[Angus_D]

OK, I'm going to have a go at fixing this. I actually have a vague hunch I know what is causing this, but I want to make sure I'm on the right track.

ONLY FOLLOW THESE INSTRUCTIONS IF YOU ARE COMFORTABLE WITH THE TERMINAL. SINCE YOU ARE MESSING WITH LOW-LEVEL BITS OF THE SYSTEM, IF YOU DON'T FULLY UNDERSTAND WHAT THESE INSTRUCTIONS TELL YOU TO DO YOU CAN SCREW UP YOUR INSTALLATION OF MAC OS X. You have been warned.

Please get AppleSCCSerial.kext.tar.gz from http://homepage.mac.com/fdobbie and decompress it. Then move the Apple-distributed AppleSCCSerial.kext out of the way to keep a copy of it (e.g. sudo mv /System/Library/Extensions/AppleSCCSerial.kext ~/AppleSCCSerial.kext-dist). Then move the AppleSCCSerial.kext you downloaded to /System/Library/Extensions (e.g. sudo mv ~/Desktop/AppleSCCSerial.kext /System/Library/Extensions/AppleSCCSerial.kext). Then reboot.

Now, fire up the Console application (in /Applications/Utilities), and go to File->Open Log. In the "Go To" field, put /var/log/system/log. Now, when you connect, you should get some lines like this:

Feb 3 15:23:52 localhost pppd[1208]: pppd 2.4.0 started by root, uid 0
Feb 3 15:23:53 localhost mach_kernel: ... about to acquirePort()
Feb 3 15:23:53 localhost mach_kernel: [acquirePort Early]
Feb 3 15:23:53 localhost mach_kernel: AppleSCCModem(1750680): acquirePort returned kIOReturnSuccess
Feb 3 15:23:53 localhost mach_kernel: AppleSCCModem(1750680): setPowerState is called -- powerStateOrdinal = 1

Could you please post these lines from your system.log, so I can see if this is where the hang is happening for you.

Edit: Add warning.

[ 02-03-2002: Message edited by: Angus_D ]

[Krypton]

I'll definitely help you!

I've sent about 20 emails to devbugs (most of them useful, some just bitching) and even to Avie Tevanian. The engineers are making no progress because they assign it to someone different each week to avoid fixing it.

[dfiler]

I would love to help but I'm always wary of being the first one try a new kernel extension. You may get more takers if you reformat the instructions with each step on its own line...

Does anyone know if the new iMac can use its modem without a multi-minute complete freeze? If not, this bug should become a standard case study in software engineering ;-)

Many months ago I really let into Apple and got as far as speaking to the public relations people. They are the people to talk to when the normal feedback process isn't work. This was back when using your modem on a dual processor machine was garaunteed to result in a kernel panic and possible data corruption.

[ 02-03-2002: Message edited by: dfiler ]

[Krypton]

Be Careful with this Kernel Extension

I went a little over the top with what I said (sorry Angus), but I think I remember what I did wrong.

When moving the files, the system added .kext-dist to the extension whereas when you place it in System/Library/Extensions it should have the extension .kext

Hope that stops someone else from making the same mistake.

[ 02-03-2002: Message edited by: Krypton ]

[Angus_D]

Umm... What?
The new extension works for me, and I was intending these instructions only for non-idiots who know their way around the terminal. Apologies for not making that clearer. If your back-up didn't work, then you obviously ****ed up. If they had the wrong extension, you obviously ****ed up.

You can't have followed the instructions properly. If it makes it easier, I can probably make an installer package.

Anyway, if that's the reaction I'm going to get, I just won't bother. This bug has stopped affecting me directly, so I don't really care. Jeez, you try to help some people...

[besson3c]

Krypton:

Angus's instructions included backing up the old .kext file so that you can easily continue to use the original in case the modified version presented problems. There was no need to copy the file from your Dad's computer.

I think common sense would indicate not to run terminal commands moving/modifying important files if you don' t really understand how to undo what you've done....

[Angus_D]

Originally posted by besson3c:
<STRONG>Angus's instructions included backing up the old .kext file so that you can easily continue to use the original in case the modified version presented problems. There was no need to copy the file from your Dad's computer.</STRONG>

I'm guessing the new one didn't work because it wasn't fully decompressed, and if he said it "had the wrong extension", I'm guessing he installed AppleSCCSerial.kext.tar.gz or something stupid like that. My backup instructions involved copying it to ~/AppleSCCSerial.kext-dist, which changed the extension (just to make it notable that it was the original dist version), and he probably copied it back with the -dist suffix, or something stupid like that.

