Fix for SMB shares after suspend #2895

verybadsoldier · 2013-06-19T20:24:56Z

It basically is supposed to do this:

-deinitialize libsmbclient before going to sleep (Linux only)
-start and stop all addon services before and after suspend (otherwise addons are able to reinitialize SMB right before actually going to sleep and break it)

More details here:
http://forum.xbmc.org/showthread.php?tid=166430

Memphiz · 2013-06-19T20:29:45Z

jenkins test this please

bulkzooi · 2013-06-19T23:53:59Z

"Good to merge" or "won't break build"? But I like this PR. Hopefully NFS isn't forgotten.

Memphiz · 2013-06-20T05:46:05Z

It didn't built yet - i have still problems with Jenkins - so basically this is a false status - still to be ignored - and if it was for real it would just indicate that it compiled on all platforms...

xbmc/powermanagement/PowerManager.cpp

  CAEFactory::Suspend();
 }

+void CPowerManager::WaitForNet()


Memphiz · 2013-06-20T13:31:16Z

delay test spam - jenkins build this please

Memphiz · 2013-06-20T14:36:19Z

@team

For clarification "Good to merge" doesn't mean this PR is approved and cleared for merging. Its something the githubapi sets for us and we are not able to change that message atm. So feel informed about that

verybadsoldier · 2013-06-21T19:07:00Z

Well I cannot offer complete failsafe solutions to the issues mentioned above regarding wait-for-net. Please help me out by providing ideas how you think the problems should be solved in a failsafe way.

I would proprose that I remove the wait-for-net functionality from this PR and just stick with the SMB-Deinit stuff. Those are two totally independent features anyway.

PS.
I did not forget NFS its just that I am not using NFS myself and I dont have a NFS server around. Once this stuff is done I might take a look at it though.

t4-ravenbird · 2013-06-22T15:08:54Z

I think removing the wait-for-net (or splitting into 2 separate commits) will be a good idea, yes.

I am also not a team-member but I raised the issues because I feel the wait-for-net part overlaps with the recently added wake-on-access feature ;
If wake-on-access is enabled then there IS already a wait-for-net performed and it does the waiting on behalf on any calling-thread and wait for the appropriate nic.
If wake-on-access is disabled, however, this wait is bypassed. An easy solution to would be to allow the wake-on-access function to wait for network even if wol-transmission is not enabled. That is in effect the same as in your wait-for-net and without the mentioned pitfalls.
I suspect however, that this must go in first ; #2770 It adds a function to determine if a host is on lan and would be needed to know when to perform a wait.

verybadsoldier · 2013-06-22T19:06:04Z

Oh fine, if the wait-for-net stuff is already there then its even better. I removed it from this PR so this is now just about deinitializing libsmbclient before going to suspend.

xhaggi · 2013-06-25T18:50:28Z

a nice sideeffect of stopping/starting the addons, the pvr addon now imports the epg data after suspend. without this the epg-data gets out of syn after a few days and you have to manually reset the epg data or restart xbmc.

xhaggi · 2013-07-05T15:27:57Z

@Memphiz would it make more sense if we split the start/stop addon stuff into a separated pull, so we can merge this in before?

Memphiz · 2013-07-05T20:36:18Z

yeah sounds good - those are 2 unrelated things imo...

verybadsoldier · 2013-07-05T20:49:49Z

The SMB-fix wont work properly without stopping the addons because a thread in a addon might re-init SMB just after it got deinited but before the system actually goes to suspend. I had this case when running the Library Watchdog addon.

xhaggi · 2013-07-05T22:09:18Z

@verybadsoldier could you please split this in logical commits one for the addon stuff and a second for the smb.

verybadsoldier · 2013-07-06T11:43:48Z

Sure, no problem. The splitted PR is here: #2946

Memphiz · 2013-07-07T19:39:43Z

Could you cherry-pick that one ontop of yours?

Memphiz@ddcfb39

If you have any chance for testing it with afp, nfs or sftp - this would be very welcome (i have no clean suspend on my rigs...). Else i would let that go in untested and wait for complaints on from users...

verybadsoldier · 2013-07-14T18:38:01Z

Ok, done. Sorry, it seems I missed your comment before.

But I cant test the other protocols either. NFS was tested by some users already because I applied the NFS-patch to some cusom OpenELEC builds already.

wsnipex · 2013-07-21T13:15:26Z

jenkins test this please

pretty pleeeease :)

wsnipex · 2013-07-21T13:16:45Z

narf, sorry for the spam, wrong trigger.

jenkins build this please

Memphiz · 2013-07-21T13:16:50Z

Wrong trigger... Build vs. Test

MartijnKaijser · 2013-08-07T06:28:26Z

@Memphiz
go for merge?

