git
/
gcc
mirror of git://gcc.gnu.org/git/gcc.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

rs6000: Don't add an immediate to r0 (PR80966) 

If there is a large stack frame the rs6000 -fstack-limit code would
calculate the new stack pointer value using two insns (an addis and
an addi), with r0 as temporary.  Such instructions do not exist.

This patch changes add<mode>3 to expand using a different strategy in
such cases; to FAIL if there is no way to do it (namely, if the source
is r0 and there is no way to get a temporary reg); and it changes
rs6000_emit_allocate_stack to assert gen_add3_insn did in fact emit
instructions.


	PR target/80966
	* config/rs6000/rs6000.c (rs6000_emit_allocate_stack): Assert that
	gen_add3_insn did not fail.
	* config/rs6000/rs6000.md (add<mode>3): If asked to add a constant to
	r0, construct that number in a temporary reg and add that reg to r0.
	If asked to put the result in r0 as well, fail.

gcc/testsuite/
	* gcc.target/powerpc/stack-limit.c: New testcase.

From-SVN: r249046
This commit is contained in:
Segher Boessenkool 2017-06-09 05:46:08 +02:00 committed by Segher Boessenkool
parent ff67045361
commit 120ef1d50a
5 changed files with 40 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
gcc/ChangeLog
View File

@ -1,3 +1,12 @@
2017-06-09 Segher Boessenkool <segher@kernel.crashing.org>
PR target/80966
* config/rs6000/rs6000.c (rs6000_emit_allocate_stack): Assert that
gen_add3_insn did not fail.
* config/rs6000/rs6000.md (add<mode>3): If asked to add a constant to
r0, construct that number in a temporary reg and add that reg to r0.
If asked to put the result in r0 as well, fail.
2017-06-08 Will Schmidt <will_schmidt@vnet.ibm.com>
* config/rs6000/rs6000.c (rs6000_gimple_fold_builtin): Add handling

8
gcc/config/rs6000/rs6000.c
View File

@ -27005,9 +27005,11 @@ rs6000_emit_allocate_stack (HOST_WIDE_INT size, rtx copy_reg, int copy_off)
&& REGNO (stack_limit_rtx) > 1
&& REGNO (stack_limit_rtx) <= 31)
{
emit_insn (gen_add3_insn (tmp_reg, stack_limit_rtx, GEN_INT (size)));
emit_insn (gen_cond_trap (LTU, stack_reg, tmp_reg,
const0_rtx));
rtx_insn *insn
= gen_add3_insn (tmp_reg, stack_limit_rtx, GEN_INT (size));
gcc_assert (insn);
emit_insn (insn);
emit_insn (gen_cond_trap (LTU, stack_reg, tmp_reg, const0_rtx));
}
else if (GET_CODE (stack_limit_rtx) == SYMBOL_REF
&& TARGET_32BIT

11
gcc/config/rs6000/rs6000.md
View File

@ -1629,6 +1629,17 @@
|| rtx_equal_p (operands[0], operands[1]))
? operands[0] : gen_reg_rtx (<MODE>mode));
/* Adding a constant to r0 is not a valid insn, so use a different
strategy in that case. */
if (REGNO (operands[1]) == 0 || REGNO (tmp) == 0)
{
if (operands[0] == operands[1])
FAIL;
rs6000_emit_move (operands[0], operands[2], <MODE>mode);
emit_insn (gen_add<mode>3 (operands[0], operands[1], operands[0]));
DONE;
}
HOST_WIDE_INT val = INTVAL (operands[2]);
HOST_WIDE_INT low = ((val & 0xffff) ^ 0x8000) - 0x8000;
HOST_WIDE_INT rest = trunc_int_for_mode (val - low, <MODE>mode);

5
gcc/testsuite/ChangeLog
View File

@ -1,3 +1,8 @@
2017-06-09 Segher Boessenkool <segher@kernel.crashing.org>
PR target/80966
* gcc.target/powerpc/stack-limit.c: New testcase.
2017-06-08 Will Schmidt <will_schmidt@vnet.ibm.com>
* testsuite/gcc.target/powerpc/fold-vec-logical-eqv-char.c: New.

10
gcc/testsuite/gcc.target/powerpc/stack-limit.c Normal file
View File

@ -0,0 +1,10 @@
/* { dg-options "-O0 -fstack-limit-register=r14" } */
// PR80966
int foo (int i)
{
char arr[135000];
arr[i] = 0;
}