webOS Nation Forums
> Homebrew
> webOS patches
>
How To: Install Mobile Hot Spot on Sprint PRE
First
...
31 36 37 38 39 40 41
42 43 44 45 46 51
...
Last
Member:
DrewPre
at: 05:19 PM 03/08/2010
for the record .... using the minidhcp from 1.3.5.x along with all the other files resolves the need for the static dns addy problem I was encountering.
I didn't run into that issue at all.. I'm using the minidhcpd from 1.4 (if it's even any different than 1.3.5?) and DNS resolution seems to be working fine. I guess YMMV.
Yes, it works, but I'm also having the hang on reboot issue. Any ideas?
Member:
mpagano
at: 05:57 PM 03/08/2010
Worked first time. Awsome. Thank you. Backing to running Mobile Hot Spot on 1.4. I missed it. Thanks again.
Member:
Centrico
at: 09:34 PM 03/08/2010
Couldn't handle the freezing during restart so I rolled back to 1.35 until the issue is resolved. (I disabled the auto updates) during my scheduled daily restart my phone would just go offline which meant no alarm to wake me for work in the morning....
ooops
Member:
Spader
at: 11:25 PM 03/08/2010
So which part of this is causing the freezing? That's got to be the last hurdle to making this work (fairly) reliably again.
My money is on the 1.3.5.1 PmNetConfigmanager file, or the modifications necessary to make this work. I am at a loss as to what else we could try. ha
Member:
pks627
at: 12:37 AM 03/09/2010
I too have MHS working, but am experiencing the hang on restart. Not a good combination with restart cron enabled. I will be checking /var/logs in the morning. Has anyone else checked these logs?
sweet! progress is being made! keep it up guys!
I'm still sticking with 1.3.5.1 until it gets simple enough for me to take it on....gettin close!
Member:
DrewPre
at: 05:22 AM 03/09/2010
I would bet that the hang is PmNetConfigManaer as well.
Question: In linux is there such a thing as a shutdown script? that could be run when the system is sent the restart command...So maybe the killall command can be used to kill PmNetConfigManager prior to restart command being executed? In fact you could/should kill PmNetConfigManager & Freetetherd, just in case. I seriously doubt it's MobileHotspotd or Minidhcp since they're only running when the app is open.
Member:
bspesq
at: 09:51 AM 03/09/2010
I'm of the opinion that it is something in the shutdown scripts. These are located in /etc/rc6.d. Just running 'reboot' from the commandline (or phone menus) will run all the scripts located in this directory. I have not isolated what in these scripts is hanging up, but executing the last script in that directory (/etc/rc6d/S90reboot, which will be the last item run because of the file naming) will do a power cycle of the phone (without hanging.)
The script that kills PmNetConfigManager does NOT seem to hang the phone if run by itself, but it's possible it's causing the hangup because of its interaction with the other shutdown scripts.
Member:
Centrico
at: 10:38 AM 03/09/2010
Originally Posted by DrewPre:
I would bet that the hang is PmNetConfigManaer as well.
Question: In linux is there such a thing as a shutdown script? that could be run when the system is sent the restart command...So maybe the killall command can be used to kill PmNetConfigManager prior to restart command being executed? In fact you could/should kill PmNetConfigManager & Freetetherd, just in case. I seriously doubt it's MobileHotspotd or Minidhcp since they're only running when the app is open.
I don't think so, I never installed PmNetConfigManager and still experienced the hanging, but I did add/ remove other files.
good work guys, and good luck
Member:
pks627
at: 03:23 PM 03/09/2010
I added some debugging statements to the scripts in /etc/rc6.d (thanks to bspesq for pointing out the location of the shutdown scripts).
I am fairly sure that the hang is occurring during the execution of S40sendsigs, which basically terminates and kills open processes.
Tried excluding minidhcpd and mobilehotspotd from kill list, but no luck.
The next step after sendsigs is unmount file system. Does anyone know the danger of unmounting file systems before all processes are killed?
Member:
Demlor
at: 03:24 PM 03/09/2010
I just sent mobilehotspotd binary and sent the other file as well to the phone, and tried to reboot and it locked up.
So I'm guessing it has to do with those files interfering with some other shutdown file.
Member:
bspesq
at: 03:56 PM 03/09/2010
Well, pks627, you're definitely on the right track. The killall5 command in those scripts is what's killing the phone. I...tried adding mobilehotspotd to the exclude list, but in doing so (stupid me for not using a real shell) screwed up the file. Could someone let me know what S40sendsigs is supposed to be linked to (something under init.d, I presume)?
Note that we also have us.ryanhope.freetetherd which is another process that's running all the time. It should probably be excluded too, in case it's causing the problem.
I seriously don't recommend it, but it is possible to fix the reboot 'hang' by removing the S40sendsigs kill script. No processes will be killed, so it's entirely possible that filesystem or file corruption will occur when apps aren't ordered to die before the OS is rebooted.
Member:
pks627
at: 04:03 PM 03/09/2010
S40sendsigs -> /init.d/sendsigs
Member:
Spader
at: 04:25 PM 03/09/2010
This is getting more and more complex... and I love it! haha
Member:
pks627
at: 04:29 PM 03/09/2010
Brainstorming session...
So, I think daemons (like freetetherd, minidhcpd, freetetherd, etc) are excluded from the kill list since as soon as you kill them, they are immediately restarted (nature of the beast). So this script uses a command called pidof to return the process ids of daemons to be excluded from the kill list. Problem is pidof freetetherd returns nothing. Don't know why.
So I suspect this script keeps killing freetetherd and it keeps coming back - forever.
Ideas?
Member:
DrewPre
at: 04:36 PM 03/09/2010
Originally Posted by zx3junglist:
I didn't run into that issue at all.. I'm using the minidhcpd from 1.4 (if it's even any different than 1.3.5?) and DNS resolution seems to be working fine. I guess YMMV.
I suspect that your version of MHS is different somehow since you didn't have the palm folder in /usr/lib/ipkg, tho.
Originally Posted by pks627:
Brainstorming session...
So, I think daemons (like freetetherd, minidhcpd, freetetherd, etc) are excluded from the kill list since as soon as you kill them, they are immediately restarted (nature of the beast). So this script uses a command called pidof to return the process ids of daemons to be excluded from the kill list. Problem is pidof freetetherd returns nothing. Don't know why.
So I suspect this script keeps killing freetetherd and it keeps coming back - forever.
Ideas?
Is there a way to edit the script to pause the startup on freetetherd? Like adding a pause 2 seconds command? Only enough time for the kill script to finish.
Member:
DrewPre
at: 04:48 PM 03/09/2010
Originally Posted by Demlor:
I just sent mobilehotspotd binary and sent the other file as well to the phone, and tried to reboot and it locked up.
So I'm guessing it has to do with those files interfering with some other shutdown file.
I don't think this should be ignored without inquiring a bit further.
What was your Pre's state prior to copying over the files? Was this a freshely 1.4 Doctored phone? What was the status of the all the files associated with MHS?? [PmNetConfigManager, MobileHotspotd, & MiniDHCP] Finally, you sent mobilehotspotd and WHAT other file over to your Pre that caused the hang? Do you mean the /etc/event.d/mobilehotspotd upstart script?
First
...
31 36 37 38 39 40 41
42 43 44 45 46 51
...
Last
webOS Nation Forums
> Homebrew
> webOS patches
>
How To: Install Mobile Hot Spot on Sprint PRE