25 Mar
2010
25 Mar
'10
10:03 a.m.
David Olofson wrote:
On Thursday 25 March 2010, at 12.49.31, Ralf Mardorf
<ralf.mardorf@alice-
dsl.net> wrote:
[...]
Thank you for the information. At least when doing test I guess the
limits are untouched.
Ralf
Btw. the
graphics has access to the main memory, unfortunately it's a shared RAM,
OTOH I used HPET so unwanted interrupts because of a shared RAM
shouldn't be the cause, if I do understand the workings of HR timers
correctly.
Interrupts can't save you if bus-master DMA is blocking the CPU's access to
RAM. Whatever IRQ source and kernel interrupt handling code you're using has
no bearing on this. If the interrupts are generated by external or internal
(to the CPU) timers shouldn't matter either. (Even if external IRQs are
delayed while the bus is busy, it doesn't make a difference, as the CPU can't
respond until the bus is free anyway.)
Either way, I don't think that should be a problem, as busmaster DMA is
normally done in short bursts, rather than large blocks. Unless you're pushing
the limits of RT-Linux or RTAI, it should affect nothing but bandwidth from a
practical POV. Then again, perhaps not all hardware is that well behaved...?