Merge commit 'dmaengine-3.13-v2' of git://git.kernel.org/pub/scm/linux/kernel/git/djbw/dmaengine

Pull dmaengine changes from Dan

1/ Bartlomiej and Dan finalized a rework of the dma address unmap
   implementation.

2/ In the course of testing 1/ a collection of enhancements to dmatest
   fell out.  Notably basic performance statistics, and fixed / enhanced
   test control through new module parameters 'run', 'wait', 'noverify',
   and 'verbose'.  Thanks to Andriy and Linus for their review.

3/ Testing the raid related corner cases of 1/ triggered bugs in the
   recently added 16-source operation support in the ioatdma driver.

4/ Some minor fixes / cleanups to mv_xor and ioatdma.

Conflicts:
	drivers/dma/dmatest.c

Signed-off-by: Vinod Koul <vinod.koul@intel.com>

6 files changed, 65 insertions, 336 deletions

diff --git a/drivers/dma/ioat/dma.c b/drivers/dma/ioat/dma.c
index a0f0fce..1a49c7776 100644
--- a/drivers/dma/ioat/dma.c
+++ b/drivers/dma/ioat/dma.c
@@ -531,21 +531,6 @@ static void ioat1_cleanup_event(unsigned long data)
 	writew(IOAT_CHANCTRL_RUN, ioat->base.reg_base + IOAT_CHANCTRL_OFFSET);
 }
 
-void ioat_dma_unmap(struct ioat_chan_common *chan, enum dma_ctrl_flags flags,
-		    size_t len, struct ioat_dma_descriptor *hw)
-{
-	struct pci_dev *pdev = chan->device->pdev;
-	size_t offset = len - hw->size;
-
-	if (!(flags & DMA_COMPL_SKIP_DEST_UNMAP))
-		ioat_unmap(pdev, hw->dst_addr - offset, len,
-			   PCI_DMA_FROMDEVICE, flags, 1);
-
-	if (!(flags & DMA_COMPL_SKIP_SRC_UNMAP))
-		ioat_unmap(pdev, hw->src_addr - offset, len,
-			   PCI_DMA_TODEVICE, flags, 0);
-}
-
 dma_addr_t ioat_get_current_completion(struct ioat_chan_common *chan)
 {
 	dma_addr_t phys_complete;
@@ -602,7 +587,7 @@ static void __cleanup(struct ioat_dma_chan *ioat, dma_addr_t phys_complete)
 		dump_desc_dbg(ioat, desc);
 		if (tx->cookie) {
 			dma_cookie_complete(tx);
-			ioat_dma_unmap(chan, tx->flags, desc->len, desc->hw);
+			dma_descriptor_unmap(tx);
 			ioat->active -= desc->hw->tx_cnt;
 			if (tx->callback) {
 				tx->callback(tx->callback_param);
@@ -833,8 +818,7 @@ int ioat_dma_self_test(struct ioatdma_device *device)
 
 	dma_src = dma_map_single(dev, src, IOAT_TEST_SIZE, DMA_TO_DEVICE);
 	dma_dest = dma_map_single(dev, dest, IOAT_TEST_SIZE, DMA_FROM_DEVICE);
-	flags = DMA_COMPL_SKIP_SRC_UNMAP | DMA_COMPL_SKIP_DEST_UNMAP |
-		DMA_PREP_INTERRUPT;
+	flags = DMA_PREP_INTERRUPT;
 	tx = device->common.device_prep_dma_memcpy(dma_chan, dma_dest, dma_src,
 						   IOAT_TEST_SIZE, flags);
 	if (!tx) {
@@ -885,8 +869,7 @@ static char ioat_interrupt_style[32] = "msix";
 module_param_string(ioat_interrupt_style, ioat_interrupt_style,
 		    sizeof(ioat_interrupt_style), 0644);
 MODULE_PARM_DESC(ioat_interrupt_style,
-		 "set ioat interrupt style: msix (default), "
-		 "msix-single-vector, msi, intx)");
+		 "set ioat interrupt style: msix (default), msi, intx");
 
