When the platform initialization fails due to missing resources, it will
return -EPROBE_DEFER after dsps_musb_init() has been called.

dsps_musb_init() calls dsps_musb_dbg_init() to allocate the debugfs
nodes. At a later point in time, the probe will be retried, and
dsps_musb_dbg_init() will be called again. debugfs_create_dir() will
fail this time, as the node already exists, and so the entire device
probe will fail with -ENOMEM.

Fix this by moving debugfs_remove_recursive() from dsps_remove() to the
plaform's exit function, so it will be cleanly torn down when the probe
fails. It also feels more natural this way, as .exit is the counterpart
to .init.
Signed-off-by: NDaniel Mack <zonque@gmail.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Otherwise the value from the last probe would be retained that possibly is
freed since (the UDC is removed) and therefore no longer relevant. Reproducible
with the dummy UDC:

  modprobe dummy_hcd
  mount -t gadgetfs gadgetfs /dev/gadget
  umount /dev/gadget
  rmmod dummy_hcd
  mount -t gadgetfs gadgetfs /dev/gadget

BUG: unable to handle kernel paging request at ffffffffa066fd9d
Call Trace:
 [<ffffffff811d0cd2>] ? d_alloc_name+0x22/0x50
 [<ffffffff812b74dc>] ? selinux_d_instantiate+0x1c/0x20
 [<ffffffffa067d687>] gadgetfs_create_file+0x27/0xa0 [gadgetfs]
 [<ffffffffa067da70>] ? setup_req.isra.4+0x80/0x80 [gadgetfs]
 [<ffffffffa067dbac>] gadgetfs_fill_super+0x13c/0x180 [gadgetfs]
 [<ffffffff811bc832>] mount_single+0x92/0xc0
 [<ffffffffa067d0f8>] gadgetfs_mount+0x18/0x20 [gadgetfs]
 [<ffffffff811bc8f9>] mount_fs+0x39/0x1b0
 [<ffffffff8116b220>] ? __alloc_percpu+0x10/0x20
 [<ffffffff811d6da3>] vfs_kern_mount+0x63/0xf0
 [<ffffffff811d93be>] do_mount+0x23e/0xac0
 [<ffffffff811660eb>] ? strndup_user+0x4b/0xf0
 [<ffffffff811d9f63>] SyS_mount+0x83/0xc0
 [<ffffffff81695b69>] system_call_fastpath+0x16/0x1b
Signed-off-by: NLubomir Rintel <lkundrak@v3.sk>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

During system resume, if the event buffers are not setup before
the gadget controller starts then we start with invalid context
and this can lead to bus access errors. This is especially true for
platforms that loose the controller context during system suspend.
e.g. AM437x.

The following backtrace was found when the system is suspended
and resumed with g_zero loaded on AM437x-evm (USB cable connected
to host all the while).

