Help - Search - Members - Calendar
Full Version: Suspend Crashes
OESF Forums > Distros, Development, and Model Specific Forums > Distro Support and Discussion > OpenBSD
Pages: 1, 2
jpmatrix
well well,
i think we still have some resume troubles: this morning i hit the ON button, then the screen resume with hard drive error messages like this one :

wdc0 channel 0: reset failed for drive 0

wd0a: device timeout reading fsbn 1595200 of 1595200-0 (wd0 bn 1595201; cn 1582 tn 8 sn 41), retrying

wdc0:0:0: not ready,st=0x88<BSY,DRQ>, err=0x00

and the harddrive green light stays on.........


i'll google search for this messages
ZDevil
QUOTE(jpmatrix @ May 12 2007, 09:50 AM)
well well,
i think we still have some resume troubles: this morning i hit the ON button, then the screen resume with hard drive error messages like this one :

wdc0 channel 0: reset failed for drive 0

wd0a: device timeout reading fsbn 1595200 of 1595200-0 (wd0 bn 1595201; cn 1582 tn 8 sn 41), retrying

wdc0:0:0: not ready,st=0x88<BSY,DRQ>, err=0x00

and the harddrive green light stays on.........


i'll google search for this messages
*


Hey, this is exactly what i've run into almost all the time! So i am not the only one... dry.gif
The constantly lit green light seems to indicate that the system is trying desparately to get the harddrive resumed, but keep failing ...
So now i am quite sure that there is something wrong with apm in the builds.
Yet I've installed the 4.1 release version. It seems zzz works again. Perhaps the release version is more stable? Will test further.
One difference i notice in the 4.1 release build: the green LED is always on even though the hard drive is not spinning.
Haven't tried "lidsuspend", which doesn't seem to work at all, i'm afraid.
(But see my next post below... rolleyes.gif)
jpmatrix
the release version seems to be rather old on
ftp://ftp.openbsd.org/pub/OpenBSD/4.1/zaurus : march 2007

latest snapshot is may 08

i wonder if i wait for the next snapshot...
mathemajikian
QUOTE(jpmatrix @ May 12 2007, 10:17 AM)
the release version seems to be rather old on
ftp://ftp.openbsd.org/pub/OpenBSD/4.1/zaurus  : march 2007

latest snapshot is may 08

i wonder if i wait for the next snapshot...
*


or better yet why not update your sources and rebuild the kernel and user world? Then you wont have to wait for the next snapshot.
ZDevil
QUOTE(jpmatrix @ May 12 2007, 12:17 PM)
the release version seems to be rather old on
ftp://ftp.openbsd.org/pub/OpenBSD/4.1/zaurus  : march 2007

latest snapshot is may 08

i wonder if i wait for the next snapshot...
*



True. But interestingly the release version does seem more stable than the snapshots: it is my very first time that i get zzz, lidsuspend and the power off button ALL working! wink.gif
Still let me see if things are really stable during the weekend.
gen2
To add another data point here, these errors

CODE
wdc0:0:0:  not ready, st = 0x88<BSD,DRQ>, err = 0x00
wdc0 channel 0: reset failed for drive 0
wd0f : device timeout reading fsbn 156672 of 156672-0 (wd0 bn 7791264 cn 7729 tn 6 sn 54)


are not exclusive to the 6 partition setup. I don't know the precise sequence of commands which wake the drive and start addressing the hardware, but I'd like to know if multiple partitions is truly the issue...seems like something else changed which has lowered a threshold somewhere.

Problem is persistent since at least 3.8, however I never saw any errors of any kind on 4.0, with a 6 partition setup. Smells like beta buglets...

Discussion of the issue dates back at least here:
http://archive.netbsd.se/?ml=openbsd-arm&a=2005-12&t=1551843

I get this using todays snapshot, one / partition and one swap, no config edits of any kind anywhere after booting (other than adding apmd):

CODE
wd0a: aborted command reading fsbn 5397260 of 5397260-0 (wd0 bn 5397323; cn 5354 tn 7 sn 50), retrying
wd0: soft error corrected


However it seems to work.

I get this using todays snapshot, using the 6 partition setup, no config edits of any kind anywhere after booting (other than adding apmd):

CODE
wd0a: aborted command reading fsbn 4982464 of 4982464-0 (wd0 bn 6002560; cn 5954 tn 14 sn 46), retrying
wd0: soft error corrected


And this too seems to work. zzz -S is causing sleep, and it wakes up, gives that error, but works.

Investigation continues...
mathemajikian
QUOTE(ZDevil @ May 12 2007, 11:42 AM)
True. But interestingly the release version does seem more stable than the snapshots: it is my very first time that i get zzz, lidsuspend and the power off button ALL working!  wink.gif
Still let me see if things are really stable during the weekend.


