[PATCH 0/3] tty: serial: 8250_omap: Enable DMA support

[PATCH 0/3] tty: serial: 8250_omap: Enable DMA support

[Vignesh R]

This patch series re enables DMA support for UART 8250_omap driver.

Tested on AM335x, AM437x that use EDMA and OMAP5 and DRA74 EVM with
SDMA.

Vignesh R (3):
  serial :8250: omap: pause DMA only if DMA transfer in progress
  tty: serial: 8250_omap: Add OMAP_DMA_TX_KICK quirk for AM437x
  tty: serial: 8250_omap: Remove rx_dma_broken flag

 drivers/tty/serial/8250/8250_omap.c | 25 ++++++++++---------------
 1 file changed, 10 insertions(+), 15 deletions(-)

-- 
2.11.0

[PATCH 1/3] serial: 8250: omap: pause DMA only if DMA transfer in progress

[Vignesh R]

It is possible that DMA transfer is already complete but, completion
handler is yet to be called, when dmaengine_pause() is called in case of
error condition(like break/rx timeout). This leads to dmaengine_pause()
API to return EINVAL (as descriptor is already NULL) causing
rx_dma_broken flag to be set and effectively disabling RX DMA.
Fix this by calling dmaengine_pause() only when transfer is in progress.

Signed-off-by: Vignesh R <vigneshr@ti.com>
---
 drivers/tty/serial/8250/8250_omap.c | 11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)

diff --git a/drivers/tty/serial/8250/8250_omap.c b/drivers/tty/serial/8250/8250_omap.c
index 61ad6c3b20a0..4ad1934ef6ed 100644
--- a/drivers/tty/serial/8250/8250_omap.c
+++ b/drivers/tty/serial/8250/8250_omap.c
@@ -790,6 +790,7 @@ static void omap_8250_rx_dma_flush(struct uart_8250_port *p)
 {
 	struct omap8250_priv	*priv = p->port.private_data;
 	struct uart_8250_dma	*dma = p->dma;
+	struct dma_tx_state     state;
 	unsigned long		flags;
 	int ret;
 
@@ -800,10 +801,12 @@ static void omap_8250_rx_dma_flush(struct uart_8250_port *p)
 		return;
 	}
 
-	ret = dmaengine_pause(dma->rxchan);
-	if (WARN_ON_ONCE(ret))
-		priv->rx_dma_broken = true;
-
+	ret = dmaengine_tx_status(dma->rxchan, dma->rx_cookie, &state);
+	if (ret == DMA_IN_PROGRESS) {
+		ret = dmaengine_pause(dma->rxchan);
+		if (WARN_ON_ONCE(ret))
+			priv->rx_dma_broken = true;
+	}
 	spin_unlock_irqrestore(&priv->rx_dma_lock, flags);
 
 	__dma_rx_do_complete(p);
-- 
2.11.0

[PATCH 2/3] tty: serial: 8250_omap: Add OMAP_DMA_TX_KICK quirk for AM437x

[Vignesh R]

UART uses as EDMA as dma engine on AM437x SoC and therefore, requires
OMAP_DMA_TX_KICK quirk just like AM33xx. So, enable OMAP_DMA_TX_KICK
quirk for AM437x platform as well. While at that, drop use of
of_machine_is_compatible() and instead pass quirks via device data.

Signed-off-by: Vignesh R <vigneshr@ti.com>
---
 drivers/tty/serial/8250/8250_omap.c | 9 +++------
 1 file changed, 3 insertions(+), 6 deletions(-)

diff --git a/drivers/tty/serial/8250/8250_omap.c b/drivers/tty/serial/8250/8250_omap.c
index 4ad1934ef6ed..97766dcd67d4 100644
--- a/drivers/tty/serial/8250/8250_omap.c
+++ b/drivers/tty/serial/8250/8250_omap.c
@@ -1078,15 +1078,15 @@ static int omap8250_no_handle_irq(struct uart_port *port)
 }
 
 static const u8 am3352_habit = OMAP_DMA_TX_KICK | UART_ERRATA_CLOCK_DISABLE;
-static const u8 am4372_habit = UART_ERRATA_CLOCK_DISABLE;
+static const u8 dra742_habit = UART_ERRATA_CLOCK_DISABLE;
 
