Skip to contentSkip to conversation

Fix detection of 2nd fan on X1C9

PATCH In review Lyude Paul 3 participants 9 messages Apr 29, 2022 – May 2, 2022

Subsystems (2)
lore
Share & tools
lore

Lyude Paul

author
#0 lore raw

cover Fix detection of 2nd fan on X1C9

Some recent changes broke detection of the second fan on the X1 Carbon 9th generation, so here's some patches to fix it.

Lyude Paul (2):
platform/x86: thinkpad_acpi: Restore X1 Carbon 9th Gen dual fan quirk platform/x86: thinkpad_acpi: Don't probe 2nd fan if enabled by quirk

 drivers/platform/x86/thinkpad_acpi.c | 26 +++++++++++++-------------
 1 file changed, 13 insertions(+), 13 deletions(-)
Signature · 1 line
2.35.1
author
#1 lore raw

PATCH 1/2 platform/x86: thinkpad_acpi: Restore X1 Carbon 9th Gen dual fan quirk

The new method of probing dual fan support introduced in:

bf779aaf56ea ("platform/x86: thinkpad_acpi: Add dual fan probe")

While this commit says this works on the X1 Carbon 9th Gen, it actually just results in hiding the second fan on my local machine. Additionally, I'm fairly sure this machine powers on quite often without either of the two fans spinning.

So let's fix this by adding back the dual fan quirk for the X1 Carbon 9th Gen.

Fixes bf779aaf56ea ("platform/x86: thinkpad_acpi: Add dual fan probe")
Cc Mark Pearson <[email protected]>
Cc Hans de Goede <[email protected]>
Cc Henrique de Moraes Holschuh <[email protected]>
Cc Mark Gross <[email protected]>
---
 drivers/platform/x86/thinkpad_acpi.c | 1 +
 1 file changed, 1 insertion(+)
drivers/platform/x86/thinkpad_acpi.c +1 -0
index c568fae56db2..9067fd0a945c 100644
@@ -8699,6 +8699,7 @@ static const struct tpacpi_quirk fan_quirk_table[] __initconst = {
 	TPACPI_Q_LNV3('N', '2', 'E', TPACPI_FAN_2CTL),	/* P1 / X1 Extreme (1st gen) */
 	TPACPI_Q_LNV3('N', '2', 'O', TPACPI_FAN_2CTL),	/* P1 / X1 Extreme (2nd gen) */
 	TPACPI_Q_LNV3('N', '3', '0', TPACPI_FAN_2CTL),	/* P15 (1st gen) / P15v (1st gen) */
+	TPACPI_Q_LNV3('N', '3', '2', TPACPI_FAN_2CTL),  /* X1 Carbon (9th gen) */
 	TPACPI_Q_LNV3('N', '3', '7', TPACPI_FAN_2CTL),  /* T15g (2nd gen) */
 	TPACPI_Q_LNV3('N', '1', 'O', TPACPI_FAN_NOFAN),	/* X1 Tablet (2nd gen) */
 };
Signature · 1 line
2.35.1
#2 lore raw

Re: [External] [PATCH 1/2] platform/x86: thinkpad_acpi: Restore X1 Carbon 9th Gen dual fan quirk

Hi Lyude

Quoted 29 lines · Lyude Paul

The new method of probing dual fan support introduced in:

bf779aaf56ea ("platform/x86: thinkpad_acpi: Add dual fan probe")

While this commit says this works on the X1 Carbon 9th Gen, it actually just results in hiding the second fan on my local machine. Additionally, I'm fairly sure this machine powers on quite often without either of the two fans spinning.

So let's fix this by adding back the dual fan quirk for the X1 Carbon 9th Gen.

Fixes bf779aaf56ea ("platform/x86: thinkpad_acpi: Add dual fan probe")
Cc Mark Pearson <[email protected]>
Cc Hans de Goede <[email protected]>
Cc Henrique de Moraes Holschuh <[email protected]>
Cc Mark Gross <[email protected]>
---
 drivers/platform/x86/thinkpad_acpi.c | 1 +
 1 file changed, 1 insertion(+)
drivers/platform/x86/thinkpad_acpi.c +0 -0
index c568fae56db2..9067fd0a945c 100644

drivers/platform/x86/thinkpad_acpi.c static const struct tpacpi_quirk fan_quirk_table[] __initconst = { · line 8705

