For the Processing Unit and Camera Terminal descriptors defined in
the UVC Gadget we currently hard-code values into their bmControls
fields, which enumerates which controls the gadget is able to
support. This isn't appropriate since only the userspace companion
program to the kernel driver will know which controls are supported.
Make the configfs attributes that point to those fields read/write
so userspace can set them to appropriate values.
Document the new behaviour at the same time so the functionality is
clear.
Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
Link: https://lore.kernel.org/r/20230309105825.216745-1-dan.scally@ideasonboard.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
We need the USB fixes here, and the USB gadget update for future
development patches to be based on.
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Mike Christie <michael.christie@oracle.com> says:
The following patches apply over Martin's 6.4 branches and Linus's tree.
They fix a couple regressions in iscsit that occur when there are TMRs
executing and a connection is closed. It also includes Dimitry's fixes in
related code paths for cmd cleanup when ERL2 is used and the write pending
hang during conn cleanup.
This version of the patchset brings it back to just regressions and fixes
for bugs we have a lot of users hitting. I'm going to fix isert and get it
hooked into iscsit properly in a second patchset, because this one was
getting so large. I've also moved my cleanup type of patches for a 3rd
patchset.
Link: https://lore.kernel.org/r/20230319015620.96006-1-michael.christie@oracle.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Currently, the USB gadget framework supports only version 1.0 of the MS OS
descriptor. OS desc has a field bcdVersion indicating its version, with
v1.0 represented by the value 0x100. However, __ffs_do_os_desc_header()
was expecting the incorrect value 0x1, so allow the correct value 0x100.
The bcdVersion field of the descriptor that is actually sent to the host
is set by composite_setup() (in composite.c) to the fixed value 0x100.
Therefore, it can be understood that __ffs_do_os_desc_header() is only
performing a format check of the OS desc passed to functionfs. If a value
other than 0x100 is accepted, there is no effect on communication over
the USB bus. Indeed, until now __ffs_do_os_desc_header() has only accepted
the incorrect value 0x1, but there was no problem with the communication
over the USB bus.
However, this can be confusing for functionfs userspace drivers. Since
bcdVersion=0x100 is used in actual communication, functionfs should accept
the value 0x100.
Note that the correct value for bcdVersion in OS desc v1.0 is 0x100, but
to avoid breaking old userspace drivers, the value 0x1 is also accepted as
an exception. At this time, a message is output to notify the user to fix
the userspace driver.
Signed-off-by: Yuta Hayama <hayama@lineo.co.jp>
Reviewed-by: Andrzej Pietrasiewicz <andrzej.p@collabora.com>
Link: https://lore.kernel.org/r/290f96db-1877-5137-373a-318e7b4f2dde@lineo.co.jp
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In the unbind callback for f_uac1 and f_uac2, a call to snd_card_free()
via g_audio_cleanup() will disconnect the card and then wait for all
resources to be released, which happens when the refcount falls to zero.
Since userspace can keep the refcount incremented by not closing the
relevant file descriptor, the call to unbind may block indefinitely.
This can cause a deadlock during reboot, as evidenced by the following
blocked task observed on my machine:
task:reboot state:D stack:0 pid:2827 ppid:569 flags:0x0000000c
Call trace:
__switch_to+0xc8/0x140
__schedule+0x2f0/0x7c0
schedule+0x60/0xd0
schedule_timeout+0x180/0x1d4
wait_for_completion+0x78/0x180
snd_card_free+0x90/0xa0
g_audio_cleanup+0x2c/0x64
afunc_unbind+0x28/0x60
...
kernel_restart+0x4c/0xac
__do_sys_reboot+0xcc/0x1ec
__arm64_sys_reboot+0x28/0x30
invoke_syscall+0x4c/0x110
...
The issue can also be observed by opening the card with arecord and
then stopping the process through the shell before unbinding:
# arecord -D hw:UAC2Gadget -f S32_LE -c 2 -r 48000 /dev/null
Recording WAVE '/dev/null' : Signed 32 bit Little Endian, Rate 48000 Hz, Stereo
^Z[1]+ Stopped arecord -D hw:UAC2Gadget -f S32_LE -c 2 -r 48000 /dev/null
# echo gadget.0 > /sys/bus/gadget/drivers/configfs-gadget/unbind
(observe that the unbind command never finishes)
Fix the problem by using snd_card_free_when_closed() instead, which will
still disconnect the card as desired, but defer the task of freeing the
resources to the core once userspace closes its file descriptor.
