The new version of alloc_ep_req() already aligns the buffer size to
wMaxPacketSize on OUT endpoints.
Signed-off-by: NFelipe F. Tonello <eu@felipetonello.com>
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

We added some new locking here, but missed an error path where we need
to unlock.

Fixes: 9acdf4df ('usb: gadget: f_midi: added spinlock on transmit function')
Acked-by: NMichal Nazarewicz <mina86@mina86.com>
Acked-by: NFelipe F. Tonello <eu@felipetonello.com>
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

buflen by default (256) is smaller than wMaxPacketSize (512) in high-speed
devices.

That caused the OUT endpoint to freeze if the host send any data packet of
length greater than 256 bytes.

This is an example dump of what happended on that enpoint:
HOST:   [DATA][Length=260][...]
DEVICE: [NAK]
HOST:   [PING]
DEVICE: [NAK]
HOST:   [PING]
DEVICE: [NAK]
...
HOST:   [PING]
DEVICE: [NAK]

This patch fixes this problem by setting the minimum usb_request's buffer size
for the OUT endpoint as its wMaxPacketSize.
Acked-by: NMichal Nazarewicz <mina86@mina86.com>
Signed-off-by: NFelipe F. Tonello <eu@felipetonello.com>
Signed-off-by: NFelipe Balbi <felipe.balbi@linux.intel.com>

Since f_midi_transmit is called by both ALSA and USB sub-systems, it can
potentially cause a race condition between both calls because f_midi_transmit
is not reentrant nor thread-safe. This is due to an implementation detail that
the transmit function looks for the next available usb request from the fifo
and only enqueues it if there is data to send, otherwise just re-uses it. So,
if both ALSA and USB frameworks calls this function at the same time,
kfifo_seek() will return the same usb_request, which will cause a race
condition.

To solve this problem a syncronization mechanism is necessary. In this case it
is used a spinlock since f_midi_transmit is also called by usb_request->complete
callback in interrupt context.

Cc: <stable@vger.kernel.org> # v4.5+
Fixes: e1e3d7ec ("usb: gadget: f_midi: pre-allocate IN requests")
Signed-off-by: NFelipe F. Tonello <eu@felipetonello.com>
Signed-off-by: NFelipe Balbi <balbi@kernel.org>

For every in_substream, there must be a corresponding gmidi_in_port
structure so it is perfectly viable and some might argue sensible to
stash pointer to the input substream in the gmidi_in_port structure.

This has an added benefit that if in_ports < MAX_PORTS, the whole
f_midi structure takes up less space because only in_ports number of
pointers for in_substream are allocated instead of MAX_PORTS lots of
them.
Signed-off-by: NMichal Nazarewicz <mina86@mina86.com>
Signed-off-by: NFelipe Balbi <balbi@kernel.org>

We added a new error path to this function and we forgot to drop the
lock.

Fixes: e1e3d7ec ('usb: gadget: f_midi: pre-allocate IN requests')
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
[mina86@mina86.com: rebased on top of refactoring commit]
Signed-off-by: NMichal Nazarewicz <mina86@mina86.com>
Signed-off-by: NFelipe Balbi <balbi@kernel.org>

Reduce number of allocations, simplify memory management and reduce
memory usage by stacking the gmidi_in_port elements at the end of the
f_midi structure using a flexible array.
Signed-off-by: NMichal Nazarewicz <mina86@mina86.com>
Signed-off-by: NFelipe Balbi <balbi@kernel.org>

In general case, all of midi->in_port pointers may be non-NULL which
implies that the ‘if (\!port)’ condition will never execute thus never
zeroing midi->in_last_port.  Fix by rewriting the loop such that the
field is set to zero if \!port or end of loop has been reached.
Signed-off-by: NMichal Nazarewicz <mina86@mina86.com>
Signed-off-by: NFelipe Balbi <balbi@kernel.org>

Move some of the f_midi_transmit to a separate f_midi_do_transmit
function so the massive indention levels are not so jarring.  This
introduces no changes in behaviour.
Signed-off-by: NMichal Nazarewicz <mina86@mina86.com>
Signed-off-by: NFelipe Balbi <balbi@kernel.org>

remove a field which is unnecessary. No functional changes.
Signed-off-by: NFelipe F. Tonello <eu@felipetonello.com>
Signed-off-by: NMichal Nazarewicz <mina86@mina86.com>
Signed-off-by: NFelipe Balbi <balbi@kernel.org>

This patch introduces pre-allocation of IN endpoint USB requests. This
improves on latency (requires no usb request allocation on transmit) and avoid
several potential probles on allocating too many usb requests (which involves
DMA pool allocation problems).

This implementation also handles better multiple MIDI Gadget ports, always
processing the last processed MIDI substream if the last USB request wasn't
enought to handle the whole stream.
Signed-off-by: NFelipe F. Tonello <eu@felipetonello.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

This ensures that the midi function will only work if the proper number of
IN and OUT requrests are allocated. Otherwise the function will work with less
requests then what the user wants.
Signed-off-by: NFelipe F. Tonello <eu@felipetonello.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

This avoids duplication of USB requests for OUT endpoint and
re-enabling endpoints.
Signed-off-by: NFelipe F. Tonello <eu@felipetonello.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

This function is shared between gadget functions, so this avoid unnecessary
duplicated code and potentially avoid memory leaks.
Reviewed-by: NRobert Baldyga <r.baldyga@samsung.com>
Signed-off-by: NFelipe F. Tonello <eu@felipetonello.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

This code is duplicated from f_midi_start_ep(midi, f, midi->out_ep).
Reviewed-by: NRobert Baldyga <r.baldyga@samsung.com>
Signed-off-by: NFelipe F. Tonello <eu@felipetonello.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

This patch fixes a memory leak that occurs when an endpoint fails to enqueue
the request. If that happens the complete function will never be called, thus
never freeing the request.
Signed-off-by: NFelipe F. Tonello <eu@felipetonello.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

This makes sure f_midi doesn't try to enqueue data when the IN endpoint is
disabled, ie, USB cable is disconnected.
Reviewed-by: NRobert Baldyga <r.baldyga@samsung.com>
Signed-off-by: NFelipe F. Tonello <eu@felipetonello.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

To simplify the configfs interface and remove boilerplate code that also
causes binary bloat.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NAndrzej Pietrasiewicz <andrzej.p@samsung.com>
Acked-by: NFelipe Balbi <balbi@ti.com>
Signed-off-by: NNicholas Bellinger <nab@linux-iscsi.org>

Since ep->driver_data is not used for endpoint claiming, neither for
enabled/disabled state storing, we can reduce number of places where
we read or modify it's value, as now it has no particular meaning for
function or framework logic.

In case of f_midi we only need to store in ep->driver_data pointer to
struct f_midi, as it's used in f_midi_complete() callback and related
functions. All other uses of ep->driver_data are now meaningless and
can be safely removed.
Signed-off-by: NRobert Baldyga <r.baldyga@samsung.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

f_midi is not checking whether there is an error on usb_ep_queue
request, ignoring potential problems, such as memory leaks.
Signed-off-by: NFelipe F. Tonello <eu@felipetonello.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Function midi registers two interfaces with single set_alt() function
which means that f_midi_set_alt() is called twice when configuration
is set. That means that endpoint initialization and ep request allocation
is done two times. To avoid this problem we do such things only once,
for interface number 1 (MIDI Streaming interface).
Signed-off-by: NRobert Baldyga <r.baldyga@samsung.com>
Acked-by: NMichal Nazarewicz <mina86@mina86.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

In case kstrdup() fails the resources to release are
midi->in_port[] and midi. No cards have been registered,
so no need to unregister any.
Signed-off-by: NAndrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

When midi function is created, 'id' attribute is initialized with
SNDRV_DEFAULT_STR1, which is NULL pointer. Trying to read this attribute
before filling it ends up with segmentation fault.

This commit fix this issue by preventing null pointer dereference. Now
f_midi_opts_id_show() returns empty string when id is a null pointer.

Reproduction path:

$ mkdir functions/midi.0
$ cat functions/midi.0/id

[   53.130132] Unable to handle kernel NULL pointer dereference at
virtual address 00000000
[   53.132630] pgd = ec6cc000
[   53.135308] [00000000] *pgd=6b759831, *pte=00000000, *ppte=00000000
[   53.141530] Internal error: Oops: 17 [#1] PREEMPT SMP ARM
[   53.146904] Modules linked in: usb_f_midi snd_rawmidi libcomposite
[   53.153071] CPU: 1 PID: 2936 Comm: cat Not tainted
3.19.0-00041-gcf4b216 #7
[   53.160010] Hardware name: SAMSUNG EXYNOS (Flattened Device Tree)
[   53.166088] task: ee234c80 ti: ec764000 task.ti: ec764000
[   53.171482] PC is at strlcpy+0x8/0x60
[   53.175128] LR is at f_midi_opts_id_show+0x28/0x3c [usb_f_midi]
[   53.181019] pc : [<c0222a9c>]    lr : [<bf01bed0>]    psr: 60000053
[   53.181019] sp : ec765ef8  ip : 00000141  fp : 00000000
[   53.192474] r10: 00019000  r9 : ed7546c0  r8 : 00010000
[   53.197682] r7 : ec765f80  r6 : eb46a000  r5 : eb46a000  r4 :
ed754734
[   53.204192] r3 : ee234c80  r2 : 00001000  r1 : 00000000  r0 :
eb46a000
[   53.210704] Flags: nZCv  IRQs on  FIQs off  Mode SVC_32  ISA ARM
Segment user
[   53.217907] Control: 10c5387d  Table: 6c6cc04a  DAC: 00000015
[   53.223636] Process cat (pid: 2936, stack limit = 0xec764238)
[   53.229364] Stack: (0xec765ef8 to 0xec766000)
[   53.233706] 5ee0:
ed754734 ed7546c0
[   53.241866] 5f00: eb46a000 bf01bed0 eb753b80 bf01cc44 eb753b98
bf01b0a4 bf01b08c c0125dd0
[   53.250025] 5f20: 00002f19 00000000 ec432e00 bf01cce8 c0530c00
00019000 00010000 ec765f80
[   53.258184] 5f40: 00010000 ec764000 00019000 c00cc4ac ec432e00
c00cc55c 00000017 000081a4
[   53.266343] 5f60: 00000001 00000000 00000000 ec432e00 ec432e00
00010000 00019000 c00cc620
[   53.274502] 5f80: 00000000 00000000 00000000 00010000 ffff1000
00019000 00000003 c000e9a8
[   53.282662] 5fa0: 00000000 c000e7e0 00010000 ffff1000 00000003
00019000 00010000 00019000
[   53.290821] 5fc0: 00010000 ffff1000 00019000 00000003 7fffe000
00000001 00000000 00000000
[   53.298980] 5fe0: 00000000 be8c68d4 0000b995 b6f0e3e6 40000070
00000003 00000000 00000000
[   53.307157] [<c0222a9c>] (strlcpy) from [<bf01bed0>]
(f_midi_opts_id_show+0x28/0x3c [usb_f_midi])
[   53.316006] [<bf01bed0>] (f_midi_opts_id_show [usb_f_midi]) from
[<bf01b0a4>] (f_midi_opts_attr_show+0x18/0x24 )
[   53.327209] [<bf01b0a4>] (f_midi_opts_attr_show [usb_f_midi]) from
[<c0125dd0>] (configfs_read_file+0x9c/0xec)
[   53.337180] [<c0125dd0>] (configfs_read_file) from [<c00cc4ac>]
(__vfs_read+0x18/0x4c)
[   53.345073] [<c00cc4ac>] (__vfs_read) from [<c00cc55c>]
(vfs_read+0x7c/0x100)
[   53.352190] [<c00cc55c>] (vfs_read) from [<c00cc620>]
(SyS_read+0x40/0x8c)
[   53.359056] [<c00cc620>] (SyS_read) from [<c000e7e0>]
(ret_fast_syscall+0x0/0x34)
[   53.366513] Code: ebffe3d3 e8bd8008 e92d4070 e1a05000 (e5d14000)
[   53.372641] ---[ end trace e4f53a4e233d98d0 ]---
Signed-off-by: NPawel Szewczyk <p.szewczyk@samsung.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Replace a misspelled function name by %s and then __func__.

This was done using Coccinelle, including the use of Levenshtein distance,
as proposed by Rasmus Villemoes.
Signed-off-by: NJulia Lawall <Julia.Lawall@lip6.fr>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

drivers/usb/gadget/function/f_midi.c:1072:21: sparse: symbol 'f_midi_alloc' was not declared. Should it be static?
drivers/usb/gadget/legacy/gmidi.c:118:30: sparse: symbol 'fi_midi' was not declared. Should it be static?
drivers/usb/gadget/legacy/gmidi.c:119:21: sparse: symbol 'f_midi' was not declared. Should it be static?
Signed-off-by: NFengguang Wu <fengguang.wu@intel.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Make the midi function available for gadgets composed with configfs.
Signed-off-by: NAndrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

In order to add configfs support the usb_gstrings_attach must be used.
Signed-off-by: NAndrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

There are no old f_midi interface users left, so remove it.
Signed-off-by: NAndrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Converting midi to the new function interface requires converting
the USB midi's function code and its users.

This patch converts the f_midi.c to the new function interface.
The file can now be compiled into a separate usb_f_midi.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: NAndrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

kstrdup() might fail, so check its return value and react appropriately.
Signed-off-by: NAndrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

The soundcard index to use for the ALSA device creation is passed as a
parameter to f_midi_bind_config(), but is assigned to midi->index only
after the call to f_midi_register_card(midi). So no matter what is passed
to f_midi_bind_config(), the actual index for snd_card_new() is always 0.
This probably works ok if at the moment of f_midi's bind there are no
other snd_cards, but if there are, it is not possible to bind f_midi.

This patch moves the assignment to a place before the call to
f_midi_register_card(midi).
Signed-off-by: NAndrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

The drivers/usb/gadget directory contains many files.
Files which are related can be distributed into separate directories.
This patch moves the USB functions implementations into a separate directory.
Signed-off-by: NAndrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Also remove superfluous snd_card_set_dev() calls.
Signed-off-by: NTakashi Iwai <tiwai@suse.de>

alloc_ep_req() is a function repeated in several modules.
Make a common implementation and use it.
Signed-off-by: NAndrzej Pietrasiewicz <andrzej.p@samsung.com>
Signed-off-by: NKyungmin Park <kyungmin.park@samsung.com>
Acked-by: NMichal Nazarewicz <mina86@mina86.com>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

HS and SS descriptors are staticaly created. They are updated during the
bind process with the endpoint address, string id or interface numbers.

After that, the descriptor chain is linked to struct usb_function which
is used by composite in order to serve the GET_DESCRIPTOR requests,
number of available configs and so on.

There is no need to assign the HS descriptor only if the UDC supports
HS speed because composite won't report those to the host if HS support
has not been reached. The same reasoning is valid for SS.

This patch makes sure each function updates HS/SS descriptors
unconditionally and uses the newly introduced helper function to create a
copy the descriptors for the speed which is supported by the UDC.

While at that, also rename f->descriptors to f->fs_descriptors in order
to make it more explicit what that means.

Cc: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: NSebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

This function works only on FS or HS. If the gadget is HS capable only
HS descriptors are assigned. If we plug it to an 1.1 host it won't work
because we have only 2.0 descriptors. This patch changes the behavior to
provide both sets.
Signed-off-by: NSebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

The HS descriptors are only created if HS is supported by the UDC but we
never free them.

Cc: stable <stable@vger.kernel.org>
Signed-off-by: NSebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Whether highspeed configuration is valid or not is something that
composite decides and not the gadget. That gadget can only provide the
required descriptors for it.
Signed-off-by: NSebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

Some gadgets provide custom entry here. Some may override it with an
etntry that is also created by composite if there was no value sumbitted
at all.
This patch removes all "custom manufacturer" strings which are the same
as these which are created by composite. Then it moves the creation of
the default manufacturer string to usb_composite_overwrite_options() in
case no command line argument has been used and the entry is still an
empty string.
By doing this we get rid of the global variable "composite_manufacturer"
in composite.
Acked-by: NMichal Nazarewicz <mina86@mina86.com>
Signed-off-by: NSebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: NFelipe Balbi <balbi@ti.com>

As the coverity checker puts it:

"Passing argument "sizeof (midi_function) /*8*/" to function "kcalloc"
and then casting the return value to "struct usb_descriptor_header **"
is suspicious. ... In this particular case sizeof(struct
usb_descriptor_header **) happens to be equal to sizeof(struct
usb_descriptor_header *), but this is not a portable assumption."

I believe we really do intend to use 'sizeof(*midi_function)' here, so
this patch makes that change.
Signed-off-by: NJesper Juhl <jj@chaosbits.net>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>