Skip to contentSkip to conversation

No network for SolidRun LX2160A Clearfog CX with kernel 6.12.113

DISCUSSION Oliver Smith 4 participants 11 messages Oct 30 – Nov 9 Arm i.MX and Layerscape

lore
Share & tools
lore
#0 lore raw

Hi,

on a SolidRun LX2160A Clearfog CX machine I've upgraded from 6.12.101 to 6.12.113 in Debian 13, and with that kernel there is no eth0 after boot. When downgrading to 6.12.101, it is there again.

I suspect this is a regression from dts changes in arch/arm64/boot/dts/freescale as there were a few for lx2160a:

* https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/log/arch/arm64/boot/dts/freescale?h=v6.12.113
* https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/log/arch/arm64/boot/dts/freescale?h=v6.12.101

[email protected] is listed in MAINTAINERS for that path and Vincent authored the related patches, therefore sending this to you. If this is the wrong place then let me know where to report this instead.

Thank you!

Best regards,
Oliver

Signature · 7 lines
- Oliver Smith <[email protected]>            https://www.sysmocom.de/
=======================================================================
* sysmocom - systems for mobile communications GmbH
* Siemensstr. 26a
* 10551 Berlin, Germany
* Sitz / Registered office: Berlin, HRB 134158 B
* Geschaeftsfuehrer / Managing Director: Harald Welte
reviewer
#1 lore raw

Hi Oliver,

Quoted 9 lines · Oliver Smith

Hi,

on a SolidRun LX2160A Clearfog CX machine I've upgraded from 6.12.101 to 6.12.113 in Debian 13, and with that kernel there is no eth0 after boot. When downgrading to 6.12.101, it is there again.

I suspect this is a regression from dts changes in arch/arm64/boot/dts/freescale as there were a few for lx2160a:

Run git bisect between 6.12.101 and 6.12.113 to find the exact commit that caused the regression.

#2 lore raw

Hi Oliver,

Quoted 10 lines · Fabio Estevam

Hi Oliver,

Quoted 7 lines · Oliver Smith

Hi,

on a SolidRun LX2160A Clearfog CX machine I've upgraded from 6.12.101 to 6.12.113 in Debian 13, and with that kernel there is no eth0 after boot. When downgrading to 6.12.101, it is there again.

Quoted 5 lines
Quoted 2 lines

I suspect this is a regression from dts changes in arch/arm64/boot/dts/freescale as there were a few for lx2160a:

Run git bisect between 6.12.101 and 6.12.113 to find the exact commit that caused the regression.

then, on the board, check, analyze and provide the outputs if needed of:

uname -r; cat /proc/device-tree/model; echo
ip -br link
dmesg | grep -i -e fsl-mc -e dpaa2 -e dpmac -e dprc cat /sys/kernel/debug/devices_deferred
ls /sys/bus/fsl-mc/devices/
busybox devmem 0x08340008 32 # or equivalent for your Debian

From U-Boot, before it boots Linux:

md.l 0x08340008 1
printenv mcinitcmd bootcmd

0x08340008 is the General Status Register (GSR) of the MC

my 2 cents,
Vincent

#3 lore raw

Hi Fabio,
Hi Vincent,

thank you for the replies!

Quoted 3 lines · Vincent Jardin
Quoted 2 lines · Fabio Estevam

Run git bisect between 6.12.101 and 6.12.113 to find the exact commit that caused the regression.

Sorry, but this is not a development machine and I don't have physical access to it, only a serial console. I can run some checks, but building multiple kernels for bisect and booting them sounds too risky and time intense unfortunately.

Quoted 8 lines · Vincent Jardin

then, on the board, check, analyze and provide the outputs if needed of:

uname -r; cat /proc/device-tree/model; echo
ip -br link
dmesg | grep -i -e fsl-mc -e dpaa2 -e dpmac -e dprc cat /sys/kernel/debug/devices_deferred
ls /sys/bus/fsl-mc/devices/
busybox devmem 0x08340008 32 # or equivalent for your Debian

Old kernel:

