http://bugs.winehq.org/show_bug.cgi?id=27349
--- Comment #9 from Anastasius Focht focht@gmx.net 2011-06-10 16:15:48 CDT --- Hello,
well after applying your attached patch and gathering new numbers:
kernel32 user32 gdi32 condition (cx < threshold) -------------------------------------------------------------------- c1: 0x43 0x55 0x50 0x5F c2: 0x23 *0x40 0x39 0x3C c3: 0x00 0x00 0x00 0x5A
(instead of 0x41)
I could not believe it hence I debugged the whole thing again and noticed the SafeDisc entry point analyzer was counting decoded/disassembled _instructions_, not bytes - sorry, that was a brain fart on my side (summer heat) :| I should have re-read my own stuff written some years ago more carefully ...
So inserting one hot-patch instruction will have no effect on majority of entry points. The first branch instruction must be 8 instructions away.
Regards