[PATCH/RFC 0/5] mmc: renesas_sdhi: add support for R-Car Gen3 SDHI DMAC

[PATCH/RFC 0/5] mmc: renesas_sdhi: add support for R-Car Gen3 SDHI DMAC

[Simon Horman]

Hi,

this series adds support for the internal DMAC used by r8a779[56] SoCs.
This is achieved by adding a new variant of the SDHI driver for this
DMA controller with compat strings for the r8a779[56] SoCs.
Compat strings for these SoCs are also removed from the existing SYS DMAC
variant of the SDHI driver.

Based on mmc/next

Headline performance boost: 9.5MB/s -> 39.7MB/s
More run-time information below.

Simon Horman (4):
  mmc: sh_mobile_sdhi: set max_segs and max_blk_count values R-Car Gen3
  mmc: tmio: add complete to DMA ops
  mmc: renesas_sdhi: add support for R-Car Gen3 SDHI DMAC
  mmc: renesas-sdhi: remove gen3 support from sysc dmac driver

Yoshihiro Shimoda (1):
  mmc: tmio: add max_segs and max_blk_count in tmio_mmc_data

 drivers/mmc/host/Kconfig                      |  17 ++
 drivers/mmc/host/Makefile                     |   8 +-
 drivers/mmc/host/renesas_sdhi.h               |   2 +
 drivers/mmc/host/renesas_sdhi_core.c          |   2 +
 drivers/mmc/host/renesas_sdhi_internal_dmac.c | 275 ++++++++++++++++++++++++++
 drivers/mmc/host/renesas_sdhi_sys_dmac.c      |  23 +--
 drivers/mmc/host/tmio_mmc.h                   |   2 +
 drivers/mmc/host/tmio_mmc_core.c              |  16 +-
 include/linux/mfd/tmio.h                      |   2 +
 9 files changed, 319 insertions(+), 28 deletions(-)
 create mode 100644 drivers/mmc/host/renesas_sdhi_internal_dmac.c

-- 
2.1.4

M3-W ES1.0

fd9437fc8884 (topic/sdhi-gen3-dma-2017) mmc: renesas-sdhi: remove gen3 support from sysc dmac driver

# dmesg | egrep '(sd|mmc)'
[    1.412911] sdhci: Secure Digital Host Controller Interface driver
[    1.419120] sdhci: Copyright(c) Pierre Ossman
[    1.423990] renesas_sdhi_internal_dmac ee100000.sd: Got CD GPIO
[    1.429951] renesas_sdhi_internal_dmac ee100000.sd: Got WP GPIO
[    1.560503] renesas_sdhi_internal_dmac ee140000.sd: mmc0 base at 0xee140000 max clock rate 200 MHz
[    1.569871] renesas_sdhi_internal_dmac ee160000.sd: Got CD GPIO
[    1.575846] renesas_sdhi_internal_dmac ee160000.sd: Got WP GPIO
[    1.590584] sdhci-pltfm: SDHCI platform and OF driver helper
[    1.638687] renesas_sdhi_internal_dmac ee100000.sd: Got CD GPIO
[    1.644661] renesas_sdhi_internal_dmac ee100000.sd: Got WP GPIO
[    1.772556] renesas_sdhi_internal_dmac ee100000.sd: mmc1 base at 0xee100000 max clock rate 200 MHz
[    1.782613] renesas_sdhi_internal_dmac ee160000.sd: Got CD GPIO
[    1.788580] renesas_sdhi_internal_dmac ee160000.sd: Got WP GPIO
[    1.842940] mmc0: new high speed MMC card at address 0001
[    1.852857] mmcblk0: mmc0:0001 eMMC   28.8 GiB
[    1.857624] mmcblk0boot0: mmc0:0001 eMMC   partition 1 4.00 MiB
[    1.867734] mmcblk0boot1: mmc0:0001 eMMC   partition 2 4.00 MiB
[    1.877858] mmcblk0rpmb: mmc0:0001 eMMC   partition 3 4.00 MiB
[    1.887028]  mmcblk0: p1
[    1.916589] renesas_sdhi_internal_dmac ee160000.sd: mmc2 base at 0xee160000 max clock rate 200 MHz
[    2.060534] mmc1: new ultra high speed SDR50 SDHC card at address e624
[    2.071600] mmcblk1: mmc1:e624 SU08G 7.40 GiB
[    2.086875]  mmcblk1: p1
[    2.260532] mmc2: new ultra high speed SDR50 SDHC card at address 0001
[    2.267407] mmcblk2: mmc2:0001 00000 29.8 GiB
[    2.281572]  mmcblk2: p1

# grep sd /proc/interrupts
100:      24972          0     GIC-0 197 Level     ee100000.sd
101:        604          0     GIC-0 199 Level     ee140000.sd
102:      24980          0     GIC-0 200 Level     ee160000.sd
178:          0          0  e6053000.gpio  12 Edge      ee100000.sd cd
197:          0          0  e6054000.gpio  15 Edge      ee160000.sd cd

# cat /sys/devices/platform/soc/ee100000.sd/mmc_host/mmc1/mmc1:e624
035344535530384780b1b8a11200d300
# dd if=/dev/mmcblk1 of=/dev/null bs=1M count=512 iflag=direct
512+0 records in
512+0 records out
536870912 bytes (537 MB) copied, 13.5376 s, 39.7 MB/s

# cat /sys/devices/platform/soc/ee160000.sd/mmc_host/mmc2/mmc2:0001
1b534d3030303030103916141700f600
# dd if=/dev/mmcblk2 of=/dev/null bs=1M count=512 iflag=direct
512+0 records in
512+0 records out
536870912 bytes (537 MB) copied, 17.5562 s, 30.6 MB/s

# grep sd /proc/interrupts
100:      49548          0     GIC-0 197 Level     ee100000.sd
101:        604          0     GIC-0 199 Level     ee140000.sd
102:      49556          0     GIC-0 200 Level     ee160000.sd
178:          0          0  e6053000.gpio  12 Edge      ee100000.sd cd
197:          0          0  e6054000.gpio  15 Edge      ee160000.sd cd

-----------------------

6e75316c15df (mmc/next) mmc: tmio: make sure SDIO gets reinitialized after resume

# dmesg | egrep '(sd|mmc)'
[    1.428112] sdhci: Secure Digital Host Controller Interface driver
[    1.434620] sdhci: Copyright(c) Pierre Ossman
[    1.439831] sh_mobile_sdhi ee100000.sd: Got CD GPIO
[    1.444764] sh_mobile_sdhi ee100000.sd: Got WP GPIO
[    1.571160] sh_mobile_sdhi ee140000.sd: mmc0 base at 0xee140000 max clock rate 200 MHz
[    1.579496] sh_mobile_sdhi ee160000.sd: Got CD GPIO
[    1.584434] sh_mobile_sdhi ee160000.sd: Got WP GPIO
[    1.598138] sdhci-pltfm: SDHCI platform and OF driver helper
[    1.647916] sh_mobile_sdhi ee100000.sd: Got CD GPIO
[    1.652857] sh_mobile_sdhi ee100000.sd: Got WP GPIO
[    1.779234] sh_mobile_sdhi ee100000.sd: mmc1 base at 0xee100000 max clock rate 200 MHz
[    1.788368] sh_mobile_sdhi ee160000.sd: Got CD GPIO
[    1.793300] sh_mobile_sdhi ee160000.sd: Got WP GPIO
[    1.832493] mmc0: new high speed MMC card at address 0001
[    1.842368] mmcblk0: mmc0:0001 eMMC   28.8 GiB
[    1.851133] mmcblk0boot0: mmc0:0001 eMMC   partition 1 4.00 MiB
[    1.857245] mmcblk0boot1: mmc0:0001 eMMC   partition 2 4.00 MiB
[    1.867323] mmcblk0rpmb: mmc0:0001 eMMC   partition 3 4.00 MiB
[    1.876809]  mmcblk0: p1
[    1.919252] sh_mobile_sdhi ee160000.sd: mmc2 base at 0xee160000 max clock rate 200 MHz
[    2.063208] mmc1: new ultra high speed SDR50 SDHC card at address e624
[    2.074300] mmcblk1: mmc1:e624 SU08G 7.40 GiB
[    2.089887]  mmcblk1: p1
[    2.263192] mmc2: new ultra high speed SDR50 SDHC card at address 0001
[    2.270073] mmcblk2: mmc2:0001 00000 29.8 GiB
[    2.284583]  mmcblk2: p1

# grep sd /proc/interrupts
100:       2276          0     GIC-0 197 Level     ee100000.sd
101:       3040          0     GIC-0 199 Level     ee140000.sd
102:       2284          0     GIC-0 200 Level     ee160000.sd
178:          0          0  e6053000.gpio  12 Edge      ee100000.sd cd
197:          0          0  e6054000.gpio  15 Edge      ee160000.sd cd

# cat /sys/devices/platform/soc/ee100000.sd/mmc_host/mmc1/mmc1:e624 
035344535530384780b1b8a11200d300

# cat /sys/devices/platform/soc/ee160000.sd/mmc_host/mmc2/mmc2:0001 
1b534d3030303030103916141700f600
# dd if=/dev/mmcblk2 of=/dev/null bs=1M count=512 iflag=direct
512+0 records in
512+0 records out
536870912 bytes (537 MB) copied, 56.7809 s, 9.5 MB/s

# grep sd /proc/interrupts
100:       2276          0     GIC-0 197 Level     ee100000.sd
101:       3040          0     GIC-0 199 Level     ee140000.sd
102:    2124012          0     GIC-0 200 Level     ee160000.sd
178:          0          0  e6053000.gpio  12 Edge      ee100000.sd cd
197:          0          0  e6054000.gpio  15 Edge      ee160000.sd cd

[PATCH/RFC 1/5] mmc: tmio: add max_segs and max_blk_count in tmio_mmc_data

[Simon Horman]

From: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>

To change each value from related driver (e.g. sh_mobile_sdhi driver),
this patch adds these member in the struct tmio_mmc_data.

Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Signed-off-by: Ai Kyuse <ai.kyuse.uw@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
---
 drivers/mmc/host/tmio_mmc_core.c | 6 +++---
 include/linux/mfd/tmio.h         | 2 ++
 2 files changed, 5 insertions(+), 3 deletions(-)

diff --git a/drivers/mmc/host/tmio_mmc_core.c b/drivers/mmc/host/tmio_mmc_core.c
index ed4068125349..0e76e099dc7f 100644
--- a/drivers/mmc/host/tmio_mmc_core.c
+++ b/drivers/mmc/host/tmio_mmc_core.c
@@ -1247,10 +1247,10 @@ int tmio_mmc_host_probe(struct tmio_mmc_host *_host,
 
 	mmc->caps |= MMC_CAP_4_BIT_DATA | pdata->capabilities;
 	mmc->caps2 |= pdata->capabilities2;
-	mmc->max_segs = 32;
+	mmc->max_segs = pdata->max_segs ? : 32;
 	mmc->max_blk_size = 512;
-	mmc->max_blk_count = (PAGE_SIZE / mmc->max_blk_size) *
-		mmc->max_segs;
+	mmc->max_blk_count = pdata->max_blk_count ? :
+		(PAGE_SIZE / mmc->max_blk_size) * mmc->max_segs;
 	mmc->max_req_size = mmc->max_blk_size * mmc->max_blk_count;
 	mmc->max_seg_size = mmc->max_req_size;
 
diff --git a/include/linux/mfd/tmio.h b/include/linux/mfd/tmio.h
index c83c16b931a8..4e261ea94d6a 100644
--- a/include/linux/mfd/tmio.h
+++ b/include/linux/mfd/tmio.h
@@ -128,6 +128,8 @@ struct tmio_mmc_data {
 	unsigned int			cd_gpio;
 	int				alignment_shift;
 	dma_addr_t			dma_rx_offset;
+	unsigned int			max_blk_count;
+	unsigned short			max_segs;
 	void (*set_pwr)(struct platform_device *host, int state);
 	void (*set_clk_div)(struct platform_device *host, int state);
 };
-- 
2.1.4

[PATCH/RFC 2/5] mmc: sh_mobile_sdhi: set max_segs and max_blk_count values R-Car Gen3

[Simon Horman]

To avoid complicate code in own dma code for gen3, this patch
adds set max_segs to 1. Then, the tmio driver will get requests
as sg_len = 1 only.

In the performance point of view, the CONFIG_MMC_BLOCK_BOUNCE should
be set. Otherwise, mmc block layer will set 4-kbyte each as a request.

Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Signed-off-by: Ai Kyuse <ai.kyuse.uw@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
---
 drivers/mmc/host/renesas_sdhi.h      | 2 ++
 drivers/mmc/host/renesas_sdhi_core.c | 2 ++
 2 files changed, 4 insertions(+)

diff --git a/drivers/mmc/host/renesas_sdhi.h b/drivers/mmc/host/renesas_sdhi.h
index eb3ea15ff92d..6f7233fe32a2 100644
--- a/drivers/mmc/host/renesas_sdhi.h
+++ b/drivers/mmc/host/renesas_sdhi.h
@@ -31,6 +31,8 @@ struct renesas_sdhi_of_data {
 	int scc_offset;
 	struct renesas_sdhi_scc *taps;
 	int taps_num;
+	unsigned int max_blk_count;
+	unsigned short max_segs;
 };
 
 int renesas_sdhi_probe(struct platform_device *pdev,
diff --git a/drivers/mmc/host/renesas_sdhi_core.c b/drivers/mmc/host/renesas_sdhi_core.c
index 82150a966391..294bc9705c5f 100644
--- a/drivers/mmc/host/renesas_sdhi_core.c
+++ b/drivers/mmc/host/renesas_sdhi_core.c
@@ -522,6 +522,8 @@ int renesas_sdhi_probe(struct platform_device *pdev,
 		mmc_data->capabilities |= of_data->capabilities;
 		mmc_data->capabilities2 |= of_data->capabilities2;
 		mmc_data->dma_rx_offset = of_data->dma_rx_offset;
+		mmc_data->max_blk_count = of_data->max_blk_count;
+		mmc_data->max_segs = of_data->max_segs;
 		dma_priv->dma_buswidth = of_data->dma_buswidth;
 		host->bus_shift = of_data->bus_shift;
 	}
-- 
2.1.4

[PATCH/RFC 2/5] mmc: tmio: add complete to DMA ops

[Simon Horman]

Add complete to DMA ops to allow DMAC implementation dependent
handling of DMA completion.

Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
---
 drivers/mmc/host/tmio_mmc.h      |  1 +
 drivers/mmc/host/tmio_mmc_core.c | 10 ++++++++--
 2 files changed, 9 insertions(+), 2 deletions(-)

diff --git a/drivers/mmc/host/tmio_mmc.h b/drivers/mmc/host/tmio_mmc.h
index 9c94b6eb9b49..82725f9de359 100644
--- a/drivers/mmc/host/tmio_mmc.h
+++ b/drivers/mmc/host/tmio_mmc.h
@@ -121,6 +121,7 @@ struct tmio_mmc_dma_ops {
 			struct tmio_mmc_data *pdata);
 	void (*release)(struct tmio_mmc_host *host);
 	void (*abort)(struct tmio_mmc_host *host);
+	void (*complete)(struct tmio_mmc_host *host);
 };
 
 struct tmio_mmc_host {
diff --git a/drivers/mmc/host/tmio_mmc_core.c b/drivers/mmc/host/tmio_mmc_core.c
index 0e76e099dc7f..26fe9bb64849 100644
--- a/drivers/mmc/host/tmio_mmc_core.c
+++ b/drivers/mmc/host/tmio_mmc_core.c
@@ -86,6 +86,12 @@ static inline void tmio_mmc_abort_dma(struct tmio_mmc_host *host)
 		host->dma_ops->abort(host);
 }
 
+static inline void tmio_mmc_complete_dma(struct tmio_mmc_host *host)
+{
+	if (host->dma_ops)
+		host->dma_ops->complete(host);
+}
+
 void tmio_mmc_enable_mmc_irqs(struct tmio_mmc_host *host, u32 i)
 {
 	host->sdcard_irq_mask &= ~(i & TMIO_MASK_IRQ);
@@ -604,11 +610,11 @@ static void tmio_mmc_data_irq(struct tmio_mmc_host *host, unsigned int stat)
 
 		if (done) {
 			tmio_mmc_disable_mmc_irqs(host, TMIO_STAT_DATAEND);
-			complete(&host->dma_dataend);
+			tmio_mmc_complete_dma(host);
 		}
 	} else if (host->chan_rx && (data->flags & MMC_DATA_READ) && !host->force_pio) {
 		tmio_mmc_disable_mmc_irqs(host, TMIO_STAT_DATAEND);
-		complete(&host->dma_dataend);
+		tmio_mmc_complete_dma(host);
 	} else {
 		tmio_mmc_do_data_irq(host);
 		tmio_mmc_disable_mmc_irqs(host, TMIO_MASK_READOP | TMIO_MASK_WRITEOP);
-- 
2.1.4

[PATCH/RFC 3/5] mmc: sh_mobile_sdhi: add some SoC specific data for R-Car Gen3

[Simon Horman]

To avoid complicate code in own dma code for gen3, this patch
adds set max_segs to 1. Then, the tmio driver will get requests
as sg_len = 1 only.

In the performance point of view, the CONFIG_MMC_BLOCK_BOUNCE should
be set. Otherwise, mmc block layer will set 4-kbyte each as a request.

Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Signed-off-by: Ai Kyuse <ai.kyuse.uw@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
---
 drivers/mmc/host/renesas_sdhi.h      | 2 ++
 drivers/mmc/host/renesas_sdhi_core.c | 2 ++
 2 files changed, 4 insertions(+)

diff --git a/drivers/mmc/host/renesas_sdhi.h b/drivers/mmc/host/renesas_sdhi.h
index eb3ea15ff92d..6f7233fe32a2 100644
--- a/drivers/mmc/host/renesas_sdhi.h
+++ b/drivers/mmc/host/renesas_sdhi.h
@@ -31,6 +31,8 @@ struct renesas_sdhi_of_data {
 	int scc_offset;
 	struct renesas_sdhi_scc *taps;
 	int taps_num;
+	unsigned int max_blk_count;
+	unsigned short max_segs;
 };
 
 int renesas_sdhi_probe(struct platform_device *pdev,
diff --git a/drivers/mmc/host/renesas_sdhi_core.c b/drivers/mmc/host/renesas_sdhi_core.c
index 82150a966391..294bc9705c5f 100644
--- a/drivers/mmc/host/renesas_sdhi_core.c
+++ b/drivers/mmc/host/renesas_sdhi_core.c
@@ -522,6 +522,8 @@ int renesas_sdhi_probe(struct platform_device *pdev,
 		mmc_data->capabilities |= of_data->capabilities;
 		mmc_data->capabilities2 |= of_data->capabilities2;
 		mmc_data->dma_rx_offset = of_data->dma_rx_offset;
+		mmc_data->max_blk_count = of_data->max_blk_count;
+		mmc_data->max_segs = of_data->max_segs;
 		dma_priv->dma_buswidth = of_data->dma_buswidth;
 		host->bus_shift = of_data->bus_shift;
 	}
-- 
2.1.4

[PATCH/RFC 3/5] mmc: tmio: add complete to DMA ops

[Simon Horman]

Add complete to DMA ops to allow DMAC implementation dependent
handling of DMA completion.

Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
---
 drivers/mmc/host/tmio_mmc.h      |  1 +
 drivers/mmc/host/tmio_mmc_core.c | 10 ++++++++--
 2 files changed, 9 insertions(+), 2 deletions(-)

diff --git a/drivers/mmc/host/tmio_mmc.h b/drivers/mmc/host/tmio_mmc.h
index 9c94b6eb9b49..82725f9de359 100644
--- a/drivers/mmc/host/tmio_mmc.h
+++ b/drivers/mmc/host/tmio_mmc.h
@@ -121,6 +121,7 @@ struct tmio_mmc_dma_ops {
 			struct tmio_mmc_data *pdata);
 	void (*release)(struct tmio_mmc_host *host);
 	void (*abort)(struct tmio_mmc_host *host);
+	void (*complete)(struct tmio_mmc_host *host);
 };
 
 struct tmio_mmc_host {
diff --git a/drivers/mmc/host/tmio_mmc_core.c b/drivers/mmc/host/tmio_mmc_core.c
index 0e76e099dc7f..26fe9bb64849 100644
--- a/drivers/mmc/host/tmio_mmc_core.c
+++ b/drivers/mmc/host/tmio_mmc_core.c
@@ -86,6 +86,12 @@ static inline void tmio_mmc_abort_dma(struct tmio_mmc_host *host)
 		host->dma_ops->abort(host);
 }
 
+static inline void tmio_mmc_complete_dma(struct tmio_mmc_host *host)
+{
+	if (host->dma_ops)
+		host->dma_ops->complete(host);
+}
+
 void tmio_mmc_enable_mmc_irqs(struct tmio_mmc_host *host, u32 i)
 {
 	host->sdcard_irq_mask &= ~(i & TMIO_MASK_IRQ);
@@ -604,11 +610,11 @@ static void tmio_mmc_data_irq(struct tmio_mmc_host *host, unsigned int stat)
 
 		if (done) {
 			tmio_mmc_disable_mmc_irqs(host, TMIO_STAT_DATAEND);
-			complete(&host->dma_dataend);
+			tmio_mmc_complete_dma(host);
 		}
 	} else if (host->chan_rx && (data->flags & MMC_DATA_READ) && !host->force_pio) {
 		tmio_mmc_disable_mmc_irqs(host, TMIO_STAT_DATAEND);
-		complete(&host->dma_dataend);
+		tmio_mmc_complete_dma(host);
 	} else {
 		tmio_mmc_do_data_irq(host);
 		tmio_mmc_disable_mmc_irqs(host, TMIO_MASK_READOP | TMIO_MASK_WRITEOP);
-- 
2.1.4

[PATCH/RFC 4/5] mmc: renesas_sdhi: add support for R-Car Gen3 SDHI DMAC

[Simon Horman]

R-Car Gen3 has a dedicated DMA controller for SDHI module. Since
the DMAC is in a part of SDHI module and is not suitable as dmaengine,
this patch adds a different code as tmio_mmc_dma_gen3.c.

