* [PATCH 1/2] dw_dmac: backlink to dw_dma in dw_dma_chan is superfluous
@ 2012-12-10 11:36 Andy Shevchenko
2012-12-10 11:36 ` [PATCH 2/2] dw_dmac: check for mapping errors Andy Shevchenko
2012-12-10 13:50 ` [PATCH 1/2] dw_dmac: backlink to dw_dma in dw_dma_chan is superfluous Viresh Kumar
0 siblings, 2 replies; 4+ messages in thread
From: Andy Shevchenko @ 2012-12-10 11:36 UTC (permalink / raw)
To: Vinod Koul, spear-devel, linux-kernel, Viresh Kumar; +Cc: Andy Shevchenko
The same information could be exctracted from the struct dma_chan.
The patch introduces two helpers dwc_get_src_data_width() and
dwc_get_dst_data_width() as well.
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
---
drivers/dma/dw_dmac.c | 27 ++++++++++++++++++++-------
drivers/dma/dw_dmac_regs.h | 3 ---
2 files changed, 20 insertions(+), 10 deletions(-)
diff --git a/drivers/dma/dw_dmac.c b/drivers/dma/dw_dmac.c
index fdae96f..a9e9a07 100644
--- a/drivers/dma/dw_dmac.c
+++ b/drivers/dma/dw_dmac.c
@@ -73,6 +73,22 @@ static inline unsigned int dwc_get_sms(struct dw_dma_slave *slave)
*/
#define NR_DESCS_PER_CHANNEL 64
+static inline unsigned int dwc_get_src_data_width(struct dma_chan *chan)
+{
+ struct dw_dma *dw = to_dw_dma(chan->device);
+ struct dw_dma_slave *dws = chan->private;
+
+ return dw->data_width[dwc_get_sms(dws)];
+}
+
+static inline unsigned int dwc_get_dst_data_width(struct dma_chan *chan)
+{
+ struct dw_dma *dw = to_dw_dma(chan->device);
+ struct dw_dma_slave *dws = chan->private;
+
+ return dw->data_width[dwc_get_dms(dws)];
+}
+
/*----------------------------------------------------------------------*/
/*
@@ -703,7 +719,6 @@ dwc_prep_dma_memcpy(struct dma_chan *chan, dma_addr_t dest, dma_addr_t src,
size_t len, unsigned long flags)
{
struct dw_dma_chan *dwc = to_dw_dma_chan(chan);
- struct dw_dma_slave *dws = chan->private;
struct dw_desc *desc;
struct dw_desc *first;
struct dw_desc *prev;
@@ -726,8 +741,8 @@ dwc_prep_dma_memcpy(struct dma_chan *chan, dma_addr_t dest, dma_addr_t src,
dwc->direction = DMA_MEM_TO_MEM;
- data_width = min_t(unsigned int, dwc->dw->data_width[dwc_get_sms(dws)],
- dwc->dw->data_width[dwc_get_dms(dws)]);
+ data_width = min(dwc_get_src_data_width(chan),
+ dwc_get_dst_data_width(chan));
src_width = dst_width = min_t(unsigned int, data_width,
dwc_fast_fls(src | dest | len));
@@ -792,7 +807,6 @@ dwc_prep_slave_sg(struct dma_chan *chan, struct scatterlist *sgl,
unsigned long flags, void *context)
{
struct dw_dma_chan *dwc = to_dw_dma_chan(chan);
- struct dw_dma_slave *dws = chan->private;
struct dma_slave_config *sconfig = &dwc->dma_sconfig;
struct dw_desc *prev;
struct dw_desc *first;
@@ -826,7 +840,7 @@ dwc_prep_slave_sg(struct dma_chan *chan, struct scatterlist *sgl,
ctllo |= sconfig->device_fc ? DWC_CTLL_FC(DW_DMA_FC_P_M2P) :
DWC_CTLL_FC(DW_DMA_FC_D_M2P);
- data_width = dwc->dw->data_width[dwc_get_sms(dws)];
+ data_width = dwc_get_src_data_width(chan);
for_each_sg(sgl, sg, sg_len, i) {
struct dw_desc *desc;
@@ -889,7 +903,7 @@ slave_sg_todev_fill_desc:
ctllo |= sconfig->device_fc ? DWC_CTLL_FC(DW_DMA_FC_P_P2M) :
DWC_CTLL_FC(DW_DMA_FC_D_P2M);
- data_width = dwc->dw->data_width[dwc_get_dms(dws)];
+ data_width = dwc_get_dst_data_width(chan);
for_each_sg(sgl, sg, sg_len, i) {
struct dw_desc *desc;
@@ -1720,7 +1734,6 @@ static int dw_probe(struct platform_device *pdev)
channel_clear_bit(dw, CH_EN, dwc->mask);
