Twice now, after installing the delta update to 10.5.6 from 10.5.5 on my Mac Pro (late 2007 version), my machine has completely frozen when the screensaver activates - a frame of the saver remains on the monitors, but the machine is completely dead (can't ssh or screen share into it).
I haven't attributed the problem to a screensaver yet...I'l have to give that a try, but after I put my computer to sleep for a while, it won't wake up. No keyboard events work. Here's part of my console log. As you can see, it's filling rather quickly. I put my computer to sleep around 12:12...so this log has been chugging for a while. The issue occurred after I installed OSX 10.5.6 Update
12/17/08 12:26:50 AM kernel AppleFWOHCI_AsyncReceive::waitForDMA - context not going inactive.
12/17/08 12:26:50 AM kernel AppleFWOHCI_AsyncTransmit::waitForDMA - context not going inactive.
12/17/08 12:26:50 AM kernel AppleFWOHCI_AsyncReceive::waitForDMA - context not going inactive.
12/17/08 12:26:50 AM kernel AppleFWOHCI_AsyncTransmit::waitForDMA - context not going inactive.
12/17/08 12:26:50 AM kernel AppleFWOHCI_AsyncReceive::waitForDMA - context not going inactive.
12/17/08 12:26:50 AM kernel AppleFWOHCI_AsyncTransmit::waitForDMA - context not going inactive.
12/17/08 12:26:50 AM kernel AppleFWOHCI_AsyncReceive::waitForDMA - context not going inactive.
12/17/08 12:26:50 AM kernel AppleFWOHCI_AsyncTransmit::waitForDMA - context not going inactive.
12/17/08 12:26:50 AM kernel AppleFWOHCI_AsyncReceive::waitForDMA - context not going inactive.
12/17/08 12:26:50 AM kernel AppleFWOHCI_AsyncTransmit::waitForDMA - context not going inactive.
12/17/08 12:26:50 AM kernel AppleFWOHCI_AsyncReceive::waitForDMA - context not going inactive.
12/17/08 12:26:50 AM kernel AppleFWOHCI_AsyncTransmit::waitForDMA - context not going inactive.
12/17/08 12:26:51 AM kernel AppleFWOHCI_AsyncReceive::waitForDMA - context not going inactive.
12/17/08 12:26:51 AM kernel AppleFWOHCI_AsyncTransmit::waitForDMA - context not going inactive.
12/17/08 12:26:51 AM kernel AppleFWOHCI_AsyncReceive::waitForDMA - context not going inactive.
12/17/08 12:26:51 AM kernel AppleFWOHCI_AsyncTransmit::waitForDMA - context not going inactive.
12/17/08 12:26:51 AM kernel AppleFWOHCI_AsyncReceive::waitForDMA - context not going inactive.
12/17/08 12:26:51 AM kernel AppleFWOHCI_AsyncTransmit::waitForDMA - context not going inactive.
12/17/08 12:26:51 AM kernel AppleFWOHCI_AsyncReceive::waitForDMA - context not going inactive.
12/17/08 12:26:51 AM kernel AppleFWOHCI_AsyncTransmit::waitForDMA - context not going inactive.
12/17/08 12:26:51 AM kernel AppleFWOHCI_AsyncReceive::waitForDMA - context not going inactive.
12/17/08 12:26:51 AM kernel AppleFWOHCI_AsyncTransmit::waitForDMA - context not going inactive.
12/17/08 12:26:51 AM kernel AppleFWOHCI_AsyncReceive::waitForDMA - context not going inactive.
My machine never sleeps (and the second time the lock-up happened, I'd just been using it not ten minutes earlier [my screensaver kicks in after about five minutes).
I'll check my log when I'm back at my machine to see if it provides any useful diagnostic information.
EDIT - I should clarify. The second time (when I was paying attention to such things), my machine "awoke" from power save fine (actually, faster than it did with 10.5.5) after being on for at least ten hours. I did some work, then walked away. The screensaver kicked in and the machine froze.
So, maybe it's not the first time that the screensaver/power save is activated, but a subsequent activation attempt.
When my screensaver activates, I am unable to get back to the main screen. The screensaver does not freeze--the picture on the screen is not static--but there is no response to a keystroke or moving the mouse. I have to do a hard reset.
I just turned the screensaver off for now. I haven't noticed any other anomalies... yet.
I'll try disabling the screen saver tonight. It seems to always happen whenever I wake up the next morning. But I simply put my computer to sleep. Good thing I'm not using it as my alarm clock right now, I'd a been way late to work the past few days.
I have the same issue when I try and put my mid 2007 iMac to sleep. Display goes black, but the machine doesn't fully power down, I can still hear the drive spinning internally.
system.log has the error "AppleFWOHCI_AsyncTransmit::waitForDMA - context not going inactive." as noted in this thread .
No response to keyboard or mouse, no network connectivity. Only way back in is a full power down n' up.
I am seeing the same issue: screensaver kicks in and I can't get ti out of it other than by a hard reset. Seems to have been fixed on my Macbook by setting the screensaver to never activate - although this feels like somewhat of a bodge. My G5 iMac however seems less happy with this approach. Logs don't appear to include entries like those reported above. Investigations continue...
Ok, so here's how I was able to reproduce the error:
This relates to what I'm seeing in the log saying:
kernel AppleFWOHCI_AsyncReceive::waitForDMA - context not going inactive.
When you bring up your console viewer, make sure you have it selected to view "All Messages". Console messages only won't display the kernel message.
At some point, that kernel message noted above starts popping up. For me, when the computer goes to sleep during this occurrence, the system won't be able to resume from keypresses or anything. You'll just need to hold the button in the back of your iMac or whatever to manually shut things down.
I'm not sure if this would also be the same thing for when the screen saver locks up, but at least for sleeping, it's a guaranteed way to lock my system up. I've tried this when the message wasn't present multiple times, and it goes to sleep and wakes up without a hitch.
I wish I knew what this message meant. Probably something to do with Direct Memory Access and an Asynchronous call waiting for some DMA response...
Yah, I have a firewire drive as well. It was also off too.
The last post I read in there (at the time) referenced that resetting the PRAM fixed the problem. Not sure yet if I need to go there yet, but here's apple's documentation for doing that:
That seemed to work. I actually did both. I shut down my computer, unplugged all my peripherals, usb & Firewire (since I also have a usb Flash card reader and firewire backup hard drive). I reset my computer and did the PRam reset. All seems ok now. I haven't seen the DMA message so far nor has my computer frozen when I try to wake it.
Try what Keith suggested and what I did. So far, that seems to have worked. I went the extra step further and unplugged the USB as well, since I had some flash drives mounted there.
I don't think this issue is as much related to "screensaver" as it is to "sleep" -- or, even more likely, to some fundamental hardware settings having been affected by the OS X update, as some earlier posters have noted. My system was totally unresponsive this morning. Last night I put it to sleep using the shutdown dialog -- there was no screensaver involved.
Earlier yesterday, I noticed that no FireWire drives would mount (one drive actually made an unusual relay-click noise when I recycled the power switch. That made me check Apple System Profiler where I found a warning present that FireWire was not available. A system restart seemed to set things right (but apparently not entirely, as I found this morning).
PRAM and NVRAM resets do seem to be a logical next step.
If I'm not mistaken, the "FW" in "AppleFWOHCI" refers to FireWire.
If you've still got access to the log, scroll back to where the "chugging" begins. Look at entries in the hour before the chugging. There will be occasional entries of: "
AppleFWOHCI_AsyncTransmit::waitForDMA" -- just "transmit", not "receive".
My log also shows an entry just prior to all this: "+FireWire (OHCI) Lucent ID 5901 built-in: 53 bus resets in last 3 minute.+"
The problem isn't related to screensaver.
Yesterday I tried resetting PRAM and NVRAM, but was still greeted with a comatose computer this morning. I just finished re-applying the 10.5.6 combo updater and will see what I find tomorrow morning.