 /**
  * ioat_dma_setup_interrupts - setup interrupt handler
@@ -904,8 +887,6 @@ int ioat_dma_setup_interrupts(struct ioatdma_device *device)
 
 	if (!strcmp(ioat_interrupt_style, "msix"))
 		goto msix;
-	if (!strcmp(ioat_interrupt_style, "msix-single-vector"))
-		goto msix_single_vector;
 	if (!strcmp(ioat_interrupt_style, "msi"))
 		goto msi;
 	if (!strcmp(ioat_interrupt_style, "intx"))
@@ -920,10 +901,8 @@ msix:
 		device->msix_entries[i].entry = i;
 
 	err = pci_enable_msix(pdev, device->msix_entries, msixcnt);
-	if (err < 0)
+	if (err)
 		goto msi;
-	if (err > 0)
-		goto msix_single_vector;
 
 	for (i = 0; i < msixcnt; i++) {
 		msix = &device->msix_entries[i];
@@ -937,29 +916,13 @@ msix:
 				chan = ioat_chan_by_index(device, j);
 				devm_free_irq(dev, msix->vector, chan);
 			}
-			goto msix_single_vector;
+			goto msi;
 		}
 	}
 	intrctrl |= IOAT_INTRCTRL_MSIX_VECTOR_CONTROL;
 	device->irq_mode = IOAT_MSIX;
 	goto done;
 
-msix_single_vector:
-	msix = &device->msix_entries[0];
-	msix->entry = 0;
-	err = pci_enable_msix(pdev, device->msix_entries, 1);
-	if (err)
-		goto msi;
-
-	err = devm_request_irq(dev, msix->vector, ioat_dma_do_interrupt, 0,
-			       "ioat-msix", device);
-	if (err) {
-		pci_disable_msix(pdev);
-		goto msi;
-	}
-	device->irq_mode = IOAT_MSIX_SINGLE;
-	goto done;
-
 msi:
 	err = pci_enable_msi(pdev);
 	if (err)
@@ -971,7 +934,7 @@ msi:
 		pci_disable_msi(pdev);
 		goto intx;
 	}
-	device->irq_mode = IOAT_MSIX;
+	device->irq_mode = IOAT_MSI;
 	goto done;
 
 intx:
diff --git a/drivers/dma/ioat/dma.h b/drivers/dma/ioat/dma.h
index 54fb7b9..11fb877 100644
--- a/drivers/dma/ioat/dma.h
+++ b/drivers/dma/ioat/dma.h
@@ -52,7 +52,6 @@
 enum ioat_irq_mode {
 	IOAT_NOIRQ = 0,
 	IOAT_MSIX,
-	IOAT_MSIX_SINGLE,
 	IOAT_MSI,
 	IOAT_INTX
 };
@@ -83,7 +82,6 @@ struct ioatdma_device {
 	struct pci_pool *completion_pool;
 #define MAX_SED_POOLS	5
 	struct dma_pool *sed_hw_pool[MAX_SED_POOLS];
-	struct kmem_cache *sed_pool;
 	struct dma_device common;
 	u8 version;
 	struct msix_entry msix_entries[4];
@@ -342,16 +340,6 @@ static inline bool is_ioat_bug(unsigned long err)
 	return !!err;
 }
 
-static inline void ioat_unmap(struct pci_dev *pdev, dma_addr_t addr, size_t len,
-			      int direction, enum dma_ctrl_flags flags, bool dst)
-{
-	if ((dst && (flags & DMA_COMPL_DEST_UNMAP_SINGLE)) ||
-	    (!dst && (flags & DMA_COMPL_SRC_UNMAP_SINGLE)))
-		pci_unmap_single(pdev, addr, len, direction);
-	else
-		pci_unmap_page(pdev, addr, len, direction);
-}
-
 int ioat_probe(struct ioatdma_device *device);
 int ioat_register(struct ioatdma_device *device);
 int ioat1_dma_probe(struct ioatdma_device *dev, int dca);
@@ -363,8 +351,6 @@ void ioat_init_channel(struct ioatdma_device *device,
 		       struct ioat_chan_common *chan, int idx);
 enum dma_status ioat_dma_tx_status(struct dma_chan *c, dma_cookie_t cookie,
 				   struct dma_tx_state *txstate);
-void ioat_dma_unmap(struct ioat_chan_common *chan, enum dma_ctrl_flags flags,
-		    size_t len, struct ioat_dma_descriptor *hw);
 bool ioat_cleanup_preamble(struct ioat_chan_common *chan,
 			   dma_addr_t *phys_complete);
 void ioat_kobject_add(struct ioatdma_device *device, struct kobj_type *type);
diff --git a/drivers/dma/ioat/dma_v2.c b/drivers/dma/ioat/dma_v2.c
index b925e1b..5d3affe 100644
--- a/drivers/dma/ioat/dma_v2.c
+++ b/drivers/dma/ioat/dma_v2.c
@@ -148,7 +148,7 @@ static void __cleanup(struct ioat2_dma_chan *ioat, dma_addr_t phys_complete)
 		tx = &desc->txd;
 		dump_desc_dbg(ioat, desc);
 		if (tx->cookie) {
-			ioat_dma_unmap(chan, tx->flags, desc->len, desc->hw);
+			dma_descriptor_unmap(tx);
 			dma_cookie_complete(tx);
 			if (tx->callback) {
 				tx->callback(tx->callback_param);
diff --git a/drivers/dma/ioat/dma_v2.h b/drivers/dma/ioat/dma_v2.h
index 212d584..4702927 100644
--- a/drivers/dma/ioat/dma_v2.h
+++ b/drivers/dma/ioat/dma_v2.h
@@ -157,7 +157,6 @@ static inline void ioat2_set_chainaddr(struct ioat2_dma_chan *ioat, u64 addr)
 
 int ioat2_dma_probe(struct ioatdma_device *dev, int dca);
 int ioat3_dma_probe(struct ioatdma_device *dev, int dca);
-void ioat3_dma_remove(struct ioatdma_device *dev);
 struct dca_provider *ioat2_dca_init(struct pci_dev *pdev, void __iomem *iobase);
 struct dca_provider *ioat3_dca_init(struct pci_dev *pdev, void __iomem *iobase);
 int ioat2_check_space_lock(struct ioat2_dma_chan *ioat, int num_descs);
diff --git a/drivers/dma/ioat/dma_v3.c b/drivers/dma/ioat/dma_v3.c
index 806b4ce..820817e9 100644
--- a/drivers/dma/ioat/dma_v3.c
+++ b/drivers/dma/ioat/dma_v3.c
@@ -67,6 +67,8 @@
 #include "dma.h"
 #include "dma_v2.h"
 
+extern struct kmem_cache *ioat3_sed_cache;
+
 /* ioat hardware assumes at least two sources for raid operations */
 #define src_cnt_to_sw(x) ((x) + 2)
 #define src_cnt_to_hw(x) ((x) - 2)
@@ -87,22 +89,8 @@ static const u8 pq_idx_to_field[] = { 1, 4, 5, 0, 1, 2, 4, 5 };
 static const u8 pq16_idx_to_field[] = { 1, 4, 1, 2, 3, 4, 5, 6, 7,
 					0, 1, 2, 3, 4, 5, 6 };
 
-/*
- * technically sources 1 and 2 do not require SED, but the op will have
- * at least 9 descriptors so that's irrelevant.
- */
-static const u8 pq16_idx_to_sed[] = { 0, 0, 0, 0, 0, 0, 0, 0, 0,
-				      1, 1, 1, 1, 1, 1, 1 };
-
 static void ioat3_eh(struct ioat2_dma_chan *ioat);
 
-static dma_addr_t xor_get_src(struct ioat_raw_descriptor *descs[2], int idx)
-{
-	struct ioat_raw_descriptor *raw = descs[xor_idx_to_desc >> idx & 1];
-
-	return raw->field[xor_idx_to_field[idx]];
-}
-
 static void xor_set_src(struct ioat_raw_descriptor *descs[2],
 			dma_addr_t addr, u32 offset, int idx)
 {
@@ -135,12 +123,6 @@ static void pq_set_src(struct ioat_raw_descriptor *descs[2],
 	pq->coef[idx] = coef;
 }
 
-static int sed_get_pq16_pool_idx(int src_cnt)
-{
-
-	return pq16_idx_to_sed[src_cnt];
-}
-
 static bool is_jf_ioat(struct pci_dev *pdev)
 {
 	switch (pdev->device) {
@@ -272,7 +254,7 @@ ioat3_alloc_sed(struct ioatdma_device *device, unsigned int hw_pool)
 	struct ioat_sed_ent *sed;
 	gfp_t flags = __GFP_ZERO | GFP_ATOMIC;
 
-	sed = kmem_cache_alloc(device->sed_pool, flags);
+	sed = kmem_cache_alloc(ioat3_sed_cache, flags);
 	if (!sed)
 		return NULL;
 
@@ -280,7 +262,7 @@ ioat3_alloc_sed(struct ioatdma_device *device, unsigned int hw_pool)
 	sed->hw = dma_pool_alloc(device->sed_hw_pool[hw_pool],
 				 flags, &sed->dma);
 	if (!sed->hw) {
-		kmem_cache_free(device->sed_pool, sed);
+		kmem_cache_free(ioat3_sed_cache, sed);
 		return NULL;
 	}
 
