https://bugs.kde.org/show_bug.cgi?id=523843
--- Comment #8 from Martin Cermak <[email protected]> --- Created attachment 196067 --> https://bugs.kde.org/attachment.cgi?id=196067&action=edit FTR patch assuming correct 16-alignment at the entry. The attached patch does pass whole the testsuite so it feels like a win at first. BUT it assumes the 16-alignment is always in place at the doHelperCall() entry .IOW it assumes that the misalignment is only caused by pushing the helper args to the stack. But adding a simple alignment check (which is currently commented out in this patch) seems to suggest, that such presumption is wrong. It's a strong assumption and the fact that the testsuite did pass with it was surprising anyway. That said, although this patch does seem to work, it's probably incorrect. The padding seems to need to be calculated dynamically. -- You are receiving this mail because: You are watching all bug changes.
