This makes it possible for engines to implement slave transfers
to be done to/from a simple kmalloc():ed memory buffer and not
just from scatterlists.
Cc: Sascha Hauer <[email protected]>
Signed-off-by: Linus Walleij <[email protected]>
---
Sacha: this has similarities to you i.MX setup_single() operation,
if you think this suits your needs for i.MX consider ACKing this
patch pls. (Circular jobs would have a similar signature I believe.)
I will likely implement this API on COH901318 and DMA40 if it is
accepted.
---
arch/arm/mach-at91/at91sam9g45_devices.c | 2 +-
drivers/dma/at_hdmac.c | 8 ++++----
drivers/dma/coh901318.c | 2 +-
drivers/dma/dmaengine.c | 9 ++++++---
drivers/dma/dw_dmac.c | 2 +-
drivers/dma/fsldma.c | 11 ++++++-----
drivers/dma/ipu/ipu_idmac.c | 4 ++--
drivers/dma/pl330.c | 2 +-
drivers/dma/shdma.c | 4 ++--
drivers/dma/ste_dma40.c | 9 +++++----
drivers/dma/timb_dma.c | 2 +-
drivers/dma/txx9dmac.c | 5 +++--
drivers/media/video/mx3_camera.c | 2 +-
drivers/mmc/host/atmel-mci.c | 2 +-
drivers/mmc/host/tmio_mmc.c | 2 +-
drivers/serial/sh-sci.c | 2 +-
drivers/video/mx3fb.c | 2 +-
include/linux/dmaengine.h | 14 +++++++++++---
sound/atmel/abdac.c | 2 +-
sound/atmel/ac97c.c | 4 ++--
sound/soc/sh/siu_pcm.c | 2 +-
sound/soc/txx9/txx9aclc.c | 2 +-
22 files changed, 54 insertions(+), 40 deletions(-)
diff --git a/arch/arm/mach-at91/at91sam9g45_devices.c b/arch/arm/mach-at91/at91sam9g45_devices.c
index 809114d..a005bc9 100644
--- a/arch/arm/mach-at91/at91sam9g45_devices.c
+++ b/arch/arm/mach-at91/at91sam9g45_devices.c
@@ -68,7 +68,7 @@ static struct platform_device at_hdmac_device = {
void __init at91_add_device_hdmac(void)
{
dma_cap_set(DMA_MEMCPY, atdma_pdata.cap_mask);
- dma_cap_set(DMA_SLAVE, atdma_pdata.cap_mask);
+ dma_cap_set(DMA_SLAVE_SGLIST, atdma_pdata.cap_mask);
platform_device_register(&at_hdmac_device);
}
#else
diff --git a/drivers/dma/at_hdmac.c b/drivers/dma/at_hdmac.c
index a0f3e6a..0df7201 100644
--- a/drivers/dma/at_hdmac.c
+++ b/drivers/dma/at_hdmac.c
@@ -615,7 +615,7 @@ err_desc_get:
/**
- * atc_prep_slave_sg - prepare descriptors for a DMA_SLAVE transaction
+ * atc_prep_slave_sg - prepare descriptors for a DMA_SLAVE_SGLIST transaction
* @chan: DMA channel
* @sgl: scatterlist to transfer to/from
* @sg_len: number of entries in @scatterlist
@@ -1093,7 +1093,7 @@ static int __init at_dma_probe(struct platform_device *pdev)
if (dma_has_cap(DMA_MEMCPY, atdma->dma_common.cap_mask))
atdma->dma_common.device_prep_dma_memcpy = atc_prep_dma_memcpy;
- if (dma_has_cap(DMA_SLAVE, atdma->dma_common.cap_mask)) {
+ if (dma_has_cap(DMA_SLAVE_SGLIST, atdma->dma_common.cap_mask)) {
atdma->dma_common.device_prep_slave_sg = atc_prep_slave_sg;
atdma->dma_common.device_control = atc_control;
}
@@ -1102,8 +1102,8 @@ static int __init at_dma_probe(struct platform_device *pdev)
dev_info(&pdev->dev, "Atmel AHB DMA Controller ( %s%s), %d channels\n",
dma_has_cap(DMA_MEMCPY, atdma->dma_common.cap_mask) ? "cpy " : "",
- dma_has_cap(DMA_SLAVE, atdma->dma_common.cap_mask) ? "slave " : "",
- atdma->dma_common.chancnt);
+ dma_has_cap(DMA_SLAVE_SGLIST, atdma->dma_common.cap_mask) ?
