Octavio,
On Sat, 22 Jun 2019, Thomas Gleixner wrote:
> On Wed, 19 Jun 2019, Octavio Alvarez wrote:
> > On 6/13/19 3:45 PM, Thomas Gleixner wrote:
> > > Can you please provide the content of /proc/interrupts with the driver
> > > loaded and working after boot (don't hibernate) for the following kernels:
> > >
> >
> > $ cat linux-master-after-boot.txt
> > CPU0 CPU1 CPU2 CPU3
>
> > 27: 1 0 0 0 PCI-MSI 3145728-edge eth0
>
> > > Linus upstream + revert
> >
> > $ cat linux-master-reverted-after-boot.txt
> > CPU0 CPU1 CPU2 CPU3
>
> > 27: 1 0 0 0 PCI-MSI 3145728-edge eth0
>
> > Meanwhile, here it is for 4.9, which is the latest Debian-provided kernel
> > and
> > worked:
> >
> > $ cat linux-4.9-after-boot.txt
> > CPU0 CPU1 CPU2 CPU3
>
> > 24: 1 0 0 0 PCI-MSI 3145728-edge eth0
>
> > I will keep trying 4.14, unless you say otherwise.
>
> It would be interesting though I don't expect too much data.
>
> So all of the above use PCI/MSI. That's at least a data point. I need to
> stare into that driver again to figure out why this might make a
> difference, but right now I'm lost.
One other data point you could provide please:
Load the driver on Linus master with the following module parameter:
disable_msi=1
That switches to INTx usage. Does the machine resume proper with that?
Thanks,
tglx