- dwc->dw = dw;
dwc->direction = DMA_TRANS_NONE;
/* hardware configuration */
diff --git a/drivers/dma/dw_dmac_regs.h b/drivers/dma/dw_dmac_regs.h
index f9532c2..577f2dd 100644
--- a/drivers/dma/dw_dmac_regs.h
+++ b/drivers/dma/dw_dmac_regs.h
@@ -214,9 +214,6 @@ struct dw_dma_chan {
/* configuration passed via DMA_SLAVE_CONFIG */
struct dma_slave_config dma_sconfig;
-
- /* backlink to dw_dma */
- struct dw_dma *dw;
};
static inline struct dw_dma_chan_regs __iomem *
--
1.7.10.4
^ permalink raw reply [flat|nested] 4+ messages in thread* [PATCH 2/2] dw_dmac: check for mapping errors
2012-12-10 11:36 [PATCH 1/2] dw_dmac: backlink to dw_dma in dw_dma_chan is superfluous Andy Shevchenko
@ 2012-12-10 11:36 ` Andy Shevchenko
2012-12-10 13:57 ` Viresh Kumar
2012-12-10 13:50 ` [PATCH 1/2] dw_dmac: backlink to dw_dma in dw_dma_chan is superfluous Viresh Kumar
1 sibling, 1 reply; 4+ messages in thread
From: Andy Shevchenko @ 2012-12-10 11:36 UTC (permalink / raw)
To: Vinod Koul, spear-devel, linux-kernel, Viresh Kumar; +Cc: Andy Shevchenko
Otherwise we get a warning in case of CONFIG_DMA_API_DEBUG=y
[ 45.775943] WARNING: at lib/dma-debug.c:933 check_unmap+0x5d6/0x6ac()
[ 45.782369] dw_dmac dw_dmac.0: DMA-API: device driver failed to check map error[device address=0x00000000356efcc0] [size=28 bytes] [mapped as single]
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
---
drivers/dma/dw_dmac.c | 14 ++++++++++++--
1 file changed, 12 insertions(+), 2 deletions(-)
diff --git a/drivers/dma/dw_dmac.c b/drivers/dma/dw_dmac.c
index a9e9a07..38eea0c 100644
--- a/drivers/dma/dw_dmac.c
+++ b/drivers/dma/dw_dmac.c
@@ -1107,6 +1107,7 @@ static int dwc_alloc_chan_resources(struct dma_chan *chan)
struct dw_desc *desc;
int i;
unsigned long flags;
+ int ret = 0;
dev_vdbg(chan2dev(chan), "%s\n", __func__);
@@ -1131,8 +1132,7 @@ static int dwc_alloc_chan_resources(struct dma_chan *chan)
desc = kzalloc(sizeof(struct dw_desc), GFP_KERNEL);
if (!desc) {
- dev_info(chan2dev(chan),
- "only allocated %d descriptors\n", i);
+ ret = -ENOMEM;
spin_lock_irqsave(&dwc->lock, flags);
break;
}
@@ -1143,6 +1143,13 @@ static int dwc_alloc_chan_resources(struct dma_chan *chan)
desc->txd.flags = DMA_CTRL_ACK;
desc->txd.phys = dma_map_single(chan2parent(chan), &desc->lli,
sizeof(desc->lli), DMA_TO_DEVICE);
+ ret = dma_mapping_error(chan2parent(chan), desc->txd.phys);
+ if (ret) {
+ kfree(desc);
+ spin_lock_irqsave(&dwc->lock, flags);
+ break;
+ }
+
dwc_desc_put(dwc, desc);
spin_lock_irqsave(&dwc->lock, flags);
@@ -1151,6 +1158,9 @@ static int dwc_alloc_chan_resources(struct dma_chan *chan)
spin_unlock_irqrestore(&dwc->lock, flags);
+ if (ret)
+ dev_info(chan2dev(chan), "only allocated %d descriptors\n", i);
+
dev_dbg(chan2dev(chan), "%s: allocated %d descriptors\n", __func__, i);
return i;
--
1.7.10.4
^ permalink raw reply [flat|nested] 4+ messages in thread* Re: [PATCH 2/2] dw_dmac: check for mapping errors
2012-12-10 11:36 ` [PATCH 2/2] dw_dmac: check for mapping errors Andy Shevchenko
@ 2012-12-10 13:57 ` Viresh Kumar
0 siblings, 0 replies; 4+ messages in thread
From: Viresh Kumar @ 2012-12-10 13:57 UTC (permalink / raw)
To: Andy Shevchenko; +Cc: Vinod Koul, spear-devel, linux-kernel
On 10 December 2012 17:06, Andy Shevchenko
<andriy.shevchenko@linux.intel.com> wrote:
> Otherwise we get a warning in case of CONFIG_DMA_API_DEBUG=y
>
> [ 45.775943] WARNING: at lib/dma-debug.c:933 check_unmap+0x5d6/0x6ac()
