[PATCH v3 1/1] serial: 8250: Toggle IER bits on only after irq has been set up

[PATCH v3 1/1] serial: 8250: Toggle IER bits on only after irq has been set up

[Ilpo Järvinen]

Invoking TIOCVHANGUP on 8250_mid port on Ice Lake-D and then reopening
the port triggers these faults during serial8250_do_startup():

  DMAR: DRHD: handling fault status reg 3
  DMAR: [DMA Write NO_PASID] Request device [00:1a.0] fault addr 0x0 [fault reason 0x05] PTE Write access is not set

If the IRQ hasn't been set up yet, the UART will have zeroes in its MSI
address/data registers. Disabling the IRQ at the interrupt controller
won't stop the UART from performing a DMA write to the address programmed
in its MSI address register (zero) when it wants to signal an interrupt.

The UARTs (in Ice Lake-D) implement PCI 2.1 style MSI without masking
capability, so there is no way to mask the interrupt at the source PCI
function level, except disabling the MSI capability entirely, but that
would cause it to fall back to INTx# assertion, and the PCI specification
prohibits disabling the MSI capability as a way to mask a function's
interrupt service request.

The MSI address register is zeroed during by the hangup as the irq is
freed. The interrupt is signalled during serial8250_do_startup()
performing a THRE test that temporarily toggles THRI in IER. The THRE
test currently occurs before UART's irq (and MSI address) is properly
set up.

Refactor serial8250_do_startup() such that irq is set up before the
THRE test. The current irq setup code is intermixed with the timer
setup code. As THRE test must be performed prior to the timer setup,
extract it into own function and call it only after the THRE test.

Reported-by: Lennert Buytenhek <buytenh@arista.com>
Tested-by: Lennert Buytenhek <buytenh@arista.com>
Fixes: 40b36daad0ac ("[PATCH] 8250 UART backup timer")
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
---

v3:
- Improved the commit with Lennert's superior descriptions.
- Added Andy's Rev-by

 drivers/tty/serial/8250/8250.h      |  2 ++
 drivers/tty/serial/8250/8250_core.c | 16 +++++++++++-----
 drivers/tty/serial/8250/8250_port.c |  8 +++++---
 3 files changed, 18 insertions(+), 8 deletions(-)

diff --git a/drivers/tty/serial/8250/8250.h b/drivers/tty/serial/8250/8250.h
index 287153d32536..dbf4c1204bf3 100644
--- a/drivers/tty/serial/8250/8250.h
+++ b/drivers/tty/serial/8250/8250.h
@@ -403,3 +403,5 @@ static inline int serial_index(struct uart_port *port)
 {
 	return port->minor - 64;
 }
+
+void univ8250_setup_timer(struct uart_8250_port *up);
diff --git a/drivers/tty/serial/8250/8250_core.c b/drivers/tty/serial/8250/8250_core.c
index 2e83e7367441..10d535640434 100644
--- a/drivers/tty/serial/8250/8250_core.c
+++ b/drivers/tty/serial/8250/8250_core.c
@@ -298,10 +298,9 @@ static void serial8250_backup_timeout(struct timer_list *t)
 		jiffies + uart_poll_timeout(&up->port) + HZ / 5);
 }
 
-static int univ8250_setup_irq(struct uart_8250_port *up)
+void univ8250_setup_timer(struct uart_8250_port *up)
 {
 	struct uart_port *port = &up->port;
-	int retval = 0;
 
 	/*
 	 * The above check will only give an accurate result the first time
@@ -322,10 +321,17 @@ static int univ8250_setup_irq(struct uart_8250_port *up)
 	 */
 	if (!port->irq)
 		mod_timer(&up->timer, jiffies + uart_poll_timeout(port));
-	else
-		retval = serial_link_irq_chain(up);
+}
+EXPORT_SYMBOL_GPL(univ8250_setup_timer);
 
