Skip to contentSkip to conversation

ASoC: topology: log all headers being parsed

PATCH In review Keyon Jie 2 participants 5 messages May 9, 2020 – May 12, 2020

Subsystems (2)
lore
Share & tools
lore

Keyon Jie

author
#0 lore raw

PATCH 1/1 ASoC: topology: log all headers being parsed

The check (tplg->pass == le32_to_cpu(hdr->type)) makes no sense as it is comparing two different enums, remove the check and log all headers that being parsed.

---
 sound/soc/soc-topology.c | 9 ++++-----
 1 file changed, 4 insertions(+), 5 deletions(-)
sound/soc/soc-topology.c +4 -5
index 49875978a1ce..58cf5a12af3f 100644
@@ -2685,11 +2685,10 @@ static int soc_valid_header(struct soc_tplg *tplg,
 		return -EINVAL;
 	}
 
-	if (tplg->pass == le32_to_cpu(hdr->type))
-		dev_dbg(tplg->dev,
-			"ASoC: Got 0x%x bytes of type %d version %d vendor %d at pass %d\n",
-			hdr->payload_size, hdr->type, hdr->version,
-			hdr->vendor_type, tplg->pass);
+	dev_dbg(tplg->dev,
+		"ASoC: Got 0x%x bytes of type %d version %d vendor %d at pass %d\n",
+		hdr->payload_size, hdr->type, hdr->version,
+		hdr->vendor_type, tplg->pass);
 
 	return 1;
 }
Signature · 1 line
2.20.1
#1 lore raw

Kind reminder to always CC: maintainers....

Quoted 3 lines · Keyon Jie

The check (tplg->pass == le32_to_cpu(hdr->type)) makes no sense as it is comparing two different enums, remove the check and log all headers that being parsed.

It's true that the TYPE_MIXER should probably not be compared to TPLG_PASS_MANIFEST, but one would think that e.g. the TYPE_MIXER is handled in the TPLG_PASS_MIXER, no? and likewise that in the TPLG_PASS_MANIFEST all TPLG_TYPE_MANIFEST are handled?

Shouldn't you add a switch case to reconcile the two lists instead of removing the test altogether?

#define SND_SOC_TPLG_TYPE_MIXER 1
#define SND_SOC_TPLG_TYPE_BYTES 2
#define SND_SOC_TPLG_TYPE_ENUM 3
#define SND_SOC_TPLG_TYPE_DAPM_GRAPH 4
#define SND_SOC_TPLG_TYPE_DAPM_WIDGET 5
#define SND_SOC_TPLG_TYPE_DAI_LINK 6
#define SND_SOC_TPLG_TYPE_PCM 7
#define SND_SOC_TPLG_TYPE_MANIFEST 8
#define SND_SOC_TPLG_TYPE_CODEC_LINK 9
#define SND_SOC_TPLG_TYPE_BACKEND_LINK 10
#define SND_SOC_TPLG_TYPE_PDATA 11
#define SND_SOC_TPLG_TYPE_DAI 12

#define SOC_TPLG_PASS_MANIFEST 0
#define SOC_TPLG_PASS_VENDOR 1
#define SOC_TPLG_PASS_MIXER 2
#define SOC_TPLG_PASS_WIDGET 3
#define SOC_TPLG_PASS_PCM_DAI 4
#define SOC_TPLG_PASS_GRAPH 5
#define SOC_TPLG_PASS_PINS 6
#define SOC_TPLG_PASS_BE_DAI 7
#define SOC_TPLG_PASS_LINK 8

Quoted 27 lines
---
  sound/soc/soc-topology.c | 9 ++++-----
  1 file changed, 4 insertions(+), 5 deletions(-)
sound/soc/soc-topology.c +4 -5
index 49875978a1ce..58cf5a12af3f 100644
@@ -2685,11 +2685,10 @@ static int soc_valid_header(struct soc_tplg *tplg,
  		return -EINVAL;
  	}
 
-	if (tplg->pass == le32_to_cpu(hdr->type))
-		dev_dbg(tplg->dev,
-			"ASoC: Got 0x%x bytes of type %d version %d vendor %d at pass %d\n",
-			hdr->payload_size, hdr->type, hdr->version,
-			hdr->vendor_type, tplg->pass);
+	dev_dbg(tplg->dev,
+		"ASoC: Got 0x%x bytes of type %d version %d vendor %d at pass %d\n",
+		hdr->payload_size, hdr->type, hdr->version,
+		hdr->vendor_type, tplg->pass);
 
  	return 1;
  }
author
#2 lore raw
Quoted 1 line · Pierre-Louis Bossart

Kind reminder to always CC: maintainers....

