GitHub/LineageOS/android_kernel_motorola_exynos9610.git
11 years agousb: gadget: s3c-hsotg: add DAINT masking
Robert Baldyga [Thu, 19 Sep 2013 09:50:20 +0000 (11:50 +0200)]
usb: gadget: s3c-hsotg: add DAINT masking

In OEPInt/IEPInt interrupts handling added bitwise and of DAINT and
DAINTMSK, because we should handle masked interrupts only.

Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: s3c-hsotg: fix endpoint interrupts handling
Robert Baldyga [Thu, 19 Sep 2013 09:50:19 +0000 (11:50 +0200)]
usb: gadget: s3c-hsotg: fix endpoint interrupts handling

When s3c_hsotg_trytx is called for ep without enqueued request, interrupts
for this ep are disabled, to prevent interrupt flooding. Interrupts are
enabled when new request for this ep is starting.

Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: s3c-hsotg: fix s3c_hsotg_write_fifo function for dedicated fifo mode
Robert Baldyga [Thu, 19 Sep 2013 09:50:18 +0000 (11:50 +0200)]
usb: gadget: s3c-hsotg: fix s3c_hsotg_write_fifo function for dedicated fifo mode

In s3c_hsotg_write_fifo function PTxFEmp/NPTxFEmp interrupts are enabled
only in shared-fifo mode. In dedicated-fifo mode they should not be used
(when enabled then cause interrupt storm).

Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: s3c-hsotg: fix "protocol stall" handling
Robert Baldyga [Thu, 19 Sep 2013 09:50:17 +0000 (11:50 +0200)]
usb: gadget: s3c-hsotg: fix "protocol stall" handling

After normal handling of SetupDone interrupt, XferCompl interrupt occurs, and
then we enqueue new setup request. But when ep0 is stalled, there is no
XferCompl, so we have to enqueue setup request immediately after stalling ep.
Otherwise incoming control requests won't be processed correctly.

Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: storage_common: pass filesem to fsg_store_cdrom
Andrzej Pietrasiewicz [Tue, 15 Oct 2013 06:33:13 +0000 (08:33 +0200)]
usb: gadget: storage_common: pass filesem to fsg_store_cdrom

If cdrom flag is set ro flag is implied. Try setting the ro first, and
only if it succeeds set the cdrom flag.

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Acked-by: Michal Nazarewicz <mina86@mina86.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: storage_common: use strtobool instead of kstrtouint
Andrzej Pietrasiewicz [Tue, 15 Oct 2013 06:33:12 +0000 (08:33 +0200)]
usb: gadget: storage_common: use strtobool instead of kstrtouint

strtobool is more flexible for the user and is more appropriate in the
context.

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Acked-by: Michal Nazarewicz <mina86@mina86.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: phy: don't return with NULL from devm_usb_get_phy()
Gabor Juhos [Fri, 11 Oct 2013 07:01:03 +0000 (09:01 +0200)]
usb: phy: don't return with NULL from devm_usb_get_phy()

The callers are expecting an ERR_PTR value in case
of an error. Change he code to return with an encoded
-ENOMEM value in the case of a failed devres_alloc call.

Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: Make VERBOSE_DEBUG enableable via Kconfig
Andreas Larsson [Fri, 11 Oct 2013 06:07:00 +0000 (08:07 +0200)]
usb: gadget: Make VERBOSE_DEBUG enableable via Kconfig

Create a way for VERBOSE_DEBUG to be enabled during
drivers/usb/gadget/ build.

Signed-off-by: Andreas Larsson <andreas@gaisler.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: s3c-hsotg: remove unused label
Robert Baldyga [Fri, 11 Oct 2013 10:48:51 +0000 (12:48 +0200)]
usb: gadget: s3c-hsotg: remove unused label

This patch removes unused label.

Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: s3c-hsotg: add multi count support
Robert Baldyga [Wed, 9 Oct 2013 07:00:02 +0000 (09:00 +0200)]
usb: gadget: s3c-hsotg: add multi count support

This patch adds Multi Count support. It adds few modifications:
- Fix s3c_hsotg_set_ep_maxpacket() function. Field wMaxPacketSize of endpoint
  descriptor is now splitted into maximum packet size value and number of
  additional transaction per microframe.
- Modify s3c_hsotg_write_fifo() function. It actually calculates transfer
  size, taking into account Multi Count value, which indicates number of
  transactions per microframe.
- Fix s3c_hsotg_start_req() function by setting number of packets to Multi
  Count field in DIEPTSIZ register for isochronous endpoints.

Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: s3c-hsotg: add isochronous transfers support
Robert Baldyga [Wed, 9 Oct 2013 06:41:57 +0000 (08:41 +0200)]
usb: gadget: s3c-hsotg: add isochronous transfers support

This patch adds isochronous transfer support. It adds few modifications:
- Modify s3c_hsotg_epint() function. Some interrupts are ignored for
  isochronous endpoints, (e.g. INTknTXFEmpMsk) becouse isochronous request is
  always transfered in single transaction, which ends with XferCompl interrupt.
- Add Odd/Even microframe toggle to allow data transfering in each microframe
  in s3c_hsotg_epint() function.
- Fix s3c_hsotg_ep_enable() function by supporting isochronous endpoint type.

Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: dwc3: core: use pm_runtime_put_sync() on remove
Felipe Balbi [Fri, 11 Oct 2013 13:34:28 +0000 (08:34 -0500)]
usb: dwc3: core: use pm_runtime_put_sync() on remove

We are going to disable runtime_pm and we're
removing the driver, we must disable the device
now.

Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: phy: Add RCAR Gen2 USB phy
Valentine Barshak [Thu, 10 Oct 2013 16:35:17 +0000 (20:35 +0400)]
usb: phy: Add RCAR Gen2 USB phy

This adds RCAR Gen2 USB phy support. The driver configures
USB channels 0/2 which are shared between PCI USB hosts and
USBHS/USBSS devices. It also controls internal USBHS phy.

Signed-off-by: Valentine Barshak <valentine.barshak@cogentembedded.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: musb: cancel work on removal
Sebastian Andrzej Siewior [Thu, 10 Oct 2013 16:26:59 +0000 (18:26 +0200)]
usb: musb: cancel work on removal

So I captured this:

|WARNING: CPU: 0 PID: 2078 at /home/bigeasy/work/new/TI/linux/lib/debugobjects.c:260 debug_print_object+0x94/0xc4()
|ODEBUG: free active (active state 0) object type: work_struct hint: musb_irq_work+0x0/0x38 [musb_hdrc]
|CPU: 0 PID: 2078 Comm: rmmod Not tainted 3.12.0-rc4+ #338
|[<c0014d38>] (unwind_backtrace+0x0/0xf4) from [<c001249c>] (show_stack+0x14/0x1c)
|[<c001249c>] (show_stack+0x14/0x1c) from [<c0037720>] (warn_slowpath_common+0x64/0x84)
|[<c0037720>] (warn_slowpath_common+0x64/0x84) from [<c00377d4>] (warn_slowpath_fmt+0x30/0x40)
|[<c00377d4>] (warn_slowpath_fmt+0x30/0x40) from [<c022ae90>] (debug_print_object+0x94/0xc4)
|[<c022ae90>] (debug_print_object+0x94/0xc4) from [<c022b7e0>] (debug_check_no_obj_freed+0x1c0/0x228)
|[<c022b7e0>] (debug_check_no_obj_freed+0x1c0/0x228) from [<c00f1f38>] (kfree+0xf8/0x228)
|[<c00f1f38>] (kfree+0xf8/0x228) from [<c02921c4>] (release_nodes+0x1a8/0x248)
|[<c02921c4>] (release_nodes+0x1a8/0x248) from [<c028f70c>] (__device_release_driver+0x98/0xf0)
|[<c028f70c>] (__device_release_driver+0x98/0xf0) from [<c028f840>] (device_release_driver+0x24/0x34)
|[<c028f840>] (device_release_driver+0x24/0x34) from [<c028ebe8>] (bus_remove_device+0x148/0x15c)
|[<c028ebe8>] (bus_remove_device+0x148/0x15c) from [<c028d120>] (device_del+0x104/0x1c0)
|[<c028d120>] (device_del+0x104/0x1c0) from [<c02911e4>] (platform_device_del+0x18/0xac)
|[<c02911e4>] (platform_device_del+0x18/0xac) from [<c029179c>] (platform_device_unregister+0xc/0x18)
|[<c029179c>] (platform_device_unregister+0xc/0x18) from [<bf1902fc>] (dsps_remove+0x20/0x4c [musb_dsps])
|[<bf1902fc>] (dsps_remove+0x20/0x4c [musb_dsps]) from [<c0290d7c>] (platform_drv_remove+0x1c/0x24)
|[<c0290d7c>] (platform_drv_remove+0x1c/0x24) from [<c028f704>] (__device_release_driver+0x90/0xf0)
|[<c028f704>] (__device_release_driver+0x90/0xf0) from [<c028f818>] (driver_detach+0xb4/0xb8)
|[<c028f818>] (driver_detach+0xb4/0xb8) from [<c028e6e8>] (bus_remove_driver+0x98/0xec)
|[<c028e6e8>] (bus_remove_driver+0x98/0xec) from [<c008fc70>] (SyS_delete_module+0x1e0/0x24c)
|[<c008fc70>] (SyS_delete_module+0x1e0/0x24c) from [<c000e680>] (ret_fast_syscall+0x0/0x48)
|---[ end trace d79045419a3e51ec ]---

The workqueue is only scheduled from the ep0 and never canceled in case
the musb is removed before the work has a chance to run.

Cc: stable@vger.kernel.org
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: mass_storage: merge usb_f_mass_storage module with u_ms module
Andrzej Pietrasiewicz [Wed, 9 Oct 2013 08:08:29 +0000 (10:08 +0200)]
usb: gadget: mass_storage: merge usb_f_mass_storage module with u_ms module

u_ms.ko is needed only together with usb_f_mass_storage.ko. Merge them.

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: f_mass_storage: remove compatibility layer
Andrzej Pietrasiewicz [Wed, 9 Oct 2013 08:08:28 +0000 (10:08 +0200)]
usb: gadget: f_mass_storage: remove compatibility layer

There are no more old interface users left. Remove it.

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: multi: convert to new interface of f_mass_storage
Andrzej Pietrasiewicz [Wed, 9 Oct 2013 08:08:27 +0000 (10:08 +0200)]
usb: gadget: multi: convert to new interface of f_mass_storage

Convert the legacy multi gadget to the new interface of f_mass_storage,
so that later the compatibility layer in f_mass_storage can be removed.

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: multi: convert to new interface of f_rndis
Andrzej Pietrasiewicz [Wed, 9 Oct 2013 08:08:26 +0000 (10:08 +0200)]
usb: gadget: multi: convert to new interface of f_rndis

Convert the legacy multi gadget to the new interface of f_rndis,
so that later the compatibility layer in f_rndis can be removed.

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: multi: convert to new interface of f_ecm
Andrzej Pietrasiewicz [Wed, 9 Oct 2013 08:08:25 +0000 (10:08 +0200)]
usb: gadget: multi: convert to new interface of f_ecm

Convert the legacy multi gadget to the new interface of f_ecm,
so that later the compatibility layer in f_ecm can be removed.

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: acm_ms: convert to new interface of f_mass_storage
Andrzej Pietrasiewicz [Wed, 9 Oct 2013 08:07:29 +0000 (10:07 +0200)]
usb: gadget: acm_ms: convert to new interface of f_mass_storage

Convert the legacy acm_ms gadget to use the new function interface
of f_mass_storage, so that later the compatibility layer in
f_mass_storage can be removed.

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: f_mass_storage: add configfs support
Andrzej Pietrasiewicz [Wed, 9 Oct 2013 08:06:05 +0000 (10:06 +0200)]
usb: gadget: f_mass_storage: add configfs support

From this commit on f_mass_storage is available through configfs.

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: storage_common: add methods to show/store 'cdrom' and 'removable'
Andrzej Pietrasiewicz [Wed, 9 Oct 2013 08:06:04 +0000 (10:06 +0200)]
usb: gadget: storage_common: add methods to show/store 'cdrom' and 'removable'

This will be required by configfs integration.

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Kyungin Park <kyungmin.park@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: storage_common: make attribute operations more generic
Andrzej Pietrasiewicz [Wed, 9 Oct 2013 08:06:03 +0000 (10:06 +0200)]
usb: gadget: storage_common: make attribute operations more generic

Show/store methods for sysfs attributes contain code which can be used
also by configfs. Make them abstract the source the lun and rw_semaphore
are taken from.

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: mass_storage: convert to new interface of f_mass_storage
Andrzej Pietrasiewicz [Wed, 9 Oct 2013 08:06:02 +0000 (10:06 +0200)]
usb: gadget: mass_storage: convert to new interface of f_mass_storage

Convert old mass_storage gadget to use the new interface of f_mass_storage
so that later the compatibility layer in f_mass_storage can be removed.

struct fsg_common is not known to mass_storage.c, so a setter method
is added to f_mass_storage.

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: f_mass_storage: convert to new function interface with backward compatib...
Andrzej Pietrasiewicz [Wed, 9 Oct 2013 08:06:01 +0000 (10:06 +0200)]
usb: gadget: f_mass_storage: convert to new function interface with backward compatibility

Converting mass storage to the new function interface requires converting
the USB mass storage's function code and its users.
This patch converts the f_mass_storage.c to the new function interface.
The file is now compiled into a separate usb_f_mass_storage.ko module.
The old function interface is provided by means of a preprocessor conditional
directives. After all users are converted, the old interface can be removed.

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Acked-by: Michal Nazarewicz <mina86@mina86.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: f_mass_storage: create fsg_common_run_thread for use in fsg_common_init
Andrzej Pietrasiewicz [Wed, 9 Oct 2013 08:06:00 +0000 (10:06 +0200)]
usb: gadget: f_mass_storage: create fsg_common_run_thread for use in fsg_common_init

fsg_common_init is a lengthy function. Factor a portion of it out.

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Acked-by: Michal Nazarewicz <mina86@mina86.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: f_mass_storage: create fsg_common_set_inquiry_string for use in fsg_comm...
Andrzej Pietrasiewicz [Wed, 9 Oct 2013 08:05:59 +0000 (10:05 +0200)]
usb: gadget: f_mass_storage: create fsg_common_set_inquiry_string for use in fsg_common_init

fsg_common_init is a lengthy function. Factor a portion of it out.

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Acked-by: Michal Nazarewicz <mina86@mina86.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: f_mass_storage: create lun creation helpers for use in fsg_common_init
Andrzej Pietrasiewicz [Wed, 9 Oct 2013 08:05:58 +0000 (10:05 +0200)]
usb: gadget: f_mass_storage: create lun creation helpers for use in fsg_common_init

fsg_common_init is a lengthy function. Factor portions of it out.

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Acked-by: Michal Nazarewicz <mina86@mina86.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: f_mass_storage: create fsg_common_set_cdev for use in fsg_common_init
Andrzej Pietrasiewicz [Wed, 9 Oct 2013 08:05:57 +0000 (10:05 +0200)]
usb: gadget: f_mass_storage: create fsg_common_set_cdev for use in fsg_common_init

fsg_common_init is a lengthy function. Factor a portion of it out.

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Acked-by: Michal Nazarewicz <mina86@mina86.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: f_mass_storage: create lun handling helpers for use in fsg_common_init
Andrzej Pietrasiewicz [Wed, 9 Oct 2013 08:05:56 +0000 (10:05 +0200)]
usb: gadget: f_mass_storage: create lun handling helpers for use in fsg_common_init

fsg_common_init is a lengthy function. Factor portions of it out.

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Acked-by: Michal Nazarewicz <mina86@mina86.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: f_mass_storage: create fsg_common_set_num_buffers for use in fsg_common_init
Andrzej Pietrasiewicz [Wed, 9 Oct 2013 08:05:55 +0000 (10:05 +0200)]
usb: gadget: f_mass_storage: create fsg_common_set_num_buffers for use in fsg_common_init

fsg_common_init is a lengthy function. Factor a portion of it out.

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Acked-by: Michal Nazarewicz <mina86@mina86.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: f_mass_storage: create fsg_common_setup for use in fsg_common_init
Andrzej Pietrasiewicz [Wed, 9 Oct 2013 08:05:54 +0000 (10:05 +0200)]
usb: gadget: f_mass_storage: create fsg_common_setup for use in fsg_common_init

fsg_common_init is a lengthy function. Factor a portion of it out.

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Acked-by: Michal Nazarewicz <mina86@mina86.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: f_mass_storage: make sysfs interface optional
Andrzej Pietrasiewicz [Wed, 9 Oct 2013 08:05:53 +0000 (10:05 +0200)]
usb: gadget: f_mass_storage: make sysfs interface optional

When configfs is in place, the luns will not be represented in sysfs,
so there will be no struct device associated with a lun.
In order to maintain compatibility and allow configfs adoption
sysfs is made optional in this patch.

As a consequence some debug macros need to be adjusted. Two new
fields are added to struct fsg_lun: name and name_pfx.
The "name" is for storing a string which is presented to the user
instead of the dev_name. The "name_pfx", if non-NULL, is prepended
to the "name" at printing time.

The name_pfx is for a future lun.0, which will be a default group in
mass_storage.<name>. By design at USB function configfs group's creation
time its name is not known (but instead set a bit later in
drivers/usb/gadget/configfs.c:function_make) and it is this name that
serves the purpose of the said name prefix. So instead of copying
a yet-unknown string a pointer to it is stored in struct fsg_lun.

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Acked-by: Michal Nazarewicz <mina86@mina86.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: f_mass_storage: create _fsg_common_free_buffers
Andrzej Pietrasiewicz [Wed, 9 Oct 2013 08:05:52 +0000 (10:05 +0200)]
usb: gadget: f_mass_storage: create _fsg_common_free_buffers

When configfs is in place, gadgets will have to be able to free
fsg buffers. Add a helper function.

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Acked-by: Michal Nazarewicz <mina86@mina86.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: s3c-hsotg: fix set_ep_maxpacket function
Robert Baldyga [Wed, 9 Oct 2013 06:20:02 +0000 (08:20 +0200)]
usb: gadget: s3c-hsotg: fix set_ep_maxpacket function

This patch fixes max packet size check in s3c_hsotg_set_ep_maxpacket()
function. According USB specification, bits 10..0 of mps specifies maximum
packet size, so there is bitwise AND between mps and 0x7ff value. Also added
check if maxpacket isn't grater than 1024 which is maximum size od single USB
transaction.

In s3c_hsotg_ep_enable() function added s3c_hsotg_set_ep_maxpacket() call
instead of setting ep.maxpacket value directly.

Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: mv_u3d_core: remove deprecated IRQF_DISABLED
Michael Opdenacker [Sun, 6 Oct 2013 06:44:35 +0000 (08:44 +0200)]
usb: gadget: mv_u3d_core: remove deprecated IRQF_DISABLED

This patch proposes to remove the use of the IRQF_DISABLED flag

It's a NOOP since 2.6.35 and it will be removed one day.
Signed-off-by: Michael Opdenacker <michael.opdenacker@free-electrons.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: g_ffs: fix compilation warning
David Cohen [Fri, 4 Oct 2013 22:30:21 +0000 (15:30 -0700)]
usb: g_ffs: fix compilation warning

If USB_FUNCTIONFS is selected without USB_FUNCTIONFS_ETH and
USB_FUNCTIONFS_RNIS, u_ether.h won't be included and then
USB_ETHERNET_MODULE_PARAMAETERS macro won't be available causing the
following warning compilation:

drivers/usb/gadget/g_ffs.c:81:1: warning: data definition has no type or
storage class [enabled by default]
drivers/usb/gadget/g_ffs.c:81:1: warning: type defaults to ‘int’ in
declaration of ‘USB_ETHERNET_MODULE_PARAMETERS’ [-Wimplicit-int]
drivers/usb/gadget/g_ffs.c:81:1: warning: function declaration isn’t a
prototype [-Wstrict-prototypes]

This patch fixes the warning by making USB_ETHERNET_MODULE_PARAMETERS to
be used iff u_ether.h is included, otherwise it is not needed.

Signed-off-by: David Cohen <david.a.cohen@linux.intel.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: s3c-hsotg: fix maxpacket size
Robert Baldyga [Wed, 2 Oct 2013 09:17:29 +0000 (11:17 +0200)]
usb: gadget: s3c-hsotg: fix maxpacket size

This patch changes ep maxpacket value from 512 to 1024, because it's needed
to handle interupt and isochronous endpoints in high-speed mode. This change
doesn't affect on driver functioning, because fifo size (3072) is still enough
for the maximum transaction payload (3*1024 for high-speed high-bandwidtch
endpoints).

Signed-off-by: Robert Baldyga <r.baldyga@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: phy: Reordering of OTG FSM variables
Anton Tikhomirov [Thu, 3 Oct 2013 03:42:04 +0000 (12:42 +0900)]
usb: phy: Reordering of OTG FSM variables

Reorder variables in struct otg_fsm as they appear in OTG and EH
supplement to USB2.0.

Signed-off-by: Anton Tikhomirov <av.tikhomirov@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: phy: Add and use missed OTG FSM inputs/outputs
Anton Tikhomirov [Thu, 3 Oct 2013 03:42:04 +0000 (12:42 +0900)]
usb: phy: Add and use missed OTG FSM inputs/outputs

Several input/output variables missed in current FSM implementation.
This patch adds and makes use of them as specified in OTG and EH
supplement to USB2.0.

Signed-off-by: Anton Tikhomirov <av.tikhomirov@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: phy: Rename "B-device session end SRP" OTG FSM input
Anton Tikhomirov [Thu, 3 Oct 2013 03:42:04 +0000 (12:42 +0900)]
usb: phy: Rename "B-device session end SRP" OTG FSM input

In accordance with OTG and EH supplement, the correct name
of the FSM input is b_ssend_srp.

Signed-off-by: Anton Tikhomirov <av.tikhomirov@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: phy: Rename OTG FSM informative variables
Anton Tikhomirov [Thu, 3 Oct 2013 03:42:04 +0000 (12:42 +0900)]
usb: phy: Rename OTG FSM informative variables

Mark informative variables with suffix '_inf' to distinguish
them from other non-informative variables with the same name.
If such non-informative varialbes were missed, they are created.

Signed-off-by: Anton Tikhomirov <av.tikhomirov@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: phy: Add and use missed OTG FSM timers
Anton Tikhomirov [Thu, 3 Oct 2013 03:42:04 +0000 (12:42 +0900)]
usb: phy: Add and use missed OTG FSM timers

a_bidl_adis_tmr and a_wait_vfall_tmr OTG timers missed in
current FSM implementation. This patch adds and makes use
of the timers as speicfied in OTG and EH supplement to USB2.0.

Signed-off-by: Anton Tikhomirov <av.tikhomirov@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: phy: Fix OTG FSM timer handling
Anton Tikhomirov [Thu, 3 Oct 2013 03:42:04 +0000 (12:42 +0900)]
usb: phy: Fix OTG FSM timer handling

Get rid of using OTG driver specific timers by passing timer
type to corresponding callbacks.

Signed-off-by: Anton Tikhomirov <av.tikhomirov@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: phy: Add and use missed helper functions
Anton Tikhomirov [Thu, 3 Oct 2013 03:42:04 +0000 (12:42 +0900)]
usb: phy: Add and use missed helper functions

This patch implements missed helper functions for start_gadget() and
start_host() OTG FSM callbacks.

Signed-off-by: Anton Tikhomirov <av.tikhomirov@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: phy: Check OTG FSM callback existance in helper functions
Anton Tikhomirov [Thu, 3 Oct 2013 03:42:04 +0000 (12:42 +0900)]
usb: phy: Check OTG FSM callback existance in helper functions

Existence of callback must be checked to avoid NULL pointer
dereferncing.

Signed-off-by: Anton Tikhomirov <av.tikhomirov@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: phy: Pass OTG FSM pointer to callback functions
Anton Tikhomirov [Thu, 3 Oct 2013 03:42:04 +0000 (12:42 +0900)]
usb: phy: Pass OTG FSM pointer to callback functions

struct otg_fsm may be embedded to device's context structure.
The callbacks may require pointer to struct otg_fsm to obtain
necessary data for its operation (example: regulator reference
for drv_vbus()).

Signed-off-by: Anton Tikhomirov <av.tikhomirov@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: udc-core: Do not report -EISNAM error from gadgetfs
Fabio Estevam [Sat, 28 Sep 2013 03:05:34 +0000 (00:05 -0300)]
usb: gadget: udc-core: Do not report -EISNAM error from gadgetfs

When mounting a gadgetfs the following error message is seen:

$ modprobe gadgetfs
gadgetfs: USB Gadget filesystem, version 24 Aug 2004
$ mkdir  /dev/gadget
$ mount -t gadgetfs none /dev/gadget
nop ci_hdrc.0: failed to start (null): -120

The error comes from gadgetfs_probe(), which returns -EISNAM (-120).

As Alan Stern explains[1], this is the normal behavior:

"It is a temporary measure, used only when the file system is set up
initially.  The real bind routine is gadgetfs_bind(), which gets called
when userspace configures the gadget.

In short, this is how it is intended to work.  It isn't a bug."

[1] http://marc.info/?l=linux-usb&m=138029668707075&w=2

So in order to prevent the error message, do not report EISNAM as an error.

Acked-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agoARM: OMAP2+: usb-host: Adapt to USB phy-nop RESET line changes
Roger Quadros [Tue, 24 Sep 2013 08:53:50 +0000 (11:53 +0300)]
ARM: OMAP2+: usb-host: Adapt to USB phy-nop RESET line changes

The USB phy-nop nop driver expects the RESET line information
to be sent as a GPIO number via platform data. Adapt to that.

Acked-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Roger Quadros <rogerq@ti.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agoARM: OMAP2+: omap-usb-host: Get rid of platform_data from struct usbhs_phy_data
Roger Quadros [Tue, 24 Sep 2013 08:53:49 +0000 (11:53 +0300)]
ARM: OMAP2+: omap-usb-host: Get rid of platform_data from struct usbhs_phy_data

The platform data bits can be inferred from the other members of
struct usbhs_phy_data. So get rid of the platform_data member.

Build the platform data for the PHY device in usbhs_init_phys() instead.

Acked-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Roger Quadros <rogerq@ti.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: phy: generic: Don't use regulator framework for RESET line
Roger Quadros [Tue, 24 Sep 2013 08:53:48 +0000 (11:53 +0300)]
usb: phy: generic: Don't use regulator framework for RESET line

Modelling the RESET line as a regulator supply wasn't a good idea
as it kind of abuses the regulator framework and also makes adaptation
code more complex.

Instead, manage the RESET gpio line directly in the driver. Update
the device tree binding information.

This also makes us easy to migrate to a dedicated GPIO RESET controller
whenever it becomes available.

Signed-off-by: Roger Quadros <rogerq@ti.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: phy: generic: Add gpio_reset to platform data
Roger Quadros [Tue, 24 Sep 2013 08:53:47 +0000 (11:53 +0300)]
usb: phy: generic: Add gpio_reset to platform data

The GPIO number of the RESET line can be passed to the
driver using the gpio_reset member.

Signed-off-by: Roger Quadros <rogerq@ti.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: musb_am335x: Remove redundant of_match_ptr
Sachin Kamat [Mon, 30 Sep 2013 04:14:43 +0000 (09:44 +0530)]
usb: musb_am335x: Remove redundant of_match_ptr

The data structure of_match_ptr() protects is always compiled in.
Hence of_match_ptr() is not needed.

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: s3c-hsotg: fix CodingStyle issues
Pavel Machek [Mon, 30 Sep 2013 12:56:02 +0000 (14:56 +0200)]
usb: gadget: s3c-hsotg: fix CodingStyle issues

checkpatch.pl has some valid complaints about style in s3c-hsotg.c:
- macro with 'if' should be really enclosed in 'do {} while (0)'
- seq_puts() is going to be slightly faster than seq_printf()
- pr_err() is shorter than printk(KERN_ERR ...)

Signed-off-by: Pavel Machek <pavel@denx.de>
[bzolnier: minor fixes]
Signed-off-by: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: musb: ux500: use SIMPLE_DEV_PM_OPS
Daniel Mack [Mon, 30 Sep 2013 19:02:09 +0000 (21:02 +0200)]
usb: musb: ux500: use SIMPLE_DEV_PM_OPS

This removes the DEV_PM_OPS macro and brings this file in line with the
other musb platform drivers.

Signed-off-by: Daniel Mack <zonque@gmail.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: musb: blackfin: use SIMPLE_DEV_PM_OPS
Daniel Mack [Mon, 30 Sep 2013 19:02:08 +0000 (21:02 +0200)]
usb: musb: blackfin: use SIMPLE_DEV_PM_OPS

This makes bfin_pm_ops const and will stub the struct out in case
CONFIG_PM_SLEEP is not set.

Signed-off-by: Daniel Mack <zonque@gmail.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: musb: am35x: use SIMPLE_DEV_PM_OPS
Daniel Mack [Mon, 30 Sep 2013 19:02:07 +0000 (21:02 +0200)]
usb: musb: am35x: use SIMPLE_DEV_PM_OPS

This makes am35x_pm_ops const and will stub the struct out in case
CONFIG_PM_SLEEP is not set.

Signed-off-by: Daniel Mack <zonque@gmail.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: phy: tegra-usb: Remove redundant of_match_ptr
Sachin Kamat [Mon, 30 Sep 2013 04:14:47 +0000 (09:44 +0530)]
usb: phy: tegra-usb: Remove redundant of_match_ptr

The data structure of_match_ptr() protects is always compiled in.
Hence of_match_ptr() is not needed.

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Cc: Stephen Warren <swarren@wwwdotorg.org>
Cc: linux-tegra@vger.kernel.org
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: phy: am335x: Remove redundant of_match_ptr
Sachin Kamat [Mon, 30 Sep 2013 04:14:46 +0000 (09:44 +0530)]
usb: phy: am335x: Remove redundant of_match_ptr

The data structure of_match_ptr() protects is always compiled in.
Hence of_match_ptr() is not needed.

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: phy: am335x-control: Remove redundant of_match_ptr
Sachin Kamat [Mon, 30 Sep 2013 04:14:45 +0000 (09:44 +0530)]
usb: phy: am335x-control: Remove redundant of_match_ptr

The data structure of_match_ptr() protects is always compiled in.
Hence of_match_ptr() is not needed.

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: musb_dsps: Remove redundant of_match_ptr
Sachin Kamat [Mon, 30 Sep 2013 04:14:44 +0000 (09:44 +0530)]
usb: musb_dsps: Remove redundant of_match_ptr

The data structure of_match_ptr() protects is always compiled in.
Hence of_match_ptr() is not needed.

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Cc: Ravi B <ravibabu@ti.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: f_mass_storage: use usb_gstrings_attach
Andrzej Pietrasiewicz [Thu, 26 Sep 2013 12:38:19 +0000 (14:38 +0200)]
usb: gadget: f_mass_storage: use usb_gstrings_attach

Prepare for handling with configfs.

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: f_mass_storage: add a level of indirection for luns storage
Andrzej Pietrasiewicz [Thu, 26 Sep 2013 12:38:18 +0000 (14:38 +0200)]
usb: gadget: f_mass_storage: add a level of indirection for luns storage

This is needed to prepare for configfs integration.

So far the luns have been allocated during gadget's initialization, based
on the nluns module parameter's value; the exact number is known when the
gadget is initialized and that number of luns is allocated in one go; they
all will be used.

When configfs is in place, the luns will be created one-by-one by the user.
Once the user is satisfied with the number of luns, they activate the
gadget. The number of luns must be <= FSG_MAX_LUN (currently 8), but other
than that it is not known up front and the user need not use contiguous
numbering (apart from the default lun #0). On the other hand, the function
code uses lun numbers to identify them and the number needs to be used
as an index into an array.

Given the above, an array needs to be allocated, but it might happen that
7 out of its 8 elements will not be used. On my machine
sizeof(struct fsg_lun) == 462, so > 3k of memory is allocated but not used
in the worst case.

By adding another level of indirection (allocating an array of pointers
to struct fsg_lun and then allocating individual luns instead of an array
of struct fsg_luns) at most 7 pointers are wasted, which is much less.

This patch also changes some for/while loops to cope with the fact
that in the luns array some entries are potentially empty.

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: f_mass_storage: factor out a header file
Andrzej Pietrasiewicz [Thu, 26 Sep 2013 12:38:17 +0000 (14:38 +0200)]
usb: gadget: f_mass_storage: factor out a header file

In order to prepare for the new function interface the f_mass_storage.c
needs to be compiled as a module, and so a header file will be required.

This patch factors out some code to a new f_mass_storage.h.

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: create a utility module for mass_storage
Andrzej Pietrasiewicz [Thu, 26 Sep 2013 12:38:16 +0000 (14:38 +0200)]
usb: gadget: create a utility module for mass_storage

Converting to configfs requires making the f_mass_storage.c a module.

But first we need to get rid of "#include "storage_common.c".

This patch makes storage_common.c a separately compiled file, which is
built as a utility module named u_ms.ko. After all mass storage users are
converted to the new function interface this module can be eliminated
by merging it with the mass storage function's module.

USB descriptors are exported so that they can be accessed from
f_mass_storage.

FSG_VENDOR_ID and FSG_PRODUCT_ID are moved to their only user.

Handling of CONFIG_USB_GADGET_DEBUG_FILES is moved to f_mass_storage.c.
The fsg_num_buffers static is moved to FSG_MODULE_PARAMETER users, so
instead of using a global variable the f_mass_storage introduces
fsg_num_buffers member in fsg_common (and fsg_config).

fsg_strings and fsg_stringtab are moved to f_mass_storage.c.

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: configfs: add a method to unregister the gadget
Andrzej Pietrasiewicz [Thu, 26 Sep 2013 12:38:15 +0000 (14:38 +0200)]
usb: gadget: configfs: add a method to unregister the gadget

Add a method to unregister the gadget using its config_item.

There can be functions (e.g. mass storage), which in some circumstances
need the gadget stopped. Add a method of stopping the gadget.

Signed-off-by: Andrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: Use ERR_CAST inlined function instead of ERR_PTR(PTR_ERR(...))
Duan Jiong [Thu, 26 Sep 2013 07:55:25 +0000 (15:55 +0800)]
usb: gadget: Use ERR_CAST inlined function instead of ERR_PTR(PTR_ERR(...))

trivial patch converting ERR_PTR(PTR_ERR()) into ERR_CAST().
No functional changes.

Signed-off-by: Duan Jiong <duanj.fnst@cn.fujitsu.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: musb: davinci: fix resources passed to MUSB driver for DM6467
Sergei Shtylyov [Sat, 21 Sep 2013 21:43:58 +0000 (01:43 +0400)]
usb: musb: davinci: fix resources passed to MUSB driver for DM6467

After commit 09fc7d22b024692b2fe8a943b246de1af307132b (usb: musb: fix incorrect
usage of  resource pointer), CPPI DMA driver on DaVinci DM6467 can't detect its
dedicated IRQ and so the MUSB IRQ  is erroneously used instead. This is because
only 2 resources are passed to the MUSB driver from the DaVinci glue layer,  so
fix  this by always copying 3 resources (it's  safe since a placeholder for the
3rd resource is always  there) and passing 'pdev->num_resources' instead of the
size of musb_resources[] to platform_device_add_resources().

Cc: <stable@vger.kernel.org> # 3.11+
Signed-off-by: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: usbtest: bmAttributes would better be masked
Huang Rui [Mon, 23 Sep 2013 16:03:43 +0000 (00:03 +0800)]
usb: usbtest: bmAttributes would better be masked

When transfer type is isochronous, the other bits (bits 5..2) of
bmAttributes in endpoint descriptor might not be set zero. So it's better
to use usb_endpoint_type routine to mask bmAttributes with
USB_ENDPOINT_XFERTYPE_MASK to judge the transfter type later.

Acked-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: dwc3: Remove additional delay of 100ms when resuming
Vivek Gautam [Mon, 23 Sep 2013 03:38:28 +0000 (09:08 +0530)]
usb: dwc3: Remove additional delay of 100ms when resuming

This delay got introduced in:
"7415f17 usb: dwc3: core: add power management support"
which reflected similar code in dwc3_core_soft_reset() function.
However, originally the delay of 100ms in dwc3_core_soft_reset() was
meant to assist USB2PHY and USB3PHY reset, not for usb_phy_init()
sequence.

We should get rid of this delay, since things will still work
fine without this.

Reviewed-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Vivek Gautam <gautam.vivek@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: mv_u3d_core: fix memory leaks
Valentin Ilie [Sat, 21 Sep 2013 09:30:15 +0000 (12:30 +0300)]
usb: gadget: mv_u3d_core: fix memory leaks

When trb_hw is NULL, trb should be free'd before return.

Signed-off-by: Valentin Ilie <valentin.ilie@gmail.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: dwc3: Remove redundant pci_set_drvdata
Sachin Kamat [Fri, 20 Sep 2013 09:46:34 +0000 (15:16 +0530)]
usb: dwc3: Remove redundant pci_set_drvdata

Driver core sets driver data to NULL upon failure or remove.

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: musb: use platform_device_register_full() to avoid directly messing with dma...
Russell King [Thu, 19 Sep 2013 23:14:38 +0000 (00:14 +0100)]
usb: musb: use platform_device_register_full() to avoid directly messing with dma masks

Use platform_device_register_full() for those drivers which can, to
avoid messing directly with DMA masks.  This can only be done when
the driver does not need to access the allocated musb platform device
from within its callbacks, which may be called during the musb
device probing.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: net2280: remove unnecessary pci_set_drvdata()
Jingoo Han [Thu, 12 Sep 2013 06:28:45 +0000 (15:28 +0900)]
usb: gadget: net2280: remove unnecessary pci_set_drvdata()

The driver core clears the driver data to NULL after device_release
or on probe failure. Thus, it is not needed to manually clear the
device driver data to NULL.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: pch_udc: remove unnecessary pci_set_drvdata()
Jingoo Han [Thu, 12 Sep 2013 06:27:58 +0000 (15:27 +0900)]
usb: gadget: pch_udc: remove unnecessary pci_set_drvdata()

The driver core clears the driver data to NULL after device_release
or on probe failure. Thus, it is not needed to manually clear the
device driver data to NULL.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: goku_udc: remove unnecessary pci_set_drvdata()
Jingoo Han [Thu, 12 Sep 2013 06:27:09 +0000 (15:27 +0900)]
usb: gadget: goku_udc: remove unnecessary pci_set_drvdata()

The driver core clears the driver data to NULL after device_release
or on probe failure. Thus, it is not needed to manually clear the
device driver data to NULL.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: amd5536udc: remove unnecessary pci_set_drvdata()
Jingoo Han [Thu, 12 Sep 2013 06:25:44 +0000 (15:25 +0900)]
usb: gadget: amd5536udc: remove unnecessary pci_set_drvdata()

The driver core clears the driver data to NULL after device_release
or on probe failure. Thus, it is not needed to manually clear the
device driver data to NULL.

Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: gadget: zero: Add flexible auto remote wakeup test method
Peter Chen [Mon, 9 Sep 2013 08:48:29 +0000 (16:48 +0800)]
usb: gadget: zero: Add flexible auto remote wakeup test method

In order to increase test coverage, we can change the interval between
two remote wakeups every time, and the interval can be any user defined
value. This change will no affect current behavior if the user does not
use two introduced module paramters.

Signed-off-by: Peter Chen <peter.chen@freescale.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: musb: Call atomic_notifier_call_chain when status is changed
Pali Rohár [Wed, 18 Sep 2013 17:03:34 +0000 (19:03 +0200)]
usb: musb: Call atomic_notifier_call_chain when status is changed

More power supply drivers depends on vbus events and without it they not
working. Power supply drivers using usb_register_notifier, so to deliver
events it is needed to call atomic_notifier_call_chain.

So without atomic notifier power supply driver isp1704 not retrieving
vbus status and reporting bogus values to userspace and also to board
platform data functions. Without proper data charger drivers trying to
charge battery also when charger is disconnected or do not start charging
when wallcharger connects.

Atomic notifier in musb driver was used before v3.5 and was replaced with
omap mailbox. This patch adding atomic_notifier_call_chain call from
function omap_musb_set_mailbox.

Signed-off-by: Pali Rohár <pali.rohar@gmail.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: musb: Add missing ATOMIC_INIT_NOTIFIER_HEAD
Pali Rohár [Wed, 18 Sep 2013 17:03:33 +0000 (19:03 +0200)]
usb: musb: Add missing ATOMIC_INIT_NOTIFIER_HEAD

&twl->phy.notifier is not initalized

Signed-off-by: Pali Rohár <pali.rohar@gmail.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: musb: name ux500 platforms more broadly
Linus Walleij [Fri, 6 Sep 2013 16:09:06 +0000 (18:09 +0200)]
usb: musb: name ux500 platforms more broadly

The Kconfig help text is talking about the U5500 which is no
longer supported by the kernel. Name the help text after the
config symbol which is more correct.

Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: phy: am335x: add wakeup support
Sebastian Andrzej Siewior [Mon, 19 Aug 2013 10:39:44 +0000 (12:39 +0200)]
usb: phy: am335x: add wakeup support

This is based on George Cherian's patch which added power & wakeup
support to am335x and does no longer apply since I took some if the code
apart in favor of the multi instance support.
This compiles and I boots and later it detects a device after I plug it in :)
Could somebody please test it so it does what it should?

Cc: George Cherian <george.cherian@ti.com>
Reviewed-by: Roger Quadros <rogerq@ti.com>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agousb: phy: nop: Defer clock prepare until PHY init
Mark Brown [Sun, 11 Aug 2013 14:26:04 +0000 (15:26 +0100)]
usb: phy: nop: Defer clock prepare until PHY init

Since we only enable the PHY clock on init and the PHY init and shutdown
does not occur in atomitc context there is no need to prepare the clock
before it is enabled.  Move the clk_prepare() operations to go along
with the enables, allowing the clock to be fully idle when not in use.

Signed-off-by: Mark Brown <broonie@linaro.org>
Signed-off-by: Felipe Balbi <balbi@ti.com>
11 years agoLinux 3.12-rc3
Linus Torvalds [Sun, 29 Sep 2013 22:02:38 +0000 (15:02 -0700)]
Linux 3.12-rc3

11 years agoMerge tag 'usb-3.12-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb
Linus Torvalds [Sun, 29 Sep 2013 20:47:35 +0000 (13:47 -0700)]
Merge tag 'usb-3.12-rc3' of git://git./linux/kernel/git/gregkh/usb

Pull USB fixes from Greg KH:
 "Here are a number of USB driver fixes for 3.12-rc3.

  These are all for host controller issues that have been reported, and
  there's a fix for an annoying error message that gets printed every
  time you remove a USB 3 device from the system that's been bugging me
  for a while"

* tag 'usb-3.12-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb:
  usb: dwc3: add support for Merrifield
  USB: fsl/ehci: fix failure of checking PHY_CLK_VALID during reinitialization
  USB: Fix breakage in ffs_fs_mount()
  fsl/usb: Resolve PHY_CLK_VLD instability issue for ULPI phy
  usb/core/devio.c: Don't reject control message to endpoint with wrong direction bit
  usb: chipidea: USB_CHIPIDEA should depend on HAS_DMA
  usb: chipidea: udc: free pending TD at removal procedure
  usb: chipidea: imx: Add usb_phy_shutdown at probe's error path
  usb: chipidea: Fix memleak for ci->hw_bank.regmap when removal
  usb: chipidea: udc: fix the oops after rmmod gadget
  USB: fix PM config symbol in uhci-hcd, ehci-hcd, and xhci-hcd
  USB: OHCI: accept very late isochronous URBs
  USB: UHCI: accept very late isochronous URBs
  USB: iMX21: accept very late isochronous URBs
  usbcore: check usb device's state before sending a Set SEL control transfer
  xhci: Fix race between ep halt and URB cancellation
  usb: Fix xHCI host issues on remote wakeup.
  xhci: Ensure a command structure points to the correct trb on the command ring
  xhci: Fix oops happening after address device timeout

11 years agoMerge tag 'tty-3.12-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty
Linus Torvalds [Sun, 29 Sep 2013 20:47:00 +0000 (13:47 -0700)]
Merge tag 'tty-3.12-rc3' of git://git./linux/kernel/git/gregkh/tty

Pull tty/serial fixes from Greg KH:
 "Here are some serial at tty driver fixes for 3.12-rc3