Clearing of DM_CM_INFO1 after DMA thanks to Dirk Behme

Cc: Dirk Behme <dirk.behme@de.bosch.com>
Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Signed-off-by: Ai Kyuse <ai.kyuse.uw@renesas.com>
---
 drivers/mmc/host/Kconfig                      |  17 ++
 drivers/mmc/host/Makefile                     |   8 +-
 drivers/mmc/host/renesas_sdhi_internal_dmac.c | 275 ++++++++++++++++++++++++++
 drivers/mmc/host/renesas_sdhi_sys_dmac.c      |   2 +-
 drivers/mmc/host/tmio_mmc.h                   |   1 +
 5 files changed, 301 insertions(+), 2 deletions(-)
 create mode 100644 drivers/mmc/host/renesas_sdhi_internal_dmac.c

diff --git a/drivers/mmc/host/Kconfig b/drivers/mmc/host/Kconfig
index 5755b69f2f72..e813d420d7b3 100644
--- a/drivers/mmc/host/Kconfig
+++ b/drivers/mmc/host/Kconfig
@@ -575,10 +575,27 @@ config MMC_SDHI
 	depends on SUPERH || ARM || ARM64
 	depends on SUPERH || ARCH_RENESAS || COMPILE_TEST
 	select MMC_TMIO_CORE
+	select MMC_SDHI_SYS_DMAC if (SUPERH || ARM)
+	select MMC_SDHI_INTERNAL_DMAC if ARM64
 	help
 	  This provides support for the SDHI SD/SDIO controller found in
 	  Renesas SuperH, ARM and ARM64 based SoCs
 
