Skip to contentSkip to conversation

vidtv: fix several things on it

PATCH New Mauro Carvalho Chehab 1 participant 10 messages Sep 21, 2020

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

cover vidtv: fix several things on it

There are still several issues at vidtv driver. This series address the ones I found so far.

After this series, the generated TS output can now be decoded by Kaffeine, and probably by other DVB players.

There is still something weird there, though: after playing some notes, there's a long silence that happens only when played with VLC. I suspect that the package times are not ok yet, but further investigation is required.

Mauro Carvalho Chehab (9):
media: vidtv: simplify PCR logic to get jiffies
media: vidtv: remove more ENDIAN_BITFIELD nonsense
media: vidtv: fix initialization of the network_id field at SDT
media: vidtv: cleanup PSI version numbers
media: vidtv: cleanup SDT string identifiers
media: vidtv: fix a typo
media: vidtv: simplify parameters for vidtv_pes_write_stuffing()
media: vidtv: rewrite the adaption field logic
media: vidtv: fix decoding with gstreamer and Vlc

 .../media/test-drivers/vidtv/vidtv_channel.c  |   7 +-
 drivers/media/test-drivers/vidtv/vidtv_mux.c  |   8 +-
 drivers/media/test-drivers/vidtv/vidtv_pes.c  | 216 +++++++++++-------
 drivers/media/test-drivers/vidtv/vidtv_pes.h  |   4 +-
 drivers/media/test-drivers/vidtv/vidtv_psi.c  |  18 +-
 drivers/media/test-drivers/vidtv/vidtv_psi.h  |   8 -
 drivers/media/test-drivers/vidtv/vidtv_ts.h   |  13 --
 7 files changed, 153 insertions(+), 121 deletions(-)
Signature · 1 line
2.26.2
#1 lore raw

PATCH 1/9 media: vidtv: simplify PCR logic to get jiffies

There's no need to have a check when setting timing.past_jiffies, as we can simply do the initialization earlier at vidtv_mux_init().

---
 drivers/media/test-drivers/vidtv/vidtv_mux.c | 7 ++-----
 1 file changed, 2 insertions(+), 5 deletions(-)
drivers/media/test-drivers/vidtv/vidtv_mux.c +2 -5
index 43b13aa3a8bc..6a3123c51cc4 100644
@@ -100,11 +100,7 @@ static void vidtv_mux_update_clk(struct vidtv_mux *m)
 	/* call this at every thread iteration */
 	u64 elapsed_time;
 
-	/* this will not hold a value yet if we have just started */
-	m->timing.past_jiffies = m->timing.current_jiffies ?
-				 m->timing.current_jiffies :
-				 get_jiffies_64();
-
+	m->timing.past_jiffies = m->timing.current_jiffies;
 	m->timing.current_jiffies = get_jiffies_64();
 
 	elapsed_time = jiffies_to_usecs(m->timing.current_jiffies -
@@ -449,6 +445,7 @@ struct vidtv_mux *vidtv_mux_init(struct dvb_frontend *fe,
 	m->pcr_pid = args.pcr_pid;
 	m->transport_stream_id = args.transport_stream_id;
 	m->priv = args.priv;
+	m->timing.current_jiffies = get_jiffies_64();
 
 	if (args.channels)
 		m->channels = args.channels;
Signature · 1 line
2.26.2
#2 lore raw

PATCH 2/9 media: vidtv: remove more ENDIAN_BITFIELD nonsense

In reply to Mauro Carvalho Chehab

Changeset 870e350d4e39 ("media: vidtv: get rid of ENDIAN_BITFIELD nonsense") was incomplete. There are still some wrong endannes logic at
the driver. Get rid of them.

