pci-imx6 how to reset completely during reboot

[rrd]

Hello,

I have noticed that on our board, which runs on an i.MX8M Plus, since switching to Yocto Scarthgap, a “Phy Link never came up” error appears on the PCIe when I reboot the device. However, if I shut it down completely with poweroff, I don't have this problem.

Unless someone already has a working solution for me, I would try to adapt the pci-imx6, as I have seen that it has been fundamentally changed since Yocto Kirkstone. Since I don't know if a backport is possible, I would rather try to continue working with the current version. Hence the question: how can a warm start be implemented in the i.MX8M Plus in terms of software? Is there perhaps a bit that I may need to set so that the PCI goes into a complete reset? In that case, I could force a complete reset before initialization during testing.

Best regards,

Ralf

[pengyong_zhang]

Hi @rrd 

Could you please re-create a new case talk about your question? And Please describe the latest status. Since I am unable to reproduce your issue here, I will contact our internal team in a new case to continue supporting your issue. Thanks for your understanding.

B.R

[pengyong_zhang]

Hi @rrd 

As I mentioned before, I still cannot reproduce your issue on my i.MX8MP EVK board. With your changes, you can now reboot the PCIe device after running the reboot command, correct? If so, could you please share your code patch? Thanks.

B.R

[rrd]

Hi@pengyong_zhang

 Yes, I will provide you with the patch later. After patching, I encounter the next problem. I am currently trying to solve that. When I use ‘systemctl suspend’ after applying my patch, pcie currently does not allow suspend mode anymore. In addition, the moal module seems to have problems so that the device no longer functions correctly.

[   75.372360] Filesystems sync: 0.014 seconds
[   75.377098] Freezing user space processes
[   75.405396] Freezing user space processes completed (elapsed 0.024 seconds)
[   75.412472] OOM killer disabled.
[   75.415717] Freezing remaining freezable tasks
[   75.416862] Freezing remaining freezable tasks completed (elapsed 0.001 seconds)
[   75.427883] printk: Suspending console(s) (use no_console_suspend to debug)
[   76.273062] wlan_pcie 0000:01:00.0: PM: pci_legacy_suspend(): woal_pcie_suspend+0x0/0x450 [moal] returns -16
[   76.273173] wlan_pcie 0000:01:00.0: PM: dpm_run_callback(): pci_pm_suspend+0x0/0x258 returns -16
[   76.273188] wlan_pcie 0000:01:00.0: PM: failed to suspend async: error -16
[   76.274115] PM: Some devices failed to suspend, or early wake event detected

Best regards

R. 

[pengyong_zhang]

Hi @rrd 

Why are you use "systemctl suspend" command apple the patch? you can re-compile the kernel Image and re-boot your board.

And I haven't received your patch file.

B.R

[rrd]

Hi @pengyong_zhang ,

i don't use 'systemctl suspend' to apply the patch. One of the mandatory requirements is that the device must be able to enter and exit suspend mode without any problems. For this reason, I tested this and found that my previous patch was causing problems. Before the patch, it was possible to go into suspend mode. Therefore, it must be due to my change. 

Best regards

R.

[rrd]

It seems to make a difference whether it is the first call to ‘systemctl suspend’ after booting or a subsequent call.

first call:

systemct[   31.455442] Scan in progress, ignore ADDBA Request event
l suspend
root@proto1-imx8mp-3:~# [   36.127317] IOCTL failed: 0000000051f95355 id=0x10000, sub_id=0x10003 action=1, status_code=0x80000007
[   36.605028] PM: suspend entry (deep)
[   36.644969] Filesystems sync: 0.036 seconds
[   36.652111] Freezing user space processes
[   36.677562] Freezing user space processes completed (elapsed 0.021 seconds)
[   36.684580] OOM killer disabled.
[   36.687827] Freezing remaining freezable tasks
[   36.693613] Freezing remaining freezable tasks completed (elapsed 0.001 seconds)
[   36.701033] printk: Suspending console(s) (use no_console_suspend to debug)
[   36.874932] PM: suspend devices took 0.160 seconds
[   36.895966] Disabling non-boot CPUs ...
[   36.896919] IRQ215: set affinity failed(-22).
[   36.896946] IRQ234: set affinity failed(-22).
[   36.897000] psci: CPU1 killed (polled 0 ms)
[   36.898387] IRQ215: set affinity failed(-22).
[   36.898417] IRQ234: set affinity failed(-22).
[   36.899480] psci: CPU2 killed (polled 0 ms)
[   36.900878] IRQ215: set affinity failed(-22).
[   36.900906] IRQ234: set affinity failed(-22).
[   36.900975] psci: CPU3 killed (polled 0 ms)
[   36.901555] Enabling non-boot CPUs ...
[   36.901980] Detected VIPT I-cache on CPU1
[   36.902013] GICv3: CPU1: found redistributor 1 region 0:0x00000000388a0000
[   36.902048] CPU1: Booted secondary processor 0x0000000001 [0x410fd034]
[   36.902722] CPU1 is up
[   36.903080] Detected VIPT I-cache on CPU2
[   36.903099] GICv3: CPU2: found redistributor 2 region 0:0x00000000388c0000
[   36.903119] CPU2: Booted secondary processor 0x0000000002 [0x410fd034]
[   36.903612] CPU2 is up
[   36.903967] Detected VIPT I-cache on CPU3
[   36.903987] GICv3: CPU3: found redistributor 3 region 0:0x00000000388e0000
[   36.904008] CPU3: Booted secondary processor 0x0000000003 [0x410fd034]
[   36.904574] CPU3 is up
[   37.113209] imx6q-pcie 33800000.pcie: PCIe Gen.1 x1 link up
[   37.113227] imx6q-pcie 33800000.pcie: PCIe Gen.1 x1 link up
[   37.113231] imx6q-pcie 33800000.pcie: Link up, Gen1
[   37.477842] xhci-hcd xhci-hcd.1.auto: xHC error in resume, USBSTS 0x401, Reinit
[   37.477858] usb usb1: root hub lost power or was reset
[   37.477863] usb usb2: root hub lost power or was reset
[   37.479659] xhci-hcd xhci-hcd.2.auto: xHC error in resume, USBSTS 0x401, Reinit
[   37.479672] usb usb3: root hub lost power or was reset
[   37.479677] usb usb4: root hub lost power or was reset
[   37.485549] woal_cancel_hs: woal_set_get_hs_params failed 
[   40.676053] wlan_wakeup_card_timeout_func: ps_state=3
[   40.676071] mlan0: 
[   40.676076] Wakeup card timeout(0)!
[   43.748025] wlan_wakeup_card_timeout_func: ps_state=3
[   43.748037] mlan0: 
[   43.748042] Wakeup card timeout(1)!
[   46.820338] wlan_wakeup_card_timeout_func: ps_state=3
[   46.820343] mlan0: 
[   46.820345] Wakeup card timeout(2)!
[   46.820365] Driver version = PCIE8997--16.92.21.p119.3-MM6X16437.p3-GPL-(FP92) 
[   46.820368] main_state = 3
[   46.820370] ioctl_pending = 2
[   46.820372] tx_pending = 0
[   46.820374] wmm_tx_pending[0] = 0
[   46.820375] wmm_tx_pending[1] = 0
[   46.820377] wmm_tx_pending[2] = 0
[   46.820379] wmm_tx_pending[3] = 0
[   46.820381] rx_pending = 0
[   46.820383] lock_count = 85
[   46.820385] malloc_count = 53
[   46.820387] mbufalloc_count = 40
[   46.820389] malloc_cons_count = 4
[   46.820391] hs_skip_count = 0
[   46.820392] hs_force_count = 0
[   46.820395] Media state = "Disconnected"
[   46.820397] carrier off
[   46.820399] tx queue 0: started
[   46.820401] tx queue 1: started
[   46.820404] tx queue 2: started
[   46.820406] tx queue 3: started
[   46.820408] mlan0: num_tx_timeout = 0
[   46.820411] uap0: num_tx_timeout = 0
[   46.820414] wfd0: num_tx_timeout = 0
[   46.820429] reg:0x00 value=0x2b421b4b
[   46.820437] reg:0x04 value=0x00100406
[   46.820444] reg:0x10 value=0x1810000c
[   46.820451] reg:0x18 value=0x1820000c
[   46.820459] reg:0x2c value=0x00000000
[   46.820466] reg:0x3c value=0x000001d6
[   46.820473] reg:0x44 value=0x00000008
[   46.820480] reg:0x80 value=0x10110040
[   46.820487] reg:0x98 value=0x00000000
[   46.820494] reg:0x170 value=0x0001001e
[   46.820501] reg:0xcf0 value=0x0
[   46.820505] reg:0xcf8 value=0x0
[   46.820510] reg:0xcfc value=0x0
[   46.922213] reg:0xcf8 value=0x0
[   46.922218] reg:0xcfc value=0x0
[   47.024390] IOCTL failed: 00000000bf55be26 id=0x90000, sub_id=0x90002 action=1, status_code=0x80000007
[   47.024400] IOCTL failed: 00000000860dac8e id=0x90000, sub_id=0x9000b action=2, status_code=0x80000007
[   47.024475] woal_reset_intf: get bss info failed 
[   47.024708] woal_cfg80211_resume: get_wakeup_reason failed 
[   47.038449] PM: resume devices took 9.916 seconds
[   47.149089] woal_close: woal_disconnect failed 
[   47.482813] OOM killer enabled.
[   47.485971] Restarting tasks ... done.
[   47.491474] random: crng reseeded on system resumption
[   47.509369] PM: suspend exit
[   47.577420] TI DP83867 30be0000.ethernet-1:00: attached PHY driver (mii_bus:phy_addr=30be0000.ethernet-1:00, irq=POLL)
[   47.862373] Failed to poll FW reset register CF4=0x99
[   47.867447] PCIe In-band Reset Fail

in the first call suspend state seems to be reached

second call:

systemctl suspend
root@proto1-imx8mp-3:~# [  336.385651] PM: suspend entry (deep)
[  336.395071] Filesystems sync: 0.005 seconds
[  336.401641] Freezing user space processes
[  336.423851] Freezing user space processes completed (elapsed 0.018 seconds)
[  336.434189] OOM killer disabled.
[  336.437425] Freezing remaining freezable tasks
[  336.443190] Freezing remaining freezable tasks completed (elapsed 0.001 seconds)
[  336.450619] printk: Suspending console(s) (use no_console_suspend to debug)
[  337.296053] wlan_pcie 0000:01:00.0: PM: pci_legacy_suspend(): woal_pcie_suspend+0x0/0x450 [moal] returns -16
[  337.296169] wlan_pcie 0000:01:00.0: PM: dpm_run_callback(): pci_pm_suspend+0x0/0x258 returns -16
[  337.296185] wlan_pcie 0000:01:00.0: PM: failed to suspend async: error -16
[  337.297108] PM: Some devices failed to suspend, or early wake event detected
[  337.333014] PM: resume devices took 0.036 seconds
[  337.393024] OOM killer enabled.
[  337.396177] Restarting tasks ... done.
[  337.401318] random: crng reseeded on system resumption
[  337.406568] PM: suspend exit
[  337.409658] PM: suspend entry (s2idle)
[  337.662584] Filesystems sync: 0.249 seconds
[  337.667273] Freezing user space processes
[  337.692910] Freezing user space processes completed (elapsed 0.021 seconds)
[  337.699890] OOM killer disabled.
[  337.703116] Freezing remaining freezable tasks
[  337.708889] Freezing remaining freezable tasks completed (elapsed 0.001 seconds)
[  337.716317] printk: Suspending console(s) (use no_console_suspend to debug)
[  338.560058] wlan_pcie 0000:01:00.0: PM: pci_legacy_suspend(): woal_pcie_suspend+0x0/0x450 [moal] returns -16
[  338.560177] wlan_pcie 0000:01:00.0: PM: dpm_run_callback(): pci_pm_suspend+0x0/0x258 returns -16
[  338.560192] wlan_pcie 0000:01:00.0: PM: failed to suspend async: error -16
[  338.560805] PM: Some devices failed to suspend, or early wake event detected
[  338.594530] PM: resume devices took 0.032 seconds
[  338.654823] OOM killer enabled.
[  338.657981] Restarting tasks ... done.
[  338.662955] random: crng reseeded on system resumption
[  338.668201] PM: suspend exit
[  338.695157] TI DP83867 30be0000.ethernet-1:00: attached PHY driver (mii_bus:phy_addr=30be0000.ethernet-1:00, irq=POLL)

However, after waking up and re-executing systemctl suspend, this is not the case.

[pengyong_zhang]

Hi @rrd 

Glad to hear that, but have you tried the code patch in pci-imx6.c file imx_pcie_shutdown I provided? What were the test results?

 if (imx_pcie->pci && imx_pcie->pci->ops && imx_pcie->pci->ops->stop_link) {
dev_info(&pdev->dev, "%s: stopping PCIe link\n", __func__);
imx_pcie->pci->ops->stop_link(imx_pcie->pci);
} else {
dev_warn(&pdev->dev, "%s: stop_link not available\n", __func__);
}

B.R

[rrd]

Hi @pengyong_zhang ,

to be honest, I didn't try adding imx_pcie->pci->ops->stop_link(imx_pcie->pci) to the code.

Let me explain why:

Using the oscilloscope, I noticed that the levels of PERST0# and W_DISABLE1# gpios (pin labels of the 88w8997) were not behaving as they should during the power-up sequence according to the data sheet. Therefore, I thought that imx_pcie->pci->ops->stop_link(imx_pcie->pci) would not change this behaviour either.

Best regards

R.

[pengyong_zhang]

Hi @rrd 

What is the check result now?

B.R

[rrd]

Hello and thank you for your enquiry.


First, here are my hardware details again:

I have an imx8mp-phycore-som from phytec.

I would like to use it with an embeddedArtists 1YM module for PCIe. This worked without any problems in the old kernel 5.15.71.
However, during today's measurements, I noticed that the timing was not quite as it should be. I was just lucky that it still worked reliably back then.

In the current BSP with kernel 6.6.23, I had problems implementing it in such a way that it could be used after a restart in the same way as after a complete shutdown. The problem was that after a restart, the message ‘Phy Link never came up’ appeared and I couldn't get the WiFi module to work reliably.

Over the last few days, I have been working intensively on getting everything to work with adjustments to the device tree, but I was unable to set the correct timing behaviour. I experimented with a fixed controller and tried various other things. At some point, I gave up and adapted the pci-imx6.c driver similar to kernel 5.15.71.

I just had the long-awaited breakthrough.

What I did: I added the W_DISABLE1 pin as gpio_desc *power_on_gpio to the imx6_pcie structure.

In imx6_pcie_host_init, this is set to high after the clock has been activated. If pcie PHY power up fails, it is set back to low. In the case of err_reg_disable, there is a 100 ms wait after disabling the regulator. In imx6_pcie_host_exit, power_on_gpio is also pulled low before disabling the clock. After disabling the regulator, there is another 100 ms wait. In imx6_pcie_probe, power_on_gpio is initialised and pulled low before the reset_gpio is handled.

This provides the necessary timing for the card to boot up cleanly.

Best

R.

[pengyong_zhang]

Hi @rrd 

I use the below dts file with my imx8mp evk board. And do not change any code.

https://github.com/nxf90153/linux-imx_myGit/blob/6beeed18c88145fac733824ffb525274c9bcfc3e/arch/arm64...

Also you can try to add the below code in below file. function imx_pcie_shutdown().

https://github.com/nxf90153/linux-imx_myGit/blob/8f83c86d42cef621156627b7e56380653fb66731/drivers/pc...

 if (imx_pcie->pci && imx_pcie->pci->ops && imx_pcie->pci->ops->stop_link) {
dev_info(&pdev->dev, "%s: stopping PCIe link\n", __func__);
imx_pcie->pci->ops->stop_link(imx_pcie->pci);
} else {
dev_warn(&pdev->dev, "%s: stop_link not available\n", __func__);
}

And you mentioned that not all boards have this issue, only your colleague's board is affected. Have you checked whether the PCIe hardware connections on the faulty board are intact?

B.R

[rrd]

Hi @pengyong_zhang

i'm on vacation this week and will test it right after I get back and let you know. I don't think it's a wiring/connection problem.

R.

[rrd]

Hi @pengyong_zhang ,

i am back from holiday and will start comparing your DT section with ours. Nevertheless, I asked my colleagues to check the quality of the soldered joints again during my holiday.

R.

[pengyong_zhang]

Hi @rrd 

What is your kernel version? I have tested your scenario on my i,MX8MP EVK board with pcie Wifi-BT card. When i use "reboot" command reboot board, there is no problem occurred. please share your log file. How did you reset your board?

B.R

[rrd]

Could you please send me your device tree settings? (Everything needed to get such a PCIe device up and running. So if you have set up a special controller, please include that entry as well). Then I could compare it with my settings...

thanks in advance
R.

[rrd]

Here are the logs

root@dut-proto1-imx8mp-3:~# # after reboot command
root@dut-proto1-imx8mp-3:~#
root@dut-proto1-imx8mp-3:~#
root@dut-proto1-imx8mp-3:~# dmesg | sed -n '/pci/,+10p'
[ 2.953057] imx6q-pcie 33800000.pcie: host bridge /soc@0/pcie@33800000 ranges:
[ 2.953099] imx6q-pcie 33800000.pcie: IO 0x001ff80000..0x001ff8ffff -> 0x0000000000
[ 2.953117] imx6q-pcie 33800000.pcie: MEM 0x0018000000..0x001fefffff -> 0x0018000000
[ 2.960606] ### panel-dut-xxxx-ourdisplay: ot_panel_enable()
[ 2.960612] panel-dut-xxxx-ourdisplay 32e60000.mipi_dsi.0: dsi ourdisplay_enable()
[ 2.960617] panel-dut-xxxx-ourdisplay 32e60000.mipi_dsi.0: ### red_gain_num: 105, red_gain_den: 100, green_gain_num: 99, green_gain_den: 100, blue_gain_num: 93, blue_gain_d
en: 100
[ 3.168756] imx6q-pcie 33800000.pcie: iATU: unroll T, 4 ob, 4 ib, align 64K, limit 16G
[ 3.245768] Console: switching to colour frame buffer device 60x50
[ 3.333707] imx-drm display-subsystem: [drm] fb0: imx-drmdrmfb frame buffer device
[ 3.341641] ### panel-dut-xxxx-ourdisplay: ot_panel_get_modes()
[ 3.344756] pps pps0: new PPS source ptp0
[ 4.168649] imx6q-pcie 33800000.pcie: Phy link never came up
[ 4.678685] bq256xx-charger 3-006b: switch cc controller to low!
[ 5.176740] imx6q-pcie 33800000.pcie: Phy link never came up
[ 5.182870] imx6q-pcie 33800000.pcie: PCI host bridge to bus 0000:00
[ 5.189320] pci_bus 0000:00: root bus resource [bus 00-ff]
[ 5.194844] pci_bus 0000:00: root bus resource [io 0x0000-0xffff]
[ 5.201052] pci_bus 0000:00: root bus resource [mem 0x18000000-0x1fefffff]
[ 5.208008] pci 0000:00:00.0: [16c3:abcd] type 01 class 0x060400
[ 5.208024] pci 0000:00:00.0: reg 0x10: [mem 0x00000000-0x000fffff]
[ 5.208037] pci 0000:00:00.0: reg 0x38: [mem 0x00000000-0x0000ffff pref]
[ 5.208090] pci 0000:00:00.0: supports D1
[ 5.208095] pci 0000:00:00.0: PME# supported from D0 D1 D3hot D3cold
[ 5.210923] pci 0000:00:00.0: BAR 0: assigned [mem 0x18000000-0x180fffff]
[ 5.244439] pci 0000:00:00.0: BAR 6: assigned [mem 0x18100000-0x1810ffff pref]
[ 5.251676] pci 0000:00:00.0: PCI bridge to [bus 01-ff]
[ 5.252117] pcieport 0000:00:00.0: PME: Signaling with IRQ 233
[ 5.275376] EXT4-fs (mmcblk2p5): mounted filesystem c69f2650-c859-4d92-bd01-c47029bcbd90 r/w with ordered data mode. Quota mode: none.
[ 5.287547] VFS: Mounted root (ext4 filesystem) on device 179:5.
[ 5.288595] devtmpfs: mounted
[ 5.304812] Freeing unused kernel memory: 8448K
[ 5.309529] Run /sbin/init as init process
[ 5.313635] with arguments:
root@dut-proto1-imx8mp-3:~# lspci
00:00.0 PCI bridge: Synopsys, Inc. DWC_usb3 / PCIe bridge (rev 01)
root@dut-proto1-imx8mp-3:~#



root@dut-proto1-imx8mp-3:~# # after normal start-up or switching off and on
root@dut-proto1-imx8mp-3:~#
root@dut-proto1-imx8mp-3:~#
root@dut-proto1-imx8mp-3:~# dmesg | sed -n '/pci/,+10p'
[ 3.011889] imx6q-pcie 33800000.pcie: host bridge /soc@0/pcie@33800000 ranges:
[ 3.011933] imx6q-pcie 33800000.pcie: IO 0x001ff80000..0x001ff8ffff -> 0x0000000000
[ 3.011951] imx6q-pcie 33800000.pcie: MEM 0x0018000000..0x001fefffff -> 0x0018000000
[ 3.019439] ### panel-dut-xxxx-ourdisplay: ot_panel_enable()
[ 3.019445] panel-dut-xxxx-ourdisplay 32e60000.mipi_dsi.0: dsi ourdisplay_enable()
[ 3.019450] panel-dut-xxxx-ourdisplay 32e60000.mipi_dsi.0: ### red_gain_num: 105, red_gain_den: 100, green_gain_num: 99, green_gain_den: 100, blue_gain_num: 93, blue_gain_d
en: 100
[ 3.227889] imx6q-pcie 33800000.pcie: iATU: unroll T, 4 ob, 4 ib, align 64K, limit 16G
[ 3.303603] Console: switching to colour frame buffer device 60x50
[ 3.328245] imx6q-pcie 33800000.pcie: PCIe Gen.1 x1 link up
[ 3.328978] imx-drm display-subsystem: [drm] fb0: imx-drmdrmfb frame buffer device
[ 3.337154] imx6q-pcie 33800000.pcie: PCIe Gen.1 x1 link up
[ 3.410294] imx6q-pcie 33800000.pcie: Link up, Gen1
[ 3.410632] ### panel-dut-xxxx-ourdisplay: ot_panel_get_modes()
[ 3.415186] imx6q-pcie 33800000.pcie: PCIe Gen.1 x1 link up
[ 3.419225] pps pps0: new PPS source ptp0
[ 3.432062] imx6q-pcie 33800000.pcie: PCI host bridge to bus 0000:00
[ 3.438463] pci_bus 0000:00: root bus resource [bus 00-ff]
[ 3.443969] pci_bus 0000:00: root bus resource [io 0x0000-0xffff]
[ 3.448507] fec 30be0000.ethernet eth0: registered PHC device 0
[ 3.450170] pci_bus 0000:00: root bus resource [mem 0x18000000-0x1fefffff]
[ 3.457621] xhci-hcd xhci-hcd.1.auto: xHCI Host Controller
[ 3.462996] pci 0000:00:00.0: [16c3:abcd] type 01 class 0x060400
[ 3.474486] pci 0000:00:00.0: reg 0x10: [mem 0x00000000-0x000fffff]
[ 3.482520] xhci-hcd xhci-hcd.1.auto: hcc params 0x0220fe6d hci version 0x110 quirks 0x000000a001000010
[ 3.488411] pci 0000:00:00.0: reg 0x38: [mem 0x00000000-0x0000ffff pref]
[ 3.497837] xhci-hcd xhci-hcd.1.auto: irq 219, io mem 0x38100000
[ 3.504556] pci 0000:00:00.0: supports D1
[ 3.510654] xhci-hcd xhci-hcd.1.auto: xHCI Host Controller
[ 3.514524] pci 0000:00:00.0: PME# supported from D0 D1 D3hot D3cold
[ 3.520018] xhci-hcd xhci-hcd.1.auto: new USB bus registered, assigned bus number 2
[ 3.529736] pci 0000:01:00.0: [1b4b:2b42] type 00 class 0x020000
[ 3.534033] xhci-hcd xhci-hcd.1.auto: Host supports USB 3.0 SuperSpeed
[ 3.540117] pci 0000:01:00.0: reg 0x10: [mem 0x00000000-0x000fffff 64bit pref]
[ 3.553842] pci 0000:01:00.0: reg 0x18: [mem 0x00000000-0x000fffff 64bit pref]
[ 3.557578] hub 1-0:1.0: 1 port detected
[ 3.565187] pci 0000:01:00.0: supports D1 D2
[ 3.569013] usb usb2: We don't know the algorithms for LPM for this host, disabling LPM.
[ 3.572988] pci 0000:01:00.0: PME# supported from D0 D1 D3hot D3cold
[ 3.581871] hub 2-0:1.0: USB hub found
[ 3.587604] pci 0000:01:00.0: 2.000 Gb/s available PCIe bandwidth, limited by 2.5 GT/s PCIe x1 link at 0000:00:00.0 (capable of 4.000 Gb/s with 5.0 GT/s PCIe x1 link)
[ 3.591223] hub 2-0:1.0: 1 port detected
[ 3.611709] xhci-hcd xhci-hcd.2.auto: xHCI Host Controller
[ 3.615897] pci 0000:00:00.0: BAR 0: assigned [mem 0x18000000-0x180fffff]
[ 3.617235] xhci-hcd xhci-hcd.2.auto: new USB bus registered, assigned bus number 3
[ 3.624017] pci 0000:00:00.0: BAR 15: assigned [mem 0x18100000-0x182fffff pref]
[ 3.632052] xhci-hcd xhci-hcd.2.auto: hcc params 0x0220fe6d hci version 0x110 quirks 0x000000a001000010
[ 3.639087] pci 0000:00:00.0: BAR 6: assigned [mem 0x18300000-0x1830ffff pref]
[ 3.648409] xhci-hcd xhci-hcd.2.auto: irq 220, io mem 0x38200000
[ 3.655674] pci 0000:01:00.0: BAR 0: assigned [mem 0x18100000-0x181fffff 64bit pref]
[ 3.661735] xhci-hcd xhci-hcd.2.auto: xHCI Host Controller
[ 3.669386] pci 0000:01:00.0: BAR 2: assigned [mem 0x18200000-0x182fffff 64bit pref]
[ 3.674847] xhci-hcd xhci-hcd.2.auto: new USB bus registered, assigned bus number 4
[ 3.682703] pci 0000:00:00.0: PCI bridge to [bus 01-ff]
[ 3.690254] xhci-hcd xhci-hcd.2.auto: Host supports USB 3.0 SuperSpeed
[ 3.695554] pci 0000:00:00.0: bridge window [mem 0x18100000-0x182fffff pref]
[ 3.709742] pcieport 0000:00:00.0: PME: Signaling with IRQ 222
[ 3.713028] hub 3-0:1.0: 1 port detected
[ 3.723089] usb usb4: We don't know the algorithms for LPM for this host, disabling LPM.
[ 3.732009] hub 4-0:1.0: USB hub found
[ 3.735803] hub 4-0:1.0: 1 port detected
[ 3.741061] Hot alarm is canceled. GPU3D clock will return to 64/64
[ 3.748608] imx-cpufreq-dt imx-cpufreq-dt: cpu speed grade 7 mkt segment 2 supported-hw 0x80 0x4
[ 3.760642] sdhci-esdhc-imx 30b50000.mmc: Got CD GPIO
[ 3.766527] nxp-pca9450 0-0025: PWRON interrupt.
[ 3.771176] nxp-pca9450 0-0025: WDOGB interrupt.
[ 3.775657] input: custom_keyboard as /devices/platform/custom_keyboard/input/input1
[ 5.791053] wlan_pcie 0000:01:00.0: enabling device (0000 -> 0002)
[ 5.797422] PCI memory map Virt0: 00000000161b0b11 PCI memory map Virt2: 00000000e0b06e9a
[ 5.805688] Attach moal handle ops, card interface type: 0x204
[ 5.811559] rps set to 0 from module param
[ 5.817047] PCIE8997: init module param from usr cfg
[ 5.822084] card_type: PCIE8997, config block: 0
[ 5.822099] cfg80211_wext=0xf
[ 5.822102] max_vir_bss=1
[ 5.822107] cal_data_cfg=none
[ 5.822110] ps_mode = 1
[ 5.822113] auto_ds = 1
[ 5.822126] fw_name=nxp/pcieuart8997_combo_v4.bin
[ 5.822150] rx_work=1 cpu_num=4
[ 5.822156] Enable moal_recv_amsdu_packet
[ 5.822182] Attach mlan adapter operations.card_type is 0x204.
[ 5.823422] Request firmware: nxp/pcieuart8997_combo_v4.bin
[ 5.823836] systemd[1]: Started Journal Service.
[ 6.235026] systemd-journald[161]: Received client request to flush runtime journal.
[ 6.511661] FW download over, size 621300 bytes
[ 7.239201] WLAN FW is active
[ 7.242211] on_time is 7239095875
[ 7.293239] EXT4-fs (mmcblk2p3): mounted filesystem aa764bef-f1a9-493d-87e5-4b2b34315827 r/w with ordered data mode. Quota mode: none.
root@dut-proto1-imx8mp-3:~# lspci
00:00.0 PCI bridge: Synopsys, Inc. DWC_usb3 / PCIe bridge (rev 01)
01:00.0 Ethernet controller: Marvell Technology Group Ltd. Device 2b42 (rev 11)
root@dut-proto1-imx8mp-3:~#

[rrd]

Thank you for your quick reply.

The old kernel version was 5.15.71.
The new kernel version is 6.6.23.

One thing to note is that this does not occur on every device (of the same model). All of the devices I have here on site are working correctly. However, my colleagues have devices of exactly the same model on which this occurs. They all have the same rauc bundle/image installed on their devices. I will probably receive one of these devices with the malfunction on my desk today. I will then post the requested output immediately.

P.S. By the way, we use the same WiFi chip that I saw in your lspci output. It is the Murata 1YM module with the 88w8997 chipset.

Best regards,
R.

[pengyong_zhang]

Hi @rrd 

Our pcie device not support warm reset. Only cold reset can Re-enumerate PCIe devices. That is why you said "if I shut it down completely with poweroff, I don't have this problem" EP and RC need to be powered off. Power on EP, then power on RC.

B.R

[rrd]

Hello and thank you for your reply.

I just don't understand why it worked in kernel 5.15.71. I'm not that familiar with pci(e).
Is the problem more likely to be caused by the WiFi card, which is controlled via PCIe, or is it caused by the PCIe itself?

We have to get this problem under control somehow... I've never seen this happen before, either with Windows or Linux, where I had to completely shut down the operating system so that the hardware would work the next time I booted up.

Can you give me some more detailed instructions on how to solve the problem?

Many thanks in advance.
Ralf