linux-kernel - Re: [RESEND BUGFIX PATCH 3/3] PCI: check whether pci device has been removed when remove a pci device by sysfs

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <503B16DD.8080900@huawei.com>
Date:	Mon, 27 Aug 2012 14:42:37 +0800
From:	Yijing Wang <wangyijing@...wei.com>
To:	Jiang Liu <liuj97@...il.com>
CC:	Bjorn Helgaas <bhelgaas@...gle.com>,
	PCI <linux-pci@...r.kernel.org>, <linux-kernel@...r.kernel.org>
Subject: Re: [RESEND BUGFIX PATCH 3/3] PCI: check whether pci device has been
 removed when remove a pci device by sysfs

On 2012/8/25 22:39, Jiang Liu wrote:
> Hi Yijing,
> 	The patch only patially fix the issue, there exists still small race
> condition window because pdev->is_added isn't a reliable flag to depend on.
> 	--Gerry
> 

Hi Gerry,
    You are right, add pdev->is_added flag check here only fix the race condition window
between remove and rescan sysfs interfaces. Maybe we need a more comprehensive solution to
solve these problems between hotplug/remove/rescan actions.Next, I will do a more detailed test for
[RFC PATCH v1 00/22] introduce PCI bus lock to serialize PCI hotplug operations patches.
That is a more comprehensive solution actually.

---------
Thanks.
Yijing

> On 08/25/2012 05:59 PM, Yijing Wang wrote:
>> We remove a pci device maybe like this
>> echo 1 > /sys/bus/pci/devices/xxxx:xx:xx.x/remove
>> Then remove_store function will be called to complete this remove work,
>> later the remove work will be queued to sysfs_workqueue by device_schedule_callback.
>> So if we remove a pci root port device and a pci endpoint device which was the root
>> port's child device concurrently.The endponit device will be removed when root port's
>> remove work completed,so when endpoint device itself's remove work start, since endpoint
>> device has been removed, it will result to oops.
>> This patch fix this.
>>
>> CallTrace:
>> kworker/u:2[220]: Oops 11003706212352 [1]
>> Modules linked in: cpufreq_conservative cpufreq_userspace cpufreq_powersave acpi
>> _cpufreq binfmt_misc fuse nls_iso8859_1 loop ipmi_si ipmi_devintf ipmi_msghandle
>> r dm_mod igb ppdev iTCO_wdt parport_pc iTCO_vendor_support i2c_i801 parport sg m
>> ptctl serio_raw i2c_core lpc_ich mfd_core hid_generic button container usbhid hi
>> d uhci_hcd ehci_hcd usbcore usb_common sd_mod crc_t10dif ext3 mbcache jbd fan pr
>> ocessor ide_pci_generic ide_core ata_piix libata mptsas mptscsih mptbase scsi_tr
>> ansport_sas scsi_mod thermal thermal_sys hwmon


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/