From: Mark Rutland <mark.rutland@arm.com> To: Jonas Jensen <jonas.jensen@gmail.com> Cc: "linux-arm-kernel@lists.infradead.org" <linux-arm-kernel@lists.infradead.org>, "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>, "arm@kernel.org" <arm@kernel.org>, "vinod.koul@intel.com" <vinod.koul@intel.com>, "djbw@fb.com" <djbw@fb.com>, "arnd@arndb.de" <arnd@arndb.de>, "linux@arm.linux.org.uk" <linux@arm.linux.org.uk> Subject: Re: [PATCH v7] dmaengine: Add MOXA ART DMA engine driver Date: Mon, 5 Aug 2013 17:57:55 +0100 [thread overview] Message-ID: <20130805165755.GD13091@e106331-lin.cambridge.arm.com> (raw) In-Reply-To: <1375713457-5562-1-git-send-email-jonas.jensen@gmail.com> On Mon, Aug 05, 2013 at 03:37:37PM +0100, Jonas Jensen wrote: > The MOXA ART SoC has a DMA controller capable of offloading expensive > memory operations, such as large copies. This patch adds support for > the controller including four channels. Two of these are used to > handle MMC copy on the UC-7112-LX hardware. The remaining two can be > used in a future audio driver or client application. > > Signed-off-by: Jonas Jensen <jonas.jensen@gmail.com> > --- > > Notes: > Thanks for the replies. > > Changes since v6: > > 1. move callback from interrupt context to tasklet > 2. remove callback and callback_param, use those provided by tx_desc > 3. don't rely on structs for register offsets > 4. remove local bool "found" variable from moxart_alloc_chan_resources() > 5. check return value of irq_of_parse_and_map > 6. use devm_request_irq instead of setup_irq > 7. elaborate commit message > > device tree bindings document: > 8. in the example, change "#dma-cells" to "<2>" > > Applies to next-20130805 > > .../devicetree/bindings/dma/moxa,moxart-dma.txt | 21 + > drivers/dma/Kconfig | 7 + > drivers/dma/Makefile | 1 + > drivers/dma/moxart-dma.c | 614 +++++++++++++++++++++ > 4 files changed, 643 insertions(+) > create mode 100644 Documentation/devicetree/bindings/dma/moxa,moxart-dma.txt > create mode 100644 drivers/dma/moxart-dma.c > > diff --git a/Documentation/devicetree/bindings/dma/moxa,moxart-dma.txt b/Documentation/devicetree/bindings/dma/moxa,moxart-dma.txt > new file mode 100644 > index 0000000..5b9f82c > --- /dev/null > +++ b/Documentation/devicetree/bindings/dma/moxa,moxart-dma.txt > @@ -0,0 +1,21 @@ > +MOXA ART DMA Controller > + > +See dma.txt first > + > +Required properties: > + > +- compatible : Must be "moxa,moxart-dma" > +- reg : Should contain registers location and length > +- interrupts : Should contain the interrupt number > +- #dma-cells : Should be 2 > + cell index 0: channel number between 0-3 > + cell index 1: line request number > + > +Example: > + > + dma: dma@90500000 { > + compatible = "moxa,moxart-dma"; > + reg = <0x90500000 0x1000>; > + interrupts = <24 0>; > + #dma-cells = <2>; > + }; Thanks for the updates on this. :) The binding and example look sensible to me; it would be nice if someone familiar with the dma subsystem could check that this has the necessary information. [...] > +static int moxart_alloc_chan_resources(struct dma_chan *chan) > +{ > + struct