Media Portal freezes on wake (2 Viewers)

[Sebastiii]

On my side i didn't notice regression with both patch

 

[azzuro]

for me, the second patch was not efficient, and few problem was always there, but it's based on personal sensor.
the first one was more speed on resume.

 

[Sebastiii]

New installer (based on 1.12 and both patch) : http://tinyurl.com/px4wwbh

 

[Sebastiii]

And i forgot to say that 1.13 pre release build is attached to Jira

 

[Sebastiii]

The freezing issue is back with the updated installer. Here are the logs. (Sorry for the delay over the weekend)

Hi,

I will push another commit because on my system, the new patch override the real PBT_APMRESUMESUSPEND sequence.
When looking at your log, effectively, on your system the PBT_APMRESUMESUSPEND is not receive and then MP can't resume GUI etc. (no clue why).

When compared to mine, the Log.Info("Main: User is providing input to the session"); can happen when my PC goes to suspend (no clue why lol because i'm didn't touch anything).

So the commit will try to handle both :

if Log.Info("Main: User is providing input to the session"); -> resume
But keep the correct sequence if the message PBT_APMRESUMESUSPEND is detected.

So please try also with the future installer
Thanks.

 

[Sebastiii]

Here we go for 1.12 installer : http://tinyurl.com/oqdrqjz
Jira has the new 1.13 Pre-release installer.

 

[mm1352000]

When compared to mine, the Log.Info("Main: User is providing input to the session"); can happen when my PC goes to suspend (no clue why lol because i'm didn't touch anything).

Please post the log so I can take a look.

 

[Sebastiii]

Hey,
I was removed all log but keep a copy in notepad lol

So :
@ 13:25 (PC goes to suspend)
[2015-11-11 13:25:34,378] [Log ] [MPMain ] [INFO ] - Main: OnSuspend - Done

But right after (but that's not true i have start suspend from remote so i wasn't present on my HTPC, the strange thing is that the same suspend doesn't give the same result lol)
[2015-11-11 13:25:34,794] [Log ] [MPMain ] [INFO ] - Main: User is providing input to the session

So MP start the new code (OnResumeSuspend(); from // GUIT_SESSION_USER_PRESENCE is only provide on Win8 and above / case 0
[2015-11-11 13:25:34,794] [Log ] [MPMain ] [INFO ] - Main: Providing input - Resuming operation of user interface

After that : the real resume that start at 16:08
[2015-11-11 16:08:07,515] [Log ] [TvClient-TvHome: HeartBeat transmitter thread] [INFO ] - RemoteControl - Connected
[2015-11-11 16:08:26,970] [Log ] [MPMain ] [INFO ] - Main: WM_POWERBROADCAST (PBT_APMRESUMESUSPEND)
Note the diff of 20 seconds (no clue why)

And here PBT_APMRESUMESUSPEND was skipped because already handled when suspend from case 0:
[2015-11-11 16:08:26,970] [Log ] [MPMain ] [INFO ] - Main: PBT_APMRESUMESUSPEND was already handled, skipping
[2015-11-11 16:08:26,986] [Log ] [MPMain ] [INFO ] - Main: WM_POWERBROADCAST (PBT_APMRESUMEAUTOMATIC)
[2015-11-11 16:08:26,986] [Log ] [MPMain ] [INFO ] - Main: PBT_APMRESUMEAUTOMATIC was already handled, skipping
[2015-11-11 16:08:26,986] [Log ] [MPMain ] [INFO ] - Main: WM_POWERBROADCAST (PBT_POWERSETTINGCHANGE)

So what i have done is to let MP try again the resume part of code (even if already executed) but only if already executed (from another test, the case 0: was not executed so it act as normal).
The thing i didn't understand is that @Jeffrios log didn't have the PBT_APMRESUMESUSPEND executed and this is why you put it in Case 0:, it's strange.
On his log it seems to always has the PBT_APMRESUMEAUTOMATIC but not the PBT_APMRESUMESUSPEND lol.

 

[mm1352000]

So what i have done is to let MP try again the resume part of code (even if already executed) but only if already executed (from another test, the case 0: was not executed so it act as normal).

Maybe try to make a change to the first patch. Change this:

if (_suspended && !_resumedSuspended)

...to...

if (_suspended && _resumedAutomatic && !_resumedSuspended)

The thing i didn't understand is that @Jeffrios log didn't have the PBT_APMRESUMESUSPEND executed and this is why you put it in Case 0:, it's strange.
On his log it seems to always has the PBT_APMRESUMEAUTOMATIC but not the PBT_APMRESUMESUSPEND lol.

Yes, exactly. That is the cause of the problem, and the reason why the patch is needed. I thought maybe the message was missing due to a GPU driver problem, but Jeffrios and keeperoftracken have completely different GPUs.

Jeffrios = AMD Radeon HD 8670D, driver date 10/23/2015, W10
keeperoftracken = NVidia ION, driver date 10/11/2015, W10

Perhaps it's related to 3rd party software. I don't know...

 

[HTPCSourcer]

I will need to start adding my logs here as I have one client that causes frequent problems when resuming from deep standby/hibernate.

Unfortunately I have little time for MediaPortal at this moment so this may take a while. What I can say though is that there apear to be GPU driver issues (Intel) under W10. Also I am using redirected paths for databases that point to a server. MP1 has an issue if the server is not available at the expected time. For true testing I would need to ensure that the server is running before waking up the "HTPC1" (see system specs for details).

That said I had tried Seb's previous 1.12-based version, which didn't run well. As we are looking at the client in my living, I need to provide a functional setup to the family, so I upgraded back to 1.13pre. If you guys could provide patches for 1.13, it would be greatly appreciated.