HD2 Android NAND Updates ONLY [No Discussion]

mskip · Oct 23, 2010

Ok this thread is for people to post Updates about NAND Development on the HD2 ONLY!!

NO DISCUSSION IN THIS THREAD - PLEASE KEEP IT CLEAN..

If anyone posts a message that doesn't contain any update or information about NAND on HD2 then it will be deleted immediately. I will be keeping this thread totally clean apart from updates and direct information on NAND so don't be offended if your post is deleted.

If you want to discuss NAND then please use the Android Lounge thread.

If anyone starts any new threads asking for updates on NAND or asking when NAND will be ready then the thread will be closed.

This is for everyone's benefit so please do not post here unless it fits the above criteria.

------------------------------------------------
------------------------------------------------

PLACES OF INFORMATION

Port progress overview: http://htc-linux.org/wiki/index.php?title=Leo/Status
Users can follow updated by checking out Cotulla's twitter: http://twitter.com/CotullaCode
IRC Logs (not just HD2, look for posts by Markinus, Netripper, LeTama, Gauner1986, Cotulla): http://irclog.netripper.com/?chan=htc-linux&msgonly=1

Mark.

Beretta93 · Oct 24, 2010

Current status for now :

Gauner1986 working on battery driver, with Cotulla's help :

Cotulla how is it?
18:51 gauner1986 i'm still crying about battery-driver
18:51 Cotulla I am very busy on this week, don't listen to news
18:52 Cotulla lol
18:52 Cotulla u got 94%?
18:52 gauner1986 90%
18:52 gauner1986
18:52 Cotulla haha
18:52 gauner1986 it's going down more over time
18:52 Cotulla like in theory
18:53 Cotulla okay I will look when I got time
18:53 gauner1986
18:54 gauner1986 seems it needs some write

And they're starting to figure out what is wrong with ppp code:

Cotulla I can't call this code "good-implemented"
19:00 Cotulla u send AT+CGACT=0,%s
19:00 Cotulla and then u need wait right response
19:00 Cotulla parse it
19:01 Cotulla as well as AT+CGDCONT=1
19:01 Cotulla also
19:01 Cotulla u sure
19:02 Cotulla execution will be stopped at middle of hackWrite() and context will be switched?
19:02 gauner1986 do you think this could be the cause of disconnects?
19:02 Cotulla in this case it may send rubbish instead of right commands
19:03 Cotulla also not sure if WM use ATD*99***1#
19:04 gauner1986 what does it use then?
19:04 gauner1986 .. / how to find out?
19:04 Cotulla as welll as sleep(5); // allow time for ip-up to run
19:04 Cotulla why 5?
19:04 Cotulla not 10?
19:05 Cotulla also in case if remove side will disconnect data?
19:05 Cotulla what will happen?
19:06 gauner1986 hm, yeah.. it definetly needs improvement
19:06 gauner1986 where can i read about the AT reply's ?
19:07 Cotulla "NW DETACH", "NW DEACT", "ME DEACT", "ME DETACH"
19:07 Cotulla this possible notify... dunno if they all used by amss
19:08 Cotulla but who manage reconnect? Android?
19:08 gauner1986 yes
19:08 Cotulla if Android it should know about disconnection
19:08 gauner1986 i guess ppp handles the disconnection, no?
19:09 gauner1986 so, if remote side hangs up, pppd will know and terminate
19:09 Cotulla how it will detect?
19:09 Cotulla data flow will stops
19:09 Cotulla and ?
19:09 gauner1986 pppd times out?
19:09 LeTama pppd dies

Eventually, not related to NAND directly but good to know, dan13jl working on Z/HD camera :

gauner1986 make me a proper battery-driver please
18:42 gauner1986
18:42 LeTama FsTluxe, let's go to -chat
18:43 dan1j3l gauner1986: currently working on camera for HD/Z builds...
18:44 gauner1986 also nice
18:45 dan1j3l apt leo-ramconsole

As far as I know, since power collapse, we also got audio quality partially fixed by LeTama, but can't find the log.