@@ -293,165 +275,7 @@ static void ioat3_free_sed(struct ioatdma_device *device, struct ioat_sed_ent *s
 		return;
 
 	dma_pool_free(device->sed_hw_pool[sed->hw_pool], sed->hw, sed->dma);
-	kmem_cache_free(device->sed_pool, sed);
-}
-
-static void ioat3_dma_unmap(struct ioat2_dma_chan *ioat,
-			    struct ioat_ring_ent *desc, int idx)
-{
-	struct ioat_chan_common *chan = &ioat->base;
-	struct pci_dev *pdev = chan->device->pdev;
-	size_t len = desc->len;
-	size_t offset = len - desc->hw->size;
-	struct dma_async_tx_descriptor *tx = &desc->txd;
-	enum dma_ctrl_flags flags = tx->flags;
-
-	switch (desc->hw->ctl_f.op) {
-	case IOAT_OP_COPY:
-		if (!desc->hw->ctl_f.null) /* skip 'interrupt' ops */
-			ioat_dma_unmap(chan, flags, len, desc->hw);
-		break;
-	case IOAT_OP_XOR_VAL:
-	case IOAT_OP_XOR: {
-		struct ioat_xor_descriptor *xor = desc->xor;
-		struct ioat_ring_ent *ext;
-		struct ioat_xor_ext_descriptor *xor_ex = NULL;
-		int src_cnt = src_cnt_to_sw(xor->ctl_f.src_cnt);
-		struct ioat_raw_descriptor *descs[2];
-		int i;
-
-		if (src_cnt > 5) {
-			ext = ioat2_get_ring_ent(ioat, idx + 1);
-			xor_ex = ext->xor_ex;
-		}
-
-		if (!(flags & DMA_COMPL_SKIP_SRC_UNMAP)) {
-			descs[0] = (struct ioat_raw_descriptor *) xor;
-			descs[1] = (struct ioat_raw_descriptor *) xor_ex;
-			for (i = 0; i < src_cnt; i++) {
-				dma_addr_t src = xor_get_src(descs, i);
-
-				ioat_unmap(pdev, src - offset, len,
-					   PCI_DMA_TODEVICE, flags, 0);
-			}
-
-			/* dest is a source in xor validate operations */
-			if (xor->ctl_f.op == IOAT_OP_XOR_VAL) {
-				ioat_unmap(pdev, xor->dst_addr - offset, len,
-					   PCI_DMA_TODEVICE, flags, 1);
-				break;
-			}
-		}
-
-		if (!(flags & DMA_COMPL_SKIP_DEST_UNMAP))
-			ioat_unmap(pdev, xor->dst_addr - offset, len,
-				   PCI_DMA_FROMDEVICE, flags, 1);
-		break;
-	}
-	case IOAT_OP_PQ_VAL:
-	case IOAT_OP_PQ: {
-		struct ioat_pq_descriptor *pq = desc->pq;
-		struct ioat_ring_ent *ext;
-		struct ioat_pq_ext_descriptor *pq_ex = NULL;
-		int src_cnt = src_cnt_to_sw(pq->ctl_f.src_cnt);
-		struct ioat_raw_descriptor *descs[2];
-		int i;
-
-		if (src_cnt > 3) {
-			ext = ioat2_get_ring_ent(ioat, idx + 1);
-			pq_ex = ext->pq_ex;
-		}
-
-		/* in the 'continue' case don't unmap the dests as sources */
-		if (dmaf_p_disabled_continue(flags))
-			src_cnt--;
-		else if (dmaf_continue(flags))
-			src_cnt -= 3;
-
-		if (!(flags & DMA_COMPL_SKIP_SRC_UNMAP)) {
-			descs[0] = (struct ioat_raw_descriptor *) pq;
-			descs[1] = (struct ioat_raw_descriptor *) pq_ex;
-			for (i = 0; i < src_cnt; i++) {
-				dma_addr_t src = pq_get_src(descs, i);
-
-				ioat_unmap(pdev, src - offset, len,
-					   PCI_DMA_TODEVICE, flags, 0);
-			}
-
-			/* the dests are sources in pq validate operations */
-			if (pq->ctl_f.op == IOAT_OP_XOR_VAL) {
-				if (!(flags & DMA_PREP_PQ_DISABLE_P))
-					ioat_unmap(pdev, pq->p_addr - offset,
-						   len, PCI_DMA_TODEVICE, flags, 0);
-				if (!(flags & DMA_PREP_PQ_DISABLE_Q))
-					ioat_unmap(pdev, pq->q_addr - offset,
-						   len, PCI_DMA_TODEVICE, flags, 0);
-				break;
-			}
-		}
-
-		if (!(flags & DMA_COMPL_SKIP_DEST_UNMAP)) {
-			if (!(flags & DMA_PREP_PQ_DISABLE_P))
-				ioat_unmap(pdev, pq->p_addr - offset, len,
-					   PCI_DMA_BIDIRECTIONAL, flags, 1);
-			if (!(flags & DMA_PREP_PQ_DISABLE_Q))
-				ioat_unmap(pdev, pq->q_addr - offset, len,
-					   PCI_DMA_BIDIRECTIONAL, flags, 1);
-		}
-		break;
-	}
-	case IOAT_OP_PQ_16S:
-	case IOAT_OP_PQ_VAL_16S: {
-		struct ioat_pq_descriptor *pq = desc->pq;
-		int src_cnt = src16_cnt_to_sw(pq->ctl_f.src_cnt);
-		struct ioat_raw_descriptor *descs[4];
-		int i;
-
-		/* in the 'continue' case don't unmap the dests as sources */
-		if (dmaf_p_disabled_continue(flags))
-			src_cnt--;
-		else if (dmaf_continue(flags))
-			src_cnt -= 3;
-
-		if (!(flags & DMA_COMPL_SKIP_SRC_UNMAP)) {
-			descs[0] = (struct ioat_raw_descriptor *)pq;
-			descs[1] = (struct ioat_raw_descriptor *)(desc->sed->hw);
-			descs[2] = (struct ioat_raw_descriptor *)(&desc->sed->hw->b[0]);
-			for (i = 0; i < src_cnt; i++) {
-				dma_addr_t src = pq16_get_src(descs, i);
-
-				ioat_unmap(pdev, src - offset, len,
-					   PCI_DMA_TODEVICE, flags, 0);
-			}
-
-			/* the dests are sources in pq validate operations */
-			if (pq->ctl_f.op == IOAT_OP_XOR_VAL) {
-				if (!(flags & DMA_PREP_PQ_DISABLE_P))
-					ioat_unmap(pdev, pq->p_addr - offset,
-						   len, PCI_DMA_TODEVICE,
-						   flags, 0);
-				if (!(flags & DMA_PREP_PQ_DISABLE_Q))
-					ioat_unmap(pdev, pq->q_addr - offset,
-						   len, PCI_DMA_TODEVICE,
-						   flags, 0);
-				break;
-			}
-		}
-
-		if (!(flags & DMA_COMPL_SKIP_DEST_UNMAP)) {
-			if (!(flags & DMA_PREP_PQ_DISABLE_P))
-				ioat_unmap(pdev, pq->p_addr - offset, len,
-					   PCI_DMA_BIDIRECTIONAL, flags, 1);
-			if (!(flags & DMA_PREP_PQ_DISABLE_Q))
-				ioat_unmap(pdev, pq->q_addr - offset, len,
-					   PCI_DMA_BIDIRECTIONAL, flags, 1);
-		}
-		break;
-	}
-	default:
-		dev_err(&pdev->dev, "%s: unknown op type: %#x\n",
-			__func__, desc->hw->ctl_f.op);
-	}
+	kmem_cache_free(ioat3_sed_cache, sed);
 }
 
