From: Daniel Mack [mailto:dan...@zonque.org] > On 06/19/2014 05:07 PM, Felipe Balbi wrote: > > On Wed, Jun 18, 2014 at 11:36:43AM +0200, Daniel Mack wrote: > >> On 06/18/2014 11:32 AM, David Laight wrote: > >>> From: Of Daniel Mack > >>>> Sent: 18 June 2014 10:28 > >>>> To: ba...@ti.com; george.cher...@ti.com; bige...@linutronix.de > >>>> Cc: sebastian.reim...@googlemail.com; linux-usb@vger.kernel.org; Daniel > >>>> Mack > >>>> Subject: [PATCH 2/2] usb: musb: cppi41: fire hrtimer according to > >>>> programmed channel length > >>>> > >>>> The musb/cppi41 code installs a hrtimer to work around DMA completion > >>>> interrupts that have fired too early on AM335x hardware. This timer > >>>> is currently programmed to first fire 140 nanoseconds after the DMA > >>>> completion callback. According to the commit which introduced it > >>>> (a655f481d83, "usb: musb: musb_cppi41: handle pre-mature TX complete > >>>> interrupt"), that value is is considered a 'rule of thumb' that worked > >>>> well with the test case described in the commit log. > >>>> > >>>> Test show, however, that for USB audio devices and much smaller packet > >>>> sizes, the timer has to fire earlier in order to correctly handle the > >>>> audio > >>>> stream. The original test case had output transfer sizes of 1514 bytes, > >>>> and > >>>> a delay of 140 nanoseconds. For audio devices with 24 bytes channel > >>>> size, 3 > >>>> nanoseconds seem to work well. > >>> > >>> You can't really mean nanoseconds? > >> > >> Microseconds of course. Thanks for the heads-up. Fixed locally. > > > > Are you sending another one or want me to fix the commit log when > > applying here ? > > I was actually waiting for feedback on the patch before resending > another version. In particular, I'd like to know whether George's test > case still works with this patch applied. > > If nobody has objections, you fixing the commit log would of course be > easiest, yes :)
IIRC there was a reference to nanoseconds inside the patch as well. Possibly a variable name. David