Memphiz · 2013-08-07T06:39:31Z

@verybadsoldier - is this the currently wanted implementation of this? I am a Bit confused with the forum thread where a different (or the Same?) Patch is discussed in conjunction with openelec... If it is the same i am Fine with merging this due to the positive feedback from oe users...

verybadsoldier · 2013-08-07T07:04:45Z

Yes its basically the same in OE. Different in the OE patch is:
-it also incorporates this PR #2946
-it does not have the SFTP and AFP part
-it has the "wait-for-net"-feature which delays the XBMC wake up (GUI thread) until the first NIC is up. It was said that a different/better mechanism is in XBMC already. This is not related to network shares but it is meant to fix access to MySQL straight after wakeup.

Fix for SMB shares after suspend

xbmc/Application.cpp

+{
+  CLog::Log(LOGDEBUG,"CApplication::CloseNetworkShares: Closing all network shares");
+
+#if defined(HAS_FILESYSTEM_SMB) && !defined(_WIN32)


t4-ravenbird reviewed Jun 20, 2013
View reviewed changes

xbmc/powermanagement/PowerManager.cpp

CAEFactory::Suspend();

}

void CPowerManager::WaitForNet()

This comment was marked as spam.

Sign in to view

deinitialize SMB before suspend

ba8bd17

deinitialize NFS,AFP and SFTP before suspend

54d194c

Memphiz added a commit that referenced this pull request Aug 7, 2013

Merge pull request #2895 from verybadsoldier/master

19a968e

Fix for SMB shares after suspend

Memphiz merged commit 19a968e into xbmc:master Aug 7, 2013

Karlson2k reviewed Aug 7, 2013
View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Fix for SMB shares after suspend #2895

Fix for SMB shares after suspend #2895

verybadsoldier commented Jun 19, 2013

Memphiz commented Jun 19, 2013

bulkzooi commented Jun 19, 2013

Memphiz commented Jun 20, 2013

This comment was marked as spam.

Memphiz commented Jun 20, 2013

Memphiz commented Jun 20, 2013

verybadsoldier commented Jun 21, 2013

t4-ravenbird commented Jun 22, 2013

verybadsoldier commented Jun 22, 2013

xhaggi commented Jun 25, 2013

xhaggi commented Jul 5, 2013

Memphiz commented Jul 5, 2013

verybadsoldier commented Jul 5, 2013

xhaggi commented Jul 5, 2013

verybadsoldier commented Jul 6, 2013

Memphiz commented Jul 7, 2013

verybadsoldier commented Jul 14, 2013

wsnipex commented Jul 21, 2013

wsnipex commented Jul 21, 2013

Memphiz commented Jul 21, 2013

MartijnKaijser commented Aug 7, 2013

Memphiz commented Aug 7, 2013

verybadsoldier commented Aug 7, 2013

This comment was marked as spam.

This comment was marked as spam.

This comment was marked as spam.

This comment was marked as spam.

This comment was marked as spam.

Fix for SMB shares after suspend #2895

Fix for SMB shares after suspend #2895

Conversation

verybadsoldier commented Jun 19, 2013

Memphiz commented Jun 19, 2013

bulkzooi commented Jun 19, 2013

Memphiz commented Jun 20, 2013

This comment was marked as spam.

Memphiz commented Jun 20, 2013

Memphiz commented Jun 20, 2013

verybadsoldier commented Jun 21, 2013

t4-ravenbird commented Jun 22, 2013

verybadsoldier commented Jun 22, 2013

xhaggi commented Jun 25, 2013

xhaggi commented Jul 5, 2013

Memphiz commented Jul 5, 2013

verybadsoldier commented Jul 5, 2013

xhaggi commented Jul 5, 2013

verybadsoldier commented Jul 6, 2013

Memphiz commented Jul 7, 2013

verybadsoldier commented Jul 14, 2013

wsnipex commented Jul 21, 2013

wsnipex commented Jul 21, 2013

Memphiz commented Jul 21, 2013

MartijnKaijser commented Aug 7, 2013

Memphiz commented Aug 7, 2013

verybadsoldier commented Aug 7, 2013

This comment was marked as spam.

This comment was marked as spam.

This comment was marked as spam.

This comment was marked as spam.

This comment was marked as spam.