 static bool desc_has_ext(struct ioat_ring_ent *desc)
@@ -577,7 +401,7 @@ static void __cleanup(struct ioat2_dma_chan *ioat, dma_addr_t phys_complete)
 		tx = &desc->txd;
 		if (tx->cookie) {
 			dma_cookie_complete(tx);
-			ioat3_dma_unmap(ioat, desc, idx + i);
+			dma_descriptor_unmap(tx);
 			if (tx->callback) {
 				tx->callback(tx->callback_param);
 				tx->callback = NULL;
@@ -1129,9 +953,6 @@ __ioat3_prep_pq16_lock(struct dma_chan *c, enum sum_check_flags *result,
 	u8 op;
 	int i, s, idx, num_descs;
 
-	/* this function only handles src_cnt 9 - 16 */
-	BUG_ON(src_cnt < 9);
-
 	/* this function is only called with 9-16 sources */
 	op = result ? IOAT_OP_PQ_VAL_16S : IOAT_OP_PQ_16S;
 
@@ -1159,8 +980,7 @@ __ioat3_prep_pq16_lock(struct dma_chan *c, enum sum_check_flags *result,
 
 		descs[0] = (struct ioat_raw_descriptor *) pq;
 
-		desc->sed = ioat3_alloc_sed(device,
-					    sed_get_pq16_pool_idx(src_cnt));
+		desc->sed = ioat3_alloc_sed(device, (src_cnt-2) >> 3);
 		if (!desc->sed) {
 			dev_err(to_dev(chan),
 				"%s: no free sed entries\n", __func__);
@@ -1218,13 +1038,21 @@ __ioat3_prep_pq16_lock(struct dma_chan *c, enum sum_check_flags *result,
 	return &desc->txd;
 }
 
