Reject URBs to _all_ devices when their host controllers are suspended;
even root hub registers will be unavailable.  Also, don't reject urbs
to root hubs in other cases; the only upstream link is through that
controller (on PCI or whatever SOC bus is in use).
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

 drivers/usb/core/hcd.c |   28 ++++++++++++----------------
 drivers/usb/core/urb.c |    3 ++-
 2 files changed, 14 insertions(+), 17 deletions(-)

The way we're looking at USB suspend lately doesn't expect drivers to
call usb_suspend_device() or usb_resume_device() directly; that'll
be implicit when no interfaces are in use.

This patch removes those APIs from visibility outside usbcore.
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>

 drivers/usb/core/hub.c |   12 ++++--------
 drivers/usb/core/usb.h |    4 ++++
 include/linux/usb.h    |    5 -----
 3 files changed, 8 insertions(+), 13 deletions(-)

This updates the PCI glue to address the new and simplified usbcore suspend
semantics, where CONFIG_USB_SUSPEND becomes irrelevant to HCDs because
hcd->hub_suspend() will always be called.

  - Removes now-unneeded recursion support

  - Go back to ignoring faults reported by the wakeup calls; we expect them
    to fail sometimes, and that's just fine.

The PCI HCDs will need simple changes to catch up to this, like being able
to ignore the setting of CONFIG_USB_SUSPEND.
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

 drivers/usb/core/hcd-pci.c |  106 +++++++++++++++++++++------------------------
 drivers/usb/core/hcd.h     |    6 +-
 2 files changed, 53 insertions(+), 59 deletions(-)

This patch associates hub suspend and resume logic (including for root hubs)
with CONFIG_PM -- instead of CONFIG_USB_SUSPEND as before -- thereby unifying
two troublesome versions of suspend logic into just one.  It'll be easier to
keep things right from now on.

  - Now usbcore _always_ calls hcd->hub_suspend as needed, instead of
    only when USB_SUSPEND is enabled:
     * Those root hub methods are now called from hub suspend/resume;
       no more skipping between layers during device suspend/resume;
     * It now handles cases allowed by sysfs or autosuspended root hubs,
       by forcing the hub interface to resume too.

  - All devices, including virtual root hubs, now get the same treatment
    on their resume paths ... including re-activating all their interfaces.

Plus it gets rid of those stub copies of usb_{suspend,resume}_device(), and
updates the Kconfig to match the new definition of USB_SUSPEND:  it provides
(a) selective suspend, downstream from hubs; and (b) remote wakeup, upstream
from any device configuration which supports it.

This calls for minor followup patches for most HCDs (and their PCI glue).
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

 drivers/usb/core/Kconfig |   11 ++-
 drivers/usb/core/hub.c   |  163 +++++++++++++++++++++++++----------------------
 2 files changed, 97 insertions(+), 77 deletions(-)

This patch collects various small updates related to root hubs, to shrink
later patches which build on them.

  - For root hub suspend/resume support:
     * Make the existing usb_hcd_resume_root_hub() routine respect pmcore
       locking, exporting and using the dpm_runtime_resume() method.
     * Add a new usb_hcd_suspend_root_hub() to pair with that routine.
       (Essential to make OHCI autosuspend behave again...)
     * HC_SUSPENDED by itself only refers to the root hub's downstream ports.
       So let HCDs see root hub URBs unless the parent device is suspended.

  - Remove an assertion we no longer need (and now, also don't want).

  - Generic suspend/resume updates to work better with swsusp.
     * Ignore the FREEZE vs SUSPEND distinction for hardware; trying to
       use it breaks the swsusp snapshots it's supposed to help (sigh).
     * On resume, mark devices as resumed right away, but then
       do nothing else if the device is marked NOTATTACHED.

These changes shouldn't be very noticable by themselves.
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

 drivers/base/power/runtime.c |    1
 drivers/usb/core/hcd.c       |   64 ++++++++++++++++++++++++++++++++++++++-----
 drivers/usb/core/hcd.h       |    1
 drivers/usb/core/hub.c       |   45 ++++++++++++++++++++++++------
 drivers/usb/core/usb.c       |   20 +++++++++----
 drivers/usb/core/usb.h       |    1
 6 files changed, 111 insertions(+), 21 deletions(-)

This cleans up a small recent FIXME, ensuring that all the HCDs provide
root hub suspend/resume methods.  It also wraps the calls to those root
suspend routines just like on the PCI "USB_SUSPEND not defined" cases,
so non-PCI bus glue won't be as tempted to behave very differently.

Several of the SOC based OHCI drivers forgot to list those methods;
the patch also adds those missing declarations.
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

 drivers/usb/core/hcd.c          |   42 +++++++++++++++++++++++++---------------
 drivers/usb/host/ohci-au1xxx.c  |    5 ++++
 drivers/usb/host/ohci-lh7a404.c |    5 ++++
 drivers/usb/host/ohci-pxa27x.c  |    1
 drivers/usb/host/ohci-s3c2410.c |    1
 drivers/usb/host/ohci-sa1111.c  |    1
 6 files changed, 40 insertions(+), 15 deletions(-)

Signed-off-by: NMatt Porter <mporter@kernel.crashing.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

 drivers/usb/core/hub.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

This simplifies some of the PM-related #ifdeffing by recognizing
that USB_SUSPEND depends on PM.  Also, OHCI drivers were often
testing for USB_SUSPEND when they should have tested just PM.
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

 drivers/usb/core/hcd.c          |    2 ++
 drivers/usb/host/ohci-hcd.c     |    2 +-
 drivers/usb/host/ohci-hub.c     |    4 ++--
 drivers/usb/host/ohci-omap.c    |    2 +-
 drivers/usb/host/ohci-pci.c     |    2 +-
 drivers/usb/host/ohci-ppc-soc.c |    4 ++--
 drivers/usb/host/ohci-pxa27x.c  |    2 +-
 drivers/usb/host/ohci-s3c2410.c |    3 +--
 drivers/usb/host/ohci-sa1111.c  |    2 +-
 9 files changed, 12 insertions(+), 11 deletions(-)

This gets rid of some inconsistently duplicated logic to resume interfaces.
Similar code was in both finish_port_resume() and in usb_generic_resume().
Now there is just one copy of that code, accessed regardless of whether
CONFIG_USB_SUSPEND is enabled.  Fault handling is also more consistent.
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch removes the extra usb_suspend_device() parameter.  The original
reason to pass that parameter was so that this routine could suspend any
active children.  A previous patch removed that functionality ... leaving
no reason to pass the parameter.  A close analogy is pci_set_power_state,
which doesn't need a pm_message_t either.

On the internal code path that comes through the driver model, the parameter
is now used to distinguish cases where USB devices need to "freeze" but not
suspend.   It also checks for an error case that's accessible through sysfs:
attempting to suspend a device before its interfaces (or for hubs, ports).
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

 drivers/usb/core/hub.c         |   34 +++++++++++++++++++++-------------
 drivers/usb/core/usb.c         |   23 +++++++++++++++++++++--
 drivers/usb/host/ehci-hcd.c    |    2 +-
 drivers/usb/host/isp116x-hcd.c |    2 +-
 drivers/usb/host/ohci-pci.c    |    2 +-
 include/linux/usb.h            |    2 +-
 6 files changed, 46 insertions(+), 19 deletions(-)

This patch removes some recursion in the CONFIG_USB_SUSPEND logic, which
suspended children (of devices or hubs) that weren't already suspended.
When it sees such cases, suspend now just fails cleanly.

That logic was not needed during system-wide sleep state transitions; and
given the current notions of how to manage selective suspend transitions,
we don't want it there either.  Where it was particularly handy was coping
with various limitations of the sysfs "echo -n N > power/state" support.
(These include assuming that "N" is always meaningful to the driver; and
that drivers can only transition to state N from state zero.)
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This updates the handling of power state for USB interfaces.

  - Formalizes an existing invariant:  interface "power state" is a boolean:
    ON when I/O is allowed, and FREEZE otherwise.  It does so by defining
    some inlined helpers, then using them.

  - Adds a useful invariant:  the only interfaces marked active are those
    bound to non-suspended drivers.  Later patches build on this invariant.

  - Simplifies the interface driver API (and removes some error paths) by
    removing the requirement that they record power state changes during
    suspend and resume callbacks.  Now usbcore does that.

A few drivers were simplified to address that last change.
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

 drivers/usb/core/hub.c       |   33 +++++++++------------
 drivers/usb/core/message.c   |    1
 drivers/usb/core/usb.c       |   65 +++++++++++++++++++++++++++++++++----------
 drivers/usb/core/usb.h       |   18 +++++++++++
 drivers/usb/input/hid-core.c |    2 -
 drivers/usb/misc/usbtest.c   |   10 ------
 drivers/usb/net/pegasus.c    |    2 -
 drivers/usb/net/usbnet.c     |    2 -
 8 files changed, 85 insertions(+), 48 deletions(-)

Now we can log the urbs travelling through usbfs
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Also fixes all drivers that set this field, and removes some other devfs
specfic USB logic.
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

 drivers/usb/class/usblp.c           |    3 +--
 drivers/usb/core/file.c             |   19 ++++---------------
 drivers/usb/image/mdc800.c          |    3 +--
 drivers/usb/input/aiptek.c          |    2 +-
 drivers/usb/input/hiddev.c          |    3 +--
 drivers/usb/media/dabusb.c          |    3 +--
 drivers/usb/misc/auerswald.c        |    3 +--
 drivers/usb/misc/idmouse.c          |    5 ++---
 drivers/usb/misc/legousbtower.c     |    5 ++---
 drivers/usb/misc/rio500.c           |    3 +--
 drivers/usb/misc/sisusbvga/sisusb.c |    5 -----
 drivers/usb/misc/usblcd.c           |    9 ++++-----
 drivers/usb/usb-skeleton.c          |    3 +--
 include/linux/usb.h                 |    7 ++-----
 14 files changed, 22 insertions(+), 51 deletions(-)

This patch adds endpoint information for both devices and interfaces to
sysfs.  Previously it was only possible to get the endpoint information
from usbfs, and never possible to get any information on endpoint 0.
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

 drivers/usb/core/sysfs.c |  195 ++++++++++++++++++++++++++++++++++++++++++++++-
 include/linux/usb.h      |    4
 2 files changed, 197 insertions(+), 2 deletions(-)

The previous patch adding the ability to nest struct class_device
changed the paramaters to the call class_device_create().  This patch
fixes up all in-kernel users of the function.
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch teaches "usb_device" about the new driver model wakeup support:

 - It updates device wakeup capabilities when entering a configuration
   with the WAKEUP attribute;

 - During suspend processing it consults the policy bit to see
   whether it should enable wakeup for that device.  (This resolves
   a FIXME to not assume the answer is always "yes"; some devices
   lie about supporting remote wakeup.)

Support for root hubs and the HCDs is separate (and more complex).
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

All the same issues - we can't just save the pointer to the thread, we
must save the pid/uid/euid combination.
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

If a process issues an URB from userspace and (starts to) terminate
before the URB comes back, we run into the issue described above.  This
is because the urb saves a pointer to "current" when it is posted to the
device, but there's no guarantee that this pointer is still valid
afterwards.

In fact, there are three separate issues:

1) the pointer to "current" can become invalid, since the task could be
   completely gone when the URB completion comes back from the device.

2) Even if the saved task pointer is still pointing to a valid task_struct,
   task_struct->sighand could have gone meanwhile.

3) Even if the process is perfectly fine, permissions may have changed,
   and we can no longer send it a signal.

So what we do instead, is to save the PID and uid's of the process, and
introduce a new kill_proc_info_as_uid() function.
Signed-off-by: NHarald Welte <laforge@gnumonks.org>
[ Fixed up types and added symbol exports ]
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

the free_irq() in USB suspend breaks resume on some setups where USB
(ohci/ehci) shares the interrupt with an other device.
Signed-off-by: NDaniel Ritz <daniel.ritz@gmx.ch>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

add the helper and use it instead of open coding the klist_node_attached() check
(which is a layering violation IMHO)

idea by Alan Stern.
Signed-off-by: NDaniel Ritz <daniel.ritz@gmx.ch>
Cc: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

This appears to help some folk, please merge.
This patch relaxes reset timings.  There are some reports that it
helps make enumeration work better on some high speed devices.
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Routine cases like handoff-to-companion shouldn't trigger diagnostics.
This gets rid of some recently added log spamming.  It's routine for
hub_port_wait_reset() to return -ENOTCONN to indicate handoff from
highspeed hubs to companions, so an error message is incorrect.
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch (as549) introduces two small changes in the HCD glue layer.
The first simply removes a redundant test.  The second allows root-hub
polling to continue for a single iteration after a host controller dies;
this is needed for the patch that follows.
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This adds think_time to the usb_tt struct and sets it appropriately
(measured in ns); this can help us implement better split transaction
scheduling.
Signed-off-by: NDan Streetman <ddstreet@ieee.org>
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Avoid an annoying message that can appear if devices are disconnected
in the middle of a USB scatterlist operation.

Message noted in http://bugzilla.kernel.org/show_bug.cgi?id=4373
(but the real issue there seems to be a SCSI level hang).
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Use a more correct calculation for highspeed bit times.

  http://bugzilla.kernel.org/show_bug.cgi?id=3604

This sort if thing might start to make a difference now that the high
speed periodic scheduler is more complete -- and even getting used.
Signed-off-by: NDavid Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch (as556) adds support for unbinding the usb_generic "driver".
That driver only binds to USB devices, as opposed to interfaces, and it
does nothing much besides marking which struct device's go with an
overall USB device plus providing suspend/resume methods.  Now that
users can unbind drivers at will using the sysfs "unbind" attribute, we
need a rational way of dealing with USB devices that are no longer under
full control of the USB stack.  The patch handles this by unconfiguring
the device, thereby removing all the interfaces and their associated
drivers and children.
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch (as555) modifies the already-awkward
usb_lock_device_for_reset routine in usbcore by adding a timeout.  The
whole point of the routine is that the caller wants to acquire some
semaphores in the wrong order; protecting against the possibility of
deadlock by timing out seems only prudent.
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch (as554) makes the hub driver disconnect any child USB devices
when it is unbound from a hub.  Normally this will never happen, but
there are a few oddball ways to unbind the hub driver while leaving the
children intact.  For example, the new "unbind" sysfs attribute can be
used for this purpose.

Given that unbinding hubs with children is now safe, the patch also
removes the code that prevented people from doing so using usbfs.
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch (as553) merely moves some code and deletes an unneeded test in
the hub driver.  This is in preparation for the patch that follows.
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

29 July 2005, Cambridge, MA:

This afternoon Alan Stern submitted a patch to remove the URB_ASYNC_UNLINK
flag from the Linux kernel.  Mr. Stern explained, "This flag is a relic
from an earlier, less-well-designed system.  For over a year it hasn't
been used for anything other than printing warning messages."

An anonymous spokesman for the Linux kernel development community
commented, "This is exactly the sort of thing we see happening all the
time.  As the kernel evolves, support for old techniques and old code can
be jettisoned and replaced by newer, better approaches.  Proprietary
operating systems do not have the freedom or flexibility to change so
quickly."

Mr. Stern, a staff member at Harvard University's Rowland Institute who
works on Linux only as a hobby, noted that the patch (labelled as548) did
not update two files, keyspan.c and option.c, in the USB drivers' "serial"
subdirectory.  "Those files need more extensive changes," he remarked.
"They examine the status field of several URBs at times when they're not
supposed to.  That will need to be fixed before the URB_ASYNC_UNLINK flag
is removed."

