Re: [PR] add a reference count to the TCB to prevent it from being deleted. [nuttx]

Fri, 12 Dec 2025 03:29:00 -0800 


acassis commented on PR #17468:
URL: https://github.com/apache/nuttx/pull/17468#issuecomment-3646100601

   > > > > > I really fail to see the rationale behind trying to turn NuttX 
into Linux. The same issue cropped up when you modified spin_lock before—if we 
are going to align everything with Linux, then why not just use a tailored 
Linux distribution instead?
   > > > > 
   > > > > 
   > > > > Good point! I think the NuttX focus should be MCUs with realtime 
requirements (something with Linux cannot offer). In the other hand many 
companies want NuttX to be a replacement to Linux itself. I think both goals 
don't need to be conflicting. But we cannot lose the main focus just to be more 
Linux-like.
   > > > > @hujun260 @anchao @xiaoxiang781216 how to fix this conflict in a way 
both goals could be meet?
   > > > 
   > > > 
   > > > Nobody want to change NuttX to Linux, but the big critical section 
need be addressed since:
   > > > ```
   > > > 1. More and more SoC equip with multple core, but SMP performance is 
very poor with one big critical section
   > > > 
   > > > 2. One big critical section also make the interrrupt latency larger 
than expect
   > > > ```
   > > > 
   > > > 
   > > >     
   > > >       
   > > >     
   > > > 
   > > >       
   > > >     
   > > > 
   > > >     
   > > >   
   > > > So, we need make a balance here.
   > > 
   > > 
   > > Ok, I think we need to understand the impact it will cause in non-SMP 
MCUs. Could you please provide some testing measurements before and after this 
modification?
   > 
   > I ran osperf on a single-core ESP32 for performance evaluation, and the 
key metrics showed little change after applying the patch. <img alt="image" 
width="563" height="520" 
src="https://private-user-images.githubusercontent.com/128452594/525789383-3701714b-b432-43b6-9022-1c35fa0ec80a.png?jwt=eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9.eyJpc3MiOiJnaXRodWIuY29tIiwiYXVkIjoicmF3LmdpdGh1YnVzZXJjb250ZW50LmNvbSIsImtleSI6ImtleTUiLCJleHAiOjE3NjU1MzkwMDIsIm5iZiI6MTc2NTUzODcwMiwicGF0aCI6Ii8xMjg0NTI1OTQvNTI1Nzg5MzgzLTM3MDE3MTRiLWI0MzItNDNiNi05MDIyLTFjMzVmYTBlYzgwYS5wbmc_WC1BbXotQWxnb3JpdGhtPUFXUzQtSE1BQy1TSEEyNTYmWC1BbXotQ3JlZGVudGlhbD1BS0lBVkNPRFlMU0E1M1BRSzRaQSUyRjIwMjUxMjEyJTJGdXMtZWFzdC0xJTJGczMlMkZhd3M0X3JlcXVlc3QmWC1BbXotRGF0ZT0yMDI1MTIxMlQxMTI1MDJaJlgtQW16LUV4cGlyZXM9MzAwJlgtQW16LVNpZ25hdHVyZT04MmFmMTdhYTFjNGI2NTVhYTBjY2EzZTJiZWVhMzUzMzc1ZjYxMDJhODIwNTc4OTlmNWZiMmRjYjFlODM5Y2FlJlgtQW16LVNpZ25lZEhlYWRlcnM9aG9zdCJ9.DdD5q2Vm495y_jwqA40b3cgkczxmFarDGi55TWzWNzU";>
   
   @anchao seems like the impact is not so negative as you thought. I think it 
is fine! What you think?


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to