4 files changed, 65 insertions, 30 deletions

diff --git a/drivers/dma/at_hdmac_regs.h b/drivers/dma/at_hdmac_regs.h
index 7f58f06..ef3f227 100644
--- a/drivers/dma/at_hdmac_regs.h
+++ b/drivers/dma/at_hdmac_regs.h
@@ -385,7 +385,7 @@ static void vdbg_dump_regs(struct at_dma_chan *atchan) {}
 static void atc_dump_lli(struct at_dma_chan *atchan, struct at_lli *lli)
 {
 	dev_crit(chan2dev(&atchan->chan_common),
-		 "  desc: s%pad d%pad ctrl0x%x:0x%x l0x%pad\n",
+		 "desc: s%pad d%pad ctrl0x%x:0x%x l%pad\n",
 		 &lli->saddr, &lli->daddr,
 		 lli->ctrla, lli->ctrlb, &lli->dscr);
 }
diff --git a/drivers/dma/coh901318.c b/drivers/dma/coh901318.c
index 74794c9..da74fd7 100644
--- a/drivers/dma/coh901318.c
+++ b/drivers/dma/coh901318.c
@@ -1319,8 +1319,8 @@ static void coh901318_list_print(struct coh901318_chan *cohc,
 	int i = 0;
 
 	while (l) {
-		dev_vdbg(COHC_2_DEV(cohc), "i %d, lli %p, ctrl 0x%x, src 0x%pad"
-			 ", dst 0x%pad, link 0x%pad virt_link_addr 0x%p\n",
+		dev_vdbg(COHC_2_DEV(cohc), "i %d, lli %p, ctrl 0x%x, src %pad"
+			 ", dst %pad, link %pad virt_link_addr 0x%p\n",
 			 i, l, l->control, &l->src_addr, &l->dst_addr,
 			 &l->link_addr, l->virt_link_addr);
 		i++;
@@ -2231,7 +2231,7 @@ coh901318_prep_memcpy(struct dma_chan *chan, dma_addr_t dest, dma_addr_t src,
 	spin_lock_irqsave(&cohc->lock, flg);
 
 	dev_vdbg(COHC_2_DEV(cohc),
-		 "[%s] channel %d src 0x%pad dest 0x%pad size %zu\n",
+		 "[%s] channel %d src %pad dest %pad size %zu\n",
 		 __func__, cohc->id, &src, &dest, size);
 
 	if (flags & DMA_PREP_INTERRUPT)
diff --git a/drivers/dma/dma-axi-dmac.c b/drivers/dma/dma-axi-dmac.c
index 7f0b9aa..2419fe5 100644
--- a/drivers/dma/dma-axi-dmac.c
+++ b/drivers/dma/dma-axi-dmac.c
@@ -72,6 +72,9 @@
 
 #define AXI_DMAC_FLAG_CYCLIC		BIT(0)
 
+/* The maximum ID allocated by the hardware is 31 */
+#define AXI_DMAC_SG_UNUSED 32U
+
 struct axi_dmac_sg {
 	dma_addr_t src_addr;
 	dma_addr_t dest_addr;
@@ -80,6 +83,7 @@ struct axi_dmac_sg {
 	unsigned int dest_stride;
 	unsigned int src_stride;
 	unsigned int id;
+	bool schedule_when_free;
 };
 
 struct axi_dmac_desc {
@@ -200,11 +204,21 @@ static void axi_dmac_start_transfer(struct axi_dmac_chan *chan)
 	}
 	sg = &desc->sg[desc->num_submitted];
 
+	/* Already queued in cyclic mode. Wait for it to finish */
+	if (sg->id != AXI_DMAC_SG_UNUSED) {
+		sg->schedule_when_free = true;
+		return;
+	}
+
 	desc->num_submitted++;
-	if (desc->num_submitted == desc->num_sgs)
-		chan->next_desc = NULL;
-	else
+	if (desc->num_submitted == desc->num_sgs) {
+		if (desc->cyclic)
+			desc->num_submitted = 0; /* Start again */
+		else
+			chan->next_desc = NULL;
+	} else {
 		chan->next_desc = desc;
+	}
 
 	sg->id = axi_dmac_read(dmac, AXI_DMAC_REG_TRANSFER_ID);
 
