https://bugs.kde.org/show_bug.cgi?id=517596
--- Comment #5 from Carl Love <[email protected]> --- Florian: Yes, looks like there is a bug. The (immediate16_lo << 24) and (immediate16_hi << 24) are missing. Thanks for finding the issue. One issue with the patch. The patch for the immediate16_low has a replicated line. You add the (immediate16_lo << 24) on the line with the (immediate16_lo << 32) and (immediate16_lo << 16): but didnt' remove the original line like you did for the immediate64_hi line + immediate64_hi = ((immediate16_hi << 56) | (immediate16_hi << 48) | (immediate16_hi << 40) | - (immediate16_hi << 32) | (immediate16_hi << 16) | << removed the original line + (immediate16_hi << 32) | (immediate16_hi << 24) | (immediate16_hi << 16) | << new line with the missing << 24 (immediate16_hi << 8) | immediate16_hi); + immediate64_lo = ((immediate16_lo << 56) | (immediate16_lo << 48) | (immediate16_lo << 40) | + (immediate16_lo << 32) | (immediate16_lo << 24) | (immediate16_lo << 16) | << new line with the missing << 24 (immediate16_lo << 32) | (immediate16_lo << 16) | << The original line was not removed, don't need this line now. (immediate16_lo << 8) | immediate16_lo); Carl On 3/30/26 5:51 AM, Florian Krohm wrote: > https://bugs.kde.org/show_bug.cgi?id=517596 > > --- Comment #3 from Florian Krohm<[email protected]> --- > Created attachment 191134 > -->https://bugs.kde.org/attachment.cgi?id=191134&action=edit > proposed fix with test > > When bm = 8, the corresponding byte will not be set to all-1. > bm = 8 yields the same result as bm = 0. As do bm = 16 + 8, bm = 32 + 8 and so > on.. > Attached patch fixes that. > I'll leave the BZ open for ppc folks to review. > -- You are receiving this mail because: You are watching all bug changes.
