OS hangs ramdonly

[franciscopaniskaseker]

We are using Cherry Trail (z8300) with AXP288C battery. We can not map any battery even using last patches from kernel next and tw.shuttle.com driver sent to us to try to solve this. We ware using kernel 4.8.0-rc1 and 4.8.0-rc1+ (from next repository tree).

We created a kernel .config with a lot of builtin modules (iio, i2c, axp etc) as recommended by documentation and with your script (specially i2cget commands) we can get battery information.

In our case we have to use i2c-6 (i2c-[46] does not exist).

We are getting a lot of OS hangs because i2cget commands (the other commands can not hang our OS). Some of them get "read failed error" from i2cget commands. We treated all of them to abort that loop after error but the problem seems relationated to use i2c-tools (i2cget in our case).

We tried to debug your script to identify which command (i2cget) hang our system, but after a lot of tests we only saw usually it will hang after charger_status_reg, but can occur between source_status_reg and charger_status_reg. Sometimes OS hang without "read failed error" too.

We tried to not use "-f" parameter, but without that device will appear always busy.

Do you know what can I do to avoid OS hang?

Oh, and we can test any driver or patch that you have for AXP288C.

thank you for all work and time.

[Icenowy]

Will it hang without axpd?
For my friend's Bay Trail tablet, Dell Venue 8 Pro, I used to met problems that the system randomly hangs. (And it does not use AXP288!)
It is solved as the kernel upgrades.
2016年8月10日 05:02于 Francisco Panis Kaseker notifications@github.com写道：We are using Cherry Trail (z8300) with AXP288C battery. We can not map any battery even using last patches from kernel next and tw.shuttle.com driver sent to us to try to solve this. We ware using kernel 4.8.0-rc1 and 4.8.0-rc1+ (from next repository tree).

We created a kernel .config with a lot of builtin modules (iio, i2c, axp etc) as recommended by documentation and with your script (specially i2cget commands) we can get battery information.

In our case we have to use i2c-6 (i2c-[46] does not exist).

We are getting a lot of OS hangs because i2cget commands (the other commands can not hang our OS). Some of them get "read failed error" from i2cget commands. We treated all of them to abort that loop after error but the problem seems relationated to use i2c-tools (i2cget in our case).

We tried to debug your script to identify which command (i2cget) hang our system, but after a lot of tests we only saw usually it will hang after charger_status_reg, but can occur between source_status_reg and charger_status_reg. Sometimes OS hang without "read failed error" too.

We tried to not use "-f" parameter, but without that device will appear always busy.

Do you know what can I do to avoid OS hang?

Oh, and we can test any driver or patch that you have for AXP288C.

thank you for all work and time.

—You are receiving this because you are subscribed to this thread.Reply to this email directly, view it on GitHub, or mute the thread.

[franciscopaniskaseker]

Do not hang without axpd. I observed that i2c designware get timeout at boot and then axp20x too (-110 ETIMED). Do you have any idea why use i2cget can hang the OS?

[Icenowy]

I have no idea.
Is your touchscreen on I2C?
2016年8月10日 08:36于 Francisco Panis Kaseker notifications@github.com写道：Do not hang without axpd. I observed that i2c designware get timeout at boot and then axp20x too (-110 ETIMED). Do you have any idea why use i2cget can hang the OS?

—You are receiving this because you commented.Reply to this email directly, view it on GitHub, or mute the thread.

[franciscopaniskaseker]

Actually this is a notebook :P And not has touch.

It has:

• sdcard+wifi+bt on SDIO
• sound (rt5645) on i2c

[Icenowy]

Ah-oh.
Is sound working properly?
2016年8月10日 08:39于 Francisco Panis Kaseker notifications@github.com写道：Actually this is a notebook :P And not has touch.

It has:

sdcard+wifi+bt on SDIOsound (rt5645) on i2c

—You are receiving this because you commented.Reply to this email directly, view it on GitHub, or mute the thread.

[franciscopaniskaseker]

We had a lot of problems to get sound in this notebook. After get the sound, we can not control with default alsa and we tried a lot of UCMs and we only get problems with them.

We did some stuffs with alsa mix control and stored alsa state sound. To change internal sound to jack and vice-versa is a big problem. We can not get internal mic working and change to hdmi output is terrible too.

This is a summary of our headache.

No support from Intel is terrible :(

Do you think maybe sound using i2c can cause problems?

[Icenowy]

No.
I doubt whether the i2c controller driver is stable.
In Baytrail, there's a long time that the mmc controller is unstable...
2016年8月10日 09:28于 Francisco Panis Kaseker notifications@github.com写道：We had a lot of problems to get sound in this notebook. After get the sound, we can not control with default alsa and we tried a lot of UCMs and we only get problems with them.

We did some stuffs with alsa mix control and stored alsa state sound. To change internal sound to jack and vice-versa is a big problem. We can not get internal mic working and change to hdmi output is terrible too.

This is a summary of our headache.

No support from Intel is terrible :(

Do you think maybe sound using i2c can cause problems?

—You are receiving this because you commented.Reply to this email directly, view it on GitHub, or mute the thread.

https://github.com/burzumishi/linux-baytrail-flexx10

If you go to the closed issues on this project, you'll see how we got audio working, mostly perfectly on the Flexx9/10 tablets, which are very similar hardware. The only thing that doesn't seem to work is automatic muting of the speaker when a headphone is inserted (it must be done manually through Pulse mixer).

And you'll really want to run Pulse, as it abstracts the very complex mixer controls on the machines into something alot more friendly.

[franciscopaniskaseker]

Thank you jingber. We will try.

[nicman23]

wait for the 4.8-rc3 as it has a lot for our devices

[franciscopaniskaseker]

Waiting :) Do you work on all winner or x-power company?

[Icenowy]

Why do you think so?
I'm only a community coder working on Allwinner mainlineing.
2016年8月19日 下午8:37于 Francisco Panis Kaseker notifications@github.com写道：Waiting :) Do you work on all winner or x-power company?

—You are receiving this because you commented.Reply to this email directly, view it on GitHub, or mute the thread.

[franciscopaniskaseker]

Actually I am asking nicman23, but about you I saw some code on kernel commited by you and I thought you work there.

[nicman23]

some code on kernel commited
wat where?

[franciscopaniskaseker]

I tested last 4.8.0 rc (from git) and I continue receiving OS hangs.

I had to disable CONFIG_I2C_DESIGNWARE_BAYTRAIL because with that semaphore I can not get any info from i2c commands (and that semaphore cause some tracebacks errors on cherrytrail).