Skip to contentSkip to conversation

dmaengine: dw-edma: HDMA support

PATCH Superseded Gustavo Pimentel 2 participants 18 messages Feb 12, 2021 – Feb 18, 2021

Subsystems (3)

v6 of 7 Superseded Latest is v7 →

lore
Share & tools
lore
#0 lore raw

v6 cover dmaengine: dw-edma: HDMA support

This patch series adds the HDMA support, as long the IP design has set the compatible register map parameter, which allows compatibility at some degree for the existing Synopsys DesignWare eDMA driver that is already available on the Kernel.

The HDMA "Hyper-DMA" IP is an enhancement of the eDMA "embedded-DMA" IP.

This new improvement comes with a PCI DVSEC that allows to the driver recognize and switch behavior if it's an eDMA or an HDMA, becoming retrocompatible, in the absence of this DVSEC, the driver will assume that is an eDMA IP.

It also adds the interleaved support, since it will be similar to the current scatter-gather implementation.

As well fixes/improves some abnormal behaviors not detected before, such as:
- crash on loading/unloading driver
- memory space definition for the data area and for the linked list space
- scatter-gather address calculation on 32 bits platforms
- minor comment and variable reordering

Changes:
V2: Applied changes based on Bjorn Helgaas' review
Rebased patches on top of v5.11-rc1 version
V3: Applied changes based on Lukas Wunner' review
V4: Fix a typo detected by kernel test robot
V5: Rework driver accordingly to Bjorn Helgaas's feedback
V6: Rework driver accordingly to Krzysztof Wilczyński's feedback

Gustavo Pimentel (15):
dmaengine: dw-edma: Add writeq() and readq() for 64 bits architectures
dmaengine: dw-edma: Fix comments offset characters' alignment
dmaengine: dw-edma: Add support for the HDMA feature
PCI: Add pci_find_vsec_capability() to find a specific VSEC
dmaengine: dw-edma: Add PCIe VSEC data retrieval support
dmaengine: dw-edma: Add device_prep_interleave_dma() support
dmaengine: dw-edma: Improve number of channels check
dmaengine: dw-edma: Reorder variables to keep consistency
dmaengine: dw-edma: Improve the linked list and data blocks definition
dmaengine: dw-edma: Change linked list and data blocks offset and
sizes
dmaengine: dw-edma: Move struct dentry variable from static definition
into dw_edma struct
dmaengine: dw-edma: Fix crash on loading/unloading driver
dmaengine: dw-edma: Change DMA abreviation from lower into upper case
dmaengine: dw-edma: Revert fix scatter-gather address calculation
dmaengine: dw-edma: Add pcim_iomap_table return check

 drivers/dma/dw-edma/dw-edma-core.c       | 178 +++++++++++-------
 drivers/dma/dw-edma/dw-edma-core.h       |  37 ++--
 drivers/dma/dw-edma/dw-edma-pcie.c       | 275 +++++++++++++++++++++-------
 drivers/dma/dw-edma/dw-edma-v0-core.c    | 300 ++++++++++++++++++++++++-------
 drivers/dma/dw-edma/dw-edma-v0-core.h    |   2 +-
 drivers/dma/dw-edma/dw-edma-v0-debugfs.c |  77 ++++----
 drivers/dma/dw-edma/dw-edma-v0-debugfs.h |   4 +-
 drivers/dma/dw-edma/dw-edma-v0-regs.h    | 291 +++++++++++++++++++-----------
 drivers/pci/pci.c                        |  34 ++++
 include/linux/pci.h                      |   2 +
 include/uapi/linux/pci_regs.h            |   6 +
 11 files changed, 851 insertions(+), 355 deletions(-)
Signature · 1 line
2.7.4
#1 lore raw

PATCH v6 1/15 dmaengine: dw-edma: Add writeq() and readq() for 64 bits architectures

Add writeq() and readq() for 64 bits architures support.

Supporting these two functions will allow the write or the read of eDMA 64 bits registers at once instead of having two consecutive operations.

Also, this improvement will allow the PCI optimization transaction messages, which will generate a 64 bits message instead of two messages of 32 bits.

---
 drivers/dma/dw-edma/dw-edma-v0-core.c    | 254 +++++++++++++++++++++++--------
 drivers/dma/dw-edma/dw-edma-v0-debugfs.c |  48 +++---
 drivers/dma/dw-edma/dw-edma-v0-regs.h    | 149 +++++++++++++-----
 3 files changed, 326 insertions(+), 125 deletions(-)
drivers/dma/dw-edma/dw-edma-v0-core.c +190 -64
index 692de47..7888eda 100644
@@ -28,29 +28,69 @@ static inline struct dw_edma_v0_regs __iomem *__dw_regs(struct dw_edma *dw)
 	return dw->rg_region.vaddr;
 }
 
-#define SET(dw, name, value)				\
+#define SET_32(dw, name, value)				\
 	writel(value, &(__dw_regs(dw)->name))
 
-#define GET(dw, name)					\
+#define GET_32(dw, name)				\
 	readl(&(__dw_regs(dw)->name))
 
-#define SET_RW(dw, dir, name, value)			\
+#define SET_RW_32(dw, dir, name, value)			\
 	do {						\
 		if ((dir) == EDMA_DIR_WRITE)		\
-			SET(dw, wr_##name, value);	\
+			SET_32(dw, wr_##name, value);	\
 		else					\
-			SET(dw, rd_##name, value);	\
+			SET_32(dw, rd_##name, value);	\
 	} while (0)
 
-#define GET_RW(dw, dir, name)				\
+#define GET_RW_32(dw, dir, name)			\
 	((dir) == EDMA_DIR_WRITE			\
-	  ? GET(dw, wr_##name)				\
-	  : GET(dw, rd_##name))
+	  ? GET_32(dw, wr_##name)			\
+	  : GET_32(dw, rd_##name))
 
-#define SET_BOTH(dw, name, value)			\
+#define SET_BOTH_32(dw, name, value)			\
 	do {						\
-		SET(dw, wr_##name, value);		\
-		SET(dw, rd_##name, value);		\
+		SET_32(dw, wr_##name, value);		\
+		SET_32(dw, rd_##name, value);		\
+	} while (0)
+
+#ifdef CONFIG_64BIT
+
+#define SET_64(dw, name, value)				\
+	writeq(value, &(__dw_regs(dw)->name))
+
+#define GET_64(dw, name)				\
+	readq(&(__dw_regs(dw)->name))
+
+#define SET_RW_64(dw, dir, name, value)			\
+	do {						\
+		if ((dir) == EDMA_DIR_WRITE)		\
+			SET_64(dw, wr_##name, value);	\
+		else					\
+			SET_64(dw, rd_##name, value);	\
+	} while (0)
+
+#define GET_RW_64(dw, dir, name)			\
+	((dir) == EDMA_DIR_WRITE			\
+	  ? GET_64(dw, wr_##name)			\
+	  : GET_64(dw, rd_##name))
+
+#define SET_BOTH_64(dw, name, value)			\
+	do {						\
+		SET_64(dw, wr_##name, value);		\
+		SET_64(dw, rd_##name, value);		\
+	} while (0)
+
+#endif /* CONFIG_64BIT */
+
+#define SET_COMPAT(dw, name, value)			\
+	writel(value, &(__dw_regs(dw)->type.unroll.name))
+
+#define SET_RW_COMPAT(dw, dir, name, value)		\
+	do {						\
+		if ((dir) == EDMA_DIR_WRITE)		\
+			SET_COMPAT(dw, wr_##name, value); \
+		else					\
+			SET_COMPAT(dw, rd_##name, value); \
 	} while (0)
 
 static inline struct dw_edma_v0_ch_regs __iomem *
@@ -115,21 +155,86 @@ static inline u32 readl_ch(struct dw_edma *dw, enum dw_edma_dir dir, u16 ch,
 	return value;
 }
 
-#define SET_CH(dw, dir, ch, name, value) \
+#define SET_CH_32(dw, dir, ch, name, value) \
 	writel_ch(dw, dir, ch, value, &(__dw_ch_regs(dw, dir, ch)->name))
 
-#define GET_CH(dw, dir, ch, name) \
+#define GET_CH_32(dw, dir, ch, name) \
 	readl_ch(dw, dir, ch, &(__dw_ch_regs(dw, dir, ch)->name))
 
-#define SET_LL(ll, value) \
+#define SET_LL_32(ll, value) \
 	writel(value, ll)
 
+#ifdef CONFIG_64BIT
+
+static inline void writeq_ch(struct dw_edma *dw, enum dw_edma_dir dir, u16 ch,
+			     u64 value, void __iomem *addr)
+{
+	if (dw->mf == EDMA_MF_EDMA_LEGACY) {
+		u32 viewport_sel;
+		unsigned long flags;
+
+		raw_spin_lock_irqsave(&dw->lock, flags);
+
+		viewport_sel = FIELD_PREP(EDMA_V0_VIEWPORT_MASK, ch);
+		if (dir == EDMA_DIR_READ)
+			viewport_sel |= BIT(31);
+
+		writel(viewport_sel,
+		       &(__dw_regs(dw)->type.legacy.viewport_sel));
+		writeq(value, addr);
+
+		raw_spin_unlock_irqrestore(&dw->lock, flags);
+	} else {
+		writeq(value, addr);
+	}
+}
+
+static inline u64 readq_ch(struct dw_edma *dw, enum dw_edma_dir dir, u16 ch,
+			   const void __iomem *addr)
+{
+	u32 value;
+
+	if (dw->mf == EDMA_MF_EDMA_LEGACY) {
+		u32 viewport_sel;
+		unsigned long flags;
+
+		raw_spin_lock_irqsave(&dw->lock, flags);
+
+		viewport_sel = FIELD_PREP(EDMA_V0_VIEWPORT_MASK, ch);
+		if (dir == EDMA_DIR_READ)
+			viewport_sel |= BIT(31);
+
+		writel(viewport_sel,
+		       &(__dw_regs(dw)->type.legacy.viewport_sel));
+		value = readq(addr);
+
+		raw_spin_unlock_irqrestore(&dw->lock, flags);
+	} else {
+		value = readq(addr);
+	}
+
+	return value;
+}
+
+#define SET_CH_64(dw, dir, ch, name, value) \
+	writeq_ch(dw, dir, ch, value, &(__dw_ch_regs(dw, dir, ch)->name))
+
+#define GET_CH_64(dw, dir, ch, name) \
+	readq_ch(dw, dir, ch, &(__dw_ch_regs(dw, dir, ch)->name))
+
+#define SET_LL_64(ll, value) \
+	writeq(value, ll)
+
+#endif /* CONFIG_64BIT */
+
 /* eDMA management callbacks */
 void dw_edma_v0_core_off(struct dw_edma *dw)
 {
-	SET_BOTH(dw, int_mask, EDMA_V0_DONE_INT_MASK | EDMA_V0_ABORT_INT_MASK);
-	SET_BOTH(dw, int_clear, EDMA_V0_DONE_INT_MASK | EDMA_V0_ABORT_INT_MASK);
-	SET_BOTH(dw, engine_en, 0);
+	SET_BOTH_32(dw, int_mask,
+		    EDMA_V0_DONE_INT_MASK | EDMA_V0_ABORT_INT_MASK);
+	SET_BOTH_32(dw, int_clear,
+		    EDMA_V0_DONE_INT_MASK | EDMA_V0_ABORT_INT_MASK);
+	SET_BOTH_32(dw, engine_en, 0);
 }
 
 u16 dw_edma_v0_core_ch_count(struct dw_edma *dw, enum dw_edma_dir dir)
@@ -137,9 +242,11 @@ u16 dw_edma_v0_core_ch_count(struct dw_edma *dw, enum dw_edma_dir dir)
 	u32 num_ch;
 
 	if (dir == EDMA_DIR_WRITE)
-		num_ch = FIELD_GET(EDMA_V0_WRITE_CH_COUNT_MASK, GET(dw, ctrl));
+		num_ch = FIELD_GET(EDMA_V0_WRITE_CH_COUNT_MASK,
+				   GET_32(dw, ctrl));
 	else
-		num_ch = FIELD_GET(EDMA_V0_READ_CH_COUNT_MASK, GET(dw, ctrl));
+		num_ch = FIELD_GET(EDMA_V0_READ_CH_COUNT_MASK,
+				   GET_32(dw, ctrl));
 
 	if (num_ch > EDMA_V0_MAX_NR_CH)
 		num_ch = EDMA_V0_MAX_NR_CH;
@@ -153,7 +260,7 @@ enum dma_status dw_edma_v0_core_ch_status(struct dw_edma_chan *chan)
 	u32 tmp;
 
 	tmp = FIELD_GET(EDMA_V0_CH_STATUS_MASK,
-			GET_CH(dw, chan->dir, chan->id, ch_control1));
+			GET_CH_32(dw, chan->dir, chan->id, ch_control1));
 
 	if (tmp == 1)
 		return DMA_IN_PROGRESS;
@@ -167,26 +274,28 @@ void dw_edma_v0_core_clear_done_int(struct dw_edma_chan *chan)
 {
 	struct dw_edma *dw = chan->chip->dw;
 
-	SET_RW(dw, chan->dir, int_clear,
-	       FIELD_PREP(EDMA_V0_DONE_INT_MASK, BIT(chan->id)));
+	SET_RW_32(dw, chan->dir, int_clear,
+		  FIELD_PREP(EDMA_V0_DONE_INT_MASK, BIT(chan->id)));
 }
 
 void dw_edma_v0_core_clear_abort_int(struct dw_edma_chan *chan)
 {
 	struct dw_edma *dw = chan->chip->dw;
 
-	SET_RW(dw, chan->dir, int_clear,
-	       FIELD_PREP(EDMA_V0_ABORT_INT_MASK, BIT(chan->id)));
+	SET_RW_32(dw, chan->dir, int_clear,
+		  FIELD_PREP(EDMA_V0_ABORT_INT_MASK, BIT(chan->id)));
 }
 
 u32 dw_edma_v0_core_status_done_int(struct dw_edma *dw, enum dw_edma_dir dir)
 {
-	return FIELD_GET(EDMA_V0_DONE_INT_MASK, GET_RW(dw, dir, int_status));
+	return FIELD_GET(EDMA_V0_DONE_INT_MASK,
+			 GET_RW_32(dw, dir, int_status));
 }
 
 u32 dw_edma_v0_core_status_abort_int(struct dw_edma *dw, enum dw_edma_dir dir)
 {
-	return FIELD_GET(EDMA_V0_ABORT_INT_MASK, GET_RW(dw, dir, int_status));
+	return FIELD_GET(EDMA_V0_ABORT_INT_MASK,
+			 GET_RW_32(dw, dir, int_status));
 }
 
 static void dw_edma_v0_core_write_chunk(struct dw_edma_chunk *chunk)
@@ -209,15 +318,23 @@ static void dw_edma_v0_core_write_chunk(struct dw_edma_chunk *chunk)
 			control |= (DW_EDMA_V0_LIE | DW_EDMA_V0_RIE);
 
 		/* Channel control */
-		SET_LL(&lli[i].control, control);
+		SET_LL_32(&lli[i].control, control);
 		/* Transfer size */
-		SET_LL(&lli[i].transfer_size, child->sz);
-		/* SAR - low, high */
-		SET_LL(&lli[i].sar_low, lower_32_bits(child->sar));
-		SET_LL(&lli[i].sar_high, upper_32_bits(child->sar));
-		/* DAR - low, high */
-		SET_LL(&lli[i].dar_low, lower_32_bits(child->dar));
-		SET_LL(&lli[i].dar_high, upper_32_bits(child->dar));
+		SET_LL_32(&lli[i].transfer_size, child->sz);
+		/* SAR */
+		#ifdef CONFIG_64BIT
+			SET_LL_64(&lli[i].sar.reg, child->sar);
+		#else /* CONFIG_64BIT */
+			SET_LL_32(&lli[i].sar.lsb, lower_32_bits(child->sar));
+			SET_LL_32(&lli[i].sar.msb, upper_32_bits(child->sar));
+		#endif /* CONFIG_64BIT */
+		/* DAR */
+		#ifdef CONFIG_64BIT
+			SET_LL_64(&lli[i].dar.reg, child->dar);
+		#else /* CONFIG_64BIT */
+			SET_LL_32(&lli[i].dar.lsb, lower_32_bits(child->dar));
+			SET_LL_32(&lli[i].dar.msb, upper_32_bits(child->dar));
+		#endif /* CONFIG_64BIT */
 		i++;
 	}
 
@@ -227,10 +344,14 @@ static void dw_edma_v0_core_write_chunk(struct dw_edma_chunk *chunk)
 		control |= DW_EDMA_V0_CB;
 
 	/* Channel control */
-	SET_LL(&llp->control, control);
-	/* Linked list  - low, high */
-	SET_LL(&llp->llp_low, lower_32_bits(chunk->ll_region.paddr));
-	SET_LL(&llp->llp_high, upper_32_bits(chunk->ll_region.paddr));
+	SET_LL_32(&llp->control, control);
+	/* Linked list */
+	#ifdef CONFIG_64BIT
+		SET_LL_64(&llp->llp.reg, chunk->ll_region.paddr);
+	#else /* CONFIG_64BIT */
+		SET_LL_32(&llp->llp.lsb, lower_32_bits(chunk->ll_region.paddr));
+		SET_LL_32(&llp->llp.msb, upper_32_bits(chunk->ll_region.paddr));
+	#endif /* CONFIG_64BIT */
 }
 
 void dw_edma_v0_core_start(struct dw_edma_chunk *chunk, bool first)
@@ -243,28 +364,33 @@ void dw_edma_v0_core_start(struct dw_edma_chunk *chunk, bool first)
 
 	if (first) {
 		/* Enable engine */
-		SET_RW(dw, chan->dir, engine_en, BIT(0));
+		SET_RW_32(dw, chan->dir, engine_en, BIT(0));
 		/* Interrupt unmask - done, abort */
-		tmp = GET_RW(dw, chan->dir, int_mask);
+		tmp = GET_RW_32(dw, chan->dir, int_mask);
 		tmp &= ~FIELD_PREP(EDMA_V0_DONE_INT_MASK, BIT(chan->id));
 		tmp &= ~FIELD_PREP(EDMA_V0_ABORT_INT_MASK, BIT(chan->id));
-		SET_RW(dw, chan->dir, int_mask, tmp);
+		SET_RW_32(dw, chan->dir, int_mask, tmp);
 		/* Linked list error */
-		tmp = GET_RW(dw, chan->dir, linked_list_err_en);
+		tmp = GET_RW_32(dw, chan->dir, linked_list_err_en);
 		tmp |= FIELD_PREP(EDMA_V0_LINKED_LIST_ERR_MASK, BIT(chan->id));
-		SET_RW(dw, chan->dir, linked_list_err_en, tmp);
+		SET_RW_32(dw, chan->dir, linked_list_err_en, tmp);
 		/* Channel control */
-		SET_CH(dw, chan->dir, chan->id, ch_control1,
-		       (DW_EDMA_V0_CCS | DW_EDMA_V0_LLE));
-		/* Linked list - low, high */
-		SET_CH(dw, chan->dir, chan->id, llp_low,
-		       lower_32_bits(chunk->ll_region.paddr));
-		SET_CH(dw, chan->dir, chan->id, llp_high,
-		       upper_32_bits(chunk->ll_region.paddr));
+		SET_CH_32(dw, chan->dir, chan->id, ch_control1,
+			  (DW_EDMA_V0_CCS | DW_EDMA_V0_LLE));
+		/* Linked list */
+		#ifdef CONFIG_64BIT
+			SET_CH_64(dw, chan->dir, chan->id, llp.reg,
+				  chunk->ll_region.paddr);
+		#else /* CONFIG_64BIT */
+			SET_CH_32(dw, chan->dir, chan->id, llp.lsb,
+				  lower_32_bits(chunk->ll_region.paddr));
+			SET_CH_32(dw, chan->dir, chan->id, llp.msb,
+				  upper_32_bits(chunk->ll_region.paddr));
+		#endif /* CONFIG_64BIT */
 	}
 	/* Doorbell */
-	SET_RW(dw, chan->dir, doorbell,
-	       FIELD_PREP(EDMA_V0_DOORBELL_CH_MASK, chan->id));
+	SET_RW_32(dw, chan->dir, doorbell,
+		  FIELD_PREP(EDMA_V0_DOORBELL_CH_MASK, chan->id));
 }
 
 int dw_edma_v0_core_device_config(struct dw_edma_chan *chan)
@@ -273,31 +399,31 @@ int dw_edma_v0_core_device_config(struct dw_edma_chan *chan)
 	u32 tmp = 0;
 
 	/* MSI done addr - low, high */
-	SET_RW(dw, chan->dir, done_imwr_low, chan->msi.address_lo);
-	SET_RW(dw, chan->dir, done_imwr_high, chan->msi.address_hi);
+	SET_RW_32(dw, chan->dir, done_imwr.lsb, chan->msi.address_lo);
+	SET_RW_32(dw, chan->dir, done_imwr.msb, chan->msi.address_hi);
 	/* MSI abort addr - low, high */
-	SET_RW(dw, chan->dir, abort_imwr_low, chan->msi.address_lo);
-	SET_RW(dw, chan->dir, abort_imwr_high, chan->msi.address_hi);
+	SET_RW_32(dw, chan->dir, abort_imwr.lsb, chan->msi.address_lo);
+	SET_RW_32(dw, chan->dir, abort_imwr.msb, chan->msi.address_hi);
 	/* MSI data - low, high */
 	switch (chan->id) {
 	case 0:
 	case 1:
-		tmp = GET_RW(dw, chan->dir, ch01_imwr_data);
+		tmp = GET_RW_32(dw, chan->dir, ch01_imwr_data);
 		break;
 
 	case 2:
 	case 3:
-		tmp = GET_RW(dw, chan->dir, ch23_imwr_data);
+		tmp = GET_RW_32(dw, chan->dir, ch23_imwr_data);
 		break;
 
 	case 4:
 	case 5:
-		tmp = GET_RW(dw, chan->dir, ch45_imwr_data);
+		tmp = GET_RW_32(dw, chan->dir, ch45_imwr_data);
 		break;
 
 	case 6:
 	case 7:
-		tmp = GET_RW(dw, chan->dir, ch67_imwr_data);
+		tmp = GET_RW_32(dw, chan->dir, ch67_imwr_data);
 		break;
 	}
 
@@ -316,22 +442,22 @@ int dw_edma_v0_core_device_config(struct dw_edma_chan *chan)
 	switch (chan->id) {
 	case 0:
 	case 1:
-		SET_RW(dw, chan->dir, ch01_imwr_data, tmp);
+		SET_RW_32(dw, chan->dir, ch01_imwr_data, tmp);
 		break;
 
 	case 2:
 	case 3:
-		SET_RW(dw, chan->dir, ch23_imwr_data, tmp);
+		SET_RW_32(dw, chan->dir, ch23_imwr_data, tmp);
 		break;
 
 	case 4:
 	case 5:
-		SET_RW(dw, chan->dir, ch45_imwr_data, tmp);
+		SET_RW_32(dw, chan->dir, ch45_imwr_data, tmp);
 		break;
 
 	case 6:
 	case 7:
-		SET_RW(dw, chan->dir, ch67_imwr_data, tmp);
+		SET_RW_32(dw, chan->dir, ch67_imwr_data, tmp);
 		break;
 	}
 
drivers/dma/dw-edma/dw-edma-v0-debugfs.c +24 -24

88 lines not on this page · raw message

drivers/dma/dw-edma/dw-edma-v0-regs.h +112 -37

215 lines not on this page · raw message

Signature · 1 line
2.7.4
#2 lore raw

PATCH v6 2/15 dmaengine: dw-edma: Fix comments offset characters' alignment

In reply to Gustavo Pimentel

Fix comments offset characters' alignment to follow the same structure of similar comments.

---
 drivers/dma/dw-edma/dw-edma-v0-regs.h | 214 +++++++++++++++++-----------------
 1 file changed, 107 insertions(+), 107 deletions(-)
drivers/dma/dw-edma/dw-edma-v0-regs.h +107 -107
index d07151d..e175f7b 100644
@@ -25,177 +25,177 @@
 #define EDMA_V0_CH_EVEN_MSI_DATA_MASK			GENMASK(15, 0)
 
 struct dw_edma_v0_ch_regs {
-	u32 ch_control1;				/* 0x000 */
-	u32 ch_control2;				/* 0x004 */
-	u32 transfer_size;				/* 0x008 */
+	u32 ch_control1;				/* 0x0000 */
+	u32 ch_control2;				/* 0x0004 */
+	u32 transfer_size;				/* 0x0008 */
 	union {
-		u64 reg;				/* 0x00c..0x010 */
+		u64 reg;				/* 0x000c..0x0010 */
 		struct {
-			u32 lsb;			/* 0x00c */
-			u32 msb;			/* 0x010 */
+			u32 lsb;			/* 0x000c */
+			u32 msb;			/* 0x0010 */
 		};
 	} sar;
 	union {
-		u64 reg;				/* 0x014..0x018 */
+		u64 reg;				/* 0x0014..0x0018 */
 		struct {
-			u32 lsb;			/* 0x014 */
-			u32 msb;			/* 0x018 */
+			u32 lsb;			/* 0x0014 */
+			u32 msb;			/* 0x0018 */
 		};
 	} dar;
 	union {
-		u64 reg;				/* 0x01c..0x020 */
+		u64 reg;				/* 0x001c..0x0020 */
 		struct {
-			u32 lsb;			/* 0x01c */
-			u32 msb;			/* 0x020 */
+			u32 lsb;			/* 0x001c */
+			u32 msb;			/* 0x0020 */
 		};
 	} llp;
 } __packed;
 
 struct dw_edma_v0_ch {
-	struct dw_edma_v0_ch_regs wr;			/* 0x200 */
-	u32 padding_1[55];				/* [0x224..0x2fc] */
-	struct dw_edma_v0_ch_regs rd;			/* 0x300 */
-	u32 padding_2[55];				/* [0x324..0x3fc] */
+	struct dw_edma_v0_ch_regs wr;			/* 0x0200 */
+	u32 padding_1[55];				/* 0x0224..0x02fc */
+	struct dw_edma_v0_ch_regs rd;			/* 0x0300 */
+	u32 padding_2[55];				/* 0x0324..0x03fc */
 } __packed;
 
 struct dw_edma_v0_unroll {
-	u32 padding_1;					/* 0x0f8 */
-	u32 wr_engine_chgroup;				/* 0x100 */
-	u32 rd_engine_chgroup;				/* 0x104 */
+	u32 padding_1;					/* 0x00f8 */
+	u32 wr_engine_chgroup;				/* 0x0100 */
+	u32 rd_engine_chgroup;				/* 0x0104 */
 	union {
-		u64 reg;				/* 0x108..0x10c */
+		u64 reg;				/* 0x0108..0x010c */
 		struct {
-			u32 lsb;			/* 0x108 */
-			u32 msb;			/* 0x10c */
+			u32 lsb;			/* 0x0108 */
+			u32 msb;			/* 0x010c */
 		};
 	} wr_engine_hshake_cnt;
-	u32 padding_2[2];				/* [0x110..0x114] */
+	u32 padding_2[2];				/* 0x0110..0x0114 */
 	union {
-		u64 reg;				/* 0x120..0x124 */
+		u64 reg;				/* 0x0120..0x0124 */
 		struct {
-			u32 lsb;			/* 0x120 */
-			u32 msb;			/* 0x124 */
+			u32 lsb;			/* 0x0120 */
+			u32 msb;			/* 0x0124 */
 		};
 	} rd_engine_hshake_cnt;
-	u32 padding_3[2];				/* [0x120..0x124] */
-	u32 wr_ch0_pwr_en;				/* 0x128 */
-	u32 wr_ch1_pwr_en;				/* 0x12c */
-	u32 wr_ch2_pwr_en;				/* 0x130 */
-	u32 wr_ch3_pwr_en;				/* 0x134 */
-	u32 wr_ch4_pwr_en;				/* 0x138 */
-	u32 wr_ch5_pwr_en;				/* 0x13c */
-	u32 wr_ch6_pwr_en;				/* 0x140 */
-	u32 wr_ch7_pwr_en;				/* 0x144 */
-	u32 padding_4[8];				/* [0x148..0x164] */
-	u32 rd_ch0_pwr_en;				/* 0x168 */
-	u32 rd_ch1_pwr_en;				/* 0x16c */
-	u32 rd_ch2_pwr_en;				/* 0x170 */
-	u32 rd_ch3_pwr_en;				/* 0x174 */
-	u32 rd_ch4_pwr_en;				/* 0x178 */
-	u32 rd_ch5_pwr_en;				/* 0x18c */
-	u32 rd_ch6_pwr_en;				/* 0x180 */
-	u32 rd_ch7_pwr_en;				/* 0x184 */
-	u32 padding_5[30];				/* [0x188..0x1fc] */
-	struct dw_edma_v0_ch ch[EDMA_V0_MAX_NR_CH];	/* [0x200..0x1120] */
+	u32 padding_3[2];				/* 0x0120..0x0124 */
+	u32 wr_ch0_pwr_en;				/* 0x0128 */
+	u32 wr_ch1_pwr_en;				/* 0x012c */
+	u32 wr_ch2_pwr_en;				/* 0x0130 */
+	u32 wr_ch3_pwr_en;				/* 0x0134 */
+	u32 wr_ch4_pwr_en;				/* 0x0138 */
+	u32 wr_ch5_pwr_en;				/* 0x013c */
+	u32 wr_ch6_pwr_en;				/* 0x0140 */
+	u32 wr_ch7_pwr_en;				/* 0x0144 */
+	u32 padding_4[8];				/* 0x0148..0x0164 */
+	u32 rd_ch0_pwr_en;				/* 0x0168 */
+	u32 rd_ch1_pwr_en;				/* 0x016c */
+	u32 rd_ch2_pwr_en;				/* 0x0170 */
+	u32 rd_ch3_pwr_en;				/* 0x0174 */
+	u32 rd_ch4_pwr_en;				/* 0x0178 */
+	u32 rd_ch5_pwr_en;				/* 0x018c */
+	u32 rd_ch6_pwr_en;				/* 0x0180 */
+	u32 rd_ch7_pwr_en;				/* 0x0184 */
+	u32 padding_5[30];				/* 0x0188..0x01fc */
+	struct dw_edma_v0_ch ch[EDMA_V0_MAX_NR_CH];	/* 0x0200..0x1120 */
 } __packed;
 
 struct dw_edma_v0_legacy {
-	u32 viewport_sel;				/* 0x0f8 */
-	struct dw_edma_v0_ch_regs ch;			/* [0x100..0x120] */
+	u32 viewport_sel;				/* 0x00f8 */
+	struct dw_edma_v0_ch_regs ch;			/* 0x0100..0x0120 */
 } __packed;
 
 struct dw_edma_v0_regs {
 	/* eDMA global registers */
-	u32 ctrl_data_arb_prior;			/* 0x000 */
-	u32 padding_1;					/* 0x004 */
-	u32 ctrl;					/* 0x008 */
-	u32 wr_engine_en;				/* 0x00c */
-	u32 wr_doorbell;				/* 0x010 */
-	u32 padding_2;					/* 0x014 */
+	u32 ctrl_data_arb_prior;			/* 0x0000 */
+	u32 padding_1;					/* 0x0004 */
+	u32 ctrl;					/* 0x0008 */
+	u32 wr_engine_en;				/* 0x000c */
+	u32 wr_doorbell;				/* 0x0010 */
+	u32 padding_2;					/* 0x0014 */
 	union {
-		u64 reg;				/* 0x018..0x01c */
+		u64 reg;				/* 0x0018..0x001c */
 		struct {
-			u32 lsb;			/* 0x018 */
-			u32 msb;			/* 0x01c */
+			u32 lsb;			/* 0x0018 */
+			u32 msb;			/* 0x001c */
 		};
 	} wr_ch_arb_weight;
-	u32 padding_3[3];				/* [0x020..0x028] */
-	u32 rd_engine_en;				/* 0x02c */
-	u32 rd_doorbell;				/* 0x030 */
-	u32 padding_4;					/* 0x034 */
+	u32 padding_3[3];				/* 0x0020..0x0028 */
+	u32 rd_engine_en;				/* 0x002c */
+	u32 rd_doorbell;				/* 0x0030 */
+	u32 padding_4;					/* 0x0034 */
 	union {
-		u64 reg;				/* 0x038..0x03c */
+		u64 reg;				/* 0x0038..0x003c */
 		struct {
-			u32 lsb;			/* 0x038 */
-			u32 msb;			/* 0x03c */
+			u32 lsb;			/* 0x0038 */
+			u32 msb;			/* 0x003c */
 		};
 	} rd_ch_arb_weight;
-	u32 padding_5[3];				/* [0x040..0x048] */
+	u32 padding_5[3];				/* 0x0040..0x0048 */
 	/* eDMA interrupts registers */
-	u32 wr_int_status;				/* 0x04c */
-	u32 padding_6;					/* 0x050 */
-	u32 wr_int_mask;				/* 0x054 */
-	u32 wr_int_clear;				/* 0x058 */
-	u32 wr_err_status;				/* 0x05c */
+	u32 wr_int_status;				/* 0x004c */
+	u32 padding_6;					/* 0x0050 */
+	u32 wr_int_mask;				/* 0x0054 */
+	u32 wr_int_clear;				/* 0x0058 */
+	u32 wr_err_status;				/* 0x005c */
 	union {
-		u64 reg;				/* 0x060..0x064 */
+		u64 reg;				/* 0x0060..0x0064 */
 		struct {
-			u32 lsb;			/* 0x060 */
-			u32 msb;			/* 0x064 */
+			u32 lsb;			/* 0x0060 */
+			u32 msb;			/* 0x0064 */
 		};
 	} wr_done_imwr;
 	union {
-		u64 reg;				/* 0x068..0x06c */
+		u64 reg;				/* 0x0068..0x006c */
 		struct {
-			u32 lsb;			/* 0x068 */
-			u32 msb;			/* 0x06c */
+			u32 lsb;			/* 0x0068 */
+			u32 msb;			/* 0x006c */
 		};
 	} wr_abort_imwr;
-	u32 wr_ch01_imwr_data;				/* 0x070 */
-	u32 wr_ch23_imwr_data;				/* 0x074 */
-	u32 wr_ch45_imwr_data;				/* 0x078 */
-	u32 wr_ch67_imwr_data;				/* 0x07c */
-	u32 padding_7[4];				/* [0x080..0x08c] */
-	u32 wr_linked_list_err_en;			/* 0x090 */
-	u32 padding_8[3];				/* [0x094..0x09c] */
-	u32 rd_int_status;				/* 0x0a0 */
-	u32 padding_9;					/* 0x0a4 */
-	u32 rd_int_mask;				/* 0x0a8 */
-	u32 rd_int_clear;				/* 0x0ac */
-	u32 padding_10;					/* 0x0b0 */
+	u32 wr_ch01_imwr_data;				/* 0x0070 */
+	u32 wr_ch23_imwr_data;				/* 0x0074 */
+	u32 wr_ch45_imwr_data;				/* 0x0078 */
+	u32 wr_ch67_imwr_data;				/* 0x007c */
+	u32 padding_7[4];				/* 0x0080..0x008c */
+	u32 wr_linked_list_err_en;			/* 0x0090 */
+	u32 padding_8[3];				/* 0x0094..0x009c */
+	u32 rd_int_status;				/* 0x00a0 */
+	u32 padding_9;					/* 0x00a4 */
+	u32 rd_int_mask;				/* 0x00a8 */
+	u32 rd_int_clear;				/* 0x00ac */
+	u32 padding_10;					/* 0x00b0 */
 	union {
-		u64 reg;				/* 0x0b4..0x0b8 */
+		u64 reg;				/* 0x00b4..0x00b8 */
 		struct {
-			u32 lsb;			/* 0x0b4 */
-			u32 msb;			/* 0x0b8 */
+			u32 lsb;			/* 0x00b4 */
+			u32 msb;			/* 0x00b8 */
 		};
 	} rd_err_status;
-	u32 padding_11[2];				/* [0x0bc..0x0c0] */
-	u32 rd_linked_list_err_en;			/* 0x0c4 */
-	u32 padding_12;					/* 0x0c8 */
+	u32 padding_11[2];				/* 0x00bc..0x00c0 */
+	u32 rd_linked_list_err_en;			/* 0x00c4 */
+	u32 padding_12;					/* 0x00c8 */
 	union {
-		u64 reg;				/* 0x0cc..0x0d0 */
+		u64 reg;				/* 0x00cc..0x00d0 */
 		struct {
-			u32 lsb;			/* 0x0cc */
-			u32 msb;			/* 0x0d0 */
+			u32 lsb;			/* 0x00cc */
+			u32 msb;			/* 0x00d0 */
 		};
 	} rd_done_imwr;
 	union {
-		u64 reg;				/* 0x0d4..0x0d8 */
+		u64 reg;				/* 0x00d4..0x00d8 */
 		struct {
-			u32 lsb;			/* 0x0d4 */
-			u32 msb;			/* 0x0d8 */
+			u32 lsb;			/* 0x00d4 */
+			u32 msb;			/* 0x00d8 */
 		};
 	} rd_abort_imwr;
-	u32 rd_ch01_imwr_data;				/* 0x0dc */
-	u32 rd_ch23_imwr_data;				/* 0x0e0 */
-	u32 rd_ch45_imwr_data;				/* 0x0e4 */
-	u32 rd_ch67_imwr_data;				/* 0x0e8 */
-	u32 padding_13[4];				/* [0x0ec..0x0f8] */
+	u32 rd_ch01_imwr_data;				/* 0x00dc */
+	u32 rd_ch23_imwr_data;				/* 0x00e0 */
+	u32 rd_ch45_imwr_data;				/* 0x00e4 */
+	u32 rd_ch67_imwr_data;				/* 0x00e8 */
+	u32 padding_13[4];				/* 0x00ec..0x00f8 */
 	/* eDMA channel context grouping */
 	union dw_edma_v0_type {
-		struct dw_edma_v0_legacy legacy;	/* [0x0f8..0x120] */
-		struct dw_edma_v0_unroll unroll;	/* [0x0f8..0x1120] */
+		struct dw_edma_v0_legacy legacy;	/* 0x00f8..0x0120 */
+		struct dw_edma_v0_unroll unroll;	/* 0x00f8..0x1120 */
 	} type;
 } __packed;
 
Signature · 1 line
2.7.4
#3 lore raw

PATCH v6 3/15 dmaengine: dw-edma: Add support for the HDMA feature

In reply to Gustavo Pimentel

Add support for the HDMA feature.

This new feature enables the current eDMA IP to use a deeper prefetch of the linked list, which reduces the algorithm execution latency observed when loading the elements of the list, causing more stable and higher data transfer.

---
 drivers/dma/dw-edma/dw-edma-core.h       | 10 ++++----
 drivers/dma/dw-edma/dw-edma-pcie.c       | 23 ++++++++---------
 drivers/dma/dw-edma/dw-edma-v0-core.c    | 42 +++++++++++++++++++++++++++++---
 drivers/dma/dw-edma/dw-edma-v0-debugfs.c |  9 +++----
 4 files changed, 60 insertions(+), 24 deletions(-)
drivers/dma/dw-edma/dw-edma-core.h +5 -5
index 31fc50d..3f9593e 100644
@@ -21,9 +21,10 @@ enum dw_edma_dir {
 	EDMA_DIR_READ
 };
 
-enum dw_edma_mode {
-	EDMA_MODE_LEGACY = 0,
-	EDMA_MODE_UNROLL
+enum dw_edma_map_format {
+	EDMA_MF_EDMA_LEGACY = 0x0,
+	EDMA_MF_EDMA_UNROLL = 0x1,
+	EDMA_MF_HDMA_COMPAT = 0x5
 };
 
 enum dw_edma_request {
@@ -123,8 +124,7 @@ struct dw_edma {
 	struct dw_edma_irq		*irq;
 	int				nr_irqs;
 
-	u32				version;
-	enum dw_edma_mode		mode;
+	enum dw_edma_map_format		mf;
 
 	struct dw_edma_chan		*chan;
 	const struct dw_edma_core_ops	*ops;
drivers/dma/dw-edma/dw-edma-pcie.c +12 -11
index 1eafc60..c130549 100644
@@ -30,8 +30,7 @@ struct dw_edma_pcie_data {
 	off_t				dt_off;
 	size_t				dt_sz;
 	/* Other */
-	u32				version;
-	enum dw_edma_mode		mode;
+	enum dw_edma_map_format		mf;
 	u8				irqs;
 };
 
@@ -49,8 +48,7 @@ static const struct dw_edma_pcie_data snps_edda_data = {
 	.dt_off				= 0x00800000,	/*  8 Mbytes */
 	.dt_sz				= 0x03800000,	/* 56 Mbytes */
 	/* Other */
-	.version			= 0,
-	.mode				= EDMA_MODE_UNROLL,
+	.mf				= EDMA_MF_EDMA_UNROLL,
 	.irqs				= 1,
 };
 
@@ -69,8 +67,8 @@ static int dw_edma_pcie_probe(struct pci_dev *pdev,
 	const struct dw_edma_pcie_data *pdata = (void *)pid->driver_data;
 	struct device *dev = &pdev->dev;
 	struct dw_edma_chip *chip;
-	int err, nr_irqs;
 	struct dw_edma *dw;
+	int err, nr_irqs;
 
 	/* Enable PCI device */
 	err = pcim_enable_device(pdev);
@@ -157,16 +155,19 @@ static int dw_edma_pcie_probe(struct pci_dev *pdev,
 	dw->dt_region.paddr += pdata->dt_off;
 	dw->dt_region.sz = pdata->dt_sz;
 
-	dw->version = pdata->version;
-	dw->mode = pdata->mode;
+	dw->mf = pdata->mf;
 	dw->nr_irqs = nr_irqs;
 	dw->ops = &dw_edma_pcie_core_ops;
 
 	/* Debug info */
-	pci_dbg(pdev, "Version:\t%u\n", dw->version);
-
-	pci_dbg(pdev, "Mode:\t%s\n",
-		dw->mode == EDMA_MODE_LEGACY ? "Legacy" : "Unroll");
+	if (dw->mf == EDMA_MF_EDMA_LEGACY)
+		pci_dbg(pdev, "Version:\teDMA Port Logic (0x%x)\n", dw->mf);
+	else if (dw->mf == EDMA_MF_EDMA_UNROLL)
+		pci_dbg(pdev, "Version:\teDMA Unroll (0x%x)\n", dw->mf);
+	else if (dw->mf == EDMA_MF_HDMA_COMPAT)
+		pci_dbg(pdev, "Version:\tHDMA Compatible (0x%x)\n", dw->mf);
+	else
+		pci_dbg(pdev, "Version:\tUnknown (0x%x)\n", dw->mf);
 
 	pci_dbg(pdev, "Registers:\tBAR=%u, off=0x%.8lx, sz=0x%zx bytes, addr(v=%p, p=%pa)\n",
 		pdata->rg_bar, pdata->rg_off, pdata->rg_sz,
drivers/dma/dw-edma/dw-edma-v0-core.c +39 -3

66 lines not on this page · raw message

drivers/dma/dw-edma/dw-edma-v0-debugfs.c +4 -5
index a5e2783..157dfc2 100644
@@ -55,7 +55,7 @@ struct debugfs_entries {
 static int dw_edma_debugfs_u32_get(void *data, u64 *val)
 {
 	void __iomem *reg = (void __force __iomem *)data;
-	if (dw->mode == EDMA_MODE_LEGACY &&
+	if (dw->mf == EDMA_MF_EDMA_LEGACY &&
 	    reg >= (void __iomem *)&regs->type.legacy.ch) {
 		void __iomem *ptr = &regs->type.legacy.ch;
 		u32 viewport_sel = 0;
@@ -174,7 +174,7 @@ static void dw_edma_debugfs_regs_wr(struct dentry *dir)
 	nr_entries = ARRAY_SIZE(debugfs_regs);
 	dw_edma_debugfs_create_x32(debugfs_regs, nr_entries, regs_dir);
 
-	if (dw->mode == EDMA_MODE_UNROLL) {
+	if (dw->mf == EDMA_MF_HDMA_COMPAT) {
 		nr_entries = ARRAY_SIZE(debugfs_unroll_regs);
 		dw_edma_debugfs_create_x32(debugfs_unroll_regs, nr_entries,
 					   regs_dir);
@@ -243,7 +243,7 @@ static void dw_edma_debugfs_regs_rd(struct dentry *dir)
 	nr_entries = ARRAY_SIZE(debugfs_regs);
 	dw_edma_debugfs_create_x32(debugfs_regs, nr_entries, regs_dir);
 
-	if (dw->mode == EDMA_MODE_UNROLL) {
+	if (dw->mf == EDMA_MF_HDMA_COMPAT) {
 		nr_entries = ARRAY_SIZE(debugfs_unroll_regs);
 		dw_edma_debugfs_create_x32(debugfs_unroll_regs, nr_entries,
 					   regs_dir);
@@ -297,8 +297,7 @@ void dw_edma_v0_debugfs_on(struct dw_edma_chip *chip)
 	if (!base_dir)
 		return;
 
-	debugfs_create_u32("version", 0444, base_dir, &dw->version);
-	debugfs_create_u32("mode", 0444, base_dir, &dw->mode);
+	debugfs_create_u32("mf", 0444, base_dir, &dw->mf);
 	debugfs_create_u16("wr_ch_cnt", 0444, base_dir, &dw->wr_ch_cnt);
 	debugfs_create_u16("rd_ch_cnt", 0444, base_dir, &dw->rd_ch_cnt);
 
Signature · 1 line
2.7.4
#4 lore raw

PATCH v6 4/15 PCI: Add pci_find_vsec_capability() to find a specific VSEC

In reply to Gustavo Pimentel

Adds another helper to ones that already exist called pci_find_vsec_capability. This helper crawls through the device PCI config space searching for a specific ID on the Vendor-Specific Extended Capabilities section.

The Vendor-Specific Extended Capability (VSEC) is a special PCI capability (acts like container) defined by PCI-SIG that allows the one or more proprietary capabilities defined by the vendor which aren't standard or shared between the manufactures.

---
 drivers/pci/pci.c             | 34 ++++++++++++++++++++++++++++++++++
 include/linux/pci.h           |  2 ++
 include/uapi/linux/pci_regs.h |  6 ++++++
 3 files changed, 42 insertions(+)
drivers/pci/pci.c +34 -0
index b9fecc2..628aa9f 100644
@@ -693,6 +693,40 @@ u8 pci_find_ht_capability(struct pci_dev *dev, int ht_cap)
 EXPORT_SYMBOL_GPL(pci_find_ht_capability);
 
 /**
+ * pci_find_vsec_capability - Find a vendor-specific extended capability
+ * @dev: PCI device to query
+ * @cap: vendor-specific capability ID code
+ *
+ * Typically this function will be called by the PCI driver, which passes
+ * through argument the 'struct pci_dev *' already pointing for the device
+ * config space that is associated with the vendor and device ID which will
+ * know which ID to search and what to do with it, however, there might be
+ * cases that this function could be called outside of this scope and
+ * therefore is the caller responsibility to check the vendor and/or
+ * device ID first.
+ *
+ * Returns the address of the vendor-specific structure that matches the
+ * requested capability ID code within the device's PCI configuration space
+ * or 0 if it does not find a match.
+ */
+u16 pci_find_vsec_capability(struct pci_dev *dev, int vsec_cap_id)
+{
+	u16 vsec = 0;
+	u32 header;
+
+	while ((vsec = pci_find_next_ext_capability(dev, vsec,
+						     PCI_EXT_CAP_ID_VNDR))) {
+		if (pci_read_config_dword(dev, vsec + PCI_VSEC_HDR,
+					  &header) == PCIBIOS_SUCCESSFUL &&
+		    PCI_VSEC_CAP_ID(header) == vsec_cap_id)
+			return vsec;
+	}
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(pci_find_vsec_capability);
+
+/**
  * pci_find_parent_resource - return resource region of parent bus of given
  *			      region
  * @dev: PCI device structure contains resources to be searched
include/linux/pci.h +2 -0
index b32126d..da6ab6a 100644
@@ -1080,6 +1080,8 @@ struct pci_bus *pci_find_next_bus(const struct pci_bus *from);
 
 u64 pci_get_dsn(struct pci_dev *dev);
 
+u16 pci_find_vsec_capability(struct pci_dev *dev, int vsec_cap_id);
+
 struct pci_dev *pci_get_device(unsigned int vendor, unsigned int device,
 			       struct pci_dev *from);
 struct pci_dev *pci_get_subsys(unsigned int vendor, unsigned int device,
include/uapi/linux/pci_regs.h +6 -0
index e709ae8..deae275 100644
@@ -983,6 +983,12 @@
 #define PCI_VSEC_HDR		4	/* extended cap - vendor-specific */
 #define  PCI_VSEC_HDR_LEN_SHIFT	20	/* shift for length field */
 
+/* Vendor-Specific Extended Capabilities */
+#define PCI_VSEC_HEADER		4	/* Vendor-Specific Header */
+#define  PCI_VSEC_CAP_ID(x)	((x) & 0xffff)
+#define  PCI_VSEC_CAP_REV(x)	(((x) >> 16) & 0xf)
+#define  PCI_VSEC_CAP_LEN(x)	(((x) >> 20) & 0xfff)
+
 /* SATA capability */
 #define PCI_SATA_REGS		4	/* SATA REGs specifier */
 #define  PCI_SATA_REGS_MASK	0xF	/* location - BAR#/inline */
Signature · 1 line
2.7.4
#5 lore raw

Re: [PATCH v6 04/15] PCI: Add pci_find_vsec_capability() to find a specific VSEC

[+cc Krzysztof, since he commented on a previous version] [+cc Lukas, who previously proposed exactly what I suggest below, sorry for repeating. I think Lukas was right to propose passing in the vendor ID because it makes it easier to read the caller.]

When you post new versions of a series, please cc people who commented on previous versions.

Quoted 4 lines · Gustavo Pimentel

Adds another helper to ones that already exist called pci_find_vsec_capability. This helper crawls through the device PCI config space searching for a specific ID on the Vendor-Specific Extended Capabilities section.

Add pci_find_vsec_capability() to locate a Vendor-Specific Extended Capability with the specified VSEC ID.

Quoted 5 lines

The Vendor-Specific Extended Capability (VSEC) is a special PCI capability (acts like container) defined by PCI-SIG that allows the one or more proprietary capabilities defined by the vendor which aren't standard or shared between the manufactures.

s/is a special ... by PCI-SIG that//
s/allows the one/allows one/
s/the manufactures/manufacturers/ (or maybe "vendors" to match previous use)

Quoted 16 lines
---
 drivers/pci/pci.c             | 34 ++++++++++++++++++++++++++++++++++
 include/linux/pci.h           |  2 ++
 include/uapi/linux/pci_regs.h |  6 ++++++
 3 files changed, 42 insertions(+)
drivers/pci/pci.c +1 -0
index b9fecc2..628aa9f 100644
@@ -693,6 +693,40 @@ u8 pci_find_ht_capability(struct pci_dev *dev, int ht_cap)
 EXPORT_SYMBOL_GPL(pci_find_ht_capability);
 
 /**
+ * pci_find_vsec_capability - Find a vendor-specific extended capability

drivers/pci/pci.c u8 pci_find_ht_capability(struct pci_dev *dev, int ht_cap) · line 706

Quoted 10 lines
drivers/pci/pci.c +10 -0
+ * @dev: PCI device to query
+ * @cap: vendor-specific capability ID code
+ *
+ * Typically this function will be called by the PCI driver, which passes
+ * through argument the 'struct pci_dev *' already pointing for the device
+ * config space that is associated with the vendor and device ID which will
+ * know which ID to search and what to do with it, however, there might be
+ * cases that this function could be called outside of this scope and
+ * therefore is the caller responsibility to check the vendor and/or
+ * device ID first.

This is important because it's a bit subtle. IIUC, each vendor (identified by Vendor ID at 0x00 in config space) can define its own VSEC IDs, so it can define up to 2^16 == 64K VSEC structures.

Of course there's not room for that many in config space; but the point is that the vendor chooses its own VSEC IDs and doesn't need to coordinate with anybody.

So a VSEC ID 0x0006 in a Synopsys device (Vendor ID 0x16c3) has nothing to do with a VSEC ID 0x0006 in an Intel device (Vendor ID 0x8086), and it's up to the caller to make sure it's using the correct one.

I wonder if it would help avoid mistakes if we made the interface look like this:

  u16 pci_find_vsec_capability(struct pci_dev *dev, u16 vendor, int vsec_cap_id)
  {
    if (vendor != dev->vendor)
      return 0;
    while ((vsec = ...))
      ...

}

so calls would look like this:

  vsec = pci_find_vsec_capability(dev, PCI_VENDOR_ID_SYNOPSYS, DW_PCIE_VSEC_DMA_ID);

which would make it more obvious that DW_PCIE_VSEC_DMA_ID is only valid in a Synopsys device.

The function comment could be something like this:

pci_find_vsec_capability - Find a vendor-specific extended capability @dev: PCI device to query
@vendor: Vendor ID for which capability is defined @vsec_cap_id: Vendor-specific capability ID

If @dev has Vendor ID @vendor, search for a VSEC capability with VSEC ID @vsec_cap_id. If found, return the capability offset in config space; otherwise return 0.