Last night I suspended my Z from within X using the lidsuspend option and when I resumed the device this morning the only problem was that the back light failed to turn back on. I saw a half second white flash on the screen, as if it was attempting to turn on, then a blank/black screen. I ssh'd into the Zaurus and issued wsconsctl display.back light=100 and everything returned to normal. I also used lidsuspend outside of X and got similar results. (The back light failed to turn back on) However, if I used the power on/off button in X or from the console the back light never failed to turn back on. So there is an issue with lidsuspend, but it dosn't result in complete catastrophic device lockup. Though out the day I've suspended and resumed the device via the on/off button and lidsuspend multiple times and never once had the Z lock up or produce the error you describe below:

wdc0 channel 0: reset failed for drive 0
wd0a: device timeout reading fsbn 1595200 of 1595200-0 (wd0 bn 1595201; cn 1582 tn 8 sn 41), retrying
wdc0:0:0: not ready,st=0x88<BSY,DRQ>, err=0x00.

Finally, in regards to your battery status concerns:

From the apmd manual
QUOTE
 
            -t seconds
            apmd periodically polls the APM driver for the current power
            state.  If the battery charge level changes substantially or the
            external power status changes, the new status is logged.  The
            polling rate defaults to once per 10 minutes, but may be speci-
            fied using the -t command-line flag.


Therefore to get better battery status reports add the following to rc.conf.local

apmd_flags="-t 300"

This changes the polling rate to once every 5 minutes instead of once every 10 minutes.

You might also want to check out: http://ports.openbsd.nu/x11/asapm
mathemajikian
CODE
wd0a: aborted command reading fsbn 5397260 of 5397260-0 (wd0 bn 5397323; cn 5354 tn 7 sn 50), retrying
wd0: soft error corrected


This is normal. If you don't want to see these messages then redirect the output.
gen2
QUOTE(mathemajikian @ May 12 2007, 05:29 PM)
CODE
wd0a: aborted command reading fsbn 5397260 of 5397260-0 (wd0 bn 5397323; cn 5354 tn 7 sn 50), retrying
wd0: soft error corrected


This is normal. If you don't want to see these messages then redirect the output.
*


Perhaps normal for 4.1, I never saw them on 4.0.
mathemajikian
QUOTE(mathemajikian @ May 12 2007, 05:23 PM)
So there is an issue with lidsuspend, but it dosn't result in complete catastrophic device lockup.

Just noticed the following from 3.9 changes located here: http://www.openbsd.org/plus39.html

QUOTE
On Zaurus, allow sysctl machdep.lidsuspend=1 to cause suspend on screen closure. Opening does not yet wake it up though.

The Z wakes up it just fails to turn the back light back on.
gen2
QUOTE(mathemajikian @ May 12 2007, 07:31 PM)
QUOTE(mathemajikian @ May 12 2007, 05:23 PM)
So there is an issue with lidsuspend, but it dosn't result in complete catastrophic device lockup.

Just noticed the following from 3.9 changes located here: http://www.openbsd.org/plus39.html

QUOTE
On Zaurus, allow sysctl machdep.lidsuspend=1 to cause suspend on screen closure. Opening does not yet wake it up though.

The Z wakes up it just fails to turn the back light back on.
*



Current testing here produces full lockup on lidsuspend, no ssh'ing in possible, no status lights a flashin. I'm running 6 partitions, sleep DOES work as advertised using # zzz -S. Unable to verify any specific link between 6 partition scheme and suspend/wake problems.

I can handle typing instead of just closing the lid. Moving on.
ZDevil
Just another update:

I have tried suspending with closing lid, zzz and power off button several times under the 4.1 release version, with at least half an hour interval between suspend and resume. Everything works so far. smile.gif
Pray that the issue won't come back again...
ZDevil
More update:

I've updated the kernel to 4.1-current (GENERIC) #100 (8 May 2007).
Suspend and resumes now works like a charm!
Even closing lid is no longer a problem.
Tried two days and haven't experienced any resume failure yet, except on one occasion after building stuff overnight the system refused to suspend. A reboot fixed that.

Thanks for the hard work!

[Update+]
Shortly after posting I ran into one incidence of failing to resume.

A handy tip:
Just create either a tiny shell script or alias as shortcuts to the backlight commands: wsconsctl display.backlight=100, wsconsctl display.brightness+=1, wsconsctl display.brightness-=1.
For example, i am using don ("Display On"), bup ("Brightness Up") and bdn ("Brightness Down") respectively. wink.gif
ZDevil
Updated to GENERIC#110 last night. Suspend by closing lid and pressing power button quite a few times. Resume hasn't failed yet.

