Validity of Ticket #1247 RTEMS does not implement locks needed by multithreaded newlib

[Ralf Corsepius]

On 02/25/2015 09:38 AM, Sebastian Huber wrote:
> On 24/02/15 22:48, Gedare Bloom wrote:
>> Sebastian,
>>
>> Do you know if #1247 is still valid for current RTEMS and newlib?
>>
>> Gedare
>
> This is still valid.

Can you prove this claim?

The problem with this bug is that nobody ever has come up with a proof, 
which rendered it a "wild unproven claim", though it's not unlikely this 
problem exists.

> I think one of the reasons nobody touched this so
> far is that the usage of locks will likely break all application
> configurations.
No. The issue behind this is nobody having a proof and therefore nobody 
having any idea about what to do.

- Nothing ("Cygwin doesn't seem to have any problems").
- Changing newlib?
- Inserting locks into RTEMS?



Ralf