Skip to contentSkip to conversation

New: broken device, retraining non-functional downstream link at 2.5GT/s]

[email protected]: Bjorn Helgaas 3 participants 3 messages Apr 23, 2024 – Apr 24, 2024 PCI

lore
Share & tools
lore
#0 lore raw

FYI. The retraining was added by a89c82249c37 ("PCI: Work around PCIe link training failures").

Paul, is this a regression? a89c82249c37 appeared in v6.5. I *assume* whatever is below bus 01 did actually work before v6.5, in spite of the fact that apparently PCI_EXP_LNKSTA_DLLLA was not set when we enumerated the 00:1c.0 Root Port?

----- Forwarded message from [email protected] -----

https://bugzilla.kernel.org/show_bug.cgi?id=218765

Created attachment 306199
--> https://bugzilla.kernel.org/attachment.cgi?id=306199&action=edit
Linux 6.9-rc5+ messages (output of dmesg)

I noticed a one second delay with Linux 6.9-rc5+:

```
[    0.000000] DMI: Dell Inc. XPS 13 9360/0596KF, BIOS 2.21.0 06/02/2022
[…]
[    0.201109] pci 0000:00:1c.0: [8086:9d10] type 01 class 0x060400 PCIe Root
Port
[    0.201109] pci 0000:00:1c.0: PCI bridge to [bus 01-39]
[    0.201109] pci 0000:00:1c.0:   bridge window [mem 0xc4000000-0xda0fffff]
[    0.201109] pci 0000:00:1c.0:   bridge window [mem 0xa0000000-0xc1ffffff
64bit pref]
[    0.201109] pci 0000:00:1c.0: broken device, retraining non-functional
downstream link at 2.5GT/s
[    1.209109] pci 0000:00:1c.0: retraining failed
[    1.209143] pci 0000:00:1c.0: PME# supported from D0 D3hot D3cold
[    1.209677] pci 0000:00:1c.4: [8086:9d14] type 01 class 0x060400 PCIe Root
Port
[    1.209700] pci 0000:00:1c.4: PCI bridge to [bus 3a]
[    1.209705] pci 0000:00:1c.4:   bridge window [mem 0xdc000000-0xdc1fffff]
[    1.209771] pci 0000:00:1c.4: PME# supported from D0 D3hot D3cold
[…]
```

Looking through all the logs since March 2024, I only found *one* other occurrence with Linux 6.9-rc4+.

```
$ lspci -tvnn
-[0000:00]-+-00.0  Intel Corporation Xeon E3-1200 v6/7th Gen Core Processor
Host Bridge/DRAM Registers [8086:5904]
           +-02.0  Intel Corporation HD Graphics 620 [8086:5916]
           +-04.0  Intel Corporation Xeon E3-1200 v5/E3-1500 v5/6th Gen Core
Processor Thermal Subsystem [8086:1903]
           +-14.0  Intel Corporation Sunrise Point-LP USB 3.0 xHCI Controller
[8086:9d2f]
           +-14.2  Intel Corporation Sunrise Point-LP Thermal subsystem
[8086:9d31]
           +-15.0  Intel Corporation Sunrise Point-LP Serial IO I2C Controller
#0 [8086:9d60]
           +-15.1  Intel Corporation Sunrise Point-LP Serial IO I2C Controller
#1 [8086:9d61]
           +-16.0  Intel Corporation Sunrise Point-LP CSME HECI #1 [8086:9d3a]
           +-1c.0-[01-39]----00.0-[02-39]--+-00.0-[03]--
           |                               +-01.0-[04-38]--
           |                               \-02.0-[39]----00.0  Intel
Corporation DSL6340 USB 3.1 Controller [Alpine Ridge] [8086:15b5]
           +-1c.4-[3a]----00.0  Qualcomm Atheros QCA6174 802.11ac Wireless
Network Adapter [168c:003e]
           +-1d.0-[3b]----00.0  SK hynix PC300 NVMe Solid State Drive 512GB
[1c5c:1284]
           +-1f.0  Intel Corporation Sunrise Point-LP LPC Controller
[8086:9d58]
           +-1f.2  Intel Corporation Sunrise Point-LP PMC [8086:9d21]
           +-1f.3  Intel Corporation Sunrise Point-LP HD Audio [8086:9d71]
           \-1f.4  Intel Corporation Sunrise Point-LP SMBus [8086:9d23]
```

The adapter wasn’t plugged in, when Linux started.

It could be related to unplugging an Dell DA300 USB Type-C adapter during system firmware (UEFI) to avoid a five second delay in Linux (ACPI). I need to test that later, but maybe you already have an idea.

----- End forwarded message -----

#1 lore raw

Dear Bjorn,

Thank you for your instant reply.

Am 23.04.24 um 17:23 schrieb Bjorn Helgaas:

Quoted 7 lines

FYI. The retraining was added by a89c82249c37 ("PCI: Work around PCIe link training failures").

Paul, is this a regression? a89c82249c37 appeared in v6.5. I *assume* whatever is below bus 01 did actually work before v6.5, in spite of the fact that apparently PCI_EXP_LNKSTA_DLLLA was not set when we enumerated the 00:1c.0 Root Port?

