This patch (as914) replaces a call to flush_scheduled_work() with
cancel_work_sync(), in order to help avoid potential deadlocks.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Jiri Kosina <[email protected]>
This patch (as901) fixes an oversight in ohci-hcd. The
hub_status_data routine must not try to access the controller's
memory-mapped registers if the controller is in a low-power state;
such attempts will cause a crash on some architectures (such as PPC).
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as904) adds code to check for endpoint descriptor bInterval
values outside the legal limits. Illegal values are set to 32 ms, which
seems like a reasonable default.
This fixes Bugzilla #8432.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as912) replaces a couple of calls to flush_workqueue()
with cancel_sync_work() and cancel_rearming_delayed_work(). Using a
more directed approach allows us to avoid some nasty deadlocks. The
prime example occurs when a first-level device (the parent is a root
hub) is removed while at the same time the root hub gets a remote
wakeup request. khubd would try to flush the autosuspend workqueue
while holding the root-hub's lock, and the remote-wakeup workqueue
routine would be waiting to lock the root hub.
The patch also reorganizes the power management portion of
usb_disconnect(), separating it out into its own routine. The
autosuspend workqueue entry is cancelled immediately instead of
waiting for the device's release routine. In addition,
synchronization with the autosuspend thread is carried out even for
root hubs (an oversight in the original code).
Signed-off-by: Alan Stern <[email protected]>
Cc: Andrew Morton <[email protected]>
Cc: Greg KH <[email protected]>
Cc: Mark Lord <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
This patch (as875) adds error-checking to the callers of hisax_register().
It also changes an error pathway in that routine, making it return an error
code rather than 0.
This fixes Bugzilla #7960.
Signed-off-by: Alan Stern <[email protected]>
Acked-by: Karsten Keil <[email protected]>
Cc: Martin Bachem <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
This patch (as910) fixes a ratelimit modification so that the
original error-handling path will be followed even when the log-rate
limitation kicks in.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as906) improves the error handling for the USB power/level
attribute file. If an error occurs, the original power-level settings
will be restored.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as881b) makes the ksuspend_usb_wq workqueue freezable. We
don't want a rogue workqueue thread running around, unexpectedly
suspending or resuming USB devices in the middle of a system sleep
transition.
This fixes Bugzilla #8498.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as907) prevents us from trying to allocate 0 bytes
when an interface has no endpoint descriptors.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as905) removes a micro-optimization from the hub port
initialization code. Previously we had been using a short timeout on
the first attempt the read the device descriptor; now we will use the
standard timeout length.
It's not clear that the short timeout ever provided any benefit. And
now we know of one case where it actually hurts: The device can't meet
the short timeout and then it gets terminally confused.
This fixes Bugzilla #8444.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as902) fixes a mistake I introduced into usb_bulk_msg().
usb_fill_int_urb() already does the bit-shifting calculation for
high-speed Interrupt intervals; it shouldn't be done twice.
Signed-off-by: Alan Stern <[email protected]>
This patch (as883) removes an out-of-date WARN_ON from the main HCD
endpoint-disable routine. The warning is triggered whenever an
endpoint is disabled while the root hub is suspended. In the past
that may not have been legal, but it definitely is legal now. Merely
unbinding a USB driver will do it.
Furthermore, I've never seen any occurrences of this warning that
really did signal an actual bug or error condition. At this point it
has outlived its purpose.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: David Brownell <[email protected]>
This patch (as879) ties up some loose ends from an earlier patch.
These are things I didn't think to include at the time but which
clearly belonged there.
If an autosuspend fails because driver activity races with
the autosuspend call, restart the autosuspend timer.
When a device is resumed by an external request, it counts
as device activity and should update the last_busy time so
that the next autoresume won't occur immediately.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as882) fixes a problem with the EHCI BIOS handoff. On my
machine, the BIOS configures the controller and the handoff fails,
leaving the controller configured. During resume-from-disk, this
confuses ehci-hcd into thinking that the controller has not been
tampered with.
