[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <BYAPR21MB16882A3C717721090777D308D7899@BYAPR21MB1688.namprd21.prod.outlook.com>
Date: Wed, 29 Mar 2023 16:55:08 +0000
From: "Michael Kelley (LINUX)" <mikelley@...rosoft.com>
To: Dexuan Cui <decui@...rosoft.com>,
"bhelgaas@...gle.com" <bhelgaas@...gle.com>,
"davem@...emloft.net" <davem@...emloft.net>,
"edumazet@...gle.com" <edumazet@...gle.com>,
Haiyang Zhang <haiyangz@...rosoft.com>,
Jake Oshins <jakeo@...rosoft.com>,
"kuba@...nel.org" <kuba@...nel.org>, "kw@...ux.com" <kw@...ux.com>,
KY Srinivasan <kys@...rosoft.com>,
"leon@...nel.org" <leon@...nel.org>,
"linux-pci@...r.kernel.org" <linux-pci@...r.kernel.org>,
"lpieralisi@...nel.org" <lpieralisi@...nel.org>,
"pabeni@...hat.com" <pabeni@...hat.com>,
"robh@...nel.org" <robh@...nel.org>,
"saeedm@...dia.com" <saeedm@...dia.com>,
"wei.liu@...nel.org" <wei.liu@...nel.org>,
Long Li <longli@...rosoft.com>,
"boqun.feng@...il.com" <boqun.feng@...il.com>
CC: "linux-hyperv@...r.kernel.org" <linux-hyperv@...r.kernel.org>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
"linux-rdma@...r.kernel.org" <linux-rdma@...r.kernel.org>,
"netdev@...r.kernel.org" <netdev@...r.kernel.org>
Subject: RE: [PATCH 6/6] PCI: hv: Use async probing to reduce boot time
From: Dexuan Cui <decui@...rosoft.com>
>
> Commit 414428c5da1c ("PCI: hv: Lock PCI bus on device eject") added the
> pci_lock_rescan_remove() and pci_unlock_rescan_remove() to address the
> race between create_root_hv_pci_bus() and hv_eject_device_work(), but it
> doesn't really work well.
>
> Now with hbus->state_lock and other fixes, the race is resolved, so
> remove pci_{lock,unlock}_rescan_remove().
Commit 414428c5da1c added the calls to pci_lock/unlock_rescan_remove()
in both create_root_hv_pci_bus() and in hv_eject_device_work(). This patch
removes the calls only in create_reboot_hv_pci_bus(), but leaves them in
hv_eject_device_work(), in hv_pci_remove(), and in pci_devices_present_work().
So evidently it is still needed to provide exclusion for other cases. Perhaps your
commit message could clarify that only the exclusion of create_root_hv_pci_bus()
is now superfluous because of the hbus->state_lock. And commit 414428c5da1c
isn't fully reverted because evidently the lock is still needed in
hv_eject_device_work().
Michael
>
> Also enable async probing to reduce boot time.
>
> Signed-off-by: Dexuan Cui <decui@...rosoft.com>
> ---
> drivers/pci/controller/pci-hyperv.c | 11 +++++++++--
> 1 file changed, 9 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/pci/controller/pci-hyperv.c b/drivers/pci/controller/pci-hyperv.c
> index 2c0b86b20408..08ab389e27cc 100644
> --- a/drivers/pci/controller/pci-hyperv.c
> +++ b/drivers/pci/controller/pci-hyperv.c
> @@ -2312,12 +2312,16 @@ static int create_root_hv_pci_bus(struct
> hv_pcibus_device *hbus)
> if (error)
> return error;
>
> - pci_lock_rescan_remove();
> + /*
> + * pci_lock_rescan_remove() and pci_unlock_rescan_remove() are
> + * unnecessary here, because we hold the hbus->state_lock, meaning
> + * hv_eject_device_work() and pci_devices_present_work() can't race
> + * with create_root_hv_pci_bus().
> + */
> hv_pci_assign_numa_node(hbus);
> pci_bus_assign_resources(bridge->bus);
> hv_pci_assign_slots(hbus);
> pci_bus_add_devices(bridge->bus);
> - pci_unlock_rescan_remove();
> hbus->state = hv_pcibus_installed;
> return 0;
> }
> @@ -4003,6 +4007,9 @@ static struct hv_driver hv_pci_drv = {
> .remove = hv_pci_remove,
> .suspend = hv_pci_suspend,
> .resume = hv_pci_resume,
> + .driver = {
> + .probe_type = PROBE_PREFER_ASYNCHRONOUS,
> + },
> };
>
> static void __exit exit_hv_pci_drv(void)
> --
> 2.25.1
Powered by blists - more mailing lists