qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: bzt <bztemail@gmail.com>
To: Peter Maydell <peter.maydell@linaro.org>
Cc: "Andrew Baumann" <Andrew.Baumann@microsoft.com>,
	"Philippe Mathieu-Daudé" <f4bug@amsat.org>,
	qemu-arm <qemu-arm@nongnu.org>,
	"QEMU Developers" <qemu-devel@nongnu.org>
Subject: Re: [Qemu-devel] [PATCH] Added periodic IRQ support for bcm2836_control local timer
Date: Sat, 9 Mar 2019 02:03:13 +0100	[thread overview]
Message-ID: <CADYoBw2XWeDUqBHVJBWxJAzQ4f0imosEeLHHoG9AfHfdavOTrA@mail.gmail.com> (raw)
In-Reply-To: <CAFEAcA-+U75+EQzcpV4kZHw+4xMSbMv3UJ0CV_Q3DVbNx5RbNg@mail.gmail.com>

Hi,

Thanks for your answers. If I don't clear the INTENABLE flag, then the
IRQ would keep firing constantly. This is not how the real hardware
works: it triggers the IRQ once, and then it inhibits. The timer won't
trigger the IRQ again until you acknowledge it by writing the INTFLAG
into the ack register. My solution emulates this behaviour. That's
what the triggered flag was for in my original patch. Should I bring
that flag back or is the current solution acceptable knowing this?

About keeping the INTFLAG on control write that's to avoid an instant
IRQ, but that's OK. I'll modify that.

Thanks,
bzt


On 3/7/19, Peter Maydell <peter.maydell@linaro.org> wrote:
> On Thu, 7 Mar 2019 at 15:57, bzt <bztemail@gmail.com> wrote:
>>
>> Nope. I meant the second patch, sent on 4th Mar, which had all the
>> things fixed you listed in your review.
>>
>> But here's the modification for your latest query. This one controls
>> the timer depending on ENABLE bit. It sets the INTFLAG even if
>> INTENABLE is not set, and only raises the IRQ if both are set.
>
> Thanks. I've listed a couple of minor things below
> which I think are not quite handling the flags right,
> but the rest looks good.
>
>> @@ -78,6 +94,17 @@ static void bcm2836_control_update(BCM2836ControlState
>> *s)
>>          s->fiqsrc[s->route_gpu_fiq] |= (uint32_t)1 << IRQ_GPU;
>>      }
>>
>> +    /* handle THE local timer interrupt for one of the cores' IRQ/FIQ */
>> +    if ((s->local_timer_control & LOCALTIMER_INTENABLE) &&
>> +        (s->local_timer_control & LOCALTIMER_INTFLAG)) {
>> +        if (s->route_localtimer & 4) {
>> +            s->fiqsrc[(s->route_localtimer & 3)] |= (uint32_t)1 <<
>> IRQ_TIMER;
>> +        } else {
>> +            s->irqsrc[(s->route_localtimer & 3)] |= (uint32_t)1 <<
>> IRQ_TIMER;
>> +        }
>> +        s->local_timer_control &= ~LOCALTIMER_INTENABLE;
>
> This shouldn't clear the INTENABLE flag.
>
>> +    }
>> +
>>      for (i = 0; i < BCM2836_NCORES; i++) {
>>          /* handle local timer interrupts for this core */
>>          if (s->timerirqs[i]) {
>> @@ -162,6 +189,55 @@ static void bcm2836_control_set_gpu_fiq(void
>> *opaque, int irq, int level)
>>      bcm2836_control_update(s);
>>  }
>
>> +static void bcm2836_control_local_timer_control(void *opaque, uint32_t
>> val)
>> +{
>> +    BCM2836ControlState *s = opaque;
>> +
>> +    s->local_timer_control = val & ~LOCALTIMER_INTFLAG;
>
> This will clear the LOCALTIMER_INTFLAG if it was already
> set -- you want to preserve it, something like
>    s->local_timer_control = (s->local_timer_control & LOCALTIMER_INTFLAG) |
>      (val & ~LOCALTIMER_INTFLAG);
>
>> +    if (val & LOCALTIMER_ENABLE) {
>> +        bcm2836_control_local_timer_set_next(s);
>> +    } else {
>> +        timer_del(&s->timer);
>> +    }
>> +}
>> +
>> +static void bcm2836_control_local_timer_ack(void *opaque, uint32_t val)
>> +{
>> +    BCM2836ControlState *s = opaque;
>> +
>> +    if (val & LOCALTIMER_INTFLAG) {
>> +        s->local_timer_control |= LOCALTIMER_INTENABLE;
>> +        s->local_timer_control &= ~LOCALTIMER_INTFLAG;
>
> This should just clear the INTFLAG, it doesn't affect INTENABLE.
>
>> +    }
>> +    if ((val & LOCALTIMER_RELOAD) &&
>> +        (s->local_timer_control & LOCALTIMER_ENABLE)) {
>> +            bcm2836_control_local_timer_set_next(s);
>> +    }
>> +}
>
> thanks
> -- PMM
>

  reply	other threads:[~2019-03-09  1:03 UTC|newest]

Thread overview: 16+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-02-26 11:38 [Qemu-devel] [PATCH] Added periodic IRQ support for bcm2836_control local timer bzt
2019-02-26 12:25 ` Peter Maydell
2019-02-27 11:54   ` bzt
2019-02-27 18:30     ` Andrew Baumann
2019-02-28 14:12       ` bzt
2019-03-04 15:55 ` Peter Maydell
2019-03-04 19:27   ` bzt
2019-03-04 20:40     ` bzt
2019-03-05 16:37     ` Peter Maydell
2019-03-07 15:27       ` bzt
2019-03-07 15:43         ` Peter Maydell
2019-03-07 15:57           ` bzt
2019-03-07 16:08             ` Peter Maydell
2019-03-09  1:03               ` bzt [this message]
2019-03-09 14:39                 ` Peter Maydell
2019-03-10 11:02                   ` bzt

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=CADYoBw2XWeDUqBHVJBWxJAzQ4f0imosEeLHHoG9AfHfdavOTrA@mail.gmail.com \
    --to=bztemail@gmail.com \
    --cc=Andrew.Baumann@microsoft.com \
    --cc=f4bug@amsat.org \
    --cc=peter.maydell@linaro.org \
    --cc=qemu-arm@nongnu.org \
    --cc=qemu-devel@nongnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).