<STRONG>I think common sense would indicate not to run terminal commands moving/modifying important files if you don' t really understand how to undo what you've done....</STRONG>

Yes, I should have mentioned that, but I was in a hurry when I wrote the instructions. I'll edit them now.

[Xeo]

It didn't work for me. I installed the extension and rebooted. I didn't write down the error from the console but it was very similar to the system.log error. This is what I got from the system.log every time I clicked the "Connect" button.

<font face = "courier">Feb 3 14:12:56 localhost pppd[287]: unrecognized option '/dev/cu.modem'</font>

I successfully copied my original back and my modem works fine again. My specs:

B&W G3 400, 640MB RAM. I don't know the modem but it's stock.

What exactly was yours supposed to do? Just narrow in further on the hang? I know there are already two lines in which the hang occurs between so that's why I'm curious if this is meant to focus it a bit more. Or was it meant to actually fix something? I couldn't really tell from your first post.

[Angus_D]

Originally posted by Xeo:
<STRONG>It didn't work for me. I installed the extension and rebooted. I didn't write down the error from the console but it was very similar to the system.log error. This is what I got from the system.log every time I clicked the "Connect" button.

&lt;font face = "courier"&gt;Feb 3 14:12:56 localhost pppd[287]: unrecognized option '/dev/cu.modem'&lt;/font&gt;</STRONG>

Hmm... Sure it's installed correctly? You should only get that if the kext isn't loaded. Oddness.

<STRONG>What exactly was yours supposed to do? Just narrow in further on the hang? I know there are already two lines in which the hang occurs between so that's why I'm curious if this is meant to focus it a bit more. Or was it meant to actually fix something? I couldn't really tell from your first post.</STRONG>

Narrow down a bit (all it does is add a few IOLog()s). It narrows it down to AppleSCCSerial::aquirePort, and once someone successfully tests that, I can narrow it down to lower than that. Hmm. Perhaps I should put up an installer package, and then I could probably make it so that you don't have to reboot to test it

[Angus_D]

OK, try AppleSCCSerial.dmg which has an installer .pkg on it.

[edddeduck]

Might give it a go on my dual tomorrow to l8 now...

Cheers Edd

btw feedback you have an e-mail or you just want the feedback to this thread?

[Xeo]

OK, I downloaded the .dmg and installed from the .pkg and it's working now. Maybe when I typed using the mv command I misspelled or something. Oh well. it's working now. Here are those lines:

Feb 3 20:19:21 localhost pppd[319]: pppd 2.4.0 started by root, uid 0
Feb 3 20:19:21 localhost mach_kernel: ... about to acquirePort()
Feb 3 20:19:21 localhost mach_kernel: [acquirePort
Feb 3 20:19:22 localhost mach_kernel: Early]
Feb 3 20:19:22 localhost mach_kernel: AppleSCCModem(2a5f480): acquirePort returned kIOReturnSuccess
Feb 3 20:19:22 localhost mach_kernel: AppleSCCModem(2a5f480): setPowerState is called -- powerStateOrdinal = 1

Keep in mind, I just restarted so there isn't any delay yet. I'll keep using it. Since I'm home from college for the next week, I have no choice but to use the modem.

I'll keep posting my results here.

[Xeo]

Well, had my first noticeable hang. It was just over a minute long. Here are all 5 of those lines, even though you really only needed to see the first two.

Feb 4 01:12:33 localhost pppd[5548]: pppd 2.4.0 started by root, uid 0
Feb 4 01:13:46 localhost mach_kernel: ... about to acquirePort()
Feb 4 01:13:46 localhost mach_kernel: [acquirePort Early]
Feb 4 01:13:46 localhost mach_kernel: AppleSCCModem(2a5f480): acquirePort returned kIOReturnSuccess
Feb 4 01:13:46 localhost mach_kernel: AppleSCCModem(2a5f480): setPowerState is called -- powerStateOrdinal = 1

Does that help to narrow it down further? I'll keep posting results as I get the hangs.

