Merging eg. xvid gives this: QA Notice: the following files contain runtime text relocations Text relocations require a lot of extra work to be preformed by the dynamic linker which will cause serious performance impact on IA-32 and might not function properly on other architectures hppa for example. If you are a programmer please take a closer look at this package and consider writing a patch which addresses this problem. TEXTREL usr/local/lib/libxvidcore.so.4.1 In fact "which will cause serious performance impact on IA-32" is nonsense, or at least misleading, as the opposite is the case. I guess portage is complaining that the lib was compiled non-pic, so loading the library will be bad for performance. But on the other hand, as it is a computational intense lib, you don't want to use pic, as this costs the ebx register on ia32 and this could have a dramatical negative impact on *runtime* performance. So my suggestion is to make the message clear, that the libs will have bad *start-up* performance by compiling non-pic. Firthermore there should be means to disable this check for certain libs where it makes actually sense to compile non-pic, as I stated. (So infact it is also not very smart to compile ffmpeg, xine, mplayer etc with pic enabled, as you trade of faster start-up for runtime performance.) Reproducible: Always Steps to Reproduce: 1. 2. 3.
*** This bug has been marked as a duplicate of 115727 ***
That is not the case at all with ffmpeg. ffmpeg after it was made pic aware was found to run in general faster. Read drepper's dsohowto.pdf for more info on the subject in general. I'll update the msg however to be more clear to the root of the problems.