On 12/12/2014 16:49, David Daney wrote: > On 12/12/2014 01:31 PM, Joshua Kinard wrote: >> Change the R4K clockevent device's name from "MIPS" to something a bit more >> meaningful, "CEVT-R4K". >> > > This is visible to userspace, so how does changing this effect the kernel <--> > userspace ABI? Or in other words, what uses this, and could changing it > possibly break things? > > David Daney I haven't noticed any issues in userspace running this change on the Octane. R4k is the only clockevent_device using the name of "MIPS", so if there is some userspace program relying on that, I would imagine it'd fail on the other clockevent_device providers as they use different names. I suspect that R4k was the original clockevent_device/timer, so it got defaulted to "MIPS". Since there's several providers now, differentiating R4K seems logical. --J >> Signed-off-by: Joshua Kinard <kumba@xxxxxxxxxx> >> --- >> arch/mips/kernel/cevt-r4k.c | 2 +- >> 1 file changed, 1 insertion(+), 1 deletion(-) >> >> diff --git a/arch/mips/kernel/cevt-r4k.c b/arch/mips/kernel/cevt-r4k.c >> index bc127e2..f531cac 100644 >> --- a/arch/mips/kernel/cevt-r4k.c >> +++ b/arch/mips/kernel/cevt-r4k.c >> @@ -178,7 +178,7 @@ int r4k_clockevent_init(void) >> >> cd = &per_cpu(mips_clockevent_device, cpu); >> >> - cd->name = "MIPS"; >> + cd->name = "CEVT-R4K"; >> cd->features = CLOCK_EVT_FEAT_ONESHOT | >> CLOCK_EVT_FEAT_C3STOP | >> CLOCK_EVT_FEAT_PERCPU; >>
