| Message ID | 20230731101442.792514-5-miquel.raynal@bootlin.com (mailing list archive) |
|---|---|
| State | Changes Requested |
| Headers | show |
| Series | dmaengine: xdma: Cyclic transfers support | expand |
Hi Miquel, kernel test robot noticed the following build warnings: [auto build test WARNING on linus/master] [also build test WARNING on v6.5-rc4 next-20230731] [cannot apply to xilinx-xlnx/master] [If your patch is applied to the wrong git tree, kindly drop us a note. And when submitting patch, we suggest to use '--base' as documented in https://git-scm.com/docs/git-format-patch#_base_tree_information] url: https://github.com/intel-lab-lkp/linux/commits/Miquel-Raynal/dmaengine-xilinx-xdma-Fix-interrupt-vector-setting/20230731-181620 base: linus/master patch link: https://lore.kernel.org/r/20230731101442.792514-5-miquel.raynal%40bootlin.com patch subject: [PATCH 4/4] dmaengine: xilinx: xdma: Support cyclic transfers config: powerpc-randconfig-r026-20230731 (https://download.01.org/0day-ci/archive/20230731/202307312225.9ewyJsVx-lkp@intel.com/config) compiler: clang version 17.0.0 (https://github.com/llvm/llvm-project.git 4a5ac14ee968ff0ad5d2cc1ffa0299048db4c88a) reproduce: (https://download.01.org/0day-ci/archive/20230731/202307312225.9ewyJsVx-lkp@intel.com/reproduce) If you fix the issue in a separate patch/commit (i.e. not just a new version of the same patch/commit), kindly add following tags | Reported-by: kernel test robot <lkp@intel.com> | Closes: https://lore.kernel.org/oe-kbuild-all/202307312225.9ewyJsVx-lkp@intel.com/ All warnings (new ones prefixed by >>): >> drivers/dma/xilinx/xdma.c:716:6: warning: variable 'desc' is used uninitialized whenever 'if' condition is false [-Wsometimes-uninitialized] 716 | if (vd) | ^~ drivers/dma/xilinx/xdma.c:718:7: note: uninitialized use occurs here 718 | if (!desc || !desc->cyclic) { | ^~~~ drivers/dma/xilinx/xdma.c:716:2: note: remove the 'if' if its condition is always true 716 | if (vd) | ^~~~~~~ 717 | desc = to_xdma_desc(vd); | ~~~~~~~~~~~~~~~~ drivers/dma/xilinx/xdma.c:703:24: note: initialize the variable 'desc' to silence this warning 703 | struct xdma_desc *desc; | ^ | = NULL 1 warning generated. vim +716 drivers/dma/xilinx/xdma.c 697 698 static enum dma_status xdma_tx_status(struct dma_chan *chan, dma_cookie_t cookie, 699 struct dma_tx_state *state) 700 { 701 struct xdma_chan *xdma_chan = to_xdma_chan(chan); 702 struct virt_dma_desc *vd; 703 struct xdma_desc *desc; 704 enum dma_status ret; 705 unsigned long flags; 706 unsigned int period_idx; 707 u32 residue = 0; 708 709 ret = dma_cookie_status(chan, cookie, state); 710 if (ret == DMA_COMPLETE || !state) 711 return ret; 712 713 spin_lock_irqsave(&xdma_chan->vchan.lock, flags); 714 715 vd = vchan_find_desc(&xdma_chan->vchan, cookie); > 716 if (vd) 717 desc = to_xdma_desc(vd); 718 if (!desc || !desc->cyclic) { 719 spin_unlock_irqrestore(&xdma_chan->vchan.lock, flags); 720 return ret; 721 } 722 723 period_idx = desc->completed_desc_num % desc->periods; 724 residue = (desc->periods - period_idx) * desc->period_size; 725 726 spin_unlock_irqrestore(&xdma_chan->vchan.lock, flags); 727 728 dma_set_residue(state, residue); 729 730 return ret; 731 } 732
On 7/31/23 03:14, Miquel Raynal wrote: > In order to use this dmaengine with sound devices, let's add cyclic > transfers support. Most of the code is reused from the existing > scatter-gather implementation, only the final linking between > descriptors, the control fields (to trigger interrupts more often) and > the interrupt handling are really different. > > This controller supports up to 32 adjacent descriptors, we assume this > is way more than enough for the purpose of cyclic transfers and limit to > 32 the number of cycled descriptors. This way, we simplify a lot the > overall handling of the descriptors. > > Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com> > --- > drivers/dma/xilinx/xdma-regs.h | 2 + > drivers/dma/xilinx/xdma.c | 173 +++++++++++++++++++++++++++++++-- > 2 files changed, 169 insertions(+), 6 deletions(-) > > diff --git a/drivers/dma/xilinx/xdma-regs.h b/drivers/dma/xilinx/xdma-regs.h > index dd98b4526b90..e641a5083e14 100644 > --- a/drivers/dma/xilinx/xdma-regs.h > +++ b/drivers/dma/xilinx/xdma-regs.h > @@ -44,6 +44,8 @@ > FIELD_PREP(XDMA_DESC_FLAGS_BITS, (flag))) > #define XDMA_DESC_CONTROL_LAST \ > XDMA_DESC_CONTROL(1, XDMA_DESC_STOPPED | XDMA_DESC_COMPLETED) > +#define XDMA_DESC_CONTROL_CYCLIC \ > + XDMA_DESC_CONTROL(1, XDMA_DESC_COMPLETED) > > /* > * Descriptor for a single contiguous memory block transfer. > diff --git a/drivers/dma/xilinx/xdma.c b/drivers/dma/xilinx/xdma.c > index 40983d9355c4..be3a212be30c 100644 > --- a/drivers/dma/xilinx/xdma.c > +++ b/drivers/dma/xilinx/xdma.c > @@ -83,6 +83,9 @@ struct xdma_chan { > * @dblk_num: Number of hardware descriptor blocks > * @desc_num: Number of hardware descriptors > * @completed_desc_num: Completed hardware descriptors > + * @cyclic: Cyclic transfer vs. scatter-gather > + * @periods: Number of periods in the cyclic transfer > + * @period_size: Size of a period in bytes in cyclic transfers > */ > struct xdma_desc { > struct virt_dma_desc vdesc; > @@ -93,6 +96,9 @@ struct xdma_desc { > u32 dblk_num; > u32 desc_num; > u32 completed_desc_num; > + bool cyclic; > + u32 periods; > + u32 period_size; > }; > > #define XDMA_DEV_STATUS_REG_DMA BIT(0) > @@ -174,6 +180,25 @@ static void xdma_link_sg_desc_blocks(struct xdma_desc *sw_desc) > desc->control = cpu_to_le32(XDMA_DESC_CONTROL_LAST); > } > > +/** > + * xdma_link_cyclic_desc_blocks - Link cyclic descriptor blocks for DMA transfer > + * @sw_desc: Tx descriptor pointer > + */ > +static void xdma_link_cyclic_desc_blocks(struct xdma_desc *sw_desc) > +{ > + struct xdma_desc_block *block; > + struct xdma_hw_desc *desc; > + int i; > + > + block = sw_desc->desc_blocks; > + for (i = 0; i < sw_desc->desc_num - 1; i++) { > + desc = block->virt_addr + i * XDMA_DESC_SIZE; > + desc->next_desc = cpu_to_le64(block->dma_addr + ((i + 1) * XDMA_DESC_SIZE)); > + } > + desc = block->virt_addr + i * XDMA_DESC_SIZE; > + desc->next_desc = cpu_to_le64(block->dma_addr); > +} > + > static inline struct xdma_chan *to_xdma_chan(struct dma_chan *chan) > { > return container_of(chan, struct xdma_chan, vchan.chan); > @@ -233,7 +258,7 @@ static void xdma_free_desc(struct virt_dma_desc *vdesc) > * @desc_num: Number of hardware descriptors > */ > static struct xdma_desc * > -xdma_alloc_desc(struct xdma_chan *chan, u32 desc_num) > +xdma_alloc_desc(struct xdma_chan *chan, u32 desc_num, bool cyclic) > { > struct xdma_desc *sw_desc; > struct xdma_hw_desc *desc; > @@ -249,13 +274,17 @@ xdma_alloc_desc(struct xdma_chan *chan, u32 desc_num) > > sw_desc->chan = chan; > sw_desc->desc_num = desc_num; > + sw_desc->cyclic = cyclic; > dblk_num = DIV_ROUND_UP(desc_num, XDMA_DESC_ADJACENT); > sw_desc->desc_blocks = kcalloc(dblk_num, sizeof(*sw_desc->desc_blocks), > GFP_NOWAIT); > if (!sw_desc->desc_blocks) > goto failed; > > - control = XDMA_DESC_CONTROL(1, 0); > + if (cyclic) > + control = XDMA_DESC_CONTROL_CYCLIC; > + else > + control = XDMA_DESC_CONTROL(1, 0); > > sw_desc->dblk_num = dblk_num; > for (i = 0; i < sw_desc->dblk_num; i++) { > @@ -269,7 +298,10 @@ xdma_alloc_desc(struct xdma_chan *chan, u32 desc_num) > desc[j].control = cpu_to_le32(control); > } > > - xdma_link_sg_desc_blocks(sw_desc); > + if (cyclic) > + xdma_link_cyclic_desc_blocks(sw_desc); > + else > + xdma_link_sg_desc_blocks(sw_desc); > > return sw_desc; > > @@ -469,7 +501,7 @@ xdma_prep_device_sg(struct dma_chan *chan, struct scatterlist *sgl, > for_each_sg(sgl, sg, sg_len, i) > desc_num += DIV_ROUND_UP(sg_dma_len(sg), XDMA_DESC_BLEN_MAX); > > - sw_desc = xdma_alloc_desc(xdma_chan, desc_num); > + sw_desc = xdma_alloc_desc(xdma_chan, desc_num, false); > if (!sw_desc) > return NULL; > sw_desc->dir = dir; > @@ -524,6 +556,89 @@ xdma_prep_device_sg(struct dma_chan *chan, struct scatterlist *sgl, > return NULL; > } > > +/** > + * xdma_prep_dma_cyclic - prepare for cyclic DMA transactions > + * @chan: DMA channel pointer > + * @address: Device DMA address to access > + * @size: Total length to transfer > + * @period_size: Period size to use for each transfer > + * @dir: Transfer direction > + * @flags: Transfer ack flags > + */ > +static struct dma_async_tx_descriptor * > +xdma_prep_dma_cyclic(struct dma_chan *chan, dma_addr_t address, > + size_t size, size_t period_size, > + enum dma_transfer_direction dir, > + unsigned long flags) > +{ > + struct xdma_chan *xdma_chan = to_xdma_chan(chan); > + struct xdma_device *xdev = xdma_chan->xdev_hdl; > + unsigned int periods = size / period_size; What if size is not multiple of period_size? > + struct dma_async_tx_descriptor *tx_desc; > + u64 addr, dev_addr, *src, *dst; > + struct xdma_desc_block *dblk; > + struct xdma_hw_desc *desc; > + struct xdma_desc *sw_desc; > + unsigned int i; > + > + /* > + * Simplify the whole logic by preventing an abnormally high number of > + * periods and periods size. > + */ > + if (period_size > XDMA_DESC_BLEN_MAX) { > + xdma_err(xdev, "period size limited to %lu bytes\n", XDMA_DESC_BLEN_MAX); > + return NULL; > + } > + > + if (periods > XDMA_DESC_ADJACENT) { > + xdma_err(xdev, "number of periods limited to %u\n", XDMA_DESC_ADJACENT); > + return NULL; > + } > + > + sw_desc = xdma_alloc_desc(xdma_chan, periods, true); > + if (!sw_desc) > + return NULL; > + > + sw_desc->periods = periods; > + sw_desc->period_size = period_size; > + sw_desc->dir = dir; > + > + if (dir == DMA_MEM_TO_DEV) { > + dev_addr = xdma_chan->cfg.dst_addr; > + src = &addr; > + dst = &dev_addr; > + } else { > + dev_addr = xdma_chan->cfg.src_addr; > + src = &dev_addr; > + dst = &addr; > + } > + > + dblk = sw_desc->desc_blocks; > + desc = dblk->virt_addr; > + for (i = 0; i < periods; i++) { > + addr = address; > + > + /* fill hardware descriptor */ > + desc->bytes = cpu_to_le32(period_size); > + desc->src_addr = cpu_to_le64(*src); > + desc->dst_addr = cpu_to_le64(*dst); > + > + desc++; > + address += period_size; > + } > + > + tx_desc = vchan_tx_prep(&xdma_chan->vchan, &sw_desc->vdesc, flags); > + if (!tx_desc) > + goto failed; > + > + return tx_desc; > + > +failed: > + xdma_free_desc(&sw_desc->vdesc); > + > + return NULL; > +} > + > /** > * xdma_device_config - Configure the DMA channel > * @chan: DMA channel > @@ -583,7 +698,36 @@ static int xdma_alloc_chan_resources(struct dma_chan *chan) > static enum dma_status xdma_tx_status(struct dma_chan *chan, dma_cookie_t cookie, > struct dma_tx_state *state) > { > - return dma_cookie_status(chan, cookie, state); > + struct xdma_chan *xdma_chan = to_xdma_chan(chan); > + struct virt_dma_desc *vd; > + struct xdma_desc *desc; > + enum dma_status ret; > + unsigned long flags; > + unsigned int period_idx; > + u32 residue = 0; > + > + ret = dma_cookie_status(chan, cookie, state); > + if (ret == DMA_COMPLETE || !state) probably do not need to check state. Or at least check before calling dma_cookie_status. > + return ret; > + > + spin_lock_irqsave(&xdma_chan->vchan.lock, flags); > + > + vd = vchan_find_desc(&xdma_chan->vchan, cookie); > + if (vd) > + desc = to_xdma_desc(vd); > + if (!desc || !desc->cyclic) { desc is un-initialized if vd is NULL. > + spin_unlock_irqrestore(&xdma_chan->vchan.lock, flags); > + return ret; > + } > + > + period_idx = desc->completed_desc_num % desc->periods; > + residue = (desc->periods - period_idx) * desc->period_size; > + > + spin_unlock_irqrestore(&xdma_chan->vchan.lock, flags); > + > + dma_set_residue(state, residue); > + > + return ret; > } > > /** > @@ -599,6 +743,7 @@ static irqreturn_t xdma_channel_isr(int irq, void *dev_id) > struct virt_dma_desc *vd; > struct xdma_desc *desc; > int ret; > + u32 st; > > spin_lock(&xchan->vchan.lock); > > @@ -617,6 +762,19 @@ static irqreturn_t xdma_channel_isr(int irq, void *dev_id) > goto out; > > desc->completed_desc_num += complete_desc_num; > + > + if (desc->cyclic) { > + ret = regmap_read(xdev->rmap, xchan->base + XDMA_CHAN_STATUS, > + &st); > + if (ret) > + goto out; > + > + regmap_write(xdev->rmap, xchan->base + XDMA_CHAN_STATUS, st); What does reading/writing channel status register do here? Thanks, Lizhi > + > + vchan_cyclic_callback(vd); > + goto out; > + } > + > /* > * if all data blocks are transferred, remove and complete the request > */ > @@ -630,7 +788,7 @@ static irqreturn_t xdma_channel_isr(int irq, void *dev_id) > complete_desc_num != XDMA_DESC_BLOCK_NUM * XDMA_DESC_ADJACENT) > goto out; > > - /* transfer the rest of data */ > + /* transfer the rest of data (SG only) */ > xdma_xfer_start(xchan); > > out: > @@ -930,8 +1088,10 @@ static int xdma_probe(struct platform_device *pdev) > > dma_cap_set(DMA_SLAVE, xdev->dma_dev.cap_mask); > dma_cap_set(DMA_PRIVATE, xdev->dma_dev.cap_mask); > + dma_cap_set(DMA_CYCLIC, xdev->dma_dev.cap_mask); > > xdev->dma_dev.dev = &pdev->dev; > + xdev->dma_dev.residue_granularity = DMA_RESIDUE_GRANULARITY_SEGMENT; > xdev->dma_dev.device_free_chan_resources = xdma_free_chan_resources; > xdev->dma_dev.device_alloc_chan_resources = xdma_alloc_chan_resources; > xdev->dma_dev.device_tx_status = xdma_tx_status; > @@ -941,6 +1101,7 @@ static int xdma_probe(struct platform_device *pdev) > xdev->dma_dev.filter.map = pdata->device_map; > xdev->dma_dev.filter.mapcnt = pdata->device_map_cnt; > xdev->dma_dev.filter.fn = xdma_filter_fn; > + xdev->dma_dev.device_prep_dma_cyclic = xdma_prep_dma_cyclic; > > ret = dma_async_device_register(&xdev->dma_dev); > if (ret) {
Hi Lizhi, > > + * xdma_prep_dma_cyclic - prepare for cyclic DMA transactions > > + * @chan: DMA channel pointer > > + * @address: Device DMA address to access > > + * @size: Total length to transfer > > + * @period_size: Period size to use for each transfer > > + * @dir: Transfer direction > > + * @flags: Transfer ack flags > > + */ > > +static struct dma_async_tx_descriptor * > > +xdma_prep_dma_cyclic(struct dma_chan *chan, dma_addr_t address, > > + size_t size, size_t period_size, > > + enum dma_transfer_direction dir, > > + unsigned long flags) > > +{ > > + struct xdma_chan *xdma_chan = to_xdma_chan(chan); > > + struct xdma_device *xdev = xdma_chan->xdev_hdl; > > + unsigned int periods = size / period_size; > What if size is not multiple of period_size? Can this really happen? I would have expected a bug if size was not a multiple of period_size, because the headers explicitly tell that we should expect an interrupt after each period and we should loop over when we reach the last one. This makes it very impractical to handle the situation you mention. > > + struct dma_async_tx_descriptor *tx_desc; > > + u64 addr, dev_addr, *src, *dst; > > + struct xdma_desc_block *dblk; > > + struct xdma_hw_desc *desc; > > + struct xdma_desc *sw_desc; > > + unsigned int i; > > + > > + /* > > + * Simplify the whole logic by preventing an abnormally high number of > > + * periods and periods size. > > + */ > > + if (period_size > XDMA_DESC_BLEN_MAX) { > > + xdma_err(xdev, "period size limited to %lu bytes\n", XDMA_DESC_BLEN_MAX); > > + return NULL; > > + } > > + > > + if (periods > XDMA_DESC_ADJACENT) { > > + xdma_err(xdev, "number of periods limited to %u\n", XDMA_DESC_ADJACENT); > > + return NULL; > > + } > > + > > + sw_desc = xdma_alloc_desc(xdma_chan, periods, true); > > + if (!sw_desc) > > + return NULL; > > + > > + sw_desc->periods = periods; > > + sw_desc->period_size = period_size; > > + sw_desc->dir = dir; > > + > > + if (dir == DMA_MEM_TO_DEV) { > > + dev_addr = xdma_chan->cfg.dst_addr; > > + src = &addr; > > + dst = &dev_addr; > > + } else { > > + dev_addr = xdma_chan->cfg.src_addr; > > + src = &dev_addr; > > + dst = &addr; > > + } > > + > > + dblk = sw_desc->desc_blocks; > > + desc = dblk->virt_addr; > > + for (i = 0; i < periods; i++) { > > + addr = address; > > + > > + /* fill hardware descriptor */ > > + desc->bytes = cpu_to_le32(period_size); > > + desc->src_addr = cpu_to_le64(*src); > > + desc->dst_addr = cpu_to_le64(*dst); > > + > > + desc++; > > + address += period_size; > > + } > > + > > + tx_desc = vchan_tx_prep(&xdma_chan->vchan, &sw_desc->vdesc, flags); > > + if (!tx_desc) > > + goto failed; > > + > > + return tx_desc; > > + > > +failed: > > + xdma_free_desc(&sw_desc->vdesc); > > + > > + return NULL; > > +} > > + > > /** > > * xdma_device_config - Configure the DMA channel > > * @chan: DMA channel > > @@ -583,7 +698,36 @@ static int xdma_alloc_chan_resources(struct dma_chan *chan) > > static enum dma_status xdma_tx_status(struct dma_chan *chan, dma_cookie_t cookie, > > struct dma_tx_state *state) > > { > > - return dma_cookie_status(chan, cookie, state); > > + struct xdma_chan *xdma_chan = to_xdma_chan(chan); > > + struct virt_dma_desc *vd; > > + struct xdma_desc *desc; > > + enum dma_status ret; > > + unsigned long flags; > > + unsigned int period_idx; > > + u32 residue = 0; > > + > > + ret = dma_cookie_status(chan, cookie, state); > > + if (ret == DMA_COMPLETE || !state) > > probably do not need to check state. Or at least check before calling dma_cookie_status. Good idea. > > + return ret; > > + > > + spin_lock_irqsave(&xdma_chan->vchan.lock, flags); > > + > > + vd = vchan_find_desc(&xdma_chan->vchan, cookie); > > + if (vd) > > + desc = to_xdma_desc(vd); > > + if (!desc || !desc->cyclic) { > > desc is un-initialized if vd is NULL. Yes, handled in v2. > > + spin_unlock_irqrestore(&xdma_chan->vchan.lock, flags); > > + return ret; > > + } > > + > > + period_idx = desc->completed_desc_num % desc->periods; > > + residue = (desc->periods - period_idx) * desc->period_size; > > + > > + spin_unlock_irqrestore(&xdma_chan->vchan.lock, flags); > > + > > + dma_set_residue(state, residue); > > + > > + return ret; > > } > > > /** > > @@ -599,6 +743,7 @@ static irqreturn_t xdma_channel_isr(int irq, void *dev_id) > > struct virt_dma_desc *vd; > > struct xdma_desc *desc; > > int ret; > > + u32 st; > > > spin_lock(&xchan->vchan.lock); > > > @@ -617,6 +762,19 @@ static irqreturn_t xdma_channel_isr(int irq, void *dev_id) > > goto out; > > > desc->completed_desc_num += complete_desc_num; > > + > > + if (desc->cyclic) { > > + ret = regmap_read(xdev->rmap, xchan->base + XDMA_CHAN_STATUS, > > + &st); > > + if (ret) > > + goto out; > > + > > + regmap_write(xdev->rmap, xchan->base + XDMA_CHAN_STATUS, st); > > What does reading/writing channel status register do here? It clears the status register to allow the next interrupt to trigger. Thanks, Miquèl
diff --git a/drivers/dma/xilinx/xdma-regs.h b/drivers/dma/xilinx/xdma-regs.h index dd98b4526b90..e641a5083e14 100644 --- a/drivers/dma/xilinx/xdma-regs.h +++ b/drivers/dma/xilinx/xdma-regs.h @@ -44,6 +44,8 @@ FIELD_PREP(XDMA_DESC_FLAGS_BITS, (flag))) #define XDMA_DESC_CONTROL_LAST \ XDMA_DESC_CONTROL(1, XDMA_DESC_STOPPED | XDMA_DESC_COMPLETED) +#define XDMA_DESC_CONTROL_CYCLIC \ + XDMA_DESC_CONTROL(1, XDMA_DESC_COMPLETED) /* * Descriptor for a single contiguous memory block transfer. diff --git a/drivers/dma/xilinx/xdma.c b/drivers/dma/xilinx/xdma.c index 40983d9355c4..be3a212be30c 100644 --- a/drivers/dma/xilinx/xdma.c +++ b/drivers/dma/xilinx/xdma.c @@ -83,6 +83,9 @@ struct xdma_chan { * @dblk_num: Number of hardware descriptor blocks * @desc_num: Number of hardware descriptors * @completed_desc_num: Completed hardware descriptors + * @cyclic: Cyclic transfer vs. scatter-gather + * @periods: Number of periods in the cyclic transfer + * @period_size: Size of a period in bytes in cyclic transfers */ struct xdma_desc { struct virt_dma_desc vdesc; @@ -93,6 +96,9 @@ struct xdma_desc { u32 dblk_num; u32 desc_num; u32 completed_desc_num; + bool cyclic; + u32 periods; + u32 period_size; }; #define XDMA_DEV_STATUS_REG_DMA BIT(0) @@ -174,6 +180,25 @@ static void xdma_link_sg_desc_blocks(struct xdma_desc *sw_desc) desc->control = cpu_to_le32(XDMA_DESC_CONTROL_LAST); } +/** + * xdma_link_cyclic_desc_blocks - Link cyclic descriptor blocks for DMA transfer + * @sw_desc: Tx descriptor pointer + */ +static void xdma_link_cyclic_desc_blocks(struct xdma_desc *sw_desc) +{ + struct xdma_desc_block *block; + struct xdma_hw_desc *desc; + int i; + + block = sw_desc->desc_blocks; + for (i = 0; i < sw_desc->desc_num - 1; i++) { + desc = block->virt_addr + i * XDMA_DESC_SIZE; + desc->next_desc = cpu_to_le64(block->dma_addr + ((i + 1) * XDMA_DESC_SIZE)); + } + desc = block->virt_addr + i * XDMA_DESC_SIZE; + desc->next_desc = cpu_to_le64(block->dma_addr); +} + static inline struct xdma_chan *to_xdma_chan(struct dma_chan *chan) { return container_of(chan, struct xdma_chan, vchan.chan); @@ -233,7 +258,7 @@ static void xdma_free_desc(struct virt_dma_desc *vdesc) * @desc_num: Number of hardware descriptors */ static struct xdma_desc * -xdma_alloc_desc(struct xdma_chan *chan, u32 desc_num) +xdma_alloc_desc(struct xdma_chan *chan, u32 desc_num, bool cyclic) { struct xdma_desc *sw_desc; struct xdma_hw_desc *desc; @@ -249,13 +274,17 @@ xdma_alloc_desc(struct xdma_chan *chan, u32 desc_num) sw_desc->chan = chan; sw_desc->desc_num = desc_num; + sw_desc->cyclic = cyclic; dblk_num = DIV_ROUND_UP(desc_num, XDMA_DESC_ADJACENT); sw_desc->desc_blocks = kcalloc(dblk_num, sizeof(*sw_desc->desc_blocks), GFP_NOWAIT); if (!sw_desc->desc_blocks) goto failed; - control = XDMA_DESC_CONTROL(1, 0); + if (cyclic) + control = XDMA_DESC_CONTROL_CYCLIC; + else + control = XDMA_DESC_CONTROL(1, 0); sw_desc->dblk_num = dblk_num; for (i = 0; i < sw_desc->dblk_num; i++) { @@ -269,7 +298,10 @@ xdma_alloc_desc(struct xdma_chan *chan, u32 desc_num) desc[j].control = cpu_to_le32(control); } - xdma_link_sg_desc_blocks(sw_desc); + if (cyclic) + xdma_link_cyclic_desc_blocks(sw_desc); + else + xdma_link_sg_desc_blocks(sw_desc); return sw_desc; @@ -469,7 +501,7 @@ xdma_prep_device_sg(struct dma_chan *chan, struct scatterlist *sgl, for_each_sg(sgl, sg, sg_len, i) desc_num += DIV_ROUND_UP(sg_dma_len(sg), XDMA_DESC_BLEN_MAX); - sw_desc = xdma_alloc_desc(xdma_chan, desc_num); + sw_desc = xdma_alloc_desc(xdma_chan, desc_num, false); if (!sw_desc) return NULL; sw_desc->dir = dir; @@ -524,6 +556,89 @@ xdma_prep_device_sg(struct dma_chan *chan, struct scatterlist *sgl, return NULL; } +/** + * xdma_prep_dma_cyclic - prepare for cyclic DMA transactions + * @chan: DMA channel pointer + * @address: Device DMA address to access + * @size: Total length to transfer + * @period_size: Period size to use for each transfer + * @dir: Transfer direction + * @flags: Transfer ack flags + */ +static struct dma_async_tx_descriptor * +xdma_prep_dma_cyclic(struct dma_chan *chan, dma_addr_t address, + size_t size, size_t period_size, + enum dma_transfer_direction dir, + unsigned long flags) +{ + struct xdma_chan *xdma_chan = to_xdma_chan(chan); + struct xdma_device *xdev = xdma_chan->xdev_hdl; + unsigned int periods = size / period_size; + struct dma_async_tx_descriptor *tx_desc; + u64 addr, dev_addr, *src, *dst; + struct xdma_desc_block *dblk; + struct xdma_hw_desc *desc; + struct xdma_desc *sw_desc; + unsigned int i; + + /* + * Simplify the whole logic by preventing an abnormally high number of + * periods and periods size. + */ + if (period_size > XDMA_DESC_BLEN_MAX) { + xdma_err(xdev, "period size limited to %lu bytes\n", XDMA_DESC_BLEN_MAX); + return NULL; + } + + if (periods > XDMA_DESC_ADJACENT) { + xdma_err(xdev, "number of periods limited to %u\n", XDMA_DESC_ADJACENT); + return NULL; + } + + sw_desc = xdma_alloc_desc(xdma_chan, periods, true); + if (!sw_desc) + return NULL; + + sw_desc->periods = periods; + sw_desc->period_size = period_size; + sw_desc->dir = dir; + + if (dir == DMA_MEM_TO_DEV) { + dev_addr = xdma_chan->cfg.dst_addr; + src = &addr; + dst = &dev_addr; + } else { + dev_addr = xdma_chan->cfg.src_addr; + src = &dev_addr; + dst = &addr; + } + + dblk = sw_desc->desc_blocks; + desc = dblk->virt_addr; + for (i = 0; i < periods; i++) { + addr = address; + + /* fill hardware descriptor */ + desc->bytes = cpu_to_le32(period_size); + desc->src_addr = cpu_to_le64(*src); + desc->dst_addr = cpu_to_le64(*dst); + + desc++; + address += period_size; + } + + tx_desc = vchan_tx_prep(&xdma_chan->vchan, &sw_desc->vdesc, flags); + if (!tx_desc) + goto failed; + + return tx_desc; + +failed: + xdma_free_desc(&sw_desc->vdesc); + + return NULL; +} + /** * xdma_device_config - Configure the DMA channel * @chan: DMA channel @@ -583,7 +698,36 @@ static int xdma_alloc_chan_resources(struct