 static const struct of_device_id omap8250_dt_ids[] = {
 	{ .compatible = "ti,omap2-uart" },
 	{ .compatible = "ti,omap3-uart" },
 	{ .compatible = "ti,omap4-uart" },
 	{ .compatible = "ti,am3352-uart", .data = &am3352_habit, },
-	{ .compatible = "ti,am4372-uart", .data = &am4372_habit, },
-	{ .compatible = "ti,dra742-uart", .data = &am4372_habit, },
+	{ .compatible = "ti,am4372-uart", .data = &am3352_habit, },
+	{ .compatible = "ti,dra742-uart", .data = &dra742_habit, },
 	{},
 };
 MODULE_DEVICE_TABLE(of, omap8250_dt_ids);
@@ -1221,9 +1221,6 @@ static int omap8250_probe(struct platform_device *pdev)
 			priv->omap8250_dma.rx_size = RX_TRIGGER;
 			priv->omap8250_dma.rxconf.src_maxburst = RX_TRIGGER;
 			priv->omap8250_dma.txconf.dst_maxburst = TX_TRIGGER;
-
-			if (of_machine_is_compatible("ti,am33xx"))
-				priv->habit |= OMAP_DMA_TX_KICK;
 			/*
 			 * pause is currently not supported atleast on omap-sdma
 			 * and edma on most earlier kernels.
-- 
2.11.0

[PATCH 3/3] tty: serial: 8250_omap: Remove rx_dma_broken flag

[Vignesh R]

8250 UART DMA support was marked broken by default as it was not
possible to pause ongoing RX DMA transfer. Now that both SDMA and
EDMA can support pause operation for RX DMA transactions, don't set
rx_dma_broken to true by default. With this patch 8250_omap driver will
use DMA by default.

Signed-off-by: Vignesh R <vigneshr@ti.com>
---
 drivers/tty/serial/8250/8250_omap.c | 5 -----
 1 file changed, 5 deletions(-)

diff --git a/drivers/tty/serial/8250/8250_omap.c b/drivers/tty/serial/8250/8250_omap.c
index 97766dcd67d4..68a6393d9636 100644
--- a/drivers/tty/serial/8250/8250_omap.c
+++ b/drivers/tty/serial/8250/8250_omap.c
@@ -1221,11 +1221,6 @@ static int omap8250_probe(struct platform_device *pdev)
 			priv->omap8250_dma.rx_size = RX_TRIGGER;
 			priv->omap8250_dma.rxconf.src_maxburst = RX_TRIGGER;
 			priv->omap8250_dma.txconf.dst_maxburst = TX_TRIGGER;
-			/*
-			 * pause is currently not supported atleast on omap-sdma
-			 * and edma on most earlier kernels.
-			 */
-			priv->rx_dma_broken = true;
 		}
 	}
 #endif
-- 
2.11.0

Re: [PATCH 0/3] tty: serial: 8250_omap: Enable DMA support

[Tony Lindgren]

* Vignesh R <vigneshr@ti.com> [170113 00:03]:
> This patch series re enables DMA support for UART 8250_omap driver.
> 
> Tested on AM335x, AM437x that use EDMA and OMAP5 and DRA74 EVM with
> SDMA.

Is 8250_omap serial console working for you on omap5 in general?

I've noticed that it's really unresponsive for me as if the FIFO
interrupt was not working. For example logging in might take several
attempts and a long time with each character showing up much later
after some timeout.

TX on the uart on omap5 seems to work OK, I do see all the console
messages fine. And these patches do not make the issue better or
worse.

Regards,

Tony

Re: [PATCH 0/3] tty: serial: 8250_omap: Enable DMA support

[Vignesh R]

On Friday 13 January 2017 11:50 PM, Tony Lindgren wrote:
> * Vignesh R <vigneshr@ti.com> [170113 00:03]:
>> This patch series re enables DMA support for UART 8250_omap driver.
>>
>> Tested on AM335x, AM437x that use EDMA and OMAP5 and DRA74 EVM with
>> SDMA.
> 
> Is 8250_omap serial console working for you on omap5 in general?
> 
> I've noticed that it's really unresponsive for me as if the FIFO
> interrupt was not working. For example logging in might take several
> attempts and a long time with each character showing up much later
> after some timeout.
> 

Yes, I did face the same issue on omap5 using 8250_omap driver for
console. Looks like this bug has existed all along.

