mach64: optimize wait_for_fifo

This is a simple optimization for fifo waiting that improves scrolling
performance by 5%. If the queue has more free entries that what we
consume, we can skip the costly register read next time.

Signed-off-by: Mikulas Patocka <[e-mail]>
Reviewed-by: Ville Syrjälä <[e-mail]>
Signed-off-by: Bartlomiej Zolnierkiewicz <[e-mail]>

drivers/video/fbdev/aty/atyfb.h

···
 	u16 pci_id;
 	u32 accel_flags;
 	int blitter_may_be_busy;
+	unsigned fifo_space;
 	int asleep;
 	int lock_blank;
 	unsigned long res_start;
···
      *  Hardware acceleration
      */
 
-static inline void wait_for_fifo(u16 entries, const struct atyfb_par *par)
+static inline void wait_for_fifo(u16 entries, struct atyfb_par *par)
 {
-	while ((aty_ld_le32(FIFO_STAT, par) & 0xffff) >
-	       ((u32) (0x8000 >> entries)));
+	unsigned fifo_space = par->fifo_space;
+	while (entries > fifo_space) {
+		fifo_space = 16 - fls(aty_ld_le32(FIFO_STAT, par) & 0xffff);
+	}
+	par->fifo_space = fifo_space - entries;
 }
 
 static inline void wait_for_idle(struct atyfb_par *par)
···
 	par->blitter_may_be_busy = 0;
 }
 
-extern void aty_reset_engine(const struct atyfb_par *par);
+extern void aty_reset_engine(struct atyfb_par *par);
 extern void aty_init_engine(struct atyfb_par *par, struct fb_info *info);
 
 void atyfb_copyarea(struct fb_info *info, const struct fb_copyarea *area);

drivers/video/fbdev/aty/mach64_accel.c

···
 	return ((rotation << 8) | DST_24_ROTATION_ENABLE);
 }
 
-void aty_reset_engine(const struct atyfb_par *par)
+void aty_reset_engine(struct atyfb_par *par)
 {
 	/* reset engine */
 	aty_st_le32(GEN_TEST_CNTL,
···
 	/* HOST errors */
 	aty_st_le32(BUS_CNTL,
 		aty_ld_le32(BUS_CNTL, par) | BUS_HOST_ERR_ACK | BUS_FIFO_ERR_ACK, par);
+
+	par->fifo_space = 0;
 }
 
 static void reset_GTC_3D_engine(const struct atyfb_par *par)