For PPC64 we skip resetting the return register if it is already set.
This is because on PPC64 there are two DWARF registers numbers that can
represent the same register. Setting the return address again confuses
the unwinder. But we do want to reset it if the register number
(non-translated by the ppc64 ebl) is equal to the actual register number
as set in the CIE as return address. This happens on older toolchains in
.debug_frame where the return address is set to 108, but the ebl abi_cfi
also sets register number 65.
Signed-off-by: Mark Wielaard <mjw@redhat.com>
+2014-01-05 Mark Wielaard <mjw@redhat.com>
+
+ * frame_unwind.c (handle_cfi): Only skip resetting return register
+ if the regno is not the actual CIE return address register.
+
2014-01-02 Mark Wielaard <mjw@redhat.com>
* linux-pid-attach.c (dwfl_linux_proc_attach): Use strtol, not atoi.
/* This is another strange PPC[64] case. There are two
registers numbers that can represent the same DWARF return
register number. We only want one to actually set the return
- register value. */
- if (ra_set)
+ register value. But we always want to override the value if
+ the register is the actual CIE return address register. */
+ if (ra_set && regno != frame->fde->cie->return_address_register)
{
unsigned r = regno;
if (ebl_dwarf_to_regno (ebl, &r) && r == ra)