Re: [PATCH 1/2] PCI/portdrv: Add necessary delay for disabling hotplug events

[Feng Tang <feng.tang@linux.alibaba.com>]

On Thu, Feb 06, 2025 at 08:26:13PM -0800, Sathyanarayanan Kuppuswamy wrote:
> 
> On 2/5/25 7:18 PM, Feng Tang wrote:
> > Hi Sathyanarayanan,
> > 
> > On Wed, Feb 05, 2025 at 10:26:59AM -0800, Sathyanarayanan Kuppuswamy wrote:
> > > On 2/3/25 9:37 PM, Feng Tang wrote:
> > > > According to PCIe 6.1 spec, section 6.7.3.2, software need to wait at
> > > > least 1 second for the command-complete event, before resending the cmd
> > > > or sending a new cmd.
> > > > 
> > > > Currently get_port_device_capability() sends slot control cmd to disable
> > > > PCIe hotplug interrupts without waiting for its completion and there was
> > > > real problem reported for the lack of waiting.
> > > Can you include the error log associated with this issue? What is the
> > > actual issue you are seeing and in which hardware?
> > For this one, we don't have specific log, as it was raised by firmware
> > developer, as in https://lore.kernel.org/lkml/Z6LRAozZm1UfgjqT@U-2FWC9VHC-2323.local/
> > 
> > When handling PCI hotplug problem, they hit issue and found their state
> > machine corrupted , and back traced to OS. They didn't expect to receive
> > 2 link control commands at almost the same time, which doesn't comply to
> 
> Which 2 commands from OS? Did you identify both commands?

Firmware developer saw 2 programming to PCIE Slot Control register of
one root port, first is writing 0x5ca, the second is writing 0x15f8. 

IIUC, the first one is to disable CCIE/HPIE here from get_port_device_capability()

Thanks,
Feng