---
 drivers/media/test-drivers/vidtv/vidtv_psi.h |  8 --------
 drivers/media/test-drivers/vidtv/vidtv_ts.h  | 13 -------------
 2 files changed, 21 deletions(-)
drivers/media/test-drivers/vidtv/vidtv_psi.h +0 -8
index c9e059bfb6eb..3f962cc78278 100644
@@ -132,17 +132,9 @@ struct vidtv_psi_table_pat {
  */
 struct vidtv_psi_table_sdt_service {
 	__be16 service_id;
-#if defined(__LITTLE_ENDIAN_BITFIELD)
 	u8 EIT_present_following:1;
 	u8 EIT_schedule:1;
 	u8 reserved:6;
-#elif defined(__BIG_ENDIAN_BITFIELD)
-	u8 reserved:6;
-	u8 EIT_schedule:1;
-	u8 EIT_present_following:1;
-#else
-#error  "Unknown bitfield ordering"
-#endif
 	__be16 bitfield; /* running_status: 3, free_ca:1, desc_loop_len:12 */
 	struct vidtv_psi_desc *descriptor;
 	struct vidtv_psi_table_sdt_service *next;
drivers/media/test-drivers/vidtv/vidtv_ts.h +0 -13
index fab523723633..83dcc9183b45 100644
@@ -24,7 +24,6 @@
 struct vidtv_mpeg_ts_adaption {
 	u8 length;
 	struct {
-#if defined(__LITTLE_ENDIAN_BITFIELD)
 		u8 extension:1;
 		u8 private_data:1;
 		u8 splicing_point:1;
@@ -33,18 +32,6 @@ struct vidtv_mpeg_ts_adaption {
 		u8 priority:1;
 		u8 random_access:1;
 		u8 discontinued:1;
-#elif defined(__BIG_ENDIAN_BITFIELD)
-		u8 discontinued:1;
-		u8 random_access:1;
-		u8 priority:1;
-		u8 PCR:1;
-		u8 OPCR:1;
-		u8 splicing_point:1;
-		u8 private_data:1;
-		u8 extension:1;
-#else
-#error  "Unknown bitfield ordering"
-#endif
 	} __packed;
 	u8 data[];
 } __packed;
Signature · 1 line
2.26.2
#3 lore raw

PATCH 3/9 media: vidtv: fix initialization of the network_id field at SDT

In reply to Mauro Carvalho Chehab

This field should point to the network ID, and has different ranges for cable, terrestrial or satellite. It also has an special range for temporary private usage.

For now, let's use the temporary private one. Once the Network Information Table (NIT) gets added, this should be reviewed.

---
 drivers/media/test-drivers/vidtv/vidtv_psi.c | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)
drivers/media/test-drivers/vidtv/vidtv_psi.c +8 -1
index 8cdf645b4fdd..5fe3fbb765b8 100644
@@ -1107,7 +1107,14 @@ struct vidtv_psi_table_sdt *vidtv_psi_sdt_table_init(u16 transport_stream_id)
 	sdt->header.section_id   = 0;
 	sdt->header.last_section = 0;
 
-	sdt->network_id = cpu_to_be16(transport_stream_id);
+	/*
+	 * FIXME: The network_id range from 0xFF01 to 0xFFFF is used to
+	 * indicate temporary private use. For now, let's use the first
+	 * value.
+	 * This can be changed to something more useful, when support for
+	 * NIT gets added
+	 */
+	sdt->network_id = cpu_to_be16(0xff01);
 	sdt->reserved = RESERVED;
 
 	vidtv_psi_sdt_table_update_sec_len(sdt);
Signature · 1 line
2.26.2
#4 lore raw

PATCH 4/9 media: vidtv: cleanup PSI version numbers

In reply to Mauro Carvalho Chehab

There's no reason to use static vars to store PSI version numbers.

Also, currently, version numbers are starting with 0x01, because there's a code being called that increases it to 1 for all table initializer code, as the code may support dynamic changes at the PS tables on some future.

So, let's just initialize them to 0x1f, in order for the versions to be reported as starting from 0.

---
 drivers/media/test-drivers/vidtv/vidtv_psi.c | 9 +++------
 1 file changed, 3 insertions(+), 6 deletions(-)
drivers/media/test-drivers/vidtv/vidtv_psi.c +3 -6
index 5fe3fbb765b8..82cf67dd27c0 100644
@@ -756,7 +756,6 @@ vidtv_psi_pat_program_assign(struct vidtv_psi_table_pat *pat,
 struct vidtv_psi_table_pat *vidtv_psi_pat_table_init(u16 transport_stream_id)
 {
 	struct vidtv_psi_table_pat *pat = kzalloc(sizeof(*pat), GFP_KERNEL);
-	static u8 pat_version;
 	const u16 SYNTAX = 0x1;
 	const u16 ZERO = 0x0;
 	const u16 ONES = 0x03;
@@ -767,7 +766,7 @@ struct vidtv_psi_table_pat *vidtv_psi_pat_table_init(u16 transport_stream_id)
 	pat->header.id           = cpu_to_be16(transport_stream_id);
 	pat->header.current_next = 0x1;
 
