How to prevent initialization of Sound Blaster AE-9 that prevents boot

deltaecho · 2024-10-30 22:59:34

I recently purchased a Sound Blaster AE-9 sound card and am attempting to get it to work on Linux. I am aware that there is no kernel support for it, but found a potential workaround that seems reasonable. I am able to boot into Windows and use it without problems, so the issue is not the card, but when I try to boot into Linux the loader hangs with messages like the following:

[    4.847093] pcieport 0000:00:03.2: PCIe Bus Error: severity=Uncorrectable (Non-Fatal), type=Transaction Layer, (Requester ID)
[    4.847155] pcieport 0000:00:03.2:   device [8086:2f0a] error status/mask=00004000/00000000
[    4.847200] pcieport 0000:00:03.2:   [14] CmpltTO                (First)
pcieport 0000:00:03.2: PCIe Bus Error: severity=Uncorrectable (Non-Fatal), type=Transaction Layer, (Requester ID)
pcieport 0000:00:03.2:   device [8086:2f0a] error status/mask=00004000/00000000
pcieport 0000:00:03.2:   [14] CmpltTO                (First)
[ ***  ] (2 of 2) A start job is running for Load Kernel Modules (4s / 1min 29s)
[   10.345547] mce: CPUs not responding to MCE broadcast (may include false positives): 0
[   10.345547] mce: CPUs not responding to MCE broadcast (may include false positives): 0
[   10.345552] Kernel panic - not syncing: Timeout: Not all CPUs entered broadcast exception handler
[   11.374142] Shutting down cpus with NMI
[   11.384440] Kernel Offset: 0x13200000 from 0xffffffff81000000 (relocation range: 0xffffffff80000000-0xffffffffbfffffff)

It looks to me like CPU 0 is attempting to initialize the sound card, identified by 8086:2f0a, and fails which causes the respective process to crash and results in a timeout. That, or the failure results in a race condition. Is there a way to disable the initialization of the sound card by Linux?

cryptearth · 2024-10-31 00:02:12

vendor-id 8086 is - as it should be easy to guess - intel - also this error occurs on 0000:00:03.2 - which is far from any pci-e plug-in slot but some of the motherboard components - or for short: this looks like an issue with your on-board audio
sanity check: do you get this error when you pull out the soundblaster?
also: why you gone for a plug-in card instead of a pure usb-only external interface?

deltaecho · 2024-10-31 03:28:30

I do not get the error when I pull the sound card out, it boots normally. On a similar note, the same error occurs with the card installed when I try to boot to the live USB. I read about some incompatibilities between the snd-hda-intel module and AE cards, so I blacklisted it but still get this error. I will check my BIOS settings to see if there is a way to disable the on-board audio.

As for why I went with an extension card, it was mostly for sentimentality. My dad bought me a sound blaster card for my first computer almost 30 years ago and it meant a lot to him. I was thinking about him when I got this one and settled on the AE-9 because I was able to get it for the price of an AE-5. I have my dad's Polk monitors and an upgraded Onkyo amplifier, and the difference in sound when paired with this card is incredible. They aren't Sennheisers but they sound magnitudes greater than most aftermarket speakers you can buy these days.

cryptearth · 2024-10-31 06:00:30

according to https://unix.stackexchange.com/question … er-bad-tlp you could try

pci=nommconf

as a kernel option
as for the reason: could be a resource conflict (on-board audio can be disabled on most systems) or some hardware/firmware bug
could also be a linux issue deep inside some low-level stuff

seth · 2024-10-31 08:24:31

"8086:2f0a" is the "Xeon E7 v3/Xeon E5 v3/Core i7 PCI Express Root Port 3" as this is a bus error. No surprises there, but I'm also not confident that this is the actual problem.

I am able to boot into Windows and use it without problems

3rd link below. Mandatory.
Disable it (it's NOT the BIOS setting!) and reboot windows and linux twice for voodo reasons.

You can suck away the device by adding "pci_stub.ids=$vid:$pid" to the https://wiki.archlinux.org/title/Kernel_parameters or likewise pass it through to a VM, https://wiki.archlinux.org/title/PCI_pa … h_via_OVMF

But I cannot seem to find the pci id for the device, creatives vendor ID is "1102", based on https://forum.manjaro.org/t/system-unre … /103027/15 the product id might be 0403, in that case it would be "pci_stub.ids=1102:0403" (if you can boot w/ nommconf you should be able to identify your device)

deltaecho · 2024-10-31 13:06:58

Thank you both! From the Windows device manager, it looks like the device ID is 0011, so I will first try to boot into Linux with

pci=nommconf

to confirm the ID and see if that will work around the current failure, and then I will try adding

pci_stubs.ids=1102:0011

to my kernel parameters if Linux reports the same ID.

Edit: Fixed typo in device specification.

Last edited by deltaecho (2024-10-31 13:25:24)

seth · 2024-10-31 13:17:49

pci_stub.ids=1102:0011

Computers are kinda sticklers for orthography

deltaecho · 2024-10-31 13:25:02

That was a typo on my part, thank you for catching it. I'll report my success shortly.

gromit · 2024-10-31 14:52:47

Also maybe check if you get the boot lockup with an older kernel aswell, as far as I understand this should not happen ..

deltaecho · 2024-11-01 13:34:35

Neither kernel parameter allowed me to boot. Is it acceptable to stub the PCIe port? If so, what is the required syntax? The device manager in Windows shows 3 devices related to my SB card: Sound Blaster AE-9, Sound Blaster AE-9s, and Sound Blaster Audio Controller. The device IDs for the first two are the same (0011), but the device ID for the audio controller is 0010. Should I stub both 0011 and 0010?

deltaecho · 2024-11-01 13:35:19

@gromit, I will try to find an older kernel to try as well.

seth · 2024-11-01 16:09:49

Should I stub both 0011 and 0010?

Yes.
If you're https://wiki.archlinux.org/title/Genera … l_messages do you get some hints at what stalls or fails during the boot?

loqs · 2024-11-01 16:18:15

[   10.345552] Kernel panic - not syncing: Timeout: Not all CPUs entered broadcast exception handler

Could one CPU be so busy handling PCI bus errors it is unable to sync in time?

seth · 2024-11-01 17:04:34

Shouldn't™ we see way more pci errors then?
@deltaecho, what if you add

pci=noaer,nomsi,noacpi,nobar,nommconf

as kernel parameter?
(The system will not work "properly", but maybe boot…)

Arch Linux

#1 2024-10-30 22:59:34

How to prevent initialization of Sound Blaster AE-9 that prevents boot

#2 2024-10-31 00:02:12

Re: How to prevent initialization of Sound Blaster AE-9 that prevents boot

#3 2024-10-31 03:28:30

Re: How to prevent initialization of Sound Blaster AE-9 that prevents boot

#4 2024-10-31 06:00:30

Re: How to prevent initialization of Sound Blaster AE-9 that prevents boot

#5 2024-10-31 08:24:31

Re: How to prevent initialization of Sound Blaster AE-9 that prevents boot

#6 2024-10-31 13:06:58

Re: How to prevent initialization of Sound Blaster AE-9 that prevents boot

#7 2024-10-31 13:17:49

Re: How to prevent initialization of Sound Blaster AE-9 that prevents boot

#8 2024-10-31 13:25:02

Re: How to prevent initialization of Sound Blaster AE-9 that prevents boot

#9 2024-10-31 14:52:47

Re: How to prevent initialization of Sound Blaster AE-9 that prevents boot

#10 2024-11-01 13:34:35

Re: How to prevent initialization of Sound Blaster AE-9 that prevents boot

#11 2024-11-01 13:35:19

Re: How to prevent initialization of Sound Blaster AE-9 that prevents boot

#12 2024-11-01 16:09:49

Re: How to prevent initialization of Sound Blaster AE-9 that prevents boot

#13 2024-11-01 16:18:15

Re: How to prevent initialization of Sound Blaster AE-9 that prevents boot

#14 2024-11-01 17:04:34

Re: How to prevent initialization of Sound Blaster AE-9 that prevents boot

Board footer