dma_chan *chan) static enum dma_status xdma_tx_status(struct dma_chan *chan, dma_cookie_t cookie, struct dma_tx_state *state) { - return dma_cookie_status(chan, cookie, state); + struct xdma_chan *xdma_chan = to_xdma_chan(chan); + struct virt_dma_desc *vd; + struct xdma_desc *desc; + enum dma_status ret; + unsigned long flags; + unsigned int period_idx; + u32 residue = 0; + + ret = dma_cookie_status(chan, cookie, state); + if (ret == DMA_COMPLETE || !state) + return ret; + + spin_lock_irqsave(&xdma_chan->vchan.lock, flags); + + vd = vchan_find_desc(&xdma_chan->vchan, cookie); + if (vd) + desc = to_xdma_desc(vd); + if (!desc || !desc->cyclic) { + spin_unlock_irqrestore(&xdma_chan->vchan.lock, flags); + return ret; + } + + period_idx = desc->completed_desc_num % desc->periods; + residue = (desc->periods - period_idx) * desc->period_size; + + spin_unlock_irqrestore(&xdma_chan->vchan.lock, flags); + + dma_set_residue(state, residue); + + return ret; } /** @@ -599,6 +743,7 @@ static irqreturn_t xdma_channel_isr(int irq, void *dev_id) struct virt_dma_desc *vd; struct xdma_desc *desc; int ret; + u32 st; spin_lock(&xchan->vchan.lock); @@ -617,6 +762,19 @@ static irqreturn_t xdma_channel_isr(int irq, void *dev_id) goto out; desc->completed_desc_num += complete_desc_num; + + if (desc->cyclic) { + ret = regmap_read(xdev->rmap, xchan->base + XDMA_CHAN_STATUS, + &st); + if (ret) + goto out; + + regmap_write(xdev->rmap, xchan->base + XDMA_CHAN_STATUS, st); + + vchan_cyclic_callback(vd); + goto out; + } + /* * if all data blocks are transferred, remove and complete the request */ @@ -630,7 +788,7 @@ static irqreturn_t xdma_channel_isr(int irq, void *dev_id) complete_desc_num != XDMA_DESC_BLOCK_NUM * XDMA_DESC_ADJACENT) goto out; - /* transfer the rest of data */ + /* transfer the rest of data (SG only) */ xdma_xfer_start(xchan); out: @@ -930,8 +1088,10 @@ static int xdma_probe(struct platform_device *pdev) dma_cap_set(DMA_SLAVE, xdev->dma_dev.cap_mask); dma_cap_set(DMA_PRIVATE, xdev->dma_dev.cap_mask); + dma_cap_set(DMA_CYCLIC, xdev->dma_dev.cap_mask); xdev->dma_dev.dev = &pdev->dev; + xdev->dma_dev.residue_granularity = DMA_RESIDUE_GRANULARITY_SEGMENT; xdev->dma_dev.device_free_chan_resources = xdma_free_chan_resources; xdev->dma_dev.device_alloc_chan_resources = xdma_alloc_chan_resources; xdev->dma_dev.device_tx_status = xdma_tx_status; @@ -941,6 +1101,7 @@ static int xdma_probe(struct platform_device *pdev) xdev->dma_dev.filter.map = pdata->device_map; xdev->dma_dev.filter.mapcnt = pdata->device_map_cnt; xdev->dma_dev.filter.fn = xdma_filter_fn; + xdev->dma_dev.device_prep_dma_cyclic = xdma_prep_dma_cyclic; ret = dma_async_device_register(&xdev->dma_dev); if (ret) {
In order to use this dmaengine with sound devices, let's add cyclic transfers support. Most of the code is reused from the existing scatter-gather implementation, only the final linking between descriptors, the control fields (to trigger interrupts more often) and the interrupt handling are really different. This controller supports up to 32 adjacent descriptors, we assume this is way more than enough for the purpose of cyclic transfers and limit to 32 the number of cycled descriptors. This way, we simplify a lot the overall handling of the descriptors. Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com> --- drivers/dma/xilinx/xdma-regs.h | 2 + drivers/dma/xilinx/xdma.c | 173 +++++++++++++++++++++++++++++++-- 2 files changed, 169 insertions(+), 6 deletions(-)