+static int src_cnt_flags(unsigned int src_cnt, unsigned long flags)
+{
+	if (dmaf_p_disabled_continue(flags))
+		return src_cnt + 1;
+	else if (dmaf_continue(flags))
+		return src_cnt + 3;
+	else
+		return src_cnt;
+}
+
 static struct dma_async_tx_descriptor *
 ioat3_prep_pq(struct dma_chan *chan, dma_addr_t *dst, dma_addr_t *src,
 	      unsigned int src_cnt, const unsigned char *scf, size_t len,
 	      unsigned long flags)
 {
-	struct dma_device *dma = chan->device;
-
 	/* specify valid address for disabled result */
 	if (flags & DMA_PREP_PQ_DISABLE_P)
 		dst[0] = dst[1];
@@ -1244,7 +1072,7 @@ ioat3_prep_pq(struct dma_chan *chan, dma_addr_t *dst, dma_addr_t *src,
 		single_source_coef[0] = scf[0];
 		single_source_coef[1] = 0;
 
-		return (src_cnt > 8) && (dma->max_pq > 8) ?
+		return src_cnt_flags(src_cnt, flags) > 8 ?
 			__ioat3_prep_pq16_lock(chan, NULL, dst, single_source,
 					       2, single_source_coef, len,
 					       flags) :
@@ -1252,7 +1080,7 @@ ioat3_prep_pq(struct dma_chan *chan, dma_addr_t *dst, dma_addr_t *src,
 					     single_source_coef, len, flags);
 
 	} else {
-		return (src_cnt > 8) && (dma->max_pq > 8) ?
+		return src_cnt_flags(src_cnt, flags) > 8 ?
 			__ioat3_prep_pq16_lock(chan, NULL, dst, src, src_cnt,
 					       scf, len, flags) :
 			__ioat3_prep_pq_lock(chan, NULL, dst, src, src_cnt,
@@ -1265,8 +1093,6 @@ ioat3_prep_pq_val(struct dma_chan *chan, dma_addr_t *pq, dma_addr_t *src,
 		  unsigned int src_cnt, const unsigned char *scf, size_t len,
 		  enum sum_check_flags *pqres, unsigned long flags)
 {
-	struct dma_device *dma = chan->device;
-
 	/* specify valid address for disabled result */
 	if (flags & DMA_PREP_PQ_DISABLE_P)
 		pq[0] = pq[1];
@@ -1278,7 +1104,7 @@ ioat3_prep_pq_val(struct dma_chan *chan, dma_addr_t *pq, dma_addr_t *src,
 	 */
 	*pqres = 0;
 
-	return (src_cnt > 8) && (dma->max_pq > 8) ?
+	return src_cnt_flags(src_cnt, flags) > 8 ?
 		__ioat3_prep_pq16_lock(chan, pqres, pq, src, src_cnt, scf, len,
 				       flags) :
 		__ioat3_prep_pq_lock(chan, pqres, pq, src, src_cnt, scf, len,
@@ -1289,7 +1115,6 @@ static struct dma_async_tx_descriptor *
 ioat3_prep_pqxor(struct dma_chan *chan, dma_addr_t dst, dma_addr_t *src,
 		 unsigned int src_cnt, size_t len, unsigned long flags)
 {
-	struct dma_device *dma = chan->device;
 	unsigned char scf[src_cnt];
 	dma_addr_t pq[2];
 
@@ -1298,7 +1123,7 @@ ioat3_prep_pqxor(struct dma_chan *chan, dma_addr_t dst, dma_addr_t *src,
 	flags |= DMA_PREP_PQ_DISABLE_Q;
 	pq[1] = dst; /* specify valid address for disabled result */
 
-	return (src_cnt > 8) && (dma->max_pq > 8) ?
+	return src_cnt_flags(src_cnt, flags) > 8 ?
 		__ioat3_prep_pq16_lock(chan, NULL, pq, src, src_cnt, scf, len,
 				       flags) :
 		__ioat3_prep_pq_lock(chan, NULL, pq, src, src_cnt, scf, len,
@@ -1310,7 +1135,6 @@ ioat3_prep_pqxor_val(struct dma_chan *chan, dma_addr_t *src,
 		     unsigned int src_cnt, size_t len,
 		     enum sum_check_flags *result, unsigned long flags)
 {
-	struct dma_device *dma = chan->device;
 	unsigned char scf[src_cnt];
 	dma_addr_t pq[2];
 
@@ -1324,8 +1148,7 @@ ioat3_prep_pqxor_val(struct dma_chan *chan, dma_addr_t *src,
 	flags |= DMA_PREP_PQ_DISABLE_Q;
 	pq[1] = pq[0]; /* specify valid address for disabled result */
 