Quoted 7 lines
drivers/platform/x86/thinkpad_acpi.c +1 -0
@@ -8699,6 +8699,7 @@ static const struct tpacpi_quirk fan_quirk_table[] __initconst = {
 	TPACPI_Q_LNV3('N', '2', 'E', TPACPI_FAN_2CTL),	/* P1 / X1 Extreme (1st gen) */
 	TPACPI_Q_LNV3('N', '2', 'O', TPACPI_FAN_2CTL),	/* P1 / X1 Extreme (2nd gen) */
 	TPACPI_Q_LNV3('N', '3', '0', TPACPI_FAN_2CTL),	/* P15 (1st gen) / P15v (1st gen) */
+	TPACPI_Q_LNV3('N', '3', '2', TPACPI_FAN_2CTL),  /* X1 Carbon (9th gen) */
 	TPACPI_Q_LNV3('N', '3', '7', TPACPI_FAN_2CTL),  /* T15g (2nd gen) */
 	TPACPI_Q_LNV3('N', '1', 'O', TPACPI_FAN_NOFAN),	/* X1 Tablet (2nd gen) */
 };

I just double checked this on my X1C9 - and it's working correctly. 2nd fan is detected correctly.

I'd rather understand why it's not working on your setup then just re-introduce the quirk.

What happens on your system when the
res = fan2_get_speed(&speed);
is called? If that is failing it means your 2nd fan isn't responding and that's not supposed to happen. Could you let me know if you get an error code, if it happens every boot, etc
I assume when the function is called later it works successfully?

Also please confirm which BIOS and EC version you have.

Thanks
Mark

#3 lore raw

Re: [External] [PATCH 1/2] platform/x86: thinkpad_acpi: Restore X1 Carbon 9th Gen dual fan quirk

Hi Mark,

Quoted 23 lines · Mark Pearson

Hi Lyude

Quoted 4 lines · Lyude Paul

The new method of probing dual fan support introduced in:

bf779aaf56ea ("platform/x86: thinkpad_acpi: Add dual fan probe")

Quoted 8 lines

While this commit says this works on the X1 Carbon 9th Gen, it actually just results in hiding the second fan on my local machine. Additionally, I'm fairly sure this machine powers on quite often without either of the two fans spinning.

So let's fix this by adding back the dual fan quirk for the X1 Carbon 9th Gen.

[..]

I just double checked this on my X1C9 - and it's working correctly. 2nd fan is detected correctly.

I'd rather understand why it's not working on your setup then just re-introduce the quirk.

Quoted 6 lines

What happens on your system when the
res = fan2_get_speed(&speed);
is called? If that is failing it means your 2nd fan isn't responding and that's not supposed to happen. Could you let me know if you get an error code, if it happens every boot, etc
I assume when the function is called later it works successfully?

I have the same issue.

To me it looks like this:

Probing for the second fan calls fan2_get_speed(), this calls fan_select_fan2() which in turn checks that tp_features.second_fan is set. But at this point in the tp_features.second_fan can not yet be set because it is either set from quirks or *after* the probing.

Maybe some of the matches for the quirk TPACPI_FAN_2FAN should also have matched this device?
It doesn't do so on my device.

Quoted 1 line

Also please confirm which BIOS and EC version you have.

Linux: 5.17.5
BIOS Revision: 1.51
Firmware Revision: 1.32

Thomas

#4 lore raw

Re: [External] [PATCH 1/2] platform/x86: thinkpad_acpi: Restore X1 Carbon 9th Gen dual fan quirk

Quoted 46 lines · Thomas Weißschuh

Hi Mark,

Quoted 23 lines · Mark Pearson

Hi Lyude

Quoted 4 lines · Lyude Paul

The new method of probing dual fan support introduced in:

bf779aaf56ea ("platform/x86: thinkpad_acpi: Add dual fan probe")

Quoted 8 lines

While this commit says this works on the X1 Carbon 9th Gen, it actually just results in hiding the second fan on my local machine. Additionally, I'm fairly sure this machine powers on quite often without either of the two fans spinning.

So let's fix this by adding back the dual fan quirk for the X1 Carbon 9th Gen.

[..]

I just double checked this on my X1C9 - and it's working correctly. 2nd fan is detected correctly.

I'd rather understand why it's not working on your setup then just re-introduce the quirk.

Quoted 6 lines

What happens on your system when the
res = fan2_get_speed(&speed);
is called? If that is failing it means your 2nd fan isn't responding and that's not supposed to happen. Could you let me know if you get an error code, if it happens every boot, etc
I assume when the function is called later it works successfully?

I have the same issue.

To me it looks like this:

Probing for the second fan calls fan2_get_speed(), this calls fan_select_fan2() which in turn checks that tp_features.second_fan is set. But at this point in the tp_features.second_fan can not yet be set because it is either set from quirks or *after* the probing.

Maybe some of the matches for the quirk TPACPI_FAN_2FAN should also have matched this device?
It doesn't do so on my device.

Quoted 6 lines
Quoted 1 line

Also please confirm which BIOS and EC version you have.

Linux: 5.17.5
BIOS Revision: 1.51
Firmware Revision: 1.32

Thanks Thomas,
I'll go do some more digging on my system and see what I've missed. Mark

author
#5 lore raw

Re: [External] [PATCH 1/2] platform/x86: thinkpad_acpi: Restore X1 Carbon 9th Gen dual fan quirk

In reply to Mark Pearson

Some answers/comments down below

Quoted 50 lines · Mark Pearson

Hi Lyude

Quoted 36 lines · Lyude Paul

The new method of probing dual fan support introduced in:

bf779aaf56ea ("platform/x86: thinkpad_acpi: Add dual fan probe")

While this commit says this works on the X1 Carbon 9th Gen, it actually just results in hiding the second fan on my local machine. Additionally, I'm fairly sure this machine powers on quite often without either of the two fans spinning.

So let's fix this by adding back the dual fan quirk for the X1 Carbon 9th Gen.

Fixes bf779aaf56ea ("platform/x86: thinkpad_acpi: Add dual fan probe")
Cc Mark Pearson <[email protected]>
Cc Hans de Goede <[email protected]>
Cc Henrique de Moraes Holschuh <[email protected]>
Cc Mark Gross <[email protected]>

---
 drivers/platform/x86/thinkpad_acpi.c | 1 +

 1 file changed, 1 insertion(+)
drivers/platform/x86/thinkpad_acpi.c +0 -0
index c568fae56db2..9067fd0a945c 100644
@@ -8699,6 +8699,7 @@ static const struct tpacpi_quirk fan_quirk_table[]
 __initconst = {

        TPACPI_Q_LNV3('N', '2', 'E', TPACPI_FAN_2CTL),  /* P1 / X1 Extreme (1st gen) */
        TPACPI_Q_LNV3('N', '2', 'O', TPACPI_FAN_2CTL),  /* P1 / X1 Extreme (2nd gen) */
        TPACPI_Q_LNV3('N', '3', '0', TPACPI_FAN_2CTL),  /* P15 (1st gen) /

Quoted 8 lines

P15v (1st gen) */
+       TPACPI_Q_LNV3('N', '3', '2', TPACPI_FAN_2CTL),  /* X1 Carbon (9th gen) */
        TPACPI_Q_LNV3('N', '3', '7', TPACPI_FAN_2CTL),  /* T15g (2nd gen) */
        TPACPI_Q_LNV3('N', '1', 'O', TPACPI_FAN_NOFAN), /* X1 Tablet (2nd gen) */
 };

I just double checked this on my X1C9 - and it's working correctly. 2nd fan is detected correctly.

Quoted 2 lines

I'd rather understand why it's not working on your setup then just re-introduce the quirk.

Of course! I figured as much, it's just easy to start conversations with a revert :P

Quoted 7 lines

What happens on your system when the
  res = fan2_get_speed(&speed);
is called? If that is failing it means your 2nd fan isn't responding and that's not supposed to happen. Could you let me know if you get an error code, if it happens every boot, etc
I assume when the function is called later it works successfully?

It definitely seems to happen every boot, not sure about the error code it returns. I will check and get you this info asap

Quoted 2 lines

Also please confirm which BIOS and EC version you have.

BIOS version N32ET75W (1.51) release date 12/02/2021, embedded controller is 0.1.32

Quoted 4 lines

Thanks
Mark

Signature · 3 lines
Cheers,
 Lyude Paul (she/her)
 Software Engineer at Red Hat
#6 lore raw

Re: [External] [PATCH 1/2] platform/x86: thinkpad_acpi: Restore X1 Carbon 9th Gen dual fan quirk

Quoted 67 lines · Lyude Paul

Some answers/comments down below

Quoted 50 lines · Mark Pearson

Hi Lyude

Quoted 36 lines · Lyude Paul

The new method of probing dual fan support introduced in:

bf779aaf56ea ("platform/x86: thinkpad_acpi: Add dual fan probe")

While this commit says this works on the X1 Carbon 9th Gen, it actually just results in hiding the second fan on my local machine. Additionally, I'm fairly sure this machine powers on quite often without either of the two fans spinning.

So let's fix this by adding back the dual fan quirk for the X1 Carbon 9th Gen.

