Message-ID: <07ee2375-382c-154a-4c47-abb1a81b3351@raspberrypi.org>
Date:   Tue, 18 Sep 2018 14:13:15 +0100
From:   Phil Elwell <phil@...pberrypi.org>
To:     Greg Kroah-Hartman <gregkh@...uxfoundation.org>
Cc:     Jiri Slaby <jslaby@...e.com>, linux-serial@...r.kernel.org,
        linux-kernel@...r.kernel.org, Alexander Graf <agraf@...e.de>,
        Stefan Wahren <stefan.wahren@...e.com>
Subject: Re: [PATCH 1/2] sc16is7xx: Fix for multi-channel stall

Hi Greg,

On 18/09/2018 14:02, Greg Kroah-Hartman wrote:
> On Wed, Sep 12, 2018 at 03:31:55PM +0100, Phil Elwell wrote:
>> The SC16IS752 is a dual-channel device. The two channels are largely
>> independent, but the IRQ signals are wired together as an open-drain,
>> active low signal which will be driven low while either of the
>> channels requires attention, which can be for significant periods of
>> time until operations complete and the interrupt can be acknowledged.
>> In that respect it is should be treated as a true level-sensitive IRQ.
>>
>> The kernel, however, needs to be able to exit interrupt context in
>> order to use I2C or SPI to access the device registers (which may
>> involve sleeping).  Therefore the interrupt needs to be masked out or
>> paused in some way.
>>
>> The usual way to manage sleeping from within an interrupt handler
>> is to use a threaded interrupt handler - a regular interrupt routine
>> does the minimum amount of work needed to triage the interrupt before
>> waking the interrupt service thread. If the threaded IRQ is marked as
>> IRQF_ONESHOT the kernel will automatically mask out the interrupt
>> until the thread runs to completion. The sc16is7xx driver used to
>> use a threaded IRQ, but a patch switched to using a kthread_worker
>> in order to set realtime priorities on the handler thread and for
>> other optimisations. The end result is non-threaded IRQ that
>> schedules some work then returns IRQ_HANDLED, making the kernel
>> think that all IRQ processing has completed.
>>
>> The work-around to prevent a constant stream of interrupts is to
>> mark the interrupt as edge-sensitive rather than level-sensitive,
>> but interpreting an active-low source as a falling-edge source
>> requires care to prevent a total cessation of interrupts. Whereas
>> an edge-triggering source will generate a new edge for every interrupt
>> condition a level-triggering source will keep the signal at the
>> interrupting level until it no longer requires attention; in other
>> words, the host won't see another edge until all interrupt conditions
>> are cleared. It is therefore vital that the interrupt handler does not
>> exit with an outstanding interrupt condition, otherwise the kernel
>> will not receive another interrupt unless some other operation causes
>> the interrupt state on the device to be cleared.
>>
>> The existing sc16is7xx driver has a very simple interrupt "thread"
>> (kthread_work job) that processes interrupts on each channel in turn
>> until there are no more. If both channels are active and the first
>> channel starts interrupting while the handler for the second channel
>> is running then it will not be detected and an IRQ stall ensues. This
>> could be handled easily if there was a shared IRQ status register, or
>> a convenient way to determine if the IRQ had been deasserted for any
>> length of time, but both appear to be lacking.
>>
>> Avoid this problem (or at least make it much less likely to happen)
>> by reducing the granularity of per-channel interrupt processing
>> to one condition per iteration, only exiting the overall loop when
>> both channels are no longer interrupting.
>>
>> Signed-off-by: Phil Elwell <phil@...pberrypi.org>
>> ---
>>  drivers/tty/serial/sc16is7xx.c | 19 +++++++++++++------
>>  1 file changed, 13 insertions(+), 6 deletions(-)
>>
>> diff --git a/drivers/tty/serial/sc16is7xx.c b/drivers/tty/serial/sc16is7xx.c
>> index 243c960..47b4115 100644
>> --- a/drivers/tty/serial/sc16is7xx.c
>> +++ b/drivers/tty/serial/sc16is7xx.c
>> @@ -657,7 +657,7 @@ static void sc16is7xx_handle_tx(struct uart_port *port)
>>  		uart_write_wakeup(port);
>>  }
>>  
>> -static void sc16is7xx_port_irq(struct sc16is7xx_port *s, int portno)
>> +static bool sc16is7xx_port_irq(struct sc16is7xx_port *s, int portno)
>>  {
>>  	struct uart_port *port = &s->p[portno].port;
>>  
>> @@ -666,7 +666,7 @@ static void sc16is7xx_port_irq(struct sc16is7xx_port *s, int portno)
>>  
>>  		iir = sc16is7xx_port_read(port, SC16IS7XX_IIR_REG);
>>  		if (iir & SC16IS7XX_IIR_NO_INT_BIT)
>> -			break;
>> +			return false;
>>  
>>  		iir &= SC16IS7XX_IIR_ID_MASK;
>>  
>> @@ -688,16 +688,23 @@ static void sc16is7xx_port_irq(struct sc16is7xx_port *s, int portno)
>>  					    port->line, iir);
>>  			break;
>>  		}
>> -	} while (1);
>> +	} while (0);
>> +	return true;
>>  }
>>  
>>  static void sc16is7xx_ist(struct kthread_work *ws)
>>  {
>>  	struct sc16is7xx_port *s = to_sc16is7xx_port(ws, irq_work);
>> -	int i;
>>  
>> -	for (i = 0; i < s->devtype->nr_uart; ++i)
>> -		sc16is7xx_port_irq(s, i);
>> +	while (1) {
>> +		bool keep_polling = false;
>> +		int i;
>> +
>> +		for (i = 0; i < s->devtype->nr_uart; ++i)
>> +			keep_polling |= sc16is7xx_port_irq(s, i);
>> +		if (!keep_polling)
>> +			break;
> 
> This makes me worried, there is no "timeout" now?  What happens if this
> never happens, will you just sit and spin forever?  What prevents that?

The patch is keeping to the spirit of the original, which also has a
potentially infinite loop (in sc16is7xx_port_irq) - this just moves it
up one level.

I could add a limit on the number of iterations, but if the limit is ever hit,
leading to an early exit, the port is basically dead because it will never
receive another interrupt.

Phil

Powered by blists - more mailing lists