[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-ppc] [PATCHv2 08/12] cuda: factor out timebase-derived counter
From: |
David Gibson |
Subject: |
Re: [Qemu-ppc] [PATCHv2 08/12] cuda: factor out timebase-derived counter value and load time |
Date: |
Sun, 11 Feb 2018 10:18:03 +1100 |
User-agent: |
Mutt/1.9.2 (2017-12-15) |
On Fri, Feb 09, 2018 at 06:51:38PM +0000, Mark Cave-Ayland wrote:
> Commit b981289c49 "PPC: Cuda: Use cuda timer to expose tbfreq to guest"
> altered
> the timer calculations from those based upon the hardware CUDA clock frequency
> to those based upon the CPU timebase frequency.
>
> In fact we can isolate the differences to 2 simple changes: one to the counter
> read value and another to the counter load time. Move these changes into
> separate functions so the implementation can be swapped later.
>
> Signed-off-by: Mark Cave-Ayland <address@hidden>
> Reviewed-by: Philippe Mathieu-Daudé <address@hidden>
Applied, thanks.
> ---
> hw/misc/macio/cuda.c | 25 ++++++++++++++++---------
> 1 file changed, 16 insertions(+), 9 deletions(-)
>
> diff --git a/hw/misc/macio/cuda.c b/hw/misc/macio/cuda.c
> index e00df4a21a..a185252144 100644
> --- a/hw/misc/macio/cuda.c
> +++ b/hw/misc/macio/cuda.c
> @@ -145,21 +145,29 @@ static void cuda_update_irq(CUDAState *s)
> }
> }
>
> -static uint64_t get_tb(uint64_t time, uint64_t freq)
> +static uint64_t get_counter_value(CUDAState *s, CUDATimer *ti)
> {
> - return muldiv64(time, freq, NANOSECONDS_PER_SECOND);
> + /* Reverse of the tb calculation algorithm that Mac OS X uses on bootup
> */
> + uint64_t tb_diff = muldiv64(qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL),
> + s->tb_frequency, NANOSECONDS_PER_SECOND) -
> + ti->load_time;
> +
> + return (tb_diff * 0xBF401675E5DULL) / (s->tb_frequency << 24);
> +}
> +
> +static uint64_t get_counter_load_time(CUDAState *s, CUDATimer *ti)
> +{
> + uint64_t load_time = muldiv64(qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL),
> + s->tb_frequency, NANOSECONDS_PER_SECOND);
> + return load_time;
> }
>
> static unsigned int get_counter(CUDAState *s, CUDATimer *ti)
> {
> int64_t d;
> unsigned int counter;
> - uint64_t tb_diff;
> - uint64_t current_time = qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL);
>
> - /* Reverse of the tb calculation algorithm that Mac OS X uses on bootup.
> */
> - tb_diff = get_tb(current_time, s->tb_frequency) - ti->load_time;
> - d = (tb_diff * 0xBF401675E5DULL) / (s->tb_frequency << 24);
> + d = get_counter_value(s, ti);
>
> if (ti->index == 0) {
> /* the timer goes down from latch to -1 (period of latch + 2) */
> @@ -178,8 +186,7 @@ static unsigned int get_counter(CUDAState *s, CUDATimer
> *ti)
> static void set_counter(CUDAState *s, CUDATimer *ti, unsigned int val)
> {
> CUDA_DPRINTF("T%d.counter=%d\n", 1 + ti->index, val);
> - ti->load_time = get_tb(qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL),
> - s->tb_frequency);
> + ti->load_time = get_counter_load_time(s, ti);
> ti->counter_value = val;
> cuda_timer_update(s, ti, ti->load_time);
> }
--
David Gibson | I'll have my music baroque, and my code
david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_
| _way_ _around_!
http://www.ozlabs.org/~dgibson
signature.asc
Description: PGP signature
- [Qemu-ppc] [PATCHv2 00/12] cuda: various fixes, tidy-ups, and move 6522 to separate device, Mark Cave-Ayland, 2018/02/09
- [Qemu-ppc] [PATCHv2 02/12] cuda: don't allow writes to port output pins, Mark Cave-Ayland, 2018/02/09
- [Qemu-ppc] [PATCHv2 03/12] cuda: don't call cuda_update() when writing to ACR register, Mark Cave-Ayland, 2018/02/09
- [Qemu-ppc] [PATCHv2 07/12] cuda: set timer 1 frequency property to CUDA_TIMER_FREQ, Mark Cave-Ayland, 2018/02/09
- [Qemu-ppc] [PATCHv2 06/12] cuda: minor cosmetic tidy-ups to get_next_irq_time(), Mark Cave-Ayland, 2018/02/09
- [Qemu-ppc] [PATCHv2 08/12] cuda: factor out timebase-derived counter value and load time, Mark Cave-Ayland, 2018/02/09
- Re: [Qemu-ppc] [PATCHv2 08/12] cuda: factor out timebase-derived counter value and load time,
David Gibson <=
- [Qemu-ppc] [PATCHv2 05/12] cuda: rename frequency property to tb_frequency, Mark Cave-Ayland, 2018/02/09
[Qemu-ppc] [PATCHv2 01/12] cuda: do not use old_mmio accesses, Mark Cave-Ayland, 2018/02/09
[Qemu-ppc] [PATCHv2 04/12] cuda: introduce CUDAState parameter to get_counter(), Mark Cave-Ayland, 2018/02/09