On 17 April 2018 at 21:37, Aaron Lindsay <alind...@codeaurora.org> wrote: > pmccntr_read and pmccntr_write contained duplicate code that was already > being handled by pmccntr_sync. Consolidate the duplicated code into two > functions: pmccntr_op_start and pmccntr_op_finish. Add a companion to > c15_ccnt in CPUARMState so that we can simultaneously save both the > architectural register value and the last underlying cycle count - this > ensure time isn't lost and will also allow us to access the 'old' > architectural register value in order to detect overflows in later > patches.
Can we handle the overflow case by setting up a timer that will expire when the counter would overflow? We'll want that anyway to be able to signal an interrupt on overflows. That would avoid the need to have an extra field in the CPU state structure, I think. Since PMCCNTR_EL0 is 64 bits it's not going to overflow very often so the slight overhead of a timer firing is not going to be a problem... thanks -- PMM
