My apologies for the short delay. Thanks for explaining why these
transforms
don't usually require explicit tests for side-effects (but occasionally do).
This does simplify things; please find attached the shorter revised patch.
This patch has been retested on x86_64-pc-linux-gnu with a make boo
The patch generalises how analyzer's call_string tracks function calls,
including the calls that doesn’t have an underlying call graph edge, making it
possible to work with function calls which are not discovered by GCC's middle
end.
Successfully bootstrapped and completed regress tests on x86_
This patch replaces the evrp_range_analyzer in the loop versioning code
with an on-demand ranger.
Everything was pretty straightforward, except that range_of_expr requires
a gimple statement as context to provide context aware ranges. I didn't see
a convient place where the statement was saved, s
On Sat, 2021-07-24 at 15:22 +0530, Ankur Saini wrote:
> The patch generalises how analyzer's call_string tracks function calls,
> including the calls that doesn’t have an underlying call graph edge,
> making it possible to work with function calls which are not discovered
> by GCC's middle end.
>
On 7/22/21 1:54 AM, Tobias Burnus wrote:
Hi Sandra,
On 21.07.21 20:01, Sandra Loosemore wrote:
Hmmm. CFI_establish explicitly says that the elem_len has to be
greater than zero. It seems somewhat confusing that it's inconsistent
with the other functions that take an elem_len argument.
Congr
