Re: locking changes in tty broke low latency feature

[Peter Hurley]

Hi Thomas,

On 02/23/2014 05:33 PM, Thomas Gleixner wrote:
> On Thu, 20 Feb 2014, Peter Hurley wrote:
>
> > On 02/19/2014 09:55 PM, Peter Hurley wrote:
> > > On 02/19/2014 06:06 PM, Hal Murray wrote:
> > > > > Can you give me an idea of your device's average and minimum required
> > > > > latency (please be specific)?  Is your target arch x86 [so I can
> > > > > evaluate the
> > > > > the impact of bus-locked instructions relative to your expected]?
> > > >
> > > > The code I'm familiar with is ntpd and gpsd.  They run on almost any
> > > > hardware
> > > > or OS and talk to a wide collection of devices.
> > > >
> > > > There is no hard requirement for latency.  They just work better with
> > > > lower
> > > > latency.  The lower the better.
> > > >
> > > > People gripe about the latency due to USB polling which is about a ms.
> > >
> > > Have you tried 3.12+ without low_latency? I ripped out a lot of locks
> > > from 3.12+ so it's possible it already meets your requirements.
> >
> > Using Alan's idea to mock up a latency test, I threw together a test jig
> > using two computers running 3.14-rc1 and my fwserial driver (modified to
> > not aggregrate writes) in raw mode where the target does this:
>
> This is a complete pointless test. Use a bog standard 8250 UART on the
> PC and connect a microcontroller on the other end which serves you an
> continous stream of data at 115200 Baud.
>
> There is no way you can keep up with that without the low latency
> option neither on old and nor on new machines if you have enough other
> stuff going on in the system.

I'm not sure exactly what you mean by 'keep up'?
115kbaud is 11.25KB/sec which is a trivial workload (unless you're
using a 1-byte read buffer).

If you have enough other stuff going on in the system (hackbench?),
even the low_latency knob won't fix the inability to keep up because all
the buffering will fill up and overrun anyway.

So what I need to understand about your setup is:
a) is throughput the actual problem or is latency? IOW, does the
   device have a minimum response time from a user-space process
   or is buffered data getting dropped?
b) is the device flow-controlled or is that not an option?

Based on those answers, if necessary, I could get you an instrumentation
patch, if your willing, so I can profile where the problem is.

And I haven't seen a bog standard 8250 UART in 3 decades. What UART
do you actually have?

Regards,
Peter Hurley



--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/