Skip to contentSkip to conversation

leds: apu: extend support for PC Engines APU1 with newer firmware

PATCH Applied Andreas Eberlein 7 participants 10 messages Feb 16, 2021 – Mar 18, 2021 LED

lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 leds: apu: extend support for PC Engines APU1 with newer firmware

The DMI_PRODUCT_NAME entry on current firmware of PC Engines APU1 changed from "APU" to "apu1"

This modification adds the missing DMI data and thereby the LED support for the PC Engines APU1 with firmware versions >= 4.6.8.

---
Changes in v2:
- Add missing brace

---
 drivers/leds/leds-apu.c | 11 ++++++++++-
 1 file changed, 10 insertions(+), 1 deletion(-)
drivers/leds/leds-apu.c +10 -1
index 7fd557ace..c409b80c2 100644
@@ -83,6 +83,7 @@ static const struct apu_led_profile apu1_led_profile[] = {
 };
 
 static const struct dmi_system_id apu_led_dmi_table[] __initconst = {
+	/* PC Engines APU with factory bios "SageBios_PCEngines_APU-45" */
 	{
 		.ident = "apu",
 		.matches = {
@@ -90,6 +91,14 @@ static const struct dmi_system_id apu_led_dmi_table[] __initconst = {
 			DMI_MATCH(DMI_PRODUCT_NAME, "APU")
 		}
 	},
+	/* PC Engines APU with "Mainline" bios >= 4.6.8 */
+	{
+		.ident = "apu",
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "PC Engines"),
+			DMI_MATCH(DMI_PRODUCT_NAME, "apu1")
+		}
+	},
 	{}
 };
 MODULE_DEVICE_TABLE(dmi, apu_led_dmi_table);
@@ -173,7 +182,7 @@ static int __init apu_led_init(void)
 	int err;
 
 	if (!(dmi_match(DMI_SYS_VENDOR, "PC Engines") &&
-	      dmi_match(DMI_PRODUCT_NAME, "APU"))) {
+	      (dmi_match(DMI_PRODUCT_NAME, "APU") || dmi_match(DMI_PRODUCT_NAME, "apu1")))) {
 		pr_err("No PC Engines APUv1 board detected. For APUv2,3 support, enable CONFIG_PCENGINES_APU2\n");
 		return -ENODEV;
 	}
Signature · 1 line
2.25.1
#1 lore raw

On 16.02.21 14:30, Andreas Eberlein wrote:

Hi,

Quoted 5 lines

The DMI_PRODUCT_NAME entry on current firmware of PC Engines APU1 changed from "APU" to "apu1"

This modification adds the missing DMI data and thereby the LED support for the PC Engines APU1 with firmware versions >= 4.6.8.

Do you have a device for more intensive testing ?

In that case I'd like to suggest splitting the driver into gpio and gpio-based LED (using leds-gpio) - just like already I did for apu2/3/4. Maybe this even could also be moveed into the apu2 driver. This probably just makes sense if there're more gpio-connected devices than just LED)

Personally, I don't have access to the old apu1 board (IIRC not even produced anymore for several years), so I didn't dare to touch anything here.

Note that apu1 vs. apu2/3/4 have completely different SOC with different gpio logic - that was one of the reasons for writing a completely new driver for apu2+ from scrath, rather than extending the old one.

Quoted 27 lines
drivers/leds/leds-apu.c +9 -0
@@ -83,6 +83,7 @@ static const struct apu_led_profile apu1_led_profile[] = {
  };
 
  static const struct dmi_system_id apu_led_dmi_table[] __initconst = {
+	/* PC Engines APU with factory bios "SageBios_PCEngines_APU-45" */
  	{
  		.ident = "apu",
  		.matches = {
@@ -90,6 +91,14 @@ static const struct dmi_system_id apu_led_dmi_table[] __initconst = {
  			DMI_MATCH(DMI_PRODUCT_NAME, "APU")
  		}
  	},
+	/* PC Engines APU with "Mainline" bios >= 4.6.8 */
+	{
+		.ident = "apu",
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "PC Engines"),
+			DMI_MATCH(DMI_PRODUCT_NAME, "apu1")
+		}
+	},
  	{}
  };
  MODULE_DEVICE_TABLE(dmi, apu_led_dmi_table);

drivers/leds/leds-apu.c static int __init apu_led_init(void) · line 188

Quoted 8 lines
drivers/leds/leds-apu.c +1 -1
@@ -173,7 +182,7 @@ static int __init apu_led_init(void)
  	int err;
 
  	if (!(dmi_match(DMI_SYS_VENDOR, "PC Engines") &&
-	      dmi_match(DMI_PRODUCT_NAME, "APU"))) {
+	      (dmi_match(DMI_PRODUCT_NAME, "APU") || dmi_match(DMI_PRODUCT_NAME, "apu1")))) {
  		pr_err("No PC Engines APUv1 board detected. For APUv2,3 support, enable CONFIG_PCENGINES_APU2\n");
  		return -ENODEV;
  	}

Looks good to me. But don't dare giving official ack, since I don't have an apu1 board for testing.

