PREEMPT_RT_FULL on x86-32 machine

PREEMPT_RT_FULL on x86-32 machine

[Pavel Machek]

[-- Attachment #1: Type: text/plain, Size: 359 bytes --]

Hi!

Is full preemption supposed to work on x86-32 machines?

Because it does not work for me. It crashes early in boot, no messages
make it to console. Similar configuration for x86-64 boots ok.

									Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 181 bytes --]

Re: PREEMPT_RT_FULL on x86-32 machine

[Jan Kiszka]

On 22.06.19 10:19, Pavel Machek wrote:
> Hi!
> 
> Is full preemption supposed to work on x86-32 machines?
> 
> Because it does not work for me. It crashes early in boot, no messages
> make it to console. Similar configuration for x86-64 boots ok.
> 

Maybe you can also tell which version(s) you tried, and in which 
configuration(s), and how the crash looked like.

Jan

-- 
Siemens AG, Corporate Technology, CT RDA IOT SES-DE
Corporate Competence Center Embedded Linux

Re: PREEMPT_RT_FULL on x86-32 machine

[Pavel Machek]

[-- Attachment #1: Type: text/plain, Size: 764 bytes --]

Hi!

> >Is full preemption supposed to work on x86-32 machines?
> >
> >Because it does not work for me. It crashes early in boot, no messages
> >make it to console. Similar configuration for x86-64 boots ok.
> >
> 
> Maybe you can also tell which version(s) you tried, and in which
> configuration(s), and how the crash looked like.

I wanted to know if the configuration is supposed to work at all
before starting heavy debugging. From your reply I assume that it
should work.

I tried 4.19.13-rt1-cip1 among others. Crash is early in boot, I can
try some early printing...

Best regards,
									Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 181 bytes --]

Re: PREEMPT_RT_FULL on x86-32 machine

[Jan Kiszka]

On 24.06.19 12:53, Pavel Machek wrote:
> Hi!
> 
>>> Is full preemption supposed to work on x86-32 machines?
>>>
>>> Because it does not work for me. It crashes early in boot, no messages
>>> make it to console. Similar configuration for x86-64 boots ok.
>>>
>>
>> Maybe you can also tell which version(s) you tried, and in which
>> configuration(s), and how the crash looked like.
> 
> I wanted to know if the configuration is supposed to work at all
> before starting heavy debugging. From your reply I assume that it
> should work.

We still have at least one 4.4-based 32-bit-only target on RT. But there is 
likely more, even though I'm promoting to use at least a 64-bit kernel on 64-bit 
capable hw for many years, irrespective of RT.

Jan

> 
> I tried 4.19.13-rt1-cip1 among others. Crash is early in boot, I can
> try some early printing...
> 
> Best regards,
> 									Pavel
> 

-- 
Siemens AG, Corporate Technology, CT RDA IOT SES-DE
Corporate Competence Center Embedded Linux

Re: PREEMPT_RT_FULL on x86-32 machine

[Sebastian Andrzej Siewior]

On 2019-06-24 12:53:40 [+0200], Pavel Machek wrote:
> Hi!
Hi,

> > >Is full preemption supposed to work on x86-32 machines?
> > >
> > >Because it does not work for me. It crashes early in boot, no messages
> > >make it to console. Similar configuration for x86-64 boots ok.
> > >
> > 
> > Maybe you can also tell which version(s) you tried, and in which
> > configuration(s), and how the crash looked like.
> 
> I wanted to know if the configuration is supposed to work at all
> before starting heavy debugging. From your reply I assume that it
> should work.
> 
> I tried 4.19.13-rt1-cip1 among others. Crash is early in boot, I can
> try some early printing...

The latest is v4.19.50-rt22 I would suggest that one. There was a bug
which was fixed either in 4.19.23-rt14 or in the following version.

If the latest one does not work please let me know and I will have a
look.

> Best regards,
> 									Pavel

Sebastian

Re: PREEMPT_RT_FULL on x86-32 machine

[Pavel Machek]

[-- Attachment #1: Type: text/plain, Size: 1158 bytes --]

Hi!

> > > >Is full preemption supposed to work on x86-32 machines?
> > > >
> > > >Because it does not work for me. It crashes early in boot, no messages
> > > >make it to console. Similar configuration for x86-64 boots ok.
> > > >
> > > 
> > > Maybe you can also tell which version(s) you tried, and in which
> > > configuration(s), and how the crash looked like.
> > 
> > I wanted to know if the configuration is supposed to work at all
> > before starting heavy debugging. From your reply I assume that it
> > should work.
> > 
> > I tried 4.19.13-rt1-cip1 among others. Crash is early in boot, I can
> > try some early printing...
> 
> The latest is v4.19.50-rt22 I would suggest that one. There was a bug
> which was fixed either in 4.19.23-rt14 or in the following version.
> 
> If the latest one does not work please let me know and I will have a
> look.

Ok, let me try again. It seems the kernel based on the latest one
works ok. Even suspend/resume works.

Thanks,
									Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 181 bytes --]