-	pat->header.version = pat_version;
+	pat->header.version = 0x1f;
 
 	pat->header.one2         = 0x03;
 	pat->header.section_id   = 0x0;
@@ -935,7 +934,6 @@ struct vidtv_psi_table_pmt *vidtv_psi_pmt_table_init(u16 program_number,
 						     u16 pcr_pid)
 {
 	struct vidtv_psi_table_pmt *pmt = kzalloc(sizeof(*pmt), GFP_KERNEL);
-	static u8 pmt_version;
 	const u16 SYNTAX = 0x1;
 	const u16 ZERO = 0x0;
 	const u16 ONES = 0x03;
@@ -953,7 +951,7 @@ struct vidtv_psi_table_pmt *vidtv_psi_pmt_table_init(u16 program_number,
 	pmt->header.id = cpu_to_be16(program_number);
 	pmt->header.current_next = 0x1;
 
-	pmt->header.version = pmt_version;
+	pmt->header.version = 0x1f;
 
 	pmt->header.one2 = ONES;
 	pmt->header.section_id   = 0;
@@ -1083,7 +1081,6 @@ void vidtv_psi_pmt_table_destroy(struct vidtv_psi_table_pmt *pmt)
 struct vidtv_psi_table_sdt *vidtv_psi_sdt_table_init(u16 transport_stream_id)
 {
 	struct vidtv_psi_table_sdt *sdt = kzalloc(sizeof(*sdt), GFP_KERNEL);
-	static u8 sdt_version;
 	const u16 SYNTAX = 0x1;
 	const u16 ONE = 0x1;
 	const u16 ONES = 0x03;
@@ -1101,7 +1098,7 @@ struct vidtv_psi_table_sdt *vidtv_psi_sdt_table_init(u16 transport_stream_id)
 	sdt->header.id = cpu_to_be16(transport_stream_id);
 	sdt->header.current_next = ONE;
 
-	sdt->header.version = sdt_version;
+	sdt->header.version = 0x1f;
 
 	sdt->header.one2  = ONES;
 	sdt->header.section_id   = 0;
Signature · 1 line
2.26.2
#5 lore raw

PATCH 5/9 media: vidtv: cleanup SDT string identifiers

In reply to Mauro Carvalho Chehab

In order to better test userspace applications, add an encoding for the strings. The original channel string is also too big. As we're just playing the first compasses of the 5th Beethoven Symphony, change the service name to just "Beethoven". Also, add a provider's name. For such purpose, let's point it to the Linux media community website.

With such changes, the SDT table now looks more like one got from a real digital TV channel:

SDT

| table_id         0x42
| section_length      44
| one                 2
| zero                1
| syntax              1
| transport_stream_id 1860
| current_next        1
| version             0
| one2                3
| section_number      0
| last_section_number 0
| network_id          65281
| reserved            255
|\
|- service 0x0880
|   EIT schedule          0
|   EIT present following 0
|   free CA mode          0
|   running status        4
|   descriptor length     27
|        0x48: service_descriptor
|           service type  1
|           provider      'LinuxTV.org'
|           name          'Beethoven'
|_  1 services
---
 drivers/media/test-drivers/vidtv/vidtv_channel.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)
drivers/media/test-drivers/vidtv/vidtv_channel.c +5 -2
index 0a62b97e6a27..f2b97cf08e87 100644
@@ -43,6 +43,8 @@ static void vidtv_channel_encoder_destroy(struct vidtv_encoder *e)
 	}
 }
 
+#define ENCODING_ISO8859_15 "\x0b"
+
 struct vidtv_channel
 *vidtv_channel_s302m_init(struct vidtv_channel *head, u16 transport_stream_id)
 {
@@ -55,7 +57,8 @@ struct vidtv_channel
 	const u16 s302m_es_pid              = 0x111; /* packet id for the ES */
 	const __be32 s302m_fid              = cpu_to_be32(VIDTV_S302M_FORMAT_IDENTIFIER);
 
-	char *name = "S302m: Sine Wave PCM Audio";
+	char *name = ENCODING_ISO8859_15 "Beethoven";
+	char *provider = ENCODING_ISO8859_15 "LinuxTV.org";
 
 	struct vidtv_channel *s302m = kzalloc(sizeof(*s302m), GFP_KERNEL);
 	struct vidtv_s302m_encoder_init_args encoder_args = {};
@@ -68,7 +71,7 @@ struct vidtv_channel
 				     vidtv_psi_service_desc_init(NULL,
 								 DIGITAL_TELEVISION_SERVICE,
 								 name,
-								 NULL);
+								 provider);
 
 	s302m->transport_stream_id = transport_stream_id;
 
