NAND Testing - 05-25 Update: New LK, Recovery.img, Kernel Updates through Recovery

[MassStash]

Originally Posted by [ACL]

Ok, so if pm.sleep_mode=1 works better for everyone then wanna just make the autobuild use that as well ? i suppose i should remove force cdma and have the initrd add that to the cmd line since we can detect the variant huh..

so much to do, so little time.. need to figure out a way to quit my day job so i can concentrate on this :-p

yea i agree, get paid to do this would be the dream eh?

[natemcnutty]

Originally Posted by [ACL]

Ok, so if pm.sleep_mode=1 works better for everyone then wanna just make the autobuild use that as well ? i suppose i should remove force cdma and have the initrd add that to the cmd line since we can detect the variant huh..

so much to do, so little time.. need to figure out a way to quit my day job so i can concentrate on this :-p

There are two catches that I've been thinking about. First, when does machine_variant get populated? Second, what about CDMA users who use a SIM card? I'd love to find a way to detect the actual SIM card, and then you don't need machine_variant and will always use a SIM if available

For the pm.sleep_mode=1, it works great for me, but I don't know if it lowers performance or messes with drivers or anything. It seems to be stable and works well, even with an overclock, but I'd be curious to see what others think about that the full package I uploaded (since I removed force_cdma and set pm.sleep_mode=1 in that one).

[Lmiller1708]

Originally Posted by natemcnutty

There are two catches that I've been thinking about. First, when does machine_variant get populated? Second, what about CDMA users who use a SIM card? I'd love to find a way to detect the actual SIM card, and then you don't need machine_variant and will always use a SIM if available

For the pm.sleep_mode=1, it works great for me, but I don't know if it lowers performance or messes with drivers or anything. It seems to be stable and works well, even with an overclock, but I'd be curious to see what others think about that the full package I uploaded (since I removed force_cdma and set pm.sleep_mode=1 in that one).

Nate your CDMA detection did not work for me.
It looks like you should remove the quotes on the 1, like this...

if [ "$CDMA" = 1 ] ; then
sed -i -e "s:$: htc_hw.force_cdma=1" /tmp/cmdline
fi

[[ACL]]

Originally Posted by natemcnutty

There are two catches that I've been thinking about. First, when does machine_variant get populated? Second, what about CDMA users who use a SIM card? I'd love to find a way to detect the actual SIM card, and then you don't need machine_variant and will always use a SIM if available

For the pm.sleep_mode=1, it works great for me, but I don't know if it lowers performance or messes with drivers or anything. It seems to be stable and works well, even with an overclock, but I'd be curious to see what others think about that the full package I uploaded (since I removed force_cdma and set pm.sleep_mode=1 in that one).

ahh ic.. well then i still should remove the force cdma and let the users do it on the froyo.conf i suppose.

No idea about sim detection. I'll need to consult one of the guys who has ril experience.

Well if you feel that sleep=1 is good let me know.. i have no service so i cant use my device full time anymore to really test it.

[psu2fan]

Nate,
does the full package in OP...

"Full Packages
[2011-01-09] - Multiupload.com - upload your files to multiple file hosting sites!
Note: This package may shut down on first boot. This is common and you just need to turn it back on once and let it settle.
This package contains the following files:
All Android media and app files
RootFS (2010-12-15)
NBH/Modules (2010-11-29)
gralloc (2010-11-29)
libGLES (2010-11-30)
System.ext2 (FRX03)"

contain the 1/07 nbh & modules as well as the newest rootfs? It's just a bit confusing with the dates you have listed. I DL'ed the nbh/mod and newest rootfs in the OP anyway and used them in place of the ones on the full package.

Also, what is everyone using for their .conf file?

[psu2fan]

FYI with the above files I'm not getting anywhere near the battery life as Nate (maybe your setup is different than mine?) 2.5 hours I'm at 68%

I can confirm as well that the windows key can wake screen, the OS is snappy after doing the Vmail trick, and only sometimes I have issues tyyyypppping.

I have RHOD400, newer one though manufactured 10/10 if it makes a difference

[coolsilver]

Mine has been stable and working great. I did kill all the running processes and was sluggish today a little. Otherwise wifi was working fine.

[m4f1050]

Installed latest package and GPS working again, but not Wifi...I still get ERROR...

For the WINDOWS key to not unlock you will need to setup one of the screen unlock passwords. I set mine up with a pattern one, now windows just wakes phone up and throws you to pattern password, haven't tried any other...

[MassStash]

i think i f'ed up my kb bl...

[natemcnutty]

Originally Posted by Lmiller1708

Nate your CDMA detection did not work for me.
It looks like you should remove the quotes on the 1, like this...

if [ "$CDMA" = 1 ] ; then
sed -i -e "s:$: htc_hw.force_cdma=1" /tmp/cmdline
fi

I didn't think the quotes should matter, but I will compile it that way next. I'm wondering if it might be because I'm using the nvram section to assign it rather than the machine_variant section. I just need to find some more time to mess with this stuff

Originally Posted by [ACL]

ahh ic.. well then i still should remove the force cdma and let the users do it on the froyo.conf i suppose.

No idea about sim detection. I'll need to consult one of the guys who has ril experience.

Well if you feel that sleep=1 is good let me know.. i have no service so i cant use my device full time anymore to really test it.

I was kind of wondering if the SIM card would get enumerated as a card slot and mess up the boot because the mmcblk0 became mmcblk1 with a SIM card in. I'm thinking of flashing maurice's nand build that dumps you straight to bash. Then I could check it from there easily.

Originally Posted by psu2fan

Nate,
does the full package in OP...

"Full Packages
[2011-01-09] - Multiupload.com - upload your files to multiple file hosting sites!
Note: This package may shut down on first boot. This is common and you just need to turn it back on once and let it settle.
This package contains the following files:
All Android media and app files
RootFS (2010-12-15)
NBH/Modules (2010-11-29)
gralloc (2010-11-29)
libGLES (2010-11-30)
System.ext2 (FRX03)"

contain the 1/07 nbh & modules as well as the newest rootfs? It's just a bit confusing with the dates you have listed. I DL'ed the nbh/mod and newest rootfs in the OP anyway and used them in place of the ones on the full package.

Also, what is everyone using for their .conf file?

Thank you so much for mentioning this! I totally forgot to fix the dates, but the RootFS and NBH/modules were updated in the full package.

Originally Posted by psu2fan

FYI with the above files I'm not getting anywhere near the battery life as Nate (maybe your setup is different than mine?) 2.5 hours I'm at 68%

I can confirm as well that the windows key can wake screen, the OS is snappy after doing the Vmail trick, and only sometimes I have issues tyyyypppping.

I have RHOD400, newer one though manufactured 10/10 if it makes a difference

I'm going to chalk it up to our batteries suck for trying to get metrics on. The algo on Wince isn't even accurate, so I can't expect Android to be. I made it 13 hours from taking it off the charger to completely dead, so I think the measurements were just off.

Originally Posted by m4f1050

Installed latest package and GPS working again, but not Wifi...I still get ERROR...

For the WINDOWS key to not unlock you will need to setup one of the screen unlock passwords. I set mine up with a pattern one, now windows just wakes phone up and throws you to pattern password, haven't tried any other...

Have you tried a new data.img? Wifi wasn't working for me at first and gave me an error, but for some reason, now I can't get it to fail again. Figures, when I'm trying to dmesg the problem, I can't repro it...