Re: [PATCH] pci: check PCI_EXP_FLAGS_SLOT before setting hotplug bridge

[Adam Lee <adam.lee@canonical.com>]

On Tue, Nov 19, 2013 at 11:08:06AM -0700, Myron Stowe wrote:
> On Mon, Nov 18, 2013 at 10:57 PM, Adam Lee <adam.lee@canonical.com> wrote:
> > On Mon, Nov 18, 2013 at 10:38:17AM -0700, Bjorn Helgaas wrote:
> >> [+cc Myron, Amos, Thomas, Ben]
> >>
> >> On Mon, Nov 18, 2013 at 2:40 AM, Adam Lee <adam.lee@canonical.com> wrote:
> >> > This patch adds the PCI_EXP_FLAGS_SLOT check back before setting
> >> > hotplug bridge, which is omitted by an API switching commit,
> >> > 59875ae489609b2267548dc85160c5f0f0c6f9d4 "PCI/core: Use PCI Express
> >> > Capability accessors".
> >> >
> >> > Some Lenovo laptops hang in booting without this fix.
> >>
> >> What kernel version hangs?  I suspect you might be missing 6d3a1741f1
> >> ("PCI: Support PCIe Capability Slot registers only for ports with
> >> slots"), because it *looks* like the current kernel should work
> >> correctly even without your patch.
> >
> > No, patching 6d3a1741f1 and d3694d4fa3 doesn't fix the hang.
> 
> Did you actually try it?  This patch did solve a problem that I
> encountered (see below).

On Tue, Nov 19, 2013 at 10:40:09AM -0700, Bjorn Helgaas wrote:
> >
> > It hangs in acpi_evaluate_integer() from
> > 59875ae489609b2267548dc85160c5f0f0c6f9d4 "PCI/core: Use PCI Express
> > Capability accessors" and before
> > ac212b6980d8d5eda705864fc5a8ecddc6d6eacc "ACPI / processor: Use common
> > hotplug infrastructure", 3.4~3.11. (double confirmed)
                             ^^^^^^^^ should be 3.7~3.10

> I don't understand what you're saying here.  We're talking about this path:
>
>     set_pcie_hotplug_bridge
>       pcie_capability_read_dword(dev, PCI_EXP_SLTCAP, ...)
>         pcie_capability_reg_implemented(dev, PCI_EXP_SLTCAP)
>           pcie_cap_has_sltctl(dev)
>
> This path doesn't invoke acpi_evaluate_integer().  How does a hang
> there relate to the patch you posted?  Are you saying you bisected the
> hang to one of the commits you mentioned?

Bjorn and Myron,

Yes, I did try, and those two commits are both coming from bisect, it is weird.
But good news, I tried the new series containing three patches, it fixed that
hang.(backported to 3.8, pcie_flags_reg is fine)

Great thanks.

-- 
Adam Lee