> TX on the uart on omap5 seems to work OK, I do see all the console
> messages fine. And these patches do not make the issue better or
> worse.
> 

True.

-- 
Regards
Vignesh

Re: [PATCH 0/3] tty: serial: 8250_omap: Enable DMA support

[Vignesh R]

On Tuesday 17 January 2017 11:36 AM, Vignesh R wrote:
> 
> 
> On Friday 13 January 2017 11:50 PM, Tony Lindgren wrote:
>> * Vignesh R <vigneshr@ti.com> [170113 00:03]:
>>> This patch series re enables DMA support for UART 8250_omap driver.
>>>
>>> Tested on AM335x, AM437x that use EDMA and OMAP5 and DRA74 EVM with
>>> SDMA.
>>
>> Is 8250_omap serial console working for you on omap5 in general?
>>
>> I've noticed that it's really unresponsive for me as if the FIFO
>> interrupt was not working. For example logging in might take several
>> attempts and a long time with each character showing up much later
>> after some timeout.
>>
> 
> Yes, I did face the same issue on omap5 using 8250_omap driver for
> console. Looks like this bug has existed all along.
> 

This looks a separate issue, I don't see any errata specific errata wrt
UART on OMAP5. I will try to debug further.

-- 
Regards
Vignesh

Re: [PATCH 0/3] tty: serial: 8250_omap: Enable DMA support

[Tony Lindgren]

* Vignesh R <vigneshr@ti.com> [170117 02:06]:
> 
> 
> On Tuesday 17 January 2017 11:36 AM, Vignesh R wrote:
> > 
> > 
> > On Friday 13 January 2017 11:50 PM, Tony Lindgren wrote:
> >> * Vignesh R <vigneshr@ti.com> [170113 00:03]:
> >>> This patch series re enables DMA support for UART 8250_omap driver.
> >>>
> >>> Tested on AM335x, AM437x that use EDMA and OMAP5 and DRA74 EVM with
> >>> SDMA.
> >>
> >> Is 8250_omap serial console working for you on omap5 in general?
> >>
> >> I've noticed that it's really unresponsive for me as if the FIFO
> >> interrupt was not working. For example logging in might take several
> >> attempts and a long time with each character showing up much later
> >> after some timeout.
> >>
> > 
> > Yes, I did face the same issue on omap5 using 8250_omap driver for
> > console. Looks like this bug has existed all along.
> > 
> 
> This looks a separate issue, I don't see any errata specific errata wrt
> UART on OMAP5. I will try to debug further.

OK maybe it's something configured or not configured by the bootloader
that we're missing in the Linux driver?

Other than the omap5 issue, 8250_OMAP seems to behave on all variants
that I've tried it with so far.

Regards,

Tony

Re: [PATCH 0/3] tty: serial: 8250_omap: Enable DMA support

[Vignesh R]

On Wednesday 18 January 2017 04:55 AM, Tony Lindgren wrote:
> * Vignesh R <vigneshr@ti.com> [170117 02:06]:
>>
>>
>> On Tuesday 17 January 2017 11:36 AM, Vignesh R wrote:
>>>
>>>
>>> On Friday 13 January 2017 11:50 PM, Tony Lindgren wrote:
>>>> * Vignesh R <vigneshr@ti.com> [170113 00:03]:
>>>>> This patch series re enables DMA support for UART 8250_omap driver.
>>>>>
>>>>> Tested on AM335x, AM437x that use EDMA and OMAP5 and DRA74 EVM with
>>>>> SDMA.
>>>>
>>>> Is 8250_omap serial console working for you on omap5 in general?
>>>>
>>>> I've noticed that it's really unresponsive for me as if the FIFO
>>>> interrupt was not working. For example logging in might take several
>>>> attempts and a long time with each character showing up much later
>>>> after some timeout.
>>>>
>>>
>>> Yes, I did face the same issue on omap5 using 8250_omap driver for
>>> console. Looks like this bug has existed all along.
>>>
>>
>> This looks a separate issue, I don't see any errata specific errata wrt
>> UART on OMAP5. I will try to debug further.
> 
> OK maybe it's something configured or not configured by the bootloader
> that we're missing in the Linux driver?
> 

omap-serial uses a UART RX FIFO trigger of 1 byte whereas 8250_omap
driver uses a RX FIFO trigger of 48 bytes. If less than 48 bytes is
received then RX Timeout interrupt is raised which will help the driver
to flush the FIFO. It seems like RX timeout does not seem to bring out
UART IP from idle state on OMAP5.

I have posted a fix based on what is done for DRA7 in hwmod code:
http://lists.infradead.org/pipermail/linux-arm-kernel/2017-January/480619.html


-- 
Regards
Vignesh

Re: [PATCH 0/3] tty: serial: 8250_omap: Enable DMA support

[Tony Lindgren]

* Vignesh R <vigneshr@ti.com> [170118 04:04]:
> 
> 
> On Wednesday 18 January 2017 04:55 AM, Tony Lindgren wrote:
> > * Vignesh R <vigneshr@ti.com> [170117 02:06]:
> >>
> >>
> >> On Tuesday 17 January 2017 11:36 AM, Vignesh R wrote:
> >>>
> >>>
> >>> On Friday 13 January 2017 11:50 PM, Tony Lindgren wrote:
> >>>> * Vignesh R <vigneshr@ti.com> [170113 00:03]:
> >>>>> This patch series re enables DMA support for UART 8250_omap driver.
> >>>>>
> >>>>> Tested on AM335x, AM437x that use EDMA and OMAP5 and DRA74 EVM with
> >>>>> SDMA.
> >>>>
> >>>> Is 8250_omap serial console working for you on omap5 in general?
> >>>>
> >>>> I've noticed that it's really unresponsive for me as if the FIFO
> >>>> interrupt was not working. For example logging in might take several
> >>>> attempts and a long time with each character showing up much later
> >>>> after some timeout.
> >>>>
> >>>
> >>> Yes, I did face the same issue on omap5 using 8250_omap driver for
> >>> console. Looks like this bug has existed all along.
> >>>
> >>
> >> This looks a separate issue, I don't see any errata specific errata wrt
> >> UART on OMAP5. I will try to debug further.
> > 
> > OK maybe it's something configured or not configured by the bootloader
> > that we're missing in the Linux driver?
> > 
> 
> omap-serial uses a UART RX FIFO trigger of 1 byte whereas 8250_omap
> driver uses a RX FIFO trigger of 48 bytes. If less than 48 bytes is
> received then RX Timeout interrupt is raised which will help the driver
> to flush the FIFO. It seems like RX timeout does not seem to bring out
> UART IP from idle state on OMAP5.
> 
> I have posted a fix based on what is done for DRA7 in hwmod code:
> http://lists.infradead.org/pipermail/linux-arm-kernel/2017-January/480619.html

OK great, thanks for fixing that! Will give it a try a bit later today.

Regards,

Tony

Re: [PATCH 0/3] tty: serial: 8250_omap: Enable DMA support

[Greg Kroah-Hartman]

On Fri, Jan 13, 2017 at 10:20:21AM -0800, Tony Lindgren wrote:
> * Vignesh R <vigneshr@ti.com> [170113 00:03]:
> > This patch series re enables DMA support for UART 8250_omap driver.
> > 
> > Tested on AM335x, AM437x that use EDMA and OMAP5 and DRA74 EVM with
> > SDMA.
> 
> Is 8250_omap serial console working for you on omap5 in general?
> 
> I've noticed that it's really unresponsive for me as if the FIFO
> interrupt was not working. For example logging in might take several
> attempts and a long time with each character showing up much later
> after some timeout.
> 
> TX on the uart on omap5 seems to work OK, I do see all the console
> messages fine. And these patches do not make the issue better or
> worse.