Fixes bf779aaf56ea ("platform/x86: thinkpad_acpi: Add dual fan probe")
Cc Mark Pearson <[email protected]>
Cc Hans de Goede <[email protected]>
Cc Henrique de Moraes Holschuh <[email protected]>
Cc Mark Gross <[email protected]>

---
 drivers/platform/x86/thinkpad_acpi.c | 1 +

 1 file changed, 1 insertion(+)
drivers/platform/x86/thinkpad_acpi.c +0 -0
index c568fae56db2..9067fd0a945c 100644
@@ -8699,6 +8699,7 @@ static const struct tpacpi_quirk fan_quirk_table[]
 __initconst = {

        TPACPI_Q_LNV3('N', '2', 'E', TPACPI_FAN_2CTL),  /* P1 / X1 Extreme (1st gen) */
        TPACPI_Q_LNV3('N', '2', 'O', TPACPI_FAN_2CTL),  /* P1 / X1 Extreme (2nd gen) */
        TPACPI_Q_LNV3('N', '3', '0', TPACPI_FAN_2CTL),  /* P15 (1st gen) /

Quoted 8 lines

P15v (1st gen) */
+       TPACPI_Q_LNV3('N', '3', '2', TPACPI_FAN_2CTL),  /* X1 Carbon (9th gen) */
        TPACPI_Q_LNV3('N', '3', '7', TPACPI_FAN_2CTL),  /* T15g (2nd gen) */
        TPACPI_Q_LNV3('N', '1', 'O', TPACPI_FAN_NOFAN), /* X1 Tablet (2nd gen) */
 };

I just double checked this on my X1C9 - and it's working correctly. 2nd fan is detected correctly.

Quoted 2 lines

I'd rather understand why it's not working on your setup then just re-introduce the quirk.

Of course! I figured as much, it's just easy to start conversations with a revert :P

Quoted 7 lines

What happens on your system when the
  res = fan2_get_speed(&speed);
is called? If that is failing it means your 2nd fan isn't responding and that's not supposed to happen. Could you let me know if you get an error code, if it happens every boot, etc
I assume when the function is called later it works successfully?

Quoted 10 lines

It definitely seems to happen every boot, not sure about the error code it returns. I will check and get you this info asap

Quoted 2 lines

Also please confirm which BIOS and EC version you have.

BIOS version N32ET75W (1.51) release date 12/02/2021, embedded controller is 0.1.32

Thanks!

Along with Thomas' notes I think I've found the problem (though still bemused why I don't see the problem on my X1C9 and I tested on multiple platforms previously...so it is somewhat weird).

Working on a fix - will try and have that out for review later today or tomorrow.

Mark

#7 lore raw

Re: [External] [PATCH 1/2] platform/x86: thinkpad_acpi: Restore X1 Carbon 9th Gen dual fan quirk

Quoted 80 lines · Mark Pearson
Quoted 67 lines · Lyude Paul

Some answers/comments down below

Quoted 50 lines · Mark Pearson

Hi Lyude

Quoted 36 lines · Lyude Paul

The new method of probing dual fan support introduced in:

bf779aaf56ea ("platform/x86: thinkpad_acpi: Add dual fan probe")

While this commit says this works on the X1 Carbon 9th Gen, it actually just results in hiding the second fan on my local machine. Additionally, I'm fairly sure this machine powers on quite often without either of the two fans spinning.

So let's fix this by adding back the dual fan quirk for the X1 Carbon 9th Gen.

Fixes bf779aaf56ea ("platform/x86: thinkpad_acpi: Add dual fan probe")
Cc Mark Pearson <[email protected]>
Cc Hans de Goede <[email protected]>
Cc Henrique de Moraes Holschuh <[email protected]>
Cc Mark Gross <[email protected]>

---
 drivers/platform/x86/thinkpad_acpi.c | 1 +

 1 file changed, 1 insertion(+)
drivers/platform/x86/thinkpad_acpi.c +0 -0
index c568fae56db2..9067fd0a945c 100644
@@ -8699,6 +8699,7 @@ static const struct tpacpi_quirk fan_quirk_table[]
 __initconst = {

        TPACPI_Q_LNV3('N', '2', 'E', TPACPI_FAN_2CTL),  /* P1 / X1 Extreme (1st gen) */
        TPACPI_Q_LNV3('N', '2', 'O', TPACPI_FAN_2CTL),  /* P1 / X1 Extreme (2nd gen) */
        TPACPI_Q_LNV3('N', '3', '0', TPACPI_FAN_2CTL),  /* P15 (1st gen) /

Quoted 8 lines

P15v (1st gen) */
+       TPACPI_Q_LNV3('N', '3', '2', TPACPI_FAN_2CTL),  /* X1 Carbon (9th gen) */
        TPACPI_Q_LNV3('N', '3', '7', TPACPI_FAN_2CTL),  /* T15g (2nd gen) */
        TPACPI_Q_LNV3('N', '1', 'O', TPACPI_FAN_NOFAN), /* X1 Tablet (2nd gen) */
 };

I just double checked this on my X1C9 - and it's working correctly. 2nd fan is detected correctly.

Quoted 2 lines

I'd rather understand why it's not working on your setup then just re-introduce the quirk.

Of course! I figured as much, it's just easy to start conversations with a revert :P

Quoted 7 lines

What happens on your system when the
  res = fan2_get_speed(&speed);
is called? If that is failing it means your 2nd fan isn't responding and that's not supposed to happen. Could you let me know if you get an error code, if it happens every boot, etc
I assume when the function is called later it works successfully?

Quoted 10 lines

It definitely seems to happen every boot, not sure about the error code it returns. I will check and get you this info asap

Quoted 2 lines

Also please confirm which BIOS and EC version you have.

BIOS version N32ET75W (1.51) release date 12/02/2021, embedded controller is 0.1.32

Quoted 9 lines

Thanks!

Along with Thomas' notes I think I've found the problem (though still bemused why I don't see the problem on my X1C9 and I tested on multiple platforms previously...so it is somewhat weird).

Working on a fix - will try and have that out for review later today or tomorrow.

Not sure exactly what the etiquette here is for the mailing list but I just posted an updated patch "platform/x86: thinkpad_acpi: Correct dual fan probe" that I think addresses all the issues raised in this patch sequence.

Please let me know any feedback or concerns. And thanks for raising this!

Mark

author
#8 lore raw

PATCH 2/2 platform/x86: thinkpad_acpi: Don't probe 2nd fan if enabled by quirk

In reply to Lyude Paul

If we already know that the system in question has a quirk telling us that the system has a second fan, there's no purpose in probing the second fan - especially when probing the second fan may not work properly with systems relying on quirks.

Also, convert all of the conditionals here into a single group of if/else statements. This is because there's no situations where there's more then one quirk on a device.

Fixes bf779aaf56ea ("platform/x86: thinkpad_acpi: Add dual fan probe")
Cc Mark Pearson <[email protected]>
Cc Hans de Goede <[email protected]>
Cc Henrique de Moraes Holschuh <[email protected]>
Cc Mark Gross <[email protected]>
---
 drivers/platform/x86/thinkpad_acpi.c | 25 ++++++++++++-------------
 1 file changed, 12 insertions(+), 13 deletions(-)
drivers/platform/x86/thinkpad_acpi.c +12 -13
index 9067fd0a945c..677822b5d4b4 100644
@@ -8747,26 +8747,25 @@ static int __init fan_init(struct ibm_init_struct *iibm)
 			unsigned int speed;
 
 			fan_status_access_mode = TPACPI_FAN_RD_TPEC;
-			if (quirks & TPACPI_FAN_Q1)
+			if (quirks & TPACPI_FAN_Q1) {
 				fan_quirk1_setup();
-			if (quirks & TPACPI_FAN_2FAN) {
+			} else if (quirks & TPACPI_FAN_2FAN) {
 				tp_features.second_fan = 1;
 				pr_info("secondary fan support enabled\n");
-			}
-			if (quirks & TPACPI_FAN_2CTL) {
+			} else if (quirks & TPACPI_FAN_2CTL) {
 				tp_features.second_fan = 1;
 				tp_features.second_fan_ctl = 1;
 				pr_info("secondary fan control enabled\n");
+			} else {
+				/* Try and probe the 2nd fan */
+				res = fan2_get_speed(&speed);
+				if (res >= 0) {
+					/* It responded - so let's assume it's there */
+					tp_features.second_fan = 1;
+					tp_features.second_fan_ctl = 1;
+					pr_info("secondary fan control detected & enabled\n");
+				}
 			}
-			/* Try and probe the 2nd fan */
-			res = fan2_get_speed(&speed);
-			if (res >= 0) {
-				/* It responded - so let's assume it's there */
-				tp_features.second_fan = 1;
-				tp_features.second_fan_ctl = 1;
-				pr_info("secondary fan control detected & enabled\n");
-			}
-
 		} else {
 			pr_err("ThinkPad ACPI EC access misbehaving, fan status and control unavailable\n");
 			return -ENODEV;
Signature · 1 line
2.35.1