authorgravatar for mlugg@mlugg.co.ukMatthew Lugg <mlugg@mlugg.co.uk> 2025-10-09 14:14:19+01:00
committergravatar for mlugg@mlugg.co.ukMatthew Lugg <mlugg@mlugg.co.uk> 2025-10-09 19:31:44+01:00
log80f6b8c4b3a6dcb17dcd9bb3ecaf528b3da9c54e
treea3d7e036a042b687b9995c3054f467673c9c0b68
parentc383cd50d5eb3b677493ab0058d860a9177a2679

std.debug: fix incorrect FP unwinding on RISC-V and SPARC

I broke this when porting this logic for the `std.debug` rework in https://github.com/ziglang/zig/pull/25227. The offset that I copied was actually being treated as relative to the address of the *saved* base pointer. I think it makes more sense to do what I did and just treat all offsets as relative to this frame's base.

1 files changed, 2 insertions(+), 0 deletions(-)

lib/std/debug.zig+2
...@@ -993,6 +993,8 @@ const StackIterator = union(enum) {...@@ -993,6 +993,8 @@ const StackIterator = union(enum) {
993993
994 /// Offset of the saved return address wrt the frame pointer.994 /// Offset of the saved return address wrt the frame pointer.
995 const ra_offset = off: {995 const ra_offset = off: {
996 if (native_arch.isRISCV()) break :off -1 * @sizeOf(usize);
997 if (native_arch.isSPARC()) break :off 15 * @sizeOf(usize);
996 if (native_arch.isPowerPC64()) break :off 2 * @sizeOf(usize);998 if (native_arch.isPowerPC64()) break :off 2 * @sizeOf(usize);
997 // On s390x, r14 is the link register and we need to grab it from its customary slot in the999 // On s390x, r14 is the link register and we need to grab it from its customary slot in the
998 // register save area (ELF ABI s390x Supplement §1.2.2.2).1000 // register save area (ELF ABI s390x Supplement §1.2.2.2).