Mac OS X 10.5 (Leopard): UI stutters and USB problems
Posted by Pierre Igot in: MacintoshNovember 21st, 2007 • 3:52 pm
There seems to be a range of new problems with Mac OS X 10.5 that people never experienced with the same equipment running Mac OS X 10.4, and I am one of the users affected by this type of problem.
The problem is a bit tricky to describe because it manifests itself in a variety of ways and cannot always be reproduced reliably.
Basically, here are the symptoms that I have noticed on my machine:
- From time to time, the whole Mac OS X UI will “stutter” or have a “hiccup,” with the mouse pointer jumping all over the place. This only lasts a few seconds and things go back to normal afterwards. If the hiccup occurs right while I am in the process of dragging something with the mouse button down or pulling down a menu, then obviously this particular item or menu will be affected as well, because the mouse pointer jumps all over the place and the item/menu tries to keep up with it. I cannot reproduce this problem 100% reliably. However, I have noticed that it occurs quite often when I invoke Snapz Pro X 2.1.2 with my chosen keyboard shortcut (command-control 3). It does not happen every time I invoke Snapz Pro X, but it definitely occurs most often the first time I invoke it (after having rebooted the computer) or if I haven’t used it in a while (i.e. in a few days). Again, the hiccup is just temporary and after a few seconds Snapz Pro X regains control and I am able to proceed. Aside from being more frequent with Snapz Pro X, the problem can occur in any application (the Finder, Pages, etc.). Typically there is a certain amount of hard disk activity that I can hear while the hiccup is happening.
- I have also noticed problems with certain USB devices. In particular, my mouse (a regular Apple Mighty Mouse) was connected to my keyboard (an aluminum Apple keyboard). When it was connected to the keyboard, I noticed that, after using Mac OS X 10.5 for a while, the mouse pointer would start behaving a bit erratically. It was quite subtle and hard to describe, but I could definitely notice that my mouse movements were no longer as accurate as they were before the problem started. It was as if the mouse pointer was a bit jumpy, but in a very subtle way. I first suspected an incompatibility with USB Overdrive (which I am using for better control of the behaviour of the mouse’s various buttons in various applications). But I turned it off and the problem was still there. I ended up disconnecting the mouse from the keyboard and connecting it to another (powered) USB port, and I haven’t had the problem since.
- The reason I changed USB ports was that, when I was first having the subtle problem with the mouse pointer, I tried unplugging the mouse from the keyboard and plugging it back in—and that was enough to cause a kernel panic! This is what first led me to suspect a problem with USB devices in Mac OS X 10.5, particularly because the log entry for the kernel panic (in Library/Logs/Panic Reporter/) contained multiple references to “
com.apple.iokit.IOUSBFamily.” I don’t really know how to read these panic logs, and the references to the USB architecture might just be a coincidence, but it made me suspicious nonetheless, especially since my problems were indeed involving USB devices, including the mouse. - I also noticed various other USB-related problems in Mac OS X 10.5, including one with my PowerWave USB audio device, which at some point disappeared from the OS X sound interface altogether. It’s working fine again now, but again, this is after having reshuffled my USB devices.
- Finally, both during my tests of earlier builds of Mac OS X 10.5 on a Mac Pro used as a test machine and after installing the final version on my regular Mac Pro workstation, I noticed that, sometimes, the Mac Pro would simply refuse to wake up from sleep. Since the problem occurred even on my test machine with no extra peripherals and no third-party software, I strongly suspect some internal problem in Mac OS X 10.5, which remains unaddressed in the currently available versions of the software. But at the same time, the panic log for these episodes also includes references to “
com.apple.iokit.IOUSBFamily,” and I cannot help but notice that, since I have reorganized my USB devices, I haven’t had a kernel panic, and I haven’t had a single “won’t wake from sleep” episode either. Which leads me to suspect that there might be a connection between this particular symptom and the other USB/UI stuttering problems. It is only speculation at this point, of course, but that’s about all I can do with the available evidence (and the lack of feedback on this issue from Apple’s engineers).
As you can imagine, it’s not exactly easy to find discussions on this topic, especially since the terminology used to describe the problem will likely vary from user to user. But I have found a number of discussion threads on Apple’s Discussions forums, including attempted descriptions of “mouse stutter and UI unresponsiveness,” a “mouse lag,” and various references to “USB problems,” a “USB sound problem,” and other “USB problems” seemingly related to Mac OS X 10.5.