Signature · 1 line
2.26.2
#6 lore raw

PATCH 6/9 media: vidtv: fix a typo

In reply to Mauro Carvalho Chehab

optinal -> optional

---
 drivers/media/test-drivers/vidtv/vidtv_pes.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/media/test-drivers/vidtv/vidtv_pes.h +1 -1
index 36443a6e30d3..84628780d8c6 100644
@@ -44,7 +44,7 @@ struct vidtv_pes_optional_pts_dts {
 struct vidtv_pes_optional {
 	/*
 	 * These flags show which components are actually
-	 * present in the "optinal fields" in the optinal PES
+	 * present in the "optional fields" in the optional PES
 	 * header and which are not
 	 *
 	 * u16 two:2;  //0x2
Signature · 1 line
2.26.2
#7 lore raw

PATCH 7/9 media: vidtv: simplify parameters for vidtv_pes_write_stuffing()

In reply to Mauro Carvalho Chehab

Instead of passing struct pes_ts_header_write_args fields as function parameters, just pass a pointer to the struct.

That would allow adding more args without needing to change the function prototype.

---
 drivers/media/test-drivers/vidtv/vidtv_pes.c | 42 +++++++++-----------
 1 file changed, 18 insertions(+), 24 deletions(-)
drivers/media/test-drivers/vidtv/vidtv_pes.c +18 -24
index 34518ef042f7..de0ce5529d06 100644
@@ -176,37 +176,34 @@ static u32 vidtv_pes_write_h(struct pes_header_write_args args)
 	return nbytes;
 }
 
-static u32 vidtv_pes_write_stuffing(void *dest_buf,
-				    u32 dest_offset,
-				    u32 n_stuffing_bytes,
-				    u32 buf_sz)
+static u32 vidtv_pes_write_stuffing(struct pes_ts_header_write_args *args,
+				    u32 dest_offset)
 {
 	u32 nbytes = 0;
 	struct vidtv_mpeg_ts_adaption ts_adap = {};
-	u32 stuff_nbytes = 0;
+	u32 stuff_nbytes;
 
-	if (!n_stuffing_bytes)
+	if (!args->n_stuffing_bytes)
 		goto out;
 
-	if (n_stuffing_bytes > PES_TS_HEADER_MAX_STUFFING_BYTES) {
+	if (args->n_stuffing_bytes > PES_TS_HEADER_MAX_STUFFING_BYTES) {
 		pr_warn_ratelimited("More than %d stuffing bytes for a PES packet!\n",
 				    PES_TS_HEADER_MAX_STUFFING_BYTES);
 
-		n_stuffing_bytes = PES_TS_HEADER_MAX_STUFFING_BYTES;
+		args->n_stuffing_bytes = PES_TS_HEADER_MAX_STUFFING_BYTES;
 	}
 
 	/* the AF will only be its 'length' field with a value of zero */
-	if (n_stuffing_bytes == 1) {
-		nbytes += vidtv_memset(dest_buf,
+	if (args->n_stuffing_bytes == 1) {
+		nbytes += vidtv_memset(args->dest_buf,
 				       dest_offset + nbytes,
-				       buf_sz,
+				       args->dest_buf_sz,
 				       0,
-				       n_stuffing_bytes);
-
+				       args->n_stuffing_bytes);
 		goto out;
 	}
 
-	stuff_nbytes = n_stuffing_bytes - sizeof(ts_adap);
+	stuff_nbytes = args->n_stuffing_bytes - sizeof(ts_adap);
 
 	/* length _immediately_ following 'adaptation_field_length' */
 	ts_adap.length = sizeof(ts_adap) -
@@ -214,24 +211,24 @@ static u32 vidtv_pes_write_stuffing(void *dest_buf,
 			 stuff_nbytes;
 
 	/* write the adap after the TS header */
-	nbytes += vidtv_memcpy(dest_buf,
+	nbytes += vidtv_memcpy(args->dest_buf,
 			       dest_offset + nbytes,
-			       buf_sz,
+			       args->dest_buf_sz,
 			       &ts_adap,
 			       sizeof(ts_adap));
 
 	/* write the stuffing bytes */
-	nbytes += vidtv_memset(dest_buf,
+	nbytes += vidtv_memset(args->dest_buf,
 			       dest_offset + nbytes,
-			       buf_sz,
+			       args->dest_buf_sz,
 			       TS_FILL_BYTE,
 			       stuff_nbytes);
 
 out:
-	if (nbytes != n_stuffing_bytes)
+	if (nbytes != args->n_stuffing_bytes)
 		pr_warn_ratelimited("write size was %d, expected %d\n",
 				    nbytes,
-				    n_stuffing_bytes);
+				    args->n_stuffing_bytes);
 
 	return nbytes;
 }
@@ -261,10 +258,7 @@ static u32 vidtv_pes_write_ts_h(struct pes_ts_header_write_args args)
 			       sizeof(ts_header));
 