+config MMC_SDHI_SYS_DMAC
+	tristate "DMA support use of SYS DMAC with SDHI SD/SDIO controller"
+	depends on MMC_SDHI
+	help
+	  This provides DMA support for the SDHI SD/SDIO controller
+	  found in SuperH and Renesas ARM based SoCs.
+
+config MMC_SDHI_INTERNAL_DMAC
+	tristate "DMA support for Internal DMAC with SDHI SD/SDIO controller"
+	depends on ARM64 || COMPILE_TEST
+	depends on MMC_SDHI
+	help
+	  This provides DMA support for the SDHI SD/SDIO controller
+	  found in Renesas arm64 based SoCs.
+
 config MMC_CB710
 	tristate "ENE CB710 MMC/SD Interface support"
 	depends on PCI
diff --git a/drivers/mmc/host/Makefile b/drivers/mmc/host/Makefile
index 4d4547116311..8c46766c000c 100644
--- a/drivers/mmc/host/Makefile
+++ b/drivers/mmc/host/Makefile
@@ -36,7 +36,13 @@ obj-$(CONFIG_MMC_S3C)   	+= s3cmci.o
 obj-$(CONFIG_MMC_SDRICOH_CS)	+= sdricoh_cs.o
 obj-$(CONFIG_MMC_TMIO)		+= tmio_mmc.o
 obj-$(CONFIG_MMC_TMIO_CORE)	+= tmio_mmc_core.o
-obj-$(CONFIG_MMC_SDHI)		+= renesas_sdhi_core.o renesas_sdhi_sys_dmac.o
+obj-$(CONFIG_MMC_SDHI)		+= renesas_sdhi_core.o
+ifeq ($(subst m,y,$(CONFIG_MMC_SDHI_SYS_DMAC)),y)
+obj-$(CONFIG_MMC_SDHI)		+= renesas_sdhi_sys_dmac.o
+endif
+ifeq ($(subst m,y,$(CONFIG_MMC_SDHI_INTERNAL_DMAC)),y)
+obj-$(CONFIG_MMC_SDHI)		+= renesas_sdhi_internal_dmac.o
+endif
 obj-$(CONFIG_MMC_CB710)		+= cb710-mmc.o
 obj-$(CONFIG_MMC_VIA_SDMMC)	+= via-sdmmc.o
 obj-$(CONFIG_SDH_BFIN)		+= bfin_sdh.o
diff --git a/drivers/mmc/host/renesas_sdhi_internal_dmac.c b/drivers/mmc/host/renesas_sdhi_internal_dmac.c
new file mode 100644
index 000000000000..a16b616a2c96
--- /dev/null
+++ b/drivers/mmc/host/renesas_sdhi_internal_dmac.c
@@ -0,0 +1,275 @@
+/*
+ * DMA support for Internal DMAC with SDHI SD/SDIO controller
+ *
+ * Copyright (C) 2016-17 Renesas Electronics Corporation
+ * Copyright (C) 2016-17 Horms Solutions, Simon Horman
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/device.h>
+#include <linux/dma-mapping.h>
+#include <linux/mfd/tmio.h>
+#include <linux/mmc/host.h>
+#include <linux/mod_devicetable.h>
+#include <linux/module.h>
+#include <linux/pagemap.h>
+#include <linux/scatterlist.h>
+
+#include "renesas_sdhi.h"
+#include "tmio_mmc.h"
+
+#define DM_CM_DTRAN_MODE	0x820
+#define DM_CM_DTRAN_CTRL	0x828
+#define DM_CM_RST		0x830
+#define DM_CM_INFO1		0x840
+#define DM_CM_INFO1_MASK	0x848
+#define DM_CM_INFO2		0x850
+#define DM_CM_INFO2_MASK	0x858
+#define DM_DTRAN_ADDR		0x880
+
+/* DM_CM_DTRAN_MODE */
+#define DTRAN_MODE_CH_NUM_CH0	0	/* "downstream" = for write commands */
+#define DTRAN_MODE_CH_NUM_CH1	BIT(16)	/* "uptream" = for read commands */
+#define DTRAN_MODE_BUS_WID_TH	(BIT(5) | BIT(4))
+#define DTRAN_MODE_ADDR_MODE	BIT(0)	/* 1 = Increment address */
+
+/* DM_CM_DTRAN_CTRL */
+#define DTRAN_CTRL_DM_START	BIT(0)
+
+/* DM_CM_RST */
+#define RST_DTRANRST1		BIT(9)
+#define RST_DTRANRST0		BIT(8)
+#define RST_RESERVED_BITS	GENMASK_ULL(32, 0)
+
+/* DM_CM_INFO1 and DM_CM_INFO1_MASK */
+#define INFO1_CLEAR		0
+#define INFO1_DTRANEND1		BIT(17)
+#define INFO1_DTRANEND0		BIT(16)
+
+/* DM_CM_INFO2 and DM_CM_INFO2_MASK */
+#define INFO2_DTRANERR1		BIT(17)
+#define INFO2_DTRANERR0		BIT(16)
+
+/*
+ * Specification of this driver:
+ * - host->chan_{rx,tx} will be used as a flag of enabling/disabling the dma
+ * - Since this SDHI DMAC register set has actual 32-bit and "bus_shift" is 2,
+ *   this driver cannot use original sd_ctrl_{write,read}32 functions.
+ */
+
+/* Definitions for sampling clocks */
+static struct renesas_sdhi_scc rcar_gen3_scc_taps[] = {
+	{
+		.clk_rate = 0,
+		.tap = 0x00000300,
+	},
+};
+
+static const struct renesas_sdhi_of_data of_rcar_gen3_compatible = {
+	.tmio_flags	= TMIO_MMC_HAS_IDLE_WAIT | TMIO_MMC_WRPROTECT_DISABLE |
+			  TMIO_MMC_CLK_ACTUAL | TMIO_MMC_MIN_RCAR2,
+	.capabilities	= MMC_CAP_SD_HIGHSPEED | MMC_CAP_SDIO_IRQ |
+			  MMC_CAP_CMD23,
+	.bus_shift	= 2,
+	.scc_offset	= 0x1000,
+	.taps		= rcar_gen3_scc_taps,
+	.taps_num	= ARRAY_SIZE(rcar_gen3_scc_taps),
+	/* Gen3 SDHI DMAC can handle 0xffffffff blk count, but seg = 1 */
+	.max_blk_count	= 0xffffffff,
+	.max_segs	= 1,
+};
+
+static const struct of_device_id renesas_sdhi_internal_dmac_of_match[] = {
+	{ .compatible = "renesas,sdhi-r8a7795", .data = &of_rcar_gen3_compatible, },
+	{ .compatible = "renesas,sdhi-r8a7796", .data = &of_rcar_gen3_compatible, },
+	{},
+};
+MODULE_DEVICE_TABLE(of, renesas_sdhi_internal_dmac_of_match);
+
+static void
+renesas_sdhi_internal_dmac_dm_write(struct tmio_mmc_host *host,
+				    int addr, u64 val)
+{
+	writeq(val, host->ctl + addr);
+}
+
+static void
+renesas_sdhi_internal_dmac_enable_dma(struct tmio_mmc_host *host, bool enable)
+{
+	if (!host->chan_tx || !host->chan_rx)
+		return;
+
+	if (!enable)
+		renesas_sdhi_internal_dmac_dm_write(host, DM_CM_INFO1,
+						    INFO1_CLEAR);
+
+	if (host->dma->enable)
+		host->dma->enable(host, enable);
+}
+
+static void
+renesas_sdhi_internal_dmac_abort_dma(struct tmio_mmc_host *host) {
+	u64 val = RST_DTRANRST1 | RST_DTRANRST0;
+
+	renesas_sdhi_internal_dmac_enable_dma(host, false);
+
+	renesas_sdhi_internal_dmac_dm_write(host, DM_CM_RST,
+					    RST_RESERVED_BITS & ~val);
+	renesas_sdhi_internal_dmac_dm_write(host, DM_CM_RST,
+					    RST_RESERVED_BITS | val);
+
+	renesas_sdhi_internal_dmac_enable_dma(host, true);
+}
+
+static void
+renesas_sdhi_internal_dmac_complete_dma(struct tmio_mmc_host *host) {
+	tasklet_schedule(&host->dma_complete);
+}
+
+static void
+renesas_sdhi_internal_dmac_start_dma(struct tmio_mmc_host *host,
+				     struct mmc_data *data)
+{
+	struct scatterlist *sg = host->sg_ptr;
+	u32 dtran_mode = DTRAN_MODE_BUS_WID_TH | DTRAN_MODE_ADDR_MODE;
+	enum dma_data_direction dir;
+	int ret;
+	u32 irq_mask;
+
+	/* This DMAC cannot handle if sg_len is not 1 */
+	WARN_ON(host->sg_len > 1);
+
+	/* This DMAC cannot handle if buffer is not 8-bytes alignment */
+	if (!IS_ALIGNED(sg->offset, 8)) {
+		host->force_pio = true;
+		renesas_sdhi_internal_dmac_enable_dma(host, false);
+		return;
+	}
+
+	if (data->flags & MMC_DATA_READ) {
+		dtran_mode |= DTRAN_MODE_CH_NUM_CH1;
+		dir = DMA_FROM_DEVICE;
+		irq_mask = TMIO_STAT_RXRDY;
+	} else {
+		dtran_mode |= DTRAN_MODE_CH_NUM_CH0;
+		dir = DMA_TO_DEVICE;
+		irq_mask = TMIO_STAT_TXRQ;
+	}
+
+	ret = dma_map_sg(&host->pdev->dev, sg, host->sg_len, dir);
+	if (ret < 0) {
+		dev_err(&host->pdev->dev, "%s: dma_map_sg failed\n", __func__);
+		return;
+	}
+
+	renesas_sdhi_internal_dmac_enable_dma(host, true);
+
+	/* disable PIO irqs to avoid "PIO IRQ in DMA mode!" */
+	tmio_mmc_disable_mmc_irqs(host, irq_mask);
+
+	/* set dma parameters */
+	renesas_sdhi_internal_dmac_dm_write(host, DM_CM_DTRAN_MODE,
+					    dtran_mode);
+	renesas_sdhi_internal_dmac_dm_write(host, DM_DTRAN_ADDR,
+					    sg->dma_address);
+}
+
+static void renesas_sdhi_internal_dmac_issue_tasklet_fn(unsigned long arg)
+{
+	struct tmio_mmc_host *host = (struct tmio_mmc_host *)arg;
+
+	tmio_mmc_enable_mmc_irqs(host, TMIO_STAT_DATAEND);
+
+	/* start the DMAC */
+	renesas_sdhi_internal_dmac_dm_write(host, DM_CM_DTRAN_CTRL,
+					    DTRAN_CTRL_DM_START);
+}
+
+static void renesas_sdhi_internal_dmac_complete_tasklet_fn(unsigned long arg)
+{
+	struct tmio_mmc_host *host = (struct tmio_mmc_host *)arg;
+	enum dma_data_direction dir;
+
+	spin_lock_irq(&host->lock);
+
+	if (!host->data)
+		goto out;
+
+	if (host->data->flags & MMC_DATA_READ)
+		dir = DMA_FROM_DEVICE;
+	else
+		dir = DMA_TO_DEVICE;
+
+	renesas_sdhi_internal_dmac_enable_dma(host, false);
+	dma_unmap_sg(&host->pdev->dev, host->sg_ptr, host->sg_len, dir);
+
+	spin_unlock_irq(&host->lock);
+
+	spin_lock_irq(&host->lock);
+	tmio_mmc_do_data_irq(host);
+out:
+	spin_unlock_irq(&host->lock);
+}
+
+static void
+renesas_sdhi_internal_dmac_request_dma(struct tmio_mmc_host *host,
+				       struct tmio_mmc_data *pdata)
+{
+	/* Each value is set to non-zero to assume "enabling" each DMA */
+	host->chan_rx = host->chan_tx = (void *)0xdeadbeaf;
+
+	tasklet_init(&host->dma_complete,
+		     renesas_sdhi_internal_dmac_complete_tasklet_fn,
+		     (unsigned long)host);
+	tasklet_init(&host->dma_issue,
+		     renesas_sdhi_internal_dmac_issue_tasklet_fn,
+		     (unsigned long)host);
+}
+
+static void
+renesas_sdhi_internal_dmac_release_dma(struct tmio_mmc_host *host)
+{
+	/* Each value is set to zero to assume "disabling" each DMA */
+	host->chan_rx = host->chan_tx = NULL;
+}
+
+static struct tmio_mmc_dma_ops renesas_sdhi_internal_dmac_dma_ops = {
+	.start = renesas_sdhi_internal_dmac_start_dma,
+	.enable = renesas_sdhi_internal_dmac_enable_dma,
+	.request = renesas_sdhi_internal_dmac_request_dma,
+	.release = renesas_sdhi_internal_dmac_release_dma,
+	.abort = renesas_sdhi_internal_dmac_abort_dma,
+	.complete = renesas_sdhi_internal_dmac_complete_dma,
+};
+
+static int renesas_sdhi_internal_dmac_probe(struct platform_device *pdev)
+{
+	return renesas_sdhi_probe(pdev, &renesas_sdhi_internal_dmac_dma_ops);
+}
+
+static const struct dev_pm_ops renesas_sdhi_internal_dmac_dev_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(pm_runtime_force_suspend,
+				pm_runtime_force_resume)
+	SET_RUNTIME_PM_OPS(tmio_mmc_host_runtime_suspend,
+			   tmio_mmc_host_runtime_resume,
+			   NULL)
+};
+
+static struct platform_driver renesas_internal_dmac_sdhi_driver = {
+	.driver		= {
+		.name	= "renesas_sdhi_internal_dmac",
+		.pm	= &renesas_sdhi_internal_dmac_dev_pm_ops,
+		.of_match_table = renesas_sdhi_internal_dmac_of_match,
+	},
+	.probe		= renesas_sdhi_internal_dmac_probe,
+	.remove		= renesas_sdhi_remove,
+};
+
+module_platform_driver(renesas_internal_dmac_sdhi_driver);
+
+MODULE_DESCRIPTION("Renesas SDHI driver for internal DMAC");
+MODULE_AUTHOR("Yoshihiro Shimoda");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/mmc/host/renesas_sdhi_sys_dmac.c b/drivers/mmc/host/renesas_sdhi_sys_dmac.c
index 7eab55adc5b2..cc3e0f94b276 100644
--- a/drivers/mmc/host/renesas_sdhi_sys_dmac.c
+++ b/drivers/mmc/host/renesas_sdhi_sys_dmac.c
@@ -1,5 +1,5 @@
 /*
- * DMA function for TMIO MMC implementations
+ * DMA support use of SYS DMAC with SDHI SD/SDIO controller
  *
  * Copyright (C) 2010-2011 Guennadi Liakhovetski
  *
diff --git a/drivers/mmc/host/tmio_mmc.h b/drivers/mmc/host/tmio_mmc.h
index 82725f9de359..3b08c76a55e7 100644
--- a/drivers/mmc/host/tmio_mmc.h
+++ b/drivers/mmc/host/tmio_mmc.h
@@ -151,6 +151,7 @@ struct tmio_mmc_host {
 	struct dma_chan		*chan_rx;
 	struct dma_chan		*chan_tx;
 	struct completion	dma_dataend;
+	struct tasklet_struct	dma_complete;
 	struct tasklet_struct	dma_issue;
 	struct scatterlist	bounce_sg;
 	u8			*bounce_buf;
-- 
2.1.4

[PATCH/RFC 5/5] mmc: renesas-sdhi: remove gen3 support from sysc dmac driver

[Simon Horman]

Gen3 SoCs are now supported by the internal dmac variant of the SDHI driver.
Remove them from the sys dmac variant where only PIO mode is supported.

Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
---
 drivers/mmc/host/renesas_sdhi_sys_dmac.c | 21 ---------------------
 1 file changed, 21 deletions(-)

diff --git a/drivers/mmc/host/renesas_sdhi_sys_dmac.c b/drivers/mmc/host/renesas_sdhi_sys_dmac.c
index cc3e0f94b276..1f8863e725aa 100644
--- a/drivers/mmc/host/renesas_sdhi_sys_dmac.c
+++ b/drivers/mmc/host/renesas_sdhi_sys_dmac.c
@@ -63,25 +63,6 @@ static const struct renesas_sdhi_of_data of_rcar_gen2_compatible = {
 	.taps_num	= ARRAY_SIZE(rcar_gen2_scc_taps),
 };
 
-/* Definitions for sampling clocks */
-static struct renesas_sdhi_scc rcar_gen3_scc_taps[] = {
-	{
-		.clk_rate = 0,
-		.tap = 0x00000300,
-	},
-};
-
-static const struct renesas_sdhi_of_data of_rcar_gen3_compatible = {
-	.tmio_flags	= TMIO_MMC_HAS_IDLE_WAIT | TMIO_MMC_WRPROTECT_DISABLE |
-			  TMIO_MMC_CLK_ACTUAL | TMIO_MMC_MIN_RCAR2,
-	.capabilities	= MMC_CAP_SD_HIGHSPEED | MMC_CAP_SDIO_IRQ |
-			  MMC_CAP_CMD23,
-	.bus_shift	= 2,
-	.scc_offset	= 0x1000,
-	.taps		= rcar_gen3_scc_taps,
-	.taps_num	= ARRAY_SIZE(rcar_gen3_scc_taps),
-};
-
 static const struct of_device_id renesas_sdhi_sys_dmac_of_match[] = {
 	{ .compatible = "renesas,sdhi-shmobile" },
 	{ .compatible = "renesas,sdhi-sh73a0", .data = &of_default_cfg, },
@@ -95,8 +76,6 @@ static const struct of_device_id renesas_sdhi_sys_dmac_of_match[] = {
 	{ .compatible = "renesas,sdhi-r8a7792", .data = &of_rcar_gen2_compatible, },
 	{ .compatible = "renesas,sdhi-r8a7793", .data = &of_rcar_gen2_compatible, },
 	{ .compatible = "renesas,sdhi-r8a7794", .data = &of_rcar_gen2_compatible, },
-	{ .compatible = "renesas,sdhi-r8a7795", .data = &of_rcar_gen3_compatible, },
-	{ .compatible = "renesas,sdhi-r8a7796", .data = &of_rcar_gen3_compatible, },
 	{},
 };
 MODULE_DEVICE_TABLE(of, renesas_sdhi_sys_dmac_of_match);
-- 
2.1.4

Re: [PATCH/RFC 0/5] mmc: renesas_sdhi: add support for R-Car Gen3 SDHI DMAC

[Simon Horman]

On Thu, Jun 08, 2017 at 03:09:21PM +0200, Simon Horman wrote:
> Hi,
> 
> this series adds support for the internal DMAC used by r8a779[56] SoCs.
> This is achieved by adding a new variant of the SDHI driver for this
> DMA controller with compat strings for the r8a779[56] SoCs.
> Compat strings for these SoCs are also removed from the existing SYS DMAC
> variant of the SDHI driver.

