Playing with the KSM 09 webinterface - Page 2

JFeole · 01-20-2018, 10:13 AM

added more detail on 3rd test output..

==

### executed:echo PRODUCT=$PRODUCT, PRODUCT_ID=$PRODUCT_ID, CPU=$CPU, PLATFORM=$PLATFORM,\WIFI_MODULE=$WIFI_MODULE, WIFI_MODULE_PATH=$WIFI_MODULE_PATH\uname=$(uname -a), FW=$(awk -F',' '{print $3}' /mnt/onboard/.kobo/version)

PRODUCT=dahlia, PRODUCT_ID=0x4213, CPU=mx50, PLATFORM=mx50-ntx,
WIFI_MODULE=dhd, WIFI_MODULE_PATH=/drivers/mx50-ntx/wifi/dhd.ko
uname=Linux (none) 2.6.35.3-850-gbc67621+ #3032 PREEMPT Mon Jan 9 13:37:40 CST 2017 armv7l GNU/Linux, FW=4.7.10413

==

### executed:epdc_test -b
Set the background to 16-bpp
screen_info.xres_virtual = 1440
screen_info.yres_virtual = 2304
screen_info.bits_per_pixel = 16
Mem-Mapping FB0
Set to region update mode
Set waveform modes

==

First press of KSM Message, OK displayed on top left screen.

Then Zoom in/out and then exit KSM Message.

Press again "ksm message". See whether the "ok" is again in the top left corner.

(No, not in Top left corner, it came back on the bottom right corner, upside down)
==

tshering · 01-20-2018, 10:28 AM

@nook_nerd
Thank you for the tests results. Could you try something in addition?
After you close/exit the first time the ksm message,

press "read rotate-value" and make a note of the result
press "set read value again"
press "ksm message"

If this time the message window is in the correct orientation, try

repeating several times "close/exit" - "set read value again" - "ksm message" (repeating makes only sense when after the first repetition the rotation is still ok.)

JFeole · 01-20-2018, 12:20 PM

Tshering,

Ok, did as instructed.

Read Rotate Value was:
-
### executed:cat /sys/class/graphics/fb0/rotate
0
-

Tried 4 or 5 times, and came back every time in correct orientation, OK was in upper left corner each time.

Hope it is useful..

Thanks,
John Feole

tshering · 01-20-2018, 01:00 PM

Quote:

Originally Posted by nook_nerd

Tshering,

Ok, did as instructed.

Read Rotate Value was:
-
### executed:cat /sys/class/graphics/fb0/rotate
0
-

Tried 4 or 5 times, and came back every time in correct orientation, OK was in upper left corner each time.

Hope it is useful..

Thanks,
John Feole

Thank you. That looks promising! Doesn't it?
After some seconds I realised that I was over enthusiastic. It seems we can prevent the rotation, when we know that the rotation has been destabilised. However, when we take the same measures without the rotation already being destabilised, we will produce the destabilisation ourselves. That means before each call of a KSM application we would have to know whether the rotation is stable or unstable. The only solution that comes to my mind now is to record whether nickel has been run in this session, and if yes presume that the rotation is unstable.
Any ideas?

boriar · 01-20-2018, 08:02 PM

Quote:

Originally Posted by tshering

Thank you. That looks promising! Doesn't it?
After some seconds I realised that I was over enthusiastic. It seems we can prevent the rotation, when we know that the rotation has been destabilised. However, when we take the same measures without the rotation already being destabilised, we will produce the destabilisation ourselves. That means before each call of a KSM application we would have to know whether the rotation is stable or unstable. The only solution that comes to my mind now is to record whether nickel has been run in this session, and if yes presume that the rotation is unstable.
Any ideas?

@tshering

Finally update FW from a factory reset and installed KSM09, lastest KOReader, CoolReader and Plato.
Make all test to proposed and obtained the same results than nook_nerd.

About the rotation issue:
May be write a token (inside a file or an empty file) and toggle/delete whenever KSM is run (or only the first run), what do you think?

Edit: I mean write a file token when KSM run the first time. The check if that file exist when KSM need to change rotation and if the file don't exist don't do anything but if it exist change the rotation value (and delete the token).
Also note that rotation occurs whenever an option is pressed, alternately upside and downside, since nickel (and only with nickel) is used once.

tshering · 01-21-2018, 07:42 AM

Quote:

Originally Posted by boriar

@tshering

Finally update FW from a factory reset and installed KSM09, lastest KOReader, CoolReader and Plato.
Make all test to proposed and obtained the same results than nook_nerd.

About the rotation issue:
May be write a token (inside a file or an empty file) and toggle/delete whenever KSM is run (or only the first run), what do you think?

Edit: I mean write a file token when KSM run the first time. The check if that file exist when KSM need to change rotation and if the file don't exist don't do anything but if it exist change the rotation value (and delete the token).
Also note that rotation occurs whenever an option is pressed, alternately upside and downside, since nickel (and only with nickel) is used once.

Thank you for the information!
I thought the rotation occurs only after nickle has been run. Therefore, I was thinking of creating a file to indicate that nickel has already been running in the current session. In addition KSM would have to check which FW is installed. For up to a certain point, the rotation problem was not there. I guess it reappeared with the change from 16 to 32 bit colour depth. But I am not certain about it, since reports came only some time later.

Did you already try to select "start plato"? If it does not work and does also not give control back to KSM, you can exit it by selecting "exit plato" in the webinterface. Or if nothing works you can "reboot" or "power off" by the webinterface.
If plato does not work, and you do not see anything identifiable on the screen, poke a little around. There are chances the screen will update in some parts, so that you can see in which direction the screen is rotated.

boriar · 01-21-2018, 08:05 AM

Quote:

Originally Posted by tshering

Thank you for the information!
I thought the rotation occurs only after nickle has been run. Therefore, I was thinking of creating a file to indicate that nickel has already been running in the current session. In addition KSM would have to check which FW is installed. For up to a certain point, the rotation problem was not there. I guess it reappeared with the change from 16 to 32 bit colour depth. But I am not certain about it, since reports came only some time later.

I agree with you about 32 bit is the culprit. I don't notice before because with your KSM-koreader-coolreader combination there's some long time ago I don't use nickel (even with amazing GeoffR patches). Do you want I check last FW with 16 bit and first with 32 bit?

Quote:

Originally Posted by tshering

Did you already try to select "start plato"? If it does not work and does also not give control back to KSM, you can exit it by selecting "exit plato" in the webinterface. Or if nothing works you can "reboot" or "power off" by the webinterface.
If plato does not work, and you do not see anything identifiable on the screen, poke a little around. There are chances the screen will update in some parts, so that you can see in which direction the screen is rotated.

I'm very interesting in that new ereader but the first iterations were too time consuming to test

. At the moment I could not start with KSM (by now I don't want to try with fmon). I will to experiment with the scrip

tshering · 01-21-2018, 08:20 AM

Quote:

Originally Posted by boriar

I agree with you about 32 bit is the culprit. I don't notice before because with your KSM-koreader-coolreader combination there's some long time ago I don't use nickel (even with amazing GeoffR patches). Do you want I check last FW with 16 bit and first with 32 bit?

Do not check. That would be too time consuming. I just assume the culprit is the 16/32 change. If it is something different, we will learn later.

Quote:

Originally Posted by boriar

At the moment I could not start with KSM (by now I don't want to try with fmon). I will to experiment with the scrip

You mean you tried to start plato from KSM (on the device), and it did not work? If you like, try the following from the webinterface:

"Kill KSM"
"set 0" (this is in the field that starts with the yellow token "rotate")
"set read value again"
"start plato as is"

boriar · 01-21-2018, 10:12 AM

Quote:

Originally Posted by tshering

You mean you tried to start plato from KSM (on the device), and it did not work? If you like, try the following from the webinterface:

"Kill KSM"
"set 0" (this is in the field that starts with the yellow token "rotate")
"set read value again"
"start plato as is"

Nothing happend!
When I try to start plato from KSM the screen flash for a second but with the webinterface and these steps... absolute nothing.
Also I have noted that trying to start plato from ksm triggered the rotation problem also.

tshering · 01-21-2018, 10:51 AM

Quote:

Originally Posted by boriar

Nothing happend!
When I try to start plato from KSM the screen flash for a second but with the webinterface and these steps... absolute nothing.
Also I have noted that trying to start plato from ksm triggered the rotation problem also.

