Re: Critical and others problems with hostmode

Re: Critical and others problems with hostmode

Pali Rohár
Karma: 1382
2012-03-29 11:19 UTC
On Thursday 29 March 2012 10:34:10 you wrote:
> On 29.03.2012 09:38, Pali Rohár wrote:
> > On Thursday 29 March 2012 09:31:34 James Brown wrote:
> >> On 29.03.2012 08:18, Pali Rohár wrote:
> >>> On Thursday 29 March 2012 07:17:49 James Brown wrote:
> >>>> Device - N900
> >>>> OS - Maemo 5
> >>>> V. 21.2011.38-1 (CSSU)
> >>>> kernel-power 1:2.6.28-10power50
> >>>> hostmode-gui 0.3
> >>>>
> >>>> With that "hostmode-gui" (h-e-n) I can easily initialize
> >>>> USB-device but it don't want to mount on my FS with that
> >>>> gui.
> >>>> So, I need to mount it (and umount) mannualy.
> >>>> Today I found a critical problem with using h-e-n. I did
> >>>> not
> >>>> use it from the last summer and I don't remember what of
> >>>> version of "kernel-power" I used that time (of course.
> >>>> not
> >>>> v.50 as now :) ) and whether I used CSSU or not but I had
> >>>> no
> >>>> that problem at that time. (But at that time it needs for
> >>>> me
> >>>> to start bme mannualy after closing h-e-n windows becouse
> >>>> it
> >>>> didn't start automatically and at one time I even had
> >>>> problems
> >>>> with (re)booting my device).
> >>>> Now, since today, when I close the "h-e-n" gui, my device
> >>>> suddenly is shutting down and I need to boot it for
> >>>> working
> >>>> with it. How can I resolve the above-mentioned issues?
> >>>
> >>> Hi,
> >>>
> >>> shutdown after closing h-e-n GUI can be caused by starting
> >>> BME
> >>> when bq27x00_battery driver is loaded. BME and
> >>> bq27x00_battery
> >>> cannot be loaded at same time. So make sure that
> >>> bq27x00_battery is blacklisted in /etc/modprobe.d/* and
> >>> also
> >>> is NOT in list /etc/modules
> >>
> >> Thanks. Yes, I have the bq27x00_battery driver is loaded:
> >> lsmod | grep bq27x00_battery
> >> bq27x00_battery 6480 0
> >> power_supply 6916 1 bq27x00_battery
> >>
> >> I enabled it for purposes getting temperature as it
> >> discribed
> >> in the page "http://wiki.maemo.org/Overclocking"
> >> So, if I want use h-e-n I need to disable it?
> >
> > You cannot run BME and bq27x00_battery together. So you can
> > load bq driver when BME is stopped and before starting BME
> > driver must be unloaded.
>
> Thank you for that information. It is very pitty that that
> issue was not described on that wiki-page.
> As I can see after removing the bq27x00_battery module my
> device become again able to reboot (earlier the command
> `reboot` from terminal or using buttong in its menu only shut
> down my device not reboot - and I didn't know why; now it
> seems that it was from loading that module). But if
> bq27x00_battery and BME cannot run together how can my device
> was able to start?

There is problem that BME and bq driver want to talking to bq
chip via i2c bus. For security reason this cannot be allowed, but
it is possible to start BME and then load driver (in this order).
This is bug in kernel and this situation is DANGEROUS!!!

>
> > So normal usage is: stop BME, load bq driver.... then remove
> > bq driver and start BME.
> >
> > h-e-n gui is stopping BME, so when h-e-n is open you can load
> > bq driver, but do not forget to unload it before closing
> > h-e-n.>
> >>> automounting FS in usb hostmode (without pressing mount
> >>> button or mounting via terminal) will work if you update
> >>> kernel-power to v50 and ke-recv to last version. Last
> >>> version of ke-recv will be updated with next CSSU update.
> >>> So
> >>> wait for next CSSU and automouting will work.
> >>
> >> Thanks for your answer.
> >> But I was not strictly correct, I cannot mount it not only
> >> automatically but also using mount-button in h-e-x. I can
> >> mount only via terminal command.
> >
> > If you can mount it via terminal, automounting with new
> > version of ke-recv will work. Wait for next CSSU. If it will
> > not work with next CSSU + kp50 write me then.
>
> Ok, if it will not work I'll write you, of course.

PS: Keep CC in mailinglist too. Informations can be usefull for
other people too.

