Re: [PATCH v7 05/12] PCI: liveupdate: Keep bus numbers constant during Live Update

On Fri, Jul 10, 2026 at 09:26:08PM +0000, David Matlack wrote:
> During a Live Update, preserved devices must be allowed to continue
> performing memory transactions so the kernel cannot change the fabric
> topology, including bus numbers, since that would require disabling
> and flushing any memory transactions first.
> 
> To keep bus numbers constant, always inherit the secondary and
> subordinate bus numbers assigned to bridges during scanning, instead of
> assigning new ones, if any PCI devices are being preserved. Note that
> the kernel inherits bus numbers even on bridges without any downstream
> endpoints that were preserved. This avoids accidentally assigning a
> bridge a new window that overlaps with a preserved device that is
> downstream of a different bridge.

I know there are many references already, but I think "preserve"
conveys more of the meaning here than "inherit".

> + * pci_liveupdate_scan_bridge_begin() - Determine if a bridge should inherit bus numbers

Wrap to fit in 80 columns.

> + * @bus: The parent bus of the bridge.
> + * @dev: The PCI bridge device.
> + * @pass: The scan pass (0 for first pass, 1 for second pass).
> + *
> + * This function is called by the PCI core when it begins scanning a bridge.
> + * It determines whether the bridge should inherit the secondary and subordinate
> + * bus numbers assigned to it by the previous kernel. This is necessary to
> + * keep bus numbers constant for preserved devices downstream of the bridge.
> + *
> + * Return: True if bus numbers should be inherited, false otherwise.
> + */
> +bool pci_liveupdate_scan_bridge_begin(struct pci_bus *bus, struct pci_dev *dev,
> +				      int pass)

I see that you named it this way to have a begin/end pair, but I think
this function name should be a predicate, i.e., it should ask a
question with a yes/no or true/false answer.  Something like
"preserve_bus_numbers".

> + * pci_liveupdate_scan_bridge_end() - Finish scanning a PCI bridge
> + * @dev: The PCI bridge device.
> + * @pass: The scan pass (0 for first pass, 1 for second pass).
> + *
> + * This function is called by the PCI core when it finishes scanning a bridge.
> + * It clears the inheritance status after the second pass so it can be
> + * re-evaluated on future scans.
> + */
> +void pci_liveupdate_scan_bridge_end(struct pci_dev *dev, int pass)
> +{
> +	/*
> +	 * Clear inherit_buses after the second pass so it can be re-evaluated
> +	 * on future scans.
> +	 */
> +	if (pass)

I think the decision based on "pass" belongs in the caller because
it's really an implementation detail of the pci_scan_bridge*() code,
and I think it's important that readers of that code understand that
we no longer need to preserve bus numbers without having to follow the
chain into liveupdate.

I guess the point here is that we've always known to preserve bus
numbers of existing devices, so after we've enumerated all the
liveupdate-preserved devices, it's safe to assign new bus numbers for
hot-adds, bus rescans, etc.  Right?

> @@ -1397,6 +1397,7 @@ static int pci_scan_bridge_extend(struct pci_bus *bus, struct pci_dev *dev,
>  				  int max, unsigned int available_buses,
>  				  int pass)
>  {
> +	bool liveupdate, assign_new_buses = pcibios_assign_all_busses();

I have a pretty strong preference for non-negated booleans, but I also
think "assign_buses" is an unfortunate historical choice.  By default
Linux owns all resources, including bus numbers, so the exception case
is that we need to *preserve* things that were previously configured.
Especially in liveupdate, "assign buses" is the wrong sense -- the
requirement is that we preserve them.

So I think here a name like "preserve_buses" will express the intent
better even though it will lead to negation here and in tests of
"!preserve_buses" below.

>  	struct pci_bus *child;
>  	u32 buses;
>  	u16 bctl;
> @@ -1406,6 +1407,10 @@ static int pci_scan_bridge_extend(struct pci_bus *bus, struct pci_dev *dev,
>  	u8 fixed_sec, fixed_sub;
>  	int next_busnr;
>  
> +	liveupdate = pci_liveupdate_scan_bridge_begin(bus, dev, pass);
> +	if (liveupdate)
> +		assign_new_buses = false;
> +
>  	/*
>  	 * Make sure the bridge is powered on to be able to access config
>  	 * space of devices below it.
> @@ -1449,8 +1454,7 @@ static int pci_scan_bridge_extend(struct pci_bus *bus, struct pci_dev *dev,
>  		goto out;
>  	}
>  
> -	if ((secondary || subordinate) &&
> -	    !pcibios_assign_all_busses() && !broken) {
> +	if ((secondary || subordinate) && !assign_new_buses && !broken) {
>  		unsigned int cmax, buses;
>  
>  		/*
> @@ -1492,8 +1496,7 @@ static int pci_scan_bridge_extend(struct pci_bus *bus, struct pci_dev *dev,
>  		 * do in the second pass.
>  		 */
>  		if (!pass) {
> -			if (pcibios_assign_all_busses() || broken)
> -
> +			if (assign_new_buses || broken)
>  				/*
>  				 * Temporarily disable forwarding of the
>  				 * configuration cycles on all bridges in
> @@ -1507,6 +1510,11 @@ static int pci_scan_bridge_extend(struct pci_bus *bus, struct pci_dev *dev,
>  			goto out;
>  		}
>  
> +		if (liveupdate) {
> +			pci_err(dev, "Cannot reconfigure bridge during Live Update, skipping\n");
> +			goto out;
> +		}
> +
>  		/* Clear errors */
>  		pci_write_config_word(dev, PCI_STATUS, 0xffff);
>  
> @@ -1567,6 +1575,7 @@ static int pci_scan_bridge_extend(struct pci_bus *bus, struct pci_dev *dev,
>  	pci_write_config_word(dev, PCI_BRIDGE_CONTROL, bctl);
>  
>  	pm_runtime_put(&dev->dev);
> +	pci_liveupdate_scan_bridge_end(dev, pass);
>  
>  	return max;
>  }
> diff --git a/include/linux/pci_liveupdate.h b/include/linux/pci_liveupdate.h
> index cfdc3d62ec02..2be98819e313 100644
> --- a/include/linux/pci_liveupdate.h
> +++ b/include/linux/pci_liveupdate.h
> @@ -17,10 +17,14 @@
>   * struct pci_liveupdate - PCI Live Update state for a struct pci_dev
>   * @outgoing: State preserved for the next kernel.
>   * @incoming: State preserved by the previous kernel.
> + * @inherit_buses: True if the PCI core should inherit the secondary and
> + *                 subordinate bus numbers assigned to this device due to
> + *                 an ongoing Live Update.
>   */
>  struct pci_liveupdate {
>  	struct pci_dev_ser *outgoing;
>  	struct pci_dev_ser *incoming;
> +	bool inherit_buses;
>  };
>  
>  struct pci_dev;
> -- 
> 2.55.0.795.g602f6c329a-goog
>