All this tends to indicate that Apple made some substantial under-the-hood changes that affect USB devices in particular (including the mouse) and might be linked to system freezes and stutters. It is particular frustrating when this happens, because Mac OS X is supposed to be a strong, stable operating system, and people shouldn’t have to worry about daily problems that require a restart or a hard reset.
Right now, I am happy that the only problem I am still experiencing is the occasional “hiccup” described above, which comes and goes without any lasting impact that I can see (although it is still worrying when it happens, and makes me fear for the health of my hard drives in particular). But it’s probably too early to tell whether I have eliminated the other problems for good, and the discussions on Apple’s forums certainly seem to indicate new USB issues in Mac OS X 10.5 that will need to be addressed by Apple sooner rather than later.
November 21st, 2007 at Nov 21, 07 | 6:24 pm
I don’t really have anything productive to add, but I have to moan publicly about how depressing this is!
It’s just such a disappointment to me how many problems I’ve had with Leopard. The reality of Mac OS X is just so, so, so far from the hype! It pains me. It’s embarrassing. I avoid calls from switchers, baffled by Mac problems they thought they weren’t supposed to have …
Will we ever have a truly reliable, relatively bug-free operating system? Will technology ever stabilize? Is it even possible? Or is computing just going to produce a perpetual flow of buggy innovations?
Yada yada yada. Sigh. Sorry, Pierre, I’m polluting your blog with pointless bellyachin’! ;-)
November 21st, 2007 at Nov 21, 07 | 7:50 pm
Interesting. I had a problem a bit like this (drags would very occasionally stop and the item under the mouse would be released) in Tiger after I installed the iStat menu bar extra, but Leopard has been smooth for me so far.
November 21st, 2007 at Nov 21, 07 | 9:20 pm
This is quite close to my own experiences with X.5 so far.
On the one hand I occasionally (actually quite rarely by now) have the impression that my mouse cursor movement isn’t as smooth as it should be. This is very hard to formulate (and my impression is that Apple will decide they can’t reproduce it). As this is transient it’s very hard to nail down. Perhaps, in some part it can be pegged to X.5 being even more of a memory hog and even keener to stall itself a bit while playing with its swap files.
On the other hands I did have my share of kernel panics while Time Machine was running already. There also pointed to USB (USBMassStorage, though). The odd thing is that while I had the problem numerous times by now (enough to turn off Time Machine to avoid strong use of the USB drive because I’m not _that_ keen on the panics), it’s completely unclear to me what actually causes it.
(Some more detail on the latter problem on my site: http://earthlingsoft.net/ssp/blog/2007/11/x5_time_warp )
November 22nd, 2007 at Nov 22, 07 | 9:18 am
Paul: No problem, you are among friends :). That said, I don’t feel that the situation with Mac OS X 10.5 is as bleak as you describe. Yes, of course, the OS still has a number of bugs, some of which I am highlighting on this blog. But on the whole I do feel that it’s a real improvement over Mac OS X 10.4. That said, I am afraid that modern technology and stability have become somewhat antithetic. Personally, I blame Microsoft for this. They are the ones who forced hundreds of millions of people to become used to buggy, flawed software and lowered the standards for everyone. But the problem also extends beyond computer-based technology. I see flawed technology all around me, from my PVR to my car to my coffee machine. The car situation is interesting, because it’s an industry where they really do have to get the basics right, for obvious safety reasons. On the other hand, there is very little innovation taking place. So to me it looks like innovation and stability/reliability is the real antithesis here. It seems impossible to have both. What is really frustrating is when you get the lack of stability/reliability without the innovation (witness current Microsoft products). Apple is redeemed by the fact that they still provide a certain degree of innovation—but not quite enough to atone for the lack of reliability. It’s a case of things being not as bad as with the competition more than anything else. Will all this ever change? It’ll take some kind of paradigm shift. I don’t think anyone knows where this shift might come from.
AlanY: Some people in the forums that I mentioned do appear to link some of the problems I described to Dashboard in Mac OS X 10.5. I have been unable to confirm this so far.
ssp: The problems with the mouse are indeed hard to describe. But they are very real. Yet I haven’t noticed a connection between the problems and overall system activity. I am on a Mac Pro with four cores and plenty of RAM (5 GB), and I am constantly monitoring the activity with MenuMeters. I haven’t noticed any peak in activity when I get the hiccups. But of course it’s possible that the variations in activity levels occur too quickly to register. I have noticed some hard drive noises during the hiccups, but it’s hard to tell whether they are the cause or the consequence. It’s also not really surprising that your panic logs refer to USBMassStorage, since that’s the USB device involved at the time of your crashes. But it does not rule out a more general USB-related problem. My hard drives are all FireWire (when they have the 2 interfaces, I use FireWire), so I haven’t experienced this myself.
November 22nd, 2007 at Nov 22, 07 | 12:36 pm
I too see LOTS of com.apple.iokit.IOUSBFamily in my logs all occurring either going to sleep or waking from sleep. So much so that for now I either don’t sleep the machine at all or completely shut it down. I have also felt the odd mouse behaviour and merely thought it was a driver issue with my Logitech MX Revolution. Most of the other quirks are app related so it’s less of a deal. I expect oddities with these full OS revs, I’ve never known an upgrade that didn’t have them no matter who’s OS it is. However I find these two items to be very annoying.
November 22nd, 2007 at Nov 22, 07 | 12:37 pm
I find it great that it’s not just me seeing the problem to be honest. I started thinking I’m just being paranoid (but then again, perhaps we bot are :P)
As I said that memory thing is just a guess. And I’m not even sure you’d see the swapping in CPU load. Is you RAM ever full with 5GB (I know my 2GB always are)? And do you get that strange phenomenon when it isn’t? That might give a simple answer to the question. As my 2GB are pretty much completely used all the time, I practically don’t have situations where the system hasn’t used swap yet. If you have those situations (you can see the amount of swapping done since starting in activity monitor) when there has been no swapping, then that theory of mine should be wrong.
November 22nd, 2007 at Nov 22, 07 | 2:11 pm
Well, I just had another KP today while fooling around in Logic Pro. Panic log refers to USB devices again (com.apple.iokit.SM56KUSBAudioFamily, com.apple.iokit.IOUSBFamily, com.apple.iokit.IOPCIFamily). No idea what it all means, but I sure would like to know whether the KPs are related to my USB devices or not.
I am also intrigued by the issue with the swapping, but it definitely seems that the “hiccups” (not the KPs) occur more often when switching to something that has not been used in a while (Snapz Pro X was one example, but I have seen occurrences of the hiccups when switching to other applications that I hadn’t used in a while). I don’t normally monitor my memory usage very closely, because I have enough of it (5 GB), but I am going to keep a closer eye on it from now on. I have MenuMeters showing used and free memory, as well as “paging activity,” whatever that means. Right now (i.e. after restarting following the KP, I have half of my RAM free, and “0” in the paging activity when idle. Activity Monitor still says “Swap used: 0 bytes” at present. Is that what I need to keep an eye on?
November 22nd, 2007 at Nov 22, 07 | 2:48 pm
I’d say the 56K in the extension’s name would hint at a modem. Are you using one?
I’m still not 100% confident about that Unixoid RAM terminology, so I’m not exactly sure what to look for. If 0 swap is used that’s probably good. But I think you might be able to have swap-ins even without having used any swap space (when a file has been loaded from disk, I think it may just be removed from memory and not written to swap because it can just be re-loaded, that might qualify as a swap-in without it having been swapped-out before), but don’t take my word for it.
November 22nd, 2007 at Nov 22, 07 | 3:17 pm
Yes, I have an Apple USB modem, which I use together with Apimac’s Caller ID to get a display of incoming phone call information on my screen. I used this modem with this software in Mac OS X 10.4, so I don’t think this can be the problem. But of course it’s another USB device thrown in the mix, and probably one with which Apple doesn’t do much testing in-house. I guess I could always try to remove it altogether for a few days and see if I still get KPs.
As for the memory stuff, I can definitely see a gradual increase in the “Page ins” number in Activity Monitor (it’s at 1.33 GB right now), but the “Page outs” number is still as “0 bytes,” whatever that means. And “Swap used” is still at 0 bytes too. I haven’t had any hiccups yet—although I have noticed some “hesitation” in the movements of the mouse pointer here and there. I’ll keep an eye on things.
November 24th, 2007 at Nov 24, 07 | 9:22 pm
[…] People that know more than me reported that bunch of graphic card-related errors show up in some log. It may be. It could also be that Leopard + iLife is not-yet-fully-charted territory, even for Apple engineers. Betalogue blog is a great example for that, as this post proves it. […]
December 19th, 2007 at Dec 19, 07 | 5:05 pm
I actually have the same issue with 10.4. I found if I have a USB mouse and keyboard plugged in, the UI will randomly stutter.
Creative Fatality mouse, Microsoft Natural keyboard.
Kind of a bummer.