That's it for now.

dusko_m · Oct 25, 2010

22:54 ogonzuhnina sorry to disturb, just wanted to ask if you got any progress/findings 'bout max battery @ android @93%
22:55 ogonzuhnina i'm not sure if it's anything new but i got similar on mu biuld (already month old or so). my battery charged to 97, than 96 now 93 max...
22:56 Markinus ogonzuhnina: yeah, I have a fix and gauner tested it, it works

seems that Markinus and gauner are close to solution on how to fix battery charging problem

one more step in right direction

ceesheim · Oct 25, 2010

and some more good news .

http://irclog.netripper.com/?chan=htc-linux&day=25

14:24 LeTama: hello guys!
14:25 gauner1986: hey letama
14:25 Jack-E: afternoon
14:25 gauner1986: markinus and i got battery problems sorted somehow...
14:25 LeTama: gauner1986: yes, saw that

Samsunguy · Oct 26, 2010

21:12 mdeejay: darkstone you fixed it

21:13 victorbd: hi mdeejay
21:13 hansineGrip: Yes, he did!

21:13 hansineGrip: Fix the stuttering, i mean.
21:14 darkstone: mdeejay o0
21:14 darkstone: checks thread
21:14 LeTama: nice!
21:14 darkstone: ah nice

21:14 darkstone: LeTama: mind pushing it?
21:15 LeTama: sure
21:15 mdeejay: fix from desire threads on XDA?
21:15 darkstone: yes
21:15 darkstone: LeTama: http://pastebin.com/raw.php?i=aRbH2cef
21:16 victorbd: can ny1 giv me an RMnet version of Mdeejay FROyO Z... im in love wid it...

21:16 darkstone: i noticed on linuxtogo that the values were correct
21:16 victorbd:

21:16 darkstone: i was wondering why there was slower read speed in evo kernel

no more music shuttering in nand

mally2 · Oct 26, 2010

00:57 darkstone Aevum: nand may not improve things instantly
00:57 Heh_ we cant find any stattion
00:58 Aevum in other words, start saving up for a desire HD
00:58 darkstone hehe
00:58 darkstone a native android device will obviously be better
00:58 darkstone since its, you know, designed for android

sad news!!!

WHO CARES WE STILL WANT NAND

ghis1 · Oct 26, 2010

14:11 => LeTama joins
14:12 LeTama hello guys
14:12 Cotulla hello LeTama
14:12 Cotulla sound is nice?
14:12 LeTama So back in business Cotulla! Great
14:12 Cotulla yeah

14:13 => rexid joins
14:13 LeTama Sound is working well yes
14:13 Cotulla what about power ?
14:13 LeTama Didn't hear my wife complain anymore
14:13 LeTama Power ?
14:13 Cotulla yes
14:14 LeTama Power is fine for me. I didn't do extensive tests, but it behaves well
14:14 LeTama the only real blackspot remaining is ppp
14:15 Cotulla oh ****\
14:15 Cotulla u tried to do
14:15 Cotulla like I said?\\
14:15 Cotulla (few days ago)

Samsunguy · Oct 27, 2010

16:46 Markinus: gauner1986: good

is finished, only some cleanupts
16:46 gauner1986: cool
16:46 => netson joins
16:46 gauner1986: what was the issue with detection?
16:46 NeoMatrixJR: ~seen WisTilt2
16:46 Markinus: works
16:46 => ElBartoME joins
16:46 apt: wistilt2 <~wisgreg@wireless251.wirelesstcp.net> was last seen on IRC in channel #htc-linux, 9h 1m 34s ago, saying: 'tks'.
16:46 Markinus: synchronisation
16:46 gauner1986: ok
16:46 [acl]: NeoMatrixJR: how goes it
16:47 gauner1986: so then we got a pretty nice port
16:47 gauner1986: battery support fully implemented
16:47 gauner1986: yay