Thanks for reminding.

Quoted 7 lines · Keyon Jie
Quoted 3 lines

The check (tplg->pass == le32_to_cpu(hdr->type)) makes no sense as it is comparing two different enums, remove the check and log all headers that being parsed.

It's true that the TYPE_MIXER should probably not be compared to TPLG_PASS_MANIFEST, but one would think that e.g. the TYPE_MIXER is handled in the TPLG_PASS_MIXER, no? and likewise that in the TPLG_PASS_MANIFEST all TPLG_TYPE_MANIFEST are handled?

Yes, all TYPE_MIXER/ENUM/BYTES are handled in the TPLG_PASS_MIXER, and in the TPLG_PASS_MANIFEST all TPLG_TYPE_MANIFEST are handled, but all these are guaranteed and more details are logged inside the corresponding xxx_load(), at here it is only checking the validity of the header, I think we can log all headers for debug purpose? Imagine that there might be some headers won't be parsed by any xxx_load().

Quoted 2 lines

Shouldn't you add a switch case to reconcile the two lists instead of removing the test altogether?

I can add a function to reconcile the two list, but since we already have "tplg->pass != SOC_TPLG_PASS_xxx" check in each xxx_load(), do you suggest to moving the logging inside these xxx_load() to reduce the redundant check?

Thanks,
~Keyon

Quoted 53 lines

#define SND_SOC_TPLG_TYPE_MIXER        1
#define SND_SOC_TPLG_TYPE_BYTES        2
#define SND_SOC_TPLG_TYPE_ENUM        3
#define SND_SOC_TPLG_TYPE_DAPM_GRAPH    4
#define SND_SOC_TPLG_TYPE_DAPM_WIDGET    5
#define SND_SOC_TPLG_TYPE_DAI_LINK    6
#define SND_SOC_TPLG_TYPE_PCM        7
#define SND_SOC_TPLG_TYPE_MANIFEST    8
#define SND_SOC_TPLG_TYPE_CODEC_LINK    9
#define SND_SOC_TPLG_TYPE_BACKEND_LINK    10
#define SND_SOC_TPLG_TYPE_PDATA        11
#define SND_SOC_TPLG_TYPE_DAI        12

#define SOC_TPLG_PASS_MANIFEST        0
#define SOC_TPLG_PASS_VENDOR        1
#define SOC_TPLG_PASS_MIXER        2
#define SOC_TPLG_PASS_WIDGET        3
#define SOC_TPLG_PASS_PCM_DAI        4
#define SOC_TPLG_PASS_GRAPH        5
#define SOC_TPLG_PASS_PINS        6
#define SOC_TPLG_PASS_BE_DAI        7
#define SOC_TPLG_PASS_LINK        8

Quoted 28 lines

---
  sound/soc/soc-topology.c | 9 ++++-----

  1 file changed, 4 insertions(+), 5 deletions(-)
sound/soc/soc-topology.c +4 -5
index 49875978a1ce..58cf5a12af3f 100644
@@ -2685,11 +2685,10 @@ static int soc_valid_header(struct soc_tplg 
 *tplg,
           return -EINVAL;
       }
-    if (tplg->pass == le32_to_cpu(hdr->type))
-        dev_dbg(tplg->dev,
-            "ASoC: Got 0x%x bytes of type %d version %d vendor %d at 
 pass %d\n",
-            hdr->payload_size, hdr->type, hdr->version,
-            hdr->vendor_type, tplg->pass);
+    dev_dbg(tplg->dev,
+        "ASoC: Got 0x%x bytes of type %d version %d vendor %d at pass 
 %d\n",
+        hdr->payload_size, hdr->type, hdr->version,
+        hdr->vendor_type, tplg->pass);
       return 1;

  }

#3 lore raw
Quoted 3 lines

I can add a function to reconcile the two list, but since we already have "tplg->pass != SOC_TPLG_PASS_xxx" check in each xxx_load(), do you suggest to moving the logging inside these xxx_load() to reduce the redundant check?

yes, you could add a static inline sof_tplg_log_hdr(hdr); after each pass check, sounds good to me.

author
#4 lore raw
Quoted 5 lines · Pierre-Louis Bossart
Quoted 3 lines

I can add a function to reconcile the two list, but since we already have "tplg->pass != SOC_TPLG_PASS_xxx" check in each xxx_load(), do you suggest to moving the logging inside these xxx_load() to reduce the redundant check?

yes, you could add a static inline sof_tplg_log_hdr(hdr); after each pass check, sounds good to me.

OK, let me drop this and will send a new one soon.

Thanks,
~Keyon