@@ -220,9 +234,11 @@ static void axi_dmac_start_transfer(struct axi_dmac_chan *chan)
 
 	/*
 	 * If the hardware supports cyclic transfers and there is no callback to
-	 * call, enable hw cyclic mode to avoid unnecessary interrupts.
+	 * call and only a single segment, enable hw cyclic mode to avoid
+	 * unnecessary interrupts.
 	 */
-	if (chan->hw_cyclic && desc->cyclic && !desc->vdesc.tx.callback)
+	if (chan->hw_cyclic && desc->cyclic && !desc->vdesc.tx.callback &&
+		desc->num_sgs == 1)
 		flags |= AXI_DMAC_FLAG_CYCLIC;
 
 	axi_dmac_write(dmac, AXI_DMAC_REG_X_LENGTH, sg->x_len - 1);
@@ -237,37 +253,52 @@ static struct axi_dmac_desc *axi_dmac_active_desc(struct axi_dmac_chan *chan)
 		struct axi_dmac_desc, vdesc.node);
 }
 
-static void axi_dmac_transfer_done(struct axi_dmac_chan *chan,
+static bool axi_dmac_transfer_done(struct axi_dmac_chan *chan,
 	unsigned int completed_transfers)
 {
 	struct axi_dmac_desc *active;
 	struct axi_dmac_sg *sg;
+	bool start_next = false;
 
 	active = axi_dmac_active_desc(chan);
 	if (!active)
-		return;
+		return false;
 
-	if (active->cyclic) {
-		vchan_cyclic_callback(&active->vdesc);
-	} else {
-		do {
-			sg = &active->sg[active->num_completed];
-			if (!(BIT(sg->id) & completed_transfers))
-				break;
-			active->num_completed++;
-			if (active->num_completed == active->num_sgs) {
+	do {
+		sg = &active->sg[active->num_completed];
+		if (sg->id == AXI_DMAC_SG_UNUSED) /* Not yet submitted */
+			break;
+		if (!(BIT(sg->id) & completed_transfers))
+			break;
+		active->num_completed++;
+		sg->id = AXI_DMAC_SG_UNUSED;
+		if (sg->schedule_when_free) {
+			sg->schedule_when_free = false;
+			start_next = true;
+		}
+
+		if (active->cyclic)
+			vchan_cyclic_callback(&active->vdesc);
+
+		if (active->num_completed == active->num_sgs) {
+			if (active->cyclic) {
+				active->num_completed = 0; /* wrap around */
+			} else {
 				list_del(&active->vdesc.node);
 				vchan_cookie_complete(&active->vdesc);
 				active = axi_dmac_active_desc(chan);
 			}
-		} while (active);
-	}
+		}
+	} while (active);
+
+	return start_next;
 }
 
 static irqreturn_t axi_dmac_interrupt_handler(int irq, void *devid)
 {
 	struct axi_dmac *dmac = devid;
 	unsigned int pending;
+	bool start_next = false;
 
 	pending = axi_dmac_read(dmac, AXI_DMAC_REG_IRQ_PENDING);
 	if (!pending)
@@ -281,10 +312,10 @@ static irqreturn_t axi_dmac_interrupt_handler(int irq, void *devid)
 		unsigned int completed;
 
 		completed = axi_dmac_read(dmac, AXI_DMAC_REG_TRANSFER_DONE);
-		axi_dmac_transfer_done(&dmac->chan, completed);
+		start_next = axi_dmac_transfer_done(&dmac->chan, completed);
 	}
 	/* Space has become available in the descriptor queue */
-	if (pending & AXI_DMAC_IRQ_SOT)
+	if ((pending & AXI_DMAC_IRQ_SOT) || start_next)
 		axi_dmac_start_transfer(&dmac->chan);
 	spin_unlock(&dmac->chan.vchan.lock);
 