[edit: just for kicks, I disconnected and reconnected again and here are the results]

Feb 4 01:31:59 localhost pppd[5610]: pppd 2.4.0 started by root, uid 0
Feb 4 01:33:26 localhost mach_kernel: ... about to acquirePort()
Feb 4 01:33:26 localhost mach_kernel: [acquirePort Early]
Feb 4 01:33:26 localhost mach_kernel: AppleSCCModem(2a5f480): acquirePort returned kIOReturnSuccess
Feb 4 01:33:26 localhost mach_kernel: AppleSCCModem(2a5f480): setPowerState is called -- powerStateOrdinal = 1

Sure enough, I got the hang between the same two lines.

[ 02-03-2002: Message edited by: Xeo ]

[Angus_D]

Hrmph. Someone before said that they had traced it to the line between ... about to aquirePort() and aquireport returned blah.

I think it's time to go back down to pppd from the kext. Perhaps I should mail Godfrey again.

[Krypton]

Feb 3 14:12:56 localhost pppd[287]: unrecognized option '/dev/cu.modem'

This is exactly the sort of error I ended up with. I didn't do something moronic like install the extension.tar.gz I followed the instructions as you wrote them.

I'm sorry if I over reacted, but loosing PPP dial up altogether is not my idea of fun.

I did back up the original, and it didn't seem to work when I put it back - that's why I was angry (I know I'm probably to blame).

I do know how to use the terminal, but somewhere along the way one of the files ended up with .kext-dist instead of .kext which threw me.

[Xeo]

Originally posted by Krypton:
<STRONG>This is exactly the sort of error I ended up with. I didn't do something moronic like install the extension.tar.gz I followed the instructions as you wrote them.</STRONG>

Then it's possible it was the kext he gave us and not our error. The new one he packaged up works fine though. Maybe it was a problem with the compression or something. Oh well. It's also possible we both screwed up the name when moving it.

<STRONG>I'm sorry if I over reacted, but loosing PPP dial up altogether is not my idea of fun.

I did back up the original, and it didn't seem to work when I put it back - that's why I was angry (I know I'm probably to blame).</STRONG>

It's not that bad. So PPP didn't work. Just use the backup and reboot. All problems fixed.

<STRONG>I do know how to use the terminal, but somewhere along the way one of the files ended up with .kext-dist instead of .kext which threw me.</STRONG>

Yes. His instructions specifically said to rename it to .kext-dist. That's just so you know it's the distribution copy and not his altered copy. It's the same as if you named it .kext-bkup or something along those lines. Then, when you move that file back to it's original directory, you're supposed to remove the -dist so it's a regular .kext file again. It was only there for "human recognition" reasons.

[Krypton]

Then, when you move that file back to it's original directory, you're supposed to remove the -dist so it's a regular .kext file again. It was only there for "human recognition" reasons.

Ah - that would be it. I was under the impression it was something the system did (I copied and pasted the commands) instead of purely to differentiate the two. Well, everything is back to normal so I'm happy

One further note - the reason I didn't see the difference between the two extensions is that my extensions folder has 'extensions hidden' so on the face of it there was nothing wrong - in future I shall know to be more careful

[Xeo]

Just thought I'd post another. Same result for me. This one was just under a minute.

Feb 4 15:16:00 localhost pppd[500]: pppd 2.4.0 started by root, uid 0
Feb 4 15:16:55 localhost mach_kernel: ... about to acquirePort()
Feb 4 15:16:55 localhost mach_kernel: [acquirePort Early]
Feb 4 15:16:55 localhost mach_kernel: AppleSCCModem(2242900): acquirePort returned kIOReturnSuccess
Feb 4 15:16:55 localhost mach_kernel: AppleSCCModem(2242900): setPowerState is called -- powerStateOrdinal = 1

Here's another:

Feb 4 16:23:36 localhost pppd[538]: pppd 2.4.0 started by root, uid 0
Feb 4 16:25:18 localhost mach_kernel: ... about to acquirePort()
Feb 4 16:25:18 localhost mach_kernel: [acquirePort Early]
Feb 4 16:25:18 localhost mach_kernel: AppleSCCModem(2242900): acquirePort returned kIOReturnSuccess
Feb 4 16:25:18 localhost mach_kernel: AppleSCCModem(2242900): setPowerState is called -- powerStateOrdinal = 1

And another:

Feb 4 16:42:53 localhost pppd[650]: pppd 2.4.0 started by root, uid 0
Feb 4 16:43:46 localhost mach_kernel: ... about to acquirePort()
Feb 4 16:43:46 localhost mach_kernel: [acquirePort Early]
Feb 4 16:43:46 localhost mach_kernel: AppleSCCModem(2242900): acquirePort returned kIOReturnSuccess
Feb 4 16:43:46 localhost mach_kernel: AppleSCCModem(2242900): setPowerState is called -- powerStateOrdinal = 1

[ 02-04-2002: Message edited by: Xeo ]

[edddeduck]

same here as above same position in the code.

