Re: [PATCH 5/6] PCI/ASPM: Actually configure the L1 substate settings to the device

[Bjorn Helgaas]

Hi James,

On Wed, Mar 08, 2017 at 06:44:36PM +0000, James Morse wrote:
> Hi!
> 
> On 03/01/17 06:34, Rajat Jain wrote:
> > Add code to actually configure the L1 substate settigns on the
> > upstream and downstream device, while taking care of the rules
> > dictated by the PCIe spec.
> 
> While testing hibernate on an arm64 juno with v4.11-rc1, I get a NULL pointer
> dereference from pcie_config_aspm_link():
> 
> 
> > diff --git a/drivers/pci/pcie/aspm.c b/drivers/pci/pcie/aspm.c
> > index a70afdf..6735f38 100644
> > --- a/drivers/pci/pcie/aspm.c
> > +++ b/drivers/pci/pcie/aspm.c
> > @@ -597,11 +683,23 @@ static void pcie_config_aspm_dev(struct pci_dev *pdev, u32 val)
> >  static void pcie_config_aspm_link(struct pcie_link_state *link, u32 state)
> >  {
> >  	u32 upstream = 0, dwstream = 0;
> > -	struct pci_dev *child, *parent = link->pdev;
> > +	struct pci_dev *child = link->downstream, *parent = link->pdev;
> 
> 
> Here link->downstream is NULL,

Sorry about the breakage.  Can you try also cherry-picking this
commit:

https://git.kernel.org/cgit/linux/kernel/git/helgaas/pci.git/commit/?h=for-linus&id=3bd7db63a841e8c5297bb18ad801df67d5e38ad2

Yinghai tripped over a similar problem in a different way, but I
suspect his fix might also fix the problem you're seeing.

If that doesn't fix it, we'll have to look farther.

I'm planning to ask Linus to pull this fix for v4.11-rc2.

Bjorn