[  120.981506] WARNING: CPU: 0 PID: 1656 at drivers/bus/omap_l3_noc.c:137 l3_interrupt_handler+0x198/0x28c()
[  120.981514] L3 custom error: MASTER:USB0 WR TARGET:GPMC
[  120.981638] Modules linked in: g_mass_storage usb_f_mass_storage libcomposite configfs bufferclass_ti(O) omaplfb(O) cryptodev(O) dwc3 snd_soc_evm snd_soc_omap snd_pe
[  120.981659] CPU: 0 PID: 1656 Comm: sh Tainted: G           O 3.12.10-gc559824 #1
[  120.981669] Backtrace:
[  120.981705] [<c0017880>] (dump_backtrace+0x0/0x10c) from [<c0017a1c>] (show_stack+0x18/0x1c)
[  120.981730]  r6:c02819ac r5:00000009 r4:ec137cb8 r3:00000000
[  120.981767] [<c0017a04>] (show_stack+0x0/0x1c) from [<c056c0b0>] (dump_stack+0x20/0x28)
[  120.981802] [<c056c090>] (dump_stack+0x0/0x28) from [<c0046d08>] (warn_slowpath_common+0x70/0x90)
[  120.981830] [<c0046c98>] (warn_slowpath_common+0x0/0x90) from [<c0046dcc>] (warn_slowpath_fmt+0x38/0x40)
[  120.981856]  r8:c0855eb0 r7:00000002 r6:f1000700 r5:00000007 r4:80080003
[  120.981886] [<c0046d94>] (warn_slowpath_fmt+0x0/0x40) from [<c02819ac>] (l3_interrupt_handler+0x198/0x28c)
[  120.981900]  r3:c0801ab8 r2:c06cb354
[  120.981936] [<c0281814>] (l3_interrupt_handler+0x0/0x28c) from [<c007162c>] (handle_irq_event_percpu+0x54/0x1b8)
[  120.981962] [<c00715d8>] (handle_irq_event_percpu+0x0/0x1b8) from [<c00717c0>] (handle_irq_event+0x30/0x40)
[  120.981993] [<c0071790>] (handle_irq_event+0x0/0x40) from [<c0074058>] (handle_fasteoi_irq+0x74/0x128)
[  120.982006]  r4:ed0056c0 r3:00000000
[  120.982033] [<c0073fe4>] (handle_fasteoi_irq+0x0/0x128) from [<c0070f34>] (generic_handle_irq+0x28/0x38)
[  120.982046]  r4:0000002a r3:c0073fe4
[  120.982085] [<c0070f0c>] (generic_handle_irq+0x0/0x38) from [<c0015560>] (handle_IRQ+0x38/0x8c)
[  120.982098]  r4:c080137c r3:00000182
[  120.982124] [<c0015528>] (handle_IRQ+0x0/0x8c) from [<c00087e0>] (gic_handle_irq+0x30/0x5c)
[  120.982145]  r6:ec137dd0 r5:c07ac480 r4:fa24010c r3:00000100
[  120.982169] [<c00087b0>] (gic_handle_irq+0x0/0x5c) from [<c056fcc0>] (__irq_svc+0x40/0x54)
[  120.982179] Exception stack(0xec137dd0 to 0xec137e18)
[  120.982195] 7dc0:                                     00000000 a00001d3 00000000 00000004
[  120.982216] 7de0: a0000153 ec1d9010 c080de90 ec137e30 c080debc 00000000 ed756e44 ec137e2c
[  120.982232] 7e00: ec137de0 ec137e18 bf1150e4 bf115474 60000153 ffffffff
[  120.982253]  r7:ec137e04 r6:ffffffff r5:60000153 r4:bf115474
[  120.982327] [<bf115438>] (dwc3_complete+0x0/0x40 [dwc3]) from [<c0338f50>] (dpm_complete+0xd4/0x19c)
[  120.982341]  r5:ed756e10 r4:ed756e64
[  120.982370] [<c0338e7c>] (dpm_complete+0x0/0x19c) from [<c0339034>] (dpm_resume_end+0x1c/0x20)
[  120.982400] [<c0339018>] (dpm_resume_end+0x0/0x20) from [<c006d4ec>] (suspend_devices_and_enter+0x118/0x33c)
[  120.982412]  r4:c0833da4 r3:00000000
[  120.982436] [<c006d3d4>] (suspend_devices_and_enter+0x0/0x33c) from [<c006d928>] (pm_suspend+0x218/0x254)
[  120.982458] [<c006d710>] (pm_suspend+0x0/0x254) from [<c006c594>] (state_store+0x70/0xc0)
[  120.982478]  r6:c057a6cc r5:c06a8320 r4:00000003 r3:0000006d
[  120.982515] [<c006c524>] (state_store+0x0/0xc0) from [<c0264cc0>] (kobj_attr_store+0x1c/0x28)
[  120.982546] [<c0264ca4>] (kobj_attr_store+0x0/0x28) from [<c012ccb8>] (sysfs_write_file+0x170/0x1a4)
[  120.982583] [<c012cb48>] (sysfs_write_file+0x0/0x1a4) from [<c00d17e4>] (vfs_write+0xb8/0x190)
[  120.982611] [<c00d172c>] (vfs_write+0x0/0x190) from [<c00d1bf8>] (SyS_write+0x44/0x78)
[  120.982641] [<c00d1bb4>] (SyS_write+0x0/0x78) from [<c0014660>] (ret_fast_syscall+0x0/0x30)
Signed-off-by: NRoger Quadros <rogerq@ti.com>
Acked-by: NFelipe Balbi <balbi@ti.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

It was impossible to enumerate on a SuperSpeed (XHCI) host
with alternate setting = 1 due to the wrongly set 'bMaxBurst'
field in the SuperSpeed Endpoint Companion descriptor.