Or maybe it's even more subtle than I thought, and I'm missing something :)

Quoted 30 lines
drivers/pci/pci.c +22 -0
+ * Returns the address of the vendor-specific structure that matches the
+ * requested capability ID code within the device's PCI configuration space
+ * or 0 if it does not find a match.
+ */
+u16 pci_find_vsec_capability(struct pci_dev *dev, int vsec_cap_id)
+{
+	u16 vsec = 0;
+	u32 header;
+
+	while ((vsec = pci_find_next_ext_capability(dev, vsec,
+						     PCI_EXT_CAP_ID_VNDR))) {
+		if (pci_read_config_dword(dev, vsec + PCI_VSEC_HDR,
+					  &header) == PCIBIOS_SUCCESSFUL &&
+		    PCI_VSEC_CAP_ID(header) == vsec_cap_id)
+			return vsec;
+	}
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(pci_find_vsec_capability);
+
+/**
  * pci_find_parent_resource - return resource region of parent bus of given
  *			      region
  * @dev: PCI device structure contains resources to be searched
include/linux/pci.h +0 -0
index b32126d..da6ab6a 100644

include/linux/pci.h struct pci_bus *pci_find_next_bus(const struct pci_bus *from); · line 1083

Quoted 4 lines
include/linux/pci.h +1 -0
@@ -1080,6 +1080,8 @@ struct pci_bus *pci_find_next_bus(const struct pci_bus *from);
 
 u64 pci_get_dsn(struct pci_dev *dev);
 
+u16 pci_find_vsec_capability(struct pci_dev *dev, int vsec_cap_id);

Can you put this declaration up with the other "find_capability" declarations just a few lines up?

Quoted 8 lines
include/linux/pci.h +0 -0
 struct pci_dev *pci_get_device(unsigned int vendor, unsigned int device,
 			       struct pci_dev *from);
 struct pci_dev *pci_get_subsys(unsigned int vendor, unsigned int device,
include/uapi/linux/pci_regs.h +0 -0
index e709ae8..deae275 100644

include/uapi/linux/pci_regs.h line 990

Quoted 8 lines
include/uapi/linux/pci_regs.h +5 -0
@@ -983,6 +983,12 @@
 #define PCI_VSEC_HDR		4	/* extended cap - vendor-specific */
 #define  PCI_VSEC_HDR_LEN_SHIFT	20	/* shift for length field */
 
+/* Vendor-Specific Extended Capabilities */
+#define PCI_VSEC_HEADER		4	/* Vendor-Specific Header */
+#define  PCI_VSEC_CAP_ID(x)	((x) & 0xffff)
+#define  PCI_VSEC_CAP_REV(x)	(((x) >> 16) & 0xf)
+#define  PCI_VSEC_CAP_LEN(x)	(((x) >> 20) & 0xfff)

The VSEC #defines are a mess. We have already have some duplication and inconsistent names. But I think what you need is already there:

  /* Vendor-Specific (VSEC, PCI_EXT_CAP_ID_VNDR) */
  #define PCI_VNDR_HEADER         4       /* Vendor-Specific Header */
  #define  PCI_VNDR_HEADER_ID(x)  ((x) & 0xffff)
  #define  PCI_VNDR_HEADER_REV(x) (((x) >> 16) & 0xf)
  #define  PCI_VNDR_HEADER_LEN(x) (((x) >> 20) & 0xfff)
Quoted 6 lines
include/uapi/linux/pci_regs.h +0 -0
 /* SATA capability */
 #define PCI_SATA_REGS		4	/* SATA REGs specifier */
 #define  PCI_SATA_REGS_MASK	0xF	/* location - BAR#/inline */
Signature · 1 line
2.7.4
#6 lore raw

RE: [PATCH v6 04/15] PCI: Add pci_find_vsec_capability() to find a specific VSEC

Quoted 7 lines · Bjorn Helgaas

[+cc Krzysztof, since he commented on a previous version] [+cc Lukas, who previously proposed exactly what I suggest below, sorry for repeating. I think Lukas was right to propose passing in the vendor ID because it makes it easier to read the caller.]

When you post new versions of a series, please cc people who commented on previous versions.

Noted.

Quoted 10 lines · Gustavo Pimentel
Quoted 4 lines

Adds another helper to ones that already exist called pci_find_vsec_capability. This helper crawls through the device PCI config space searching for a specific ID on the Vendor-Specific Extended Capabilities section.

Add pci_find_vsec_capability() to locate a Vendor-Specific Extended Capability with the specified VSEC ID.

Quoted 1 line
Quoted 8 lines
Quoted 4 lines

The Vendor-Specific Extended Capability (VSEC) is a special PCI capability (acts like container) defined by PCI-SIG that allows the one or more proprietary capabilities defined by the vendor which aren't standard or shared between the manufactures.

s/is a special ... by PCI-SIG that//
s/allows the one/allows one/
s/the manufactures/manufacturers/ (or maybe "vendors" to match previous use)

Ok, I'll include those changes.

Quoted 67 lines
Quoted 17 lines
---
 drivers/pci/pci.c             | 34 ++++++++++++++++++++++++++++++++++
 include/linux/pci.h           |  2 ++
 include/uapi/linux/pci_regs.h |  6 ++++++
 3 files changed, 42 insertions(+)
drivers/pci/pci.c +1 -0
index b9fecc2..628aa9f 100644
@@ -693,6 +693,40 @@ u8 pci_find_ht_capability(struct pci_dev *dev, int ht_cap)
 EXPORT_SYMBOL_GPL(pci_find_ht_capability);
 
 /**
+ * pci_find_vsec_capability - Find a vendor-specific extended capability

drivers/pci/pci.c u8 pci_find_ht_capability(struct pci_dev *dev, int ht_cap) · line 706

Quoted 10 lines
drivers/pci/pci.c +10 -0
+ * @dev: PCI device to query
+ * @cap: vendor-specific capability ID code
+ *
+ * Typically this function will be called by the PCI driver, which passes
+ * through argument the 'struct pci_dev *' already pointing for the device
+ * config space that is associated with the vendor and device ID which will
+ * know which ID to search and what to do with it, however, there might be
+ * cases that this function could be called outside of this scope and
+ * therefore is the caller responsibility to check the vendor and/or
+ * device ID first.

This is important because it's a bit subtle. IIUC, each vendor (identified by Vendor ID at 0x00 in config space) can define its own VSEC IDs, so it can define up to 2^16 == 64K VSEC structures.

Of course there's not room for that many in config space; but the point is that the vendor chooses its own VSEC IDs and doesn't need to coordinate with anybody.

So a VSEC ID 0x0006 in a Synopsys device (Vendor ID 0x16c3) has nothing to do with a VSEC ID 0x0006 in an Intel device (Vendor ID 0x8086), and it's up to the caller to make sure it's using the correct one.

I wonder if it would help avoid mistakes if we made the interface look like this:

  u16 pci_find_vsec_capability(struct pci_dev *dev, u16 vendor, int vsec_cap_id)
  {
    if (vendor != dev->vendor)
      return 0;
    while ((vsec = ...))
      ...

}

so calls would look like this:

  vsec = pci_find_vsec_capability(dev, PCI_VENDOR_ID_SYNOPSYS, DW_PCIE_VSEC_DMA_ID);

which would make it more obvious that DW_PCIE_VSEC_DMA_ID is only valid in a Synopsys device.

The function comment could be something like this:

pci_find_vsec_capability - Find a vendor-specific extended capability @dev: PCI device to query
@vendor: Vendor ID for which capability is defined @vsec_cap_id: Vendor-specific capability ID

Quoted 6 lines

If @dev has Vendor ID @vendor, search for a VSEC capability with VSEC ID @vsec_cap_id. If found, return the capability offset in config space; otherwise return 0.

Or maybe it's even more subtle than I thought, and I'm missing something :)

Okay, seems reasonable.

Quoted 32 lines
Quoted 32 lines
drivers/pci/pci.c +22 -0
+ * Returns the address of the vendor-specific structure that matches the
+ * requested capability ID code within the device's PCI configuration space
+ * or 0 if it does not find a match.
+ */
+u16 pci_find_vsec_capability(struct pci_dev *dev, int vsec_cap_id)
+{
+	u16 vsec = 0;
+	u32 header;
+
+	while ((vsec = pci_find_next_ext_capability(dev, vsec,
+						     PCI_EXT_CAP_ID_VNDR))) {
+		if (pci_read_config_dword(dev, vsec + PCI_VSEC_HDR,
+					  &header) == PCIBIOS_SUCCESSFUL &&
+		    PCI_VSEC_CAP_ID(header) == vsec_cap_id)
+			return vsec;
+	}
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(pci_find_vsec_capability);
+
+/**
  * pci_find_parent_resource - return resource region of parent bus of given
  *			      region
  * @dev: PCI device structure contains resources to be searched
include/linux/pci.h +0 -0
index b32126d..da6ab6a 100644
@@ -1080,6 +1080,8 @@ struct pci_bus *pci_find_next_bus(const struct pci_bus *from);
 
Quoted 6 lines

include/linux/pci.h struct pci_bus *pci_find_next_bus(const struct pci_bus *from); · line 1083

Quoted 3 lines
include/linux/pci.h +1 -0
 u64 pci_get_dsn(struct pci_dev *dev);
 
+u16 pci_find_vsec_capability(struct pci_dev *dev, int vsec_cap_id);

Can you put this declaration up with the other "find_capability" declarations just a few lines up?

Sure.

Quoted 21 lines
Quoted 9 lines
include/linux/pci.h +0 -0
 struct pci_dev *pci_get_device(unsigned int vendor, unsigned int device,
 			       struct pci_dev *from);
 struct pci_dev *pci_get_subsys(unsigned int vendor, unsigned int device,
include/uapi/linux/pci_regs.h +0 -0
index e709ae8..deae275 100644

include/uapi/linux/pci_regs.h line 990

Quoted 8 lines
include/uapi/linux/pci_regs.h +5 -0
@@ -983,6 +983,12 @@
 #define PCI_VSEC_HDR		4	/* extended cap - vendor-specific */
 #define  PCI_VSEC_HDR_LEN_SHIFT	20	/* shift for length field */
 
+/* Vendor-Specific Extended Capabilities */
+#define PCI_VSEC_HEADER		4	/* Vendor-Specific Header */
+#define  PCI_VSEC_CAP_ID(x)	((x) & 0xffff)
+#define  PCI_VSEC_CAP_REV(x)	(((x) >> 16) & 0xf)
+#define  PCI_VSEC_CAP_LEN(x)	(((x) >> 20) & 0xfff)

The VSEC #defines are a mess. We have already have some duplication and inconsistent names. But I think what you need is already there:

Quoted 5 lines
  /* Vendor-Specific (VSEC, PCI_EXT_CAP_ID_VNDR) */
  #define PCI_VNDR_HEADER         4       /* Vendor-Specific Header */
  #define  PCI_VNDR_HEADER_ID(x)  ((x) & 0xffff)
  #define  PCI_VNDR_HEADER_REV(x) (((x) >> 16) & 0xf)
  #define  PCI_VNDR_HEADER_LEN(x) (((x) >> 20) & 0xfff)

Indeed, I'll use this instead.

Quoted 7 lines
Quoted 7 lines
include/uapi/linux/pci_regs.h +0 -0
 /* SATA capability */
 #define PCI_SATA_REGS		4	/* SATA REGs specifier */
 #define  PCI_SATA_REGS_MASK	0xF	/* location - BAR#/inline */
Signature · 1 line
2.7.4

-Gustavo

#7 lore raw

PATCH v6 5/15 dmaengine: dw-edma: Add PCIe VSEC data retrieval support

In reply to Gustavo Pimentel

The latest eDMA IP development implements a Vendor-Specific Extended Capability that contains the eDMA BAR, offset, map format, and the number of read/write channels available.

---
 drivers/dma/dw-edma/dw-edma-core.c |  20 ++++---
 drivers/dma/dw-edma/dw-edma-pcie.c | 114 ++++++++++++++++++++++++++++---------
 2 files changed, 99 insertions(+), 35 deletions(-)
drivers/dma/dw-edma/dw-edma-core.c +12 -8
index b971505..b65c32e1 100644
@@ -863,15 +863,19 @@ int dw_edma_probe(struct dw_edma_chip *chip)
 
 	raw_spin_lock_init(&dw->lock);
 
-	/* Find out how many write channels are supported by hardware */
-	dw->wr_ch_cnt = dw_edma_v0_core_ch_count(dw, EDMA_DIR_WRITE);
-	if (!dw->wr_ch_cnt)
-		return -EINVAL;
+	if (!dw->wr_ch_cnt) {
+		/* Find out how many write channels are supported by hardware */
+		dw->wr_ch_cnt = dw_edma_v0_core_ch_count(dw, EDMA_DIR_WRITE);
+		if (!dw->wr_ch_cnt)
+			return -EINVAL;
+	}
 
-	/* Find out how many read channels are supported by hardware */
-	dw->rd_ch_cnt = dw_edma_v0_core_ch_count(dw, EDMA_DIR_READ);
-	if (!dw->rd_ch_cnt)
-		return -EINVAL;
+	if (!dw->rd_ch_cnt) {
+		/* Find out how many read channels are supported by hardware */
+		dw->rd_ch_cnt = dw_edma_v0_core_ch_count(dw, EDMA_DIR_READ);
+		if (!dw->rd_ch_cnt)
+			return -EINVAL;
+	}
 
 	dev_vdbg(dev, "Channels:\twrite=%d, read=%d\n",
 		 dw->wr_ch_cnt, dw->rd_ch_cnt);
drivers/dma/dw-edma/dw-edma-pcie.c +87 -27

177 lines not on this page · raw message

Signature · 1 line
2.7.4
#8 lore raw

PATCH v6 6/15 dmaengine: dw-edma: Add device_prep_interleave_dma() support

In reply to Gustavo Pimentel

Add device_prep_interleave_dma() support to Synopsys DMA driver.

This feature implements a similar data transfer mechanism to the scatter-gather implementation.