@@ -334,12 +365,16 @@ static void axi_dmac_issue_pending(struct dma_chan *c)
 static struct axi_dmac_desc *axi_dmac_alloc_desc(unsigned int num_sgs)
 {
 	struct axi_dmac_desc *desc;
+	unsigned int i;
 
 	desc = kzalloc(sizeof(struct axi_dmac_desc) +
 		sizeof(struct axi_dmac_sg) * num_sgs, GFP_NOWAIT);
 	if (!desc)
 		return NULL;
 
+	for (i = 0; i < num_sgs; i++)
+		desc->sg[i].id = AXI_DMAC_SG_UNUSED;
+
 	desc->num_sgs = num_sgs;
 
 	return desc;
diff --git a/drivers/dma/pch_dma.c b/drivers/dma/pch_dma.c
index f9028e9..afd8f27 100644
--- a/drivers/dma/pch_dma.c
+++ b/drivers/dma/pch_dma.c
@@ -123,7 +123,7 @@ struct pch_dma_chan {
 struct pch_dma {
 	struct dma_device	dma;
 	void __iomem *membase;
-	struct pci_pool		*pool;
+	struct dma_pool		*pool;
 	struct pch_dma_regs	regs;
 	struct pch_dma_desc_regs ch_regs[MAX_CHAN_NR];
 	struct pch_dma_chan	channels[MAX_CHAN_NR];
@@ -437,7 +437,7 @@ static struct pch_dma_desc *pdc_alloc_desc(struct dma_chan *chan, gfp_t flags)
 	struct pch_dma *pd = to_pd(chan->device);
 	dma_addr_t addr;
 
-	desc = pci_pool_zalloc(pd->pool, flags, &addr);
+	desc = dma_pool_zalloc(pd->pool, flags, &addr);
 	if (desc) {
 		INIT_LIST_HEAD(&desc->tx_list);
 		dma_async_tx_descriptor_init(&desc->txd, chan);
@@ -549,7 +549,7 @@ static void pd_free_chan_resources(struct dma_chan *chan)
 	spin_unlock_irq(&pd_chan->lock);
 
 	list_for_each_entry_safe(desc, _d, &tmp_list, desc_node)
-		pci_pool_free(pd->pool, desc, desc->txd.phys);
+		dma_pool_free(pd->pool, desc, desc->txd.phys);
 
 	pdc_enable_irq(chan, 0);
 }
@@ -880,7 +880,7 @@ static int pch_dma_probe(struct pci_dev *pdev,
 		goto err_iounmap;
 	}
 
-	pd->pool = pci_pool_create("pch_dma_desc_pool", pdev,
+	pd->pool = dma_pool_create("pch_dma_desc_pool", &pdev->dev,
 				   sizeof(struct pch_dma_desc), 4, 0);
 	if (!pd->pool) {
 		dev_err(&pdev->dev, "Failed to alloc DMA descriptors\n");
@@ -931,7 +931,7 @@ static int pch_dma_probe(struct pci_dev *pdev,
 	return 0;
 
 err_free_pool:
-	pci_pool_destroy(pd->pool);
+	dma_pool_destroy(pd->pool);
 err_free_irq:
 	free_irq(pdev->irq, pd);
 err_iounmap:
@@ -963,7 +963,7 @@ static void pch_dma_remove(struct pci_dev *pdev)
 			tasklet_kill(&pd_chan->tasklet);
 		}
 
-		pci_pool_destroy(pd->pool);
+		dma_pool_destroy(pd->pool);
 		pci_iounmap(pdev, pd->membase);
 		pci_release_regions(pdev);
 		pci_disable_device(pdev);