[PATCH backport 4.4.y-4.8.z] dmaengine: at_xdmac: fix spurious flag status for mem2mem transfers

[PATCH backport 4.4.y-4.8.z] dmaengine: at_xdmac: fix spurious flag status for mem2mem transfers

[Nicolas Ferre]

From: Ludovic Desroches <ludovic.desroches@atmel.com>

commit 95da0c19d164f6df0b71a5187950f47d4b746e91 upstream.

When setting the channel configuration register, the perid field is not
set to 0 since it is useless for mem2mem transfers. Unfortunately, a
device has 0 as perid. It could cause spurious flags status because
the controller could mix some events from the two channels.
For that reason, use the highest perid value for mem2mem transfers since it
doesn't match the perid of other devices.

Signed-off-by: Ludovic Desroches <ludovic.desroches@atmel.com>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
[nicolas.ferre@atmel.com: backport to 4.4.y stable kernel]
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
---
 drivers/dma/at_xdmac.c | 18 +++++++++++++++---
 1 file changed, 15 insertions(+), 3 deletions(-)

diff --git a/drivers/dma/at_xdmac.c b/drivers/dma/at_xdmac.c
index 9d05d7dbcfa9..66c073fc8afc 100644
--- a/drivers/dma/at_xdmac.c
+++ b/drivers/dma/at_xdmac.c
@@ -864,8 +864,12 @@ at_xdmac_interleaved_queue_desc(struct dma_chan *chan,
 	 * access. Hopefully we can access DDR through both ports (at least on
 	 * SAMA5D4x), so we can use the same interface for source and dest,
 	 * that solves the fact we don't know the direction.
+	 * ERRATA: Even if useless for memory transfers, the PERID has to not
+	 * match the one of another channel. If not, it could lead to spurious
+	 * flag status.
 	 */
-	u32			chan_cc = AT_XDMAC_CC_DIF(0)
+	u32			chan_cc = AT_XDMAC_CC_PERID(0x3f)
+					| AT_XDMAC_CC_DIF(0)
 					| AT_XDMAC_CC_SIF(0)
 					| AT_XDMAC_CC_MBSIZE_SIXTEEN
 					| AT_XDMAC_CC_TYPE_MEM_TRAN;
@@ -1042,8 +1046,12 @@ at_xdmac_prep_dma_memcpy(struct dma_chan *chan, dma_addr_t dest, dma_addr_t src,
 	 * access DDR through both ports (at least on SAMA5D4x), so we can use
 	 * the same interface for source and dest, that solves the fact we
 	 * don't know the direction.
+	 * ERRATA: Even if useless for memory transfers, the PERID has to not
+	 * match the one of another channel. If not, it could lead to spurious
+	 * flag status.
 	 */
-	u32			chan_cc = AT_XDMAC_CC_DAM_INCREMENTED_AM
+	u32			chan_cc = AT_XDMAC_CC_PERID(0x3f)
+					| AT_XDMAC_CC_DAM_INCREMENTED_AM
 					| AT_XDMAC_CC_SAM_INCREMENTED_AM
 					| AT_XDMAC_CC_DIF(0)
 					| AT_XDMAC_CC_SIF(0)
@@ -1144,8 +1152,12 @@ static struct at_xdmac_desc *at_xdmac_memset_create_desc(struct dma_chan *chan,
 	 * access. Hopefully we can access DDR through both ports (at least on
 	 * SAMA5D4x), so we can use the same interface for source and dest,
 	 * that solves the fact we don't know the direction.
+	 * ERRATA: Even if useless for memory transfers, the PERID has to not
+	 * match the one of another channel. If not, it could lead to spurious
+	 * flag status.
 	 */
-	u32			chan_cc = AT_XDMAC_CC_DAM_UBS_AM
+	u32			chan_cc = AT_XDMAC_CC_PERID(0x3f)
+					| AT_XDMAC_CC_DAM_UBS_AM
 					| AT_XDMAC_CC_SAM_INCREMENTED_AM
 					| AT_XDMAC_CC_DIF(0)
 					| AT_XDMAC_CC_SIF(0)
-- 
2.9.0

Re: [PATCH backport 4.4.y-4.8.z] dmaengine: at_xdmac: fix spurious flag status for mem2mem transfers

[Greg Kroah-Hartman]

On Tue, Nov 15, 2016 at 04:30:48PM +0100, Nicolas Ferre wrote:
> From: Ludovic Desroches <ludovic.desroches@atmel.com>
> 
> commit 95da0c19d164f6df0b71a5187950f47d4b746e91 upstream.
> 
> When setting the channel configuration register, the perid field is not
> set to 0 since it is useless for mem2mem transfers. Unfortunately, a
> device has 0 as perid. It could cause spurious flags status because
> the controller could mix some events from the two channels.
> For that reason, use the highest perid value for mem2mem transfers since it
> doesn't match the perid of other devices.
> 
> Signed-off-by: Ludovic Desroches <ludovic.desroches@atmel.com>
> Signed-off-by: Vinod Koul <vinod.koul@intel.com>
> [nicolas.ferre@atmel.com: backport to 4.4.y stable kernel]

Why does your subject say 4.4.y-4.8.y when this patch showed up in
4.5-rc1?

confused,

greg k-h

Re: [PATCH backport 4.4.y-4.8.z] dmaengine: at_xdmac: fix spurious flag status for mem2mem transfers

[Nicolas Ferre]

Le 15/11/2016 � 20:27, Greg Kroah-Hartman a �crit :
> On Tue, Nov 15, 2016 at 04:30:48PM +0100, Nicolas Ferre wrote:
>> From: Ludovic Desroches <ludovic.desroches@atmel.com>
>>
>> commit 95da0c19d164f6df0b71a5187950f47d4b746e91 upstream.
>>
>> When setting the channel configuration register, the perid field is not
>> set to 0 since it is useless for mem2mem transfers. Unfortunately, a
>> device has 0 as perid. It could cause spurious flags status because
>> the controller could mix some events from the two channels.
>> For that reason, use the highest perid value for mem2mem transfers since it
>> doesn't match the perid of other devices.
>>
>> Signed-off-by: Ludovic Desroches <ludovic.desroches@atmel.com>
>> Signed-off-by: Vinod Koul <vinod.koul@intel.com>
>> [nicolas.ferre@atmel.com: backport to 4.4.y stable kernel]
> 
> Why does your subject say 4.4.y-4.8.y when this patch showed up in
> 4.5-rc1?

True. I mixed-up kernel revision numbers while checking where this patch
should apply (4.4.8 vs. 4.8.8) and didn't double check with original
commit ID: wrong!

Sorry for the confusion Gerg.

Best regards,
-- 
Nicolas Ferre