Quoted 14 lines
root@lx2-1:~# uname -r; cat /proc/device-tree/model; echo
6.12.101+deb13-arm64
SolidRun LX2160A Clearfog CX
root@lx2-1:~# ip -br link
lo               UNKNOWN        00:00:00:00:00:00 <LOOPBACK,UP,LOWER_UP>
eth0             UP             d0:63:b4:03:b4:f7 <BROADCAST,MULTICAST,UP,LOWER_UP>
br0              UP             52:a0:18:ae:7c:34 <BROADCAST,MULTICAST,UP,LOWER_UP>
lxcbr0           DOWN           10:66:6a:00:00:00 <NO-CARRIER,BROADCAST,MULTICAST,UP>
vethhEvvMx@if2   UP             fe:19:6b:c5:90:0e <BROADCAST,MULTICAST,UP,LOWER_UP>
veth5HPqsX@if2   UP             fe:de:f1:c5:ae:3a <BROADCAST,MULTICAST,UP,LOWER_UP>
vethFfnWAR@if2   UP             fe:ec:61:dc:5a:aa <BROADCAST,MULTICAST,UP,LOWER_UP>
root@lx2-1:~# dmesg | grep -i -e fsl-mc -e dpaa2 -e dpmac -e dprc
root@lx2-1:~# cat /sys/kernel/debug/devices_deferred
root@lx2-1:~# ls /sys/bus/fsl-mc/devices/
Quoted 8 lines
dpbp.0	  dpcon.12  dpcon.18  dpcon.23	dpcon.29  dpcon.6  dpio.10  dpio.2  dpio.8    dpmac.5  dpmcp.10  dpmcp.16  dpmcp.21  dpmcp.27  dpmcp.32  dpmcp.6  dprtc.0
dpbp.1	  dpcon.13  dpcon.19  dpcon.24	dpcon.3   dpcon.7  dpio.11  dpio.3  dpio.9    dpmac.6  dpmcp.11  dpmcp.17  dpmcp.22  dpmcp.28  dpmcp.33  dpmcp.7  dpseci.0
dpcon.0   dpcon.14  dpcon.2   dpcon.25	dpcon.30  dpcon.8  dpio.12  dpio.4  dpmac.10  dpmac.7  dpmcp.12  dpmcp.18  dpmcp.23  dpmcp.29  dpmcp.34  dpmcp.8
dpcon.1   dpcon.15  dpcon.20  dpcon.26	dpcon.31  dpcon.9  dpio.13  dpio.5  dpmac.17  dpmac.8  dpmcp.13  dpmcp.19  dpmcp.24  dpmcp.3   dpmcp.35  dpmcp.9
dpcon.10  dpcon.16  dpcon.21  dpcon.27	dpcon.4   dpio.0   dpio.14  dpio.6  dpmac.3   dpmac.9  dpmcp.14  dpmcp.2   dpmcp.25  dpmcp.30  dpmcp.4	 dpni.0
dpcon.11  dpcon.17  dpcon.22  dpcon.28	dpcon.5   dpio.1   dpio.15  dpio.7  dpmac.4   dpmcp.1  dpmcp.15  dpmcp.20  dpmcp.26  dpmcp.31  dpmcp.5	 dprc.1
root@lx2-1:~# busybox devmem 0x08340008 32
0x0000003F

New kernel:

Quoted 19 lines
root@lx2-1:~# uname -r; cat /proc/device-tree/model; echo
6.12.113+deb13-arm64
SolidRun LX2160A Clearfog CX
root@lx2-1:~# ip -br link
lo               UNKNOWN        00:00:00:00:00:00 <LOOPBACK,UP,LOWER_UP>
br0              UP             52:a0:18:ae:7c:34 <BROADCAST,MULTICAST,UP,LOWER_UP>
lxcbr0           DOWN           10:66:6a:00:00:00 <NO-CARRIER,BROADCAST,MULTICAST,UP>
vethXFQUqu@if2   UP             fe:0a:7b:88:52:cd <BROADCAST,MULTICAST,UP,LOWER_UP>
vethUG8N9y@if2   UP             fe:ff:eb:09:16:a5 <BROADCAST,MULTICAST,UP,LOWER_UP>
vethoesNCY@if2   UP             fe:8e:29:91:43:6b <BROADCAST,MULTICAST,UP,LOWER_UP>
root@lx2-1:~# dmesg | grep -i -e fsl-mc -e dpaa2 -e dpmac -e dprc
[    0.008101] fsl-mc MSI: msi-controller@6020000 domain created
[    0.107175] fsl_mc_bus 80c000000.fsl-mc: fsl-mc: MC boot completed with error 0x3f
[    0.114747] fsl_mc_bus 80c000000.fsl-mc: probe with driver fsl_mc_bus failed with error -22
root@lx2-1:~# cat /sys/kernel/debug/devices_deferred
8000000.crypto
root@lx2-1:~# ls /sys/bus/fsl-mc/devices/
root@lx2-1:~# busybox devmem 0x08340008 32
0x0000003F
Quoted 7 lines · Vincent Jardin

From U-Boot, before it boots Linux:

md.l 0x08340008 1
printenv mcinitcmd bootcmd

0x08340008 is the General Status Register (GSR) of the MC

uboot:

Quoted 6 lines

=> md.l 0x08340008 1
08340008: 00000001 ....
=> printenv mcinitcmd bootcmd
mcinitcmd=mmc dev 1; mmc read 0x80a00000 0x5000 0x1200;mmc read 0x80e00000 0x7000 0x800;env exists secureboot && mmc read 0x80640000 0x3200 0x20 && mmc read 0x80680000 0x3400 0x20 && esbc_validate 0x80640000 && esbc_validate 0x80680000 ;fsl_mc start mc 0x80a00000 0x80e00000 bootcmd=mmc dev 1; env exists mcinitcmd && mmcinfo; mmc read 0x80d00000 0x6800 0x800; env exists mcinitcmd && env exists secureboot && mmc read 0x806C0000 0x3600 0x20 && esbc_validate 0x806C0000;env exists mcinitcmd && fsl_mc lazyapply dpl 0x80d00000;run distro_bootcmd;run sd2_bootcmd;env exists secureboot && esbc_halt; =>

Best regards,
Oliver

Signature · 7 lines
- Oliver Smith <[email protected]>            https://www.sysmocom.de/
=======================================================================
* sysmocom - systems for mobile communications GmbH
* Siemensstr. 26a
* 10551 Berlin, Germany
* Sitz / Registered office: Berlin, HRB 134158 B
* Geschaeftsfuehrer / Managing Director: Harald Welte
reviewer
#4 lore raw

Hi Oliver,

Quoted 4 lines · Oliver Smith

Sorry, but this is not a development machine and I don't have physical access to it, only a serial console. I can run some checks, but building multiple kernels for bisect and booting them sounds too risky and time intense unfortunately.

I suggest reverting the 6.12 commit below:

commit 9a884fefd537e6a54aaaaa6d180d1f03f42baa8e

Author: Ioana Ciornei <[email protected]>
Date:   Wed Apr 1 17:45:08 2026 +0300

bus: fsl-mc: wait for the MC firmware to complete its boot

[ Upstream commit 208858b1b48eba83d073542372329cf8ed606526 ]

There are use cases in which the Management Complex firmware boot process is started by the bootloader which does not wait for the boot to complete. This is mainly done in order to reduce the overall boot time of a DPAA2 based SoC.

In this kind of circumstance, the fsl-mc bus driver needs to make sure that the MC firmware boot process is finished before proceeding to the usual operations such as interrogating the firmware to gather all existent DPAA2 objects, creating the fsl-mc devices on the bus etc.

Add this kind of check early in the boot process of the fsl-mc bus and defer the probe in case the firmware is still in its boot process.

#5 lore raw
In reply to Oliver Smith

Hi Ioana,

Can you have a look ?

It seems to be related to the backport of:
https://lore.kernel.org/all/[email protected]/

I do not understand the backport flow of the kernel: it was not flagged as a fix so I do not understand how it landed into Oliver's 6.12 kernel.

best regards,
Vincent

Quoted 99 lines · Oliver Smith

Hi Fabio,
Hi Vincent,

thank you for the replies!

Quoted 3 lines · Vincent Jardin
Quoted 2 lines · Fabio Estevam

Run git bisect between 6.12.101 and 6.12.113 to find the exact commit that caused the regression.

Sorry, but this is not a development machine and I don't have physical access to it, only a serial console. I can run some checks, but building multiple kernels for bisect and booting them sounds too risky and time intense unfortunately.

Quoted 8 lines · Vincent Jardin

then, on the board, check, analyze and provide the outputs if needed of:

uname -r; cat /proc/device-tree/model; echo
ip -br link
dmesg | grep -i -e fsl-mc -e dpaa2 -e dpmac -e dprc cat /sys/kernel/debug/devices_deferred
ls /sys/bus/fsl-mc/devices/
busybox devmem 0x08340008 32 # or equivalent for your Debian

Old kernel:

Quoted 14 lines
root@lx2-1:~# uname -r; cat /proc/device-tree/model; echo
6.12.101+deb13-arm64
SolidRun LX2160A Clearfog CX
root@lx2-1:~# ip -br link
lo               UNKNOWN        00:00:00:00:00:00 <LOOPBACK,UP,LOWER_UP>
eth0             UP             d0:63:b4:03:b4:f7 <BROADCAST,MULTICAST,UP,LOWER_UP>
br0              UP             52:a0:18:ae:7c:34 <BROADCAST,MULTICAST,UP,LOWER_UP>
lxcbr0           DOWN           10:66:6a:00:00:00 <NO-CARRIER,BROADCAST,MULTICAST,UP>
vethhEvvMx@if2   UP             fe:19:6b:c5:90:0e <BROADCAST,MULTICAST,UP,LOWER_UP>
veth5HPqsX@if2   UP             fe:de:f1:c5:ae:3a <BROADCAST,MULTICAST,UP,LOWER_UP>
vethFfnWAR@if2   UP             fe:ec:61:dc:5a:aa <BROADCAST,MULTICAST,UP,LOWER_UP>
root@lx2-1:~# dmesg | grep -i -e fsl-mc -e dpaa2 -e dpmac -e dprc
root@lx2-1:~# cat /sys/kernel/debug/devices_deferred
root@lx2-1:~# ls /sys/bus/fsl-mc/devices/
Quoted 8 lines
dpbp.0	  dpcon.12  dpcon.18  dpcon.23	dpcon.29  dpcon.6  dpio.10  dpio.2  dpio.8    dpmac.5  dpmcp.10  dpmcp.16  dpmcp.21  dpmcp.27  dpmcp.32  dpmcp.6  dprtc.0
dpbp.1	  dpcon.13  dpcon.19  dpcon.24	dpcon.3   dpcon.7  dpio.11  dpio.3  dpio.9    dpmac.6  dpmcp.11  dpmcp.17  dpmcp.22  dpmcp.28  dpmcp.33  dpmcp.7  dpseci.0
dpcon.0   dpcon.14  dpcon.2   dpcon.25	dpcon.30  dpcon.8  dpio.12  dpio.4  dpmac.10  dpmac.7  dpmcp.12  dpmcp.18  dpmcp.23  dpmcp.29  dpmcp.34  dpmcp.8
dpcon.1   dpcon.15  dpcon.20  dpcon.26	dpcon.31  dpcon.9  dpio.13  dpio.5  dpmac.17  dpmac.8  dpmcp.13  dpmcp.19  dpmcp.24  dpmcp.3   dpmcp.35  dpmcp.9
dpcon.10  dpcon.16  dpcon.21  dpcon.27	dpcon.4   dpio.0   dpio.14  dpio.6  dpmac.3   dpmac.9  dpmcp.14  dpmcp.2   dpmcp.25  dpmcp.30  dpmcp.4	 dpni.0
dpcon.11  dpcon.17  dpcon.22  dpcon.28	dpcon.5   dpio.1   dpio.15  dpio.7  dpmac.4   dpmcp.1  dpmcp.15  dpmcp.20  dpmcp.26  dpmcp.31  dpmcp.5	 dprc.1
root@lx2-1:~# busybox devmem 0x08340008 32
0x0000003F

New kernel:

Quoted 19 lines
root@lx2-1:~# uname -r; cat /proc/device-tree/model; echo
6.12.113+deb13-arm64
SolidRun LX2160A Clearfog CX
root@lx2-1:~# ip -br link
lo               UNKNOWN        00:00:00:00:00:00 <LOOPBACK,UP,LOWER_UP>
br0              UP             52:a0:18:ae:7c:34 <BROADCAST,MULTICAST,UP,LOWER_UP>
lxcbr0           DOWN           10:66:6a:00:00:00 <NO-CARRIER,BROADCAST,MULTICAST,UP>
vethXFQUqu@if2   UP             fe:0a:7b:88:52:cd <BROADCAST,MULTICAST,UP,LOWER_UP>
vethUG8N9y@if2   UP             fe:ff:eb:09:16:a5 <BROADCAST,MULTICAST,UP,LOWER_UP>
vethoesNCY@if2   UP             fe:8e:29:91:43:6b <BROADCAST,MULTICAST,UP,LOWER_UP>
root@lx2-1:~# dmesg | grep -i -e fsl-mc -e dpaa2 -e dpmac -e dprc
[    0.008101] fsl-mc MSI: msi-controller@6020000 domain created
[    0.107175] fsl_mc_bus 80c000000.fsl-mc: fsl-mc: MC boot completed with error 0x3f
[    0.114747] fsl_mc_bus 80c000000.fsl-mc: probe with driver fsl_mc_bus failed with error -22
root@lx2-1:~# cat /sys/kernel/debug/devices_deferred
8000000.crypto
root@lx2-1:~# ls /sys/bus/fsl-mc/devices/
root@lx2-1:~# busybox devmem 0x08340008 32
0x0000003F
Quoted 7 lines · Vincent Jardin

From U-Boot, before it boots Linux:

md.l 0x08340008 1
printenv mcinitcmd bootcmd

0x08340008 is the General Status Register (GSR) of the MC

uboot:

Quoted 6 lines

=> md.l 0x08340008 1
08340008: 00000001 ....
=> printenv mcinitcmd bootcmd
mcinitcmd=mmc dev 1; mmc read 0x80a00000 0x5000 0x1200;mmc read 0x80e00000 0x7000 0x800;env exists secureboot && mmc read 0x80640000 0x3200 0x20 && mmc read 0x80680000 0x3400 0x20 && esbc_validate 0x80640000 && esbc_validate 0x80680000 ;fsl_mc start mc 0x80a00000 0x80e00000 bootcmd=mmc dev 1; env exists mcinitcmd && mmcinfo; mmc read 0x80d00000 0x6800 0x800; env exists mcinitcmd && env exists secureboot && mmc read 0x806C0000 0x3600 0x20 && esbc_validate 0x806C0000;env exists mcinitcmd && fsl_mc lazyapply dpl 0x80d00000;run distro_bootcmd;run sd2_bootcmd;env exists secureboot && esbc_halt; =>

Best regards,
Oliver

Signature · 7 lines
- Oliver Smith <[email protected]>            https://www.sysmocom.de/
=======================================================================
* sysmocom - systems for mobile communications GmbH
* Siemensstr. 26a
* 10551 Berlin, Germany
* Sitz / Registered office: Berlin, HRB 134158 B
* Geschaeftsfuehrer / Managing Director: Harald Welte
#6 lore raw
Quoted 1 line · Vincent Jardin

Hi Ioana,

Hi Vincent,

Quoted 5 lines

Can you have a look ?

It seems to be related to the backport of:
https://lore.kernel.org/all/[email protected]/

This patch should only prevent the fsl-mc bus to probe if the firmware ran into an error in its boot process or if it's still creating the DPAA2 objects requested through the DPL (layout) file.

Based on the GSR value, it seems that the system in question was running into a DPL processing error even before the update in Linux kernel:

	root@lx2-1:~# busybox devmem 0x08340008 32
	0x0000003F

The 0x3F value means "MC layout process error".

Please look into the MC firmware's log. It should help in determining what exactly is not ok in the DPL file processing.

	$ cat /dev/dpaa2_mc_console

The dpaa2_mc_console device file should be created even if the fsl-mc bus failed to probe. In case you encounter any problems from Linux prompt, the same log can be retrieved from the u-boot prompt:

=> fsl_mc dump_log

But the command above should be issued once the DPL file is applied, not before it.

Quoted 4 lines

I do not understand the backport flow of the kernel: it was not flagged as a fix so I do not understand how it landed into Oliver's 6.12 kernel.

There is a script / process called AUTOSEL which selects patches from the mainline tree to be backported to LTS kernels. This is happening even though there is no Fixes tag in the initial commit.

https://lwn.net/ml/all/aBj_SEgFTXfrPVuj@lappy/

Ioana

#7 lore raw

Hi Iona,
Hi Vincent,

Quoted 10 lines · Ioana Ciornei

On Sat, Oct 31, 2026 at 03:42:28PM +0100, Vincent Jardin wrote:
[…]

Quoted 2 lines

It seems to be related to the backport of:

   https://lore.kernel.org/all/[email protected]/

[…]> The 0x3F value means "MC layout process error".

Please look into the MC firmware's log. It should help in determining what exactly is not ok in the DPL file processing.

	$ cat /dev/dpaa2_mc_console

With both the old and the new kernel I get the exact same log:

Quoted 27 lines
root@lx2-1:~# cat /dev/dpaa2_mc_console
[W, PLATFORM]  UART print failed, all debug data will be printed to buffer.
[E, open_device:2371, DPMAC]  no device
[E, dpni_probe_cb_v1:4704, DPNI]  Can't create DPNI 0001
[E, subnode_process:155]  Probing module 'dpni' return error code -19. Continue dpl processing...
[E, open_device:2371, DPMAC]  no device
[E, dpni_probe_cb_v1:4704, DPNI]  Can't create DPNI 0002
[E, subnode_process:155]  Probing module 'dpni' return error code -19. Continue dpl processing...
[E, open_device:2371, DPMAC]  no device
[E, dpni_probe_cb_v1:4704, DPNI]  Can't create DPNI 0003
[E, subnode_process:155]  Probing module 'dpni' return error code -19. Continue dpl processing...
[E, open_device:2371, DPMAC]  no device
[E, dpni_probe_cb_v1:4704, DPNI]  Can't create DPNI 0004
[E, subnode_process:155]  Probing module 'dpni' return error code -19. Continue dpl processing...
[E, open_device:2371, DPMAC]  no device
[E, dpni_probe_cb_v1:4704, DPNI]  Can't create DPNI 0005
[E, subnode_process:155]  Probing module 'dpni' return error code -19. Continue dpl processing...
[E, open_device:2371, DPMAC]  no device
[E, dpni_probe_cb_v1:4704, DPNI]  Can't create DPNI 0006
[E, subnode_process:155]  Probing module 'dpni' return error code -19. Continue dpl processing...
[E, open_device:2371, DPMAC]  no device
[E, dpni_probe_cb_v1:4704, DPNI]  Can't create DPNI 0007
[E, subnode_process:155]  Probing module 'dpni' return error code -19. Continue dpl processing...
[E, open_device:2371, DPMAC]  no device
[E, dpni_probe_cb_v1:4704, DPNI]  Can't create DPNI 0008
[E, subnode_process:155]  Probing module 'dpni' return error code -19. Continue dpl processing...
[E, open_device:2371, DPMAC]  no device
Quoted 8 lines
[E, dpmac_probe_cb:766, DPMAC]  Can't create DPMAC 0002
[E, subnode_process:155]  Probing module 'dpmac' return error code -19. Continue dpl processing...
[E, open_device:2371, DPMAC]  no device
[E, dpmac_probe_cb:766, DPMAC]  Can't create DPMAC 0012
[E, subnode_process:155]  Probing module 'dpmac' return error code -19. Continue dpl processing...
[E, dpl_process:527]  Error while parsing 'connections'. Skip processing the rest of DPL.
[E, main:186]  DPL processing failed; continuing...
Running MC app, waiting for events ...

