https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101373
Torbjorn SVENSSON <azoff at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |azoff at gcc dot gnu.org
--- Comment #20 from Torbjorn SVENSSON <azoff at gcc dot gnu.org> ---
The g++.dg/torture/pr101373.C test case XPASS on arm-none-eabi for Cortex-M.
Does this line make sense in embedded context?
// { dg-xfail-run-if "PR100409" { *-*-* } }
If I run the test case on x86_64-pc-linux-gnu, it SIGSEGV on the nullptr
dereference, i.e. the throw statement in foo() is ignored. Is this correct?
On arm-non-eabi, the result depends on how the memory space is defined, but
generally speaking, reading address 0 does not cause the application to abort.
Most Cortex-M targets, that I've come across, have the initial $SP value
defined at address 0 and the entry point defined at address 4.
How can this test case be update to reflect that it sometimes is expected to
XFAIL and sometimes to PASS?