On Fri, Jul 29, 2016 at 5:32 PM, Muni Sekhar <munisekharrms@gmail.com> wrote:
Hi All,

I have a doubt regarding the workqueue scheduling.

I am using the workqueue for processing the Rx Interrupt data. I am
calling schedule_work() on receiving the Rx interrupt from hardware.

I calculated the time between calling the schedule_work() and
workqueue task actually getting executed, Here I see many cases of
less than 100 us(It is fairly good).

But sometimes I see 10’s of ms and a lot in the 100’s of ms. I have
seen over 0.5 secs too. I would like to know why does sometimes kernel
takes longer time(in milli seconds) to schedule it? Is there any way
to reduce this time gap?


My code:

static void my_workqueuetask(struct work_struct *work)
{
    printk("In %s() \n",__func__);

    mutex_lock(&bh_mutex);

    //Do something here

    mutex_unlock(&bh_mutex);
}


static irqreturn_t my_irq_handler(int irq, void *dev)
{
        ……;

        if(Rx Interrupt)
             schedule_work(&rx_work);

return IRQ_HANDLED;
}

--
Thanks,
Sekhar


Hi Sekhar

In general, in non real time kernel, there is no way you can make sure your work will be scheduled after N seconds.

Have you done simple statistic calculation how much % it is slower than 100 us, and how much % it is faster than 100 us?

IMHO, if slower % is lower than 1% of overall sample, I think it is still acceptable, but this is up to your judgement

BTW, like other said too, that mutex_lock, have you also measured how long, by average, the lock is taken?

--
regards,

Mulyadi Santosa
Freelance Linux trainer and consultant

blog: the-hydra.blogspot.com
training: mulyaditraining.blogspot.com