+ "slave " : "", atdma->dma_common.chancnt);
dma_async_device_register(&atdma->dma_common);
diff --git a/drivers/dma/coh901318.c b/drivers/dma/coh901318.c
index 557e227..4bdd155 100644
--- a/drivers/dma/coh901318.c
+++ b/drivers/dma/coh901318.c
@@ -1518,7 +1518,7 @@ static int __init coh901318_probe(struct platform_device *pdev)
base);
dma_cap_zero(base->dma_slave.cap_mask);
- dma_cap_set(DMA_SLAVE, base->dma_slave.cap_mask);
+ dma_cap_set(DMA_SLAVE_SGLIST, base->dma_slave.cap_mask);
base->dma_slave.device_alloc_chan_resources = coh901318_alloc_chan_resources;
base->dma_slave.device_free_chan_resources = coh901318_free_chan_resources;
diff --git a/drivers/dma/dmaengine.c b/drivers/dma/dmaengine.c
index 9d31d5e..0d37c76 100644
--- a/drivers/dma/dmaengine.c
+++ b/drivers/dma/dmaengine.c
@@ -300,7 +300,6 @@ static int __init dma_channel_table_init(void)
*/
clear_bit(DMA_INTERRUPT, dma_cap_mask_all.bits);
clear_bit(DMA_PRIVATE, dma_cap_mask_all.bits);
- clear_bit(DMA_SLAVE, dma_cap_mask_all.bits);
for_each_dma_cap_mask(cap, dma_cap_mask_all) {
channel_table[cap] = alloc_percpu(struct dma_chan_tbl_ent);
@@ -690,9 +689,13 @@ int dma_async_device_register(struct dma_device *device)
!device->device_prep_dma_memset);
BUG_ON(dma_has_cap(DMA_INTERRUPT, device->cap_mask) &&
!device->device_prep_dma_interrupt);
- BUG_ON(dma_has_cap(DMA_SLAVE, device->cap_mask) &&
+ BUG_ON(dma_has_cap(DMA_SLAVE_SGLIST, device->cap_mask) &&
!device->device_prep_slave_sg);
- BUG_ON(dma_has_cap(DMA_SLAVE, device->cap_mask) &&
+ BUG_ON(dma_has_cap(DMA_SLAVE_SGLIST, device->cap_mask) &&
+ !device->device_control);
+ BUG_ON(dma_has_cap(DMA_SLAVE_BUFFER, device->cap_mask) &&
+ !device->device_prep_slave_buffer);
+ BUG_ON(dma_has_cap(DMA_SLAVE_BUFFER, device->cap_mask) &&
!device->device_control);
BUG_ON(!device->device_alloc_chan_resources);
diff --git a/drivers/dma/dw_dmac.c b/drivers/dma/dw_dmac.c
index a3991ab..e38ea35 100644
--- a/drivers/dma/dw_dmac.c
+++ b/drivers/dma/dw_dmac.c
@@ -1334,7 +1334,7 @@ static int __init dw_probe(struct platform_device *pdev)
channel_clear_bit(dw, MASK.ERROR, dw->all_chan_mask);
dma_cap_set(DMA_MEMCPY, dw->dma.cap_mask);
- dma_cap_set(DMA_SLAVE, dw->dma.cap_mask);
+ dma_cap_set(DMA_SLAVE_SGLIST, dw->dma.cap_mask);
dw->dma.dev = &pdev->dev;
dw->dma.device_alloc_chan_resources = dwc_alloc_chan_resources;
dw->dma.device_free_chan_resources = dwc_free_chan_resources;
diff --git a/drivers/dma/fsldma.c b/drivers/dma/fsldma.c
index cea08be..95b012a 100644
--- a/drivers/dma/fsldma.c
+++ b/drivers/dma/fsldma.c
@@ -584,16 +584,17 @@ fail:
}
/**
- * fsl_dma_prep_slave_sg - prepare descriptors for a DMA_SLAVE transaction
+ * fsl_dma_prep_slave_sg - prepare descriptors for a DMA_SLAVE_SGLIST
+ * transaction
* @chan: DMA channel
* @sgl: scatterlist to transfer to/from
* @sg_len: number of entries in @scatterlist
* @direction: DMA direction
* @flags: DMAEngine flags
*
- * Prepare a set of descriptors for a DMA_SLAVE transaction. Following the
- * DMA_SLAVE API, this gets the device-specific information from the
- * chan->private variable.