i expect a release very soon

ghis1 · Oct 28, 2010

15:07 luc__ New battery driver seems to work very good, and the standby also.
15:07 luc__ 9 hours on standby overnight just 4%
15:08 <= programmer8922 left/quit
15:10 luc__ one more thing from 99% to 100% takes 20min and the green light turns on at 100%
15:10 => TheDeadCPU joins
15:11 luc__ Markinus: WELL DONE!
15:12 drellisdee oh a kernel question too phh. Code aurora has added a commit to use 800mhz on 7627. It reads pll config and sets freq table from it. My plls are 960,245,1200. The 800mhz uses plls 960,245,800. Other than writing a val of 41 for the mult when its reading the init of plls is there an way to change it?
15:16 => LeTama joins
15:16 <= Neo31 left/quit
15:16 => Neo31 joins
15:20 gauner1986 hey letama
15:20 <= bbz left/quit
15:20 <= Andreyxxl[HD2EU] left/quit
15:20 LeTama hello gauner1986, hello all
15:20 LeTama what's up ? saw that battery driver looks good ?
15:21 <= iDavid left/quit
15:21 gauner1986 yeah
15:21 gauner1986 it's really good
15:21 manusfreedom hi, can we have an autobuild for a RMNET version of the kernel (evo based)?
15:21 gauner1986 but nand corruption is bugging me now
15:21 gauner1986

15:22 LeTama gauner1986: that's bad. Saw that Cotulla had it too ?
15:22 gauner1986 yes
15:23 gauner1986 hope he's working on it
15:23 gauner1986

15:23 LeTama

Sweet

9 hours on standby overnight just 4%
-- Who said we can't get better battery life with NAND???

Though some of them seems to have little problems with corruption

quertox · Oct 28, 2010

http://twitter.com/hd2_android

# Battery driver improved. Charging works correctly now. vor ungefähr 22 Stunden via HTC Peep

rajkosto · Oct 28, 2010

CM6.1 RC1 came out today and is running smoothly in my nand build... no problems other than we still have to use ppp rather than rmnet

confuse1990 · Oct 29, 2010

11:47 gauner1986 cass: ah nice.. see you found out about ppp issues?
11:48 Cass yeah we got some issues noted
11:48 Cass have a crude workaround
11:49 Cass not elegant at all
11:49 Cass but seems to work
11:49 gauner1986 checker scirpt if ppp dies?
11:49 Cass yep
11:49 gauner1986 what if we disable data connection?
11:49 Cass and letamas tool to push at command when it does
11:49 gauner1986 and how do you restart ppp?
11:49 Cass service in init.rc

not the best issue, but im happy about it

luckych · Oct 29, 2010

[Delenn] corruption solved for u, Cass?
16:20 Cass yes
16:21 Cass used config to put all dalvik-cache on /data
16:21 Cass cant corrumpt any more
16:21 [Delenn] so is it android fault?
16:21 [Delenn] or yaffs?
16:21 Cass also was same when i had /cache on proper partition too
16:21 Cass its build fault i think

guess they fixed the corruption

shadiku · Oct 29, 2010

Two new commits from LeTama.

e7a2876: htcleo: change himem/lomem repartition to have more low memory
1b31881: htcleo: fix htcleo_defconfig to match kernel Kconfig

Samsunguy · Oct 31, 2010

What you all have been waiting for about NAND

1026 now
19:01 gauner1986 maybe wait until the media **** has finished scanning....
19:04 gauner1986 okay.. 2d and 3d performance seem fine now
19:04 =>*gnutoo joins
19:04 gauner1986 but I/O performance is ****ty
19:06 gauner1986 is nand slower than sd ?
19:07 <=*Sharky^ger left/quit
19:07 <=*FsTluxe left/quit
19:08 <=*GNUtoo|laptop left/quit
19:08 gauner1986 rajko: what do you have in quadrant on CPU and IO ?
19:13 =>*arrrghhh joins
19:14 =>*baluns joins
19:18 HardDisk_WP dcordes, lemme check
19:18 HardDisk_WP yep, process died
19:18 HardDisk_WP restarting#
19:19 =>*unilinky joins
19:20 HardDisk_WP dcordes, thx 4 notice#
19:22 <=*yggdrasi1 left/quit
19:26 =>*slim_ joins
19:28 Rajko io is 600
19:28 Rajko memory is 1000
19:28 gauner1986 cpu = ?
19:28 Rajko cpu varies between runs
19:28 Rajko so why bother
19:28 gauner1986 yeah.. but it stays around some value
19:28 gauner1986 i worry because mdeejay had 4200
19:29 gauner1986 and i have 3100
19:29 Rajko run it again
19:29 Rajko i have 4032 first run
19:29 Rajko 4220 second
19:29 gauner1986 i have around 3100 always
19:29 gauner1986 okay, maybe it's due to the undervolt
19:30 gauner1986 but it shouldn't be
19:30 Rajko 4586 3rd run
19:30 <=*aro left/quit
19:32 gauner1986 hm
19:32 gauner1986 okay.. so nand is just slower than sd it seems
19:32 Rajko im running from nand
19:32 gauner1986 i mean in terms of I/O
19:32 Rajko yeah thats obvious
19:32 Rajko since writes are slower
19:32 Rajko 600 io here