---
 drivers/dma/dw-edma/dw-edma-core.c | 85 ++++++++++++++++++++++++++++++--------
 drivers/dma/dw-edma/dw-edma-core.h | 13 ++++--
 2 files changed, 78 insertions(+), 20 deletions(-)
drivers/dma/dw-edma/dw-edma-core.c +68 -17
index b65c32e1..0fe3835 100644
@@ -329,7 +329,7 @@ dw_edma_device_transfer(struct dw_edma_transfer *xfer)
 	struct dw_edma_chunk *chunk;
 	struct dw_edma_burst *burst;
 	struct dw_edma_desc *desc;
-	u32 cnt;
+	u32 cnt = 0;
 	int i;
 
 	if (!chan->configured)
@@ -352,12 +352,19 @@ dw_edma_device_transfer(struct dw_edma_transfer *xfer)
 		return NULL;
 	}
 
-	if (xfer->cyclic) {
+	if (xfer->type == EDMA_XFER_CYCLIC) {
 		if (!xfer->xfer.cyclic.len || !xfer->xfer.cyclic.cnt)
 			return NULL;
-	} else {
+	} else if (xfer->type == EDMA_XFER_SCATTER_GATHER) {
 		if (xfer->xfer.sg.len < 1)
 			return NULL;
+	} else if (xfer->type == EDMA_XFER_INTERLEAVED) {
+		if (!xfer->xfer.il->numf)
+			return NULL;
+		if (xfer->xfer.il->numf > 0 && xfer->xfer.il->frame_size > 0)
+			return NULL;
+	} else {
+		return NULL;
 	}
 
 	desc = dw_edma_alloc_desc(chan);
@@ -368,18 +375,28 @@ dw_edma_device_transfer(struct dw_edma_transfer *xfer)
 	if (unlikely(!chunk))
 		goto err_alloc;
 
-	src_addr = chan->config.src_addr;
-	dst_addr = chan->config.dst_addr;
+	if (xfer->type == EDMA_XFER_INTERLEAVED) {
+		src_addr = xfer->xfer.il->src_start;
+		dst_addr = xfer->xfer.il->dst_start;
+	} else {
+		src_addr = chan->config.src_addr;
+		dst_addr = chan->config.dst_addr;
+	}
 
-	if (xfer->cyclic) {
+	if (xfer->type == EDMA_XFER_CYCLIC) {
 		cnt = xfer->xfer.cyclic.cnt;
-	} else {
+	} else if (xfer->type == EDMA_XFER_SCATTER_GATHER) {
 		cnt = xfer->xfer.sg.len;
 		sg = xfer->xfer.sg.sgl;
+	} else if (xfer->type == EDMA_XFER_INTERLEAVED) {
+		if (xfer->xfer.il->numf > 0)
+			cnt = xfer->xfer.il->numf;
+		else
+			cnt = xfer->xfer.il->frame_size;
 	}
 
 	for (i = 0; i < cnt; i++) {
-		if (!xfer->cyclic && !sg)
+		if (xfer->type == EDMA_XFER_SCATTER_GATHER && !sg)
 			break;
 
 		if (chunk->bursts_alloc == chan->ll_max) {
@@ -392,19 +409,21 @@ dw_edma_device_transfer(struct dw_edma_transfer *xfer)
 		if (unlikely(!burst))
 			goto err_alloc;
 
-		if (xfer->cyclic)
+		if (xfer->type == EDMA_XFER_CYCLIC)
 			burst->sz = xfer->xfer.cyclic.len;
-		else
+		else if (xfer->type == EDMA_XFER_SCATTER_GATHER)
 			burst->sz = sg_dma_len(sg);
+		else if (xfer->type == EDMA_XFER_INTERLEAVED)
+			burst->sz = xfer->xfer.il->sgl[i].size;
 
 		chunk->ll_region.sz += burst->sz;
 		desc->alloc_sz += burst->sz;
 
 		if (chan->dir == EDMA_DIR_WRITE) {
 			burst->sar = src_addr;
-			if (xfer->cyclic) {
+			if (xfer->type == EDMA_XFER_CYCLIC) {
 				burst->dar = xfer->xfer.cyclic.paddr;
-			} else {
+			} else if (xfer->type == EDMA_XFER_SCATTER_GATHER) {
 				burst->dar = dst_addr;
 				/* Unlike the typical assumption by other
 				 * drivers/IPs the peripheral memory isn't
@@ -416,9 +435,9 @@ dw_edma_device_transfer(struct dw_edma_transfer *xfer)
 			}
 		} else {
 			burst->dar = dst_addr;
-			if (xfer->cyclic) {
+			if (xfer->type == EDMA_XFER_CYCLIC) {
 				burst->sar = xfer->xfer.cyclic.paddr;
-			} else {
+			} else if (xfer->type == EDMA_XFER_SCATTER_GATHER) {
 				burst->sar = src_addr;
 				/* Unlike the typical assumption by other
 				 * drivers/IPs the peripheral memory isn't
@@ -430,10 +449,24 @@ dw_edma_device_transfer(struct dw_edma_transfer *xfer)
 			}
 		}
 
-		if (!xfer->cyclic) {
+		if (xfer->type == EDMA_XFER_SCATTER_GATHER) {
 			src_addr += sg_dma_len(sg);
 			dst_addr += sg_dma_len(sg);
 			sg = sg_next(sg);
+		} else if (xfer->type == EDMA_XFER_INTERLEAVED &&
+			   xfer->xfer.il->frame_size > 0) {
+			struct dma_interleaved_template *il = xfer->xfer.il;
+			struct data_chunk *dc = &il->sgl[i];
+
+			if (il->src_sgl) {
+				src_addr += burst->sz;
+				src_addr += dmaengine_get_src_icg(il, dc);
+			}
+
+			if (il->dst_sgl) {
+				dst_addr += burst->sz;
+				dst_addr += dmaengine_get_dst_icg(il, dc);
+			}
 		}
 	}
 
@@ -459,7 +492,7 @@ dw_edma_device_prep_slave_sg(struct dma_chan *dchan, struct scatterlist *sgl,
 	xfer.xfer.sg.sgl = sgl;
 	xfer.xfer.sg.len = len;
 	xfer.flags = flags;
-	xfer.cyclic = false;
+	xfer.type = EDMA_XFER_SCATTER_GATHER;
 
 	return dw_edma_device_transfer(&xfer);
 }
@@ -478,7 +511,23 @@ dw_edma_device_prep_dma_cyclic(struct dma_chan *dchan, dma_addr_t paddr,
 	xfer.xfer.cyclic.len = len;
 	xfer.xfer.cyclic.cnt = count;
 	xfer.flags = flags;
-	xfer.cyclic = true;
+	xfer.type = EDMA_XFER_CYCLIC;
+
+	return dw_edma_device_transfer(&xfer);
+}
+
+static struct dma_async_tx_descriptor *
+dw_edma_device_prep_interleaved_dma(struct dma_chan *dchan,
+				    struct dma_interleaved_template *ilt,
+				    unsigned long flags)
+{
+	struct dw_edma_transfer xfer;
+
+	xfer.dchan = dchan;
+	xfer.direction = ilt->dir;
+	xfer.xfer.il = ilt;
+	xfer.flags = flags;
+	xfer.type = EDMA_XFER_INTERLEAVED;
 
 	return dw_edma_device_transfer(&xfer);
 }
@@ -738,6 +787,7 @@ static int dw_edma_channel_setup(struct dw_edma_chip *chip, bool write,
 	dma_cap_set(DMA_SLAVE, dma->cap_mask);
 	dma_cap_set(DMA_CYCLIC, dma->cap_mask);
 	dma_cap_set(DMA_PRIVATE, dma->cap_mask);
+	dma_cap_set(DMA_INTERLEAVE, dma->cap_mask);
 	dma->directions = BIT(write ? DMA_DEV_TO_MEM : DMA_MEM_TO_DEV);
 	dma->src_addr_widths = BIT(DMA_SLAVE_BUSWIDTH_4_BYTES);
 	dma->dst_addr_widths = BIT(DMA_SLAVE_BUSWIDTH_4_BYTES);
@@ -756,6 +806,7 @@ static int dw_edma_channel_setup(struct dw_edma_chip *chip, bool write,
 	dma->device_tx_status = dw_edma_device_tx_status;
 	dma->device_prep_slave_sg = dw_edma_device_prep_slave_sg;
 	dma->device_prep_dma_cyclic = dw_edma_device_prep_dma_cyclic;
+	dma->device_prep_interleaved_dma = dw_edma_device_prep_interleaved_dma;
 
 	dma_set_max_seg_size(dma->dev, U32_MAX);
 
drivers/dma/dw-edma/dw-edma-core.h +10 -3
index 3f9593e..f72ebaa 100644
@@ -39,6 +39,12 @@ enum dw_edma_status {
 	EDMA_ST_BUSY
 };
 
+enum dw_edma_xfer_type {
+	EDMA_XFER_SCATTER_GATHER = 0,
+	EDMA_XFER_CYCLIC,
+	EDMA_XFER_INTERLEAVED
+};
+
 struct dw_edma_chan;
 struct dw_edma_chunk;
 
@@ -146,12 +152,13 @@ struct dw_edma_cyclic {
 struct dw_edma_transfer {
 	struct dma_chan			*dchan;
 	union dw_edma_xfer {
-		struct dw_edma_sg	sg;
-		struct dw_edma_cyclic	cyclic;
+		struct dw_edma_sg		sg;
+		struct dw_edma_cyclic		cyclic;
+		struct dma_interleaved_template *il;
 	} xfer;
 	enum dma_transfer_direction	direction;
 	unsigned long			flags;
-	bool				cyclic;
+	enum dw_edma_xfer_type		type;
 };
 
 static inline
Signature · 1 line
2.7.4
#9 lore raw

PATCH v6 7/15 dmaengine: dw-edma: Improve number of channels check

In reply to Gustavo Pimentel

It was added some extra checks to ensure that the driver doesn't try to use more DMA channels than actually are available in hardware.

---
 drivers/dma/dw-edma/dw-edma-core.c | 21 +++++++++------------
 drivers/dma/dw-edma/dw-edma-core.h |  2 ++
 2 files changed, 11 insertions(+), 12 deletions(-)
drivers/dma/dw-edma/dw-edma-core.c +9 -12
index 0fe3835..5495cf7 100644
@@ -914,19 +914,16 @@ int dw_edma_probe(struct dw_edma_chip *chip)
 
 	raw_spin_lock_init(&dw->lock);
 
-	if (!dw->wr_ch_cnt) {
-		/* Find out how many write channels are supported by hardware */
-		dw->wr_ch_cnt = dw_edma_v0_core_ch_count(dw, EDMA_DIR_WRITE);
-		if (!dw->wr_ch_cnt)
-			return -EINVAL;
-	}
+	dw->wr_ch_cnt = min_t(u16, dw->wr_ch_cnt,
+			      dw_edma_v0_core_ch_count(dw, EDMA_DIR_WRITE));
+	dw->wr_ch_cnt = min_t(u16, dw->wr_ch_cnt, EDMA_MAX_WR_CH);
 
-	if (!dw->rd_ch_cnt) {
-		/* Find out how many read channels are supported by hardware */
-		dw->rd_ch_cnt = dw_edma_v0_core_ch_count(dw, EDMA_DIR_READ);
-		if (!dw->rd_ch_cnt)
-			return -EINVAL;
-	}
+	dw->rd_ch_cnt = min_t(u16, dw->rd_ch_cnt,
+			      dw_edma_v0_core_ch_count(dw, EDMA_DIR_READ));
+	dw->rd_ch_cnt = min_t(u16, dw->rd_ch_cnt, EDMA_MAX_RD_CH);
+
+	if (!dw->wr_ch_cnt && !dw->rd_ch_cnt)
+		return -EINVAL;
 
 	dev_vdbg(dev, "Channels:\twrite=%d, read=%d\n",
 		 dw->wr_ch_cnt, dw->rd_ch_cnt);
drivers/dma/dw-edma/dw-edma-core.h +2 -0
index f72ebaa..650b1c7 100644
@@ -15,6 +15,8 @@
 #include "../virt-dma.h"
 
 #define EDMA_LL_SZ					24
+#define EDMA_MAX_WR_CH					8
+#define EDMA_MAX_RD_CH					8
 
 enum dw_edma_dir {
 	EDMA_DIR_WRITE = 0,
Signature · 1 line
2.7.4
#10 lore raw

PATCH v6 8/15 dmaengine: dw-edma: Reorder variables to keep consistency

In reply to Gustavo Pimentel

In the driver code structure, I tried to keep the code style consistency by writing the write channels instructions first, and then follow by the read channels instructions, mimicking the hardware implementation.

However, this code style failed in some cases. This patch fixes that and no functional changes are expected.

