Watchdog is not a problem - it just reboots device because whole system became irresponsible. If you look into bug 9150 you can find that w/out watchdog reboot the battery depletes VERY fast because N900 loops, and loops, and loops.

A reliable SGX reset is needed and restart of X11.

ddwalker:
Do you have phone orientation set to swap while receiving calls? Have you tried setting the orientation to forced landscape and made a call then? Do you have other apps that rotate the screen?


Another good idea may in fact be to change the device to debug mode, make a call while logged in via ssh, and watch what happens in the log after the call goes off. It's quite possible that the reason the watchdog is calling for a reset is not being flushed to the log before the reset, and it will be more obvious after turning debug mode off.

egoshin:
Please stop saying this has to do with 9150. It has NOTHING to do with that bug. His bug is that the watchdog is resetting his device for some unknown reason. If the watchdog were resetting because of SGX failures my device would be resetting. No one in 9150 is complaining of their device resetting. This is not the same bug.

Just because his syslog shows two SGX resets before the reboot does NOT imply it's the SGX causing the issue. I've seen SGX resets randomly happen all the time. I see them regularly whenever the device changes orientation, which is common for anyone using the default settings when getting or ending a call. Those two messages probably have nothing to do with why his device resets.

Prove it, please, if you have any evidence of "If the watchdog were resetting because of SGX failures my device would be resetting."

It is very significant to understand a problem. For now it seems that one N900 is little more loaded then another N900 and in one N900 watchdog happens but another one happily avoid that because CPU has a little resource for BME to feed watchdog. Some silicon deviations may pay here too.

No, it is not. If you see SGX failure in syslog it means a trouble. And it is a very low probability what two troubles happen in the same time and are not linked.

It doesn't. I can write a trivial program that triggers lots of them and the system will recover nicely. Also, I sometimes get sgx resets while on normal operation, and I never experienced #9150 (despite my long battles with the stupid sgx).

I do have proof: Run syslog on your N900. Load ANY app that changes device orientation, and change orientation a few times, then look at your log. You will very likely see an SGX reset message or two. It happens all the time. The watchdog does not watch for those messages.

Bug 9150 is about the sgx kernel process spinning in a tight loop and never recovering from it's call to reset the SGX subsystem/hardware. It is not about the N900 resetting. Nobody voting for it, or watching, or commenting on the bug has ever reported N900 reboots. Not one. Sgx hanging does not trigger watchdog reboots. Period.

Now how about you prove something you stated in bug 9150? PROVE you can reset the SGX driver and restart X when it's in this hung state. You said it was possible 3 times in your last comment on the bug, yet provide no method for how to do so. The Nokia engineer responding to the bug even seems to indicate it's not possible, yet you still keep insisting it is. Please prove it.

Wrong and wrong. The warning is NOT an indication of SGX failure. It's an indicator that the video chip is being reset for some reason. There are very many valid reasons to reset the chip in non-failure modes. Changing video out resets the chip. Changing orientation resets the chip. Changing some parameters for video playback optimization resets the chip. That's not failure, that's normal operation.

You need to educate yourself about this issue before spreading more FUD here, claiming others don't know what they're talking about. You clearly are not grasping what's going on in either case.

These are two separate bugs. There may be some cross-over relation, but just because you're seeing SGX resets during a time you would normally see them (device image rotation), does not indicate they are the same bug. Even the result (one spinning in a loop resetting the SGX sub-system, the other rebooting) indicate they are two very different issues.

Simply put, you are wrong here. Stop spreading FUD.

I don't have the phone orientation in swap mode. I set the phone in portrait mode. I also set dorian in portrait mode. I'll switch both to landscape mode to see if it resolves the problem.

I also set the phone option of "Launch by turning". I'll turn that off as well.

I'd suspect the launch by turning will work without the orientation change. Again, I'm not sure any of this is the culprit, just things that may be that can't hurt to try. Installing ssh and logging in remotely when it happens would give a good deal of detail too if you disable the watchdog temporarily.

Two other questions: When was your device manufactured (sticker on the box or under the battery should say), and how easy is this for you to reproduce? Does it happen every time, or just sometimes?

Here is the latest development. As I mentioned, I turned off portrait mode of the phone, "launch by turning" of the phone, and portrait mode of dorian. It lasted 4 or 5 days without rebooting. One thing to mention is that during this period of time, I used cell phone data instead of wifi most of the time.

A day ago, I started to use dorian in portrait mode again. And also I started to use wifi. Today, it rebooted. And here is the log:

Dec 29 13:32:28 Nokia-N900-42-11 systemui-tklock[1123]: Method call received from: :1.12, iface: com.nokia.system_ui.request, method: tklock_closeDec 29 13:32:28 Nokia-N900-42-11 kernel: [38717.841705] kb_lock (GPIO 113) is now open
Dec 29 13:32:33 Nokia-N900-42-11 browser[1819]: GLIB CRITICAL ** Gtk - gtk_widget_show: assertion `GTK_IS_WIDGET (widget)' failed
Dec 29 13:32:33 Nokia-N900-42-11 browser[1819]: GLIB WARNING ** GLib-GObject - /home/bifh4/fremantle-arm-fremantle1.2.cs2007q3/work/glib2.0-2.20.3/g
object/gsignal.c:1107: no emission of signal "show" to stop for instance `0xe71a0'
Dec 29 13:33:32 Nokia-N900-42-11 kernel: [38781.630798] kb_lock (GPIO 113) is now closed
Dec 29 13:33:32 Nokia-N900-42-11 systemui-tklock[1123]: Method call received from: :1.12, iface: com.nokia.system_ui.request, method: tklock_open
Dec 29 13:33:32 Nokia-N900-42-11 kernel: [38781.951110] kb_lock (GPIO 113) is now open
Dec 29 13:33:34 Nokia-N900-42-11 kernel: [38784.115234] kb_lock (GPIO 113) is now closed
Dec 29 13:33:34 Nokia-N900-42-11 systemui-tklock[1123]: Method call received from: :1.12, iface: com.nokia.system_ui.request, method: tklock_close
Dec 29 13:33:34 Nokia-N900-42-11 kernel: [38784.474548] kb_lock (GPIO 113) is now open
Dec 29 13:33:43 Nokia-N900-42-11 cellular: csd[800]: ISI_SMS .308964> ind_reg_status(): Net registration (ind) status:1 rc:0
Dec 29 13:33:43 Nokia-N900-42-11 cellular: csd[800]: ISI_SMS .309635> set_timeout(): Timeout 2133 s event type:-1
Dec 29 13:33:55 Nokia-N900-42-11 cellular: csd[800]: ISI_SMS .213535> ind_reg_status(): Net registration (ind) status:1 rc:0
Dec 29 13:33:55 Nokia-N900-42-11 cellular: csd[800]: ISI_SMS .214207> set_timeout(): Timeout 2121 s event type:-1
Dec 29 13:35:24 Nokia-N900-42-11 cellular: csd[800]: ISI_SMS .929386> ind_reg_status(): Net registration (ind) status:1 rc:0
Dec 29 13:35:24 Nokia-N900-42-11 cellular: csd[800]: ISI_SMS .932804> set_timeout(): Timeout 2032 s event type:-1
Dec 29 13:35:34 Nokia-N900-42-11 cellular: csd[800]: ISI_SMS .172886> ind_reg_status(): Net registration (ind) status:1 rc:0
Dec 29 13:35:34 Nokia-N900-42-11 cellular: csd[800]: ISI_SMS .175480> set_timeout(): Timeout 2022 s event type:-1
Dec 29 13:35:55 Nokia-N900-42-11 kernel: [38925.169708] kb_lock (GPIO 113) is now closed
Dec 29 13:35:55 Nokia-N900-42-11 systemui-tklock[1123]: Method call received from: :1.12, iface: com.nokia.system_ui.request, method: tklock_open
Dec 29 13:35:55 Nokia-N900-42-11 kernel: [38925.427612] kb_lock (GPIO 113) is now open
Dec 29 13:36:06 Nokia-N900-42-11 cellular: csd[800]: ISI_SMS .106998> ind_reg_status(): Net registration (ind) status:1 rc:0
Dec 29 13:36:06 Nokia-N900-42-11 cellular: csd[800]: ISI_SMS .109684> set_timeout(): Timeout 1990 s event type:-1
Dec 29 13:36:22 Nokia-N900-42-11 cellular: csd[800]: ISI_SMS .199558> ind_reg_status(): Net registration (ind) status:1 rc:0
Dec 29 13:36:22 Nokia-N900-42-11 cellular: csd[800]: ISI_SMS .202213> set_timeout(): Timeout 1974 s event type:-1
Dec 29 13:37:53 Nokia-N900-42-11 cellular: csd[800]: ISI_SMS .701389> ind_reg_status(): Net registration (ind) status:1 rc:0
Dec 29 13:37:53 Nokia-N900-42-11 cellular: csd[800]: ISI_SMS .704044> set_timeout(): Timeout 1883 s event type:-1
Dec 29 13:38:12 Nokia-N900-42-11 cellular: csd[800]: ISI_SMS .228275> ind_reg_status(): Net registration (ind) status:1 rc:0
Dec 29 13:38:12 Nokia-N900-42-11 cellular: csd[800]: ISI_SMS .230900> set_timeout(): Timeout 1864 s event type:-1
Dec 29 13:38:37 Nokia-N900-42-11 cellular: csd[800]: ISI_SMS .035191> ind_reg_status(): Net registration (ind) status:1 rc:0
Dec 29 13:38:37 Nokia-N900-42-11 cellular: csd[800]: ISI_SMS .037846> set_timeout(): Timeout 1839 s event type:-1
Dec 29 13:38:48 Nokia-N900-42-11 cellular: csd[800]: ISI_SMS .773869> ind_reg_status(): Net registration (ind) status:1 rc:0
Dec 29 13:38:48 Nokia-N900-42-11 cellular: csd[800]: ISI_SMS .777622> set_timeout(): Timeout 1828 s event type:-1
Dec 29 13:39:12 Nokia-N900-42-11 cellular: csd[800]: ISI_SMS .326389> ind_reg_status(): Net registration (ind) status:1 rc:0
Dec 29 13:39:12 Nokia-N900-42-11 cellular: csd[800]: ISI_SMS .329044> set_timeout(): Timeout 1804 s event type:-1
Dec 29 13:40:53 Nokia-N900-42-11 cellular: csd[800]: ISI_SMS .373844> ind_reg_status(): Net registration (ind) status:1 rc:0
Dec 29 13:40:53 Nokia-N900-42-11 cellular: csd[800]: ISI_SMS .377628> set_timeout(): Timeout 1703 s event type:-1
Dec 29 13:42:38 Nokia-N900-42-11 cellular: csd[800]: ISI_SMS .037174> ind_reg_status(): Net registration (ind) status:1 rc:0
Dec 29 13:42:38 Nokia-N900-42-11 cellular: csd[800]: ISI_SMS .038334> set_timeout(): Timeout 1598 s event type:-1
Dec 29 13:43:11 Nokia-N900-42-11 syslogd 1.5.0#5maemo7+0m5: restart.
Dec 29 13:43:11 Nokia-N900-42-11 BME: pre-start
Dec 29 13:43:11 Nokia-N900-42-11 BME: start
Dec 29 13:43:11 Nokia-N900-42-11 init: console main process (743) terminated with status 2

Hi, do you have that application installed that automatically rotates the screen (I forget the name) but when I had it installed, whenever there was an application that automatically orientated itself, the phone would reboot. Hope this helps!