 	/* write stuffing, if any */
-	nbytes += vidtv_pes_write_stuffing(args.dest_buf,
-					   args.dest_offset + nbytes,
-					   args.n_stuffing_bytes,
-					   args.dest_buf_sz);
+	nbytes += vidtv_pes_write_stuffing(&args, args.dest_offset + nbytes);
 
 	return nbytes;
 }
Signature · 1 line
2.26.2
#8 lore raw

PATCH 8/9 media: vidtv: rewrite the adaption field logic

In reply to Mauro Carvalho Chehab

When compared with a stream generated via ffmpeg, it can be noticed that the PES doesn't contain any PCR info. That could cause problems with userspace decoding. So, rewrite the logic that fills the adaptation info, in order to allow it to add PCR frames without breaking frame alignment.

---
 drivers/media/test-drivers/vidtv/vidtv_mux.c |   1 +
 drivers/media/test-drivers/vidtv/vidtv_pes.c | 198 ++++++++++++-------
 drivers/media/test-drivers/vidtv/vidtv_pes.h |   2 +
 3 files changed, 128 insertions(+), 73 deletions(-)
drivers/media/test-drivers/vidtv/vidtv_mux.c +1 -0
index 6a3123c51cc4..082740ae9d44 100644
@@ -254,6 +254,7 @@ static u32 vidtv_mux_packetize_access_units(struct vidtv_mux *m,
 		args.access_unit_len = au->nbytes;
 		args.dest_offset     = m->mux_buf_offset;
 		args.pts             = au->pts;
+		args.pcr	     = m->timing.clk;
 
 		m->mux_buf_offset += vidtv_pes_write_into(args);
 
drivers/media/test-drivers/vidtv/vidtv_pes.c +125 -73

295 lines not on this page · raw message

drivers/media/test-drivers/vidtv/vidtv_pes.h +2 -0
index 84628780d8c6..0ea9e863024d 100644
@@ -125,6 +125,7 @@ struct pes_ts_header_write_args {
 	u8 *continuity_counter;
 	bool wrote_pes_header;
 	u32 n_stuffing_bytes;
+	u64 pcr;
 };
 
 /**
@@ -168,6 +169,7 @@ struct pes_write_args {
 	u64 dts;
 
 	u32 n_pes_h_s_bytes;
+	u64 pcr;
 };
 
 /**
Signature · 1 line
2.26.2
#9 lore raw

PATCH 9/9 media: vidtv: fix decoding with gstreamer and Vlc

In reply to Mauro Carvalho Chehab

Neither Vlc nor Gstreamer likes the PES_scrambling_control bits.

In the case of GST, this can be seen with:

	$ GST_DEBUG=2 LANG=C gst-play-1.0 pcm_audio.ts
	...
	0:00:00.097973439 12308 0x55f7ddd155e0 WARN               pesparser pesparse.c:411:mpegts_parse_pes_header: Wrong '0x10' marker before PES_scrambling_control (0x40)
	0:00:00.097987026 12308 0x55f7ddd155e0 WARN                 tsdemux tsdemux.c:2314:gst_ts_demux_parse_pes_header: Error parsing PES header. pid: 0x111 stream_type: 0x6
	...

So, change, it. After such change, the stream now plays fine with Vlc, Gstreamer, ffmpeg - and with programs that use such libraries, like Kaffeine.

---
 drivers/media/test-drivers/vidtv/vidtv_pes.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/media/test-drivers/vidtv/vidtv_pes.c +1 -1
index 57c7ed4fb3b5..2caad64349b3 100644
@@ -137,7 +137,7 @@ static u32 vidtv_pes_write_h(struct pes_header_write_args args)
 	struct vidtv_pes_optional pes_optional    = {};
 	struct pes_header_write_args pts_dts_args = args;
 	u32 stream_id = (args.encoder_id == S302M) ? PRIVATE_STREAM_1_ID : args.stream_id;
-	u16 pes_opt_bitfield = 0x2 << 13;
+	u16 pes_opt_bitfield = 0x01 << 15;
 
 	pes_header.bitfield = cpu_to_be32((PES_START_CODE_PREFIX << 8) | stream_id);
 
Signature · 1 line
2.26.2