RE: [Resend PATCH 2/2 v3] pci-hyperv: lock pci bus on device eject

RE: [Resend PATCH 2/2 v3] pci-hyperv: lock pci bus on device eject

[KY Srinivasan]

> -----Original Message-----
> From: Long Li [mailto:longli@exchange.microsoft.com]
> Sent: Monday, January 23, 2017 9:46 PM
> To: KY Srinivasan <kys@microsoft.com>; Haiyang Zhang
> <haiyangz@microsoft.com>; Bjorn Helgaas <bhelgaas@google.com>
> Cc: devel@linuxdriverproject.org; linux-pci@vger.kernel.org; linux-
> kernel@vger.kernel.org; Long Li <longli@microsoft.com>
> Subject: [Resend PATCH 2/2 v3] pci-hyperv: lock pci bus on device eject
> 
> [This sender failed our fraud detection checks and may not be who they
> appear to be. Learn about spoofing at http://aka.ms/LearnAboutSpoofing]
> 
> From: Long Li <longli@microsoft.com>
> 
> A PCI_EJECT message can arrive at the same time we are calling
> pci_scan_child_bus in the workqueue for the previous PCI_BUS_RELATIONS
> message or in create_root_hv_pci_bus(), in this case we could potentailly
> modify the bus from multiple places. Properly lock the bus access.
> 
> Thanks Dexuan Cui <decui@microsoft.com> for pointing out the race
> condition in create_root_hv_pci_bus().
> 
> Signed-off-by: Long Li <longli@microsoft.com>
> Reported-by: Xiaofeng Wang <xiaofwan@redhat.com>

Acked-by: K. Y. Srinivasan <kys@microsoft.com>

> ---
>  drivers/pci/host/pci-hyperv.c | 4 ++++
>  1 file changed, 4 insertions(+)
> 
> diff --git a/drivers/pci/host/pci-hyperv.c b/drivers/pci/host/pci-hyperv.c
> index 4a37598..33c75c9 100644
> --- a/drivers/pci/host/pci-hyperv.c
> +++ b/drivers/pci/host/pci-hyperv.c
> @@ -1198,9 +1198,11 @@ static int create_root_hv_pci_bus(struct
> hv_pcibus_device *hbus)
>         hbus->pci_bus->msi = &hbus->msi_chip;
>         hbus->pci_bus->msi->dev = &hbus->hdev->device;
> 
> +       pci_lock_rescan_remove();
>         pci_scan_child_bus(hbus->pci_bus);
>         pci_bus_assign_resources(hbus->pci_bus);
>         pci_bus_add_devices(hbus->pci_bus);
> +       pci_unlock_rescan_remove();
>         hbus->state = hv_pcibus_installed;
>         return 0;
>  }
> @@ -1590,8 +1592,10 @@ static void hv_eject_device_work(struct
> work_struct *work)
>         pdev = pci_get_domain_bus_and_slot(hpdev->hbus->sysdata.domain,
> 0,
>                                            wslot);
>         if (pdev) {
> +               pci_lock_rescan_remove();
>                 pci_stop_and_remove_bus_device(pdev);
>                 pci_dev_put(pdev);
> +               pci_unlock_rescan_remove();
>         }
> 
>         memset(&ctxt, 0, sizeof(ctxt));
> --
> 1.8.5.6

RE: [Resend PATCH 2/2 v3] pci-hyperv: lock pci bus on device eject

[KY Srinivasan]

> -----Original Message-----
> From: Long Li [mailto:longli@exchange.microsoft.com]
> Sent: Monday, January 23, 2017 9:46 PM
> To: KY Srinivasan <kys@microsoft.com>; Haiyang Zhang
> <haiyangz@microsoft.com>; Bjorn Helgaas <bhelgaas@google.com>
> Cc: devel@linuxdriverproject.org; linux-pci@vger.kernel.org; linux-
> kernel@vger.kernel.org; Long Li <longli@microsoft.com>
> Subject: [Resend PATCH 2/2 v3] pci-hyperv: lock pci bus on device eject
>=20
> [This sender failed our fraud detection checks and may not be who they
> appear to be. Learn about spoofing at http://aka.ms/LearnAboutSpoofing]
>=20
> From: Long Li <longli@microsoft.com>
>=20
> A PCI_EJECT message can arrive at the same time we are calling
> pci_scan_child_bus in the workqueue for the previous PCI_BUS_RELATIONS
> message or in create_root_hv_pci_bus(), in this case we could potentailly
> modify the bus from multiple places. Properly lock the bus access.
>=20
> Thanks Dexuan Cui <decui@microsoft.com> for pointing out the race
> condition in create_root_hv_pci_bus().
>=20
> Signed-off-by: Long Li <longli@microsoft.com>
> Reported-by: Xiaofeng Wang <xiaofwan@redhat.com>