Only used 6.5 and onward on the Dell XPS 9360, so I cannot say if it worked before. I have to find time to boot an old Linux kernel image. (I’d like to emphasize again, that this only happens having USB-C adapter connected during system firmware phase, and disconnecting it before Linux boots – for example, when in GRUB.)

Kind regards,

Paul

Quoted 58 lines

----- Forwarded message from [email protected] -----

https://bugzilla.kernel.org/show_bug.cgi?id=218765

Created attachment 306199
--> https://bugzilla.kernel.org/attachment.cgi?id=306199&action=edit
Linux 6.9-rc5+ messages (output of dmesg)

I noticed a one second delay with Linux 6.9-rc5+:

```
[    0.000000] DMI: Dell Inc. XPS 13 9360/0596KF, BIOS 2.21.0 06/02/2022
[…]
[    0.201109] pci 0000:00:1c.0: [8086:9d10] type 01 class 0x060400 PCIe Root Port
[    0.201109] pci 0000:00:1c.0: PCI bridge to [bus 01-39]
[    0.201109] pci 0000:00:1c.0:   bridge window [mem 0xc4000000-0xda0fffff]
[    0.201109] pci 0000:00:1c.0:   bridge window [mem 0xa0000000-0xc1ffffff 64bit pref]
[    0.201109] pci 0000:00:1c.0: broken device, retraining non-functional downstream link at 2.5GT/s
[    1.209109] pci 0000:00:1c.0: retraining failed
[    1.209143] pci 0000:00:1c.0: PME# supported from D0 D3hot D3cold
[    1.209677] pci 0000:00:1c.4: [8086:9d14] type 01 class 0x060400 PCIe Root Port
[    1.209700] pci 0000:00:1c.4: PCI bridge to [bus 3a]
[    1.209705] pci 0000:00:1c.4:   bridge window [mem 0xdc000000-0xdc1fffff]
[    1.209771] pci 0000:00:1c.4: PME# supported from D0 D3hot D3cold
[…]
```

Looking through all the logs since March 2024, I only found *one* other occurrence with Linux 6.9-rc4+.

```
$ lspci -tvnn
-[0000:00]-+-00.0  Intel Corporation Xeon E3-1200 v6/7th Gen Core Processor Host Bridge/DRAM Registers [8086:5904]
            +-02.0  Intel Corporation HD Graphics 620 [8086:5916]
            +-04.0  Intel Corporation Xeon E3-1200 v5/E3-1500 v5/6th Gen Core Processor Thermal Subsystem [8086:1903]
            +-14.0  Intel Corporation Sunrise Point-LP USB 3.0 xHCI Controller [8086:9d2f]
            +-14.2  Intel Corporation Sunrise Point-LP Thermal subsystem [8086:9d31]
            +-15.0  Intel Corporation Sunrise Point-LP Serial IO I2C Controller #0 [8086:9d60]
            +-15.1  Intel Corporation Sunrise Point-LP Serial IO I2C Controller #1 [8086:9d61]
            +-16.0  Intel Corporation Sunrise Point-LP CSME HECI #1 [8086:9d3a]
            +-1c.0-[01-39]----00.0-[02-39]--+-00.0-[03]--
            |                               +-01.0-[04-38]--
            |                               \-02.0-[39]----00.0  Intel Corporation DSL6340 USB 3.1 Controller [Alpine Ridge] [8086:15b5]
            +-1c.4-[3a]----00.0  Qualcomm Atheros QCA6174 802.11ac Wireless Network Adapter [168c:003e]
            +-1d.0-[3b]----00.0  SK hynix PC300 NVMe Solid State Drive 512GB [1c5c:1284]
            +-1f.0  Intel Corporation Sunrise Point-LP LPC Controller [8086:9d58]
            +-1f.2  Intel Corporation Sunrise Point-LP PMC [8086:9d21]
            +-1f.3  Intel Corporation Sunrise Point-LP HD Audio [8086:9d71]
            \-1f.4  Intel Corporation Sunrise Point-LP SMBus [8086:9d23]
```

The adapter wasn’t plugged in, when Linux started.

It could be related to unplugging an Dell DA300 USB Type-C adapter during system firmware (UEFI) to avoid a five second delay in Linux (ACPI). I need to test that later, but maybe you already have an idea.

----- End forwarded message -----

#2 lore raw
In reply to Bjorn Helgaas
Quoted 7 lines · Bjorn Helgaas

FYI. The retraining was added by a89c82249c37 ("PCI: Work around PCIe link training failures").

Paul, is this a regression? a89c82249c37 appeared in v6.5. I *assume* whatever is below bus 01 did actually work before v6.5, in spite of the fact that apparently PCI_EXP_LNKSTA_DLLLA was not set when we enumerated the 00:1c.0 Root Port?

Bjorn,

Has <https://patchwork.kernel.org/project/linux-pci/list/?series=824858>
gone anywhere? I believe this would've addressed the issue of the delay.

Over the last couple of months Ilpo and myself have been discussing the
issue of clearing leftover LBMS state and Smita has literally just posted a proposed solution for that, which I haven't yet had time to go through. The combination of a new day job, overall shortage of time, and hardware playing tricks with me has contributed to me not having made anything in this area recently. I haven't lost interest though and I'll see what I can do.

Maciej