This reverts commit 786dc1d3d7.
As Al so eloquently points out, the patch is crap. The old code was fine,
the new code was bogus.
It never dereferenced a user pointer, the "->" operator was to an array
member, which gives the _address_ of the member (in user space), not an
actual dereference at all.
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Extend the get_sb() filesystem operation to take an extra argument that
permits the VFS to pass in the target vfsmount that defines the mountpoint.
The filesystem is then required to manually set the superblock and root dentry
pointers. For most filesystems, this should be done with simple_set_mnt()
which will set the superblock pointer and then set the root dentry to the
superblock's s_root (as per the old default behaviour).
The get_sb() op now returns an integer as there's now no need to return the
superblock pointer.
This patch permits a superblock to be implicitly shared amongst several mount
points, such as can be done with NFS to avoid potential inode aliasing. In
such a case, simple_set_mnt() would not be called, and instead the mnt_root
and mnt_sb would be set directly.
The patch also makes the following changes:
(*) the get_sb_*() convenience functions in the core kernel now take a vfsmount
pointer argument and return an integer, so most filesystems have to change
very little.
(*) If one of the convenience function is not used, then get_sb() should
normally call simple_set_mnt() to instantiate the vfsmount. This will
always return 0, and so can be tail-called from get_sb().
(*) generic_shutdown_super() now calls shrink_dcache_sb() to clean up the
dcache upon superblock destruction rather than shrink_dcache_anon().
This is required because the superblock may now have multiple trees that
aren't actually bound to s_root, but that still need to be cleaned up. The
currently called functions assume that the whole tree is rooted at s_root,
and that anonymous dentries are not the roots of trees which results in
dentries being left unculled.
However, with the way NFS superblock sharing are currently set to be
implemented, these assumptions are violated: the root of the filesystem is
simply a dummy dentry and inode (the real inode for '/' may well be
inaccessible), and all the vfsmounts are rooted on anonymous[*] dentries
with child trees.
[*] Anonymous until discovered from another tree.
(*) The documentation has been adjusted, including the additional bit of
changing ext2_* into foo_* in the documentation.
[akpm@osdl.org: convert ipath_fs, do other stuff]
Signed-off-by: David Howells <dhowells@redhat.com>
Acked-by: Al Viro <viro@zeniv.linux.org.uk>
Cc: Nathan Scott <nathans@sgi.com>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Yeah, it's a hack, but it is only temporary until Alan's patches
reworking this area make it in. We really should not care what devices
below us are doing, especially when we do not really know what type of
devices they are. This patch relies on the fact that the endpoint
devices do not have a driver assigned to us.
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Now /sys/class/usb is dynamically created when we have something to put
in it, and removed when all devices go away.
Just trying to cut down on the clutter in sysfs...
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
This moves the usb class devices that control the usbfs nodes to show up
in the proper place in the larger device tree.
No userspace changes is needed, this is compatible due to the symlinks
generated by the driver core.
Cc: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
This will allow for us to give endpoints a major/minor to create a
"usbfs2-like" way to access endpoints directly from userspace in an
easier manner than the current usbfs provides us.
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Instead of a kobject, will make things easier in the future (don't know
what I was thinking when I did this originally...)
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
This patch (as703) improves the error handling when a Set-Configuration
request fails. The old interfaces are all unregistered before the
request is sent, and if the request fails then we don't know what config
the device is using. So it makes no sense to leave actconfig pointing
to the old configuration with its invalid interfaces.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
This patch (as700) modifies the hub driver to take advantage of the new
usb_reset_composite_device API. The existing code had special-case
calls stuck into usb_reset_device, just before and after the reset.
With the new version there's no need for special-case stuff; it all
happens naturally in the form of pre_reset and post_reset notifications.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
This patch (as699) adds usb_reset_composite_device(), a routine for
sending a USB port reset to a device with multiple interfaces owned by
different drivers. Drivers are notified about impending and completed
resets through two new methods in the usb_driver structure.
The patch modifieds the usbfs ioctl code to make it use the new routine
instead of usb_reset_device(). Follow-up patches will modify the hub,
usb-storage, and usbhid drivers so they can utilize this new API.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
RNDIS devices don't get configured owing to a typo in
choose_configuration(). This patch from Giridhar Pemmasani fixes the
typo.
From: Giridhar Pemmasani <giri@lmc.cs.sunysb.edu>
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2.6.16 introduces USB power budgeting in the Linux kernel, and since then, a
fair number of users have observed that some of their devices no longer work in
unpowered hubs (this is not a bug, the devices claim that they need more than
100mA).
The very least we can do is print an informational message to the kernel log
when this happens, otherwise it is not at all clear why the device was not
accepted.
Signed-off-by: Daniel Drake <dsd@gentoo.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
This patch increases an arbitrary limit on the size of
individual isochronous packets submitted via usbfs. The
limit is still arbitrary, but it's now large enough to
support the maximum packet size used by high-bandwidth
isochronous transfers.
Signed-off-by: Micah Dowty <micah@vmware.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
This patch removes the artificial 4088-byte limit that usbfs
currently places on Control transfers. The USB spec does not
specify a strict limit on the size of an entire control transfer.
It does, however, state that the data stage "follows the same
protocol rules as bulk transfers." (USB 2, 8.5.3)
The level of support for large control transfers in real host
controllers varies, but it's important to support at least 4K
transfers. Windows enforces a maximum control transfer size
of 4K, so there exists some hardware that requires a full 4096
byte data stage. Without this patch, we fall short of that by
8 bytes on architectures with a 4K page size, and it becomes
impossible to support such hardware with a user-space driver.
Since any limit placed on control transfers by usbfs would be
arbitrary, this patch replaces the PAGE_SIZE limit with the same
arbitrary limit used by bulk transfers.
Signed-off-by: Micah Dowty <micah@vmware.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Some hubs claim not to support port-power switching, and right now the
hub driver believes them and does not enable power to their ports.
However it turns out that even though they don't actually switch power,
they do ignore all events on a port until told to turn on the power!
This problem has been reported by several users.
This revised patch (as672b) makes the hub driver always try to turn on
port power to all hubs, regardless of what the hub descriptor says. It
also adds a comment explaining the need for this.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
This ugly hack was long overdue to die.
It was a way to print out Sparc interrupts in a more freindly format,
since IRQ numbers were arbitrary opaque 32-bit integers which vectored
into PIL levels. These 32-bit integers were not necessarily in the
0-->NR_IRQS range, but the PILs they vectored to were.
The idea now is that we will increase NR_IRQS a little bit and use a
virtual<-->real IRQ number mapping scheme similar to PowerPC.
That makes this IRQ printing hack irrelevant, and furthermore only a
handful of drivers actually used __irq_itoa() making it even less
useful.
Signed-off-by: David S. Miller <davem@davemloft.net>
Replace all module uses with the new vfs_kern_mount() interface, and fix up
simple_pin_fs().
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
The choose_configuration() routine contains code the determine the
device's power source, so that configurations requiring external power
can be ruled out if the device is running on bus power. Unfortunately
it turns out that some devices have errors in their config descriptors
and other devices don't like the GET_DEVICE_STATUS request.
Since that information wasn't used for anything else, this patch (as673)
removes the code, leaving only a comment. It fixes bugzilla entry
#6448.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
This fixes a small regression in USB controller power usage for many
OHCI controllers, notably including every non-PCI version of OHCI: on
those systems, the runtime autosuspend mechanism is no longer enabled.
The change moves to saner defaults. All root hubs are expected to handle
remote wakeup (and hence autosuspend), although drivers for buggy silicon
may override that default.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