--
Pali Rohár
pali.rohar@gmail.com
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)

iEYEABECAAYFAk90RTUACgkQi/DJPQPkQ1LaVQCgqzEdSikhfu6sSrUn1XJn9VHx
igAAoJHaeqJy4DNVbL9hc/q/pnbLrn66
=IxkE
-----END PGP SIGNATURE-----

  •  Reply

Re: Critical and others problems with hostmode

James Brown
Karma: 30
2012-03-29 12:28 UTC
On 29.03.2012 11:19, Pali Rohár wrote:
> On Thursday 29 March 2012 10:34:10 you wrote:
>> On 29.03.2012 09:38, Pali Rohár wrote:
>>> On Thursday 29 March 2012 09:31:34 James Brown wrote:
>>>> On 29.03.2012 08:18, Pali Rohár wrote:
>>>>> On Thursday 29 March 2012 07:17:49 James Brown wrote:
>>>>>> Device - N900
>>>>>> OS - Maemo 5
>>>>>> V. 21.2011.38-1 (CSSU)
>>>>>> kernel-power 1:2.6.28-10power50
>>>>>> hostmode-gui 0.3
>>>>>>
>>>>>> With that "hostmode-gui" (h-e-n) I can easily initialize
>>>>>> USB-device but it don't want to mount on my FS with that
>>>>>> gui.
>>>>>> So, I need to mount it (and umount) mannualy.
>>>>>> Today I found a critical problem with using h-e-n. I did
>>>>>> not
>>>>>> use it from the last summer and I don't remember what of
>>>>>> version of "kernel-power" I used that time (of course.
>>>>>> not
>>>>>> v.50 as now :) ) and whether I used CSSU or not but I had
>>>>>> no
>>>>>> that problem at that time. (But at that time it needs for
>>>>>> me
>>>>>> to start bme mannualy after closing h-e-n windows becouse
>>>>>> it
>>>>>> didn't start automatically and at one time I even had
>>>>>> problems
>>>>>> with (re)booting my device).
>>>>>> Now, since today, when I close the "h-e-n" gui, my device
>>>>>> suddenly is shutting down and I need to boot it for
>>>>>> working
>>>>>> with it. How can I resolve the above-mentioned issues?
>>>>>
>>>>> Hi,
>>>>>
>>>>> shutdown after closing h-e-n GUI can be caused by starting
>>>>> BME
>>>>> when bq27x00_battery driver is loaded. BME and
>>>>> bq27x00_battery
>>>>> cannot be loaded at same time. So make sure that
>>>>> bq27x00_battery is blacklisted in /etc/modprobe.d/* and
>>>>> also
>>>>> is NOT in list /etc/modules
>>>>
>>>> Thanks. Yes, I have the bq27x00_battery driver is loaded:
>>>> lsmod | grep bq27x00_battery
>>>> bq27x00_battery 6480 0
>>>> power_supply 6916 1 bq27x00_battery
>>>>
>>>> I enabled it for purposes getting temperature as it
>>>> discribed
>>>> in the page "http://wiki.maemo.org/Overclocking"
>>>> So, if I want use h-e-n I need to disable it?
>>>
>>> You cannot run BME and bq27x00_battery together. So you can
>>> load bq driver when BME is stopped and before starting BME
>>> driver must be unloaded.
>>
>> Thank you for that information. It is very pitty that that
>> issue was not described on that wiki-page.
>> As I can see after removing the bq27x00_battery module my
>> device become again able to reboot (earlier the command
>> `reboot` from terminal or using buttong in its menu only shut
>> down my device not reboot - and I didn't know why; now it
>> seems that it was from loading that module). But if
>> bq27x00_battery and BME cannot run together how can my device
>> was able to start?
>
> There is problem that BME and bq driver want to talking to bq
> chip via i2c bus. For security reason this cannot be allowed, but
> it is possible to start BME and then load driver (in this order).
> This is bug in kernel and this situation is DANGEROUS!!!
>
>>
>>> So normal usage is: stop BME, load bq driver.... then remove
>>> bq driver and start BME.
>>>
>>> h-e-n gui is stopping BME, so when h-e-n is open you can load
>>> bq driver, but do not forget to unload it before closing
>>> h-e-n.>
>>>>> automounting FS in usb hostmode (without pressing mount
>>>>> button or mounting via terminal) will work if you update
>>>>> kernel-power to v50 and ke-recv to last version. Last
>>>>> version of ke-recv will be updated with next CSSU update.
>>>>> So
>>>>> wait for next CSSU and automouting will work.
>>>>
>>>> Thanks for your answer.
>>>> But I was not strictly correct, I cannot mount it not only
>>>> automatically but also using mount-button in h-e-x. I can
>>>> mount only via terminal command.
>>>
>>> If you can mount it via terminal, automounting with new
>>> version of ke-recv will work. Wait for next CSSU. If it will
>>> not work with next CSSU + kp50 write me then.
>>
>> Ok, if it will not work I'll write you, of course.
>
> PS: Keep CC in mailinglist too. Informations can be usefull for
> other people too.
>


