That's true. But if it isn't defined until some kernel after 2.4.9 then we can't use it. Anyway there's no harm in taking it out. Philip Pokorny wrote: > I think "unlikely" is a macro or a hint to GCC that the if condition > being tested is unlikely to be true and therefore it should set the > branch prediction bits in the code generated to improve performance for > the more probable case of the expression being false... > > author at mordac.netroedge.com wrote: > >> Update of /home/cvs/i2c/kernel >> In directory mordac.netroedge.com:/tmp/cvs-serv18981 >> >> Modified Files: >> Tag: lk2-4 >> i2c-proc.c Log Message: >> (mds) fix problem with multiple chips from same driver >> found by Khali. This was caused by not mallocing space for >> and copying the ctl_template. >> This also fixes incomplete removal of the sysctl table >> when i2c-proc is rmmodded. >> Also remove unlikely() calls since unlikely() is not available >> in kernel 2.4.9. >> >> > >
