linux-kernel - 2.6.24-rc4: pci_remove_bus_device() => pci_scan_child_bus() => pci_bus_add

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-next>] [day] [month] [year] [list]

Message-ID: <48591941.4070408@extricom.com>
Date:	Wed, 18 Jun 2008 17:18:41 +0300
From:	Eran Liberty <liberty@...ricom.com>
To:	linux-kernel@...r.kernel.org
Subject: 2.6.24-rc4: pci_remove_bus_device() => pci_scan_child_bus() => pci_bus_add_devices
 bug?

Short:
1. Calling pci_remove_bus_device() on some of the devices on a bus will 
remove them. (ok)
2. Calling pci_scan_child_bus() on that bus will scan the bus for 
devices and add them all. even the ones already in the bus device list, 
which will now be duplicated in the bus device list! (no so ok)
3. Calling pci_bus_add_devices() Will try to add the device, assign 
resource, and create procfs files. The devices which were not originally 
removed are were being assigned resource they don't need and procfs 
files already exist and will collide (Bug)

Background:
I have a programmable component  which implements a pci device. Upon 
warm update i need to remove my devices from the bus, upgrade, and add 
them back.
I have bashed my head against pci.h API for the last week trying to make 
this work with no success. Right now I have devised a hideous hack to 
circumvent pci.h and directly remap the pci device registers.

demo code:

reload_my_hardware_design() {
    struct pci_dev *dev = NULL;
    struct pci_bus *bus = NULL;

    while ((dev = pci_get_device(PCI_VENDOR_ID_MY_ID,PCI_ANY_ID,NULL))) {
         pci_remove_bus_device(dev);
    }

    ...
    reload the hardware
    ...

    bus = pci_find_bus(0,0); /* though hard coded in this example this 
is the right bus*/
    pci_scan_child_bus(bus);
    pci_bus_add_devices(bus);
}

Is this a bug or am I doing something wrong?

Liberty
--
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/