-
-	return (src_cnt > 8) && (dma->max_pq > 8) ?
+	return src_cnt_flags(src_cnt, flags) > 8 ?
 		__ioat3_prep_pq16_lock(chan, result, pq, &src[1], src_cnt - 1,
 				       scf, len, flags) :
 		__ioat3_prep_pq_lock(chan, result, pq, &src[1], src_cnt - 1,
@@ -1444,9 +1267,7 @@ static int ioat_xor_val_self_test(struct ioatdma_device *device)
 					   DMA_TO_DEVICE);
 	tx = dma->device_prep_dma_xor(dma_chan, dest_dma, dma_srcs,
 				      IOAT_NUM_SRC_TEST, PAGE_SIZE,
-				      DMA_PREP_INTERRUPT |
-				      DMA_COMPL_SKIP_SRC_UNMAP |
-				      DMA_COMPL_SKIP_DEST_UNMAP);
+				      DMA_PREP_INTERRUPT);
 
 	if (!tx) {
 		dev_err(dev, "Self-test xor prep failed\n");
@@ -1507,9 +1328,7 @@ static int ioat_xor_val_self_test(struct ioatdma_device *device)
 					   DMA_TO_DEVICE);
 	tx = dma->device_prep_dma_xor_val(dma_chan, dma_srcs,
 					  IOAT_NUM_SRC_TEST + 1, PAGE_SIZE,
-					  &xor_val_result, DMA_PREP_INTERRUPT |
-					  DMA_COMPL_SKIP_SRC_UNMAP |
-					  DMA_COMPL_SKIP_DEST_UNMAP);
+					  &xor_val_result, DMA_PREP_INTERRUPT);
 	if (!tx) {
 		dev_err(dev, "Self-test zero prep failed\n");
 		err = -ENODEV;
@@ -1545,6 +1364,8 @@ static int ioat_xor_val_self_test(struct ioatdma_device *device)
 		goto free_resources;
 	}
 
+	memset(page_address(dest), 0, PAGE_SIZE);
+
 	/* test for non-zero parity sum */
 	op = IOAT_OP_XOR_VAL;
 
@@ -1554,9 +1375,7 @@ static int ioat_xor_val_self_test(struct ioatdma_device *device)
 					   DMA_TO_DEVICE);
 	tx = dma->device_prep_dma_xor_val(dma_chan, dma_srcs,
 					  IOAT_NUM_SRC_TEST + 1, PAGE_SIZE,
-					  &xor_val_result, DMA_PREP_INTERRUPT |
-					  DMA_COMPL_SKIP_SRC_UNMAP |
-					  DMA_COMPL_SKIP_DEST_UNMAP);
+					  &xor_val_result, DMA_PREP_INTERRUPT);
 	if (!tx) {
 		dev_err(dev, "Self-test 2nd zero prep failed\n");
 		err = -ENODEV;
@@ -1630,52 +1449,36 @@ static int ioat3_dma_self_test(struct ioatdma_device *device)
 
 static int ioat3_irq_reinit(struct ioatdma_device *device)
 {
-	int msixcnt = device->common.chancnt;
 	struct pci_dev *pdev = device->pdev;
-	int i;
-	struct msix_entry *msix;
-	struct ioat_chan_common *chan;
-	int err = 0;
+	int irq = pdev->irq, i;
+
+	if (!is_bwd_ioat(pdev))
+		return 0;
 
 	switch (device->irq_mode) {
 	case IOAT_MSIX:
+		for (i = 0; i < device->common.chancnt; i++) {
+			struct msix_entry *msix = &device->msix_entries[i];
+			struct ioat_chan_common *chan;
 
-		for (i = 0; i < msixcnt; i++) {
-			msix = &device->msix_entries[i];
 			chan = ioat_chan_by_index(device, i);
 			devm_free_irq(&pdev->dev, msix->vector, chan);
 		}
 
 		pci_disable_msix(pdev);
 		break;
-
-	case IOAT_MSIX_SINGLE:
-		msix = &device->msix_entries[0];
-		chan = ioat_chan_by_index(device, 0);
-		devm_free_irq(&pdev->dev, msix->vector, chan);
-		pci_disable_msix(pdev);
-		break;
-
 	case IOAT_MSI:
-		chan = ioat_chan_by_index(device, 0);
-		devm_free_irq(&pdev->dev, pdev->irq, chan);
 		pci_disable_msi(pdev);
-		break;
-
+		/* fall through */
 	case IOAT_INTX:
-		chan = ioat_chan_by_index(device, 0);
-		devm_free_irq(&pdev->dev, pdev->irq, chan);
+		devm_free_irq(&pdev->dev, irq, device);
 		break;
-
 	default:
 		return 0;
 	}
-
 	device->irq_mode = IOAT_NOIRQ;
 
