Oz 3.5.4 Rc Bug & Work-around List

[lardman]

Where's the proper place to report this (outside of here, of course!)?

Here is almost certainly not the best place to report it as few of the core developers inhabit this forum.

The bugs.openembedded.org bugtracker is also used for distribution bugs (atm) - you can select whether a bug is about OE or the distro when you submit it.


Si

[lardman]

Is some reason this isn't enabled by default? Does it actually work for *anyone* out of the box?

It depends on the processor - the SA processor equipped machines shouldn't have this enabled, while the PXA ones need it.

I have a feeling it works 'out of the box' for my C750, but that may just have been a left-over from a previous install.

It would be possible (if it's not done already) to make the conf file a machine-depends afaik.


Si

[Viking]

Hello
i tried out the 3.5.4 snaptshot  with 2,6 kernel on my akita c-1000
//edit: ( exactly opie-image-akita-20051209185141 )
very nice
 but i had problems getting my usb-pegasus-ethnet-adapter  to work
so i recompiled the .dev kernel ( 14+15rc2)
i had also a problem with the apm
after resume there was no backlight ( all seems working )
trying to tap the lightapplet on the dark, black screen an hiiting it will result
to an nice working zaurus.

a better solution is adding a new resume/suspendscript
/etc/apm/scripts.d/backlight

------------
#!/bin/sh
if [ "$1" = suspend ] ; then

echo "1"  > /sys/class/backlight/corgi-bl/power
elif [ "$1" = resume ] ; then
echo "0"  > /sys/class/backlight/corgi-bl/power
echo "25"  > /sys/class/backlight/corgi-bl/brightness
fi

: exit 0
------------------------------
don't forget setting the symlinks in
/etc/apm/resume.d
/etc/apm/suspend.d
pointing to the backlight script


but i have no working sound (:-((

[lardman]

Did these problems occur before your kernel rebuild?

Is something (a script etc.) altering the values in /sys/class/backlight/corgi-bl/brightness and /sys/class/backlight/corgi-bl/power after resume, or is this the kernel forgetting?

The scripts shouldn't be necessary as I think the kernel should remember what its settings were before the suspend.

You might want to talk to RP on irc about this.


Si

[Viking]

Did these problems occur before your kernel rebuild?
yes, but at that time i didn't focus on getting running apm. ( network connect was primary goal)


Is something (a script etc.) altering the values in /sys/class/backlight/corgi-bl/brightness and /sys/class/backlight/corgi-bl/power after resume, or is this the kernel forgetting?

no i dont think so ..

The scripts shouldn't be necessary as I think the kernel should remember what its settings were before the suspend.

hmm ..

You might want to talk to RP on irc about this.


Si
[div align=\"right\"][a href=\"index.php?act=findpost&pid=108107\"][{POST_SNAPBACK}][/a][/div]

[Viking]

the best solution is to wait for the next kernel patch

http://www.rpsys.net/openzaurus/temp/pxa_i2c_fix-r0.patch
 works fine on my C-1000
Index: linux-2.6.15-rc2/drivers/i2c/busses/i2c-pxa.c
===================================================================
--- linux-2.6.15-rc2.orig/drivers/i2c/busses/i2c-pxa.c  2005-12-19 09:51:31.000000000 +0000
+++ linux-2.6.15-rc2/drivers/i2c/busses/i2c-pxa.c       2005-12-22 18:21:20.000000000 +0000
@@ -900,6 +900,10 @@
        struct pxa_i2c *i2c = adap->algo_data;
        int ret, i;

+       printk("ICR: %x\n", ICR);
+       if (!(ICR & ICR_IUE))
+               i2c_pxa_reset(i2c);
+
        for (i = adap->retries; i >= 0; i--) {
                ret = i2c_pxa_do_xfer(i2c, msgs, num);
                if (ret != I2C_RETRY)

       

[lardman]

Your chat with RP paid-off then I take it

Si

[Viking]

Anoter problem occured    
on boot i can mount my sd-card
i can write sync and read from it.
after a suspend/resume cycle
i can read directories that i had read from bevore i have done a suspend/resume cyle
( maybe the information is buffered in memory )
but if i try to open an unread directory/file or do a  write ro sync
the process stuck and i cannot kill him .
the only thing I can do is a reset.

any ideas ?