I'm not familiar with the MC firmware, but all this explains what happens, there's the firmware error and then because of the mentioned patch, fsl_mc_bus_probe() aborts early and then I get no ethernet device.

As mentioned earlier, dmesg also has:

Quoted 2 lines
Quoted 2 lines
[    0.107175] fsl_mc_bus 80c000000.fsl-mc: fsl-mc: MC boot completed with error 0x3f
[    0.114747] fsl_mc_bus 80c000000.fsl-mc: probe with driver fsl_mc_bus failed with error -22

Could you give me a pointer for getting rid of the firmware errors?

(And I do wonder how many other people may run into the same problem, maybe it would be better to just print a warning instead of aborting there...?)

Thanks for all the help everybody!

Best regards,
Oliver

Signature · 7 lines
- Oliver Smith <[email protected]>            https://www.sysmocom.de/
=======================================================================
* sysmocom - systems for mobile communications GmbH
* Siemensstr. 26a
* 10551 Berlin, Germany
* Sitz / Registered office: Berlin, HRB 134158 B
* Geschaeftsfuehrer / Managing Director: Harald Welte
#8 lore raw
Quoted 8 lines · Oliver Smith

Hi Iona,
Hi Vincent,

Quoted 4 lines · Ioana Ciornei

On Sat, Oct 31, 2026 at 03:42:28PM +0100, Vincent Jardin wrote:
[…]

Quoted 2 lines

It seems to be related to the backport of:

   https://lore.kernel.org/all/[email protected]/
Quoted 41 lines
Quoted 6 lines

[…]> The 0x3F value means "MC layout process error".

Please look into the MC firmware's log. It should help in determining what exactly is not ok in the DPL file processing.

	$ cat /dev/dpaa2_mc_console

With both the old and the new kernel I get the exact same log:

Quoted 33 lines
root@lx2-1:~# cat /dev/dpaa2_mc_console
[W, PLATFORM]  UART print failed, all debug data will be printed to buffer.
[E, open_device:2371, DPMAC]  no device
[E, dpni_probe_cb_v1:4704, DPNI]  Can't create DPNI 0001
[E, subnode_process:155]  Probing module 'dpni' return error code -19. Continue dpl processing...
[E, open_device:2371, DPMAC]  no device
[E, dpni_probe_cb_v1:4704, DPNI]  Can't create DPNI 0002
[E, subnode_process:155]  Probing module 'dpni' return error code -19. Continue dpl processing...
[E, open_device:2371, DPMAC]  no device
[E, dpni_probe_cb_v1:4704, DPNI]  Can't create DPNI 0003
[E, subnode_process:155]  Probing module 'dpni' return error code -19. Continue dpl processing...
[E, open_device:2371, DPMAC]  no device
[E, dpni_probe_cb_v1:4704, DPNI]  Can't create DPNI 0004
[E, subnode_process:155]  Probing module 'dpni' return error code -19. Continue dpl processing...
[E, open_device:2371, DPMAC]  no device
[E, dpni_probe_cb_v1:4704, DPNI]  Can't create DPNI 0005
[E, subnode_process:155]  Probing module 'dpni' return error code -19. Continue dpl processing...
[E, open_device:2371, DPMAC]  no device
[E, dpni_probe_cb_v1:4704, DPNI]  Can't create DPNI 0006
[E, subnode_process:155]  Probing module 'dpni' return error code -19. Continue dpl processing...
[E, open_device:2371, DPMAC]  no device
[E, dpni_probe_cb_v1:4704, DPNI]  Can't create DPNI 0007
[E, subnode_process:155]  Probing module 'dpni' return error code -19. Continue dpl processing...
[E, open_device:2371, DPMAC]  no device
[E, dpni_probe_cb_v1:4704, DPNI]  Can't create DPNI 0008
[E, subnode_process:155]  Probing module 'dpni' return error code -19. Continue dpl processing...
[E, open_device:2371, DPMAC]  no device
[E, dpmac_probe_cb:766, DPMAC]  Can't create DPMAC 0002
[E, subnode_process:155]  Probing module 'dpmac' return error code -19. Continue dpl processing...
[E, open_device:2371, DPMAC]  no device
[E, dpmac_probe_cb:766, DPMAC]  Can't create DPMAC 0012
[E, subnode_process:155]  Probing module 'dpmac' return error code -19. Continue dpl processing...
[E, dpl_process:527]  Error while parsing 'connections'. Skip processing the rest of DPL.