Edd

[xi_hyperon]

I had my first experience with a PPP hang yesterday, so prior to that I had not paid much attention to the threads related to the problem. My question is, and maybe it's a stupid one, but why has this problem been around for so long? Is Apple having a hard time tracking it down? I remember seeing threads about this earlier last year.

Of course, now that I've experienced it myself, I'm much more interested in the discovery of a solution.

[JKT]

Originally posted by xi_hyperon:
<STRONG>I had my first experience with a PPP hang yesterday, so prior to that I had not paid much attention to the threads related to the problem. My question is, and maybe it's a stupid one, but why has this problem been around for so long? Is Apple having a hard time tracking it down? I remember seeing threads about this earlier last year.

Of course, now that I've experienced it myself, I'm much more interested in the discovery of a solution. </STRONG>

Yes, Apple is having trouble tracking it down. They apparently can't reproduce it on their own networks (duh, its a PPP/modem problem!) if the Darwin bug report 2691330 is anything to go by. (N.B. This bug report keeps dissappearing from the list and reappearing every now and then so it might not be there at the moment)

Please feel free to contribute voiciferously to the discussion at Apple support here - the more people posting the sooner it will be cured. And don't forget OS X Feedback either.

P.S. I've been suffering from it since day one of 10.0 and I sincerely hope for your sake that yours is just a one off occurrence because this is the biggest PITA of my computing life... even DOS wasn't as bad as this.

[Angus_D]

Thanks guys... Conferring with Apple people on this (actually bugging them individually, hopefully will get a response).

[dillerX]

This may sound stupid, but mine does not freeze if I leave Mail open while disconnected. If I leave that program open, the computer re-connects right away. If I disconnect, close the Mail program, and then try to re-connect, the computer hangs.

It has NEVER hung as long as I keep that program open.

400MHz G4 AGP
960RAM

Just my experience, probably worth nothing.

[Krypton]

Thanks guys... Conferring with Apple people on this (actually bugging them individually, hopefully will get a response).

Progress at last (wish I could get hold of them) - I can only manage to get hold of executives, and that's through guessing their email addresses.

[Angus_D]

Originally posted by Krypton:
<STRONG>Progress at last (wish I could get hold of them) - I can only manage to get hold of executives, and that's through guessing their email addresses.</STRONG>

Best people are Godfrey van der Linden (gvdl at apple dot com), who is a driver guy, and Christoph Allie (callie at apple dot com), who is the Apple PPP engineer (and owners of the initials C.A. on 2691330).

I'm just off to bug Ron Dumont and Beatrice Katianda about the lameness of Darwin Bugs and how 2691330 keeps on disappearing.

[Angus_D]

Originally posted by dillerX:
<STRONG>This may sound stupid, but mine does not freeze if I leave Mail open while disconnected. If I leave that program open, the computer re-connects right away. If I disconnect, close the Mail program, and then try to re-connect, the computer hangs. Just my experience, probably worth nothing.</STRONG>

Not worth anything, no, and please don't clutter the thread with these ye olde wive's tales of how to fix this bug, they aren't terribly useful (and are probably wrong). Thanks.

[Krypton]

Best people are Godfrey van der Linden (gvdl at apple dot com), who is a driver guy, and Christoph Allie (callie at apple dot com), who is the Apple PPP engineer (and owners of the initials C.A. on 2691330).

I'm beginning to think devbugs have filtered out my emails after I gave up on being patient, and just grilled them over this issue!

[Xeo]

Are you going to alter the kext again to narrow it down further? I'm eager to help while I'm home using my modem (seeing how I get the damned hang every time).

[benh57]

2691330 has been updated and is currently viewable.
http://www.opensource.apple.com/bugs...y/2691330.html

Sounds like they are trying the "latency" tool.

