On 01 Mar 2014 10:04:23 +0000, "Summi" said: > a) inline functions in .c file --> Will it be correct to have inline function > declared in .c file? I am speaking with respect to device driver meant for the > linux kernel. I have a function, which is quite short in terms of function body > and it is exported from one module to another. If a function is truly inlined, what address is used to be exported? (Though gcc stops you from that idiocy - if you take the address of the function, gcc will emit a non-inlined version whether you like it or not. Of course, external references to that function will involve a function call and not be inlined.) Most of the kernel sticks a 'static inline' in a .h file that's included in every .c that needs it, and doesn't bother with the EXPORT_SYMBOL.
Attachment:
pgp_X7PZiuUAq.pgp
Description: PGP signature
_______________________________________________ Kernelnewbies mailing list Kernelnewbies@xxxxxxxxxxxxxxxxx http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