Testcase:
<host> modprobe -r usbtest; modprobe usbtest alt=1
<device> modprobe g_zero
plug device to SuperSpeed port on the host.

Without this patch the host always complains like so
"usb 12-2: Not enough bandwidth for new device state.
 usb 12-2: Not enough bandwidth for altsetting 1"

Bug was introduced by commit cf9a08ae in v3.9

Fixes: cf9a08ae (usb: gadget: convert source sink and loopback to
new function interface)

Cc: 3.9+ <stable@vger.kernel.org> # 3.9+
Reviewed-by: NFelipe Balbi <balbi@ti.com>
Acked-by: NSebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: NRoger Quadros <rogerq@ti.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Make dwc3_gadget_resize_tx_fifos() iterate only over IN
endpoints that are actually present, based on the
num_in_eps parameter. This terminates the loop so as to
prevent dereferencing a potential NULL dwc->eps[i] where
i >= (num_in_eps + num_out_eps).
Signed-off-by: NJack Pham <jackp@codeaurora.org>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

on early initialization we could fall into
a situation where the mailbox is called before
MUSB's clocks are running, in order to avoid
that, make sure mailbox is always wrapped with
pm_runtime calls.
Reported-by: NStefan Roese <sr@denx.de>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

commi 30a70b02 (usb: musb: fix obex in g_nokia.ko
causing kernel panic) removed phy_power_on()
and phy_power_off() calls from runtime PM callbacks
but it failed to note that the driver depended
on pm_runtime_get_sync() calls to power up the PHY,
thus leaving some platforms without any means to
have a working PHY.

Fix that by enabling the phy during omap2430_musb_init()
and killing it in omap2430_musb_exit().

Fixes: 30a70b02 (usb: musb: fix obex in g_nokia.ko causing kernel panic)
Cc: <stable@vger.kernel.org> # v3.14
Cc: Pali Rohár <pali.rohar@gmail.com>
Cc: Ivaylo Dimitrov <ivo.g.dimitrov.75@gmail.com>
Reported-by: NMichael Scott <hashcode0f@gmail.com>
Tested-by: NMichael Scott <hashcode0f@gmail.com>
Tested-by: NStefan Roese <sr@denx.de>
Reported-by: NRabin Vincent <rabin@rab.in>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

In order to support local WRITE_INSERT + READ_STRIP operations for
non PI enabled fabrics, the fabric driver needs to be able signal
what protection offload operations are supported.

This is done at session initialization time so the modes can be
signaled by individual se_wwn + se_portal_group endpoints, as well
as optionally across different transports on the same endpoint.

For iser-target, set TARGET_PROT_ALL if the underlying ib_device
has already signaled PI offload support, and allow this to be
exposed via a new iscsit_transport->iscsit_get_sup_prot_ops()
callback.

For loopback, set TARGET_PROT_ALL to signal SCSI initiator mode
operation.

For all other drivers, set TARGET_PROT_NORMAL to disable fabric
level PI.

Cc: Martin K. Petersen <martin.petersen@oracle.com>
Cc: Sagi Grimberg <sagig@mellanox.com>
Cc: Or Gerlitz <ogerlitz@mellanox.com>
Cc: Quinn Tran <quinn.tran@qlogic.com>
Cc: Giridhar Malavali <giridhar.malavali@qlogic.com>
Signed-off-by: NNicholas Bellinger <nab@linux-iscsi.org>

Now that TASK_ABORTED status is not generated for all cases by
TMR ABORT_TASK + LUN_RESET, a new TFO->abort_task() caller is
necessary in order to give fabric drivers a chance to unmap
hardware / software resources before the se_cmd descriptor is
released via the normal TFO->release_cmd() codepath.