Try this during a PPP hang:

sudo latency -rt -st 20000 -it 000 -l /var/tmp/latency.log


If you don't get anything in latency.log, try reducing the 20000 value.

More usage in "man latency".

[Angus_D]

Asking Godfrey for his modified /usr/bin/latency.

[drjoe]

If I disconnect, close the Mail program, and then try to re-connect, the computer hangs. Just my experience, probably worth nothing.



Not worth anything, no, and please don't clutter the thread with these ye olde wive's tales of how to fix this bug, they aren't terribly useful (and are probably wrong). Thanks.


*****

Well it seems to me that since nobody has been able to isolate the exact problem here, and that noone has been able to een empirically solve the problem, users who in good faith put forward fixes should at least not be publicly humiliated for their attempts to put forward solutions. These solutions may not hae valid technical underpinnings but if they DO WORK they should be accepted as at least good faith attempts to solve a real problem.
In my own case I have put into effect a simple solution that least for me stops PPP hangs on modem connections. Sent this in to Apple and at least got a response that was civil and not "Go away boy, I'm to busy to listen to you - gotta get back to my techno-poseur posture" The end result is that I don't have PPP hangs anymore. Bottom Line.

[boots]

Originally posted by drjoe:
<STRONG>
Well it seems to me that since nobody has been able to isolate the exact problem here, and that noone has been able to een empirically solve the problem, users who in good faith put forward fixes should at least not be publicly humiliated for their attempts to put forward solutions. These solutions may not hae valid technical underpinnings but if they DO WORK they should be accepted as at least good faith attempts to solve a real problem.
In my own case I have put into effect a simple solution that least for me stops PPP hangs on modem connections. Sent this in to Apple and at least got a response that was civil and not "Go away boy, I'm to busy to listen to you - gotta get back to my techno-poseur posture" The end result is that I don't have PPP hangs anymore. Bottom Line.</STRONG>

I think there is just a lot of frustration. If you go to the apple support page (actually the discussion pages) linked above, you'll see that people have proposed everything from simple "if I have program X open, it never hangs" to "Offer a chicken to the computer gods on the second newmoon at 12:52am but only in months that have two U's in them."

None of them are universal fixes. If you have suggestions of the above nature, I suggest you post them on the Apple Forum. If there is concensus that it works, it will get passed on to the developers.

[drjoe]

Originally posted by boots:
<STRONG>

None of them are universal fixes. If you have suggestions of the above nature, I suggest you post them on the Apple Forum. If there is concensus that it works, it will get passed on to the developers.</STRONG>

I don't have to send it in to Apple Support. I emailed Christoph Allie and got a polite and immediate response. So it presumably in the right hands. A universal solution would imply that something is known of the etiology of this problem. Since THAT is not known a reasoned solution may be a long ways off. But that does not mean that an effective solution is a long ways off. It may be that a serendipitous solution may be found. If it is reliable and reproducible that, as a first approximation, it may be useful. It may also, in the hands of people who are professionals in the field be useful in figuring out the fundamental problem. Frustration or not incivility really should not be tolerated. Anyway, for better or worse, here goes:

PPP hangs occurrED for me all the time. The system was stable but this was a real thorn in my side. A few timers a day. I recently was trying out a new piece of software Audiocorder, and was loading the entire first act of an opera and neglected to specify the scratch disk as other than my boot partition. When all the space ran out a kernel panic intervened and down we sent. Upon reboot fsck saw a number of invalid sister links and would not [could not] repair them. Neither could DFA in 9.2.2, DFA on the OSX install disk. When ran Diskwarrior in 9.2.2 the MACOSX partition was a mess but I dutifully plunged ahead and rebuilt the directory and optimized it. Took an age. Without any real hope that the problems were soled I rebooted and up it came. It wasn't till later that I realized that the PPP hangs were gone. That was last week and after 100+ trips back and forth to the connection I have not had a hang since. If they recur I will not be surprised but for now the problem is gone and certainly SOMETHING had changed in the PPP behavior. I offer this up as perhaps a starting point. I have no clue as to why rebuilding the directory tree has had this effect but it has - and without sacrificing any chickens.

[Angus_D]

Originally posted by drjoe:
<STRONG>I don't have to send it in to Apple Support.</STRONG>

Of course you don't. It was suggested that you post it to the Apple forums, where people have suggested millions of ways of "fixing" this issue, but very few of them have a grounding in reality and none of them work for everybody (or, in most cases, anybody other than the originator, pointing to: COINCIDENCE).

I asked people NOT to post these types of things because they are not on-topic for this thread, which was started as an attempt at actually fixing this problem, not "working around it" using some convoluted system of steps that probably don't actually work. If you want to blatantly ignore my request, go ahead, but I can't see it helping anybody (and will annoy me).

[OverclockedHomoSapien]

I found that if I apply gentle but firm pressure to my left nut while connecting or disconnecting my modem connection, then OS X never hands. As you can all imagine, it took a little luck and a lot of time to discover this remedy.

BTW, I emailed everyone who's addy I know at Apple, explaining things to them, and also providing links to all the discussion threads I could find on this bug. Also for months I've been sending feedback to Apple via the OS X feedback form, on a daily basis (1-2 per day). If everyone does this then Apple will hopefully fix the bug someday.

Sorry, but I don't know Unix very well so I'm not going to use that kext file of yours, Agnus. But I must tell you, I really appreciate you putting your time and effort into solving this problem. It's shocking that Apple has allowed a bug of this nature to persist for so long, but perhaps with your help Agnus, we can put enough pressure on Apple to get it fixed.

How hard can it be to fix it? Just sick a platoon of OS X programmers on it, have them dial up a connection from someone's residence, and squash that bug! I get the feeling that Apple isn't giving this bug much attention or effort, because it's so reproducible it's difficult to imagine them having a hard time finding it.

But then I'm not a programmer so I don't really know what I'm talking about

[ 02-06-2002: Message edited by: OverclockedHomoSapien ]

[Angus_D]

Grr... Looks like the semantics of IOLog have changed. New package is up (http://homepage.mac.com/fdobbie as before), this time I'm interested in the line that looks like this:
Feb 6 22:46:30 localhost mach_kernel: acquirePort seems to have taken about xx seconds

It could still be acquirePort!

[benh57]

The way i read the bug he is using the standard "latency" tool, just using the st argument. Does the hang show up in the latency stats? With "st" it prints the lagging calls to the log. Open the latency.log in console and watch them fly by if your number is low.

I don't actually use PPP, just interested in helping.

-B

[ 02-06-2002: Message edited by: benh57 ]

[Xeo]

I'll have to give it a shot when I get back home. Right now I'm at my dad's on his cable modem so I can't try it.

[Angus_D]

I got a 14 second hang, and it looks like it is definitely occurring there. Time to add some more logging, I think there may be some loops that are spinning out of control.

[Xeo]

Hang between the same two lines, but you're right on the "acuirePort seems to have taken" line. This was a nice long one, eh? 8 minutes!

Feb 7 15:38:48 localhost pppd[612]: pppd 2.4.0 started by root, uid 0
Feb 7 15:46:49 localhost mach_kernel: ... about to acquirePort()
Feb 7 15:46:49 localhost mach_kernel: [acquirePort Early]
Feb 7 15:46:49 localhost mach_kernel: acquirePort seems to have taken about 481 seconds
Feb 7 15:46:49 localhost mach_kernel: AppleSCCModem(2a5f400): acquirePort returned kIOReturnSuccess
Feb 7 15:46:49 localhost mach_kernel: AppleSCCModem(2a5f400): setPowerState is called -- powerStateOrdinal = 1

[Angus_D]

Originally posted by Xeo:
<STRONG>Hang between the same two lines</STRONG>

This is misleading. What's happening is that IOLog() isn't being flushed to the syslog process quick enough, so the original "about to acquirePort()" message doesn't get logged until after the hang. This changed somewhere along the road, and threw me a bit. Time to add even more logging!

[jmzorko]

Hello, all ...

I've experienced this hang on my Pismo 500 since day one of the PB, and i've had the hang happen for as long as 30 minutes (after 30 minutes, I just swear a lot and restart &lt;g&gt; ). When I tried to trace it down, it seemed to be happening _inside_ acquirePort(), but I think that may have been with 10.1.1 or something. Also, I used to get the hang to happen, like I said, for many minutes at a time, but since going to 10.1.2, i've only gotten it to hang for 20-30 seconds max. I'm not sure if it's 10.1.2, or that this bug just knows that i'm looking for it :-)

Does anyone think it may have something to do with specific modem hardware? I ask because it doesn't seem to happen anywhere _near_ as often on my tangerine iBook as on my Pismo ...

Also, i've gotten the hang to happen w/o using PPP at all ... just compile the following code and run it (you may need to be su):


#include &lt;stdio.h&gt;
#include &lt;ctype.h&gt;
#include &lt;stdlib.h&gt;
#include &lt;string.h&gt;
#include &lt;unistd.h&gt;
#include &lt;signal.h&gt;
#include &lt;errno.h&gt;
#include &lt;fcntl.h&gt;
#include &lt;syslog.h&gt;
#include &lt;netdb.h&gt;
#include &lt;utmp.h&gt;
#include &lt;pwd.h&gt;
#include &lt;setjmp.h&gt;
#include &lt;sys/param.h&gt;
#include &lt;sys/types.h&gt;
#include &lt;sys/wait.h&gt;
#include &lt;sys/time.h&gt;
#include &lt;sys/resource.h&gt;
#include &lt;sys/stat.h&gt;
#include &lt;sys/socket.h&gt;
#include &lt;netinet/in.h&gt;
#include &lt;arpa/inet.h&gt;
#include &lt;sys/syslog.h&gt;

main ()
{
int iModem;
syslog(LOG_INFO, "Will do open\n");
iModem = open("/dev/cu.modem", O_NONBLOCK | O_RDWR, 0);

if ( iModem == -1 )
{
syslog(LOG_INFO, "Error opening the port = %d\n", errno);
}
else
{
syslog(LOG_INFO, "Open done\n");
}

return 0;
}


Regards,

John

Falling You - exploring the beauty of voice and sound http://www.mp3.com/fallingyou

[benh57]

Ouch, that certainly did the trick on my DP800. After 15 minutes, the clock updated and the cursor changed from disk to mouse, but i still couldn't use the mac or ssh in. And energy saver idle time worked!

I gave up and restarted after 30 mins. Upon restart the program works fine.
Anyway, i'm running "latency" now and i'll try it periodically today to try to reproduce.

[jmzorko]

Originally posted by benh57:
<STRONG>Ouch, that certainly did the trick on my DP800. After 15 minutes, the clock updated and the cursor changed from disk to mouse, but i still couldn't use the mac or ssh in. And energy saver idle time worked!

I gave up and restarted after 30 mins. Upon restart the program works fine.
Anyway, i'm running "latency" now and i'll try it periodically today to try to reproduce.</STRONG>

Has anyone noticed if the power management (system sleep) features have anything to do with the hangs when using the modem? In other words, does the hang happen soon after the screen is awakened, or the hard disks spin up, or the laptop is opened? I ask this because, when I experience the hang, it is always after opening up my Pismo or iBook (as one of the first things I do usually upon opening is connect). Also, if this is the case, does the length of time the system / screen is asleep seem to have anything to do with it? I ask this because if I just close the lid and open it again right away and try, it doesn't seem to happen, but it almost always happens when I open the lid after the drive home (about an hour from San Jose to Santa Cruz / Bonny Doon).

Regards,

John

Falling You - exploring the beauty of voice and sound
http://www.mp3.com/fallingyou

[John Gee]

I've installed the debugging module, no hangs so far (but computer has not gone longer than a day without a reboot since the install). Will mail as soon as get the "usual" bad behaviour. On my computer (DP450) it lasts between 30 seconds and two minutes, so I had been ignoring it with gritted teeth until I stumbled across this thread - I'm happy to assist in the hopes of a fix!

[Xeo]

Originally posted by John Gee:
<STRONG>I've installed the debugging module, no hangs so far (but computer has not gone longer than a day without a reboot since the install). Will mail as soon as get the "usual" bad behaviour. On my computer (DP450) it lasts between 30 seconds and two minutes, so I had been ignoring it with gritted teeth until I stumbled across this thread - I'm happy to assist in the hopes of a fix!</STRONG>

I'm not exactly sure how you found this thread since it died about a week ago . Angus_D restarted the discussion just last night in a new thread.

Just to keep things straight, I'll close this thread so the new kext can be discussed in the new thread. You have the current one though since it's the only one in his Public folder. We could definitely use another tester though. So far, everyone likes to complain about this bug but no one wants to help track it down. I'm the only tester for the new one so far (and now you).