Did you look into the error log of plato?

boriar · 01-21-2018, 01:50 PM

Quote:

Originally Posted by tshering

Did you look into the error log of plato?

Sorry for delay.
That's the only thing in crash.log:

Code:

Can't load metadata: Can't open file.
kbmenupngs/exit_nickel.png
kbmenupngs/poweroff.png
kbmenupngs/switchtokoreader.png
kbmenupngs/toggle_nightmode.png
kbmenupngs/koreader.png
kbmenupngs/launcher.png
kbmenupngs/plato.png
thread 'main' panicked at 'index 8 out of range for slice of length 7', /checkout/src/libcore/slice/mod.rs:745:4
note: Run with `RUST_BACKTRACE=1` for a backtrace.
Aborted

tshering · 01-21-2018, 01:54 PM

Maybe we should abstain from changing back to 16 bit color depth, once nickel has changed the value to 32 bit. Originally, I decided to change back to 16 bit, since at this point this allowed to easily run Koreader. More recent versions of Koreader however can run with 32 bit. And if not changing the color depth helps prevent the rotation problem, I do not care if I break compatibility with older Koreader versions.

The current experiment has the purpose to see whether staying at 32 bit depth prevents the rotation problem.

I assume that you use usbnet. Doing it with wifi instead, would need some additional steps. If you cannot use usbnet and want to participate, tell me.

Read the first three steps in advance, because step 3 and 4 must be executed quickly. From then onwards we can proceed as slow as we like.

Write into the navigation bar of your browser http://192.168.2.101:8087/rmc_test.html, but do not yet press enter.
Power on your device.
When you see the KSM menu, press return (in the browser). The browser should now show a page with some buttons. (If nothing happens, press enter again. Maybe the first time was to early.)
The first button is "kill KSM", press it. The screen should now be completely unresponsive.
You can now zoom in or out [Ctrl-+, Ctrl--], so that you can see the whole page, including the output field at the bottom.

Press "basic info". Copy the text from the output field to the text file.

Code:

### executed:echo $PRODUCT, $PRODUCT_ID, $CPU, $PLATFORM, $(uname -a), FW $(awk -F',' '{print $3}' /mnt/onboard/.kobo/version)
kraken, 0x4173, mx50, mx50-ntx, Linux (none) 2.6.35.3-850-gbc67621+ #619 PREEMPT Thu Dec 22 15:29:00 CST 2016 armv7l GNU/Linux, FW 4.7.10413

Press "read rotate value"
Copy the text from the output field to the text file.
Code:
```
### executed:cat /sys/class/graphics/fb0/rotate
0
```
And remember the value, since we will need it later.
Press "start nickel, basic". Nickel will start. Give it some time to settle itself. Do not open any book, stay a the home screen. Anyway, you must return to the home screen before you do the next step.
Press "exit nickel". You will not see any change on the screen. The only difference is that the screen is again unresponsive.
Press "set 0", "set 1", "set 2", "set 3", depending on the value that was returned by "read rotate-value" (if the value was 0, do "set 0"if 1, do "set 1", and so on)
Press "set read value again"
Press "ksm message". There should appear the KSM message application, showing "ok" in the top left corner. [on the Kobo] Press the middle of the screen to open the bottom menu bar. Zoom in and out.
[on the Kobo] Press "close", or [on the PC] press "exit ksm message". The screen will now again be unresponsive.
Press again "ksm message". See whether the "ok" is again in the top left corner.
If everything worked, repeat the sequence "ksm message" - "exit ksm message" several times to see, whether the screen orientation stays the same.
Press "reboot" or (but only when things went well) press "ksm power off"

Please report what happened and post the results that you copied to the text file.

tshering · 01-21-2018, 02:00 PM

Quote:

Originally Posted by boriar

Sorry for delay.
That's the only thing in crash.log:

Code:

Can't load metadata: Can't open file.
kbmenupngs/exit_nickel.png
kbmenupngs/poweroff.png
kbmenupngs/switchtokoreader.png
kbmenupngs/toggle_nightmode.png
kbmenupngs/koreader.png
kbmenupngs/launcher.png
kbmenupngs/plato.png
thread 'main' panicked at 'index 8 out of range for slice of length 7', /checkout/src/libcore/slice/mod.rs:745:4
note: Run with `RUST_BACKTRACE=1` for a backtrace.
Aborted

There is the index error, but of course, I do not know the details. Do you have the newest version of plato?

JFeole · 01-21-2018, 02:04 PM

For this new experiment, can you clarify this step?
-
Press "read rotate value"
Copy the text from the output field to the text file.
Code:
### executed:cat /sys/class/graphics/fb0/rotate
0

Press "set 0", "set 1", "set 2", "set 3", depending on the value that was returned by "read rotate-value"
-
Do you mean, if it comes back with 0, do set 0?? if 1, do set 1??

tshering · 01-21-2018, 02:08 PM

Quote:

Originally Posted by nook_nerd

For this new experiment, can you clarify this step?
-
Press "read rotate value"
Copy the text from the output field to the text file.
Code:
### executed:cat /sys/class/graphics/fb0/rotate
0

Press "set 0", "set 1", "set 2", "set 3", depending on the value that was returned by "read rotate-value"
-
Do you mean, if it comes back with 0, do set 0?? if 1, do set 1??

If it comes back with 0, do set 0! Thank you, I will add this to the description.

01-20-2018, 10:28 AM	#17
tshering Wizard Posts: 3,489 Karma: 2914715 Join Date: Jun 2012 Device: kobo touch	@nook_nerd Thank you for the tests results. Could you try something in addition? After you close/exit the first time the ksm message, press "read rotate-value" and make a note of the result press "set read value again" press "ksm message" If this time the message window is in the correct orientation, try repeating several times "close/exit" - "set read value again" - "ksm message" (repeating makes only sense when after the first repetition the rotation is still ok.) Last edited by tshering; 01-20-2018 at 10:30 AM.

01-21-2018, 01:54 PM	#27
tshering Wizard Posts: 3,489 Karma: 2914715 Join Date: Jun 2012 Device: kobo touch	rotation problem, experiment 2 Maybe we should abstain from changing back to 16 bit color depth, once nickel has changed the value to 32 bit. Originally, I decided to change back to 16 bit, since at this point this allowed to easily run Koreader. More recent versions of Koreader however can run with 32 bit. And if not changing the color depth helps prevent the rotation problem, I do not care if I break compatibility with older Koreader versions. The current experiment has the purpose to see whether staying at 32 bit depth prevents the rotation problem. I assume that you use usbnet. Doing it with wifi instead, would need some additional steps. If you cannot use usbnet and want to participate, tell me. Read the first three steps in advance, because step 3 and 4 must be executed quickly. From then onwards we can proceed as slow as we like. Write into the navigation bar of your browser http://192.168.2.101:8087/rmc_test.html, but do not yet press enter. Power on your device. When you see the KSM menu, press return (in the browser). The browser should now show a page with some buttons. (If nothing happens, press enter again. Maybe the first time was to early.) The first button is "kill KSM", press it. The screen should now be completely unresponsive. You can now zoom in or out [Ctrl-+, Ctrl--], so that you can see the whole page, including the output field at the bottom. Press "basic info". Copy the text from the output field to the text file. Code: ### executed:echo $PRODUCT, $PRODUCT_ID, $CPU, $PLATFORM, $(uname -a), FW $(awk -F',' '{print $3}' /mnt/onboard/.kobo/version) kraken, 0x4173, mx50, mx50-ntx, Linux (none) 2.6.35.3-850-gbc67621+ #619 PREEMPT Thu Dec 22 15:29:00 CST 2016 armv7l GNU/Linux, FW 4.7.10413 Press "read rotate value" Copy the text from the output field to the text file. Code: ### executed:cat /sys/class/graphics/fb0/rotate 0 And remember the value, since we will need it later. Press "start nickel, basic". Nickel will start. Give it some time to settle itself. Do not open any book, stay a the home screen. Anyway, you must return to the home screen before you do the next step. Press "exit nickel". You will not see any change on the screen. The only difference is that the screen is again unresponsive. Press "set 0", "set 1", "set 2", "set 3", depending on the value that was returned by "read rotate-value" (if the value was 0, do "set 0"if 1, do "set 1", and so on) Press "set read value again" Press "ksm message". There should appear the KSM message application, showing "ok" in the top left corner. [on the Kobo] Press the middle of the screen to open the bottom menu bar. Zoom in and out. [on the Kobo] Press "close", or [on the PC] press "exit ksm message". The screen will now again be unresponsive. Press again "ksm message". See whether the "ok" is again in the top left corner. If everything worked, repeat the sequence "ksm message" - "exit ksm message" several times to see, whether the screen orientation stays the same. Press "reboot" or (but only when things went well) press "ksm power off" Please report what happened and post the results that you copied to the text file. Last edited by tshering; 01-21-2018 at 02:09 PM.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Aura H2O If I use KSM can I switch between pocket etc?	masterz87	Kobo Reader	1	09-28-2017 05:15 PM
Mesure battery duration for KSM	embryo	Kobo Developer's Corner	11	09-21-2017 04:00 PM
KSM 08, kobo database, and Calibre	drawson1	Kobo Reader	4	01-30-2017 04:06 AM
USB identification with KSM	Jellby	Kobo Developer's Corner	4	01-09-2015 07:30 AM
Greetings from KSM	ksmurali	Introduce Yourself	4	05-25-2011 04:35 AM

