Fix inline fixnum LDB on PowerPC for certain bytespecs

[sbcl.git] / src / compiler / ppc / arith.lisp

diff --git a/src/compiler/ppc/arith.lisp b/src/compiler/ppc/arith.lisp
index 6e90c89..5f4bfe1 100644 (file)
--- a/src/compiler/ppc/arith.lisp
+++ b/src/compiler/ppc/arith.lisp
@@ -679,10 +679,22 @@
   (:result-types tagged-num)
   (:policy :fast-safe)
   (:generator 2
-    (inst rlwinm res x
-          (mod (- 32 posn) 32)          ; effectively rotate right
-          (- 32 size n-fixnum-tag-bits)
-          (- 31 n-fixnum-tag-bits))))
+    (let ((phantom-bits (- (+ size posn) 30)))
+      (cond
+        ((plusp phantom-bits)
+         ;; The byte to be loaded into RES includes sign bits which are not
+         ;; present in the input X physically.  RLWINM as used below would
+         ;; mask these out with 0 even for negative inputs.
+         (inst srawi res x phantom-bits)
+         (inst rlwinm res x
+               (mod (- 32 posn (- phantom-bits)) 32)
+               (- 32 size n-fixnum-tag-bits)
+               (- 31 n-fixnum-tag-bits)))
+        (t
+         (inst rlwinm res x
+               (mod (- 32 posn) 32)     ; effectively rotate right
+               (- 32 size n-fixnum-tag-bits)
+               (- 31 n-fixnum-tag-bits)))))))
 
 (define-vop (ldb-c/signed)
   (:translate %%ldb)