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

[natemcnutty]

Originally Posted by damage_them

@ nate i just realized that my keyboard is not mapping right..i think i saw acl mention something about that...im rhod400..was this suppose to detect it or is your tinboot specifically for rhod500??

That's one of those things I'm just figuring out. From reading through the IRC logs, it looked like Wozz got a single NBH for all devices, but he was still working on getting the keyboard layouts. So I guess it is currently only using the rhod500 (vzw) layout.

[[ACL]]

Originally Posted by natemcnutty

ACL, I'd be happy to maintain this, and rootfs really doesn't change much. From reading through init (init - eclair-rootfs in XDANDROID Eclair - Gitorious), it looks like I just change lines 430 through 505 to basically replace "if /bin/grep -c 'physkeyboard=fuze' /proc/cmdline >/dev/null" with "if -f /sdcard/rhod500" or and each subsequent model?

I will start packaging the CustomRUU with the NBH and everything else that is needed in a folder named SD Card. With Dropbox, I don't have to worry about space or bandwidth, so it just means people will have to download a slightly bigger file.

As for why the 09-29 changes aren't working for me, I'm looking into that today. Unfortunately, the wife wants me to go shopping with her which could take a while. I'll have some free time while I watch the U of O vs Stanford game in a few hours though

you got it.. those changes for the rootfs should work just fine. I dont think i can get those committed, the best i can do is convince them to be more nand friendly but its not worth it since Rhod is the only nand device. We may need to just do it ourselves like the hd2 guys.

as for the changes on 9/30. You mean the kernel changes here ?

Commit in linux-msm-rhod-nand in Linux on Qualcomm's MSM - Gitorious

Those changes actually enable 3d. The latest rootfs has the gralloc lib needed for 3d to work, so make sure you have that. Also we decided to have 3d always enabled on all builds (including haret) unless the force_3d=0. So if you think 3d is breaking the startup, mod your cmd line on tinboot for no 3d and it will default back to the old settings.

If that does not work, then i need some dmesgs / logcats. Give it a try.

[[ACL]]

Originally Posted by natemcnutty

That's one of those things I'm just figuring out. From reading through the IRC logs, it looked like Wozz got a single NBH for all devices, but he was still working on getting the keyboard layouts. So I guess it is currently only using the rhod500 (vzw) layout.

Lets just mod the rootfs and call it a day .. you already know how to do it so feel free to butcher it.

[natemcnutty]

Originally Posted by [ACL]

you got it.. those changes for the rootfs should work just fine. I dont think i can get those committed, the best i can do is convince them to be more nand friendly but its not worth it since Rhod is the only nand device. We may need to just do it ourselves like the hd2 guys.

as for the changes on 9/30. You mean the kernel changes here ?

Commit in linux-msm-rhod-nand in Linux on Qualcomm's MSM - Gitorious

Those changes actually enable 3d. The latest rootfs has the gralloc lib needed for 3d to work, so make sure you have that. Also we decided to have 3d always enabled on all builds (including haret) unless the force_3d=0. So if you think 3d is breaking the startup, mod your cmd line on tinboot for no 3d and it will default back to the old settings.

If that does not work, then i need some dmesgs / logcats. Give it a try.

Excellent. I'm heading over to a friends for the game, and I'll make changes to rootfs over there. I did notice that init uses $notnand=0 to qualify if we are nand booting, so maybe we could utilize that unless it is actually looking for data stored on nand as well.

Also, is it possible to write a provisioning program that would store values that can be used on next boot? We could have it present the user with options on phone model, GSM vs CMDA, and anything else we want on first boot, and then use those values for subsequent boots.

As far as the changes on 9-30, it must require me to make changes to tinboot, but apparently I'm not making the right changes. I am on the absolute latest rootfs direct from XDAndroid like you recommended. So far, I have tried leaving force_hw3d=1 in tinboot, setting force_hw3d=0, and leaving it out entirely, but the phone still cannot get past the animation. dmesg isn't helpful at all for this one, but I'll try to get a logcat for you. Adding a custom startup command of "su logcat >> /sdcard/logcat.txt" should work right?