  The serial driver fixes some kref leaks, documentation is moved to the
  proper places, and the tty and n_tty fixes resolve some reported
  regressions.  There is still one outstanding tty regression fix that
  isn't in here yet, as I want to test it out some more, it will be sent
  for 3.12-rc4 if it checks out"

* tag 'tty-3.12-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty:
  tty: ar933x_uart: move devicetree binding documentation
  tty: Fix SIGTTOU not sent with tcflush()
  n_tty: Fix EOF push index when termios changes
  serial: pch_uart: remove unnecessary tty_port_tty_get
  serial: pch_uart: fix tty-kref leak in dma-rx path
  serial: pch_uart: fix tty-kref leak in rx-error path
  serial: tegra: fix tty-kref leak

11 years agoMerge tag 'staging-3.12-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh...
Linus Torvalds [Sun, 29 Sep 2013 20:46:18 +0000 (13:46 -0700)]
Merge tag 'staging-3.12-rc3' of git://git./linux/kernel/git/gregkh/staging

Pull staging fixes from Greg KH:
 "Here are some staging driver fixes, MAINTAINER updates, and a new
  device id.  All of these have been in the linux-next tree, and are
  pretty simple patches"

* tag 'staging-3.12-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging:
  staging: r8188eu: Add new device ID
  staging: imx-drm: Fix probe failure
  staging: vt6656: [BUG] iwctl_siwencodeext return if device not open
  staging: vt6656: [BUG] main_usb.c oops on device_close move flag earlier.
  staging: vt6656: rxtx.c [BUG] s_vGetFreeContext dead lock on null apTD.
  Staging: rtl8192u: r819xU_cmdpkt: checking NULL value after doing dev_alloc_skb
  staging: usbip: Orphan usbip
  staging: r8188eu: Add files for new drive: Cocci spatch "noderef"
  staging: r8188eu: Cocci spatch "noderef"
  staging: octeon-usb: Cocci spatch "noderef"
  staging: r8188eu: Add files for new drive: Cocci spatch "noderef"
  MAINTAINERS: staging: dgnc and dgap drivers: add maintainer
  staging: lustre: Cocci spatch "noderef"

11 years agoMerge tag 'driver-core-3.12-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sun, 29 Sep 2013 20:45:23 +0000 (13:45 -0700)]
Merge tag 'driver-core-3.12-rc3' of git://git./linux/kernel/git/gregkh/driver-core

Pull driver core / sysfs fixes from Greg KH:
 "Here are 2 fixes for 3.12-rc3.  One fixes a sysfs problem with
  mounting caused by 3.12-rc1, and the other is a bug reported by the
  chromeos developers with the driver core.

  Both have been in linux-next for a bit"

* tag 'driver-core-3.12-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core:
  driver core : Fix use after free of dev->parent in device_shutdown
  sysfs: Allow mounting without CONFIG_NET

11 years agoMerge tag 'char-misc-3.12-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregk...
Linus Torvalds [Sun, 29 Sep 2013 20:44:48 +0000 (13:44 -0700)]
Merge tag 'char-misc-3.12-rc3' of git://git./linux/kernel/git/gregkh/char-misc

Pull char/misc driver fixes from Greg KH:
 "Here are some HyperV and MEI driver fixes for 3.12-rc3.  They resolve
  some issues that people have been reporting for them"

* tag 'char-misc-3.12-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/char-misc:
  Drivers: hv: vmbus: Terminate vmbus version negotiation on timeout
  Drivers: hv: util: Correctly support ws2008R2 and earlier
  mei: cancel stall timers in mei_reset
  mei: bus: stop wait for read during cl state transition
  mei: make me client counters less error prone

11 years agoMerge branch 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 29 Sep 2013 17:04:03 +0000 (10:04 -0700)]
Merge branch 'perf-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull perf revert from Ingo Molnar:
 "This fixes the 'perf top' regression Markus Trippelsdorf reported"

* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  Revert "perf symbols: Demangle cloned functions"

11 years agoMerge branch 'drm-fixes' of git://people.freedesktop.org/~airlied/linux
Linus Torvalds [Sun, 29 Sep 2013 17:02:40 +0000 (10:02 -0700)]
Merge branch 'drm-fixes' of git://people.freedesktop.org/~airlied/linux

Pull drm fixes from Dave Airlie:
 "Nothing too major, radeon still has some dpm changes for off by
  default.

  Radeon, intel, msm:
   - radeon: a few more dpm fixes (still off by default), uvd fixes
   - i915: runtime warn backtrace and regression fix
   - msm: iommu changes fallout"

* 'drm-fixes' of git://people.freedesktop.org/~airlied/linux: (27 commits)
  drm/msm: use drm_gem_dumb_destroy helper
  drm/msm: deal with mach/iommu.h removal
  drm/msm: Remove iommu include from mdp4_kms.c
  drm/msm: Odd PTR_ERR usage
  drm/i915: Fix up usage of SHRINK_STOP
  drm/radeon: fix hdmi audio on DCE3.0/3.1 asics
  drm/i915: preserve pipe A quirk in i9xx_set_pipeconf
  drm/i915/tv: clear adjusted_mode.flags
  drm/i915/dp: increase i2c-over-aux retry interval on AUX DEFER
  drm/radeon/cik: fix overflow in vram fetch
  drm/radeon: add missing hdmi callbacks for rv6xx
  drm/i915: Use a temporary va_list for two-pass string handling
  drm/radeon/uvd: lower msg&fb buffer requirements on UVD3
  drm/radeon: disable tests/benchmarks if accel is disabled
  drm/radeon: don't set default clocks for SI when DPM is disabled
  drm/radeon/dpm/ci: filter clocks based on voltage/clk dep tables
  drm/radeon/dpm/si: filter clocks based on voltage/clk dep tables
  drm/radeon/dpm/ni: filter clocks based on voltage/clk dep tables
  drm/radeon/dpm/btc: filter clocks based on voltage/clk dep tables
  drm/radeon/dpm: fetch the max clk from voltage dep tables helper
  ...

11 years agoRevert "perf symbols: Demangle cloned functions"
Ingo Molnar [Sun, 29 Sep 2013 14:12:54 +0000 (16:12 +0200)]
Revert "perf symbols: Demangle cloned functions"

This reverts commit de95ab53645a2f0015e0f68ee723f18dce2b8b51.

Markus Trippelsdorf reported that this commit broke 'perf top':

 > I just see a gray screen with no text at all. Sometimes the
 > following error messages are printed:
 >
 >  *** Error in `perf': invalid fastbin entry (free): 0x00000000029b18c0
 >  ***
 >  *** Error in `perf': malloc(): memory corruption (fast): 0x0000000000ee0b10 ***

While this code is fixable, the commit itself fails on several levels:

 - it should have been a separate helper function
 - why the heck does it do strchr() twice
 - it casts a const char * over into char *
 - sloppy style
 - it's not even a regression fix!

So lets revert it and re-try the patch in v3.13.

Reported-by: Markus Trippelsdorf <markus@trippelsdorf.de>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
11 years agoMerge branch 'msm-fixes-3.12-rc2' of git://people.freedesktop.org/~robclark/linux...
Dave Airlie [Sun, 29 Sep 2013 00:06:28 +0000 (10:06 +1000)]
Merge branch 'msm-fixes-3.12-rc2' of git://people.freedesktop.org/~robclark/linux into drm-fixes

A small fix + deal with fallout of iommu changes + use new
drm_gem_dumb_destroy helper.