> [ 45.782369] dw_dmac dw_dmac.0: DMA-API: device driver failed to check map error[device address=0x00000000356efcc0] [size=28 bytes] [mapped as single]
>
> Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
> ---
> drivers/dma/dw_dmac.c | 14 ++++++++++++--
> 1 file changed, 12 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/dma/dw_dmac.c b/drivers/dma/dw_dmac.c
> index a9e9a07..38eea0c 100644
> --- a/drivers/dma/dw_dmac.c
> +++ b/drivers/dma/dw_dmac.c
> @@ -1107,6 +1107,7 @@ static int dwc_alloc_chan_resources(struct dma_chan *chan)
> struct dw_desc *desc;
> int i;
> unsigned long flags;
> + int ret = 0;
>
> dev_vdbg(chan2dev(chan), "%s\n", __func__);
>
> @@ -1131,8 +1132,7 @@ static int dwc_alloc_chan_resources(struct dma_chan *chan)
>
> desc = kzalloc(sizeof(struct dw_desc), GFP_KERNEL);
> if (!desc) {
> - dev_info(chan2dev(chan),
> - "only allocated %d descriptors\n", i);
> + ret = -ENOMEM;
> spin_lock_irqsave(&dwc->lock, flags);
> break;
> }
> @@ -1143,6 +1143,13 @@ static int dwc_alloc_chan_resources(struct dma_chan *chan)
> desc->txd.flags = DMA_CTRL_ACK;
> desc->txd.phys = dma_map_single(chan2parent(chan), &desc->lli,
> sizeof(desc->lli), DMA_TO_DEVICE);
> + ret = dma_mapping_error(chan2parent(chan), desc->txd.phys);
> + if (ret) {
> + kfree(desc);
> + spin_lock_irqsave(&dwc->lock, flags);
> + break;
> + }
> +
> dwc_desc_put(dwc, desc);
>
> spin_lock_irqsave(&dwc->lock, flags);
> @@ -1151,6 +1158,9 @@ static int dwc_alloc_chan_resources(struct dma_chan *chan)
>
> spin_unlock_irqrestore(&dwc->lock, flags);
>
> + if (ret)
> + dev_info(chan2dev(chan), "only allocated %d descriptors\n", i);
> +
> dev_dbg(chan2dev(chan), "%s: allocated %d descriptors\n", __func__, i);
>
> return i;
I would use goto with appropriate label: to avoid lock/unlock of spin_lock.
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH 1/2] dw_dmac: backlink to dw_dma in dw_dma_chan is superfluous
2012-12-10 11:36 [PATCH 1/2] dw_dmac: backlink to dw_dma in dw_dma_chan is superfluous Andy Shevchenko
2012-12-10 11:36 ` [PATCH 2/2] dw_dmac: check for mapping errors Andy Shevchenko
@ 2012-12-10 13:50 ` Viresh Kumar
1 sibling, 0 replies; 4+ messages in thread
From: Viresh Kumar @ 2012-12-10 13:50 UTC (permalink / raw)
To: Andy Shevchenko; +Cc: Vinod Koul, spear-devel, linux-kernel
On 10 December 2012 17:06, Andy Shevchenko
<andriy.shevchenko@linux.intel.com> wrote:
> The same information could be exctracted from the struct dma_chan.
>
> The patch introduces two helpers dwc_get_src_data_width() and
> dwc_get_dst_data_width() as well.
>
> Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
> ---
> drivers/dma/dw_dmac.c | 27 ++++++++++++++++++++-------
> drivers/dma/dw_dmac_regs.h | 3 ---
> 2 files changed, 20 insertions(+), 10 deletions(-)
>
> diff --git a/drivers/dma/dw_dmac.c b/drivers/dma/dw_dmac.c
> index fdae96f..a9e9a07 100644
> --- a/drivers/dma/dw_dmac.c
> +++ b/drivers/dma/dw_dmac.c
> @@ -73,6 +73,22 @@ static inline unsigned int dwc_get_sms(struct dw_dma_slave *slave)
> */
> #define NR_DESCS_PER_CHANNEL 64
>
> +static inline unsigned int dwc_get_src_data_width(struct dma_chan *chan)
> +{
> + struct dw_dma *dw = to_dw_dma(chan->device);
> + struct dw_dma_slave *dws = chan->private;
I never liked the tabs used before the variable names in this driver.
Though have kept them while fixing existing routines.
