On Tue, Dec 31, 2019 at 05:47:54PM +0100, Richard Biener wrote:
> >Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk?
>
> Ok.
Thanks.
> >One thing I haven't done anything about yet is that there is
> >FAIL: gcc.dg/tree-ssa/popcount4ll.c scan-tree-dump-times optimized
> >".POPC
On December 31, 2019 12:00:56 PM GMT+01:00, Jakub Jelinek
wrote:
>Hi!
>
>This patch fixes various issues in the popcount{64,32}c matcher.
>
>The first issue is that it blindly uses tree_to_uhwi on all the
>INTEGER_CST
>operands. That is fine for those where we know their type, after the
>prec <=
This code:
/* Make sure we don't accidentally use the old condition. */
cond_expr = NULL_TREE;
was misplaced, since it triggered even when we needed to force the
original unmodified cond_expr into a mask temporary and then invert it.
Tested on aarch64-linux-gnu and applied as obviou
Hi!
My PR90677 fix actually made building older GCCs with newer ones worse.
The problem is that identifier_global_value used earlier returned either the
type decl on success or NULL_TREE on failure and the caller in that case
just defers handling it until later, and that is also what the C
identif
Hi!
This patch fixes various issues in the popcount{64,32}c matcher.
The first issue is that it blindly uses tree_to_uhwi on all the INTEGER_CST
operands. That is fine for those where we know their type, after the
prec <= 64 && TYPE_UNSIGNED (type)
verification, but shift counts can have differe
Hi Jerry,
OK for trunk?
Looks good. I also think that your approach that DEC stuff should
be checked later is good. If it passes the testsuite, that's good
enough for a commit.
Thanks for the patch!
Regards
Thomas
The first level is ordinary clone, corresponding to a non-self caller,
and the param_ipa_cp_max_recursive_depth-1 is for recursive cloning.
Then it's ok that the least value is 1, with which disabling does happen.
Feng
From: luoxhu
Sent: Tuesday, December
