Re: [Qemu-arm] [PATCH 5/5] aspeed/timer: Use signed muldiv for timer resets

[Christian Svensson]

Hi all,

I have a new patch but I'm not sure how you want me to post it.

Should I do a "PATCH v2" with a single patch and this thread as the thread ID?

Thanks,

- Chris

On Thu, Mar 14, 2019 at 10:05 AM Cédric Le Goater <address@hidden> wrote:

Christian,

Could you please provide a fix for this patch ? patchew complains, see
attached log.

Thanks,

C.

On 3/14/19 9:42 AM, Cédric Le Goater wrote:
> From: Christian Svensson <address@hidden>
>
> If the host decrements the counter register that results in a negative
> delta. This is then passed to muldiv64 which only handles unsigned
> numbers resulting in bogus results.
>
> This fix ensures the data being operated on is signed before it is
> ultimately casted to the final unsigned value.
>
> Test case: kexec a kernel using aspeed_timer and it will freeze on the
> second bootup when the kernel initializes the timer. With this patch
> that no longer happens and the timer appears to run OK.
>
> Signed-off-by: Christian Svensson <address@hidden>
> [clg: - checkpatch fixes ]
> Signed-off-by: Cédric Le Goater <address@hidden>
> ---
>  hw/timer/aspeed_timer.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/hw/timer/aspeed_timer.c b/hw/timer/aspeed_timer.c
> index 9988b8fbbf17..0b16eac8970c 100644
> --- a/hw/timer/aspeed_timer.c
> +++ b/hw/timer/aspeed_timer.c
> @@ -275,7 +275,8 @@ static void aspeed_timer_set_value(AspeedTimerCtrlState *s, int timer, int reg,
>              int64_t delta = (int64_t) value - (int64_t) calculate_ticks(t, now);
>              uint32_t rate = calculate_rate(t);
> 
> -            t->start += muldiv64(delta, NANOSECONDS_PER_SECOND, rate);
> +            t->start = (int64_t)t->start +
> +                ((__int128_t)delta * NANOSECONDS_PER_SECOND / rate);
>              aspeed_timer_mod(t);
>          }
>          break;
>