The above two "Can't create DPMAC" errors suggest that the DPL (Data Path Layout) file described some MACs which are not available on your board, meaning that the RCW (Reset Configuration Word) used does not instantiate those MACs.

Quoted 8 lines
Quoted 2 lines

[E, main:186] DPL processing failed; continuing... Running MC app, waiting for events ...

I'm not familiar with the MC firmware, but all this explains what happens, there's the firmware error and then because of the mentioned patch, fsl_mc_bus_probe() aborts early and then I get no ethernet device.

Quoted 5 lines

As mentioned earlier, dmesg also has:

Quoted 2 lines
Quoted 2 lines
[    0.107175] fsl_mc_bus 80c000000.fsl-mc: fsl-mc: MC boot completed with error 0x3f
[    0.114747] fsl_mc_bus 80c000000.fsl-mc: probe with driver fsl_mc_bus failed with error -22

Could you give me a pointer for getting rid of the firmware errors?

I need to have a look at the DPL file used on your board. Have you built it using Solidrun's lx2160a_build project (https://github.com/SolidRun/lx2160a_build)?
Can you get me a link to the source file?

Otherwise you can dump it from u-boot by using fdt print before the fsl_mc lazyapply command:

	=> fdt addr <addr> && fdt print && fsl_mc lazyapply dpl <addr>;

It would also help to know the RCW and the SerDes protocol combination that you are using. You can find those in the u-boot log. I pasted an example below:

	Reset Configuration Word (RCW):
	       00000000: 50838338 24500050 00000000 00000000
	       00000010: 00000000 0e010000 00000000 00000000
	       00000020: 0fc001a0 00002580 00000000 08000086
	       00000030: 09240000 00000001 00000000 00000000
	       00000040: 00000000 00000000 00000000 00000000
	       00000050: 00000000 00000000 00000000 00000000
	       00000060: 00000000 00000000 00027000 00000000
	       00000070: 08a70000 00150020
	Model: SolidRun LX2160ACEX7 COM express type 7 based board
	Board: LX2160ACE Rev2.0-CEX7, SD
	SERDES1 Reference: Clock1 = 161.13MHz Clock2 = 100MHz
	SERDES2 Reference: Clock1 = 100MHz Clock2 = 100MHz
	SERDES3 Reference: Clock1 = 100MHz Clock2 = 100Hz
	DRAM:  15.9 GiB
	DDR    15.9 GiB (DDR4, 64-bit, CL=22, ECC off)
	       DDR Controller Interleaving Mode: 256B
	WDT:   Started with servicing (30s timeout)
	Using SERDES1 Protocol: 7 (0x7)
	Using SERDES2 Protocol: 5 (0x5)
	Using SERDES3 Protocol: 2 (0x2)
Quoted 3 lines

(And I do wonder how many other people may run into the same problem, maybe it would be better to just print a warning instead of aborting there...?)

On the other hand, before the patch your system was basically limping and its state wasn't really consistent with the configuration files passed to the firmware.

Ioana

#9 lore raw

Hi Ioana,

Quoted 3 lines · Ioana Ciornei

I need to have a look at the DPL file used on your board. Have you built it using Solidrun's lx2160a_build project (https://github.com/SolidRun/lx2160a_build)?
Can you get me a link to the source file?

This image was used in 2022:
https://solid-run-images.sos-de-fra-1.exo.io/LX2k/lx2160a_build/lx2160acex7_2000_700_3200_8_5_2-f57e117.img.xz

It is now a dead link, but from the commit hash in the file, apparently it has been built from:

https://github.com/SolidRun/lx2160a_build/tree/f57e11707eb4aa5df1908ff64ba52b9f6141ab1f

Quoted 4 lines

Otherwise you can dump it from u-boot by using fdt print before the fsl_mc lazyapply command:

	=> fdt addr <addr> && fdt print && fsl_mc lazyapply dpl <addr>;

If the above is not useful: What address would I need to use here?