So, i would like you to use spaces for new functions.
> +
> + return dw->data_width[dwc_get_sms(dws)];
> +}
> +
> +static inline unsigned int dwc_get_dst_data_width(struct dma_chan *chan)
> +{
> + struct dw_dma *dw = to_dw_dma(chan->device);
> + struct dw_dma_slave *dws = chan->private;
> +
> + return dw->data_width[dwc_get_dms(dws)];
> +}
Can you try merging these routines into a single routine, by adding a
single param?
> /*----------------------------------------------------------------------*/
>
> /*
> @@ -703,7 +719,6 @@ dwc_prep_dma_memcpy(struct dma_chan *chan, dma_addr_t dest, dma_addr_t src,
> size_t len, unsigned long flags)
> {
> struct dw_dma_chan *dwc = to_dw_dma_chan(chan);
> - struct dw_dma_slave *dws = chan->private;
> struct dw_desc *desc;
> struct dw_desc *first;
> struct dw_desc *prev;
> @@ -726,8 +741,8 @@ dwc_prep_dma_memcpy(struct dma_chan *chan, dma_addr_t dest, dma_addr_t src,
>
> dwc->direction = DMA_MEM_TO_MEM;
>
> - data_width = min_t(unsigned int, dwc->dw->data_width[dwc_get_sms(dws)],
> - dwc->dw->data_width[dwc_get_dms(dws)]);
> + data_width = min(dwc_get_src_data_width(chan),
> + dwc_get_dst_data_width(chan));
min_t() is more efficient then min(), that's why it was used.
> src_width = dst_width = min_t(unsigned int, data_width,
> dwc_fast_fls(src | dest | len));
> @@ -792,7 +807,6 @@ dwc_prep_slave_sg(struct dma_chan *chan, struct scatterlist *sgl,
> unsigned long flags, void *context)
> {
> struct dw_dma_chan *dwc = to_dw_dma_chan(chan);
> - struct dw_dma_slave *dws = chan->private;
> struct dma_slave_config *sconfig = &dwc->dma_sconfig;
> struct dw_desc *prev;
> struct dw_desc *first;
> @@ -826,7 +840,7 @@ dwc_prep_slave_sg(struct dma_chan *chan, struct scatterlist *sgl,
> ctllo |= sconfig->device_fc ? DWC_CTLL_FC(DW_DMA_FC_P_M2P) :
> DWC_CTLL_FC(DW_DMA_FC_D_M2P);
>
> - data_width = dwc->dw->data_width[dwc_get_sms(dws)];
> + data_width = dwc_get_src_data_width(chan);
>
> for_each_sg(sgl, sg, sg_len, i) {
> struct dw_desc *desc;
> @@ -889,7 +903,7 @@ slave_sg_todev_fill_desc:
> ctllo |= sconfig->device_fc ? DWC_CTLL_FC(DW_DMA_FC_P_P2M) :
> DWC_CTLL_FC(DW_DMA_FC_D_P2M);
>
> - data_width = dwc->dw->data_width[dwc_get_dms(dws)];
> + data_width = dwc_get_dst_data_width(chan);
>
> for_each_sg(sgl, sg, sg_len, i) {
> struct dw_desc *desc;
> @@ -1720,7 +1734,6 @@ static int dw_probe(struct platform_device *pdev)
>
> channel_clear_bit(dw, CH_EN, dwc->mask);
>
> - dwc->dw = dw;
> dwc->direction = DMA_TRANS_NONE;
>
> /* hardware configuration */
> diff --git a/drivers/dma/dw_dmac_regs.h b/drivers/dma/dw_dmac_regs.h
> index f9532c2..577f2dd 100644
> --- a/drivers/dma/dw_dmac_regs.h
> +++ b/drivers/dma/dw_dmac_regs.h
> @@ -214,9 +214,6 @@ struct dw_dma_chan {
>
> /* configuration passed via DMA_SLAVE_CONFIG */
> struct dma_slave_config dma_sconfig;
> -
> - /* backlink to dw_dma */
> - struct dw_dma *dw;
> };
>
> static inline struct dw_dma_chan_regs __iomem *
> --
> 1.7.10.4
>
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2012-12-10 13:57 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2012-12-10 11:36 [PATCH 1/2] dw_dmac: backlink to dw_dma in dw_dma_chan is superfluous Andy Shevchenko
2012-12-10 11:36 ` [PATCH 2/2] dw_dmac: check for mapping errors Andy Shevchenko
2012-12-10 13:57 ` Viresh Kumar
2012-12-10 13:50 ` [PATCH 1/2] dw_dmac: backlink to dw_dma in dw_dma_chan is superfluous Viresh Kumar
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
Powered by JetHome