From: Bjorn Helgaas <helgaas@kernel.org>
To: David Matlack <dmatlack@google.com>
Cc: kexec@lists.infradead.org, linux-doc@vger.kernel.org,
linux-kernel@vger.kernel.org, linux-mm@kvack.org,
linux-pci@vger.kernel.org,
Adithya Jayachandran <ajayachandra@nvidia.com>,
Alexander Graf <graf@amazon.com>,
Alex Williamson <alex@shazbot.org>,
Bjorn Helgaas <bhelgaas@google.com>, Chris Li <chrisl@kernel.org>,
David Rientjes <rientjes@google.com>,
Jacob Pan <jacob.pan@linux.microsoft.com>,
Jason Gunthorpe <jgg@nvidia.com>,
Jonathan Corbet <corbet@lwn.net>, Josh Hilke <jrhilke@google.com>,
Leon Romanovsky <leonro@nvidia.com>,
Lukas Wunner <lukas@wunner.de>, Mike Rapoport <rppt@kernel.org>,
Parav Pandit <parav@nvidia.com>,
Pasha Tatashin <pasha.tatashin@soleen.com>,
Pranjal Shrivastava <praan@google.com>,
Pratyush Yadav <pratyush@kernel.org>,
Saeed Mahameed <saeedm@nvidia.com>,
Samiullah Khawaja <skhawaja@google.com>,
Shuah Khan <skhan@linuxfoundation.org>,
Vipin Sharma <vipinsh@google.com>, William Tu <witu@nvidia.com>,
Yi Liu <yi.l.liu@intel.com>
Subject: Re: [PATCH v7 05/12] PCI: liveupdate: Keep bus numbers constant during Live Update
Date: Thu, 23 Jul 2026 18:01:27 -0500 [thread overview]
Message-ID: <20260723230127.GA869843@bhelgaas> (raw)
In-Reply-To: <20260710212616.1351130-6-dmatlack@google.com>
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
>
next prev parent reply other threads:[~2026-07-23 23:01 UTC|newest]
Thread overview: 47+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-07-10 21:26 [PATCH v7 00/12] PCI: liveupdate: PCI core support for Live Update David Matlack
2026-07-10 21:26 ` [PATCH v7 01/12] PCI: liveupdate: Set up FLB handler for the PCI core David Matlack
2026-07-17 19:28 ` Pasha Tatashin
2026-07-17 19:30 ` Jason Gunthorpe
2026-07-17 19:40 ` Pasha Tatashin
2026-07-17 19:42 ` Pasha Tatashin
2026-07-23 22:21 ` Bjorn Helgaas
2026-07-10 21:26 ` [PATCH v7 02/12] PCI: liveupdate: Track outgoing preserved PCI devices David Matlack
2026-07-17 19:38 ` Pasha Tatashin
2026-07-23 22:29 ` Bjorn Helgaas
2026-07-10 21:26 ` [PATCH v7 03/12] PCI: liveupdate: Track incoming " David Matlack
2026-07-17 21:46 ` Pasha Tatashin
2026-07-20 21:54 ` David Matlack
2026-07-20 22:44 ` Pasha Tatashin
2026-07-20 23:07 ` David Matlack
2026-07-21 17:55 ` Pasha Tatashin
2026-07-21 20:25 ` David Matlack
2026-07-21 21:35 ` Pasha Tatashin
2026-07-21 23:02 ` Samiullah Khawaja
2026-07-21 23:16 ` David Matlack
2026-07-21 23:18 ` David Matlack
2026-07-21 23:46 ` Pasha Tatashin
2026-07-22 17:00 ` David Matlack
2026-07-17 22:38 ` Alex Williamson
2026-07-20 22:00 ` David Matlack
2026-07-22 16:10 ` David Matlack
2026-07-10 21:26 ` [PATCH v7 04/12] PCI: liveupdate: Document driver binding responsibilities David Matlack
2026-07-10 21:26 ` [PATCH v7 05/12] PCI: liveupdate: Keep bus numbers constant during Live Update David Matlack
2026-07-17 21:48 ` Pasha Tatashin
2026-07-23 23:01 ` Bjorn Helgaas [this message]
2026-07-10 21:26 ` [PATCH v7 06/12] PCI: liveupdate: Auto-preserve upstream bridges across " David Matlack
2026-07-17 22:00 ` Pasha Tatashin
2026-07-10 21:26 ` [PATCH v7 07/12] PCI: Refactor matching logic for pci_dev_acs_ops David Matlack
2026-07-17 22:02 ` Pasha Tatashin
2026-07-10 21:26 ` [PATCH v7 08/12] PCI: liveupdate: Inherit ACS flags in incoming preserved devices David Matlack
2026-07-17 22:08 ` Pasha Tatashin
2026-07-10 21:26 ` [PATCH v7 09/12] PCI: liveupdate: Inherit ARI Forwarding Enable on preserved bridges David Matlack
2026-07-17 23:29 ` Pasha Tatashin
2026-07-20 23:19 ` David Matlack
2026-07-21 18:17 ` Pasha Tatashin
2026-07-21 20:26 ` David Matlack
2026-07-10 21:26 ` [PATCH v7 10/12] PCI: liveupdate: Freeze preservation status during shutdown David Matlack
2026-07-17 23:30 ` Pasha Tatashin
2026-07-10 21:26 ` [PATCH v7 11/12] PCI: liveupdate: Do not disable bus mastering on preserved devices during kexec David Matlack
2026-07-17 23:32 ` Pasha Tatashin
2026-07-10 21:26 ` [PATCH v7 12/12] Documentation: PCI: Add documentation for Live Update David Matlack
2026-07-17 23:38 ` Pasha Tatashin
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20260723230127.GA869843@bhelgaas \
--to=helgaas@kernel.org \
--cc=ajayachandra@nvidia.com \
--cc=alex@shazbot.org \
--cc=bhelgaas@google.com \
--cc=chrisl@kernel.org \
--cc=corbet@lwn.net \
--cc=dmatlack@google.com \
--cc=graf@amazon.com \
--cc=jacob.pan@linux.microsoft.com \
--cc=jgg@nvidia.com \
--cc=jrhilke@google.com \
--cc=kexec@lists.infradead.org \
--cc=leonro@nvidia.com \
--cc=linux-doc@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=linux-pci@vger.kernel.org \
--cc=lukas@wunner.de \
--cc=parav@nvidia.com \
--cc=pasha.tatashin@soleen.com \
--cc=praan@google.com \
--cc=pratyush@kernel.org \
--cc=rientjes@google.com \
--cc=rppt@kernel.org \
--cc=saeedm@nvidia.com \
--cc=skhan@linuxfoundation.org \
--cc=skhawaja@google.com \
--cc=vipinsh@google.com \
--cc=witu@nvidia.com \
--cc=yi.l.liu@intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox