target/arm: Use uint32_t in t32_expandimm_imm()

In t32_expandimm_imm(), we take an 8 bit value XY and construct a
32-bit value which might be of the form XY, 00XY00XY, XY00XY00, or
XYXYXYXY.  We do this with multiplications, and we use an 'int' type.
For the cases where we're setting the high byte of the 32-bit value
to XY, this means that we do an integer multiplication that might
overflow, and rely on the -fwrapv semantics to keep this from being
undefined behaviour.

It's clearer to use an unsigned type here, because we're really
doing operations on the value considered as a set of bits. The
result is the same.

The return value from the function remains 'int', because this
is a decodetree !function function, and follows the API for those
functions.

Signed-off-by: Stephen Longfield <slongfield@google.com>
Signed-off-by: Roque Arcudia Hernandez <roqueh@google.com>
Message-id: 20250219165534.3387376-1-slongfield@google.com
[PMM: Rewrote the commit message]
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

diff --git a/target/arm/tcg/translate.c b/target/arm/tcg/translate.c
index 68ac39341530f6d6e7421d42943ac25497d0d8f3..d8225b77c8cc3d3abd96cd71560cdc5c4a7e3b75 100644 (file)
--- a/target/arm/tcg/translate.c
+++ b/target/arm/tcg/translate.c
@@ -3510,7 +3510,7 @@ static int t32_expandimm_rot(DisasContext *s, int x)
 /* Return the unrotated immediate from T32ExpandImm.  */
 static int t32_expandimm_imm(DisasContext *s, int x)
 {
-    int imm = extract32(x, 0, 8);
+    uint32_t imm = extract32(x, 0, 8);
 
     switch (extract32(x, 8, 4)) {
     case 0: /* XY */