The problem is fixed by turning off the Configured Flag whenever a
BIOS handoff is attempted, whether it succeeds or not.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: David Brownell <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as909) fixes a couple of refcounting errors in the sd
driver's suspend and resume methods.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: James Bottomley <[email protected]>
This patch (as903) adds a "busnum" sysfs attribute for USB devices.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as897) changes the autosuspend timer code to use the
standard types and macros in dealing with jiffies values.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as877) adds a "last_busy" field to struct usb_device, for
use by the autosuspend framework. Now if an autosuspend call comes at
a time when the device isn't busy but hasn't yet been idle for long
enough, the timer can be set to exactly the desired value. And we
will be ready to handle things like HID drivers, which can't maintain
a useful usage count and must rely on the time-of-last-use to decide
when to autosuspend.
The patch also makes some related minor improvements:
Move the calls to the autosuspend condition-checking routine
into usb_suspend_both(), which is the only place where it
really matters.
If the autosuspend timer is already running, don't stop
and restart it.
Replace immediate returns with gotos so that the optional
debugging ouput won't be bypassed.
If autoresume is disabled but the device is already awake,
don't return an error for an autoresume call.
Don't try to autoresume a device if it isn't suspended.
(Yes, this undercuts the previous change -- so sue me.)
Don't duplicate existing code in the autosuspend work routine.
Fix the kerneldoc in usb_autopm_put_interface(): If an
autoresume call fails, the usage counter is left unchanged.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as874) adds another piece to the user-visible part of the
USB autosuspend interface. The new power/level sysfs attribute allows
users to force the device on (with autosuspend off), force the device
to sleep (with autoresume off), or return to normal automatic operation.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as872) adds some WARN_ON()s to various error checks which
are never supposed to fail. Unsettlingly, one of them has shown up in
a user's log! Maybe making the warning more visible and having the
call-stack information available will help pinpoint the source of the
problem.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as870) adds a delay to ehci-hcd's bus_resume routine.
Apparently there are controllers and/or BIOSes out there which need
such a delay to get the ports back into their correct state. This
fixes Bugzilla #8190.
Signed-off-by: Alan Stern <[email protected]>
Cc: David Brownell <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as867) adds an entry for the new power/autosuspend
attribute in Documentation/ABI/testing, and it changes the behavior of
the delay value. Now a delay of 0 means to autosuspend as soon as
possible, and negative values will prevent autosuspend.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as866) adds new entry points for external USB device
suspend and resume requests, as opposed to internally-generated
autosuspend or autoresume. It also changes the existing
remote-wakeup code paths to use the new routines, since remote wakeup
is not the same as autoresume.
As part of the change, it turns out to be necessary to do remote
wakeup of root hubs from a workqueue. We had been using khubd, but it
does autoresume rather than an external resume. Using the
ksuspend_usb_wq workqueue for this purpose seemed a logical choice.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as865) fixes a race in the HCD removal code discovered by
Milan Plzik. Arrival of an interrupt after the root hub was
unregistered could cause the root-hub status timer to start up, even
after it was supposed to have been shut down. The problem is fixed by
moving the del_timer_sync() call to after the HCD's stop() method, at
which time IRQ generation should be disabled.
Cc: Milan Plzik <[email protected]>
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as864) moves the work routine for USB autosuspend from one
source file to another. This permits the removal of one whole global
symbol (!) and should smooth the way for more changes in the future.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as896b) fixes an oversight in the design of
device_schedule_callback(). It is necessary to acquire a reference to the
module owning the callback routine, to prevent the module from being
unloaded before the callback can run.
Signed-off-by: Alan Stern <[email protected]>
Cc: Satyam Sharma <[email protected]>
Cc: Neil Brown <[email protected]>
Cc: Cornelia Huck <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as857) modifies the SG_GET_RESERVED_SIZE and
SG_SET_RESERVED_SIZE ioctls in the sg driver, capping the values at
the device's request_queue's max_sectors value. This will permit
cdrecord to obtain a legal value for the maximum transfer length,
fixing Bugzilla #7026.
The patch also caps the initial reserved_size value. There's no
reason to have a reserved buffer larger than max_sectors, since it
would be impossible to use the extra space.
The corresponding ioctls in the block layer are modified similarly,
and the initial value for the reserved_size is set as large as
possible. This will effectively make it default to max_sectors.
Note that the actual value is meaningless anyway, since block devices
don't have a reserved buffer.
Finally, the BLKSECTGET ioctl is added to sg, so that there will be a
uniform way for users to determine the actual max_sectors value for
any raw SCSI transport.
Signed-off-by: Alan Stern <[email protected]>
Acked-by: Jens Axboe <[email protected]>
Acked-by: Douglas Gilbert <[email protected]>
Signed-off-by: James Bottomley <[email protected]>
There is one significant difference between the behavior of root hubs
(as embodied in host controller hardware) and external hubs: When a
remote-wakeup signal is received, an external hub sends an interrupt
message at the _end_ of the resume sequence but a root hub generates
and interrupt at the _beginning_ of the resume sequence. The host
system must poll for the end of the sequence.
When ehci-hcd was converted to interrupt-driven operation instead of
using polling, the remaining need for this particular poll was
overlooked. This patch (as894) fixes the problem.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
The USB HID driver fails to reset its error-retry timeout when there
has been a long time interval between I/O errors with no successful URB
completions in the meantime. As a result, the very next error would
trigger an immediate reset, even if it was a chance event occurring
long after the previous error.
More USB keyboards and mice than one might expect end up getting I/O
errors. Almost always this results from hardware problems of one sort of
another. For example, people attach the device to a USB extension cable,
which degrades the signal. Or they simply have poor quality cables to
begin with. Or they use a KVM switch which doesn't handle USB messages
correctly. Etc...
There have been reports from several users in which these I/O
errors would occur more or less randomly, at intervals ranging from
seconds to minutes. The error-handling code in hid-core.c was originally
meant for situations where a single outage would persist for a few hundred
ms (electromagnetic interference, for example). It didn't work right when
these more sporadic errors occurred, because of a flaw in the logic
which this patch fixes.
This patch (as873) fixes that oversight.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Jiri Kosina <[email protected]>
This patch (as871) fixes a problem introduced by an earlier change.
It turns out that some systems really do need to have a terminating
skeleton QH present whenever FSBR is on. I don't know any way to tell
which systems do need it and which don't; the easiest answer is to
have it there always.
This fixes the NumLock-hang bug reported by Jiri Slaby.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as872) adds a device table entry and a new quirk flag to
the usblp driver for the Seiko Epson Receipt printer. This printer
returns Vendor-Specific values for bInterfaceClass and
bInterfaceSubClass, but the bInterfaceProtocol value is valid and it
works with usblp. The new quirks flag tells the driver to ignore the
Class and SubClass values in the interface descriptor.
Signed-off-by: Alan Stern <[email protected]>
Cc: Vojtech Pavlik <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as869) reinstates the mutual exclusion between sysfs
attribute method calls and attribute unregistration. The
previously-reported deadlocks have been fixed, and this exclusion is
by far the simplest way to avoid races during driver unbinding.
The check for orphaned read-buffers has been moved down slightly, so
that the remainder of a partially-read buffer will still be available
to userspace even after the attribute has been unregistered.
Signed-off-by: Alan Stern <[email protected]>
Cc: Hugh Dickins <[email protected]>
Cc: Cornelia Huck <[email protected]>
Cc: Oliver Neukum <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
This patch (as868) adds a helper routine for device drivers that need
to set up a callback to perform some action in a different process's
context. This is intended for use by attribute methods that want to
unregister themselves or their parent device. Attribute method calls
are mutually exclusive with unregistration, so such actions cannot be
taken directly.
Two attribute methods are converted to use the new helper routine: one
for SCSI device deletion and one for System/390 ccwgroup devices.
Signed-off-by: Alan Stern <[email protected]>
Cc: Hugh Dickins <[email protected]>
Cc: Cornelia Huck <[email protected]>
Cc: Oliver Neukum <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
This patch (as863) fixes a problem encountered sometimes when resuming
a port on a UHCI controller. The hardware may turn off the
Resume-Detect bit before turning off the Suspend bit, leading usbcore
to think that the port is still suspended and the resume has failed.
The patch makes uhci_finish_suspend() wait until both bits are safely
off.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as862) fixes a couple of bugs in the way usbcore handles
intervals for interrupt URBs. usb_interrupt_msg (and usb_bulk_msg for
backward compatibility) don't set the interval correctly for
high-speed devices. proc_do_submiturb() doesn't set it correctly when
a bulk URB is submitted to an interrupt endpoint.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as861) adds sysfs attributes to expose the autosuspend
delay value for each USB device. If the user changes the delay from 0
(no autosuspend) to a positive value, an autosuspend is attempted.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as860) adds two new sysfs routines:
sysfs_add_file_to_group() and sysfs_remove_file_from_group().
