Re: boot interrupt quirk (also in 4.19.y) breaks serial ports (was: [PATCH v2 0/2] pci: Add boot interrupt quirk mechanism for Xeon chipsets)

[Bjorn Helgaas]

On Fri, Sep 23, 2022 at 09:20:30PM +0200, Grzegorz Halat wrote:
> On Wed, Sep 16 2020 at 12:12, Stefan Bühler wrote:
> > this quirk breaks our serial ports PCIe card (i.e. we don't see any
> > output from the connected devices; no idea whether anything we send
> > reaches them):
> 
> I have the same problem, also with a PCI serial adapter from Oxford
> Semiconductor.  I've bisected the kernel and it was introduced in
> b88bf6c3b6ff.  When the system is booted with "pci=noioapicquirk"
> then the PCI card works fine.  The CPU is Intel Xeon E5-2680 v3 @
> 2.50GHz.
> 
> Sean, do you have any news about this issue?

I think Sean has moved on and isn't available to work this issue.

To help whoever *can* work on it, would you mind attaching the
complete dmesg log and "sudo lspci -vv" output for your system?  Then
we can potentially tweak the quirk to not break your system.

Bjorn