Is Alan Mizrahi (original author) still here ?

--mtx

Signature · 8 lines
---
Hinweis: unverschlüsselte E-Mails können leicht abgehört und manipuliert
werden ! Für eine vertrauliche Kommunikation senden Sie bitte ihren
GPG/PGP-Schlüssel zu.
---
Enrico Weigelt, metux IT consult
Free software and Linux embedded engineering
[email protected] -- +49-151-27565287
Applied
#2 lore raw
Quoted 14 lines

drivers/leds/leds-apu.c static int __init apu_led_init(void) · line 188

Quoted 10 lines
  MODULE_DEVICE_TABLE(dmi, apu_led_dmi_table);
drivers/leds/leds-apu.c +1 -1
@@ -173,7 +182,7 @@ static int __init apu_led_init(void)
  	int err;
  	if (!(dmi_match(DMI_SYS_VENDOR, "PC Engines") &&
-	      dmi_match(DMI_PRODUCT_NAME, "APU"))) {
+	      (dmi_match(DMI_PRODUCT_NAME, "APU") || dmi_match(DMI_PRODUCT_NAME, "apu1")))) {
  		pr_err("No PC Engines APUv1 board detected. For APUv2,3 support, enable CONFIG_PCENGINES_APU2\n");
  		return -ENODEV;
  	}
 

Looks good to me. But don't dare giving official ack, since I don't have an apu1 board for testing.

Quoted 1 line

Is Alan Mizrahi (original author) still here ?

People are usually not that careful with thier acks.

Applied, thanks.

BR,
Pavel

Signature · 1 line
http://www.livejournal.com/~pavelmachek
#3 lore raw
In reply to Enrico Weigelt, metux IT consult

Enrico Weigelt, metux IT consult <[email protected]> [2021-02-17 12:02:47]:

Hi,

(adding Zbynek to Cc: who seems to own a complete range of APU boards and has done the runtime testing already)

Quoted 2 lines

Looks good to me. But don't dare giving official ack, since I don't have an apu1 board for testing.

@Zbynek can you perhaps provide your Tested-by: Your Name <your@email> tag as it seems, that you've tested[1] this extensively on several boards/versions already? I think, that you deserve credits as well. Thanks!

1. https://github.com/openwrt/openwrt/pull/3135#issuecomment-748953745

Cheers,

Petr

Tested-by
#4 lore raw

Petr Štetiar <[email protected]>

Hi Peter,

Thanks for adding to the discussion. I tested the proposed modification on APU1 with different versions of bios. The LED subsystem now behaves the same as the APU2 and higher. If it needs more tests on various boards
from PCengines, I'm available.

Zbyněk

#5 lore raw

On 19.02.21 21:51, Zbyněk Kocur wrote:

Hi Zbyněk,

Quoted 3 lines

Thanks for adding to the discussion. I tested the proposed modification on APU1 with different versions of bios. The LED subsystem now behaves the same as the APU2 and higher. If it needs more tests on various boards
from PCengines, I'm available.

Do you also happen to have different apu2/3/4 boards (various hw revs and bios versions) for testing ? I've still got some open issues, eg. regarding pcie reset lines, etc.

--mtx

Signature · 8 lines
---
Hinweis: unverschlüsselte E-Mails können leicht abgehört und manipuliert
werden ! Für eine vertrauliche Kommunikation senden Sie bitte ihren
GPG/PGP-Schlüssel zu.
---
Enrico Weigelt, metux IT consult
Free software and Linux embedded engineering
[email protected] -- +49-151-27565287
#6 lore raw

Hello,

I checked my stock and I have following models:

- apu2e2
- apu2e4
- apu3d4
- apu4c4

The bios in all models can be upgraded from version 4.0.6 to higher. The latest is 4.13.0.3.

Zbyněk
---
email: [email protected]
mobile: +420 728 946 803
phone: +420 224 354 054
web: https://comtel.fel.cvut.cz
Department of Telecommunications Engineering
Faculty of Electrical Engineering

Quoted 26 lines · metux IT consult

On 19.02.21 21:51, Zbyněk Kocur wrote:

Hi Zbyněk,

Quoted 3 lines

Thanks for adding to the discussion. I tested the proposed modification on APU1 with different versions of bios. The LED subsystem now behaves the same as the APU2 and higher. If it needs more tests on various boards
from PCengines, I'm available.