[[ACL]]

Originally Posted by natemcnutty

Excellent. I'm heading over to a friends for the game, and I'll make changes to rootfs over there. I did notice that init uses $notnand=0 to qualify if we are nand booting, so maybe we could utilize that unless it is actually looking for data stored on nand as well.

Also, is it possible to write a provisioning program that would store values that can be used on next boot? We could have it present the user with options on phone model, GSM vs CMDA, and anything else we want on first boot, and then use those values for subsequent boots.

As far as the changes on 9-30, it must require me to make changes to tinboot, but apparently I'm not making the right changes. I am on the absolute latest rootfs direct from XDAndroid like you recommended. So far, I have tried leaving force_hw3d=1 in tinboot, setting force_hw3d=0, and leaving it out entirely, but the phone still cannot get past the animation. dmesg isn't helpful at all for this one, but I'll try to get a logcat for you. Adding a custom startup command of "su logcat >> /sdcard/logcat.txt" should work right?

The only part that changes on tinboot is the cmd line. Here is what i use and it works for me

"gsensor_axis=2,1,3 msmvkeyb_toggle=off htc_hw.force_cdma=1 hw3d.force=1 physkeyboard=rhod400 panic=0 pm.sleep_mode=2"

I pushed some other changes today so you may need to try those out. In anycase ive never heard of bootanimation crashing anything on our builds. So im really baffled by your issue

[sykosoft]

The problem I have (and why I have to use haret) is because being a rhod400, I sometimes use CDMA, and when travelling, I use GSM. I can't use both with the tinboot solution, it ONLY allows me to use CDMA. Any possible solution to this?

[natemcnutty]

Originally Posted by [ACL]

The only part that changes on tinboot is the cmd line. Here is what i use and it works for me

"gsensor_axis=2,1,3 msmvkeyb_toggle=off htc_hw.force_cdma=1 hw3d.force=1 physkeyboard=rhod400 panic=0 pm.sleep_mode=2"

I pushed some other changes today so you may need to try those out. In anycase ive never heard of bootanimation crashing anything on our builds. So im really baffled by your issue

I did compile with the new changes, and still the same problem. About 20 flashes in now, and so far, I've tried with hw3d.force=0, hw3d.force=1, and removing it entirely (and every combination I can think of). I've even tried your exact startup and no dice (moved files to root of card for that). Any chance you can post your froyo.user.conf? To add to the fun, every once in a while, I actually get the animation going forever (let it run for 30 minutes). I even pulled my data off the card and formatted it while I was doing all of this to make sure the card wasn't corrupt.

On the rootfs part, I'm still messing with this. I'm re-working their script a bit so I can keep my changes

dmesg doesn't help me catch whatever is causing this because it stops logging before it crashes out, but it crashes before I get to a point I could turn it again. I think the only option is ADB that I still haven't been able to get working

[damage_them]

hey im back on wozzers test 7 for now..hopefully u guys get everything sort out and get everything running properly...good luck and keep up the good work guys!!

[natemcnutty]

Wirelessly posted (Opera/9.80 (Windows Mobile; Opera Mini/5.1.21594/20.2485; U; en) Presto/2.5.25)

Originally Posted by damage_them

hey im back on wozzers test 7 for now..hopefully u guys get everything sort out and get everything running properly...good luck and keep up the good work guys!!

Is there something better on Test 7 than the 09-25 build? Just curious if anything was working in that one that isn't anymore

Also, I think I got the keyboard config thing down with a modded rootfs. I'll finish my testing when I get home tonight and upload it along with the latest compile from ACL's changes.

I still haven't gotten the hw3d changes to work, but I'm not giving up. Just went back to WM for the drive today so I could have bluetooth

[damage_them]

Originally Posted by natemcnutty

Wirelessly posted (Opera/9.80 (Windows Mobile; Opera Mini/5.1.21594/20.2485; U; en) Presto/2.5.25)



Is there something better on Test 7 than the 09-25 build? Just curious if anything was working in that one that isn't anymore

Also, I think I got the keyboard config thing down with a modded rootfs. I'll finish my testing when I get home tonight and upload it along with the latest compile from ACL's changes.

I still haven't gotten the hw3d changes to work, but I'm not giving up. Just went back to WM for the drive today so I could have bluetooth

nope..its actually way slower..but the keyboard layout is why i switched back ..