+ * Prepare a set of descriptors for a DMA_SLAVE_SGLIST transaction.
+ * Following the DMA_SLAVE API, this gets the device-specific information
+ * from the chan->private variable.
*/
static struct dma_async_tx_descriptor *fsl_dma_prep_slave_sg(
struct dma_chan *dchan, struct scatterlist *sgl, unsigned int sg_len,
@@ -1327,7 +1328,7 @@ static int __devinit fsldma_of_probe(struct platform_device *op,
dma_cap_set(DMA_MEMCPY, fdev->common.cap_mask);
dma_cap_set(DMA_INTERRUPT, fdev->common.cap_mask);
- dma_cap_set(DMA_SLAVE, fdev->common.cap_mask);
+ dma_cap_set(DMA_SLAVE_SGLIST, fdev->common.cap_mask);
fdev->common.device_alloc_chan_resources = fsl_dma_alloc_chan_resources;
fdev->common.device_free_chan_resources = fsl_dma_free_chan_resources;
fdev->common.device_prep_dma_interrupt = fsl_dma_prep_interrupt;
diff --git a/drivers/dma/ipu/ipu_idmac.c b/drivers/dma/ipu/ipu_idmac.c
index cb26ee9..27d8226 100644
--- a/drivers/dma/ipu/ipu_idmac.c
+++ b/drivers/dma/ipu/ipu_idmac.c
@@ -1665,7 +1665,7 @@ static int __init ipu_idmac_init(struct ipu *ipu)
struct dma_device *dma = &idmac->dma;
int i;
- dma_cap_set(DMA_SLAVE, dma->cap_mask);
+ dma_cap_set(DMA_SLAVE_SGLIST, dma->cap_mask);
dma_cap_set(DMA_PRIVATE, dma->cap_mask);
/* Compulsory common fields */
@@ -1675,7 +1675,7 @@ static int __init ipu_idmac_init(struct ipu *ipu)
dma->device_tx_status = idmac_tx_status;
dma->device_issue_pending = idmac_issue_pending;
- /* Compulsory for DMA_SLAVE fields */
+ /* Compulsory for DMA_SLAVE_SGLIST fields */
dma->device_prep_slave_sg = idmac_prep_slave_sg;
dma->device_control = idmac_control;
diff --git a/drivers/dma/pl330.c b/drivers/dma/pl330.c
index 7c50f6d..3974357 100644
--- a/drivers/dma/pl330.c
+++ b/drivers/dma/pl330.c
@@ -727,7 +727,7 @@ pl330_probe(struct amba_device *adev, struct amba_id *id)
break;
case MEMTODEV:
case DEVTOMEM:
- dma_cap_set(DMA_SLAVE, pd->cap_mask);
+ dma_cap_set(DMA_SLAVE_SGLIST, pd->cap_mask);
break;
default:
dev_err(&adev->dev, "DEVTODEV Not Supported\n");
diff --git a/drivers/dma/shdma.c b/drivers/dma/shdma.c
index fb64cf3..5b09f81 100644
--- a/drivers/dma/shdma.c
+++ b/drivers/dma/shdma.c
@@ -1040,7 +1040,7 @@ static int __init sh_dmae_probe(struct platform_device *pdev)
dma_cap_set(DMA_MEMCPY, shdev->common.cap_mask);
if (dmars)
- dma_cap_set(DMA_SLAVE, shdev->common.cap_mask);
+ dma_cap_set(DMA_SLAVE_SGLIST, shdev->common.cap_mask);
shdev->common.device_alloc_chan_resources
= sh_dmae_alloc_chan_resources;
@@ -1049,7 +1049,7 @@ static int __init sh_dmae_probe(struct platform_device *pdev)
shdev->common.device_tx_status = sh_dmae_tx_status;
shdev->common.device_issue_pending = sh_dmae_memcpy_issue_pending;
- /* Compulsory for DMA_SLAVE fields */
+ /* Compulsory for DMA_SLAVE_SGLIST fields */
shdev->common.device_prep_slave_sg = sh_dmae_prep_slave_sg;
shdev->common.device_control = sh_dmae_control;
diff --git a/drivers/dma/ste_dma40.c b/drivers/dma/ste_dma40.c
index 17e2600..9c6cde2 100644
--- a/drivers/dma/ste_dma40.c
+++ b/drivers/dma/ste_dma40.c
@@ -1213,14 +1213,15 @@ static int d40_config_memcpy(struct d40_chan *d40c)
{
dma_cap_mask_t cap = d40c->chan.device->cap_mask;
- if (dma_has_cap(DMA_MEMCPY, cap) && !dma_has_cap(DMA_SLAVE, cap)) {
+ if (dma_has_cap(DMA_MEMCPY, cap) &&
+ !dma_has_cap(DMA_SLAVE_SGLIST, cap)) {
d40c->dma_cfg = *d40c->base->plat_data->memcpy_conf_log;
d40c->dma_cfg.src_dev_type = STEDMA40_DEV_SRC_MEMORY;
d40c->dma_cfg.dst_dev_type = d40c->base->plat_data->
memcpy[d40c->chan.chan_id];
} else if (dma_has_cap(DMA_MEMCPY, cap) &&
- dma_has_cap(DMA_SLAVE, cap)) {
+ dma_has_cap(DMA_SLAVE_SGLIST, cap)) {
d40c->dma_cfg = *d40c->base->plat_data->memcpy_conf_phy;
} else {
dev_err(&d40c->chan.dev->device, "[%s] No memcpy\n",
@@ -2276,7 +2277,7 @@ static int __init d40_dmaengine_init(struct d40_base *base,
0, base->num_log_chans);
dma_cap_zero(base->dma_slave.cap_mask);
- dma_cap_set(DMA_SLAVE, base->dma_slave.cap_mask);
+ dma_cap_set(DMA_SLAVE_SGLIST, base->dma_slave.cap_mask);
base->dma_slave.device_alloc_chan_resources = d40_alloc_chan_resources;
base->dma_slave.device_free_chan_resources = d40_free_chan_resources;
@@ -2329,7 +2330,7 @@ static int __init d40_dmaengine_init(struct d40_base *base,
0, num_reserved_chans);
dma_cap_zero(base->dma_both.cap_mask);
- dma_cap_set(DMA_SLAVE, base->dma_both.cap_mask);
+ dma_cap_set(DMA_SLAVE_SGLIST, base->dma_both.cap_mask);
dma_cap_set(DMA_MEMCPY, base->dma_both.cap_mask);
base->dma_both.device_alloc_chan_resources = d40_alloc_chan_resources;
diff --git a/drivers/dma/timb_dma.c b/drivers/dma/timb_dma.c
index 2ec1ed5..75149883 100644
--- a/drivers/dma/timb_dma.c
+++ b/drivers/dma/timb_dma.c
@@ -744,7 +744,7 @@ static int __devinit td_probe(struct platform_device *pdev)
td->dma.device_tx_status = td_tx_status;
td->dma.device_issue_pending = td_issue_pending;
- dma_cap_set(DMA_SLAVE, td->dma.cap_mask);
+ dma_cap_set(DMA_SLAVE_SGLIST, td->dma.cap_mask);
dma_cap_set(DMA_PRIVATE, td->dma.cap_mask);
td->dma.device_prep_slave_sg = td_prep_slave_sg;
td->dma.device_control = td_control;
diff --git a/drivers/dma/txx9dmac.c b/drivers/dma/txx9dmac.c
index cbd83e36..c2d1343 100644
--- a/drivers/dma/txx9dmac.c
+++ b/drivers/dma/txx9dmac.c
@@ -1164,7 +1164,7 @@ static int __init txx9dmac_chan_probe(struct platform_device *pdev)
