aarch64: hevcdsp: Make returns match the call site

For cases when returning early without updating any pixels, we
previously returned to return address in the caller's scope,
bypassing one function entirely. While this may seem like a neat
optimization, it makes the return stack predictor mispredict
the returns - which potentially can cost more performance than
it gains.

Secondly, if the armv9.3 feature GCS (Guarded Control Stack) is
enabled, then returns _must_ match the expected value; this feature
is being enabled across linux distributions, and by fixing the
hevc assembly, we can enable the security feature on ffmpeg as well.

(cherry picked from commit 1f7ed8a78d)

libavcodec/aarch64/hevcdsp_deblock_neon.S

@@ -511,8 +511,11 @@ function hevc_loop_filter_luma_body_\bitdepth\()_neon, export=0
         sqxtun          v6.8b, v6.8h
         sqxtun          v7.8b, v7.8h
 .endif
+        // Use x15 to signal whether any pixels should be updated or not.
+        mov             x15, #1
+        ret
+3:      mov             x15, #0
         ret
-3:      ret             x6
 endfunc
 .endm
 
@@ -562,6 +565,7 @@ function ff_hevc_\dir\()_loop_filter_luma_\bitdepth\()_neon, export=1
 .endif
 .endif
         bl              hevc_loop_filter_luma_body_\bitdepth\()_neon
+        cbz             x15, 9f
 .if \bitdepth > 8
 .ifc \dir, v
         transpose_8x8H  v0, v1, v2, v3, v4, v5, v6, v7, v16, v17
@@ -587,6 +591,7 @@ function ff_hevc_\dir\()_loop_filter_luma_\bitdepth\()_neon, export=1
         st1             {v6.8b}, [x10], x1
         st1             {v7.8b}, [x10]
 .endif
+9:
         ret             x6
 endfunc
 .endm