Do you also happen to have different apu2/3/4 boards (various hw revs and bios versions) for testing ? I've still got some open issues, eg. regarding pcie reset lines, etc.

--mtx

Signature · 8 lines
---
Hinweis: unverschlüsselte E-Mails können leicht abgehört und manipuliert
werden ! Für eine vertrauliche Kommunikation senden Sie bitte ihren
GPG/PGP-Schlüssel zu.
---
Enrico Weigelt, metux IT consult
Free software and Linux embedded engineering
[email protected] -- +49-151-27565287
#8 lore raw
In reply to Enrico Weigelt, metux IT consult

Hello,

Quoted 8 lines · metux IT consult
Quoted 5 lines · Andreas Eberlein

The DMI_PRODUCT_NAME entry on current firmware of PC Engines APU1 changed from "APU" to "apu1"

This modification adds the missing DMI data and thereby the LED support for the PC Engines APU1 with firmware versions >= 4.6.8.

Do you have a device for more intensive testing ?

I have an apu1d4 on my desk, which is supposed to be used soon, but available for testing at the moment. I put the latest coreboot bios version 4.13.0.3 on it yesterday.

Quoted 8 lines

In that case I'd like to suggest splitting the driver into gpio and gpio-based LED (using leds-gpio) - just like already I did for apu2/3/4. Maybe this even could also be moveed into the apu2 driver. This probably just makes sense if there're more gpio-connected devices than just LED)

Personally, I don't have access to the old apu1 board (IIRC not even produced anymore for several years), so I didn't dare to touch anything here.

If you give me a hint, which tree or patchset should be tested, and some hints what should be tested, I can try.

Quoted 3 lines

Note that apu1 vs. apu2/3/4 have completely different SOC with different gpio logic - that was one of the reasons for writing a completely new driver for apu2+ from scrath, rather than extending the old one.

Thanks for that work. I have to admit someone from the fli4l linux router distribution team also wrote LED and button drivers for the APU boards, but never managed to upstream those. :-/

If someone is interested, those are spread in our Subversion repository, but the apu drivers are here:

https://repo.nettworks.org/svn/fli4l/branches/4.0/trunk/src/packages/src/src/fli4l/hwsupp/pcengines-apu/

Personally, I'd rather have mainline drivers for all that boards. Don't know if it still makes sense for the older wrap or alix boards, though. I also have those lying around. ;-)

Greets
Alex

Signature · 4 lines
/"\ ASCII RIBBON | »With the first link, the chain is forged. The first
\ / CAMPAIGN     | speech censured, the first thought forbidden, the
 X  AGAINST      | first freedom denied, chains us all irrevocably.«
/ \ HTML MAIL    | (Jean-Luc Picard, quoting Judge Aaron Satie)
#9 lore raw

On 06.03.21 17:51, Alexander Dahl wrote:

Hi,

Quoted 2 lines

If you give me a hint, which tree or patchset should be tested, and some hints what should be tested, I can try.

haven't written anything for apu1 yet (as I dont have one), but I wrote the drivers for apu2/3/4.

My idea (which I never actually started on) was writing a separate gpio
driver (not LED) for the old Soc in apu1 and add instantiation w/ leds-gpio, keys, etc, into the pcengines-apu2 driver.

Quoted 8 lines

Thanks for that work. I have to admit someone from the fli4l linux router distribution team also wrote LED and button drivers for the APU boards, but never managed to upstream those. :-/

If someone is interested, those are spread in our Subversion repository, but the apu drivers are here:

https://repo.nettworks.org/svn/fli4l/branches/4.0/trunk/src/packages/src/src/fli4l/hwsupp/pcengines-apu/

hmm, maybe I could pick up pieces for the FCH functionality that's not supported yet (eg. wdt) ... not sure how much they differ between different SoC versions.

Quoted 3 lines

Personally, I'd rather have mainline drivers for all that boards. Don't know if it still makes sense for the older wrap or alix boards, though. I also have those lying around. ;-)

Well, somebody has to make his hands dirty, write those drivers, bring them to mainline, and maintain them.

--mtx

Signature · 9 lines
-- 
---
Hinweis: unverschlüsselte E-Mails können leicht abgehört und manipuliert
werden ! Für eine vertrauliche Kommunikation senden Sie bitte ihren
GPG/PGP-Schlüssel zu.
---
Enrico Weigelt, metux IT consult
Free software and Linux embedded engineering
[email protected] -- +49-151-27565287