EINVAL; + spin_lock_irqsave(&od->irq_lock, flags); - if (!c->paused) { - omap_dma_stop(c); - c->paused = true; + if (!c->desc) + goto out; + + if (c->cyclic) + can_pause = true; + + /* + * We do not allow DMA_MEM_TO_DEV transfers to be paused. + * From the AM572x TRM, 16.1.4.18 Disabling a Channel During Transfer: + * "When a channel is disabled during a transfer, the channel undergoes + * an abort, unless it is hardware-source-synchronized …". + * A source-synchronised channel is one where the fetching of data is + * under control of the device. In other words, a device-to-memory + * transfer. So, a destination-synchronised channel (which would be a + * memory-to-device transfer) undergoes an abort if the the CCR_ENABLE + * bit is cleared. + * From 16.1.4.20.4.6.2 Abort: "If an abort trigger occurs, the channel + * aborts immediately after completion of current read/write + * transactions and then the FIFO is cleaned up." The term "cleaned up" + * is not defined. TI recommends to check that RD_ACTIVE and WR_ACTIVE + * are both clear _before_ disabling the channel, otherwise data loss + * will occur. + * The problem is that if the channel is active, then device activity + * can result in DMA activity starting between reading those as both + * clear and the write to DMA_CCR to clear the enable bit hitting the + * hardware. If the DMA hardware can't drain the data in its FIFO to the + * destination, then data loss "might" occur (say if we write to an UART + * and the UART is not accepting any further data). + */ + else if (c->desc->dir == DMA_DEV_TO_MEM) + can_pause = true; + + if (can_pause && !c->paused) { + ret = omap_dma_stop(c); + if (!ret) + c->paused = true; } +out: + spin_unlock_irqrestore(&od->irq_lock, flags); - return 0; + return ret; } static int omap_dma_resume(struct dma_chan *chan) { struct omap_chan *c = to_omap_dma_chan(chan); + struct omap_dmadev *od = to_omap_dma_dev(chan->device); + unsigned long flags; + int ret = -EINVAL; - /* Pause/Resume only allowed with cyclic mode */ - if (!c->cyclic) - return -EINVAL; + spin_lock_irqsave(&od->irq_lock, flags); - if (c->paused) { + if (c->paused && c->desc) { mb(); /* Restore channel link register */ @@ -1298,9 +1350,11 @@ static int omap_dma_resume(struct dma_chan *chan) omap_dma_start(c, c->desc); c->paused = false; + ret = 0; } + spin_unlock_irqrestore(&od->irq_lock, flags); - return 0; + return ret; } static int omap_dma_chan_init(struct omap_dmadev *od)