This patch adds TFO->aborted_task() in core_tmr_abort_task()
in place of the original transport_send_task_abort(), and
also updates all fabric drivers to implement this caller.

The fabric drivers that include changes to perform cleanup
via ->aborted_task() are:

  - iscsi-target
  - iser-target
  - srpt
  - tcm_qla2xxx

The fabric drivers that currently set ->aborted_task() to
NOPs are:

  - loopback
  - tcm_fc
  - usb-gadget
  - sbp-target
  - vhost-scsi

For the latter five, there appears to be no additional cleanup
required before invoking TFO->release_cmd() to release the
se_cmd descriptor.

v2 changes:
  - Move ->aborted_task() call into transport_cmd_finish_abort (Alex)

Cc: Alex Leung <amleung21@yahoo.com>
Cc: Mark Rustad <mark.d.rustad@intel.com>
Cc: Roland Dreier <roland@kernel.org>
Cc: Vu Pham <vu@mellanox.com>
Cc: Chris Boot <bootc@bootc.net>
Cc: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: Michael S. Tsirkin <mst@redhat.com>
Cc: Giridhar Malavali <giridhar.malavali@qlogic.com>
Cc: Saurav Kashyap <saurav.kashyap@qlogic.com>
Cc: Quinn Tran <quinn.tran@qlogic.com>
Cc: Sagi Grimberg <sagig@mellanox.com>
Cc: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: NNicholas Bellinger <nab@linux-iscsi.org>

The USB_QUIRK_RESET flag indicates that a USB device changes its
identity in some way when it is reset.  It may lose its firmware, its
descriptors may change, or it may switch back to a default mode of
operation.

If a device does this, the kernel needs to avoid resetting it.  Resets
are likely to fail, or worse, succeed while changing the device's
state in a way the system can't detect.

This means we should disable the reset-resume mechanism whenever this
quirk flag is present.  An attempted reset-resume will fail, the
device will be logically disconnected, and later on the hub driver
will rediscover and re-enumerate the device.  This will cause the
appropriate udev events to be generated, so that userspace will have a
chance to switch the device into its normal operating mode, if
necessary.
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
CC: Oliver Neukum <oliver@neukum.org>
Reviewed-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This patch move simply OF helper function to extcon core and change function
name as following:
- of_extcon_get_extcon_dev() -> extcon_get_edev_by_phandle()
Signed-off-by: NChanwoo Choi <cw00.choi@samsung.com>
Acked-by: NFelipe Balbi <balbi@ti.com>

When a driver doesn't have pre_reset, post_reset, or reset_resume
methods, the USB core unbinds that driver when its device undergoes a
reset or a reset-resume, and then rebinds it afterward.

The existing straightforward implementation can lead to problems,
because each interface gets unbound and rebound before the next
interface is handled.  If a driver claims additional interfaces, the
claim may fail because the old binding instance may still own the
additional interface when the new instance tries to claim it.

This patch fixes the problem by first unbinding all the interfaces
that are marked (i.e., their needs_binding flag is set) and then
rebinding all of them.

The patch also makes the helper functions in driver.c a little more
uniform and adjusts some out-of-date comments.
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
Reported-and-tested-by: N"Poulain, Loic" <loic.poulain@intel.com>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Add new ulpi IDs which are available on Xilinx Zynq boards.
Signed-off-by: NMichal Simek <michal.simek@xilinx.com>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This makes sure that the name coming out of configfs cannot be used
accidentally as a format string.
Signed-off-by: NKees Cook <keescook@chromium.org>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This patch adds missing spin_unlock and mutex_unlock calls in
error handling code.
Signed-off-by: NRobert Baldyga <r.baldyga@samsung.com>
Acked-by: NMichal Nazarewicz <mina86@mina86.com>
Acked-by: NFelipe Balbi <balbi@ti.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This patch fixes the following Coccinelle warning:

drivers/usb/gadget/composite.c:1142:9-16: WARNING: \
	ERR_CAST can be used with uc
Signed-off-by: NFelipe Balbi <balbi@ti.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This patch fixes the following Coccinelle warning:

drivers/usb/gadget/inode.c:442:8-15: WARNING \
	opportunity for memdup_user
Signed-off-by: NFelipe Balbi <balbi@ti.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