A later patch adds code that uses the new routines.
Signed-off-by: Alan Stern <[email protected]>
Cc: Maneesh Soni <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as859) makes the default USB autosuspend delay a module
parameter of usbcore. By setting the delay value at boot time, users
will be able to prevent the system from autosuspending devices which
for some reason can't handle it.
The patch also stores the autosuspend delay as a per-device value. A
later patch will allow the user to change the value, tailoring the
delay for each individual device. A delay value of 0 will prevent
autosuspend.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as858) makes some minor cleanups to sysfs.c in usbcore.
Unnecessary tests are removed and a few temp variables are added.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as856) attempts to improve the performance of uhci-hcd by
removing the asynchronous skeleton Queue Headers. They don't contain
any useful information but the controller has to read through them at
least once every millisecond, incurring a non-zero DMA overhead.
Now all the asynchronous queues are combined, along with the period-1
interrupt queue, into a single list with a single skeleton QH. The
start of the low-speed control, full-speed control, and bulk sublists
is determined by linear search. Since there should rarely be more
than a couple of QHs in the list, the searches should incur a much
smaller total load than keeping the skeleton QHs.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as855) adds some convenience macros to uhci-hcd, to help
simplify the code for computing hardware DMA pointers.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Some USB devices do have a configuration 0, in contravention of the
USB spec. Normally 0 is supposed to indicate that a device is
unconfigured.
While we can't change what the device is doing, we can change usbcore.
This patch (as852) allows usb_set_configuration() to accept a config
value of -1 as indicating that the device should be unconfigured. The
request actually sent to the device will still contain 0 as the value.
But even if the device does have a configuration 0, dev->actconfig
will be set to NULL and dev->state will be set to USB_STATE_ADDRESS.
Without some sort of special-case handling like this, there is no way
to unconfigure these non-compliant devices.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as849) fixes a bug in the USB hub driver. A single
pre-allocated buffer is used for all port status reads, but nothing
guarantees exclusive use of the buffer. A mutex is added to provide
this guarantee.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as848) adds a useful little debugging message to let us
know when ehci-hcd's bus_suspend method runs. The other HCDs have
similar messages; now ehci-hcd doesn't need to feel left out.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as847) makes some small changes to the hub driver's
suspend method:
For root hubs, the status URB should be unlinked and other
activity stopped _before_ the bus_suspend method is called.
The test for hdev->bus being NULL has been removed, since
it can never succeed.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as846) adds the IGNORE_RESIDUE flag to the unusual_devs
entry for Sony-Ericsson's P990i phone.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Phil Dibowitz <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as850b) disables remote wakeup (and everything else!) on
all EHCI ports when the shutdown() method is called. If remote wakeup
is left active then some systems will reboot instead of powering off.
This fixes Bugzilla #7828.
Signed-off-by: Alan Stern <[email protected]>
Cc: David Brownell <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch will affect the CDB in INQUIRY commands sent to LUNs above 0
when LUN-0 reports a scsi_level of 0; the LUN bits will no longer be set
in the second byte of the CDB. This is as it should be. Nevertheless,
it's possible that some wacky device might be adversely affected. I doubt
anyone will complain...
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: James Bottomley <[email protected]>
This patch (as833) fixes the "SCSI revision" output for
/proc/scsi/scsi. If the scsi_level value is 0 (UNKNOWN), we want it
to show up as "0", not "ffffffff".
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: James Bottomley <[email protected]>
A particular USB device has been reporting short inquiry lengths. The
SCSI code cannot operate properly unless we get an inquiry length of
36 or above (because of the way we parse vendor and product), so
assume at least 36 bytes are valid even if the device reports fewer.
This is wrong, but it's no worse than what we're doing now (using the
garbage beyond the last reported valid byte).
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: James Bottomley <[email protected]>
According to the Bulk-Only spec, usb-storage is supposed to use the
_first_ bulk-in and bulk-out endpoints it finds, not the _last_. And
while we're at it, we ought to test the direction of the interrupt
endpoint as well. This patch (as842) makes both changes.
Signed-off-by: Alan Stern <[email protected]>
Cc: Matthew Dharm <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Now that port status change notifications are interrupt-driven,
ehci-hcd needs to tell usbcore when a remote-wakeup resume operation
is finished -- we can no longer rely on the core to poll and find
out. This patch (as843) uses the root-hub status timer to force a
poll after the resume is complete.
The patch also changes the test for detecting when the TDRSMDN resume
period has expired. It's necessary to use time_after_eq() instead of
time_after(), since the polling is triggered precisely by a timer.
The same change is made for TDRSTR reset expiration, for consistency.
Signed-off-by: Alan Stern <[email protected]>
Cc: David Brownell <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as837) fixes several mistakes in the AIO interface of the
gadgetfs driver:
The ki_retry method is not supposed to do a put on the kiocb.
The extra call to aio_put_req() causes memory corruption.
(Note: This call was removed before, by patch as691, and then
mysteriously re-introduced later.)
Even if a read transfer is cancelled, we can and should send
to the user all the data that did manage to get transferred.
Testing for AIO cancellation in the I/O completion handler
is both racy and (now) unnecessary. aio_complete() does its
own checking, in a safe manner.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: David Brownell <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as710) adds a sysfs class-device attribute file named
"companion" for EHCI controllers. The file contains a list of port
numbers that are dedicated to the companion controller; by writing a
port number to the file the user can force a high-speed device
attached directly to the computer to run at full speed. (As far as I
know it is not possible to do this for a device attached to an
external hub.) A port is removed from the file by writing the
negative of its port number.
Several users have asked for this facility and it seems like a useful
thing to have. Every now and then one runs across a device which
behaves much better at full speed than at high speed.
Signed-off-by: Alan Stern <[email protected]>
Cc: David Brownell <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as709) changes the way ehci-hcd presents port status
values for ports owned by the companion controller. It no longer
hides the information; in particular, it allows the core to see the
disconnect event that occurs when a full- or low-speed device is
switched over to the companion. This is required for the next patch
in this series.
Signed-off-by: Alan Stern <[email protected]>
Cc: David Brownell <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as708) introduces a local variable to hold the port
status-register address in ehci-hub.c. There's not much improvement
in the object code, but it sure is a lot easier to read.
Signed-off-by: Alan Stern <[email protected]>
Cc: David Brownell <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as841) removes from usbcore a couple of support routines
meant to help with bandwidth allocation. With the changes to uhci-hcd
in the previous patch, these routines are no longer used anywhere.
Also removed is the CONFIG_USB_BANDWIDTH option; it no longer does
anything and is no longer needed since the HCDs now handle bandwidth
issues correctly.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: David Brownell <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as840) fixes the bandwidth allocation mechanism in
uhci-hcd. It has never worked correctly.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as835) removes from usb-storage the code which sets all
devices to a SCSI level of at least SCSI-2. The original reasons for
doing this no longer apply, and in fact it prevents certain kinds of
ATA pass-thru commands from being used.
The patch also marks CB and CBI devices that are SCSI-0 (legacy SCSI)
as being single-LUN, since the combined SCSI-over-USB transport
protocol has no way to convey LUN information to these devices.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Matthew Dharm <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Isochronous queues don't need a dummy TD because the Queue Header
isn't managed by the hardware. This patch (as836) removes the
unnecessary dummy TDs.
The patch also fixes a long-standing typo in a comment (a "don't" was
missing -- potentially very confusing!).
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as768) improves the debugging checks for the uhci-hcd
frame list. The number of entries displayed is limited to 10, and the
driver now checks for the correct Skeleton QH link value at the end of
each chain of Isochronous TDs. The code to compute these link values
is now used in two spots, so it is moved into its own separate
subroutine.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as830) removes some unnecessary error checking. According
to the kerneldoc, schedule_work() can't fail.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as831) adds device_may_wakeup() support to uhci-hcd; it
has been lacking for a long time.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Instead of matching all motherboards whose name contains "A7V8X" for a
remote-wakeup hardware bug, this patch (as829) matches only those
boards whose name is exactly equal to "A7V8X". Later motherboards
don't seem to have the bug.
(In fact, it's possible that only one motherboard in the world has the
bug. With only one user reporting problems, it's hard to tell.)
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Certain boards seem to like to issue false overcurrent notifications,
for example on ports that don't have anything connected to them. This
looks like a hardware error, at the level of noise to those ports'
overcurrent input signals (or non-debounced VBUS comparators). This
surfaces to users as truly massive amounts of syslog spam from khubd
(which is appropriate for real hardware problems, except for the
volume from multiple ports).
Using this new "ignore_oc" flag helps such systems work more sanely,
by preventing such indications from getting to khubd (and spamming
syslog). The downside is of course that true overcurrent errors will
be masked; they'll appear as spontaneous disconnects, without the
diagnostics that will let users troubleshoot issues like
short-circuited cables. In addition, controllers with no devices
attached will be forced to poll for new devices rather than relying on
interrupts, since each overcurrent event would generate a new
interrupt.
This patch (as826) is essentially a copy of David Brownell's ignore_oc
patch for ehci-hcd, ported to uhci-hcd.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
The patch (as824b) makes percpu_free() ignore NULL arguments, as one would
expect for a deallocation routine. (Note that free_percpu is #defined as
percpu_free in include/linux/percpu.h.) A few callers are updated to remove
now-unneeded tests for NULL. A few other callers already seem to assume
that passing a NULL pointer to percpu_free() is okay!
The patch also removes an unnecessary NULL check in percpu_depopulate().
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Thanks to several earlier patches, usb_autosuspend_device() and
usb_autoresume_device() are never called with a second argument other
than 1. This patch (as819) removes the now-redundant argument.
It also consolidates some common code between those two routines,
putting it into a new subroutine called usb_autopm_do_device(). And
it includes a sizable kerneldoc update for the affected functions.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as818b) simplifies autosuspend processing by keeping track
of the number of unsuspended children of each USB hub. This will
permit us to avoid a good deal of unnecessary work all the time; we
will no longer have to create a bunch of workqueue entries to carry
out autosuspend requests, only to have them fail because one of the
hub's children isn't suspended.
The basic idea is simple. There already is a usage counter in the
usb_device structure for preventing autosuspends. The patch just
increments that counter for every unsuspended child. There's only one
tricky part: When a device disconnects we need to remember whether it
was suspended at the time (leave the counter alone) or not (decrement
the counter).
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as817) simplifies the remote-wakeup processing in the hub
driver. Now instead of using a specialized code path, it relies on
the standard USB resume routines. The hub_port_resume() function does
an initial get_port_status() to see whether the port has already
resumed itself (as it does when a remote-wakeup request is sent).
This will slow down handling of other resume events slightly, but not
enough to matter.
The patch also changes the hub_port_status() routine, making it return
an error if a short reply is received.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as816) changes an existing flag in the usb_device
structure to a bitflag, preparing the way for more bitflags to come
in the future.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Unlike UHCI, OHCI does not exert any DMA load on the system when no
devices are connected. Consequently there is no advantage to doing
an autostop other than the power savings, so we shouldn't compile the
necessary code unless CONFIG_PM is enabled.
This patch (as820) makes the root-hub suspend and resume routines
conditional on CONFIG_PM. It also prevents autostop from activating
if the device_may_wakeup flag isn't set; some people use this flag to
alert the driver about Resume-Detect bugs in the hardware.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as742b) adds autosuspend/autoresume support to the USB hub
driver. The largest aspect of the change is that we no longer need a
special flag for root hubs that want to be resumed. Now every hub is
autoresumed whenever khubd needs to access it.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as738b) fixes numerous problems in the controller/root-hub
suspend/resume/remote-wakeup support in ehci-hcd:
The bus_resume() routine should wake up only the ports that
were suspended by bus_suspend(). Ports that were already
suspended should remain that way.
The interrupt mask is used to detect loss of power in the
bus_resume() routine (if the mask is 0 then power was lost).
However bus_suspend() always sets the mask to 0. Instead the
mask should retain its normal value, with port-change-detect
interrupts disabled if remote wakeup is turned off.
The interrupt mask should be reset to its correct value at the
end of bus_resume() regardless of whether power was lost.
bus_resume() reinitializes the operational registers if power
was lost. However those registers are not in the aux power
well, hence they can lose their values whenever the controller
is put into D3. They should always be reinitialized.