moxart_dma_chan *mchan = to_moxart_dma_chan(chan); > + int i; > + > + for (i = 0; i < APB_DMA_MAX_CHANNEL; i++) { > + if (i == mchan->ch_num > + && !mchan->allocated) { > + dev_dbg(chan2dev(chan), "%s: allocating channel #%d\n", > + __func__, mchan->ch_num); > + mchan->allocated = true; > + return 0; > + } > + } Come to think of it, why do you need to iterate over all of the channels to handle a particular channel number that you already know, and already have the struct for? I'm not familiar with the dma subsystem, and I couldn't spot when the dma channel is actually assigned/selected prior to this. [...] > +static enum dma_status moxart_tx_status(struct dma_chan *chan, > + dma_cookie_t cookie, > + struct dma_tx_state *txstate) > +{ > + enum dma_status ret; > + > + ret = dma_cookie_status(chan, cookie, txstate); > + if (ret == DMA_SUCCESS || !txstate) > + return ret; > + > + return ret; No special status handling? This function is equivalent to: return dma_cookie_status(chan, cookie, txstate); [...] > +static int moxart_probe(struct platform_device *pdev) > +{ > + struct device *dev = &pdev->dev; > + struct device_node *node = dev->of_node; > + struct resource *res; > + static void __iomem *dma_base_addr; > + int ret, i; > + unsigned int irq; > + struct moxart_dma_chan *mchan; > + struct moxart_dma_container *mdc; > + > + mdc = devm_kzalloc(dev, sizeof(*mdc), GFP_KERNEL); > + if (!mdc) { > + dev_err(dev, "can't allocate DMA container\n"); > + return -ENOMEM; > + } > + > + irq = irq_of_parse_and_map(node, 0); > + if (irq <= 0) { > + dev_err(dev, "irq_of_parse_and_map failed\n"); > + return -EINVAL; > + } > + > + res = platform_get_resource(pdev, IORESOURCE_MEM, 0); > + dma_base_addr = devm_ioremap_resource(dev, res); > + if (IS_ERR(dma_base_addr)) { > + dev_err(dev, "devm_ioremap_resource failed\n"); > + return PTR_ERR(dma_base_addr); > + } > + > + mdc->ctlr = pdev->id; > + spin_lock_init(&mdc->dma_lock); > + > + dma_cap_zero(mdc->dma_slave.cap_mask); > + dma_cap_set(DMA_SLAVE, mdc->dma_slave.cap_mask); > + > + moxart_dma_init(&mdc->dma_slave, dev); > + > + mchan = &mdc->slave_chans[0]; > + for (i = 0; i < APB_DMA_MAX_CHANNEL; i++, mchan++) { > + mchan->ch_num = i; > + mchan->base = dma_base_addr + 0x80 + i * REG_CHAN_SIZE; > + mchan->allocated = 0; > + > + dma_cookie_init(&mchan->chan); > + mchan->chan.device = &mdc->dma_slave; > + list_add_tail(&mchan->chan.device_node, > + &mdc->dma_slave.channels); > + > + dev_dbg(dev, "%s: mchans[%d]: mchan->ch_num=%d mchan->base=%p\n", > + __func__, i, mchan->ch_num, mchan->base); > + } > + > + ret = dma_async_device_register(&mdc->dma_slave); What if this fails? > + platform_set_drvdata(pdev, mdc); > + > + of_dma_controller_register(node, moxart_of_xlate, &moxart_dma_info); > + > + tasklet_init(&mdc->tasklet, moxart_dma_tasklet, (unsigned long)mdc); > + > + devm_request_irq(dev, irq, moxart_dma_interrupt, 0, > + "moxart-dma-engine", mdc); The return value of devm_request_irq should be checked; it might fail. > + > + dev_dbg(dev, "%s: IRQ=%d\n", __func__, irq); > + > + return ret; > +} Thanks, Mark.