Sorry, I managed to post duplicate patches in this series.
I will repost.

Re: [PATCH/RFC 3/5] mmc: sh_mobile_sdhi: add some SoC specific data for R-Car Gen3

[Geert Uytterhoeven]

Hi Simon,

On Thu, Jun 8, 2017 at 3:09 PM, Simon Horman <horms+renesas@verge.net.au> wrote:
> To avoid complicate code in own dma code for gen3, this patch
> adds set max_segs to 1. Then, the tmio driver will get requests
> as sg_len = 1 only.

The above doesn't seem to match with the actual code changes?

> In the performance point of view, the CONFIG_MMC_BLOCK_BOUNCE should
> be set. Otherwise, mmc block layer will set 4-kbyte each as a request.

FWIW, CONFIG_MMC_BLOCK_BOUNCE is no more in -next.

> Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
> Signed-off-by: Ai Kyuse <ai.kyuse.uw@renesas.com>
> Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
> ---
>  drivers/mmc/host/renesas_sdhi.h      | 2 ++
>  drivers/mmc/host/renesas_sdhi_core.c | 2 ++
>  2 files changed, 4 insertions(+)
>
> diff --git a/drivers/mmc/host/renesas_sdhi.h b/drivers/mmc/host/renesas_sdhi.h
> index eb3ea15ff92d..6f7233fe32a2 100644
> --- a/drivers/mmc/host/renesas_sdhi.h
> +++ b/drivers/mmc/host/renesas_sdhi.h
> @@ -31,6 +31,8 @@ struct renesas_sdhi_of_data {
>         int scc_offset;
>         struct renesas_sdhi_scc *taps;
>         int taps_num;
> +       unsigned int max_blk_count;
> +       unsigned short max_segs;
>  };
>
>  int renesas_sdhi_probe(struct platform_device *pdev,
> diff --git a/drivers/mmc/host/renesas_sdhi_core.c b/drivers/mmc/host/renesas_sdhi_core.c
> index 82150a966391..294bc9705c5f 100644
> --- a/drivers/mmc/host/renesas_sdhi_core.c
> +++ b/drivers/mmc/host/renesas_sdhi_core.c
> @@ -522,6 +522,8 @@ int renesas_sdhi_probe(struct platform_device *pdev,
>                 mmc_data->capabilities |= of_data->capabilities;
>                 mmc_data->capabilities2 |= of_data->capabilities2;
>                 mmc_data->dma_rx_offset = of_data->dma_rx_offset;
> +               mmc_data->max_blk_count = of_data->max_blk_count;
> +               mmc_data->max_segs = of_data->max_segs;

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: [PATCH/RFC 2/5] mmc: tmio: add complete to DMA ops

[Geert Uytterhoeven]

Hi Simon,

On Thu, Jun 8, 2017 at 3:09 PM, Simon Horman <horms+renesas@verge.net.au> wrote:
> Add complete to DMA ops to allow DMAC implementation dependent
> handling of DMA completion.
>
> Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
> ---
>  drivers/mmc/host/tmio_mmc.h      |  1 +
>  drivers/mmc/host/tmio_mmc_core.c | 10 ++++++++--
>  2 files changed, 9 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/mmc/host/tmio_mmc.h b/drivers/mmc/host/tmio_mmc.h
> index 9c94b6eb9b49..82725f9de359 100644
> --- a/drivers/mmc/host/tmio_mmc.h
> +++ b/drivers/mmc/host/tmio_mmc.h
> @@ -121,6 +121,7 @@ struct tmio_mmc_dma_ops {
>                         struct tmio_mmc_data *pdata);
>         void (*release)(struct tmio_mmc_host *host);
>         void (*abort)(struct tmio_mmc_host *host);
> +       void (*complete)(struct tmio_mmc_host *host);
>  };
>
>  struct tmio_mmc_host {
> diff --git a/drivers/mmc/host/tmio_mmc_core.c b/drivers/mmc/host/tmio_mmc_core.c
> index 0e76e099dc7f..26fe9bb64849 100644
> --- a/drivers/mmc/host/tmio_mmc_core.c
> +++ b/drivers/mmc/host/tmio_mmc_core.c
> @@ -86,6 +86,12 @@ static inline void tmio_mmc_abort_dma(struct tmio_mmc_host *host)
>                 host->dma_ops->abort(host);
>  }
>
> +static inline void tmio_mmc_complete_dma(struct tmio_mmc_host *host)
> +{
> +       if (host->dma_ops)
> +               host->dma_ops->complete(host);

I believe this will cause a crash and thus break bisection, as no driver
fills in the complete pointer yet.

> +}
> +
>  void tmio_mmc_enable_mmc_irqs(struct tmio_mmc_host *host, u32 i)
>  {
>         host->sdcard_irq_mask &= ~(i & TMIO_MASK_IRQ);
> @@ -604,11 +610,11 @@ static void tmio_mmc_data_irq(struct tmio_mmc_host *host, unsigned int stat)
>
>                 if (done) {
>                         tmio_mmc_disable_mmc_irqs(host, TMIO_STAT_DATAEND);
> -                       complete(&host->dma_dataend);
> +                       tmio_mmc_complete_dma(host);

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: [PATCH/RFC 5/5] mmc: renesas-sdhi: remove gen3 support from sysc dmac driver

[Geert Uytterhoeven]

Hi Simon,

On Thu, Jun 8, 2017 at 3:09 PM, Simon Horman <horms+renesas@verge.net.au> wrote:
> Gen3 SoCs are now supported by the internal dmac variant of the SDHI driver.
> Remove them from the sys dmac variant where only PIO mode is supported.
>
> Signed-off-by: Simon Horman <horms+renesas@verge.net.au>

Is it possible to fall back to PIO with the new DMAC driver?
If not, is this desirable?

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: [PATCH/RFC 5/5] mmc: renesas-sdhi: remove gen3 support from sysc dmac driver

[Magnus Damm]

Hi Geert and Simon,

On Fri, Jun 9, 2017 at 4:40 PM, Geert Uytterhoeven <geert@linux-m68k.org> wrote:
> Hi Simon,
>
> On Thu, Jun 8, 2017 at 3:09 PM, Simon Horman <horms+renesas@verge.net.au> wrote:
>> Gen3 SoCs are now supported by the internal dmac variant of the SDHI driver.
>> Remove them from the sys dmac variant where only PIO mode is supported.
>>
>> Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
>
> Is it possible to fall back to PIO with the new DMAC driver?
> If not, is this desirable?

I believe R-Car Gen3 SDHI may depending on SoC use either on-chip bus
mastering DMA or SYS-DMAC via DMA Engine. In both cases fall back to
PIO is most likely valid. For any given SoC which is better of on-chip
DMA or SYS-DMAC needs to be investigated together with IPMMU.

Cheers,

/ magnus

Re: [PATCH/RFC 2/5] mmc: tmio: add complete to DMA ops

[Simon Horman]

On Fri, Jun 09, 2017 at 09:35:20AM +0200, Geert Uytterhoeven wrote:
> Hi Simon,
> 
> On Thu, Jun 8, 2017 at 3:09 PM, Simon Horman <horms+renesas@verge.net.au> wrote:
> > Add complete to DMA ops to allow DMAC implementation dependent
> > handling of DMA completion.
> >
> > Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
> > ---
> >  drivers/mmc/host/tmio_mmc.h      |  1 +
> >  drivers/mmc/host/tmio_mmc_core.c | 10 ++++++++--
> >  2 files changed, 9 insertions(+), 2 deletions(-)
> >
> > diff --git a/drivers/mmc/host/tmio_mmc.h b/drivers/mmc/host/tmio_mmc.h
> > index 9c94b6eb9b49..82725f9de359 100644
> > --- a/drivers/mmc/host/tmio_mmc.h
> > +++ b/drivers/mmc/host/tmio_mmc.h
> > @@ -121,6 +121,7 @@ struct tmio_mmc_dma_ops {
> >                         struct tmio_mmc_data *pdata);
> >         void (*release)(struct tmio_mmc_host *host);
> >         void (*abort)(struct tmio_mmc_host *host);
> > +       void (*complete)(struct tmio_mmc_host *host);
> >  };
> >
> >  struct tmio_mmc_host {
> > diff --git a/drivers/mmc/host/tmio_mmc_core.c b/drivers/mmc/host/tmio_mmc_core.c
> > index 0e76e099dc7f..26fe9bb64849 100644
> > --- a/drivers/mmc/host/tmio_mmc_core.c
> > +++ b/drivers/mmc/host/tmio_mmc_core.c
> > @@ -86,6 +86,12 @@ static inline void tmio_mmc_abort_dma(struct tmio_mmc_host *host)
> >                 host->dma_ops->abort(host);
> >  }
> >
> > +static inline void tmio_mmc_complete_dma(struct tmio_mmc_host *host)
> > +{
> > +       if (host->dma_ops)
> > +               host->dma_ops->complete(host);
> 
> I believe this will cause a crash and thus break bisection, as no driver
> fills in the complete pointer yet.

Indeed. And it seems to cause a panic on non-Gen3 users of the driver.

