Closed GoogleCodeExporter closed 9 years ago
First of all, thanks for the report.
Unfortunately, I haven't been able to reconstruct this bug with my installation
of
Windows XP SP3, so I'd like you to do the following:
Please install a debug version (see attachment) and send me the debug log if
it's
not too much trouble. For that, download DebugView
(http://www.microsoft.com/technet/
sysinternals/utilities/debugview.mspx) and start it before you update to the
debug
version of the driver. Then try to install the driver (you can use setup.exe),
and
after that, start cmicontrol, press "Apply", close it and do a standby / normal
operation cycle. Then restart cmicontrol and press "Apply". That should
hopefully
give me a clue on what went wrong.
There's no hurry, so take all the time you need.
Original comment by dogb...@gmail.com
on 19 Mar 2009 at 6:21
Attachments:
I installed debuView, started the program then updated your new debug build of
the
driver. It asked me to restart which I did. I noticed that nothing was
generated in
Debug View though. After restart I started DebugView again, and then the
control
panel. Noticed the version was now 1.2.4-debug.
Did an apply/close cycle, and noticed that still nothing in DebugView. I
enabled the
option for "Capture Kernel" which was not enabled by default, and did another
apply/close cycle in the control panel. This time it generated content. I
assume
this is what you wanted.
So I did a standby/resume cycle. Launched the control panel again, and it
showed the
incorrect settings / problem I reported. Did another apply/close and verfied
DebugView showed updated content. I saved this log and attached it here.
Hopefully
this is what you needed, let me know if I need to do anything else! Thanks!
Original comment by protonus.ws@gmail.com
on 19 Mar 2009 at 9:11
Attachments:
Thanks for the log. Superficially, everything looks alright.
Does the computer account you are using have administrator rights?
After a standby cycle, can you please try to disable and enable the device in
the
device manager while running DebugView?
Original comment by dogb...@gmail.com
on 19 Mar 2009 at 9:35
Yes, I am logging in as an administrator user.
I created two new logs for you. The first, log2.log, was me rebooting the PC
to fix
the control panel and then setting the values normally and hitting ok. Then I
started the DebugView. I then reopened the control panel and hit apply/ok.
Then I
went into standby. Then resumed. Then went back into the control panel and
verified
the settings were corrupt as usual. I hit apply/ok again. By this point, this
is the
same thing that you had me try before, except I noticed it had generated a lot
more
entries this time. After this, I then went into device manager, and hit
disable on
the device except I accidently clicked cancel instead of OK. I did it again
and hit
ok. It asked for a reboot but I did NOT reboot. I then Enabled the device
again.
Again I did NOT reboot. I closed device manager, and saved the log as log2.log.
The second new log, log3.log, I did right after the aforementioned one.
Without a
reboot or doing anything else, I started DebugView again, and went into Device
Manager, and disabled the device. I did NOT reboot when asked. I then enabled
the
device, and again, did not reboot. I then closed device manager and saved it as
log3.log. I think that is all you asked for me to do in this second log. But I
included the first one too as hopefully it may contain more info then the
original
log. Hope this helps!
Original comment by protonus.ws@gmail.com
on 20 Mar 2009 at 2:46
Attachments:
Thank you very much for the logs - that has finally given me the right idea.
My guess is that after a standby cycle, access to the PCI I/O is somehow
blocked
(0xFF values in the registers - every bit is '1'). If you disable and enable
the
driver manually, the I/O stuff should be completely reinitialized and possibly
prove
that this might be a workaround for this bug.
However, when you are asked for a reboot after disabling the device, some app
seems
to be using the soundcard and therefore blocking the driver from unloading. Can
you
try to identify this app and close it before trying to unload the driver?
Again, thanks for your effort - I'm happy about every bug I can squash.
Original comment by dogb...@gmail.com
on 20 Mar 2009 at 10:46
Any news so far?
Original comment by dogb...@gmail.com
on 26 Mar 2009 at 8:31
Hi, sorry it took so long just had a busy week here. I did the following:
1. Verified I had audio / normal ACP (i'll start abbreviating Audio Control
Panel)
settings, after normal boot.
2. I closed all apps and processes that were non essential to Windows. Very
clean -
18 processes total.
3. I launched debug view to start a log.
4. I put the computer into standby and then resumed it.
5. Verified after resume, per usual I had no Audio and ACP settings were whacky
again
like "usual". Verified changing them didn't "stick" per usual.
6. I went into device manager and disabled the audio device. Did NOT restart
like it
asked. I then enabled it again and closed device manager etc.
7. I went back into the ACP - the settings were still whacky, and changing them
did
not "stick" still after applied...
8. I closed and saved the log i'm attaching here...
So it seems like disabling/re-enabling the device even with all non-essential
programs/processes closed, isn't working as a work around. I'm rather
confident I
had no applications or processes running that would be using audio or affecting
this.
This PC is rather clean/barebones too. But perhaps this log will show something
new/useful? Let me know if you need me to do more testing! Thanks!
Original comment by protonus.ws@gmail.com
on 27 Mar 2009 at 6:34
Attachments:
Here is a log from process explorer showing all running processes I have open
at the
time of this testing - only 17 processes total. Very clean I think. The SVC
hosts
are just running normal Windows XP Services as far as I see.
Original comment by protonus.ws@gmail.com
on 27 Mar 2009 at 6:40
Attachments:
Thanks again for the detailed log.
You are right: I can't make out any process which might use the soundcard and
block
the driver from unloading, so there has to be some other cause.
I took another dive into the source code: the C-Media hardware is pretty old-
fashioned, and its registers aren't memory mapped like most other PCI devices,
but
accessible only over I/O. This is pretty much straightforward and doesn't
require
initialization which suggests that the blocking of the I/O access after a
standby
cycle is caused either by a buggy BIOS or a buggy chipset driver.
The only advise I can give you for now is to try some different versions of the
chipset driver and to fiddle around with the BIOS settings and/or versions. In
order
to test whether this has an effect or not, I suggest employing tools which
allow
direct access to the PCI I/O registers from user mode, for instance WPCREDIT.
The
vendor/device ID of the C-Media device should be 13F6:0111 or 13F6:0110. If you
read
out the PCI registers of the C-Media device after a standby cycle and most of
them
are 0xFF, then the hardware is inaccessible.
Good luck!
Original comment by dogb...@gmail.com
on 28 Mar 2009 at 5:30
Original comment by dogb...@gmail.com
on 11 Sep 2010 at 6:07
I ended up getting a new HTPC, and throwing a video card in it that does audio
over HDMI, so this is no longer an issue for me as I'm no longer using that
PC/sound card/this driver. Thanks.
Original comment by protonus.ws@gmail.com
on 13 Sep 2010 at 3:30
Original issue reported on code.google.com by
protonus.ws@gmail.com
on 15 Mar 2009 at 9:26Attachments: