On 08/11/2017 01:51 AM, Ilya Enkovich wrote:
> 2017-08-10 10:40 GMT+03:00 Martin Liška <mli...@suse.cz>:
>> Hello.
>>
>> In order to prevent the ICE, CHKP should not isntrument variables of void 
>> type.
> 
> Hi,
> 
> There was another thread for this PR where I proposed a way to handle such 
> vars
> via size relocations. But there was no feedback in that thread for two
> months and
> skipping void var is better than ICE. The patch is OK then.

Yep, probably does not worth doing that with relocations.
Thanks for review.

Martin

> 
> Thanks,
> Ilya
> 
>>
>> Patch can bootstrap on ppc64le-redhat-linux and survives regression tests.
>>
>> Ready to be installed?
>> Martin
>>
>> gcc/ChangeLog:
>>
>> 2017-08-09  Martin Liska  <mli...@suse.cz>
>>
>>         PR tree-opt/79987
>>         * tree-chkp.c (chkp_get_bounds_for_decl_addr): Do not instrument
>>         variables of void type.
>>
>> gcc/testsuite/ChangeLog:
>>
>> 2017-08-09  Martin Liska  <mli...@suse.cz>
>>
>>         PR tree-opt/79987
>>         * gcc.target/i386/mpx/pr79987.c: New test.
>> ---
>>  gcc/testsuite/gcc.target/i386/mpx/pr79987.c | 5 +++++
>>  gcc/tree-chkp.c                             | 3 +++
>>  2 files changed, 8 insertions(+)
>>  create mode 100644 gcc/testsuite/gcc.target/i386/mpx/pr79987.c
>>
>>

Reply via email to