I think the fix should be:

diff --git a/drivers/mmc/host/tmio_mmc_core.c b/drivers/mmc/host/tmio_mmc_core.c
index 26fe9bb64849..9ec2e933b6f0 100644
--- a/drivers/mmc/host/tmio_mmc_core.c
+++ b/drivers/mmc/host/tmio_mmc_core.c
@@ -88,7 +88,7 @@ static inline void tmio_mmc_abort_dma(struct tmio_mmc_host *host)
 
 static inline void tmio_mmc_complete_dma(struct tmio_mmc_host *host)
 {
-	if (host->dma_ops)
+	if (host->dma_ops && host->dma_ops->complete)
 		host->dma_ops->complete(host);
 }
 

> 
> > +}
> > +
> >  void tmio_mmc_enable_mmc_irqs(struct tmio_mmc_host *host, u32 i)
> >  {
> >         host->sdcard_irq_mask &= ~(i & TMIO_MASK_IRQ);
> > @@ -604,11 +610,11 @@ static void tmio_mmc_data_irq(struct tmio_mmc_host *host, unsigned int stat)
> >
> >                 if (done) {
> >                         tmio_mmc_disable_mmc_irqs(host, TMIO_STAT_DATAEND);
> > -                       complete(&host->dma_dataend);
> > +                       tmio_mmc_complete_dma(host);
> 
> Gr{oetje,eeting}s,
> 
>                         Geert
> 
> --
> Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org
> 
> In personal conversations with technical people, I call myself a hacker. But
> when I'm talking to journalists I just say "programmer" or something like that.
>                                 -- Linus Torvalds
> 

Re: [PATCH/RFC 2/5] mmc: tmio: add complete to DMA ops

[Geert Uytterhoeven]

Hi Simon,

On Fri, Jun 16, 2017 at 8:06 AM, Simon Horman <horms@verge.net.au> wrote:
> On Fri, Jun 09, 2017 at 09:35:20AM +0200, Geert Uytterhoeven wrote:
>> On Thu, Jun 8, 2017 at 3:09 PM, Simon Horman <horms+renesas@verge.net.au> wrote:
>> > Add complete to DMA ops to allow DMAC implementation dependent
>> > handling of DMA completion.
>> >
>> > Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
>> > ---
>> >  drivers/mmc/host/tmio_mmc.h      |  1 +
>> >  drivers/mmc/host/tmio_mmc_core.c | 10 ++++++++--
>> >  2 files changed, 9 insertions(+), 2 deletions(-)
>> >
>> > diff --git a/drivers/mmc/host/tmio_mmc.h b/drivers/mmc/host/tmio_mmc.h
>> > index 9c94b6eb9b49..82725f9de359 100644
>> > --- a/drivers/mmc/host/tmio_mmc.h
>> > +++ b/drivers/mmc/host/tmio_mmc.h
>> > @@ -121,6 +121,7 @@ struct tmio_mmc_dma_ops {
>> >                         struct tmio_mmc_data *pdata);
>> >         void (*release)(struct tmio_mmc_host *host);
>> >         void (*abort)(struct tmio_mmc_host *host);
>> > +       void (*complete)(struct tmio_mmc_host *host);
>> >  };
>> >
>> >  struct tmio_mmc_host {
>> > diff --git a/drivers/mmc/host/tmio_mmc_core.c b/drivers/mmc/host/tmio_mmc_core.c
>> > index 0e76e099dc7f..26fe9bb64849 100644
>> > --- a/drivers/mmc/host/tmio_mmc_core.c
>> > +++ b/drivers/mmc/host/tmio_mmc_core.c
>> > @@ -86,6 +86,12 @@ static inline void tmio_mmc_abort_dma(struct tmio_mmc_host *host)
>> >                 host->dma_ops->abort(host);
>> >  }
>> >
>> > +static inline void tmio_mmc_complete_dma(struct tmio_mmc_host *host)
>> > +{
>> > +       if (host->dma_ops)
>> > +               host->dma_ops->complete(host);
>>
>> I believe this will cause a crash and thus break bisection, as no driver
>> fills in the complete pointer yet.
>
> Indeed. And it seems to cause a panic on non-Gen3 users of the driver.
>
> I think the fix should be:
>
> diff --git a/drivers/mmc/host/tmio_mmc_core.c b/drivers/mmc/host/tmio_mmc_core.c
> index 26fe9bb64849..9ec2e933b6f0 100644
> --- a/drivers/mmc/host/tmio_mmc_core.c
> +++ b/drivers/mmc/host/tmio_mmc_core.c
> @@ -88,7 +88,7 @@ static inline void tmio_mmc_abort_dma(struct tmio_mmc_host *host)
>
>  static inline void tmio_mmc_complete_dma(struct tmio_mmc_host *host)
>  {
> -       if (host->dma_ops)
> +       if (host->dma_ops && host->dma_ops->complete)
>                 host->dma_ops->complete(host);
>  }

Is it? Now nothing will be completed?

The old code did:

     complete(&host->dma_dataend);

so perhaps

    else
            complete(&host->dma_dataend);

with the "else" branch to be deleted when all drivers that need it provide
their own complete callback?

>> > +}
>> > +
>> >  void tmio_mmc_enable_mmc_irqs(struct tmio_mmc_host *host, u32 i)
>> >  {
>> >         host->sdcard_irq_mask &= ~(i & TMIO_MASK_IRQ);
>> > @@ -604,11 +610,11 @@ static void tmio_mmc_data_irq(struct tmio_mmc_host *host, unsigned int stat)
>> >
>> >                 if (done) {
>> >                         tmio_mmc_disable_mmc_irqs(host, TMIO_STAT_DATAEND);
>> > -                       complete(&host->dma_dataend);
>> > +                       tmio_mmc_complete_dma(host);

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: [PATCH/RFC 5/5] mmc: renesas-sdhi: remove gen3 support from sysc dmac driver

[Simon Horman]

On Fri, Jun 09, 2017 at 05:40:15PM +0900, Magnus Damm wrote:
> Hi Geert and Simon,
> 
> On Fri, Jun 9, 2017 at 4:40 PM, Geert Uytterhoeven <geert@linux-m68k.org> wrote:
> > Hi Simon,
> >
> > On Thu, Jun 8, 2017 at 3:09 PM, Simon Horman <horms+renesas@verge.net.au> wrote:
> >> Gen3 SoCs are now supported by the internal dmac variant of the SDHI driver.
> >> Remove them from the sys dmac variant where only PIO mode is supported.
> >>
> >> Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
> >
> > Is it possible to fall back to PIO with the new DMAC driver?
> > If not, is this desirable?
> 
> I believe R-Car Gen3 SDHI may depending on SoC use either on-chip bus
> mastering DMA or SYS-DMAC via DMA Engine. In both cases fall back to
> PIO is most likely valid. For any given SoC which is better of on-chip
> DMA or SYS-DMAC needs to be investigated together with IPMMU.

My answer is a bit different:

With this patchset there are two variants of the SDHI driver, one that
supports SYS-DMAC via DMA Engine and one that supports on-chip bus
mastering DMA.

My understanding is that prior to this patch set the SYS-DMAC and only
variant of the SDHI driver supported fallback to PIO. And as the internal
DMAC variant uses the same core code my assumption is that it should also
have that property.  I would need to investigate things in more detail to
confirm that assumption.

In terms of supporting Gen3 with SYS-DMAC or on-chip bus mastering DMA,
I believe those cases should use compat strings and soc-match in each of
the two variants of the SDHI driver as appropriate.

Re: [PATCH/RFC 2/5] mmc: tmio: add complete to DMA ops

[Simon Horman]

On Fri, Jun 16, 2017 at 08:25:12AM +0200, Geert Uytterhoeven wrote:
> Hi Simon,
> 
> On Fri, Jun 16, 2017 at 8:06 AM, Simon Horman <horms@verge.net.au> wrote:
> > On Fri, Jun 09, 2017 at 09:35:20AM +0200, Geert Uytterhoeven wrote:
> >> On Thu, Jun 8, 2017 at 3:09 PM, Simon Horman <horms+renesas@verge.net.au> wrote:
> >> > Add complete to DMA ops to allow DMAC implementation dependent
> >> > handling of DMA completion.
> >> >
> >> > Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
> >> > ---
> >> >  drivers/mmc/host/tmio_mmc.h      |  1 +
> >> >  drivers/mmc/host/tmio_mmc_core.c | 10 ++++++++--
> >> >  2 files changed, 9 insertions(+), 2 deletions(-)
> >> >
> >> > diff --git a/drivers/mmc/host/tmio_mmc.h b/drivers/mmc/host/tmio_mmc.h
> >> > index 9c94b6eb9b49..82725f9de359 100644
> >> > --- a/drivers/mmc/host/tmio_mmc.h
> >> > +++ b/drivers/mmc/host/tmio_mmc.h
> >> > @@ -121,6 +121,7 @@ struct tmio_mmc_dma_ops {
> >> >                         struct tmio_mmc_data *pdata);
> >> >         void (*release)(struct tmio_mmc_host *host);
> >> >         void (*abort)(struct tmio_mmc_host *host);
> >> > +       void (*complete)(struct tmio_mmc_host *host);
> >> >  };
> >> >
> >> >  struct tmio_mmc_host {
> >> > diff --git a/drivers/mmc/host/tmio_mmc_core.c b/drivers/mmc/host/tmio_mmc_core.c
> >> > index 0e76e099dc7f..26fe9bb64849 100644
> >> > --- a/drivers/mmc/host/tmio_mmc_core.c
> >> > +++ b/drivers/mmc/host/tmio_mmc_core.c
> >> > @@ -86,6 +86,12 @@ static inline void tmio_mmc_abort_dma(struct tmio_mmc_host *host)
> >> >                 host->dma_ops->abort(host);
> >> >  }
> >> >
> >> > +static inline void tmio_mmc_complete_dma(struct tmio_mmc_host *host)
> >> > +{
> >> > +       if (host->dma_ops)
> >> > +               host->dma_ops->complete(host);
> >>
> >> I believe this will cause a crash and thus break bisection, as no driver
> >> fills in the complete pointer yet.
> >
> > Indeed. And it seems to cause a panic on non-Gen3 users of the driver.
> >
> > I think the fix should be:
> >
> > diff --git a/drivers/mmc/host/tmio_mmc_core.c b/drivers/mmc/host/tmio_mmc_core.c
> > index 26fe9bb64849..9ec2e933b6f0 100644
> > --- a/drivers/mmc/host/tmio_mmc_core.c
> > +++ b/drivers/mmc/host/tmio_mmc_core.c
> > @@ -88,7 +88,7 @@ static inline void tmio_mmc_abort_dma(struct tmio_mmc_host *host)
> >
> >  static inline void tmio_mmc_complete_dma(struct tmio_mmc_host *host)
> >  {
> > -       if (host->dma_ops)
> > +       if (host->dma_ops && host->dma_ops->complete)
> >                 host->dma_ops->complete(host);
> >  }
> 
> Is it? Now nothing will be completed?
> 
> The old code did:
> 
>      complete(&host->dma_dataend);
> 
> so perhaps
> 
>     else
>             complete(&host->dma_dataend);
> 
> with the "else" branch to be deleted when all drivers that need it provide
> their own complete callback?

Thanks, I see that is still a problem.

As there is only one other driver and its closerly related
perhaps it could just implement the callback in this patch?

> >> > +}
> >> > +
> >> >  void tmio_mmc_enable_mmc_irqs(struct tmio_mmc_host *host, u32 i)
> >> >  {
> >> >         host->sdcard_irq_mask &= ~(i & TMIO_MASK_IRQ);
> >> > @@ -604,11 +610,11 @@ static void tmio_mmc_data_irq(struct tmio_mmc_host *host, unsigned int stat)
> >> >
> >> >                 if (done) {
> >> >                         tmio_mmc_disable_mmc_irqs(host, TMIO_STAT_DATAEND);
> >> > -                       complete(&host->dma_dataend);
> >> > +                       tmio_mmc_complete_dma(host);
> 
> Gr{oetje,eeting}s,
> 
>                         Geert
> 
> --
> Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org
> 
> In personal conversations with technical people, I call myself a hacker. But
> when I'm talking to journalists I just say "programmer" or something like that.
>                                 -- Linus Torvalds
> 

Re: [PATCH/RFC 2/5] mmc: tmio: add complete to DMA ops

[Geert Uytterhoeven]

Hi Simon,

On Fri, Jun 16, 2017 at 9:04 AM, Simon Horman <horms@verge.net.au> wrote:
> On Fri, Jun 16, 2017 at 08:25:12AM +0200, Geert Uytterhoeven wrote:
>> On Fri, Jun 16, 2017 at 8:06 AM, Simon Horman <horms@verge.net.au> wrote:
>> > On Fri, Jun 09, 2017 at 09:35:20AM +0200, Geert Uytterhoeven wrote:
>> >> On Thu, Jun 8, 2017 at 3:09 PM, Simon Horman <horms+renesas@verge.net.au> wrote:
>> >> > Add complete to DMA ops to allow DMAC implementation dependent
>> >> > handling of DMA completion.
>> >> >
>> >> > Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
>> >> > ---
>> >> >  drivers/mmc/host/tmio_mmc.h      |  1 +
>> >> >  drivers/mmc/host/tmio_mmc_core.c | 10 ++++++++--
>> >> >  2 files changed, 9 insertions(+), 2 deletions(-)
>> >> >
>> >> > diff --git a/drivers/mmc/host/tmio_mmc.h b/drivers/mmc/host/tmio_mmc.h
>> >> > index 9c94b6eb9b49..82725f9de359 100644
>> >> > --- a/drivers/mmc/host/tmio_mmc.h
>> >> > +++ b/drivers/mmc/host/tmio_mmc.h
>> >> > @@ -121,6 +121,7 @@ struct tmio_mmc_dma_ops {
>> >> >                         struct tmio_mmc_data *pdata);
>> >> >         void (*release)(struct tmio_mmc_host *host);
>> >> >         void (*abort)(struct tmio_mmc_host *host);
>> >> > +       void (*complete)(struct tmio_mmc_host *host);
>> >> >  };
>> >> >
>> >> >  struct tmio_mmc_host {
>> >> > diff --git a/drivers/mmc/host/tmio_mmc_core.c b/drivers/mmc/host/tmio_mmc_core.c
>> >> > index 0e76e099dc7f..26fe9bb64849 100644
>> >> > --- a/drivers/mmc/host/tmio_mmc_core.c
>> >> > +++ b/drivers/mmc/host/tmio_mmc_core.c
>> >> > @@ -86,6 +86,12 @@ static inline void tmio_mmc_abort_dma(struct tmio_mmc_host *host)
>> >> >                 host->dma_ops->abort(host);
>> >> >  }
>> >> >
>> >> > +static inline void tmio_mmc_complete_dma(struct tmio_mmc_host *host)
>> >> > +{
>> >> > +       if (host->dma_ops)
>> >> > +               host->dma_ops->complete(host);
>> >>
>> >> I believe this will cause a crash and thus break bisection, as no driver
>> >> fills in the complete pointer yet.
>> >
>> > Indeed. And it seems to cause a panic on non-Gen3 users of the driver.
>> >
>> > I think the fix should be:
>> >
>> > diff --git a/drivers/mmc/host/tmio_mmc_core.c b/drivers/mmc/host/tmio_mmc_core.c
>> > index 26fe9bb64849..9ec2e933b6f0 100644
>> > --- a/drivers/mmc/host/tmio_mmc_core.c
>> > +++ b/drivers/mmc/host/tmio_mmc_core.c
>> > @@ -88,7 +88,7 @@ static inline void tmio_mmc_abort_dma(struct tmio_mmc_host *host)
>> >
>> >  static inline void tmio_mmc_complete_dma(struct tmio_mmc_host *host)
>> >  {
>> > -       if (host->dma_ops)
>> > +       if (host->dma_ops && host->dma_ops->complete)
>> >                 host->dma_ops->complete(host);
>> >  }
>>
>> Is it? Now nothing will be completed?
>>
>> The old code did:
>>
>>      complete(&host->dma_dataend);
>>
>> so perhaps
>>
>>     else
>>             complete(&host->dma_dataend);
>>
>> with the "else" branch to be deleted when all drivers that need it provide
>> their own complete callback?
>
> Thanks, I see that is still a problem.
>
> As there is only one other driver and its closerly related
> perhaps it could just implement the callback in this patch?

Yes, that sounds good to me.
Else you need 3 patches:
  1. Introduced .complete field,
  2. Update driver to provide .complete handledr,
  3. Update generic code.

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: [PATCH/RFC 2/5] mmc: tmio: add complete to DMA ops

[Simon Horman]

On Fri, Jun 16, 2017 at 09:09:31AM +0200, Geert Uytterhoeven wrote:
> Hi Simon,
> 
> On Fri, Jun 16, 2017 at 9:04 AM, Simon Horman <horms@verge.net.au> wrote:
> > On Fri, Jun 16, 2017 at 08:25:12AM +0200, Geert Uytterhoeven wrote:
> >> On Fri, Jun 16, 2017 at 8:06 AM, Simon Horman <horms@verge.net.au> wrote:
> >> > On Fri, Jun 09, 2017 at 09:35:20AM +0200, Geert Uytterhoeven wrote:
> >> >> On Thu, Jun 8, 2017 at 3:09 PM, Simon Horman <horms+renesas@verge.net.au> wrote:
> >> >> > Add complete to DMA ops to allow DMAC implementation dependent
> >> >> > handling of DMA completion.
> >> >> >
> >> >> > Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
> >> >> > ---
> >> >> >  drivers/mmc/host/tmio_mmc.h      |  1 +
> >> >> >  drivers/mmc/host/tmio_mmc_core.c | 10 ++++++++--
> >> >> >  2 files changed, 9 insertions(+), 2 deletions(-)
> >> >> >
> >> >> > diff --git a/drivers/mmc/host/tmio_mmc.h b/drivers/mmc/host/tmio_mmc.h
> >> >> > index 9c94b6eb9b49..82725f9de359 100644
> >> >> > --- a/drivers/mmc/host/tmio_mmc.h
> >> >> > +++ b/drivers/mmc/host/tmio_mmc.h
> >> >> > @@ -121,6 +121,7 @@ struct tmio_mmc_dma_ops {
> >> >> >                         struct tmio_mmc_data *pdata);
> >> >> >         void (*release)(struct tmio_mmc_host *host);
> >> >> >         void (*abort)(struct tmio_mmc_host *host);
> >> >> > +       void (*complete)(struct tmio_mmc_host *host);
> >> >> >  };
> >> >> >
> >> >> >  struct tmio_mmc_host {
> >> >> > diff --git a/drivers/mmc/host/tmio_mmc_core.c b/drivers/mmc/host/tmio_mmc_core.c
> >> >> > index 0e76e099dc7f..26fe9bb64849 100644
> >> >> > --- a/drivers/mmc/host/tmio_mmc_core.c
> >> >> > +++ b/drivers/mmc/host/tmio_mmc_core.c
> >> >> > @@ -86,6 +86,12 @@ static inline void tmio_mmc_abort_dma(struct tmio_mmc_host *host)
> >> >> >                 host->dma_ops->abort(host);
> >> >> >  }
> >> >> >
> >> >> > +static inline void tmio_mmc_complete_dma(struct tmio_mmc_host *host)
> >> >> > +{
> >> >> > +       if (host->dma_ops)
> >> >> > +               host->dma_ops->complete(host);
> >> >>
> >> >> I believe this will cause a crash and thus break bisection, as no driver
> >> >> fills in the complete pointer yet.
> >> >
> >> > Indeed. And it seems to cause a panic on non-Gen3 users of the driver.
> >> >
> >> > I think the fix should be:
> >> >
> >> > diff --git a/drivers/mmc/host/tmio_mmc_core.c b/drivers/mmc/host/tmio_mmc_core.c
> >> > index 26fe9bb64849..9ec2e933b6f0 100644
> >> > --- a/drivers/mmc/host/tmio_mmc_core.c
> >> > +++ b/drivers/mmc/host/tmio_mmc_core.c
> >> > @@ -88,7 +88,7 @@ static inline void tmio_mmc_abort_dma(struct tmio_mmc_host *host)
> >> >
> >> >  static inline void tmio_mmc_complete_dma(struct tmio_mmc_host *host)
> >> >  {
> >> > -       if (host->dma_ops)
> >> > +       if (host->dma_ops && host->dma_ops->complete)
> >> >                 host->dma_ops->complete(host);
> >> >  }
> >>
> >> Is it? Now nothing will be completed?
> >>
> >> The old code did:
> >>
> >>      complete(&host->dma_dataend);
> >>
> >> so perhaps
> >>
> >>     else
> >>             complete(&host->dma_dataend);
> >>
> >> with the "else" branch to be deleted when all drivers that need it provide
> >> their own complete callback?
> >
> > Thanks, I see that is still a problem.
> >
> > As there is only one other driver and its closerly related
> > perhaps it could just implement the callback in this patch?
> 
> Yes, that sounds good to me.
> Else you need 3 patches:
>   1. Introduced .complete field,
>   2. Update driver to provide .complete handledr,
>   3. Update generic code.

Yes, and 3 patch seems a bit excessive in this case.

Sorry for letting this bug creep in to renesas-drivers.
Do you want a patch against your tree? I will plan to post v2 of
this patchset soon in any case.

Re: [PATCH/RFC 2/5] mmc: tmio: add complete to DMA ops

[Geert Uytterhoeven]

Hi Simon,

On Fri, Jun 16, 2017 at 9:13 AM, Simon Horman <horms@verge.net.au> wrote:
> Sorry for letting this bug creep in to renesas-drivers.
> Do you want a patch against your tree? I will plan to post v2 of
> this patchset soon in any case.

You can post such a patch to please users of renesas-drivers, but I won't
apply it now. I assume the issue will be fixed when next renesas-drivers
release is created ;-)

Thanks!

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: [PATCH/RFC 2/5] mmc: tmio: add complete to DMA ops

[Simon Horman]

On Fri, Jun 16, 2017 at 09:23:15AM +0200, Geert Uytterhoeven wrote:
> Hi Simon,
> 
> On Fri, Jun 16, 2017 at 9:13 AM, Simon Horman <horms@verge.net.au> wrote:
> > Sorry for letting this bug creep in to renesas-drivers.
> > Do you want a patch against your tree? I will plan to post v2 of
> > this patchset soon in any case.
> 
> You can post such a patch to please users of renesas-drivers, but I won't
> apply it now. I assume the issue will be fixed when next renesas-drivers
> release is created ;-)

That is my assumption too :)

I'll see about providing a fix patch too.

Re: [PATCH/RFC 3/5] mmc: sh_mobile_sdhi: add some SoC specific data for R-Car Gen3

[Simon Horman]

On Fri, Jun 09, 2017 at 09:34:46AM +0200, Geert Uytterhoeven wrote:
> Hi Simon,
> 
> On Thu, Jun 8, 2017 at 3:09 PM, Simon Horman <horms+renesas@verge.net.au> wrote:
> > To avoid complicate code in own dma code for gen3, this patch
> > adds set max_segs to 1. Then, the tmio driver will get requests
> > as sg_len = 1 only.
> 
> The above doesn't seem to match with the actual code changes?
> 
> > In the performance point of view, the CONFIG_MMC_BLOCK_BOUNCE should
> > be set. Otherwise, mmc block layer will set 4-kbyte each as a request.
> 
> FWIW, CONFIG_MMC_BLOCK_BOUNCE is no more in -next.

Thanks.

The code has been shuffled around to the point where this patch doesn't
seem to make much sense as is. I squashed it into the previous patch like this:

mmc: tmio, renesas_sdhi: add max_segs and max_blk_count in tmio_mmc_data

Allow TMIO and SDHI driver implementations to provide values for
max_segs and max_blk_count.

A follow-up patch will set these values for Renesas Gen3 SoCs
the using an SDHI driver.

Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Signed-off-by: Ai Kyuse <ai.kyuse.uw@renesas.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>

diff --git a/drivers/mmc/host/renesas_sdhi.h b/drivers/mmc/host/renesas_sdhi.h
index eb3ea15ff92d..6f7233fe32a2 100644
--- a/drivers/mmc/host/renesas_sdhi.h
+++ b/drivers/mmc/host/renesas_sdhi.h
@@ -31,6 +31,8 @@ struct renesas_sdhi_of_data {
 	int scc_offset;
 	struct renesas_sdhi_scc *taps;
 	int taps_num;
+	unsigned int max_blk_count;
+	unsigned short max_segs;
 };
 
 int renesas_sdhi_probe(struct platform_device *pdev,
diff --git a/drivers/mmc/host/renesas_sdhi_core.c b/drivers/mmc/host/renesas_sdhi_core.c
index f4690cba3443..75c7bf4f7ca2 100644
--- a/drivers/mmc/host/renesas_sdhi_core.c
+++ b/drivers/mmc/host/renesas_sdhi_core.c
@@ -523,6 +523,8 @@ int renesas_sdhi_probe(struct platform_device *pdev,
 		mmc_data->capabilities |= of_data->capabilities;
 		mmc_data->capabilities2 |= of_data->capabilities2;
 		mmc_data->dma_rx_offset = of_data->dma_rx_offset;
+		mmc_data->max_blk_count = of_data->max_blk_count;
+		mmc_data->max_segs = of_data->max_segs;
 		dma_priv->dma_buswidth = of_data->dma_buswidth;
 		host->bus_shift = of_data->bus_shift;
 	}
diff --git a/drivers/mmc/host/tmio_mmc_core.c b/drivers/mmc/host/tmio_mmc_core.c
index fbe38464e7d7..7cb8d3510396 100644
--- a/drivers/mmc/host/tmio_mmc_core.c
+++ b/drivers/mmc/host/tmio_mmc_core.c
@@ -1248,10 +1248,10 @@ int tmio_mmc_host_probe(struct tmio_mmc_host *_host,
 
 	mmc->caps |= MMC_CAP_4_BIT_DATA | pdata->capabilities;
 	mmc->caps2 |= pdata->capabilities2;
-	mmc->max_segs = 32;
+	mmc->max_segs = pdata->max_segs ? : 32;
 	mmc->max_blk_size = 512;
-	mmc->max_blk_count = (PAGE_SIZE / mmc->max_blk_size) *
-		mmc->max_segs;
+	mmc->max_blk_count = pdata->max_blk_count ? :
+		(PAGE_SIZE / mmc->max_blk_size) * mmc->max_segs;
 	mmc->max_req_size = mmc->max_blk_size * mmc->max_blk_count;
 	mmc->max_seg_size = mmc->max_req_size;
 
diff --git a/include/linux/mfd/tmio.h b/include/linux/mfd/tmio.h
index c83c16b931a8..4e261ea94d6a 100644
--- a/include/linux/mfd/tmio.h
+++ b/include/linux/mfd/tmio.h
@@ -128,6 +128,8 @@ struct tmio_mmc_data {
 	unsigned int			cd_gpio;
 	int				alignment_shift;
 	dma_addr_t			dma_rx_offset;
+	unsigned int			max_blk_count;
+	unsigned short			max_segs;
 	void (*set_pwr)(struct platform_device *host, int state);
 	void (*set_clk_div)(struct platform_device *host, int state);
 };

Re: [PATCH/RFC 2/5] mmc: tmio: add complete to DMA ops

[Simon Horman]

On Fri, Jun 16, 2017 at 09:33:31AM +0200, Simon Horman wrote:
> On Fri, Jun 16, 2017 at 09:23:15AM +0200, Geert Uytterhoeven wrote:
> > Hi Simon,
> > 
> > On Fri, Jun 16, 2017 at 9:13 AM, Simon Horman <horms@verge.net.au> wrote:
> > > Sorry for letting this bug creep in to renesas-drivers.
> > > Do you want a patch against your tree? I will plan to post v2 of
> > > this patchset soon in any case.
> > 
> > You can post such a patch to please users of renesas-drivers, but I won't
> > apply it now. I assume the issue will be fixed when next renesas-drivers
> > release is created ;-)
> 
> That is my assumption too :)

Revised patches posted as "[PATCH mmc/next v2 0/4] mmc: renesas_sdhi: add
support for R-Car Gen3 SDHI DMAC"

Pushed to topic/sdhi-gen3-dma-2017-v2 in my renesas tree.
Do you need a fresh pull request?

> I'll see about providing a fix patch too.

Fix is:

From: Simon Horman <horms+renesas@verge.net.au>
Subject: [PATCH] mmc: renesas-sdhi: implement complete to DMA op

Implement complete DMA op to resolve regression introduced by
the introduction of that DMA op.

This resolves a panic when initialising SDHI on non Gen3 platforms
which use DMA with SDHI.

This is planned to be squashed into
"mmc: tmio, renesas-sdhi: add complete to DMA ops" for
submission to upstream.

Fixes: f6bcd9f5e392 ("mmc: tmio: add complete to DMA ops")
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
---
 drivers/mmc/host/renesas_sdhi_sys_dmac.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/drivers/mmc/host/renesas_sdhi_sys_dmac.c b/drivers/mmc/host/renesas_sdhi_sys_dmac.c
index 90f8d11c4868..2897ee0b444a 100644
--- a/drivers/mmc/host/renesas_sdhi_sys_dmac.c
+++ b/drivers/mmc/host/renesas_sdhi_sys_dmac.c
@@ -106,6 +106,11 @@ static void renesas_sdhi_sys_dmac_abort_dma(struct tmio_mmc_host *host)
 	renesas_sdhi_sys_dmac_enable_dma(host, true);
 }
 
+static void renesas_sdhi_sys_dmac_complete_dma(struct tmio_mmc_host *host)
+{
+	complete(&host->dma_dataend);
+}
+
 static void renesas_sdhi_sys_dmac_dma_callback(void *arg)
 {
 	struct tmio_mmc_host *host = arg;
@@ -427,6 +432,7 @@ static const struct tmio_mmc_dma_ops renesas_sdhi_sys_dmac_dma_ops = {
 	.request = renesas_sdhi_sys_dmac_request_dma,
 	.release = renesas_sdhi_sys_dmac_release_dma,
 	.abort = renesas_sdhi_sys_dmac_abort_dma,
+	.complete = renesas_sdhi_sys_dmac_complete_dma,
 };
 
 static int renesas_sdhi_sys_dmac_probe(struct platform_device *pdev)
-- 
2.1.4

Re: [PATCH/RFC 2/5] mmc: tmio: add complete to DMA ops

[Geert Uytterhoeven]

Hi Simon,

On Fri, Jun 16, 2017 at 12:03 PM, Simon Horman <horms@verge.net.au> wrote:
> On Fri, Jun 16, 2017 at 09:33:31AM +0200, Simon Horman wrote:
>> On Fri, Jun 16, 2017 at 09:23:15AM +0200, Geert Uytterhoeven wrote:
>> > Hi Simon,
>> >
>> > On Fri, Jun 16, 2017 at 9:13 AM, Simon Horman <horms@verge.net.au> wrote:
>> > > Sorry for letting this bug creep in to renesas-drivers.
>> > > Do you want a patch against your tree? I will plan to post v2 of
>> > > this patchset soon in any case.
>> >
>> > You can post such a patch to please users of renesas-drivers, but I won't
>> > apply it now. I assume the issue will be fixed when next renesas-drivers
>> > release is created ;-)
>>
>> That is my assumption too :)
>
> Revised patches posted as "[PATCH mmc/next v2 0/4] mmc: renesas_sdhi: add
> support for R-Car Gen3 SDHI DMAC"

Thanks!

> Pushed to topic/sdhi-gen3-dma-2017-v2 in my renesas tree.
> Do you need a fresh pull request?

Not needed, I already updated my list of topics to merge.

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds