diff --git a/libavformat/isom.h b/libavformat/isom.h
index 41a9c64c11..78495fd336 100644
--- a/libavformat/isom.h
+++ b/libavformat/isom.h
@@ -291,6 +291,8 @@ typedef struct MOVContext {
     int decryption_key_len;
     int enable_drefs;
     int32_t movie_display_matrix[3][3]; ///< display matrix from mvhd
+    int have_read_mfra_size;
+    uint32_t mfra_size;
 } MOVContext;
 
 int ff_mp4_read_descr_len(AVIOContext *pb);
diff --git a/libavformat/mov.c b/libavformat/mov.c
index 49c2d05d21..e33031f158 100644
--- a/libavformat/mov.c
+++ b/libavformat/mov.c
@@ -5097,14 +5097,16 @@ static int mov_read_sidx(MOVContext *c, AVIOContext *pb, MOVAtom atom)
     if (!is_complete && (pb->seekable & AVIO_SEEKABLE_NORMAL)) {
         int64_t ret;
         int64_t original_pos = avio_tell(pb);
-        int32_t mfra_size;
-        if ((ret = avio_seek(pb, stream_size - 4, SEEK_SET)) < 0)
-            return ret;
-        mfra_size = avio_rb32(pb);
-        if (offset + mfra_size == stream_size)
+        if (!c->have_read_mfra_size) {
+            if ((ret = avio_seek(pb, stream_size - 4, SEEK_SET)) < 0)
+                return ret;
+            c->mfra_size = avio_rb32(pb);
+            c->have_read_mfra_size = 1;
+            if ((ret = avio_seek(pb, original_pos, SEEK_SET)) < 0)
+                return ret;
+        }
+        if (offset + c->mfra_size == stream_size)
             is_complete = 1;
-        if ((ret = avio_seek(pb, original_pos, SEEK_SET)) < 0)
-            return ret;
     }
 
     if (is_complete) {