xen-devel.lists.xenproject.org archive mirror
 help / color / mirror / Atom feed
From: Boris Ostrovsky <boris.ostrovsky@oracle.com>
To: shuo.a.liu@intel.com, xen-devel@lists.xenproject.org
Cc: Juergen Gross <jgross@suse.com>
Subject: Re: Why xen-pirq chip use startup_irq() for .irq_enable?
Date: Thu, 27 Jul 2017 12:06:10 -0400	[thread overview]
Message-ID: <68e1206c-219b-85d8-15e8-cc591158207d@oracle.com> (raw)
In-Reply-To: <20170727130414.GA8141@shuo-intel.sh.intel.com>

(Adjusting addressees: David is no longer maintaining Xen code, Juergen is)

On 07/27/2017 09:04 AM, shuo.a.liu@intel.com wrote:
> Hi,
> Here is a device has xen-pirq-MSI interrupt. I found dom0 might lost
> interrupt during driver irq_disable/irq_enable.
> There is a pair of irq_disable/enable in driver. Here is the scenario,
>  1. irq_disable(dev_irq) -> disable_dynirq -> mask_evtchn(dev_irq
> channel)
>  2. dev interrupt raised by HW and Xen mark its evtchn as *pending*
> status.
>  3. irq_enable(dev_irq) -> startup_pirq -> eoi_pirq ->
>     clear_evtchn(channel of dev_irq) -> clear *pending* status
>  4. consume_one_event process the dev irq event without pending bit
> assert
>     which result in interrupt lost once.
>  5. No HW interrupt raising anymore.
>
> The first question here is why using startup_irq for .irq_enable
> rather than
> enable_dynirq ? startup_irq will do eoi_pirq who clear the mask bit
> and pending
> bit of the channel while enable_dynirq just only unmask the channel.

Seems like enable_dynirq() would indeed be the right choice. What is a
bit strange is that scenario that you are describing looks pretty common
so we should have hit this problem before.

>
> Second question is that what's the purpose of eoi_pirq in startup_irq?

When we are actually creating new pirq we want to make sure there are no
pending interrupts left over from previous use of the pirq.

-boris

>
> BTW, i can resolve my problem by below patch. Does it make sence?
>
> ---
> drivers/xen/events/events_base.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/xen/events/events_base.c
> b/drivers/xen/events/events_base.c
> index 4bf7a34..341c456 100644
> --- a/drivers/xen/events/events_base.c
> +++ b/drivers/xen/events/events_base.c
> @@ -582,7 +582,7 @@ static void shutdown_pirq(struct irq_data *data)
>
> static void enable_pirq(struct irq_data *data)
> {
> -    startup_pirq(data);
> +    enable_dynirq(data);
> }
>
> static void disable_pirq(struct irq_data *data)


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
https://lists.xen.org/xen-devel

  reply	other threads:[~2017-07-27 16:04 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-07-27 13:04 Why xen-pirq chip use startup_irq() for .irq_enable? shuo.a.liu
2017-07-27 16:06 ` Boris Ostrovsky [this message]
2017-07-28  1:25   ` shuo.a.liu
2017-07-28 21:55     ` Boris Ostrovsky
2017-07-29 15:32       ` shuo.a.liu

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=68e1206c-219b-85d8-15e8-cc591158207d@oracle.com \
    --to=boris.ostrovsky@oracle.com \
    --cc=jgross@suse.com \
    --cc=shuo.a.liu@intel.com \
    --cc=xen-devel@lists.xenproject.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).