When a port-change interrupt occurs and the root hub is
suspended, the interrupt handler should request a root-hub
resume instead of starting up the controller all by itself.
There's no need for the interrupt handler to request a
root-hub resume every time a suspended port sends a
remote-wakeup request.
The pci_resume() method doesn't need to check for connected
ports when deciding whether or not to reset the controller.
It can make that decision based on whether Vaux power was
maintained.
Even when the controller does not need to be reset,
pci_resume() must undo the effect of pci_suspend() by
re-enabling the interrupt mask.
If power was lost, pci_resume() must not call ehci_run().
At this point the root hub is still supposed to be suspended,
not running. It's enough to rewrite the command register and
set the configured_flag.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
The net2280 driver is too eager to send zero-length packets when
IN tokens are received on ep0. No such packet should be sent (the
driver should NAK) before the gadget driver has queued the proper
response. Otherwise deferred responses are impossible.
This patch (as823) makes net2280 avoid sending ZLPs for IN transfers
on ep0 until a response has been submitted, and avoids stalling when an
OUT packet is received before a request has been submitted for an OUT
transfer on ep0.
Signed-off-by: Alan Stern <[email protected]>
Cc: David Brownell <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as821) fixes a compiler warning when CONFIG_PM isn't on
("usb_autosuspend_work" defined but not used).
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
All the other root-hub suspend or resume log messages, in ohci-hcd or
any of the other host controller drivers, use the debug priority
level. This patch (as815) makes the one single exception behave like
all the rest.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as809b) moves the declaration of the hub driver's private
data structure from hub.h into the hub.c source file. Lots of other
files import hub.h; they have no need to know about the details of the
hub driver's private data.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as814) adds usb_autopm_set_interface() to the autosuspend
API. It also provides convenient wrapper routines,
usb_autopm_enable() and usb_autopm_disable(), for drivers that want
to specify directly whether autosuspend should be allowed.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as813) gathers together common code for USB interface
autosuspend/autoresume.
It also adds some simple checking at the time an autosuspend request
is made, to see whether the request will fail. This way we don't
add a workqueue entry when it would end up doing nothing.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as811) removes some stale testing code from the root-hub
resume routine in ohci-hcd. It also adds a spin_lock_irq() call that
inadvertently got left out of an error pathway.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as808b) moves the Root Hub Status Change interrupt-disable
code in ohci-hcd back into the interrupt handler proper, to avoid the
chance of adverse interactions with mediocre hardware implementations.
It also deletes the root-hub status timer from within the interrupt-enable
routine. There's no need to poll for status any more once interrupts are
re-enabled.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as806) fixes a compiler warning when ohci-hcd is built
with CONFIG_PM turned off.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as804) makes USB driver matching ignore the interface
class, subclass, and protocol if the device class is Vendor Specific.
Drivers can override this policy by specifying a Vendor ID as part
of the match; then vendor-specific matches are allowed.
Linus Walleij has reported a problem this patch fixes. When a
particular mass-storage device is switched from mass-storage mode to
Media Transfer Protocol, the interface class remains set to mass-storage
and usb-storage binds to it erroneously, even though the device class
changes to Vendor-Specific.
This may cause a problem for some drivers until their match records can
be updated to include Vendor IDs. But if it does, then those records
were broken to begin with.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
The USB HID driver doesn't include any code to handle a STALL on the
interrupt endpoint. While this may be uncommon, it does happen
sometimes. This patch (as805) adds a fix.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
When a suspended OHCI controller sees a port's status change, it sets
both the Root-Hub-Status-Change and the Resume-Detect bits in the
Interrupt Status register. Processing both these bits, the driver
tries to resume the root hub twice!
This patch (as807) fixes the bug by ignoring RD if RHSC is set. It
also prints a slightly more informative log message when a
remote-wakeup event occurs.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as822) prevents the OHCI autostop mechanism from kicking in
if the root hub is not able or not allowed to issue wakeup requests.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
If a drive reports that no media is present, there's no point in
continuing to ask it about media status. This patch (as696) cuts the
TUR polling short as soon as the drive reports no media instead of
going a full 3 iterations.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: James Bottomley <[email protected]>
This patch (as810c) copies a minimum of 36 bytes of INQUIRY data, even if
the device claims that not all of them are valid. Often badly behaved
devices put plausible data in the Vendor, Product, and Revision strings but
set the Additional Length byte to a small value. Using potentially valid
data is certainly better than allocating a short buffer and then reading
beyond the end of it, which is what we do now.
Signed-off-by: Alan Stern <[email protected]>
Cc: James Bottomley <[email protected]>
Cc: Greg KH <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
This patch (as812) changes the kerneldoc comments explaining the return
values from queue_work(), queue_delayed_work(), and
queue_delayed_work_on(). The updated comments explain more accurately the
meaning of the return code and avoid suggesting that a 0 value means the
routine was unsuccessful.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
This patch (as797) fixes device_add() in the driver core. It needs to
pay attention when the driver for a new device reports an error.
At the same time, since bus_remove_device() undoes the effects of both
bus_add_device() and bus_attach_device(), it needs to check whether
the bus_attach_device step failed.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as803) adds an unusual_devs entry for the Nokia 6234
mobile phone.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as800) straightens out the USB endpoint class device
creation routine, fixing a refcount bug in the process.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as799) fixes a nasty refcount error in the USB endpoint class.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as798) adds a workaround to uhci-hcd. At least one Asus
motherboard is wired in such a way that any device attached to a
suspended UHCI controller will prevent the system from entering
suspend-to-RAM by immediately waking it up. The only way around the
problem is to turn the controller off instead of suspending it.
This fixes Bugzilla #6193.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as796) adds an unusual_devs entry for the Nokia 6131, which
doesn't like large transfer sizes.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
This patch (as762) changes the cpufreq_transition_notifier_list from a
blocking_notifier_head to an srcu_notifier_head. This will prevent errors
caused attempting to call down_read() to access the notifier chain at a
time when interrupts must remain disabled, during system suspend.
It's not clear to me whether this is really necessary; perhaps the chain
could be made into an atomic_notifier. However a couple of the callout
routines do use blocking operations, so this approach seems safer.
The head of the notifier chain needs to be initialized before use; this is
done by an __init routine at core_initcall time. If this turns out not to
be a good choice, it can easily be changed.
Signed-off-by: Alan Stern <[email protected]>
Cc: "Paul E. McKenney" <[email protected]>
Cc: Jesse Brandeburg <[email protected]>
Cc: Dave Jones <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Currently the init_srcu_struct() routine has no way to report out-of-memory
errors. This patch (as761) makes it return -ENOMEM when the per-cpu data
allocation fails.
The patch also makes srcu_init_notifier_head() report a BUG if a notifier
head can't be initialized. Perhaps it should return -ENOMEM instead, but
in the most likely cases where this might occur I don't think any recovery
is possible. Notifier chains generally are not created dynamically.
[[email protected]: avoid statement-with-side-effect in macro]
Signed-off-by: Alan Stern <[email protected]>
Acked-by: Paul E. McKenney <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
This patch (as751) adds a new type of notifier chain, based on the SRCU
(Sleepable Read-Copy Update) primitives recently added to the kernel. An
SRCU notifier chain is much like a blocking notifier chain, in that it must
be called in process context and its callout routines are allowed to sleep.
The difference is that the chain's links are protected by the SRCU
mechanism rather than by an rw-semaphore, so calling the chain has
extremely low overhead: no memory barriers and no cache-line bouncing. On
the other hand, unregistering from the chain is expensive and the chain
head requires special runtime initialization (plus cleanup if it is to be
deallocated).
SRCU notifiers are appropriate for notifiers that will be called very
frequently and for which unregistration occurs very seldom. The proposed
"task notifier" scheme qualifies, as may some of the network notifiers.
Signed-off-by: Alan Stern <[email protected]>
Acked-by: Paul E. McKenney <[email protected]>
Acked-by: Chandra Seetharaman <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
This patch (as776) adds a new chapter to Documentation/CodingStyle,
explaining the circumstances under which a function should return 0 for
failure and non-zero for success as opposed to a negative error code for
failure and 0 for success.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Strictly speaking, the Valid bit in SCSI sense data is supposed to
be set only when the Information field contains a valid number. This
patch (as793) turns off the Valid bit when the Information field
hasn't been set.
Signed-off-by: Alan Stern <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Alan Stern