This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.
Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
---|---|---|
Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |
Other format: | [Raw text] |
The problem in this PR is that we have distinct symbols which are in the same partition:
# MPT.764_162 = VDEF <MPT.764_161(D)> x = {}; # MPT.764_167 = VDEF <MPT.764_162> y = {};
DSE, being MPT unaware, thought the store to `x' was dead. With the patch below, we now verify that the base addresses of the LHS are the same.
This patch fixes the PR and does not introduce any regressions.
x.i = bar; x.j = foo;
? get_base_address will strip the component_ref I think. I guess you should use get_ref_base_and_extent and verify that base, offset and extent are the same.
Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
---|---|---|
Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |