[PATCH v1] mailbox: cmdq: add instruction time-out interrupt support
Chun-Kuang Hu
chunkuang.hu at kernel.org
Fri Jul 16 09:42:43 PDT 2021
Hi, Yongqiang:
Yongqiang Niu <yongqiang.niu at mediatek.com> 於 2021年7月16日 週五 下午2:58寫道:
>
> add time-out cycle setting to make sure time-out interrupt irq
> will happened when instruction time-out for wait and poll
>
> Signed-off-by: Yongqiang Niu <yongqiang.niu at mediatek.com>
> ---
> drivers/mailbox/mtk-cmdq-mailbox.c | 11 +++++++++++
> 1 file changed, 11 insertions(+)
>
> diff --git a/drivers/mailbox/mtk-cmdq-mailbox.c b/drivers/mailbox/mtk-cmdq-mailbox.c
> index de4793e..9a76bcd 100644
> --- a/drivers/mailbox/mtk-cmdq-mailbox.c
> +++ b/drivers/mailbox/mtk-cmdq-mailbox.c
> @@ -35,6 +35,7 @@
> #define CMDQ_THR_END_ADDR 0x24
> #define CMDQ_THR_WAIT_TOKEN 0x30
> #define CMDQ_THR_PRIORITY 0x40
> +#define CMDQ_THR_INSTN_TIMEOUT_CYCLES 0x50
>
> #define GCE_GCTL_VALUE 0x48
>
> @@ -53,6 +54,15 @@
> #define CMDQ_JUMP_BY_OFFSET 0x10000000
> #define CMDQ_JUMP_BY_PA 0x10000001
>
> +/*
> + * instruction time-out
> + * cycles to issue instruction time-out interrupt for wait and poll instructions
> + * GCE axi_clock 156MHz
> + * 1 cycle = 6.41ns
> + * instruction time out 2^22*2*6.41ns = 53ms
I think every client has different timeout value, so it's not a good
idea to have a unique timeout value in mailbox controller. Client
could use timer or something similar to detect timeout.
Regards,
Chun-Kuang.
> + */
> +#define CMDQ_INSTN_TIMEOUT_CYCLES 22
> +
> struct cmdq_thread {
> struct mbox_chan *chan;
> void __iomem *base;
> @@ -368,6 +378,7 @@ static int cmdq_mbox_send_data(struct mbox_chan *chan, void *data)
> writel((task->pa_base + pkt->cmd_buf_size) >> cmdq->shift_pa,
> thread->base + CMDQ_THR_END_ADDR);
>
> + writel(CMDQ_INSTN_TIMEOUT_CYCLES, thread->base + CMDQ_THR_INSTN_TIMEOUT_CYCLES);
> writel(thread->priority, thread->base + CMDQ_THR_PRIORITY);
> writel(CMDQ_THR_IRQ_EN, thread->base + CMDQ_THR_IRQ_ENABLE);
> writel(CMDQ_THR_ENABLED, thread->base + CMDQ_THR_ENABLE_TASK);
> --
> 1.8.1.1.dirty
>
More information about the Linux-mediatek
mailing list