Quoted 3 lines

It would also help to know the RCW and the SerDes protocol combination that you are using. You can find those in the u-boot log. I pasted an example below:

Reset Configuration Word (RCW):

       00000000: 50838338 24500050 00000000 00000000
       00000010: 00000000 0e010000 00000000 00000000
       00000020: 10c001a0 00002580 00000000 00000000
       00000030: 09240000 00000000 00000000 00000000
       00000040: 00000000 00000000 00000000 00000000
       00000050: 00000000 00000000 00000000 00000000
       00000060: 00000000 00000000 00027000 00000000
       00000070: 08a80001 00151020

Model: SolidRun LX2160ACEX7 COM express type 7 based board
Board: LX2160ACE Rev2.0-CEX7, eMMC
SERDES1 Reference: Clock1 = 161.13MHz Clock2 = 100MHz SERDES2 Reference: Clock1 = 100MHz Clock2 = 100MHz SERDES3 Reference: Clock1 = 100MHz Clock2 = 100Hz
DRAM: 63.9 GiB
DDR 63.9 GiB (DDR4, 64-bit, CL=22, ECC off)
DDR Controller Interleaving Mode: 256B
DDR Chip-Select Interleaving Mode: CS0+CS1
dev_get_priv: null device
dev_get_priv: null device
Using SERDES1 Protocol: 8 (0x8)
Using SERDES2 Protocol: 5 (0x5)
Using SERDES3 Protocol: 2 (0x2)

Best regards,
Oliver

Signature · 7 lines
- Oliver Smith <[email protected]>            https://www.sysmocom.de/
=======================================================================
* sysmocom - systems for mobile communications GmbH
* Siemensstr. 26a
* 10551 Berlin, Germany
* Sitz / Registered office: Berlin, HRB 134158 B
* Geschaeftsfuehrer / Managing Director: Harald Welte
#10 lore raw
Quoted 7 lines · Oliver Smith

Hi Ioana,

Quoted 3 lines · Ioana Ciornei

I need to have a look at the DPL file used on your board. Have you built it using Solidrun's lx2160a_build project (https://github.com/SolidRun/lx2160a_build)?
Can you get me a link to the source file?

Quoted 8 lines

This image was used in 2022:
https://solid-run-images.sos-de-fra-1.exo.io/LX2k/lx2160a_build/lx2160acex7_2000_700_3200_8_5_2-f57e117.img.xz

It is now a dead link, but from the commit hash in the file, apparently it has been built from:

https://github.com/SolidRun/lx2160a_build/tree/f57e11707eb4aa5df1908ff64ba52b9f6141ab1f

This helps, thanks.

The problem is that the DPL file lists 1 DPNI object and multiple DPMAC objects [1]. The problem arises when the same DPL file tries to describe extra objects [2] and even connect these objects [3].

The newer versions of the same DPL file do not have this issue [4].

For a targetted fix, without the need to rewrite the entire firmware image, you would need to:
- clone mc-utils (https://github.com/nxp-qoriq/mc-utils.git)
- apply [4]
- build (make -C config/ )
- Update the DPL file from the board with the newly compiled one - config/lx2160a/LX2160A-CEX7/clearfog-cx-s1_8-s2_0-dpl.dtb.

Quoted 6 lines
Quoted 4 lines

Otherwise you can dump it from u-boot by using fdt print before the fsl_mc lazyapply command:

	=> fdt addr <addr> && fdt print && fsl_mc lazyapply dpl <addr>;

If the above is not useful: What address would I need to use here?

The address should be the one that's already used to keep the DPL file. I am sure there is already a 'fsl_mc lazyapply' command in the u-boot boot path which had the necessary address.

Ioana

[1] https://github.com/SolidRun/lx2160a_build/blob/f57e11707eb4aa5df1908ff64ba52b9f6141ab1f/patches/mc-utils/0001-lx2160acex7-add-8x10G-dual-40G-and-dual-100G-DPL-DPC.patch#L359
[2] https://github.com/SolidRun/lx2160a_build/blob/f57e11707eb4aa5df1908ff64ba52b9f6141ab1f/patches/mc-utils/0001-lx2160acex7-add-8x10G-dual-40G-and-dual-100G-DPL-DPC.patch#L420
[3] https://github.com/SolidRun/lx2160a_build/blob/f57e11707eb4aa5df1908ff64ba52b9f6141ab1f/patches/mc-utils/0001-lx2160acex7-add-8x10G-dual-40G-and-dual-100G-DPL-DPC.patch#L832
[4] https://github.com/SolidRun/lx2160a_build/blob/develop-ls-6.6.52-2.2.0/patches/mc-utils/0001-add-solidrun-lx2160-cex7-based-clearfog-cx-dpl-dpc.patch