[gcc r10-9108] dse: Cope with bigger-than-integer modes [PR98037]

[gcc r10-9108] dse: Cope with bigger-than-integer modes [PR98037]

[Richard Sandiford]

https://gcc.gnu.org/g:fb9ee3e7419178e6530c09b99272022d062b50f3

commit r10-9108-gfb9ee3e7419178e6530c09b99272022d062b50f3
Author: Richard Sandiford <richard.sandiford@arm.com>
Date:   Wed Dec 2 16:20:34 2020 +0000

    dse: Cope with bigger-than-integer modes [PR98037]
    
    dse.c:find_shift_sequence tries to represent a store and load
    back as a shift right followed by a truncation.  It therefore
    needs to find an integer mode in which to do the shift right.
    The loop it uses has the form:
    
      FOR_EACH_MODE_FROM (new_mode_iter,
                          smallest_int_mode_for_size (GET_MODE_BITSIZE (read_mode)))
    
    which implicitly assumes that read_mode has an equivalent integer mode.
    As shown in the testcase, not all modes have such an integer mode.
    
    This patch just makes the code start from the smallest integer mode and
    skip modes that are too small.  The loop already breaks at the first
    mode wider than word_mode.
    
    gcc/
            PR rtl-optimization/98037
            * dse.c (find_shift_sequence): Iterate over all integers and
            skip modes that are too small.
    
    gcc/testsuite/
            PR rtl-optimization/98037
            * gcc.target/aarch64/sve/acle/general/pr98037.c: New test.
    
    (cherry picked from commit f835e9f6562dda9c8a1384be2c9d4e45c112ed8e)

Diff:
---
 gcc/dse.c                                                   | 5 +++--
 gcc/testsuite/gcc.target/aarch64/sve/acle/general/pr98037.c | 6 ++++++
 2 files changed, 9 insertions(+), 2 deletions(-)

diff --git a/gcc/dse.c b/gcc/dse.c
index bbe792e48e8..c8dcf46ed29 100644
--- a/gcc/dse.c
+++ b/gcc/dse.c
@@ -1728,8 +1728,7 @@ find_shift_sequence (poly_int64 access_size,
      the machine.  */
 
   opt_scalar_int_mode new_mode_iter;
-  FOR_EACH_MODE_FROM (new_mode_iter,
-		      smallest_int_mode_for_size (access_size * BITS_PER_UNIT))
+  FOR_EACH_MODE_IN_CLASS (new_mode_iter, MODE_INT)
     {
       rtx target, new_reg, new_lhs;
       rtx_insn *shift_seq, *insn;
@@ -1738,6 +1737,8 @@ find_shift_sequence (poly_int64 access_size,
       new_mode = new_mode_iter.require ();
       if (GET_MODE_BITSIZE (new_mode) > BITS_PER_WORD)
 	break;
+      if (maybe_lt (GET_MODE_SIZE (new_mode), access_size))
+	continue;
 
       /* If a constant was stored into memory, try to simplify it here,
 	 otherwise the cost of the shift might preclude this optimization
diff --git a/gcc/testsuite/gcc.target/aarch64/sve/acle/general/pr98037.c b/gcc/testsuite/gcc.target/aarch64/sve/acle/general/pr98037.c
new file mode 100644
index 00000000000..b91e940b18e
--- /dev/null
+++ b/gcc/testsuite/gcc.target/aarch64/sve/acle/general/pr98037.c
@@ -0,0 +1,6 @@
+/* { dg-options "-msve-vector-bits=1024 -O3" } */
+
+typedef __SVInt8_t vec __attribute__((arm_sve_vector_bits(1024)));
+struct pair { vec v[2]; };
+void use (struct pair *);
+vec f (struct pair p) { vec v = p.v[1]; use (&p); return v; }