* 'msm-fixes-3.12-rc2' of git://people.freedesktop.org/~robclark/linux:
  drm/msm: use drm_gem_dumb_destroy helper
  drm/msm: deal with mach/iommu.h removal
  drm/msm: Remove iommu include from mdp4_kms.c
  drm/msm: Odd PTR_ERR usage

11 years agoMerge branches 'sched-urgent-for-linus', 'timers-urgent-for-linus' and 'x86-urgent...
Linus Torvalds [Sat, 28 Sep 2013 21:22:17 +0000 (14:22 -0700)]
Merge branches 'sched-urgent-for-linus', 'timers-urgent-for-linus' and 'x86-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull scheduler, timer and x86 fixes from Ingo Molnar:
 - A context tracking ARM build and functional fix
 - A handful of ARM clocksource/clockevent driver fixes
 - An AMD microcode patch level sysfs reporting fixlet

* 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  arm: Fix build error with context tracking calls

* 'timers-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  clocksource: em_sti: Set cpu_possible_mask to fix SMP broadcast
  clocksource: of: Respect device tree node status
  clocksource: exynos_mct: Set IRQ affinity when the CPU goes online
  arm: clocksource: mvebu: Use the main timer as clock source from DT

* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/microcode/AMD: Fix patch level reporting for family 15h

11 years agoMerge branch 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sat, 28 Sep 2013 21:21:13 +0000 (14:21 -0700)]
Merge branch 'perf-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull perf fixes from Ingo Molnar:
 "A couple of tooling fixlets and a PMU detection printout fix"

* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  perf/x86: Fix PMU detection printout when no PMU is detected
  perf symbols: Demangle cloned functions
  perf machine: Fix path unpopulated in machine__create_modules()
  perf tools: Explicitly add libdl dependency
  perf probe: Fix probing symbols with optimization suffix
  perf trace: Add mmap2 handler
  perf kmem: Make it work again on non NUMA machines

11 years agoMerge tag 'xfs-for-linus-v3.12-rc3' of git://oss.sgi.com/xfs/xfs
Linus Torvalds [Sat, 28 Sep 2013 20:52:05 +0000 (13:52 -0700)]
Merge tag 'xfs-for-linus-v3.12-rc3' of git://oss.sgi.com/xfs/xfs

Pull xfs bugfixes from Ben Myers:
 - fix for directory node collapse regression
 - fix for recovery over stale on disk structures
 - fix for eofblocks ioctl
 - fix asserts in xfs_inode_free
 - lock the ail before removing an item from it

* tag 'xfs-for-linus-v3.12-rc3' of git://oss.sgi.com/xfs/xfs:
  xfs: fix node forward in xfs_node_toosmall
  xfs: log recovery lsn ordering needs uuid check
  xfs: fix XFS_IOC_FREE_EOFBLOCKS definition
  xfs: asserting lock not held during freeing not valid
  xfs: lock the AIL before removing the buffer item

11 years agoMerge branch 'i2c/for-current' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa...
Linus Torvalds [Sat, 28 Sep 2013 20:44:09 +0000 (13:44 -0700)]
Merge branch 'i2c/for-current' of git://git./linux/kernel/git/wsa/linux

Pull i2c fixes from Wolfram Sang:
 "Some driver bugfixes for the I2C subsystem"

* 'i2c/for-current' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux:
  i2c: ismt: initialize DMA buffer
  i2c: designware: 10-bit addressing mode enabling if I2C_DYNAMIC_TAR_UPDATE is set
  i2c: mv64xxx: Do not use writel_relaxed()
  i2c: mv64xxx: Fix some build warnings
  i2c: s3c2410: fix clk_disable/clk_unprepare WARNings

11 years agoMerge tag 'pm+acpi-3.12-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael...
Linus Torvalds [Sat, 28 Sep 2013 20:27:31 +0000 (13:27 -0700)]
Merge tag 'pm+acpi-3.12-rc3' of git://git./linux/kernel/git/rafael/linux-pm

Pull ACPI and power management fixes from Rafael Wysocki:
 "These fix one recent cpufreq regression, a few older bugs that may
  harm users and a kerneldoc typo.

  Specifics:

   1) After the recent locking changes in the cpufreq core it is
      possible to trigger BUG_ON(!policy) in lock_policy_rwsem_read() if
      cpufreq_get() is called before registering a cpufreq driver.  Fix
      from Viresh Kumar.

   2) If intel_pstate has been loaded already, it doesn't make sense to
      do anything in acpi_cpufreq_init() and moreover doing something in
      there in that case may be harmful, so make that function return
      immediately if another cpufreq driver is already present.  From
      Yinghai Lu.

   3) The ACPI IPMI driver sometimes attempts to acquire a mutex from
      interrupt context, which can be avoided by replacing that mutex
      with a spinlock.  From Lv Zheng.

   4) A NULL pointer may be dereferenced by the exynos5440 cpufreq
      driver if a memory allocation made by it fails.  Fix from Sachin
      Kamat.

   5) Hanjun Guo's commit fixes a typo in the kerneldoc comment
      documenting acpi_bus_unregister_driver()"

* tag 'pm+acpi-3.12-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  ACPI / scan: fix typo in comments of acpi_bus_unregister_driver()
  cpufreq: exynos5440: Fix potential NULL pointer dereference
  cpufreq: check cpufreq driver is valid and cpufreq isn't disabled in cpufreq_get()
  acpi-cpufreq: skip loading acpi_cpufreq after intel_pstate
  ACPI / IPMI: Fix atomic context requirement of ipmi_msg_handler()

11 years agoPCI: Workaround missing pci_set_master in pci drivers
Yinghai Lu [Sat, 28 Sep 2013 20:13:07 +0000 (13:13 -0700)]
PCI: Workaround missing pci_set_master in pci drivers

Ben Herrenschmidt found that commit 928bea964827 ("PCI: Delay enabling
bridges until they're needed") breaks PCI in some powerpc environments.

The reason is that the PCIe port driver will call pci_enable_device() on
the bridge, so the device is enabled, but skips pci_set_master because
pcie_port_auto and no acpi on powerpc.

Because of that, pci_enable_bridge() later on (called as a result of the
child device driver doing pci_enable_device) will see the bridge as
already enabled and will not call pci_set_master() on it.

Fixed by add checking in pci_enable_bridge, and call pci_set_master
if driver skip that.

That will make the code more robot and wade off problem for missing
pci_set_master in drivers.

Reported-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Yinghai Lu <yinghai@kernel.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agoMerge branch 'lockref' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux
Linus Torvalds [Sat, 28 Sep 2013 19:36:19 +0000 (12:36 -0700)]
Merge branch 'lockref' of git://git./linux/kernel/git/s390/linux

Pull s390 lockref enablement from Heiko Carstens:
 "Enabling the new lockless lockref variant on s390 would have been
  trivial until Tony Luck added a cpu_relax() call into the
  CMPXCHG_LOOP(), with commit d472d9d98b46 ("lockref: Relax in cmpxchg
  loop")

  As already mentioned cpu_relax() is very expensive on s390 since it
  yields() the current virtual cpu.  So we are talking of several
  thousand cycles.  Considering this enabling the lockless lockref
  variant would contradict the intention of the new semantics.  And also
  some quick measurements show performance regressions of 50% and more.

  Simply removing the cpu_relax() call again seems also not very
  desireable since Waiman Long reported that for some workloads the call
  improved performance by 5%."

* 'lockref' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux:
  s390: enable ARCH_USE_CMPXCHG_LOCKREF
  lockref: use arch_mutex_cpu_relax() in CMPXCHG_LOOP()
  mutex: replace CONFIG_HAVE_ARCH_MUTEX_CPU_RELAX with simple ifdef