WARNING: multiple messages have this Message-ID (diff)
From: mark.rutland@arm.com (Mark Rutland) To: linux-arm-kernel@lists.infradead.org Subject: [PATCH v7] dmaengine: Add MOXA ART DMA engine driver Date: Mon, 5 Aug 2013 17:57:55 +0100 [thread overview] Message-ID: <20130805165755.GD13091@e106331-lin.cambridge.arm.com> (raw) In-Reply-To: <1375713457-5562-1-git-send-email-jonas.jensen@gmail.com> On Mon, Aug 05, 2013 at 03:37:37PM +0100, Jonas Jensen wrote: > The MOXA ART SoC has a DMA controller capable of offloading expensive > memory operations, such as large copies. This patch adds support for > the controller including four channels. Two of these are used to > handle MMC copy on the UC-7112-LX hardware. The remaining two can be > used in a future audio driver or client application. > > Signed-off-by: Jonas Jensen <jonas.jensen@gmail.com> > --- > > Notes: > Thanks for the replies. > > Changes since v6: > > 1. move callback from interrupt context to tasklet > 2. remove callback and callback_param, use those provided by tx_desc > 3. don't rely on structs for register offsets > 4. remove local bool "found" variable from moxart_alloc_chan_resources() > 5. check return value of irq_of_parse_and_map > 6. use devm_request_irq instead of setup_irq > 7. elaborate commit message > > device tree bindings document: > 8. in the example, change "#dma-cells" to "<2>" > > Applies to next-20130805 > > .../devicetree/bindings/dma/moxa,moxart-dma.txt | 21 + > drivers/dma/Kconfig | 7 + > drivers/dma/Makefile | 1 + > drivers/dma/moxart-dma.c | 614 +++++++++++++++++++++ > 4 files changed, 643 insertions(+) > create mode 100644 Documentation/devicetree/bindings/dma/moxa,moxart-dma.txt > create mode 100644 drivers/dma/moxart-dma.c > > diff --git a/Documentation/devicetree/bindings/dma/moxa,moxart-dma.txt b/Documentation/devicetree/bindings/dma/moxa,moxart-dma.txt > new file mode 100644 > index 0000000..5b9f82c > --- /dev/null > +++ b/Documentation/devicetree/bindings/dma/moxa,moxart-dma.txt > @@ -0,0 +1,21 @@ > +MOXA ART DMA Controller > + > +See dma.txt first > + > +Required properties: > + > +- compatible : Must be "moxa,moxart-dma" > +- reg : Should contain registers location and length > +- interrupts : Should contain the interrupt number > +- #dma-cells : Should be 2 > + cell index 0: channel number between 0-3 > + cell index 1: line request number > + > +Example: > + > + dma: dma at 90500000 { > + compatible = "moxa,moxart-dma"; > + reg = <0x90500000 0x1000>; > + interrupts = <24 0>; > + #dma-cells = <2>; > + }; Thanks for the updates on this. :) The binding and example look sensible to me; it would be nice if someone familiar with the dma subsystem could check that this has the necessary information. [...] > +static int moxart_alloc_chan_resources(struct dma_chan *chan) > +{ > + struct moxart_dma_chan *mchan = to_moxart_dma_chan(chan); > + int i; > + > + for (i = 0; i < APB_DMA_MAX_CHANNEL; i++) { > + if (i == mchan->ch_num > + && !mchan->allocated) { > + dev_dbg(chan2dev(chan), "%s: allocating channel #%d\n", > + __func__, mchan->ch_num); > + mchan->allocated = true; > + return 0; > + } > + } Come to think of it, why