Use more accurate int16_t type

Use more accurate voice sample type.
int16_t is architecture agnostic one.

src/chan_quectel.c

@@ -1363,7 +1363,7 @@ static size_t pvt_get_audio_frame_size_r(unsigned int ptime, const unsigned int
 {
     size_t res  = ptime;
     res        *= sr / 1000;
-    res        *= sizeof(short);
+    res        *= sizeof(int16_t);
 
     return res;
 }

src/channel.c

@@ -562,7 +562,7 @@ static struct ast_frame* channel_read_uac(struct cpvt* cpvt, struct pvt* pvt, si
                         write_conference(pvt, buf, res);
                     }
 
-                    PVT_STAT(pvt, a_read_bytes) += res * sizeof(short);
+                    PVT_STAT(pvt, a_read_bytes) += res * sizeof(int16_t);
                     PVT_STAT(pvt, read_frames)++;
                     if (res < frames) {
                         PVT_STAT(pvt, read_sframes)++;
@@ -623,7 +623,7 @@ static struct ast_frame* channel_read(struct ast_channel* channel)
     }
 
     if (CONF_UNIQ(pvt, uac) > TRIBOOL_FALSE && CPVT_IS_MASTER(cpvt)) {
-        f = channel_read_uac(cpvt, pvt, frame_size / sizeof(short), fmt);
+        f = channel_read_uac(cpvt, pvt, frame_size / sizeof(int16_t), fmt);
     } else {
         f = channel_read_tty(cpvt, pvt, frame_size, fmt);
     }
@@ -633,7 +633,7 @@ static struct ast_frame* channel_read(struct ast_channel* channel)
         const int fd = ast_channel_fd(channel, 0);
         ast_debug(5, "[%s] Read - idx:%d state:%s audio:%d:%d - returning SILENCE frame\n", PVT_ID(pvt), cpvt->call_idx, call_state2str(cpvt->state), fd,
                   pvt->audio_fd);
-        return cpvt_prepare_silence_voice_frame(cpvt, frame_size / sizeof(short), fmt);
+        return cpvt_prepare_silence_voice_frame(cpvt, frame_size / sizeof(int16_t), fmt);
     } else {
         ast_debug(8, "[%s] Read - idx:%d state:%s samples:%d\n", PVT_ID(pvt), cpvt->call_idx, call_state2str(cpvt->state), f->samples);
         return f;
@@ -778,7 +778,7 @@ static int channel_write_uac(struct ast_channel* attribute_unused(channel), stru
         default:
             if (res >= 0) {
                 PVT_STAT(pvt, write_frames)  += 1;
-                PVT_STAT(pvt, a_write_bytes) += res * sizeof(short);
+                PVT_STAT(pvt, a_write_bytes) += res * sizeof(int16_t);
                 if (res != samples) {
                     PVT_STAT(pvt, write_tframes)++;
                     ast_log(LOG_WARNING, "[%s][ALSA][PLAYBACK] Write: %d/%d\n", PVT_ID(pvt), res, samples);

src/cpvt.c

@@ -418,7 +418,7 @@ struct ast_frame* cpvt_prepare_voice_frame(struct cpvt* const cpvt, void* const
     f->frametype       = AST_FRAME_VOICE;
     f->subclass.format = (struct ast_format*)fmt;
     f->samples         = samples;
-    f->datalen         = samples * sizeof(short);
+    f->datalen         = samples * sizeof(int16_t);
     f->data.ptr        = buf;
     f->offset          = AST_FRIENDLY_OFFSET;
     f->src             = AST_MODULE;