dma_cap_set(DMA_MEMCPY, dc->dma.cap_mask);
} else {
dc->dma.device_prep_slave_sg = txx9dmac_prep_slave_sg;
- dma_cap_set(DMA_SLAVE, dc->dma.cap_mask);
+ dma_cap_set(DMA_SLAVE_SGLIST, dc->dma.cap_mask);
dma_cap_set(DMA_PRIVATE, dc->dma.cap_mask);
}
@@ -1208,7 +1208,8 @@ static int __init txx9dmac_chan_probe(struct platform_device *pdev)
dev_dbg(&pdev->dev, "TXx9 DMA Channel (dma%d%s%s)\n",
dc->dma.dev_id,
dma_has_cap(DMA_MEMCPY, dc->dma.cap_mask) ? " memcpy" : "",
- dma_has_cap(DMA_SLAVE, dc->dma.cap_mask) ? " slave" : "");
+ dma_has_cap(DMA_SLAVE_SGLIST, dc->dma.cap_mask) ?
+ " slave" : "");
return 0;
}
diff --git a/drivers/media/video/mx3_camera.c b/drivers/media/video/mx3_camera.c
index a9be14c..fd6f735 100644
--- a/drivers/media/video/mx3_camera.c
+++ b/drivers/media/video/mx3_camera.c
@@ -780,7 +780,7 @@ static int acquire_dma_channel(struct mx3_camera_dev *mx3_cam)
}
dma_cap_zero(mask);
- dma_cap_set(DMA_SLAVE, mask);
+ dma_cap_set(DMA_SLAVE_SGLIST, mask);
dma_cap_set(DMA_PRIVATE, mask);
chan = dma_request_channel(mask, chan_filter, &rq);
if (!chan)
diff --git a/drivers/mmc/host/atmel-mci.c b/drivers/mmc/host/atmel-mci.c
index 95ef864..fae4b39 100644
--- a/drivers/mmc/host/atmel-mci.c
+++ b/drivers/mmc/host/atmel-mci.c
@@ -1728,7 +1728,7 @@ static void atmci_configure_dma(struct atmel_mci *host)
/* Try to grab a DMA channel */
dma_cap_zero(mask);
- dma_cap_set(DMA_SLAVE, mask);
+ dma_cap_set(DMA_SLAVE_SGLIST, mask);
host->dma.chan =
dma_request_channel(mask, filter, pdata->dma_slave);
}
diff --git a/drivers/mmc/host/tmio_mmc.c b/drivers/mmc/host/tmio_mmc.c
index ee7d0a5..905ffb2 100644
--- a/drivers/mmc/host/tmio_mmc.c
+++ b/drivers/mmc/host/tmio_mmc.c
@@ -591,7 +591,7 @@ static void tmio_mmc_request_dma(struct tmio_mmc_host *host,
dma_cap_mask_t mask;
dma_cap_zero(mask);
- dma_cap_set(DMA_SLAVE, mask);
+ dma_cap_set(DMA_SLAVE_SGLIST, mask);
host->chan_tx = dma_request_channel(mask, tmio_mmc_filter,
pdata->dma->chan_priv_tx);
diff --git a/drivers/serial/sh-sci.c b/drivers/serial/sh-sci.c
index c291b3a..0efab71 100644
--- a/drivers/serial/sh-sci.c
+++ b/drivers/serial/sh-sci.c
@@ -1316,7 +1316,7 @@ static void sci_request_dma(struct uart_port *port)
return;
dma_cap_zero(mask);
- dma_cap_set(DMA_SLAVE, mask);
+ dma_cap_set(DMA_SLAVE_SGLIST, mask);
param = &s->param_tx;
diff --git a/drivers/video/mx3fb.c b/drivers/video/mx3fb.c
index 7cfc170..ced7d86 100644
--- a/drivers/video/mx3fb.c
+++ b/drivers/video/mx3fb.c
@@ -1482,7 +1482,7 @@ static int mx3fb_probe(struct platform_device *pdev)
rq.mx3fb = mx3fb;
dma_cap_zero(mask);
- dma_cap_set(DMA_SLAVE, mask);
