<AWT Dev> [8u-backport] RFR: 8166673: The new implementation of Robot.waitForIdle() may hang

classic Classic list List threaded Threaded
5 messages Options
Reply | Threaded
Open this post in threaded view
|

<AWT Dev> [8u-backport] RFR: 8166673: The new implementation of Robot.waitForIdle() may hang

Fairoz Matte
Hi,

Please review the backport of "JDK-8166673: The new implementation of Robot.waitForIdle() may hang" to 8u
Changes apply cleanly, there was a typo error in test case name "WaitForIdleSyncroizedOnString.java" it has been corrected to "WaitForIdleSynchronizedOnString.java"
As the test case doesn't have any impact on JDK8, it passes before the patch.
I have verified the issue by executing test case found in one of the duplicate issue JDK-8194862

Webrev - http://cr.openjdk.java.net/~fmatte/8166673/webrev.00/ 
JBS bug - https://bugs.openjdk.java.net/browse/JDK-8166673
JDK9 changeset - http://hg.openjdk.java.net/jdk9/client/jdk/rev/fec7a8a6a46a
JDK9 review thread - http://mail.openjdk.java.net/pipermail/awt-dev/2016-October/012114.html

Thanks,
Fairoz
Reply | Threaded
Open this post in threaded view
|

Re: <AWT Dev> [8u-backport] RFR: 8166673: The new implementation of Robot.waitForIdle() may hang

Sergey Bylokhov
Hi, Fairoz.

On 28/09/2018 08:32, Fairoz Matte wrote:
> As the test case doesn't have any impact on JDK8, it passes before the patch.

The tests works before the fix because in jdk8 the code which is updated
in SunToolkit is not used by the Robot class, so it is unclear how the
application triggers the bug.

> I have verified the issue by executing test case found in one of the duplicate issue JDK-8194862

Can you please add a stacktrace of the deadlock to the bug descripton?

>
> Webrev - http://cr.openjdk.java.net/~fmatte/8166673/webrev.00/
> JBS bug - https://bugs.openjdk.java.net/browse/JDK-8166673
> JDK9 changeset - http://hg.openjdk.java.net/jdk9/client/jdk/rev/fec7a8a6a46a
> JDK9 review thread - http://mail.openjdk.java.net/pipermail/awt-dev/2016-October/012114.html
>
> Thanks,
> Fairoz
>


--
Best regards, Sergey.
Reply | Threaded
Open this post in threaded view
|

Re: <AWT Dev> [8u-backport] RFR: 8166673: The new implementation of Robot.waitForIdle() may hang

Fairoz Matte
Hi Sergey,

Thanks for looking into this.

> -----Original Message-----
> From: Sergey Bylokhov
> Sent: Saturday, September 29, 2018 12:52 AM
> To: Fairoz Matte <[hidden email]>; [hidden email]
> Subject: Re: <AWT Dev> [8u-backport] RFR: 8166673: The new
> implementation of Robot.waitForIdle() may hang
>
> Hi, Fairoz.
>
> On 28/09/2018 08:32, Fairoz Matte wrote:
> > As the test case doesn't have any impact on JDK8, it passes before the
> patch.
>
> The tests works before the fix because in jdk8 the code which is updated in
> SunToolkit is not used by the Robot class, so it is unclear how the application
> triggers the bug.
>
> > I have verified the issue by executing test case found in one of the
> > duplicate issue JDK-8194862
>
> Can you please add a stacktrace of the deadlock to the bug descripton?
>

I have updated bug (JDK-8194862) with stacktraces of the threads causing deadlock.

Thanks,
Fairoz

> >
> > Webrev - http://cr.openjdk.java.net/~fmatte/8166673/webrev.00/
> > JBS bug - https://bugs.openjdk.java.net/browse/JDK-8166673
> > JDK9 changeset -
> > http://hg.openjdk.java.net/jdk9/client/jdk/rev/fec7a8a6a46a
> > JDK9 review thread -
> > http://mail.openjdk.java.net/pipermail/awt-dev/2016-
> October/012114.htm
> > l
> >
> > Thanks,
> > Fairoz
> >
>
>
> --
> Best regards, Sergey.
Reply | Threaded
Open this post in threaded view
|

Re: <AWT Dev> [8u-backport] RFR: 8166673: The new implementation of Robot.waitForIdle() may hang

Sergey Bylokhov
Hi, Fairoz.

On 28/09/2018 21:29, Fairoz Matte wrote:
> I have updated bug (JDK-8194862) with stacktraces of the threads causing deadlock.

The stacktrace shows the deadlock between Component$AWTTreeLock and
AquaFileSystemModel$fileCacheLock. I am not sure that the changes in the
SunToolkit are related to this deadlock, and I think that JDK-8194862 is
not a duplicate of JDK-8166673


--
Best regards, Sergey.
Reply | Threaded
Open this post in threaded view
|

Re: <AWT Dev> [8u-backport] RFR: 8166673: The new implementation of Robot.waitForIdle() may hang

Fairoz Matte
Hi Sergey,

Thanks for the feedback,
Let me reinvestigate JDK-8194862 issue further.

Thanks,
Fairoz

> -----Original Message-----
> From: Sergey Bylokhov
> Sent: Monday, October 01, 2018 6:50 AM
> To: Fairoz Matte <[hidden email]>; [hidden email]
> Subject: Re: <AWT Dev> [8u-backport] RFR: 8166673: The new
> implementation of Robot.waitForIdle() may hang
>
> Hi, Fairoz.
>
> On 28/09/2018 21:29, Fairoz Matte wrote:
> > I have updated bug (JDK-8194862) with stacktraces of the threads causing
> deadlock.
>
> The stacktrace shows the deadlock between Component$AWTTreeLock and
> AquaFileSystemModel$fileCacheLock. I am not sure that the changes in the
> SunToolkit are related to this deadlock, and I think that JDK-8194862 is not a
> duplicate of JDK-8166673
>
>
> --
> Best regards, Sergey.