Dead lock in my code

Hi ,
I am working on a embedded system. the modules work fine in one thread but as I decided to migrate to multithread in esp32 I face deadlock. Any idea or help to make the neckbottles clear and how to debug these problems ?

If you are able to use a debugger to determine where each thread is running when it gets stuck, then that will tell you where the problem is.

Thanks, According ro my debugs I found out that I lock a struct for reading and also writing. maybe splitting the read and write part fix my problem. I checked tokio streamer and found out that read and write parts are different. Do you have any suggestion ?

Trying to lock the same lock twice on a single thread will indeed result in a deadlock. How you can avoid this will depend on what operations you can make on the IO resource. Splitting reader and writer can be one solution.

What are other options ? I am looking for all solutions to choose the best.

Well, if you can read/write from it with &self access, then you don't need a mutex at all.

Another option is to release the write lock before reading, and vice-versa. Or you could use a re-entrant lock.

How does re-enterant lock work ? I did not found any good resources in google.

It can be locked multiple times from a single thread. (But only from one thread at the time.)

this is the diagram of the situation that my system lock.

That is indeed a diagram. Do you have any questions related to it?

Sorry, I forgot to share my question. As I said I am looking for all solutions and also some examples if exist. Here is What I found out and also you said. Any other solution or best practice will be appreciated.

  • put data in buffer and return to open the lock. ( the code get messy, May be a best practice is needed)
  • use channels
  • split read and write parts
  • re-entrant lock (I don't have much information, I would be glad if U share your knowledge)
  • use async libraries to make sure locks wait ( not sure)

This topic was automatically closed 90 days after the last reply. We invite you to open a new topic if you have further questions or comments.