-	err = ioat_dma_setup_interrupts(device);
-
-	return err;
+	return ioat_dma_setup_interrupts(device);
 }
 
 static int ioat3_reset_hw(struct ioat_chan_common *chan)
@@ -1718,14 +1521,12 @@ static int ioat3_reset_hw(struct ioat_chan_common *chan)
 	}
 
 	err = ioat2_reset_sync(chan, msecs_to_jiffies(200));
-	if (err) {
-		dev_err(&pdev->dev, "Failed to reset!\n");
-		return err;
-	}
-
-	if (device->irq_mode != IOAT_NOIRQ && is_bwd_ioat(pdev))
+	if (!err)
 		err = ioat3_irq_reinit(device);
 
+	if (err)
+		dev_err(&pdev->dev, "Failed to reset: %d\n", err);
+
 	return err;
 }
 
@@ -1835,21 +1636,15 @@ int ioat3_dma_probe(struct ioatdma_device *device, int dca)
 		char pool_name[14];
 		int i;
 
-		/* allocate sw descriptor pool for SED */
-		device->sed_pool = kmem_cache_create("ioat_sed",
-				sizeof(struct ioat_sed_ent), 0, 0, NULL);
-		if (!device->sed_pool)
-			return -ENOMEM;
-
 		for (i = 0; i < MAX_SED_POOLS; i++) {
 			snprintf(pool_name, 14, "ioat_hw%d_sed", i);
 
 			/* allocate SED DMA pool */
-			device->sed_hw_pool[i] = dma_pool_create(pool_name,
+			device->sed_hw_pool[i] = dmam_pool_create(pool_name,
 					&pdev->dev,
 					SED_SIZE * (i + 1), 64, 0);
 			if (!device->sed_hw_pool[i])
-				goto sed_pool_cleanup;
+				return -ENOMEM;
 
 		}
 	}
@@ -1875,28 +1670,4 @@ int ioat3_dma_probe(struct ioatdma_device *device, int dca)
 		device->dca = ioat3_dca_init(pdev, device->reg_base);
 
 	return 0;
-
-sed_pool_cleanup:
-	if (device->sed_pool) {
-		int i;
-		kmem_cache_destroy(device->sed_pool);
-
-		for (i = 0; i < MAX_SED_POOLS; i++)
-			if (device->sed_hw_pool[i])
-				dma_pool_destroy(device->sed_hw_pool[i]);
-	}
-
-	return -ENOMEM;
-}
-
-void ioat3_dma_remove(struct ioatdma_device *device)
-{
-	if (device->sed_pool) {
-		int i;
-		kmem_cache_destroy(device->sed_pool);
-
-		for (i = 0; i < MAX_SED_POOLS; i++)
-			if (device->sed_hw_pool[i])
-				dma_pool_destroy(device->sed_hw_pool[i]);
-	}
 }
diff --git a/drivers/dma/ioat/pci.c b/drivers/dma/ioat/pci.c
index 2c8d560..1d051cd 100644
--- a/drivers/dma/ioat/pci.c
+++ b/drivers/dma/ioat/pci.c
@@ -123,6 +123,7 @@ module_param(ioat_dca_enabled, int, 0644);
 MODULE_PARM_DESC(ioat_dca_enabled, "control support of dca service (default: 1)");
 
 struct kmem_cache *ioat2_cache;
+struct kmem_cache *ioat3_sed_cache;
 
 #define DRV_NAME "ioatdma"
 
@@ -207,9 +208,6 @@ static void ioat_remove(struct pci_dev *pdev)
 	if (!device)
 		return;
 
-	if (device->version >= IOAT_VER_3_0)
-		ioat3_dma_remove(device);
-
 	dev_err(&pdev->dev, "Removing dma and dca services\n");
 	if (device->dca) {
 		unregister_dca_provider(device->dca, &pdev->dev);
@@ -221,7 +219,7 @@ static void ioat_remove(struct pci_dev *pdev)
 
 static int __init ioat_init_module(void)
 {
-	int err;
+	int err = -ENOMEM;
 
 	pr_info("%s: Intel(R) QuickData Technology Driver %s\n",
 		DRV_NAME, IOAT_DMA_VERSION);
@@ -231,9 +229,21 @@ static int __init ioat_init_module(void)
 	if (!ioat2_cache)
 		return -ENOMEM;
 
+	ioat3_sed_cache = KMEM_CACHE(ioat_sed_ent, 0);
+	if (!ioat3_sed_cache)
+		goto err_ioat2_cache;
+
 	err = pci_register_driver(&ioat_pci_driver);
 	if (err)
-		kmem_cache_destroy(ioat2_cache);
+		goto err_ioat3_cache;
+
+	return 0;
+
+ err_ioat3_cache:
+	kmem_cache_destroy(ioat3_sed_cache);
+
+ err_ioat2_cache:
+	kmem_cache_destroy(ioat2_cache);
 
 	return err;
 }