VIking

[obergix]

the best solution is to wait for the next kernel patch

http://www.rpsys.net/openzaurus/temp/pxa_i2c_fix-r0.patch
 works fine on my C-1000

[div align=\"right\"][a href=\"index.php?act=findpost&pid=108123\"][{POST_SNAPBACK}][/a][/div]

Is it available in monotone for regenerating in bitbake ? and in which branch ?

[obergix]

Hello
i tried out the 3.5.4 snaptshot  with 2,6 kernel on my akita c-1000
//edit: ( exactly opie-image-akita-20051209185141 )
very nice

...SNIP...
[div align=\"right\"][{POST_SNAPBACK}][/a][/div]

For the records, this 2.6 image for akita (available at [a href=\"http://ewi546.ewi.utwente.nl/mirror/sam.rpsys.net/openzaurus/autobuild/]http://ewi546.ewi.utwente.nl/mirror/sam.rp...urus/autobuild/[/url] ) was suggested in https://www.oesf.org/forums/index.php?showt...ndpost&p=107230 ... and probably a good place for followup on 2.6 with akita, since the 3.5.4 won't apparently include that version of kernel for akitas, so not really about the 3.5.4RC bugs

[Misiurei]

Still struggling with wlan:
3.4.RC on tosa, gpe. I have flashed R.C onto a cleaned tosa (after nando backup), before I ve been using opie, everything worked fine, I am only not satisfied with the poor text-programmes. For I am a philosopher who is writing, this is important. So gpe because abiword is available. But 3.5.3 was too bugy, installing abiword made the gui instable. So i have tried 3.4. It is really a big step forward, congratulations and thank you for your work!! Only one thing drives me crazy:

Wlan does not work, behaviour: network is detected, tosa is registered, web browser can display google, but nothing more. ipkg doesn’t update (starts with the process, then hangs.).
Is there a workaround?

I ve done some more testing using ping and traceroute. I can ping every http-adress I want, but traceroute stops at about position 12. or 13. It simply doesn't go further. IPKG doesn't work, also sylpheed. Is the ip6-bug responsible for that?

sebastian

[donallen]

I tried 3.5.4RC GPE on my C1000. Looks very nice. WLAN with D-Link DCF-660W worked. The main issue was that the power button did not work. I could not suspend the machine either running on A/C or on battery. I finally pulled the battery for 10 seconds, reinserted and found the machine still didn't respond to the power button, so I couldn't turn it on, either on A/C or battery. I got things working again by installing the pdaxrom RC12 files on my SD card and going through the usual routine to flash the rom, which I did successfully.

[lardman]

I tried 3.5.4RC GPE on my C1000. Looks very nice. WLAN with D-Link DCF-660W worked. The main issue was that the power button did not work. I could not suspend the machine either running on A/C or on battery.

'apm -s' ?

This is the kind of problem which should be added to the bug tracker....

[obergix]

I tried 3.5.4RC GPE on my C1000. Looks very nice. WLAN with D-Link DCF-660W worked. The main issue was that the power button did not work. I could not suspend the machine either running on A/C or on battery.

'apm -s' ?

This is the kind of problem which should be added to the bug tracker....
[div align=\"right\"][{POST_SNAPBACK}][/a][/div]

Already reported, more or less : [a href=\"http://bugs.openembedded.org/show_bug.cgi?id=497]http://bugs.openembedded.org/show_bug.cgi?id=497[/url]

[obergix]

I tried 3.5.4RC GPE on my C1000. Looks very nice. WLAN with D-Link DCF-660W worked. The main issue was that the power button did not work. I could not suspend the machine either running on A/C or on battery. I finally pulled the battery for 10 seconds, reinserted and found the machine still didn't respond to the power button, so I couldn't turn it on, either on A/C or battery. I got things working again by installing the pdaxrom RC12 files on my SD card and going through the usual routine to flash the rom, which I did successfully.
[div align=\"right\"][{POST_SNAPBACK}][/a][/div]

FYI, APM works well (correcting these issues) in the .dev 2.6 generated kernel (tagged revision : -rjan2006prebreakage) ... so we can expect this to be corrected in the 3.5.4.1 release for the C1000 (see announcement in [a href=\"http://openzaurus.org/wordpress/2006/01/09/openzaurus-354-release-plans/)]http://openzaurus.org/wordpress/2006/01/09...release-plans/)[/url].