On Sat, Sep 26, 2015 at 9:08 AM, Peter Maydell <peter.mayd...@linaro.org> wrote: > On 26 September 2015 at 01:54, mar.krzeminski <mar.krzemin...@gmail.com> > wrote: >> Hello again, >> >> My next question is still related with M3 and A9 board what I want to model. >> This time my peripheral has some interrupts that are connected both to A9 >> processor(gic), >> and M3 processor (nvic). Additionally those interrupts have same number. >> Currently I use only two in my model so I added to my device another >> interrupt that does the same, >> but are connected to different processor, and it seem that works. >> Is there any way that I can do it better - to connect one interrupt source >> to two receivers (A9 and M3)? > > This is what qemu_irq_split() is for. >
Yes this is right. There was the same problem with R5 and A9 double interrupt wiring, and to reduce the verbosity, I implemented automatic IRQ splitting in the qdev layer: https://github.com/Xilinx/qemu/blob/pub/2015.2.plnx/hw/core/qdev.c Line 461. Not sure what others think of this, but it would be nice to just double connect on machine level which is an intuitive API for this behaviour. It would make short work of bringing the zynqmp R5 intc online. Regards, Peter > -- PMM >