Thanks. Sorry, I sent that message you personally not to the list by
mistake.
  •  Reply

Re: Critical and others problems with hostmode

James Brown
Karma: 30
2012-03-29 16:49 UTC
On 29.03.2012 11:19, Pali Rohár wrote:
> On Thursday 29 March 2012 10:34:10 you wrote:
>> On 29.03.2012 09:38, Pali Rohár wrote:
>>> On Thursday 29 March 2012 09:31:34 James Brown wrote:
>>>> On 29.03.2012 08:18, Pali Rohár wrote:
>>>>> On Thursday 29 March 2012 07:17:49 James Brown wrote:
>>>>>> Device - N900
>>>>>> OS - Maemo 5
>>>>>> V. 21.2011.38-1 (CSSU)
>>>>>> kernel-power 1:2.6.28-10power50
>>>>>> hostmode-gui 0.3
>>>>>>
>>>>>> With that "hostmode-gui" (h-e-n) I can easily initialize
>>>>>> USB-device but it don't want to mount on my FS with that
>>>>>> gui.
>>>>>> So, I need to mount it (and umount) mannualy.
>>>>>> Today I found a critical problem with using h-e-n. I did
>>>>>> not
>>>>>> use it from the last summer and I don't remember what of
>>>>>> version of "kernel-power" I used that time (of course.
>>>>>> not
>>>>>> v.50 as now :) ) and whether I used CSSU or not but I had
>>>>>> no
>>>>>> that problem at that time. (But at that time it needs for
>>>>>> me
>>>>>> to start bme mannualy after closing h-e-n windows becouse
>>>>>> it
>>>>>> didn't start automatically and at one time I even had
>>>>>> problems
>>>>>> with (re)booting my device).
>>>>>> Now, since today, when I close the "h-e-n" gui, my device
>>>>>> suddenly is shutting down and I need to boot it for
>>>>>> working
>>>>>> with it. How can I resolve the above-mentioned issues?
>>>>>
>>>>> Hi,
>>>>>
>>>>> shutdown after closing h-e-n GUI can be caused by starting
>>>>> BME
>>>>> when bq27x00_battery driver is loaded. BME and
>>>>> bq27x00_battery
>>>>> cannot be loaded at same time. So make sure that
>>>>> bq27x00_battery is blacklisted in /etc/modprobe.d/* and
>>>>> also
>>>>> is NOT in list /etc/modules
>>>>
>>>> Thanks. Yes, I have the bq27x00_battery driver is loaded:
>>>> lsmod | grep bq27x00_battery
>>>> bq27x00_battery 6480 0
>>>> power_supply 6916 1 bq27x00_battery
>>>>
>>>> I enabled it for purposes getting temperature as it
>>>> discribed
>>>> in the page "http://wiki.maemo.org/Overclocking"
>>>> So, if I want use h-e-n I need to disable it?
>>>
>>> You cannot run BME and bq27x00_battery together. So you can
>>> load bq driver when BME is stopped and before starting BME
>>> driver must be unloaded.
>>
>> Thank you for that information. It is very pitty that that
>> issue was not described on that wiki-page.
>> As I can see after removing the bq27x00_battery module my
>> device become again able to reboot (earlier the command
>> `reboot` from terminal or using buttong in its menu only shut
>> down my device not reboot - and I didn't know why; now it
>> seems that it was from loading that module). But if
>> bq27x00_battery and BME cannot run together how can my device
>> was able to start?
>
> There is problem that BME and bq driver want to talking to bq
> chip via i2c bus. For security reason this cannot be allowed, but
> it is possible to start BME and then load driver (in this order).
> This is bug in kernel and this situation is DANGEROUS!!!
>
>>
>>> So normal usage is: stop BME, load bq driver.... then remove
>>> bq driver and start BME.
>>>
>>> h-e-n gui is stopping BME, so when h-e-n is open you can load
>>> bq driver, but do not forget to unload it before closing
>>> h-e-n.>

Is there any way to get temperature without the bq27x00_battery module
and, so, without stopping BME?

