g2gps
g2gps
Ok, thanks for the suggestions. I'll put together a PR early next week, when I have a chance to test this out properly.
> ping @g2gps I was hoping for some feedback from @xiaoxiang781216 in regards to my comment above, before continuing with the PR.
> > > ping @g2gps > > > > > > I was hoping for some feedback from @xiaoxiang781216 in regards to my comment above, before continuing with the PR....
A couple of questions have come up in the details whilst revisiting this: - The `vexriscv_smp` core doesn't implement the performance counter by default. Opensbi handles this case on startup...
> Let chip developer add riscv_perf.c to their Make.defs if they want to use the common implementation. Yes, that would work for the general implementation. However, this won't work for...
> How about we provide riscv_perf_cycle.c and riscv_perf_time.c? Let's chip developer add one of these as needed. This makes sense. I've updated as suggested. I've testing the CSR_TIME implementation on...
I did a quick test with `rv-virt:nsh`, by adding ` riscv_perf_cycle.c` and enabling `CONFIG_SCHED_IRQMONITOR`, `CONFIG_ARCH_PERF_EVENTS` ``` nsh> cat /proc/irqs IRQ HANDLER ARGUMENT COUNT RATE TIME 11 80005f16 00000000 2 4.385...
> need fix this ci error: > > ``` > Normalize arty_a7/knsh-tickless > 67d66 > < CONFIG_PERF_OVERFLOW_CORRECTION=y > ``` Thanks, SYSTEM_TIME64 should have been enabled here as well. Since this...
I set up an example program to look at seek's behaviour: ```c #ifdef __NuttX__ #include #else #define FAR #endif #include #include #include #include #include #include #include char* vbuffer; FILE *test_open(const...