+ dma_cap_set(DMA_SLAVE_SGLIST, mask);
dma_cap_set(DMA_PRIVATE, mask);
rq.id = IDMAC_SDC_0;
chan = dma_request_channel(mask, chan_filter, &rq);
diff --git a/include/linux/dmaengine.h b/include/linux/dmaengine.h
index c61d4ca..5811d2b 100644
--- a/include/linux/dmaengine.h
+++ b/include/linux/dmaengine.h
@@ -66,11 +66,12 @@ enum dma_transaction_type {
DMA_INTERRUPT,
DMA_PRIVATE,
DMA_ASYNC_TX,
- DMA_SLAVE,
+ DMA_SLAVE_SGLIST,
+ DMA_SLAVE_BUFFER,
};
/* last transaction type for creation of the capabilities mask */
-#define DMA_TX_TYPE_END (DMA_SLAVE + 1)
+#define DMA_TX_TYPE_END (DMA_SLAVE_BUFFER + 1)
/**
@@ -421,7 +422,10 @@ struct dma_tx_state {
* @device_prep_dma_pq_val: prepares a pqzero_sum operation
* @device_prep_dma_memset: prepares a memset operation
* @device_prep_dma_interrupt: prepares an end of chain interrupt operation
- * @device_prep_slave_sg: prepares a slave dma operation
+ * @device_prep_slave_buffer: prepares a slave DMA operation to/from a
+ * buffer
+ * @device_prep_slave_sg: prepares a slave dma operation to/from a
+ * scatterlist
* @device_control: manipulate all pending operations on a channel, returns
* zero or error code
* @device_tx_status: poll for transaction completion, the optional
@@ -474,6 +478,10 @@ struct dma_device {
struct dma_async_tx_descriptor *(*device_prep_dma_interrupt)(
struct dma_chan *chan, unsigned long flags);
+ struct dma_async_tx_descriptor *(*device_prep_slave_buffer)(
+ struct dma_chan *chan, char *buffer,
+ enum dma_data_direction direction,
+ unsigned long flags);
struct dma_async_tx_descriptor *(*device_prep_slave_sg)(
struct dma_chan *chan, struct scatterlist *sgl,
unsigned int sg_len, enum dma_data_direction direction,
diff --git a/sound/atmel/abdac.c b/sound/atmel/abdac.c
index f2f41c8..642bdea 100644
--- a/sound/atmel/abdac.c
+++ b/sound/atmel/abdac.c
@@ -473,7 +473,7 @@ static int __devinit atmel_abdac_probe(struct platform_device *pdev)
dws->tx_reg = regs->start + DAC_DATA;
dma_cap_zero(mask);
- dma_cap_set(DMA_SLAVE, mask);
+ dma_cap_set(DMA_SLAVE_SGLIST, mask);
dac->dma.chan = dma_request_channel(mask, filter, dws);
}
diff --git a/sound/atmel/ac97c.c b/sound/atmel/ac97c.c
index 10c3a87..1f0d328 100644
--- a/sound/atmel/ac97c.c
+++ b/sound/atmel/ac97c.c
@@ -1013,7 +1013,7 @@ static int __devinit atmel_ac97c_probe(struct platform_device *pdev)
dws->rx_reg = regs->start + AC97C_CARHR + 2;
dma_cap_zero(mask);
- dma_cap_set(DMA_SLAVE, mask);
+ dma_cap_set(DMA_SLAVE_SGLIST, mask);
chip->dma.rx_chan = dma_request_channel(mask, filter,
dws);
@@ -1030,7 +1030,7 @@ static int __devinit atmel_ac97c_probe(struct platform_device *pdev)