Acked-by: K. Y. Srinivasan <kys@microsoft.com>

> ---
>  drivers/pci/host/pci-hyperv.c | 4 ++++
>  1 file changed, 4 insertions(+)
>=20
> diff --git a/drivers/pci/host/pci-hyperv.c b/drivers/pci/host/pci-hyperv.=
c
> index 4a37598..33c75c9 100644
> --- a/drivers/pci/host/pci-hyperv.c
> +++ b/drivers/pci/host/pci-hyperv.c
> @@ -1198,9 +1198,11 @@ static int create_root_hv_pci_bus(struct
> hv_pcibus_device *hbus)
>         hbus->pci_bus->msi =3D &hbus->msi_chip;
>         hbus->pci_bus->msi->dev =3D &hbus->hdev->device;
>=20
> +       pci_lock_rescan_remove();
>         pci_scan_child_bus(hbus->pci_bus);
>         pci_bus_assign_resources(hbus->pci_bus);
>         pci_bus_add_devices(hbus->pci_bus);
> +       pci_unlock_rescan_remove();
>         hbus->state =3D hv_pcibus_installed;
>         return 0;
>  }
> @@ -1590,8 +1592,10 @@ static void hv_eject_device_work(struct
> work_struct *work)
>         pdev =3D pci_get_domain_bus_and_slot(hpdev->hbus->sysdata.domain,
> 0,
>                                            wslot);
>         if (pdev) {
> +               pci_lock_rescan_remove();
>                 pci_stop_and_remove_bus_device(pdev);
>                 pci_dev_put(pdev);
> +               pci_unlock_rescan_remove();
>         }
>=20
>         memset(&ctxt, 0, sizeof(ctxt));
> --
> 1.8.5.6

[Resend PATCH 2/2 v3] pci-hyperv: lock pci bus on device eject

[Long Li]

A PCI_EJECT message can arrive at the same time we are calling pci_scan_child_bus in the workqueue for the previous PCI_BUS_RELATIONS message or in create_root_hv_pci_bus(), in this case we could potentailly modify the bus from multiple places. Properly lock the bus access.

Thanks Dexuan Cui <decui@microsoft.com> for pointing out the race condition in create_root_hv_pci_bus().

Signed-off-by: Long Li <longli@microsoft.com>
Reported-by: Xiaofeng Wang <xiaofwan@redhat.com>
Acked-by: K. Y. Srinivasan <kys@microsoft.com>
---
 drivers/pci/host/pci-hyperv.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/drivers/pci/host/pci-hyperv.c b/drivers/pci/host/pci-hyperv.c
index 4a37598..33c75c9 100644
--- a/drivers/pci/host/pci-hyperv.c
+++ b/drivers/pci/host/pci-hyperv.c
@@ -1198,9 +1198,11 @@ static int create_root_hv_pci_bus(struct hv_pcibus_device *hbus)
 	hbus->pci_bus->msi = &hbus->msi_chip;
 	hbus->pci_bus->msi->dev = &hbus->hdev->device;
 
+	pci_lock_rescan_remove();
 	pci_scan_child_bus(hbus->pci_bus);
 	pci_bus_assign_resources(hbus->pci_bus);
 	pci_bus_add_devices(hbus->pci_bus);
+	pci_unlock_rescan_remove();
 	hbus->state = hv_pcibus_installed;
 	return 0;
 }
@@ -1590,8 +1592,10 @@ static void hv_eject_device_work(struct work_struct *work)
 	pdev = pci_get_domain_bus_and_slot(hpdev->hbus->sysdata.domain, 0,
 					   wslot);
 	if (pdev) {
+		pci_lock_rescan_remove();
 		pci_stop_and_remove_bus_device(pdev);
 		pci_dev_put(pdev);
+		pci_unlock_rescan_remove();
 	}
 
 	memset(&ctxt, 0, sizeof(ctxt));
-- 
1.8.5.6

[Resend PATCH 2/2 v3] pci-hyperv: lock pci bus on device eject

[Long Li]

A PCI_EJECT message can arrive at the same time we are calling pci_scan_chi=
ld_bus in the workqueue for the previous PCI_BUS_RELATIONS message or in cr=
eate_root_hv_pci_bus(), in this case we could potentailly modify the bus fr=
om multiple places. Properly lock the bus access.

Thanks Dexuan Cui <decui@microsoft.com> for pointing out the race condition=
 in create_root_hv_pci_bus().

Signed-off-by: Long Li <longli@microsoft.com>
Reported-by: Xiaofeng Wang <xiaofwan@redhat.com>
Acked-by: K. Y. Srinivasan <kys@microsoft.com>
---
 drivers/pci/host/pci-hyperv.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/drivers/pci/host/pci-hyperv.c b/drivers/pci/host/pci-hyperv.c
index 4a37598..33c75c9 100644
--- a/drivers/pci/host/pci-hyperv.c
+++ b/drivers/pci/host/pci-hyperv.c
@@ -1198,9 +1198,11 @@ static int create_root_hv_pci_bus(struct hv_pcibus_d=
evice *hbus)
 	hbus->pci_bus->msi =3D &hbus->msi_chip;
 	hbus->pci_bus->msi->dev =3D &hbus->hdev->device;
=20
+	pci_lock_rescan_remove();
 	pci_scan_child_bus(hbus->pci_bus);
 	pci_bus_assign_resources(hbus->pci_bus);
 	pci_bus_add_devices(hbus->pci_bus);
+	pci_unlock_rescan_remove();
 	hbus->state =3D hv_pcibus_installed;
 	return 0;
 }
@@ -1590,8 +1592,10 @@ static void hv_eject_device_work(struct work_struct =
*work)
 	pdev =3D pci_get_domain_bus_and_slot(hpdev->hbus->sysdata.domain, 0,
 					   wslot);
 	if (pdev) {
+		pci_lock_rescan_remove();
 		pci_stop_and_remove_bus_device(pdev);
 		pci_dev_put(pdev);
+		pci_unlock_rescan_remove();
 	}
=20
 	memset(&ctxt, 0, sizeof(ctxt));
--=20
1.8.5.6

Re: [Resend PATCH 2/2 v3] pci-hyperv: lock pci bus on device eject

[Greg KH]

On Fri, Feb 24, 2017 at 09:49:17PM +0000, Long Li wrote:
> A PCI_EJECT message can arrive at the same time we are calling pci_scan_child_bus in the workqueue for the previous PCI_BUS_RELATIONS message or in create_root_hv_pci_bus(), in this case we could potentailly modify the bus from multiple places. Properly lock the bus access.

Properly wrap your changelog comments at 72 columns like your editor is
telling you to do...

RE: [Resend PATCH 2/2 v3] pci-hyperv: lock pci bus on device eject

[Long Li]

Ok, I will resend.

> -----Original Message-----
> From: Greg KH [mailto:greg@kroah.com]
> Sent: Saturday, February 25, 2017 12:02 AM
> To: Long Li <longli@microsoft.com>
> Cc: KY Srinivasan <kys@microsoft.com>; Haiyang Zhang
> <haiyangz@microsoft.com>; Bjorn Helgaas <bhelgaas@google.com>;
> devel@linuxdriverproject.org; linux-kernel@vger.kernel.org; linux-
> pci@vger.kernel.org
> Subject: Re: [Resend PATCH 2/2 v3] pci-hyperv: lock pci bus on device eject
> 
> On Fri, Feb 24, 2017 at 09:49:17PM +0000, Long Li wrote:
> > A PCI_EJECT message can arrive at the same time we are calling
> pci_scan_child_bus in the workqueue for the previous PCI_BUS_RELATIONS
> message or in create_root_hv_pci_bus(), in this case we could potentailly
> modify the bus from multiple places. Properly lock the bus access.
> 
> Properly wrap your changelog comments at 72 columns like your editor is
> telling you to do...

RE: [Resend PATCH 2/2 v3] pci-hyperv: lock pci bus on device eject

[Long Li]

Ok, I will resend.

> -----Original Message-----
> From: Greg KH [mailto:greg@kroah.com]
> Sent: Saturday, February 25, 2017 12:02 AM
> To: Long Li <longli@microsoft.com>
> Cc: KY Srinivasan <kys@microsoft.com>; Haiyang Zhang
> <haiyangz@microsoft.com>; Bjorn Helgaas <bhelgaas@google.com>;
> devel@linuxdriverproject.org; linux-kernel@vger.kernel.org; linux-
> pci@vger.kernel.org
> Subject: Re: [Resend PATCH 2/2 v3] pci-hyperv: lock pci bus on device eje=
ct
>=20
> On Fri, Feb 24, 2017 at 09:49:17PM +0000, Long Li wrote:
> > A PCI_EJECT message can arrive at the same time we are calling
> pci_scan_child_bus in the workqueue for the previous PCI_BUS_RELATIONS
> message or in create_root_hv_pci_bus(), in this case we could potentailly
> modify the bus from multiple places. Properly lock the bus access.
>=20
> Properly wrap your changelog comments at 72 columns like your editor is
> telling you to do...