do you need to iterate over all of the channels to handle a particular channel number that you already know, and already have the struct for? I'm not familiar with the dma subsystem, and I couldn't spot when the dma channel is actually assigned/selected prior to this. [...] > +static enum dma_status moxart_tx_status(struct dma_chan *chan, > + dma_cookie_t cookie, > + struct dma_tx_state *txstate) > +{ > + enum dma_status ret; > + > + ret = dma_cookie_status(chan, cookie, txstate); > + if (ret == DMA_SUCCESS || !txstate) > + return ret; > + > + return ret; No special status handling? This function is equivalent to: return dma_cookie_status(chan, cookie, txstate); [...] > +static int moxart_probe(struct platform_device *pdev) > +{ > + struct device *dev = &pdev->dev; > + struct device_node *node = dev->of_node; > + struct resource *res; > + static void __iomem *dma_base_addr; > + int ret, i; > + unsigned int irq; > + struct moxart_dma_chan *mchan; > + struct moxart_dma_container *mdc; > + > + mdc = devm_kzalloc(dev, sizeof(*mdc), GFP_KERNEL); > + if (!mdc) { > + dev_err(dev, "can't allocate DMA container\n"); > + return -ENOMEM; > + } > + > + irq = irq_of_parse_and_map(node, 0); > + if (irq <= 0) { > + dev_err(dev, "irq_of_parse_and_map failed\n"); > + return -EINVAL; > + } > + > + res = platform_get_resource(pdev, IORESOURCE_MEM, 0); > + dma_base_addr = devm_ioremap_resource(dev, res); > + if (IS_ERR(dma_base_addr)) { > + dev_err(dev, "devm_ioremap_resource failed\n"); > + return PTR_ERR(dma_base_addr); > + } > + > + mdc->ctlr = pdev->id; > + spin_lock_init(&mdc->dma_lock); > + > + dma_cap_zero(mdc->dma_slave.cap_mask); > + dma_cap_set(DMA_SLAVE, mdc->dma_slave.cap_mask); > + > + moxart_dma_init(&mdc->dma_slave, dev); > + > + mchan = &mdc->slave_chans[0]; > + for (i = 0; i < APB_DMA_MAX_CHANNEL; i++, mchan++) { > + mchan->ch_num = i; > + mchan->base = dma_base_addr + 0x80 + i * REG_CHAN_SIZE; > + mchan->allocated = 0; > + > + dma_cookie_init(&mchan->chan); > + mchan->chan.device = &mdc->dma_slave; > + list_add_tail(&mchan->chan.device_node, > + &mdc->dma_slave.channels); > + > + dev_dbg(dev, "%s: mchans[%d]: mchan->ch_num=%d mchan->base=%p\n", > + __func__, i, mchan->ch_num, mchan->base); > + } > + > + ret = dma_async_device_register(&mdc->dma_slave); What if this fails? > + platform_set_drvdata(pdev, mdc); > + > + of_dma_controller_register(node, moxart_of_xlate, &moxart_dma_info); > + > + tasklet_init(&mdc->tasklet, moxart_dma_tasklet, (unsigned long)mdc); > + > + devm_request_irq(dev, irq, moxart_dma_interrupt, 0, > + "moxart-dma-engine", mdc); The return value of devm_request_irq should be checked; it might fail. > + > + dev_dbg(dev, "%s: IRQ=%d\n", __func__, irq); > + > + return ret; > +} Thanks, Mark.