GREAT JOB! THANK YOU FOR THE CONSTANT AND PERSISTENT UPDATING! wink.gif

I really hope I can contribute my small part soon by submitting packages to the repository.
I've been collecting OBSD tips and making notes of them. Will try to make a newbie starter guide in addition to the great installation documentations.
jpmatrix
Great piece of news ! I think i ll be back to openbsd as soon as i come back from holidays smile.gif
any news about bluetooth and sd card support in latest builds ?
jpmatrix
well, it seems suspend/resume troubles are now gone smile.gif
i've zzz -S a couple of time and resume always worked for me smile.gif let's touch some wood ! (french proverb wink.gif

however i've some troubles when i suspend and the usb host cable is plugged (and perhaps pppd is running): when i resume i got lot of messages which fills the screen and i cannot do anything other than reboot.....

messages are:
uhub0: device problem, disabling port 2

would someone help?
mathemajikian
QUOTE(jpmatrix @ Jun 4 2007, 10:04 AM)
however i've some troubles when i suspend and the usb host cable is plugged (and perhaps pppd is running): when i resume i got lot of messages which fills the screen and i cannot do anything other than reboot.....

messages are:
uhub0: device problem, disabling port 2

would someone help?

Disconnect usb host cable before suspending? smile.gif
jpmatrix
QUOTE(mathemajikian @ Jun 4 2007, 01:32 PM)
QUOTE(jpmatrix @ Jun 4 2007, 10:04 AM)


however i've some troubles when i suspend and the usb host cable is plugged (and perhaps pppd is running): when i resume i got lot of messages which fills the screen and i cannot do anything other than reboot.....

messages are:
uhub0: device problem, disabling port 2

would someone help?

Disconnect usb host cable before suspending? smile.gif
*



lol i didn't thought about it biggrin.gif
you know this weekend i let the zaurus plugged to the nokia and internet up for quite all saturday and sunday smile.gif compilation stories wink.gif then i said to myself : perhaps the zaurus would need some pause to rest a little wink.gif so sometimes i just switched the backlight off (i created scripts) .
but sometimes you also need to zzz it no? perhaps the suspend/resume process should be modified to disconnect the usb bus before and reconnect it on resume?
jpmatrix
guys,
i'm running into a new problem : suspend just doesn't work !
nothing happens as i type zzz -S ........
just a "suspending system..." message then back to prompt......

any clue before i reboot ?
gaarder
I've tried zzz -S, zzz -z, with and without sd card, cf card, and cf wifi card, with apm -A, apm -At 300, apm -t 300, machdep.maxspeed=520 or #machdep.maxspeed=520...

my new z (3200) fails to shut down its charge light during an overnight suspend charge session (~6hrs)

any idea? when I boot back to sharp rom the charge light usually disappears after a few hours.

Also, if suspend charging under OpenBSD works for anyone, can someone post their configurations?

I am using OpenBSD #current 114 dual booting OpenBSD and sharp original rom. Resume works, but I prefer not to use lidsuspend.
jpmatrix
QUOTE(gaarder @ Jun 13 2007, 05:39 AM)
I've tried zzz -S, zzz -z, with and without sd card, cf card, and cf wifi card, with apm -A, apm -At 300, apm -t 300, machdep.maxspeed=520 or #machdep.maxspeed=520...

my new z (3200) fails to shut down its charge light during an overnight suspend charge session (~6hrs)

any idea? when I boot back to sharp rom the charge light usually disappears after a few hours.

Also, if suspend charging under OpenBSD works for anyone, can someone post their configurations?

I am using OpenBSD #current 114 dual booting OpenBSD and sharp original rom. Resume works, but I prefer not to use lidsuspend.
*


well,
with apmd -A that didn't work for me too...
but with apmd="" or -t 300 it does work
but it probably depends on what you're doing with your Zaurus (wifi ? bluetooth?)
when i let the zaurus up to compile, i'm used to plug my nokia via its usb cable so power requirements are at low level
jpmatrix
QUOTE(jpmatrix @ Jun 12 2007, 01:46 PM)
guys,
i'm running into a new problem : suspend just doesn't work !
nothing happens as i type zzz -S ........
just a "suspending system..." message then back to prompt......

any clue before i reboot ?
*


any news about how to solve this bug ? i'm having it again....
mathemajikian
QUOTE(jpmatrix @ Jun 16 2007, 12:49 PM) *
QUOTE(jpmatrix @ Jun 12 2007, 01:46 PM)
guys,
i'm running into a new problem : suspend just doesn't work !
nothing happens as i type zzz -S ........
just a "suspending system..." message then back to prompt......

any clue before i reboot ?
*


any news about how to solve this bug ? i'm having it again....


Here's some new news:

"Revision 1.3, Mon Oct 8 20:18:19 2007 UTC (2 weeks, 4 days ago)
Branch: MAIN
CVS Tags: HEAD
Changes since 1.2: +2 -1 lines
Diff to previous 1.2
Align the suspend saved data to a cache line boundary, apparently prevents issues upon resume."
This is a "lo-fi" version of our main content. To view the full version with more information, formatting and images, please click here.
Invision Power Board © 2001-2014 Invision Power Services, Inc.