dws->tx_reg = regs->start + AC97C_CATHR + 2;
dma_cap_zero(mask);
- dma_cap_set(DMA_SLAVE, mask);
+ dma_cap_set(DMA_SLAVE_SGLIST, mask);
chip->dma.tx_chan = dma_request_channel(mask, filter,
dws);
diff --git a/sound/soc/sh/siu_pcm.c b/sound/soc/sh/siu_pcm.c
index b0ccd0b..510d177 100644
--- a/sound/soc/sh/siu_pcm.c
+++ b/sound/soc/sh/siu_pcm.c
@@ -352,7 +352,7 @@ static int siu_pcm_open(struct snd_pcm_substream *ss)
struct sh_dmae_slave *param;
dma_cap_zero(mask);
- dma_cap_set(DMA_SLAVE, mask);
+ dma_cap_set(DMA_SLAVE_SGLIST, mask);
dev_dbg(dev, "%s, port=%d@%p\n", __func__, port, port_info);
diff --git a/sound/soc/txx9/txx9aclc.c b/sound/soc/txx9/txx9aclc.c
index 0e345230..11c9525 100644
--- a/sound/soc/txx9/txx9aclc.c
+++ b/sound/soc/txx9/txx9aclc.c
@@ -331,7 +331,7 @@ static int txx9aclc_dma_init(struct txx9aclc_soc_device *dev,
/* Try to grab a DMA channel */
dma_cap_zero(mask);
- dma_cap_set(DMA_SLAVE, mask);
+ dma_cap_set(DMA_SLAVE_SGLIST, mask);
dmadata->dma_chan = dma_request_channel(mask, filter, dmadata);
if (!dmadata->dma_chan) {
dev_err(dev->soc_dev.dev,
--
1.7.2
On Tue, Aug 10, 2010 at 11:46:06PM +0200, Linus Walleij wrote:
> This makes it possible for engines to implement slave transfers
> to be done to/from a simple kmalloc():ed memory buffer and not
> just from scatterlists.
Why is this needed? Drivers can just pass in a single-entry scatterlist
to the existing API to achieve the same functionality, and a couple of
them already do so.
Rabin
2010/8/11 Rabin Vincent <[email protected]>:
> On Tue, Aug 10, 2010 at 11:46:06PM +0200, Linus Walleij wrote:
>> This makes it possible for engines to implement slave transfers
>> to be done to/from a simple kmalloc():ed memory buffer and not
>> just from scatterlists.
>
> Why is this needed? ?Drivers can just pass in a single-entry scatterlist
> to the existing API to achieve the same functionality, and a couple of
> them already do so.
Because of the overhead, simply. Especially if you want to trigger
many jobs after each other. (This is necessary in device/slave-DMA
since every transaction may fail...) It's not just constructing the
sg-headers and freeing them again and again, it's also list traversals
here and there since the driver must assume it can be a linked sglist and
then two other list traversals for each
dma_map_sg()/dma_unmap_sg() pair and ... yeah that's basically
it.
And the number of extra code lines needed.
Then it's something conceptwise of creating a list that you know
is just always one element that is just not elegant, like taking a queue
numer and standing in queue when there is only one customer but
hey, maybe it's just me.
