You are not logged in.

#1 2018-11-23 20:26:48

InvisibleRasta
Member
Registered: 2017-04-12
Posts: 111

PCIe Bus Error: severity=Corrected, type=Physical Layer, (Receiver ID)

Hi guys,I am getting this spammed in dmesg, system works perfectly but this is quite annoying cause TTY'sare full spam mode

[ 1722.172207] pcieport 0000:00:1c.0: AER: Corrected error received: 0000:00:1c.0
[ 1722.172214] pcieport 0000:00:1c.0: PCIe Bus Error: severity=Corrected, type=Physical Layer, (Receiver ID)
[ 1722.172219] pcieport 0000:00:1c.0:   device [8086:a110] error status/mask=00000001/00000000
[ 1722.172223] pcieport 0000:00:1c.0:    [ 0] RxErr                  (First)
[ 1722.376811] pcieport 0000:00:1c.0: AER: Corrected error received: 0000:00:1c.0
[ 1722.376817] pcieport 0000:00:1c.0: PCIe Bus Error: severity=Corrected, type=Physical Layer, (Receiver ID)
[ 1722.376821] pcieport 0000:00:1c.0:   device [8086:a110] error status/mask=00000001/00000000
[ 1722.376824] pcieport 0000:00:1c.0:    [ 0] RxErr                  (First)

Offline

#2 2018-11-23 21:30:02

seth
Member
Registered: 2012-09-03
Posts: 58,745

Re: PCIe Bus Error: severity=Corrected, type=Physical Layer, (Receiver ID)

You can suppress this w/ the pci=noaer kernel parameter, but your system does not work "perfectly", some device on that bus has communication issues, causing some error correction process.
Check what's on that bus and see whether there's some problem (bad seating, underpowered, dirt/dust, dead cable…)

Edit: if there's no device on the bus and the slots aren't compromised, this might be a BIOS bug as well.

Last edited by seth (2018-11-23 21:31:08)

Offline

#3 2018-11-28 02:10:46

blahhumbug
Member
Registered: 2016-10-08
Posts: 64

Re: PCIe Bus Error: severity=Corrected, type=Physical Layer, (Receiver ID)

I've had period where different kernels or bios versions have caused this more than others on my XPS 13.   

These are PCIe physical layer receiver errors.   PCIe is inherently designed to be a lossy protocol and so it can handle some amount of noise on the bus.   When those problems are detected, this error occurs, but all of the packets that were not received properly are NAKed by the link layer, and the packets get reissued.  This is why this is listed as a "Corrected error" and your system is still working.

You only need to worry about functionality when you See "Uncorrected error".

However, if you are getting a storm of these in your logs, then this means your PCIe bus is going to have lower bandwidth, because packets have to constantly be replayed.  If this is a desktop, you can determine which card is attached to this root-port, and try re-seating it.   Also, check for bios updates, as often the PCIe phy layers are tuned in post-silicon via bios configuration and so bios updates can improve PCIe reliability.

Edit: If your bios supports it, you can also try to force that root-port to run at a slower speed (usually the default is AUTO), then see if the errors remain.

Last edited by blahhumbug (2018-11-28 02:29:33)

Offline

Board footer

Powered by FluxBB