01-20-2018, 10:13 AM	#16
JFeole Connoisseur Posts: 93 Karma: 28862 Join Date: Jul 2010 Device: Kobo_Aura_H2O, Kobo_Aura_One, Kobo Sage	added more detail on 3rd test output.. == ### executed:echo PRODUCT=$PRODUCT, PRODUCT_ID=$PRODUCT_ID, CPU=$CPU, PLATFORM=$PLATFORM,\WIFI_MODULE=$WIFI_MODULE, WIFI_MODULE_PATH=$WIFI_MODULE_PATH\uname=$(uname -a), FW=$(awk -F',' '{print $3}' /mnt/onboard/.kobo/version) PRODUCT=dahlia, PRODUCT_ID=0x4213, CPU=mx50, PLATFORM=mx50-ntx, WIFI_MODULE=dhd, WIFI_MODULE_PATH=/drivers/mx50-ntx/wifi/dhd.ko uname=Linux (none) 2.6.35.3-850-gbc67621+ #3032 PREEMPT Mon Jan 9 13:37:40 CST 2017 armv7l GNU/Linux, FW=4.7.10413 == ### executed:epdc_test -b Set the background to 16-bpp screen_info.xres_virtual = 1440 screen_info.yres_virtual = 2304 screen_info.bits_per_pixel = 16 Mem-Mapping FB0 Set to region update mode Set waveform modes == First press of KSM Message, OK displayed on top left screen. Then Zoom in/out and then exit KSM Message. Press again "ksm message". See whether the "ok" is again in the top left corner. (No, not in Top left corner, it came back on the bottom right corner, upside down) ==

01-20-2018, 12:20 PM	#18
JFeole Connoisseur Posts: 93 Karma: 28862 Join Date: Jul 2010 Device: Kobo_Aura_H2O, Kobo_Aura_One, Kobo Sage	Tshering, Ok, did as instructed. Read Rotate Value was: - ### executed:cat /sys/class/graphics/fb0/rotate 0 - Tried 4 or 5 times, and came back every time in correct orientation, OK was in upper left corner each time. Hope it is useful.. Thanks, John Feole

01-21-2018, 02:04 PM	#29
JFeole Connoisseur Posts: 93 Karma: 28862 Join Date: Jul 2010 Device: Kobo_Aura_H2O, Kobo_Aura_One, Kobo Sage	For this new experiment, can you clarify this step? - Press "read rotate value" Copy the text from the output field to the text file. Code: ### executed:cat /sys/class/graphics/fb0/rotate 0 Press "set 0", "set 1", "set 2", "set 3", depending on the value that was returned by "read rotate-value" - Do you mean, if it comes back with 0, do set 0?? if 1, do set 1??