One way of achieving it for all present drivers is to wrap the passed
buffer in a single sglist and pass to the sglist function if the single
buffer call is not supported in the driver. Maybe it'd be better if I
coded up the patch like that so all driver can rely on this function
to be present?
Yours,
Linus Walleij
On Wed, Aug 11, 2010 at 09:02:04PM +0200, Linus Walleij wrote:
> 2010/8/11 Rabin Vincent <[email protected]>:
>
> > On Tue, Aug 10, 2010 at 11:46:06PM +0200, Linus Walleij wrote:
> >> This makes it possible for engines to implement slave transfers
> >> to be done to/from a simple kmalloc():ed memory buffer and not
> >> just from scatterlists.
> >
> > Why is this needed? ?Drivers can just pass in a single-entry scatterlist
> > to the existing API to achieve the same functionality, and a couple of
> > them already do so.
>
> Because of the overhead, simply. Especially if you want to trigger
> many jobs after each other. (This is necessary in device/slave-DMA
> since every transaction may fail...) It's not just constructing the
> sg-headers and freeing them again and again, it's also list traversals
> here and there since the driver must assume it can be a linked sglist and
> then two other list traversals for each
> dma_map_sg()/dma_unmap_sg() pair and ... yeah that's basically
> it.
>
> And the number of extra code lines needed.
>
> Then it's something conceptwise of creating a list that you know
> is just always one element that is just not elegant, like taking a queue
> numer and standing in queue when there is only one customer but
> hey, maybe it's just me.
>
> One way of achieving it for all present drivers is to wrap the passed
> buffer in a single sglist and pass to the sglist function if the single
> buffer call is not supported in the driver. Maybe it'd be better if I
> coded up the patch like that so all driver can rely on this function
> to be present?
+1 for this approach as it seems easier to implement and has benefits
for all drivers.
Sascha
--
Pengutronix e.K. | |
Industrial Linux Solutions | http://www.pengutronix.de/ |
Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0 |
Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 |
On Wed, Aug 11, 2010 at 09:02:04PM +0200, Linus Walleij wrote:
> 2010/8/11 Rabin Vincent <[email protected]>:
> > On Tue, Aug 10, 2010 at 11:46:06PM +0200, Linus Walleij wrote:
> >> This makes it possible for engines to implement slave transfers
> >> to be done to/from a simple kmalloc():ed memory buffer and not
> >> just from scatterlists.
> >
> > Why is this needed? ?Drivers can just pass in a single-entry scatterlist
> > to the existing API to achieve the same functionality, and a couple of
> > them already do so.
>
> Because of the overhead, simply. Especially if you want to trigger
> many jobs after each other. (This is necessary in device/slave-DMA
> since every transaction may fail...) It's not just constructing the
> sg-headers and freeing them again and again,
Note that the single length SG list can just be created on the stack.
For example, sound/soc/sh/siu_pcm.c.
> it's also list traversals
> here and there since the driver must assume it can be a linked sglist and
> then two other list traversals for each
> dma_map_sg()/dma_unmap_sg() pair and ... yeah that's basically
> it.
These list traversals of course run only one iteration for a single
length SG list.
>
> And the number of extra code lines needed.
It's about five lines, but yes, these are duplicated in drivers.
>
> Then it's something conceptwise of creating a list that you know
> is just always one element that is just not elegant, like taking a queue
> numer and standing in queue when there is only one customer but
> hey, maybe it's just me.
While I don't know about the overhead benefits, such an API would
probably be nice to have to at least avoid duplicating the sglist
building sequence.
Since it can be easily implemented as a wrapper over the existing API
with no change to exsting DMA drivers, why does a new cap need to be
added? Your suggestion below sounds like a better approach.
>
> One way of achieving it for all present drivers is to wrap the passed
> buffer in a single sglist and pass to the sglist function if the single
> buffer call is not supported in the driver. Maybe it'd be better if I
> coded up the patch like that so all driver can rely on this function
> to be present?
Rabin