Greg Kroah-Hartman, the kernel maintainer responsible for overseeing all
of Linux's USB drivers, did not respond to our inquiries or return our
calls.  His only comment was "Applied, thanks."
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch (as551) fixes another little problem recently added to the
USB core.  Someone didn't fix the type of the first argument to
unregister_chrdev_region.
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch introduces a /sys/class/usb_device/ class
where every connected usb-device will show up:

  tree /sys/class/usb_device/
  /sys/class/usb_device/
  |-- usb1.1
  |   |-- dev
  |   `-- device -> ../../../devices/pci0000:00/0000:00:1d.0/usb1
  |-- usb2.1
  |   |-- dev
  |   `-- device -> ../../../devices/pci0000:00/0000:00:1d.1/usb2
  ...

The presence of the "dev" file lets udev create real device nodes.
  kay@pim:~/src/linux-2.6> tree /dev/bus/usb/
  /dev/bus/usb/
  |-- 1
  |   `-- 1
  |-- 2
  |   `-- 1
  ...

udev rule:
  SUBSYSTEM="usb_device", PROGRAM="/sbin/usb_device %k", NAME="%c"
  (echo $1 | /bin/sed 's/usb\([0-9]*\)\.\([0-9]*\)/bus\/usb\/\1\/\2/')

This makes libusb pick up the real nodes instead of the mounted usbfs:
  export USB_DEVFS_PATH=/dev/bus/usb

Background:
  All this makes it possible to manage usb devices with udev instead of
  the devfs solution. We are currently working on a pam_console/resmgr
  replacement driven by udev and a pam-helper. It applies ACL's to device
  nodes, which is required for modern desktop functionalty like
  "Fast User Switching" or multiple local login support.

New patch with its own major. I've succesfully disabled usbfs and use real
nodes only on my box. With: "export USB_DEVFS_PATH=/dev/bus/usb" libusb picks
up the udev managed nodes instead of reading usbfs files.

This makes udev to provide symlinks for libusb to pick up:
  SUBSYSTEM="usb_device", PROGRAM="/sbin/usbdevice %k", SYMLINK="%c"

/sbin/usbdevice:
  #!/bin/sh
  echo $1 | /bin/sed 's/usbdev\([0-9]*\)\.\([0-9]*\)/bus\/usb\/\1\/\2/'
Signed-off-by: NKay Sievers <kay.sievers@suse.de>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch removes CONFIG_PCI_NAMES.
Signed-off-by: NAdrian Bunk <bunk@stusta.de>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This patch converts kcalloc(1, ...) calls to use the new kzalloc() function.
Signed-off-by: NPekka Enberg <penberg@cs.helsinki.fi>
Cc: Greg KH <greg@kroah.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>

Driver core: link device and all class devices derived from it.

To ease the task of locating class devices derived from a certain
device create symlinks from parent device to its class devices.
Change USB host class device name from usbX to usb_hostX to avoid
conflict when creating aforementioned links.

Tweaked by Greg to have the symlink be "class_name:class_device_name" in
order to prevent duplicate links.
Signed-off-by: NDmitry Torokhov <dtor@mail.ru>
Signed-off-by: NGreg Kroah-Hartman <gregkh@suse.de>

This adds type-checking to pm_message_t, so that people can't confuse it
with int or u32.  It also allows us to fix "disk yoyo" during suspend (disk
spinning down/up/down).

[We've tried that before; since that cpufreq problems were fixed and I've
tried make allyes config and fixed resulting damage.]
Signed-off-by: NPavel Machek <pavel@suse.cz>
Signed-off-by: NAlexander Nyberg <alexn@telia.com>
Signed-off-by: NAndrew Morton <akpm@osdl.org>
Signed-off-by: NLinus Torvalds <torvalds@osdl.org>