On Sat, Aug 25, 2012 at 05:59:44PM +0800, Yijing Wang wrote:
> Date:	Sat, 25 Aug 2012 17:59:44 +0800
> From: Yijing Wang <wangyijing@huawei.com>
> To: Bjorn Helgaas <bhelgaas@google.com>, Rusty Russell
>  <rusty@rustcorp.com.au>, Mauro Carvalho Chehab <mchehab@redhat.com>
> CC: PCI <linux-pci@vger.kernel.org>, Jiang Liu <liuj97@gmail.com>, Huang
>  Ying <ying.huang@intel.com>, Hanjun Guo <guohanjun@huawei.com>,
>  linux-kernel@vger.kernel.org
> Subject: [RESEND BUGFIX PATCH 1/3] PCI/AER: fix pci_ops return NULL when
>  hotplug a pci bus which was doing aer error inject
> User-Agent: Mozilla/5.0 (Windows NT 6.1; rv:14.0) Gecko/20120713
>  Thunderbird/14.0
> 
> When we inject aer errors to the target pci device by aer_inject module, the pci_ops of pci
> bus which the target device is on will be assign to pci_ops_aer.So if the target pci device
> is a bridge, once we hotplug the pci bus(child bus) which the target device bridges to, child
> bus's pci_ops will be assigned to pci_ops_aer too.Now every access to the child bus's device
> will result to system panic, because it return NULL pci_ops in pci_read_aer.
> This patch fix this.
> 
> CallTrace:
> bash[5908]: NaT consumption 17179869216 [1]
> Modules linked in: aer_inject cpufreq_conservative cpufreq_userspace cpufreq_pow
> ersave acpi_cpufreq binfmt_misc fuse nls_iso8859_1 loop ipmi_si(+) ipmi_devintf
> ipmi_msghandler dm_mod ppdev iTCO_wdt iTCO_vendor_support sg igb parport_pc i2c_
> i801 mptctl i2c_core serio_raw hid_generic lpc_ich mfd_core parport button conta
> iner usbhid hid uhci_hcd ehci_hcd usbcore usb_common sd_mod crc_t10dif ext3 mbca
> che jbd fan processor ide_pci_generic ide_core ata_piix libata mptsas mptscsih m
> ptbase scsi_transport_sas scsi_mod thermal thermal_sys hwmon
> 
[...]
> 
> Signed-off-by: Yijing Wang <wangyijing@huawei.com>
> Signed-off-by: Jiang Liu <liuj97@gmail.com>
> ---
>  drivers/pci/pcie/aer/aer_inject.c |   21 +++++++++++++++++++++
>  1 files changed, 21 insertions(+), 0 deletions(-)
> 
> diff --git a/drivers/pci/pcie/aer/aer_inject.c b/drivers/pci/pcie/aer/aer_inject.c
> index 5222986..fc28785 100644
> --- a/drivers/pci/pcie/aer/aer_inject.c
> +++ b/drivers/pci/pcie/aer/aer_inject.c
> @@ -109,6 +109,19 @@ static struct aer_error *__find_aer_error_by_dev(struct pci_dev *dev)
>  	return __find_aer_error((u16)domain, dev->bus->number, dev->devfn);
>  }
> 
> +static bool pci_is_upstream_bus(struct pci_bus *bus, struct pci_bus *up_bus)
> +{
> +	struct pci_bus *pbus = bus->parent;
> +
> +	while (pbus) {
> +		if (pbus == up_bus)
> +			return true;
> +		pbus = pbus->parent;
> +	}
> +
> +	return false;
> +}
> +
>  /* inject_lock must be held before calling */
>  static struct pci_ops *__find_pci_bus_ops(struct pci_bus *bus)
>  {
> @@ -118,6 +131,13 @@ static struct pci_ops *__find_pci_bus_ops(struct pci_bus *bus)
>  		if (bus_ops->bus == bus)
>  			return bus_ops->ops;
>  	}
> +
> +	/* here can't find bus_ops, fall back to get bus_ops of upstream bus */
> +	list_for_each_entry(bus_ops, &pci_bus_ops_list, list) {
> +		if (pci_is_upstream_bus(bus, bus_ops->bus))
> +			return bus_ops->ops;
> +	}
> +
>  	return NULL;
>  }
> 
At least, when returning NULL, a proper check and protection is needed.