[PATCH] Fix PR target/37483
Adam Nemet
anemet@caviumnetworks.com
Tue Sep 16 20:28:00 GMT 2008
Jakub Jelinek writes:
> When you are already reshuffling the code, it might not be a bad idea
> to use if (!(t_unconditional
> || rtx_cost (...) < COSTS_N_INSNS (2)))
Sure. Here is the updated patch.
Bootstrapped on mips-linux-gnu and regtested on mips-elf. Also bootstrapped
and regtested on x86_64-linux-gnu.
OK?
2008-09-16 Jakub Jelinek <jakub@redhat.com>
Adam Nemet <anemet@caviumnetworks.com>
PR target/37483
* ifcvt.c (noce_try_sign_mask): Use if_info->test_bb instead of
if_info->insn_b's bb as argument to optimize_bb_for_speed_p.
Rearrange code to better match the original comment. Check
t_unconditional first. Improve comment.
testsuite/
* gcc.c-torture/compile/pr37483.c: New test.
* gcc.c-torture/compile/pr37395.c: New test.
Index: ifcvt.c
===================================================================
*** ifcvt.c (revision 139918)
--- ifcvt.c (working copy)
*************** noce_try_sign_mask (struct noce_if_info
*** 1869,1875 ****
rtx cond, t, m, c, seq;
enum machine_mode mode;
enum rtx_code code;
! bool b_unconditional;
cond = if_info->cond;
code = GET_CODE (cond);
--- 1869,1875 ----
rtx cond, t, m, c, seq;
enum machine_mode mode;
enum rtx_code code;
! bool t_unconditional;
cond = if_info->cond;
code = GET_CODE (cond);
*************** noce_try_sign_mask (struct noce_if_info
*** 1898,1913 ****
if (GET_MODE (m) != mode)
return FALSE;
! /* This is only profitable if T is cheap, or T is unconditionally
! executed/evaluated in the original insn sequence. The latter
! happens if INSN_B was taken from TEST_BB, or if there was no
! INSN_B which can happen for e.g. conditional stores to memory. */
! b_unconditional = (if_info->insn_b == NULL_RTX
! || BLOCK_FOR_INSN (if_info->insn_b) == if_info->test_bb);
! if (rtx_cost (t, SET, optimize_bb_for_speed_p (BLOCK_FOR_INSN (if_info->insn_b)))
! >= COSTS_N_INSNS (2)
! && (!b_unconditional
! || t != if_info->b))
return FALSE;
start_sequence ();
--- 1898,1916 ----
if (GET_MODE (m) != mode)
return FALSE;
! /* This is only profitable if T is unconditionally executed/evaluated in the
! original insn sequence or T is cheap. The former happens if B is the
! non-zero (T) value and if INSN_B was taken from TEST_BB, or there was no
! INSN_B which can happen for e.g. conditional stores to memory. For the
! cost computation use the block TEST_BB where the evaluation will end up
! after the transformation. */
! t_unconditional =
! (t == if_info->b
! && (if_info->insn_b == NULL_RTX
! || BLOCK_FOR_INSN (if_info->insn_b) == if_info->test_bb));
! if (!(t_unconditional
! || (rtx_cost (t, SET, optimize_bb_for_speed_p (if_info->test_bb))
! < COSTS_N_INSNS (2))))
return FALSE;
start_sequence ();
Index: testsuite/gcc.c-torture/compile/pr37483.c
===================================================================
*** testsuite/gcc.c-torture/compile/pr37483.c (revision 0)
--- testsuite/gcc.c-torture/compile/pr37483.c (revision 0)
***************
*** 0 ****
--- 1,10 ----
+ /* PR target/37483 */
+
+ unsigned long long
+ foo (unsigned count, int i)
+ {
+ unsigned long long value;
+ if (i == 0)
+ value = (value & 0xFFFFFFFF) >> count;
+ return value;
+ }
Index: testsuite/gcc.c-torture/compile/pr37395.c
===================================================================
*** testsuite/gcc.c-torture/compile/pr37395.c (revision 0)
--- testsuite/gcc.c-torture/compile/pr37395.c (revision 0)
***************
*** 0 ****
--- 1,11 ----
+ /* PR target/37395 */
+
+ int
+ f (int j)
+ {
+ int i;
+ asm volatile ("" : "=r"(i));
+ if (i >= 0)
+ j = 0;
+ return j;
+ }
More information about the Gcc-patches
mailing list