Ok, I'm dropping them from my review queue then.

thanks,

greg k-h

Re: [PATCH 0/3] tty: serial: 8250_omap: Enable DMA support

[Tony Lindgren]

* Greg Kroah-Hartman <gregkh@linuxfoundation.org> [170119 05:25]:
> On Fri, Jan 13, 2017 at 10:20:21AM -0800, Tony Lindgren wrote:
> > * Vignesh R <vigneshr@ti.com> [170113 00:03]:
> > > This patch series re enables DMA support for UART 8250_omap driver.
> > > 
> > > Tested on AM335x, AM437x that use EDMA and OMAP5 and DRA74 EVM with
> > > SDMA.
> > 
> > Is 8250_omap serial console working for you on omap5 in general?
> > 
> > I've noticed that it's really unresponsive for me as if the FIFO
> > interrupt was not working. For example logging in might take several
> > attempts and a long time with each character showing up much later
> > after some timeout.
> > 
> > TX on the uart on omap5 seems to work OK, I do see all the console
> > messages fine. And these patches do not make the issue better or
> > worse.
> 
> Ok, I'm dropping them from my review queue then.

It turned out to be not related to this set of patches, Vignesh
posted a fix for it. For this set:

Acked-by: Tony Lindgren <tony@atomide.com>

Re: [PATCH 0/3] tty: serial: 8250_omap: Enable DMA support

[Greg Kroah-Hartman]

On Thu, Jan 19, 2017 at 06:54:23AM -0800, Tony Lindgren wrote:
> * Greg Kroah-Hartman <gregkh@linuxfoundation.org> [170119 05:25]:
> > On Fri, Jan 13, 2017 at 10:20:21AM -0800, Tony Lindgren wrote:
> > > * Vignesh R <vigneshr@ti.com> [170113 00:03]:
> > > > This patch series re enables DMA support for UART 8250_omap driver.
> > > > 
> > > > Tested on AM335x, AM437x that use EDMA and OMAP5 and DRA74 EVM with
> > > > SDMA.
> > > 
> > > Is 8250_omap serial console working for you on omap5 in general?
> > > 
> > > I've noticed that it's really unresponsive for me as if the FIFO
> > > interrupt was not working. For example logging in might take several
> > > attempts and a long time with each character showing up much later
> > > after some timeout.
> > > 
> > > TX on the uart on omap5 seems to work OK, I do see all the console
> > > messages fine. And these patches do not make the issue better or
> > > worse.
> > 
> > Ok, I'm dropping them from my review queue then.
> 
> It turned out to be not related to this set of patches, Vignesh
> posted a fix for it. For this set:
> 
> Acked-by: Tony Lindgren <tony@atomide.com>

Ugh, these are now gone from my queue, can someone resend them with this
acked-by added?

thanks,

greg k-h

Re: [PATCH 0/3] tty: serial: 8250_omap: Enable DMA support

[Vignesh R]

On Thursday 19 January 2017 09:05 PM, Greg Kroah-Hartman wrote:
> On Thu, Jan 19, 2017 at 06:54:23AM -0800, Tony Lindgren wrote:
>> * Greg Kroah-Hartman <gregkh@linuxfoundation.org> [170119 05:25]:
>>> On Fri, Jan 13, 2017 at 10:20:21AM -0800, Tony Lindgren wrote:
>>>> * Vignesh R <vigneshr@ti.com> [170113 00:03]:
>>>>> This patch series re enables DMA support for UART 8250_omap driver.
>>>>>
>>>>> Tested on AM335x, AM437x that use EDMA and OMAP5 and DRA74 EVM with
>>>>> SDMA.
>>>>
>>>> Is 8250_omap serial console working for you on omap5 in general?
>>>>
>>>> I've noticed that it's really unresponsive for me as if the FIFO
>>>> interrupt was not working. For example logging in might take several
>>>> attempts and a long time with each character showing up much later
>>>> after some timeout.
>>>>
>>>> TX on the uart on omap5 seems to work OK, I do see all the console
>>>> messages fine. And these patches do not make the issue better or
>>>> worse.
>>>
>>> Ok, I'm dropping them from my review queue then.
>>
>> It turned out to be not related to this set of patches, Vignesh
>> posted a fix for it. For this set:
>>
>> Acked-by: Tony Lindgren <tony@atomide.com>
> 
> Ugh, these are now gone from my queue, can someone resend them with this
> acked-by added?
> 

I have resent this series with the Ack:
http://marc.info/?l=linux-serial&m=148490028003788&w=2

-- 
Regards
Vignesh