* [PATCH] dmaengine: ti: k3-udma-glue: move psi-l pairing in channel en/dis functions
@ 2020-10-30 20:30 Grygorii Strashko
2020-11-02 7:46 ` Peter Ujfalusi
2020-11-09 11:54 ` Vinod Koul
0 siblings, 2 replies; 4+ messages in thread
From: Grygorii Strashko @ 2020-10-30 20:30 UTC (permalink / raw)
To: Vinod Koul, Peter Ujfalusi, dmaengine
Cc: Sekhar Nori, linux-kernel, Vignesh Raghavendra, Grygorii Strashko
The NAVSS UDMA will stuck if target IP module is disabled by PM while PSI-L
threads are paired UDMA<->IP and no further transfers is possible. This
could be the case for IPs J721E Main CPSW (cpsw9g).
Hence, to avoid such situation do PSI-L threads pairing only when UDMA
channel is going to be enabled as at this time DMA consumer module expected
to be active already.
Signed-off-by: Grygorii Strashko <grygorii.strashko@ti.com>
---
drivers/dma/ti/k3-udma-glue.c | 64 +++++++++++++++++++++--------------
1 file changed, 38 insertions(+), 26 deletions(-)
diff --git a/drivers/dma/ti/k3-udma-glue.c b/drivers/dma/ti/k3-udma-glue.c
index a367584f0d7b..dfb65e382ab9 100644
--- a/drivers/dma/ti/k3-udma-glue.c
+++ b/drivers/dma/ti/k3-udma-glue.c
@@ -303,19 +303,6 @@ struct k3_udma_glue_tx_channel *k3_udma_glue_request_tx_chn(struct device *dev,
goto err;
}
- ret = xudma_navss_psil_pair(tx_chn->common.udmax,
- tx_chn->common.src_thread,
- tx_chn->common.dst_thread);
- if (ret) {
- dev_err(dev, "PSI-L request err %d\n", ret);
- goto err;
- }
-
- tx_chn->psil_paired = true;
-
- /* reset TX RT registers */
- k3_udma_glue_disable_tx_chn(tx_chn);
-
k3_udma_glue_dump_tx_chn(tx_chn);
return tx_chn;
@@ -378,6 +365,18 @@ EXPORT_SYMBOL_GPL(k3_udma_glue_pop_tx_chn);
int k3_udma_glue_enable_tx_chn(struct k3_udma_glue_tx_channel *tx_chn)
{
+ int ret;
+
+ ret = xudma_navss_psil_pair(tx_chn->common.udmax,
+ tx_chn->common.src_thread,
+ tx_chn->common.dst_thread);
+ if (ret) {
+ dev_err(tx_chn->common.dev, "PSI-L request err %d\n", ret);
+ return ret;
+ }
+
+ tx_chn->psil_paired = true;
+
xudma_tchanrt_write(tx_chn->udma_tchanx, UDMA_CHAN_RT_PEER_RT_EN_REG,
UDMA_PEER_RT_EN_ENABLE);
@@ -398,6 +397,13 @@ void k3_udma_glue_disable_tx_chn(struct k3_udma_glue_tx_channel *tx_chn)
xudma_tchanrt_write(tx_chn->udma_tchanx,
UDMA_CHAN_RT_PEER_RT_EN_REG, 0);
k3_udma_glue_dump_tx_rt_chn(tx_chn, "txchn dis2");
+
+ if (tx_chn->psil_paired) {
+ xudma_navss_psil_unpair(tx_chn->common.udmax,
+ tx_chn->common.src_thread,
+ tx_chn->common.dst_thread);
+ tx_chn->psil_paired = false;
+ }
}
EXPORT_SYMBOL_GPL(k3_udma_glue_disable_tx_chn);
@@ -815,19 +821,6 @@ k3_udma_glue_request_rx_chn_priv(struct device *dev, const char *name,
goto err;
}
- ret = xudma_navss_psil_pair(rx_chn->common.udmax,
- rx_chn->common.src_thread,
- rx_chn->common.dst_thread);
- if (ret) {
- dev_err(dev, "PSI-L request err %d\n", ret);
- goto err;
- }
-
- rx_chn->psil_paired = true;
-
- /* reset RX RT registers */
- k3_udma_glue_disable_rx_chn(rx_chn);
-
k3_udma_glue_dump_rx_chn(rx_chn);
return rx_chn;
@@ -1052,12 +1045,24 @@ EXPORT_SYMBOL_GPL(k3_udma_glue_rx_flow_disable);
int k3_udma_glue_enable_rx_chn(struct k3_udma_glue_rx_channel *rx_chn)
{
+ int ret;
+
if (rx_chn->remote)
return -EINVAL;
if (rx_chn->flows_ready < rx_chn->flow_num)
return -EINVAL;
+ ret = xudma_navss_psil_pair(rx_chn->common.udmax,
+ rx_chn->common.src_thread,
+ rx_chn->common.dst_thread);
+ if (ret) {
+ dev_err(rx_chn->common.dev, "PSI-L request err %d\n", ret);
+ return ret;
+ }
+
+ rx_chn->psil_paired = true;
+
xudma_rchanrt_write(rx_chn->udma_rchanx, UDMA_CHAN_RT_CTL_REG,
UDMA_CHAN_RT_CTL_EN);
@@ -1078,6 +1083,13 @@ void k3_udma_glue_disable_rx_chn(struct k3_udma_glue_rx_channel *rx_chn)
xudma_rchanrt_write(rx_chn->udma_rchanx, UDMA_CHAN_RT_CTL_REG, 0);
k3_udma_glue_dump_rx_rt_chn(rx_chn, "rxrt dis2");
+
+ if (rx_chn->psil_paired) {
+ xudma_navss_psil_unpair(rx_chn->common.udmax,
+ rx_chn->common.src_thread,
+ rx_chn->common.dst_thread);
+ rx_chn->psil_paired = false;
+ }
}
EXPORT_SYMBOL_GPL(k3_udma_glue_disable_rx_chn);
--
2.17.1
^ permalink raw reply related [flat|nested] 4+ messages in thread
* Re: [PATCH] dmaengine: ti: k3-udma-glue: move psi-l pairing in channel en/dis functions
2020-10-30 20:30 [PATCH] dmaengine: ti: k3-udma-glue: move psi-l pairing in channel en/dis functions Grygorii Strashko
@ 2020-11-02 7:46 ` Peter Ujfalusi
2020-11-04 11:32 ` Grygorii Strashko
2020-11-09 11:54 ` Vinod Koul
1 sibling, 1 reply; 4+ messages in thread
From: Peter Ujfalusi @ 2020-11-02 7:46 UTC (permalink / raw)
To: Grygorii Strashko, Vinod Koul, dmaengine
Cc: Sekhar Nori, linux-kernel, Vignesh Raghavendra
On 30/10/2020 22.30, Grygorii Strashko wrote:
> The NAVSS UDMA will stuck if target IP module is disabled by PM while PSI-L
> threads are paired UDMA<->IP and no further transfers is possible. This
> could be the case for IPs J721E Main CPSW (cpsw9g).
>
> Hence, to avoid such situation do PSI-L threads pairing only when UDMA
> channel is going to be enabled as at this time DMA consumer module expected
> to be active already.
Is this patch on top of the AM64 (BCDMA/PKTDMA) series or not?
Will it cause any conflict?
Acked-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
> Signed-off-by: Grygorii Strashko <grygorii.strashko@ti.com>
> ---
> drivers/dma/ti/k3-udma-glue.c | 64 +++++++++++++++++++++--------------
> 1 file changed, 38 insertions(+), 26 deletions(-)
>
> diff --git a/drivers/dma/ti/k3-udma-glue.c b/drivers/dma/ti/k3-udma-glue.c
> index a367584f0d7b..dfb65e382ab9 100644
> --- a/drivers/dma/ti/k3-udma-glue.c
> +++ b/drivers/dma/ti/k3-udma-glue.c
> @@ -303,19 +303,6 @@ struct k3_udma_glue_tx_channel *k3_udma_glue_request_tx_chn(struct device *dev,
> goto err;
> }
>
> - ret = xudma_navss_psil_pair(tx_chn->common.udmax,
> - tx_chn->common.src_thread,
> - tx_chn->common.dst_thread);
> - if (ret) {
> - dev_err(dev, "PSI-L request err %d\n", ret);
> - goto err;
> - }
> -
> - tx_chn->psil_paired = true;
> -
> - /* reset TX RT registers */
> - k3_udma_glue_disable_tx_chn(tx_chn);
> -
> k3_udma_glue_dump_tx_chn(tx_chn);
>
> return tx_chn;
> @@ -378,6 +365,18 @@ EXPORT_SYMBOL_GPL(k3_udma_glue_pop_tx_chn);
>
> int k3_udma_glue_enable_tx_chn(struct k3_udma_glue_tx_channel *tx_chn)
> {
> + int ret;
> +
> + ret = xudma_navss_psil_pair(tx_chn->common.udmax,
> + tx_chn->common.src_thread,
> + tx_chn->common.dst_thread);
> + if (ret) {
> + dev_err(tx_chn->common.dev, "PSI-L request err %d\n", ret);
> + return ret;
> + }
> +
> + tx_chn->psil_paired = true;
> +
> xudma_tchanrt_write(tx_chn->udma_tchanx, UDMA_CHAN_RT_PEER_RT_EN_REG,
> UDMA_PEER_RT_EN_ENABLE);
>
> @@ -398,6 +397,13 @@ void k3_udma_glue_disable_tx_chn(struct k3_udma_glue_tx_channel *tx_chn)
> xudma_tchanrt_write(tx_chn->udma_tchanx,
> UDMA_CHAN_RT_PEER_RT_EN_REG, 0);
> k3_udma_glue_dump_tx_rt_chn(tx_chn, "txchn dis2");
> +
> + if (tx_chn->psil_paired) {
> + xudma_navss_psil_unpair(tx_chn->common.udmax,
> + tx_chn->common.src_thread,
> + tx_chn->common.dst_thread);
> + tx_chn->psil_paired = false;
> + }
> }
> EXPORT_SYMBOL_GPL(k3_udma_glue_disable_tx_chn);
>
> @@ -815,19 +821,6 @@ k3_udma_glue_request_rx_chn_priv(struct device *dev, const char *name,
> goto err;
> }
>
> - ret = xudma_navss_psil_pair(rx_chn->common.udmax,
> - rx_chn->common.src_thread,
> - rx_chn->common.dst_thread);
> - if (ret) {
> - dev_err(dev, "PSI-L request err %d\n", ret);
> - goto err;
> - }
> -
> - rx_chn->psil_paired = true;
> -
> - /* reset RX RT registers */
> - k3_udma_glue_disable_rx_chn(rx_chn);
> -
> k3_udma_glue_dump_rx_chn(rx_chn);
>
> return rx_chn;
> @@ -1052,12 +1045,24 @@ EXPORT_SYMBOL_GPL(k3_udma_glue_rx_flow_disable);
>
> int k3_udma_glue_enable_rx_chn(struct k3_udma_glue_rx_channel *rx_chn)
> {
> + int ret;
> +
> if (rx_chn->remote)
> return -EINVAL;
>
> if (rx_chn->flows_ready < rx_chn->flow_num)
> return -EINVAL;
>
> + ret = xudma_navss_psil_pair(rx_chn->common.udmax,
> + rx_chn->common.src_thread,
> + rx_chn->common.dst_thread);
> + if (ret) {
> + dev_err(rx_chn->common.dev, "PSI-L request err %d\n", ret);
> + return ret;
> + }
> +
> + rx_chn->psil_paired = true;
> +
> xudma_rchanrt_write(rx_chn->udma_rchanx, UDMA_CHAN_RT_CTL_REG,
> UDMA_CHAN_RT_CTL_EN);
>
> @@ -1078,6 +1083,13 @@ void k3_udma_glue_disable_rx_chn(struct k3_udma_glue_rx_channel *rx_chn)
> xudma_rchanrt_write(rx_chn->udma_rchanx, UDMA_CHAN_RT_CTL_REG, 0);
>
> k3_udma_glue_dump_rx_rt_chn(rx_chn, "rxrt dis2");
> +
> + if (rx_chn->psil_paired) {
> + xudma_navss_psil_unpair(rx_chn->common.udmax,
> + rx_chn->common.src_thread,
> + rx_chn->common.dst_thread);
> + rx_chn->psil_paired = false;
> + }
> }
> EXPORT_SYMBOL_GPL(k3_udma_glue_disable_rx_chn);
>
>
- Péter
Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki.
Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH] dmaengine: ti: k3-udma-glue: move psi-l pairing in channel en/dis functions
2020-11-02 7:46 ` Peter Ujfalusi
@ 2020-11-04 11:32 ` Grygorii Strashko
0 siblings, 0 replies; 4+ messages in thread
From: Grygorii Strashko @ 2020-11-04 11:32 UTC (permalink / raw)
To: Peter Ujfalusi, Vinod Koul, dmaengine
Cc: Sekhar Nori, linux-kernel, Vignesh Raghavendra
On 02/11/2020 09:46, Peter Ujfalusi wrote:
>
>
> On 30/10/2020 22.30, Grygorii Strashko wrote:
>> The NAVSS UDMA will stuck if target IP module is disabled by PM while PSI-L
>> threads are paired UDMA<->IP and no further transfers is possible. This
>> could be the case for IPs J721E Main CPSW (cpsw9g).
>>
>> Hence, to avoid such situation do PSI-L threads pairing only when UDMA
>> channel is going to be enabled as at this time DMA consumer module expected
>> to be active already.
>
> Is this patch on top of the AM64 (BCDMA/PKTDMA) series or not?
> Will it cause any conflict?
No. It was not based on top of AM64 series and I've not checked for conflicts.
>
> Acked-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
>
>
>> Signed-off-by: Grygorii Strashko <grygorii.strashko@ti.com>
>> ---
>> drivers/dma/ti/k3-udma-glue.c | 64 +++++++++++++++++++++--------------
>> 1 file changed, 38 insertions(+), 26 deletions(-)
...
--
Best regards,
grygorii
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH] dmaengine: ti: k3-udma-glue: move psi-l pairing in channel en/dis functions
2020-10-30 20:30 [PATCH] dmaengine: ti: k3-udma-glue: move psi-l pairing in channel en/dis functions Grygorii Strashko
2020-11-02 7:46 ` Peter Ujfalusi
@ 2020-11-09 11:54 ` Vinod Koul
1 sibling, 0 replies; 4+ messages in thread
From: Vinod Koul @ 2020-11-09 11:54 UTC (permalink / raw)
To: Grygorii Strashko
Cc: Peter Ujfalusi, dmaengine, Sekhar Nori, linux-kernel,
Vignesh Raghavendra
On 30-10-20, 22:30, Grygorii Strashko wrote:
> The NAVSS UDMA will stuck if target IP module is disabled by PM while PSI-L
> threads are paired UDMA<->IP and no further transfers is possible. This
> could be the case for IPs J721E Main CPSW (cpsw9g).
>
> Hence, to avoid such situation do PSI-L threads pairing only when UDMA
> channel is going to be enabled as at this time DMA consumer module expected
> to be active already.
Applied, thanks
--
~Vinod
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2020-11-09 11:54 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-10-30 20:30 [PATCH] dmaengine: ti: k3-udma-glue: move psi-l pairing in channel en/dis functions Grygorii Strashko
2020-11-02 7:46 ` Peter Ujfalusi
2020-11-04 11:32 ` Grygorii Strashko
2020-11-09 11:54 ` Vinod Koul
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).