Any thread issues with va_list etc.

Chris Thomas

In my app I have verbose logging going out over the network and more concise logging to SRAM to provide post-mortem debugging.

Both of these go thru an implementation of vsnprintf, which in turn uses va_start and va_end.

I have been getting instances of verbose log messages going into the other log.

Before I wade in and extend my library modules with portable locking mechanisms to ensure use of va_start/va_end is serialized, is this actually a problem in ti compilers (7.3.14 on 6412)?

over 9 years ago

0 George Mock over 9 years ago

TI__Guru**** 238965 points

I have only seen va_arg and friends use a local variable for the argument pointer. In that case, there can be no thread issues.

The function vsnprintf and friends does use some global state, so you can see thread issues there. To handle that please read the section titled Handling Reentrancy in the C6000 compiler manual.

Thanks and regards,

-George

0 Chris Thomas over 9 years ago in reply to George Mock

Expert 1310 points

I had already sidestepped the vsnprintf issues with a PD version by Patrick Powell.

But I have found the issue, I forgot to include the header and the compiler made something up, it is a bit of a shame as I am running with warnings as errors and with a good deal of MISRA enabled.

What is the ti way of saying "-Werror=implicit-function-declaration" ?

0 Archaeologist over 9 years ago in reply to Chris Thomas

TI__Guru* 84225 points

-pdsw225

Code Composer Studio™︎

Code Composer Studio forum

Any thread issues with va_list etc.