this patch fixes the following Coccinelle warning:

drivers/usb/gadget/f_subset.c:279:8-14: WARNING: \
	PTR_RET can be used
Signed-off-by: NFelipe Balbi <balbi@ti.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This patch fixes the following Coccinelle error:

drivers/usb/gadget/lpc32xx_udc.c:3313:1-7: ERROR: \
	missing clk_put; clk_get on line 3139 and \
	execution via conditional on line 3146
Signed-off-by: NFelipe Balbi <balbi@ti.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Remove out-commented and ifdeffed debugging code.
Signed-off-by: NJohan Hovold <jhovold@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <greg@kroah.com>

Add missing newlines to dev_<level> messages.

Also make some messages less verbose where appropriate.
Signed-off-by: NJohan Hovold <jhovold@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <greg@kroah.com>

Add missing braces to conditional branches and one loop in usb-serial
core and generic implementation.
Signed-off-by: NJohan Hovold <jhovold@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <greg@kroah.com>

Do not discard buffered data and make sure to try to resubmit the write
urbs on errors.

Currently a recoverable error would lead to more data than necessary
being dropped.

Also upgrade error messages from debug to error log level.
Signed-off-by: NJohan Hovold <jhovold@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <greg@kroah.com>

Make sure to try to resubmit the read urb on errors.

Currently a recoverable error would lead to reduced throughput as only
one urb will be used until the port is closed and reopened (or
resumed or unthrottled).

Also upgrade error messages from debug to error log level.
Signed-off-by: NJohan Hovold <jhovold@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <greg@kroah.com>

Drivers are allowed to override the default bulk-out buffer size
(endpoint maximum packet size) in order to increase throughput, but it
does not make much sense to allow buffers smaller than the default.

Note that this is already how bulk_in_size is defined.
Signed-off-by: NJohan Hovold <jhovold@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <greg@kroah.com>

Remove erroneous call to usb_clear_halt which is blocking and cannot be
used in interrupt context.

This code has possibly never been executed as it would cause an oops if
it was. Simply treat a stalled-endpoint error as any other error
condition.
Signed-off-by: NJohan Hovold <jhovold@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <greg@kroah.com>

The USB controller in TI-NSPIRE calculators (LSI Zevio SoC) are based off either
Freescale's USB OTG controller or the USB controller found in the IMX233, both
of which are Chipidea compatible.

This patch adds a device tree binding for the controller.
Signed-off-by: NPeter Chen <peter.chen@freescale.com>
Signed-off-by: NDaniel Tang <dt.tangr@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Use dev_name() for ci_hdrc name to distinguish USBs
Signed-off-by: NPeter Chen <peter.chen@freescale.com>
Signed-off-by: NAlexander Shiyan <shc_work@mail.ru>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Matthieu CASTET and Michael Grzeschik mentioned isr_tr_complete_handler
is a bit messy at below:
http://marc.info/?l=linux-usb&m=139047775001152&w=2

This commit creates a new function isr_setup_packet_handler to handle
setup packet, it makes isr_tr_complete_handler easy to read.

This is no functional change at this commit, tested with g_mass_storage
and g_ether.

Cc: Michael Grzeschik <mgr@pengutronix.de>
Cc: Matthieu CASTET <matthieu.castet@parrot.com>
Signed-off-by: NPeter Chen <peter.chen@freescale.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Reprogramming the DMA after tear down is initiated leads to warning.
This is mainly seen with ISOCH since we do a delayed completion for
ISOCH transfers. In ISOCH transfers dma_completion should not reprogram
if the channel tear down is initiated.
Signed-off-by: NGeorge Cherian <george.cherian@ti.com>
Signed-off-by: NVinod Koul <vinod.koul@intel.com>

Since it is needed outside usbcore and exposed in include/linux/usb.h,
it conflicts with enum dev_state in rt2x00 wireless driver.

Mark it as usb specific to avoid conflicts in the future.
Signed-off-by: NValentina Manea <valentina.manea.m@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Rename struct omap_control_usb to struct omap_control_phy since it can
be used to control PHY of USB, SATA and PCIE. Also move the driver and
include files under *phy* and made the corresponding changes in the users
of phy-omap-control.
Signed-off-by: NKishon Vijay Abraham I <kishon@ti.com>
Signed-off-by: NRoger Quadros <rogerq@ti.com>
Acked-by: NFelipe Balbi <balbi@ti.com>

A device should not be able to be used concurrently both by
the server and the client. Claiming the port used by the
shared device ensures no interface drivers bind to it and
that it is not usable from the server.
Signed-off-by: NValentina Manea <valentina.manea.m@gmail.com>
Acked-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Submit multiple concurrent urbs for HWA isochronous transfer result data
frame reads.  This keeps the read pipeline full and significantly
improves performance in cases where the frame reads cannot be combined
because they are not contiguous or multiples of the max packet size.
Signed-off-by: NThomas Pugliese <thomas.pugliese@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

When reading the transfer result data for an isochronous in request, if
the current frame actual_length is contiguous with the next frame and
actual_length is a multiple of the DTI endpoint max packet size, combine
the current frame with the next frame in a single URB.  This reduces the
number of URBs that must be submitted in that case which increases
performance and reduces CPU interrupt overhead.
Signed-off-by: NThomas Pugliese <thomas.pugliese@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The HWA driver does not do anything with transfer notifications after
receiving the first one and the Alereon HWA allows them to be disabled
as a performance optimization.  This patch sends a vendor specific
command to the Alereon HWA on startup to disable transfer notifications.
If the command is successful, the DTI system is started immediately
since that would normally be started upon the first reception of a
transfer notification which will no longer be sent.
Signed-off-by: NThomas Pugliese <thomas.pugliese@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Data for transfer segments in the WA_SEG_DTI_PENDING state is actively
being read by the driver.  Let the buffer read callback handle the
transfer cleanup since cleaning it up in wa_urb_dequeue will cause the
read callback to access invalid memory if the transfer is completed.
Signed-off-by: NThomas Pugliese <thomas.pugliese@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This patch locks rpipe->seg_lock around the entire transfer segment
cleanup loop in wa_urb_dequeue instead of just one case of the switch
statement.  This fixes a race between __wa_xfer_delayed_run and
wa_urb_dequeue where a transfer segment in the WA_SEG_DELAYED state
could be removed from the rpipe seg_list twice leading to memory
corruption.  It also switches the spin_lock call to use the non-irqsave
version since the xfer->lock is already held and irqs already disabled.
Signed-off-by: NThomas Pugliese <thomas.pugliese@gmail.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This reverts commit 247bf557.

This commit, together with commit 3804fad4
"USBNET: ax88179_178a: enable tso if usb host supports sg dma" were
origially added to get xHCI 1.0 hosts and usb ethernet ax88179_178a devices
working together with scatter gather. xHCI 1.0 hosts pose some requirement on how transfer
buffers are aligned, setting this requirement for 1.0 hosts caused USB 3.0 mass
storage devices to fail more frequently.

USB 3.0 mass storage devices used to work before 3.14-rc1.  Theoretically,
the TD fragment rules could have caused an occasional disk glitch.
Now the devices *will* fail, instead of theoretically failing.
>From a user perspective, this looks like a regression; the USB device obviously
fails on 3.14-rc1, and may sometimes silently fail on prior kernels.

The proper soluition is to implement the TD fragment rules required, but for now
this patch needs to be reverted to get USB 3.0 mass storage devices working at the
level they used to.
Signed-off-by: NMathias Nyman <mathias.nyman@linux.intel.com>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The DELAY_INIT quirk only reduces the frequency of enumeration failures
with the Logitech HD Pro C920 and C930e webcams, but does not quite
eliminate them. We have found that adding a delay of 100ms between the
first and second Get Configuration request makes the device enumerate
perfectly reliable even after several weeks of extensive testing. The
reasons for that are anyone's guess, but since the DELAY_INIT quirk
already delays enumeration by a whole second, wating for another 10th of
that isn't really a big deal for the one other device that uses it, and
it will resolve the problems with these webcams.
Signed-off-by: NJulius Werner <jwerner@chromium.org>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>