---
 drivers/dma/dw-edma/dw-edma-pcie.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)
drivers/dma/dw-edma/dw-edma-pcie.c +5 -5
index 1ddea34..41384ff 100644
@@ -20,8 +20,8 @@
 #define DW_PCIE_VSEC_DMA_ID			0x6
 #define DW_PCIE_VSEC_DMA_BAR			GENMASK(10, 8)
 #define DW_PCIE_VSEC_DMA_MAP			GENMASK(2, 0)
-#define DW_PCIE_VSEC_DMA_RD_CH			GENMASK(25, 16)
 #define DW_PCIE_VSEC_DMA_WR_CH			GENMASK(9, 0)
+#define DW_PCIE_VSEC_DMA_RD_CH			GENMASK(25, 16)
 
 struct dw_edma_pcie_data {
 	/* eDMA registers location */
@@ -39,8 +39,8 @@ struct dw_edma_pcie_data {
 	/* Other */
 	enum dw_edma_map_format		mf;
 	u8				irqs;
-	u16				rd_ch_cnt;
 	u16				wr_ch_cnt;
+	u16				rd_ch_cnt;
 };
 
 static const struct dw_edma_pcie_data snps_edda_data = {
@@ -59,8 +59,8 @@ static const struct dw_edma_pcie_data snps_edda_data = {
 	/* Other */
 	.mf				= EDMA_MF_EDMA_UNROLL,
 	.irqs				= 1,
-	.rd_ch_cnt			= 0,
 	.wr_ch_cnt			= 0,
+	.rd_ch_cnt			= 0,
 };
 
 static int dw_edma_pcie_irq_vector(struct device *dev, unsigned int nr)
@@ -99,8 +99,8 @@ static void dw_edma_pcie_get_vsec_dma_data(struct pci_dev *pdev,
 	pdata->rg_bar = FIELD_GET(DW_PCIE_VSEC_DMA_BAR, val);
 
 	pci_read_config_dword(pdev, vsec + 0xc, &val);
-	pdata->rd_ch_cnt = FIELD_GET(DW_PCIE_VSEC_DMA_RD_CH, val);
 	pdata->wr_ch_cnt = FIELD_GET(DW_PCIE_VSEC_DMA_WR_CH, val);
+	pdata->rd_ch_cnt = FIELD_GET(DW_PCIE_VSEC_DMA_RD_CH, val);
 
 	pci_read_config_dword(pdev, vsec + 0x14, &val);
 	off = val;
@@ -216,8 +216,8 @@ static int dw_edma_pcie_probe(struct pci_dev *pdev,
 	dw->mf = vsec_data.mf;
 	dw->nr_irqs = nr_irqs;
 	dw->ops = &dw_edma_pcie_core_ops;
-	dw->rd_ch_cnt = vsec_data.rd_ch_cnt;
 	dw->wr_ch_cnt = vsec_data.wr_ch_cnt;
+	dw->rd_ch_cnt = vsec_data.rd_ch_cnt;
 
 	/* Debug info */
 	if (dw->mf == EDMA_MF_EDMA_LEGACY)
Signature · 1 line
2.7.4
#11 lore raw

PATCH v6 9/15 dmaengine: dw-edma: Improve the linked list and data blocks definition

In reply to Gustavo Pimentel

In the previous implementation, the driver assumed that there existed only two memory spaces that would equally distribute the amount of read/write channels.

This might not be the case on some other implementations, therefore this patch change this requirement so that each write/read channel has its own linked list and data space well defined, which allows different sizes and locations.

---
 drivers/dma/dw-edma/dw-edma-core.c |  51 +++++-----
 drivers/dma/dw-edma/dw-edma-core.h |   9 +-
 drivers/dma/dw-edma/dw-edma-pcie.c | 185 ++++++++++++++++++++++++++-----------
 3 files changed, 160 insertions(+), 85 deletions(-)
drivers/dma/dw-edma/dw-edma-core.c +24 -27
index 5495cf7..48887b5 100644
@@ -81,8 +81,13 @@ static struct dw_edma_chunk *dw_edma_alloc_chunk(struct dw_edma_desc *desc)
 	 *  - Even chunks originate CB equal to 1
 	 */
 	chunk->cb = !(desc->chunks_alloc % 2);
-	chunk->ll_region.paddr = dw->ll_region.paddr + chan->ll_off;
-	chunk->ll_region.vaddr = dw->ll_region.vaddr + chan->ll_off;
+	if (chan->dir == EDMA_DIR_WRITE) {
+		chunk->ll_region.paddr = dw->ll_region_wr[chan->id].paddr;
+		chunk->ll_region.vaddr = dw->ll_region_wr[chan->id].vaddr;
+	} else {
+		chunk->ll_region.paddr = dw->ll_region_rd[chan->id].paddr;
+		chunk->ll_region.vaddr = dw->ll_region_rd[chan->id].vaddr;
+	}
 
 	if (desc->chunk) {
 		/* Create and add new element into the linked list */
@@ -691,24 +696,13 @@ static int dw_edma_channel_setup(struct dw_edma_chip *chip, bool write,
 	struct device *dev = chip->dev;
 	struct dw_edma *dw = chip->dw;
 	struct dw_edma_chan *chan;
-	size_t ll_chunk, dt_chunk;
 	struct dw_edma_irq *irq;
 	struct dma_device *dma;
-	u32 i, j, cnt, ch_cnt;
 	u32 alloc, off_alloc;
+	u32 i, j, cnt;
 	int err = 0;
 	u32 pos;
 
-	ch_cnt = dw->wr_ch_cnt + dw->rd_ch_cnt;
-	ll_chunk = dw->ll_region.sz;
-	dt_chunk = dw->dt_region.sz;
-
-	/* Calculate linked list chunk for each channel */
-	ll_chunk /= roundup_pow_of_two(ch_cnt);
-
-	/* Calculate linked list chunk for each channel */
-	dt_chunk /= roundup_pow_of_two(ch_cnt);
-
 	if (write) {
 		i = 0;
 		cnt = dw->wr_ch_cnt;
@@ -740,14 +734,14 @@ static int dw_edma_channel_setup(struct dw_edma_chip *chip, bool write,
 		chan->request = EDMA_REQ_NONE;
 		chan->status = EDMA_ST_IDLE;
 
-		chan->ll_off = (ll_chunk * i);
-		chan->ll_max = (ll_chunk / EDMA_LL_SZ) - 1;
-
-		chan->dt_off = (dt_chunk * i);
+		if (write)
+			chan->ll_max = (dw->ll_region_wr[j].sz / EDMA_LL_SZ);
+		else
+			chan->ll_max = (dw->ll_region_rd[j].sz / EDMA_LL_SZ);
+		chan->ll_max -= 1;
 
-		dev_vdbg(dev, "L. List:\tChannel %s[%u] off=0x%.8lx, max_cnt=%u\n",
-			 write ? "write" : "read", j,
-			 chan->ll_off, chan->ll_max);
+		dev_vdbg(dev, "L. List:\tChannel %s[%u] max_cnt=%u\n",
+			 write ? "write" : "read", j, chan->ll_max);
 
 		if (dw->nr_irqs == 1)
 			pos = 0;
@@ -772,12 +766,15 @@ static int dw_edma_channel_setup(struct dw_edma_chip *chip, bool write,
 		chan->vc.desc_free = vchan_free_desc;
 		vchan_init(&chan->vc, dma);
 
-		dt_region->paddr = dw->dt_region.paddr + chan->dt_off;
-		dt_region->vaddr = dw->dt_region.vaddr + chan->dt_off;
-		dt_region->sz = dt_chunk;
-
-		dev_vdbg(dev, "Data:\tChannel %s[%u] off=0x%.8lx\n",
-			 write ? "write" : "read", j, chan->dt_off);
+		if (write) {
+			dt_region->paddr = dw->dt_region_wr[j].paddr;
+			dt_region->vaddr = dw->dt_region_wr[j].vaddr;
+			dt_region->sz = dw->dt_region_wr[j].sz;
+		} else {
+			dt_region->paddr = dw->dt_region_rd[j].paddr;
+			dt_region->vaddr = dw->dt_region_rd[j].vaddr;
+			dt_region->sz = dw->dt_region_rd[j].sz;
+		}
 
 		dw_edma_v0_core_device_config(chan);
 	}
drivers/dma/dw-edma/dw-edma-core.h +4 -5
index 650b1c7..cba5436 100644
@@ -91,11 +91,8 @@ struct dw_edma_chan {
 	int				id;
 	enum dw_edma_dir		dir;
 
-	off_t				ll_off;
 	u32				ll_max;
 
-	off_t				dt_off;
-
 	struct msi_msg			msi;
 
 	enum dw_edma_request		request;
@@ -126,8 +123,10 @@ struct dw_edma {
 	u16				rd_ch_cnt;
 
 	struct dw_edma_region		rg_region;	/* Registers */
-	struct dw_edma_region		ll_region;	/* Linked list */
-	struct dw_edma_region		dt_region;	/* Data */
+	struct dw_edma_region		ll_region_wr[EDMA_MAX_WR_CH];
+	struct dw_edma_region		ll_region_rd[EDMA_MAX_RD_CH];
+	struct dw_edma_region		dt_region_wr[EDMA_MAX_WR_CH];
+	struct dw_edma_region		dt_region_rd[EDMA_MAX_RD_CH];
 
 	struct dw_edma_irq		*irq;
 	int				nr_irqs;
drivers/dma/dw-edma/dw-edma-pcie.c +132 -53

253 lines not on this page · raw message

Signature · 1 line
2.7.4
#12 lore raw

PATCH v6 10/15 dmaengine: dw-edma: Change linked list and data blocks offset and sizes

In reply to Gustavo Pimentel

Changes the linked list and data blocks offset and sizes to follow the recommendation given by the hardware team for the IPK solution.

Although the previous data blocks offset and sizes are still valid and functional, using them that might present some issues related to the IPK solution, since this solution is based on FPGA and might be subjected to timmings constrains.

---
 drivers/dma/dw-edma/dw-edma-pcie.c | 32 ++++++++++++++++----------------
 1 file changed, 16 insertions(+), 16 deletions(-)
drivers/dma/dw-edma/dw-edma-pcie.c +16 -16
index 4e404f9..502de71 100644
@@ -59,29 +59,29 @@ static const struct dw_edma_pcie_data snps_edda_data = {
 	.rg.sz				= 0x00002000,	/*  8 Kbytes */
 	/* eDMA memory linked list location */
 	.ll_wr = {
-		/* Channel 0 - BAR 2, offset 0 Mbytes, size 2 Mbytes */
-		DW_BLOCK(BAR_2, 0x00000000, 0x00200000)
-		/* Channel 1 - BAR 2, offset 2 Mbytes, size 2 Mbytes */
-		DW_BLOCK(BAR_2, 0x00200000, 0x00200000)
+		/* Channel 0 - BAR 2, offset 0 Mbytes, size 2 Kbytes */
+		DW_BLOCK(BAR_2, 0x00000000, 0x00000800)
+		/* Channel 1 - BAR 2, offset 2 Mbytes, size 2 Kbytes */
+		DW_BLOCK(BAR_2, 0x00200000, 0x00000800)
 	},
 	.ll_rd = {
-		/* Channel 0 - BAR 2, offset 4 Mbytes, size 2 Mbytes */
-		DW_BLOCK(BAR_2, 0x00400000, 0x00200000)
-		/* Channel 1 - BAR 2, offset 6 Mbytes, size 2 Mbytes */
-		DW_BLOCK(BAR_2, 0x00600000, 0x00200000)
+		/* Channel 0 - BAR 2, offset 4 Mbytes, size 2 Kbytes */
+		DW_BLOCK(BAR_2, 0x00400000, 0x00000800)
+		/* Channel 1 - BAR 2, offset 6 Mbytes, size 2 Kbytes */
+		DW_BLOCK(BAR_2, 0x00600000, 0x00000800)
 	},
 	/* eDMA memory data location */
 	.dt_wr = {
-		/* Channel 0 - BAR 2, offset 8 Mbytes, size 14 Mbytes */
-		DW_BLOCK(BAR_2, 0x00800000, 0x00e00000)
-		/* Channel 1 - BAR 2, offset 22 Mbytes, size 14 Mbytes */
-		DW_BLOCK(BAR_2, 0x01600000, 0x00e00000)
+		/* Channel 0 - BAR 2, offset 8 Mbytes, size 2 Kbytes */
+		DW_BLOCK(BAR_2, 0x00800000, 0x00000800)
+		/* Channel 1 - BAR 2, offset 9 Mbytes, size 2 Kbytes */
+		DW_BLOCK(BAR_2, 0x00900000, 0x00000800)
 	},
 	.dt_rd = {
-		/* Channel 0 - BAR 2, offset 36 Mbytes, size 14 Mbytes */
-		DW_BLOCK(BAR_2, 0x02400000, 0x00e00000)
-		/* Channel 1 - BAR 2, offset 50 Mbytes, size 14 Mbytes */
-		DW_BLOCK(BAR_2, 0x03200000, 0x00e00000)
+		/* Channel 0 - BAR 2, offset 10 Mbytes, size 2 Kbytes */
+		DW_BLOCK(BAR_2, 0x00a00000, 0x00000800)
+		/* Channel 1 - BAR 2, offset 11 Mbytes, size 2 Kbytes */
+		DW_BLOCK(BAR_2, 0x00b00000, 0x00000800)
 	},
 	/* Other */
 	.mf				= EDMA_MF_EDMA_UNROLL,
Signature · 1 line
2.7.4
#13 lore raw

PATCH v6 11/15 dmaengine: dw-edma: Move struct dentry variable from static definition into dw_edma struct

In reply to Gustavo Pimentel

Move struct dentry variable from static definition (dw-edma-v0-debugfs.c) into dw_edma struct (dw-edma-core.h)

Also the variable was renamed from base_dir to debugfs.

---
 drivers/dma/dw-edma/dw-edma-core.c       |  2 +-
 drivers/dma/dw-edma/dw-edma-core.h       |  3 +++
 drivers/dma/dw-edma/dw-edma-v0-core.c    |  4 ++--
 drivers/dma/dw-edma/dw-edma-v0-core.h    |  2 +-
 drivers/dma/dw-edma/dw-edma-v0-debugfs.c | 22 +++++++++++++---------
 drivers/dma/dw-edma/dw-edma-v0-debugfs.h |  4 ++--
 6 files changed, 22 insertions(+), 15 deletions(-)
drivers/dma/dw-edma/dw-edma-core.c +1 -1
index 48887b5..8d8292e 100644
@@ -1003,7 +1003,7 @@ int dw_edma_remove(struct dw_edma_chip *chip)
 	dma_async_device_unregister(&dw->rd_edma);
 
 	/* Turn debugfs off */
-	dw_edma_v0_core_debugfs_off();
+	dw_edma_v0_core_debugfs_off(chip);
 
 	return 0;
 }
drivers/dma/dw-edma/dw-edma-core.h +3 -0
index cba5436..60316d4 100644
@@ -137,6 +137,9 @@ struct dw_edma {
 	const struct dw_edma_core_ops	*ops;
 
 	raw_spinlock_t			lock;		/* Only for legacy */
+#ifdef CONFIG_DEBUG_FS
+	struct dentry			*debugfs;
+#endif /* CONFIG_DEBUG_FS */
 };
 
 struct dw_edma_sg {
drivers/dma/dw-edma/dw-edma-v0-core.c +2 -2
index 5b0541a..329fc2e 100644
@@ -506,7 +506,7 @@ void dw_edma_v0_core_debugfs_on(struct dw_edma_chip *chip)
 	dw_edma_v0_debugfs_on(chip);
 }
 
-void dw_edma_v0_core_debugfs_off(void)
+void dw_edma_v0_core_debugfs_off(struct dw_edma_chip *chip)
 {
-	dw_edma_v0_debugfs_off();
+	dw_edma_v0_debugfs_off(chip);
 }
drivers/dma/dw-edma/dw-edma-v0-core.h +1 -1
index abae152..2afa626 100644
@@ -23,6 +23,6 @@ void dw_edma_v0_core_start(struct dw_edma_chunk *chunk, bool first);
 int dw_edma_v0_core_device_config(struct dw_edma_chan *chan);
 /* eDMA debug fs callbacks */
 void dw_edma_v0_core_debugfs_on(struct dw_edma_chip *chip);
-void dw_edma_v0_core_debugfs_off(void);
+void dw_edma_v0_core_debugfs_off(struct dw_edma_chip *chip);
 
 #endif /* _DW_EDMA_V0_CORE_H */
drivers/dma/dw-edma/dw-edma-v0-debugfs.c +13 -9
index 157dfc2..4b3bcff 100644
@@ -38,7 +38,6 @@
 #define CHANNEL_STR				"channel"
 #define REGISTERS_STR				"registers"
 
-static struct dentry				*base_dir;
 static struct dw_edma				*dw;
 static struct dw_edma_v0_regs			__iomem *regs;
 
@@ -272,7 +271,7 @@ static void dw_edma_debugfs_regs(void)
 	struct dentry *regs_dir;
 	int nr_entries;
 
-	regs_dir = debugfs_create_dir(REGISTERS_STR, base_dir);
+	regs_dir = debugfs_create_dir(REGISTERS_STR, dw->debugfs);
 	if (!regs_dir)
 		return;
 
@@ -293,18 +292,23 @@ void dw_edma_v0_debugfs_on(struct dw_edma_chip *chip)
 	if (!regs)
 		return;
 
-	base_dir = debugfs_create_dir(dw->name, NULL);
-	if (!base_dir)
+	dw->debugfs = debugfs_create_dir(dw->name, NULL);
+	if (!dw->debugfs)
 		return;
 
-	debugfs_create_u32("mf", 0444, base_dir, &dw->mf);
-	debugfs_create_u16("wr_ch_cnt", 0444, base_dir, &dw->wr_ch_cnt);
-	debugfs_create_u16("rd_ch_cnt", 0444, base_dir, &dw->rd_ch_cnt);
+	debugfs_create_u32("mf", 0444, dw->debugfs, &dw->mf);
+	debugfs_create_u16("wr_ch_cnt", 0444, dw->debugfs, &dw->wr_ch_cnt);
+	debugfs_create_u16("rd_ch_cnt", 0444, dw->debugfs, &dw->rd_ch_cnt);
 
 	dw_edma_debugfs_regs();
 }
 
-void dw_edma_v0_debugfs_off(void)
+void dw_edma_v0_debugfs_off(struct dw_edma_chip *chip)
 {
-	debugfs_remove_recursive(base_dir);
+	dw = chip->dw;
+	if (!dw)
+		return;
+
+	debugfs_remove_recursive(dw->debugfs);
+	dw->debugfs = NULL;
 }
drivers/dma/dw-edma/dw-edma-v0-debugfs.h +2 -2
index 5450a0a..d0ff25a 100644
@@ -13,13 +13,13 @@
 
 #ifdef CONFIG_DEBUG_FS
 void dw_edma_v0_debugfs_on(struct dw_edma_chip *chip);
-void dw_edma_v0_debugfs_off(void);
+void dw_edma_v0_debugfs_off(struct dw_edma_chip *chip);
 #else
 static inline void dw_edma_v0_debugfs_on(struct dw_edma_chip *chip)
 {
 }
 
-static inline void dw_edma_v0_debugfs_off(void)
+static inline void dw_edma_v0_debugfs_off(struct dw_edma_chip *chip)
 {
 }
 #endif /* CONFIG_DEBUG_FS */
Signature · 1 line
2.7.4
#14 lore raw

PATCH v6 12/15 dmaengine: dw-edma: Fix crash on loading/unloading driver

In reply to Gustavo Pimentel

When the driver is compiled as a module and loaded if we try to unload it, the Kernel shows a crash log. This Kernel crash is due to the dma_async_device_unregister() call done after deleting the channels, this patch fixes this issue.

---
 drivers/dma/dw-edma/dw-edma-core.c | 11 +++++------
 1 file changed, 5 insertions(+), 6 deletions(-)
drivers/dma/dw-edma/dw-edma-core.c +5 -6
index 8d8292e..f7a1930 100644
@@ -986,22 +986,21 @@ int dw_edma_remove(struct dw_edma_chip *chip)
 	/* Power management */
 	pm_runtime_disable(dev);
 
+	/* Deregister eDMA device */
+	dma_async_device_unregister(&dw->wr_edma);
 	list_for_each_entry_safe(chan, _chan, &dw->wr_edma.channels,
 				 vc.chan.device_node) {
-		list_del(&chan->vc.chan.device_node);
 		tasklet_kill(&chan->vc.task);
+		list_del(&chan->vc.chan.device_node);
 	}
 
+	dma_async_device_unregister(&dw->rd_edma);
 	list_for_each_entry_safe(chan, _chan, &dw->rd_edma.channels,
 				 vc.chan.device_node) {
-		list_del(&chan->vc.chan.device_node);
 		tasklet_kill(&chan->vc.task);
+		list_del(&chan->vc.chan.device_node);
 	}
 
-	/* Deregister eDMA device */
-	dma_async_device_unregister(&dw->wr_edma);
-	dma_async_device_unregister(&dw->rd_edma);
-
 	/* Turn debugfs off */
 	dw_edma_v0_core_debugfs_off(chip);
 
Signature · 1 line
2.7.4
#15 lore raw

PATCH v6 13/15 dmaengine: dw-edma: Change DMA abreviation from lower into upper case

In reply to Gustavo Pimentel

To keep code consistent, some comments with dma keyword written in lower case are now in upper case.

---
 drivers/dma/dw-edma/dw-edma-core.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
drivers/dma/dw-edma/dw-edma-core.c +3 -3
index f7a1930..a299eed 100644
@@ -341,15 +341,15 @@ dw_edma_device_transfer(struct dw_edma_transfer *xfer)
 		return NULL;
 
 	switch (chan->config.direction) {
-	case DMA_DEV_TO_MEM: /* local dma */
+	case DMA_DEV_TO_MEM: /* local DMA */
 		if (dir == DMA_DEV_TO_MEM && chan->dir == EDMA_DIR_READ)
 			break;
 		return NULL;
-	case DMA_MEM_TO_DEV: /* local dma */
+	case DMA_MEM_TO_DEV: /* local DMA */
 		if (dir == DMA_MEM_TO_DEV && chan->dir == EDMA_DIR_WRITE)
 			break;
 		return NULL;
-	default: /* remote dma */
+	default: /* remote DMA */
 		if (dir == DMA_MEM_TO_DEV && chan->dir == EDMA_DIR_READ)
 			break;
 		if (dir == DMA_DEV_TO_MEM && chan->dir == EDMA_DIR_WRITE)
Signature · 1 line
2.7.4
#16 lore raw

PATCH v6 14/15 dmaengine: dw-edma: Revert fix scatter-gather address calculation

In reply to Gustavo Pimentel

Reverting the applied patch because it caused a regression on ARC700 platform (32 bits).

Fixes 05655541c950 ("dmaengine: dw-edma: Fix scatter-gather address calculation")
---
 drivers/dma/dw-edma/dw-edma-core.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)
drivers/dma/dw-edma/dw-edma-core.c +4 -4
index a299eed..c198451 100644
@@ -429,7 +429,8 @@ dw_edma_device_transfer(struct dw_edma_transfer *xfer)
 			if (xfer->type == EDMA_XFER_CYCLIC) {
 				burst->dar = xfer->xfer.cyclic.paddr;
 			} else if (xfer->type == EDMA_XFER_SCATTER_GATHER) {
-				burst->dar = dst_addr;
+				src_addr += sg_dma_len(sg);
+				burst->dar = sg_dma_address(sg);
 				/* Unlike the typical assumption by other
 				 * drivers/IPs the peripheral memory isn't
 				 * a FIFO memory, in this case, it's a
@@ -443,7 +444,8 @@ dw_edma_device_transfer(struct dw_edma_transfer *xfer)
 			if (xfer->type == EDMA_XFER_CYCLIC) {
 				burst->sar = xfer->xfer.cyclic.paddr;
 			} else if (xfer->type == EDMA_XFER_SCATTER_GATHER) {
-				burst->sar = src_addr;
+				dst_addr += sg_dma_len(sg);
+				burst->sar = sg_dma_address(sg);
 				/* Unlike the typical assumption by other
 				 * drivers/IPs the peripheral memory isn't
 				 * a FIFO memory, in this case, it's a
@@ -455,8 +457,6 @@ dw_edma_device_transfer(struct dw_edma_transfer *xfer)
 		}
 
 		if (xfer->type == EDMA_XFER_SCATTER_GATHER) {
-			src_addr += sg_dma_len(sg);
-			dst_addr += sg_dma_len(sg);
 			sg = sg_next(sg);
 		} else if (xfer->type == EDMA_XFER_INTERLEAVED &&
 			   xfer->xfer.il->frame_size > 0) {
Signature · 1 line
2.7.4
#17 lore raw

PATCH v6 15/15 dmaengine: dw-edma: Add pcim_iomap_table return check

In reply to Gustavo Pimentel

Currently, is missing a null check on a pcim_iomap_table() return value and this can lead to a null pointer dereference if the desired BAR wasn't mapped previously.
Fix this by adding a null check and returning -ENOMEM.

Addresses-Coverity: ("Dereference null return")

---
 drivers/dma/dw-edma/dw-edma-pcie.c | 15 +++++++++++++++
 1 file changed, 15 insertions(+)
drivers/dma/dw-edma/dw-edma-pcie.c +15 -0
index 502de71..70d0d08 100644
@@ -238,6 +238,9 @@ static int dw_edma_pcie_probe(struct pci_dev *pdev,
 	dw->rd_ch_cnt = vsec_data.rd_ch_cnt;
 
 	dw->rg_region.vaddr = pcim_iomap_table(pdev)[vsec_data.rg.bar];
+	if (!dw->rg_region.vaddr)
+		return -ENOMEM;
+
 	dw->rg_region.vaddr += vsec_data.rg.off;
 	dw->rg_region.paddr = pdev->resource[vsec_data.rg.bar].start;
 	dw->rg_region.paddr += vsec_data.rg.off;
@@ -250,12 +253,18 @@ static int dw_edma_pcie_probe(struct pci_dev *pdev,
 		struct dw_edma_block *dt_block = &vsec_data.dt_wr[i];
 
 		ll_region->vaddr = pcim_iomap_table(pdev)[ll_block->bar];
+		if (!ll_region->vaddr)
+			return -ENOMEM;
+
 		ll_region->vaddr += ll_block->off;
 		ll_region->paddr = pdev->resource[ll_block->bar].start;
 		ll_region->paddr += ll_block->off;
 		ll_region->sz = ll_block->sz;
 
 		dt_region->vaddr = pcim_iomap_table(pdev)[dt_block->bar];
+		if (!dt_region->vaddr)
+			return -ENOMEM;
+
 		dt_region->vaddr += dt_block->off;
 		dt_region->paddr = pdev->resource[dt_block->bar].start;
 		dt_region->paddr += dt_block->off;
@@ -269,12 +278,18 @@ static int dw_edma_pcie_probe(struct pci_dev *pdev,
 		struct dw_edma_block *dt_block = &vsec_data.dt_rd[i];
 
 		ll_region->vaddr = pcim_iomap_table(pdev)[ll_block->bar];
+		if (!ll_region->vaddr)
+			return -ENOMEM;
+
 		ll_region->vaddr += ll_block->off;
 		ll_region->paddr = pdev->resource[ll_block->bar].start;
 		ll_region->paddr += ll_block->off;
 		ll_region->sz = ll_block->sz;
 
 		dt_region->vaddr = pcim_iomap_table(pdev)[dt_block->bar];
+		if (!dt_region->vaddr)
+			return -ENOMEM;
+
 		dt_region->vaddr += dt_block->off;
 		dt_region->paddr = pdev->resource[dt_block->bar].start;
 		dt_region->paddr += dt_block->off;
Signature · 1 line
2.7.4