>>>>> automounting FS in usb hostmode (without pressing mount
>>>>> button or mounting via terminal) will work if you update
>>>>> kernel-power to v50 and ke-recv to last version. Last
>>>>> version of ke-recv will be updated with next CSSU update.
>>>>> So
>>>>> wait for next CSSU and automouting will work.
>>>>
>>>> Thanks for your answer.
>>>> But I was not strictly correct, I cannot mount it not only
>>>> automatically but also using mount-button in h-e-x. I can
>>>> mount only via terminal command.
>>>
>>> If you can mount it via terminal, automounting with new
>>> version of ke-recv will work. Wait for next CSSU. If it will
>>> not work with next CSSU + kp50 write me then.
>>
>> Ok, if it will not work I'll write you, of course.
>
> PS: Keep CC in mailinglist too. Informations can be usefull for
> other people too.
>

Strange. As I see, when I use option "reply to list" in my Icedove
3.0.11 it places "pali.rohar@gmail.com" into "to" not
"maemo-users@maemo.org"
But with another lists it works as expected, not so strange. Probably
the settings of the list are something wrong?
  •  Reply

Re: Critical and others problems with hostmode

Pali Rohár
Karma: 1382
2012-03-29 17:00 UTC
On Thursday 29 March 2012 16:49:01 James Brown wrote:
> Is there any way to get temperature without the
> bq27x00_battery module and, so, without stopping BME?
>

Yes, directly via i2c bus from userspace. You need i2cget utility
and run it as root. Look at shadowjk bq27200.sh information
script at http://enivax.net/jk/n900/bq27200.sh

--
Pali Rohár
pali.rohar@gmail.com
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)

iEYEABECAAYFAk90lTgACgkQi/DJPQPkQ1KAjgCfR0bFyqLOlEm7TwEkp+9bitJo
XFwAn32hwYcrClrpBemwIMrRk7nhkdtj
=kTbA
-----END PGP SIGNATURE-----

  •  Reply

Re: Critical and others problems with hostmode

Paul Hartman

2012-03-29 18:47 UTC
On Thu, Mar 29, 2012 at 11:49 AM, James Brown <jbrownfirst@gmail.com> wrote:
>
> Strange. As I see, when I use option "reply to list" in my Icedove
> 3.0.11 it places "pali.rohar@gmail.com" into "to" not
> "maemo-users@maemo.org"
> But with another lists it works as expected, not so strange. Probably
> the settings of the list are something wrong?

Looks like this list does not set a reply-to header
  •  Reply

Re: Critical and others problems with hostmode

James Brown
Karma: 30
2012-03-30 16:37 UTC
On 29.03.2012 17:00, Pali Rohár wrote:
> On Thursday 29 March 2012 16:49:01 James Brown wrote:
>> Is there any way to get temperature without the
>> bq27x00_battery module and, so, without stopping BME?
>>
>
> Yes, directly via i2c bus from userspace. You need i2cget utility
> and run it as root. Look at shadowjk bq27200.sh information
> script at http://enivax.net/jk/n900/bq27200.sh
>
>

Very thanks. That script indicates temperature.
Interesting, what means "Cycle Count since Learning: 29 Total Cycle
Count since last full reset: 134"? There are no HDD in the device :)
  •  Reply

Re: Critical and others problems with hostmode

Jan Knutar
Karma: 281
2012-03-30 20:24 UTC
On Friday 30 March 2012, James Brown wrote:
> On 29.03.2012 17:00, Pali Rohár wrote:
> > On Thursday 29 March 2012 16:49:01 James Brown wrote:
> >> Is there any way to get temperature without the
> >> bq27x00_battery module and, so, without stopping BME?
> >
> > Yes, directly via i2c bus from userspace. You need i2cget utility
> > and run it as root. Look at shadowjk bq27200.sh information
> > script at http://enivax.net/jk/n900/bq27200.sh
>
> Very thanks. That script indicates temperature.
> Interesting, what means "Cycle Count since Learning: 29 Total Cycle
> Count since last full reset: 134"? There are no HDD in the device :)

Battery charge/discharge cycles. The battery fuel gauge chip will count
the capacity from full charge down to empty, and learn the battery's
capacity. Due to the default settings, it takes about 4-5 such learning
cycles before it learns the capacity correctly the first time. It's also
actually a bit rare to have the chip learn capacity normally, because
bme shuts down device before the threshold at which the chip considers
the battery empty.

In any case, 29 means 29 charge/discharge cycles since the last cycle
during which it could learn the battery capacity. 134 total
charge/discharge cycles.

It gets reset if the battery is removed for a longer time.
  •  Reply