Introduce a new inline function called vc_compile_le16 and do the shifts
and ORs there. Depending on LE x BE.
I tried cpu_to_le16, but it ends up with worse assembly on BE for
whatever reason -- the compiler seems to be unable to optimize the swap.
Signed-off-by: Jiri Slaby <jslaby@xxxxxxx>
---
drivers/tty/vt/vc_screen.c | 29 +++++++++++++++++------------
1 file changed, 17 insertions(+), 12 deletions(-)
diff --git a/drivers/tty/vt/vc_screen.c b/drivers/tty/vt/vc_screen.c
index 10a26fd5f1b7..36b967825f68 100644
--- a/drivers/tty/vt/vc_screen.c
+++ b/drivers/tty/vt/vc_screen.c
@@ -481,6 +481,19 @@ static u16 *vcs_write_buf_noattr(struct vc_data *vc, const char *con_buf,
return org;
}
+/*
+ * Compilers (gcc 10) are unable to optimize the swap in cpu_to_le16. So do it
+ * the poor man way.
+ */
+static inline u16 vc_compile_le16(u8 hi, u8 lo)
+{
+#ifdef __BIG_ENDIAN
+ return (lo << 8u) | hi;
+#else
+ return (hi << 8u) | lo;
+#endif
+}
+
static u16 *vcs_write_buf(struct vc_data *vc, const char *con_buf,
unsigned int pos, unsigned int count, bool viewed, u16 **org0)
{
@@ -513,13 +526,8 @@ static u16 *vcs_write_buf(struct vc_data *vc, const char *con_buf,
if (pos & 1) {
count--;
c = *con_buf++;
-#ifdef __BIG_ENDIAN
- vcs_scr_writew(vc, c |
- (vcs_scr_readw(vc, org) & 0xff00), org);
-#else
- vcs_scr_writew(vc, (c << 8) |
- (vcs_scr_readw(vc, org) & 0xff), org);
-#endif
+ vcs_scr_writew(vc, vc_compile_le16(c, vcs_scr_readw(vc, org)),
+ org);
org++;
pos++;
if (++col == maxcol) {
@@ -551,11 +559,8 @@ static u16 *vcs_write_buf(struct vc_data *vc, const char *con_buf,
/* odd pos -- the remaining character */
c = *con_buf++;
-#ifdef __BIG_ENDIAN
- vcs_scr_writew(vc, (vcs_scr_readw(vc, org) & 0xff) | (c << 8), org);
-#else
- vcs_scr_writew(vc, (vcs_scr_readw(vc, org) & 0xff00) | c, org);
-#endif
+ vcs_scr_writew(vc, vc_compile_le16(vcs_scr_readw(vc, org) >> 8, c),
+ org);
return org;
}
--
2.28.0
