GitLab

During the course of a hotplug removal of a PCI bus, release_pcibus_dev()
attempts to remove attribute files from a kobject directory that no longer
exists.  This patch moves these calls to pci_remove_bus(), where they can work
as intended.
Signed-off-by: John Rose <johnrose@austin.ibm.com>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

Linus has requested that the sysfs rom attribute be changed to require
enabling before it works. echo "0" to the attribute to disable,
echoing anything else enables the rom output. The concern is that
something like a file browser could inadvertently read the attribute
and change the state of the hardware without the user's knowledge.

The attached patch includes the previous patch plus the enabling logic.

[root@smirl 0000:02:02.0]#
[root@smirl 0000:02:02.0]# cat rom
cat: rom: Invalid argument
[root@smirl 0000:02:02.0]# echo "1" >rom
[root@smirl 0000:02:02.0]# hexdump -C -n 20 rom
00000000  55 aa 60 e9 d6 01 00 00  00 00 00 00 00 00 00 00  |U.`.............|
00000010  00 00 00 00                                       |....|
00000014
[root@smirl 0000:02:02.0]# echo "0" >rom
[root@smirl 0000:02:02.0]# hexdump -C -n 20 rom
hexdump: rom: Invalid argument
[root@smirl 0000:02:02.0]#
Signed-off-by: Jon Smirl <jonsmirl@gmail.com>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

The following patch implements a pci_remove_bus() that can be used by
hotplug drivers for the removal of root buses.  It also defines a
release function that frees the device struct for pci_bus->bridge when a
root bus class device is unregistered.
Signed-off-by: John Rose <johnrose@austin.ibm.com>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

John Rose wrote:
> The function release_pcibus_dev() in probe.c defines the release procedure for
> device class pcibus_class.  I want to suggest that this function be scrapped :)
>
> This release function is called in the code path of class_device_unregister().
> The pcibus_class devices aren't currently unregistered anywhere, from what I
> can tell, so this release function is currently unused.  The runtime removal of
> PCI buses from logical partitions on PPC64 requires the unregistration of these
> class devices.  The natural place to do this IMHO is in pci_remove_bus_device()
> in remove.c.

You're right that the class device isn't currently unregistered, and
that was an oversight in the patch I originally sent.  Attatched is a
patch that remedies that situation.  pci_remove_bus_device() *is* the
natural place to unregister the class_dev, and that's just what the
patch does.


> The problem is that this calls pci_destroy_dev(), which calls put() on the same
> "bridge" device that the release function does.  This should only be done once
> in the course of removing a pci_bus, and I doubt that we want to change
> pci_destroy_dev().   The kfree() of the pci_bus struct is also done in both
> pci_remove_bus_device() and release_pcibus_dev().

Yep.  The patch pulls the kfree() out of pci_remove_bus_device() and
calls class_device_unregister() in it's place.  This will put() the
bridge device and free the pci_bus as needed.  put() does need to be
called twice because the bridge device is get()'d twice: once when the
device is registered and once when it's bus device grabs a reference to it.

From: long <tlnguyen@snoqualmie.dp.intel.com>


Add support for Message Signalled Interrupt delivery on ia32.

With a fix from Zwane Mwaikambo <zwane@arm.linux.org.uk>

This fixes a build problem on sparc, and hopefully will prevent this from
happening in the future...