next prev parent reply other threads:[~2013-08-05 16:58 UTC|newest] Thread overview: 80+ messages / expand[flat|nested] mbox.gz Atom feed top 2013-07-10 8:51 [PATCH] dmaengine: Add MOXA ART DMA engine driver Jonas Jensen 2013-07-10 8:51 ` Jonas Jensen 2013-07-10 9:30 ` Russell King - ARM Linux 2013-07-10 9:30 ` Russell King - ARM Linux 2013-07-10 9:48 ` Jonas Jensen 2013-07-10 9:48 ` Jonas Jensen 2013-07-10 12:43 ` [PATCH v2] " Jonas Jensen 2013-07-10 12:43 ` Jonas Jensen 2013-07-17 10:06 ` [PATCH v3] " Jonas Jensen 2013-07-17 10:06 ` Jonas Jensen 2013-07-29 13:44 ` [PATCH v4] " Jonas Jensen 2013-07-29 13:44 ` Jonas Jensen 2013-07-29 16:35 ` Arnd Bergmann 2013-07-29 16:35 ` Arnd Bergmann 2013-08-02 12:28 ` Jonas Jensen 2013-08-02 12:28 ` Jonas Jensen 2013-08-02 19:28 ` Arnd Bergmann 2013-08-02 19:28 ` Arnd Bergmann 2013-08-02 12:03 ` [PATCH v5] " Jonas Jensen 2013-08-02 12:03 ` Jonas Jensen 2013-08-02 13:28 ` [PATCH v6] " Jonas Jensen 2013-08-02 13:28 ` Jonas Jensen 2013-08-02 13:51 ` Russell King - ARM Linux 2013-08-02 13:51 ` Russell King - ARM Linux 2013-08-02 14:09 ` Mark Rutland 2013-08-02 14:09 ` Mark Rutland 2013-08-05 14:37 ` [PATCH v7] " Jonas Jensen 2013-08-05 14:37 ` Jonas Jensen 2013-08-05 16:57 ` Mark Rutland [this message] 2013-08-05 16:57 ` Mark Rutland 2013-08-05 20:49 ` Arnd Bergmann 2013-08-05 20:49 ` Arnd Bergmann 2013-08-06 12:38 ` [PATCH v8] " Jonas Jensen 2013-08-06 12:38 ` Jonas Jensen 2013-08-06 18:42 ` Arnd Bergmann 2013-08-06 18:42 ` Arnd Bergmann 2013-08-07 15:13 ` Mark Rutland 2013-08-07 15:13 ` Mark Rutland 2013-10-07 13:42 ` Jonas Jensen 2013-10-07 13:42 ` Jonas Jensen 2013-10-07 13:13 ` [PATCH v9] " Jonas Jensen 2013-10-07 13:13 ` Jonas Jensen 2013-10-07 14:10 ` [PATCH v10] " Jonas Jensen 2013-10-07 14:10 ` Jonas Jensen 2013-10-07 15:12 ` Mark Rutland 2013-10-07 15:12 ` Mark Rutland 2013-10-07 15:12 ` Mark Rutland 2013-10-08 9:53 ` Jonas Jensen 2013-10-08 9:53 ` Jonas Jensen 2013-10-08 9:53 ` Jonas Jensen 2013-10-08 12:55 ` Mark Rutland 2013-10-08 12:55 ` Mark Rutland 2013-10-08 12:55 ` Mark Rutland 2013-10-08 8:42 ` [PATCH v11] " Jonas Jensen 2013-10-08 8:42 ` Jonas Jensen 2013-11-13 13:59 ` Vinod Koul 2013-11-13 13:59 ` Vinod Koul 2013-11-13 17:16 ` Arnd Bergmann 2013-11-13 17:16 ` Arnd Bergmann 2013-12-06 14:27 ` [PATCH v12] " Jonas Jensen 2013-12-06 14:27 ` Jonas Jensen 2013-12-11 15:13 ` [PATCH v13] " Jonas Jensen 2013-12-11 15:13 ` Jonas Jensen 2013-12-11 21:27 ` Arnd Bergmann 2013-12-11 21:27 ` Arnd Bergmann 2013-12-12 9:16 ` Andy Shevchenko 2013-12-12 9:16 ` Andy Shevchenko 2013-12-12 12:32 ` [PATCH v14] " Jonas Jensen 2013-12-12 12:32 ` Jonas Jensen 2013-12-13 16:02 ` Lars-Peter Clausen 2013-12-13 16:02 ` Lars-Peter Clausen 2013-12-16 10:24 ` [PATCH v15] " Jonas Jensen 2013-12-16 10:24 ` Jonas Jensen 2014-01-17 8:46 ` [PATCH v16] " Jonas Jensen 2014-01-17 8:46 ` Jonas Jensen [not found] ` <1389948365-13999-1-git-send-email-jonas.jensen-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> 2014-01-17 13:29 ` Fwd: " Jonas Jensen 2014-01-17 14:42 ` Arnd Bergmann 2014-01-17 14:42 ` Arnd Bergmann 2014-01-20 7:07 ` Vinod Koul 2014-01-20 7:07 ` Vinod Koul
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=20130805165755.GD13091@e106331-lin.cambridge.arm.com \ --to=mark.rutland@arm.com \ --cc=arm@kernel.org \ --cc=arnd@arndb.de \ --cc=djbw@fb.com \ --cc=jonas.jensen@gmail.com \ --cc=linux-arm-kernel@lists.infradead.org \ --cc=linux-kernel@vger.kernel.org \ --cc=linux@arm.linux.org.uk \ --cc=vinod.koul@intel.com \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.