-	return retval;
+static int univ8250_setup_irq(struct uart_8250_port *up)
+{
+	struct uart_port *port = &up->port;
+
+	if (port->irq)
+		return serial_link_irq_chain(up);
+
+	return 0;
 }
 
 static void univ8250_release_irq(struct uart_8250_port *up)
diff --git a/drivers/tty/serial/8250/8250_port.c b/drivers/tty/serial/8250/8250_port.c
index 39b35a61958c..6e8e16227a3a 100644
--- a/drivers/tty/serial/8250/8250_port.c
+++ b/drivers/tty/serial/8250/8250_port.c
@@ -2294,6 +2294,10 @@ int serial8250_do_startup(struct uart_port *port)
 	if (port->irq && (up->port.flags & UPF_SHARE_IRQ))
 		up->port.irqflags |= IRQF_SHARED;
 
+	retval = up->ops->setup_irq(up);
+	if (retval)
+		goto out;
+
 	if (port->irq && !(up->port.flags & UPF_NO_THRE_TEST)) {
 		unsigned char iir1;
 
@@ -2336,9 +2340,7 @@ int serial8250_do_startup(struct uart_port *port)
 		}
 	}
 
-	retval = up->ops->setup_irq(up);
-	if (retval)
-		goto out;
+	univ8250_setup_timer(up);
 
 	/*
 	 * Now, initialize the UART
-- 
2.30.2

Re: [PATCH v3 1/1] serial: 8250: Toggle IER bits on only after irq has been set up

[Ilpo Järvinen]

[-- Attachment #1: Type: text/plain, Size: 4743 bytes --]

On Mon, 19 Sep 2022, Ilpo Järvinen wrote:

> Invoking TIOCVHANGUP on 8250_mid port on Ice Lake-D and then reopening
> the port triggers these faults during serial8250_do_startup():
> 
>   DMAR: DRHD: handling fault status reg 3
>   DMAR: [DMA Write NO_PASID] Request device [00:1a.0] fault addr 0x0 [fault reason 0x05] PTE Write access is not set
> 
> If the IRQ hasn't been set up yet, the UART will have zeroes in its MSI
> address/data registers. Disabling the IRQ at the interrupt controller
> won't stop the UART from performing a DMA write to the address programmed
> in its MSI address register (zero) when it wants to signal an interrupt.
> 
> The UARTs (in Ice Lake-D) implement PCI 2.1 style MSI without masking
> capability, so there is no way to mask the interrupt at the source PCI
> function level, except disabling the MSI capability entirely, but that
> would cause it to fall back to INTx# assertion, and the PCI specification
> prohibits disabling the MSI capability as a way to mask a function's
> interrupt service request.
> 
> The MSI address register is zeroed during by the hangup as the irq is
> freed. The interrupt is signalled during serial8250_do_startup()
> performing a THRE test that temporarily toggles THRI in IER. The THRE
> test currently occurs before UART's irq (and MSI address) is properly
> set up.
> 
> Refactor serial8250_do_startup() such that irq is set up before the
> THRE test. The current irq setup code is intermixed with the timer
> setup code. As THRE test must be performed prior to the timer setup,
> extract it into own function and call it only after the THRE test.
> 
> Reported-by: Lennert Buytenhek <buytenh@arista.com>
> Tested-by: Lennert Buytenhek <buytenh@arista.com>
> Fixes: 40b36daad0ac ("[PATCH] 8250 UART backup timer")
> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
> Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
> ---
> 
> v3:
> - Improved the commit with Lennert's superior descriptions.
> - Added Andy's Rev-by
> 
>  drivers/tty/serial/8250/8250.h      |  2 ++
>  drivers/tty/serial/8250/8250_core.c | 16 +++++++++++-----
>  drivers/tty/serial/8250/8250_port.c |  8 +++++---
>  3 files changed, 18 insertions(+), 8 deletions(-)
> 
> diff --git a/drivers/tty/serial/8250/8250.h b/drivers/tty/serial/8250/8250.h
> index 287153d32536..dbf4c1204bf3 100644
> --- a/drivers/tty/serial/8250/8250.h
> +++ b/drivers/tty/serial/8250/8250.h
> @@ -403,3 +403,5 @@ static inline int serial_index(struct uart_port *port)
>  {
>  	return port->minor - 64;
>  }
> +
> +void univ8250_setup_timer(struct uart_8250_port *up);
> diff --git a/drivers/tty/serial/8250/8250_core.c b/drivers/tty/serial/8250/8250_core.c
> index 2e83e7367441..10d535640434 100644
> --- a/drivers/tty/serial/8250/8250_core.c
> +++ b/drivers/tty/serial/8250/8250_core.c
> @@ -298,10 +298,9 @@ static void serial8250_backup_timeout(struct timer_list *t)
>  		jiffies + uart_poll_timeout(&up->port) + HZ / 5);
>  }
>  
> -static int univ8250_setup_irq(struct uart_8250_port *up)
> +void univ8250_setup_timer(struct uart_8250_port *up)
>  {
>  	struct uart_port *port = &up->port;
> -	int retval = 0;
>  
>  	/*
>  	 * The above check will only give an accurate result the first time
> @@ -322,10 +321,17 @@ static int univ8250_setup_irq(struct uart_8250_port *up)
>  	 */
>  	if (!port->irq)
>  		mod_timer(&up->timer, jiffies + uart_poll_timeout(port));
> -	else
> -		retval = serial_link_irq_chain(up);
> +}
> +EXPORT_SYMBOL_GPL(univ8250_setup_timer);
>  
> -	return retval;
> +static int univ8250_setup_irq(struct uart_8250_port *up)
> +{
> +	struct uart_port *port = &up->port;
> +
> +	if (port->irq)
> +		return serial_link_irq_chain(up);
> +
> +	return 0;
>  }
>  
>  static void univ8250_release_irq(struct uart_8250_port *up)
> diff --git a/drivers/tty/serial/8250/8250_port.c b/drivers/tty/serial/8250/8250_port.c
> index 39b35a61958c..6e8e16227a3a 100644
> --- a/drivers/tty/serial/8250/8250_port.c
> +++ b/drivers/tty/serial/8250/8250_port.c
> @@ -2294,6 +2294,10 @@ int serial8250_do_startup(struct uart_port *port)
>  	if (port->irq && (up->port.flags & UPF_SHARE_IRQ))
>  		up->port.irqflags |= IRQF_SHARED;
>  
> +	retval = up->ops->setup_irq(up);
> +	if (retval)
> +		goto out;
> +
>  	if (port->irq && !(up->port.flags & UPF_NO_THRE_TEST)) {
>  		unsigned char iir1;
>  
> @@ -2336,9 +2340,7 @@ int serial8250_do_startup(struct uart_port *port)
>  		}
>  	}
>  
> -	retval = up->ops->setup_irq(up);
> -	if (retval)
> -		goto out;
> +	univ8250_setup_timer(up);
>  
>  	/*
>  	 * Now, initialize the UART

Hi Greg,

Please scratch this patch. It seems to create a circular dependency issue 
with allmodconfig. I'll send v4 once the problem is sorted out.

-- 
 i.

Re: [PATCH v3 1/1] serial: 8250: Toggle IER bits on only after irq has been set up

[Andy Shevchenko]

On Wed, Sep 21, 2022 at 04:20:50PM +0300, Ilpo Järvinen wrote:
> On Mon, 19 Sep 2022, Ilpo Järvinen wrote:

...

> > -	retval = up->ops->setup_irq(up);
> > -	if (retval)
> > -		goto out;
> > +	univ8250_setup_timer(up);

> Please scratch this patch. It seems to create a circular dependency issue
> with allmodconfig. I'll send v4 once the problem is sorted out.

In accordance with what you told me privately, I think you will need yet
another ops, i.e. ->setup_timer().

-- 
With Best Regards,
Andy Shevchenko