https://gcc.gnu.org/bugzilla/show_bug.cgi?id=78969
Jakub Jelinek <jakub at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |jakub at gcc dot gnu.org
--- Comment #7 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
This is because the PHI at that point is only created during CFG changes at the
end of VRP1 pass.
After creating ASSERT_EXPRs, we still have:
<bb 2> [1.00%]:
goto <bb 4>; [100.00%]
<bb 3> [99.00%]:
# RANGE [0, 1000] NONZERO 1023
idx_7 = ASSERT_EXPR <idx_1, idx_1 <= 999>;
__builtin_snprintf (p_4(D), 4, "%d", idx_7);
# RANGE [1, 1000] NONZERO 1023
idx_6 = idx_7 + 1;
<bb 4> [100.00%]:
# RANGE [0, 1000] NONZERO 1023
# idx_1 = PHI <0(2), idx_6(3)>
if (idx_1 <= 999)
goto <bb 3>; [99.00%]
else
goto <bb 5>; [1.00%]
<bb 5> [1.00%]:
return;
Then VRP1 correctly determines:
idx_1: [0, 1000]
.MEM_2: VARYING
idx_6: [1, 1000]
idx_7: [0, 999] EQUIVALENCES: { idx_1 } (1 elements)
then the ASSERT_EXPRs are removed, which means whenever idx_7 is used we now
use idx_1, and finally the loop is changed, idx_8 and idx_10 is created:
<bb 2> [1.00%]:
goto <bb 6>; [100.00%]
<bb 3> [99.00%]:
# RANGE [0, 1000] NONZERO 1023
# idx_10 = PHI <idx_1(4), idx_8(6)>
__builtin_snprintf (p_4(D), 4, "%d", idx_10);
# RANGE [1, 1000] NONZERO 1023
idx_6 = idx_10 + 1;
<bb 4> [99.00%]:
# RANGE [0, 1000] NONZERO 1023
# idx_1 = PHI <idx_6(3)>
if (idx_1 != 1000)
goto <bb 3>; [98.99%]
else
goto <bb 5>; [1.01%]
<bb 5> [1.00%]:
return;
<bb 6> [1.00%]:
# RANGE [0, 1000] NONZERO 1023
# idx_8 = PHI <0(2)>
goto <bb 3>; [100.00%]
But there is no obvious connection between idx_10 (which indeed nicely could
hold the [0, 999] range) and idx_7 (the already removed ASSERT_EXPR);
similarly, idx_8 could very well have RANGE [0, 0] but doesn't (though in that
case it isn't that big a deal, because it is going to be removed immediately
afterwards).
