> I have heard reports that SCHED_FIFO has several bugs.
I test out SCHED_FIFO code under 2.2.17 daily, using buffer
latencies in the 3ms range and doing work that isn't disk
intensive (i.e. sfront networking development, exercising
audio I/O, UDP send/receive, and moderately demanding
floating-point CPU usage). And it works really well for me:
I can switch from being root to being a non-root, and hear
night-and-day differences in dropouts.
So, SCHED_FIFO wasn't always broken -- I haven't kept up
with the details of how it broke under 2.4 and whether or
not its truly fixed, but if your goal is to just get a
system up where SCHED_FIFO is known to work in order to
do initial code debugging, 2.2.17 is known to work ...
--jl
-------------------------------------------------------------------------
John Lazzaro -- Research Specialist -- CS Division -- EECS -- UC Berkeley
lazzaro [at] cs [dot] berkeley [dot] edu www.cs.berkeley.edu/~lazzaro
-------------------------------------------------------------------------