Sleep while holding a mutex

Sleep while holding a mutex

[Russell Johnson]

[-- Attachment #1.1: Type: text/plain, Size: 744 bytes --]

In one of my functions, I lock a mutex, call evl_timedwat_event(), and then
unlock the mutex. Whenever I call the timedwait, I get an error from the EVL
Core saying that there was a sleep while holding a mutex. I was following
the example on [1]. I also noticed under a test at [2] that there is a
comment saying to disable the T_WOLI mode on the thread because there will
be a sleep during a mutex. The test appears to verify the timedwait call. So
do I need to just ignore that error whenever I am using an EVL event? Or do
I have something else going wrong potentially?

 

[1] https://evlproject.org/core/user-api/event/

[2]
https://source.denx.de/Xenomai/xenomai4/libevl/-/blob/next/tests/monitor-eve
nt.c#L40

 

Thanks,

 

Russell

 


[-- Attachment #1.2: Type: text/html, Size: 2687 bytes --]

[-- Attachment #2: smime.p7s --]
[-- Type: application/pkcs7-signature, Size: 6759 bytes --]

Re: Sleep while holding a mutex

[Philippe Gerum]

Russell Johnson <russell.johnson@kratosdefense.com> writes:

> [[S/MIME Signed Part:Undecided]]
> In one of my functions, I lock a mutex, call evl_timedwat_event(), and then unlock the mutex. Whenever I call the timedwait, I get an error from the EVL Core saying that there was a sleep while holding a mutex. I was following the example on [1]. I also noticed under a test at [2] that there is a comment saying to disable the T_WOLI mode on the thread because there will be a sleep during a mutex. The test appears to verify the timedwait
> call. So do I need to just ignore that error whenever I am using an EVL event? Or do I have something else going wrong potentially?
>
>  
>
> [1] https://evlproject.org/core/user-api/event/
>
> [2] https://source.denx.de/Xenomai/xenomai4/libevl/-/blob/next/tests/monitor-event.c#L40
>
>  
>
> Thanks,
>
>  
>
> Russell
>
> [[End of S/MIME Signed Part]]

The documentation is right, but the code is wrong. This patch fixes it:
https://source.denx.de/Xenomai/xenomai4/linux-evl/-/commit/1dc0d91aa3a5c2055fe8e492df0b229546d4f5da

The test suite was updated accordingly to check the right behavior:
https://source.denx.de/Xenomai/xenomai4/libevl/-/commit/1b30808a6b0095736c5e1059377e2cb07225e9f6

-- 
Philippe.

RE: [External] - Re: Sleep while holding a mutex

[Russell Johnson]

[-- Attachment #1: Type: text/plain, Size: 214 bytes --]

Sounds good. I did notice when compiling the updated libevl that there was
an error with not using the return value of the write call in the sigdebug
handler in the updated test. Just a heads up.

Thanks,

Russell

[-- Attachment #2: smime.p7s --]
[-- Type: application/pkcs7-signature, Size: 6759 bytes --]