So running Android from the sd card is a whole lot faster than NAND at this point of course

sp33dy_gtx · Oct 31, 2010

status

Development Status:

Bad Audio Quality Fixed So Far
Battery Driver Fully Implemented
Nand Corruption Fixed

What to do:

Fixing Some Minor Bug's
Fixing RMNET PPP Data Issues

What is understand: nand is coming and for post above it will be slow..

Beretta93 · Nov 1, 2010

An important information I think that everyone must understand before assuming any soon or distant release :

- In order to "install" any Android build into NAND, you need to flash Cotulla's MAGLDR within the phone first. The logs that are posted on this thread are about developping Android kernel with specific drivers for NAND builds. But EVEN if this kernel shows steady improvement, and even if it was perfect by tomorrow, you would still need MAGLDR to launch it, so the key release before any NAND build is MAGLDR, and its development has no direct link with the work on kernel that rajko, cass, gauner1986 or LeTama are doing.

-In order to flash MAGLDR within the phone, you will need a flashing utility working under Windows environment. And AFAIK by reading the logs, it is not ready yet. So even before any MAGLDR 10+ release (which legens says will be the good one), common mortals will still need this utility be ready to process to the flash.

To conclude, even if information about kernel improvement is important, let's not just jump to any conclusion such as "release should be soon" or so, just by reading good news from the logs.

Even if it is not update, I thought this had to be explained here. I hope OP will not mind.

For the guy below : public is no dev. Do you have MAGLDR? Do you have flashing utility? I don't, but I have Rajko's NAND build and can't run it. That's what I'm talking about.

luckych · Nov 1, 2010

Latest update on Data problem:

Cass Letama - adding sleep(10); before the first send_modem command fixes the cell switching issue i mentioned before.. cant make it happen no more
01:38 => RaiderX joins
01:38 Cass need to figure why data does not work on boot now unless switch data off/on

Luckych

shahidpak · Nov 5, 2010

Rajko well your awful week starts now for me
13:05 => zoraz joins
13:06 => L_miller joins
13:09 <= wootty left/quit
13:11 Markinus Rajko: Hi, on my mobiles it's running still great. Booted ca 30 times and installed a lot of software. sync with pc. Till now no problems. .
13:11 Rajko well your phone's busted
13:11 Rajko since its supposed to **** up
13:12 gauner1986

13:13 gauner1986 the day will come

I think NAND corruption is over ,,,,...

Samsunguy · Nov 6, 2010

MAGLDR is a bootloader for HTC Leo / HD2 that allows direct OS booting from NAND, developed by DFT team members. Currently it is still a work in progress and not available to the public.

MAGLDR is closed source. It will eventually be able to do more than just boot a Linux kernel from NAND. No release date is known as of yet. NAND support in the Linux kernel is done, DFT team (Cotulla et al.) is now working on bootloader/flasher.

* Choose between WM 6.5 / Linux-based OS
* WM 6.5 can be installed on NAND while Android could be launched from SD, or two Linux-based OS's can be launched; one from NAND and one from SD.
* Select different kernels (you can load kernel from list)
* TTY mode for developers. (ability to dump dmesg after crash and etc)
* Built-in Tetris
* Support YAFFS2 partitions flash via USB / from SD. Up to 8 YAFFS2 partitions.

* Bad Audio Quality Fixed So Far
* Battery Driver Fully Implemented
* Nand Corruption Fixed

What to do:

* Fixing Some Minor Bug's
* Fixing RMNET PPP Data Issues

HD2 Android NAND Updates ONLY [No Discussion]

Retired Senior Moderator

Senior Member

Senior Member

Retired Forum Moderator

Senior Member

Senior Member

Senior Member

Senior Member

Senior Member

Member

Senior Member

Member

Senior Member

Senior Member

Senior Member

Senior Member

Senior Member

Senior Member

Senior Member

Senior Member

Similar threads

Top Liked Posts