Fixes: 132fcb4608 ("usb: gadget: Add Audio Class 2.0 Driver")
Cc: stable@vger.kernel.org
Signed-off-by: Alvin Šipraga <alsi@bang-olufsen.dk>
Reviewed-by: Ruslan Bilovol <ruslan.bilovol@gmail.com>
Reviewed-by: John Keeping <john@metanate.com>
Link: https://lore.kernel.org/r/20230302163648.3349669-1-alvin@pqrs.dk
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The function gether_setup_name_default() is called by various USB
ethernet gadget drivers. Calling this function will select a random
host and device MAC addresses. A properly working driver should be
silent and not warn the user about default MAC addresses selection.
Given that the MAC addresses are also printed when the function
gether_register_netdev() is called, remove these unnecessary warnings.
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Reviewed-by: Andrzej Pietrasiewicz <andrzej.p@collabora.com>
Link: https://lore.kernel.org/r/20230209125319.18589-2-jonathanh@nvidia.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The USB ethernet gadget driver implements its own print macros which
call printk. Device drivers should use the device prints that print the
device name. Fortunately, the same macro names are defined in the header
file 'linux/usb/composite.h' and these use the device prints. Therefore,
remove the local definitions in the USB ethernet gadget driver and use
those in 'linux/usb/composite.h'. The only difference is that now the
device name is printed instead of the ethernet interface name.
Tested using ethernet gadget on Jetson AGX Orin.
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Tested-by: Jon Hunter <jonathanh@nvidia.com>
Link: https://lore.kernel.org/r/20230209125319.18589-1-jonathanh@nvidia.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Allow users to create new color matching descriptors in addition to
the default one. These must be associated with a UVC format in order
to be transmitted to the host, which is achieved by symlinking from
the format to the newly created color matching descriptor - extend
the uncompressed and mjpeg formats to support that linking operation.
Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
Link: https://lore.kernel.org/r/20230202114142.300858-7-dan.scally@ideasonboard.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
As currently implemented the default color matching descriptor is
appended after _all_ the formats and frames that the gadget is
configured with. According to the UVC specifications however this
is supposed to be on a per-format basis (section 3.9.2.6):
"Only one instance is allowed for a given format and if present,
the Color Matching descriptor shall be placed following the Video
and Still Image Frame descriptors for that format."
Associate the default color matching descriptor with struct
uvcg_format and copy it once-per-format instead of once only.
Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
Link: https://lore.kernel.org/r/20230202114142.300858-4-dan.scally@ideasonboard.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Color matching descriptors are meant to be a per-format piece of data
and we need to be able to support different descriptors for different
formats. As a preliminary step towards that goal, switch the default
color matching configfs functionality to point to an instance of a
new struct uvcg_color_matching. Use the same default values for its
attributes as the currently hard-coded ones so that the interface to
userspace is consistent.
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
Link: https://lore.kernel.org/r/20230202114142.300858-3-dan.scally@ideasonboard.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
We need the USB fixes in here, and this resolves a merge conflict with
the i915 driver as reported in linux-next
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The f_uvc code includes an interrupt endpoint against the VideoControl
interface. According to section 2.4.2 of the UVC specification however
this endpoint is optional in at least some cases:
"This endpoint is optional, but may be mandatory under certain
conditions"
The conditions enumerated are whether...
1. The device supports hardware triggers
2. The device implements any AutoUpdate controls
3. The device implements any Asynchronous controls
As all of those things are implementation dependent, this endpoint
might be unnecessary for some users. Further to that it is unusable
in the current implementation as there is no mechanism within the
UVC gadget driver that allows data to be sent over that endpoint.
Disable the interrupt endpoint by default, but check whether the
user has asked for it to be enabled in configfs and continue to
generate it if so.
Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
Link: https://lore.kernel.org/r/20230130105045.120886-4-dan.scally@ideasonboard.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The f_uvc code defines an endpoint named "uvc_control_ep" but it
is configured with a non-zero endpoint address and has its
bmAttributes flagged as USB_ENDPOINT_XFER_INT - this cannot be the
VideoControl interface's control endpoint, as the default endpoint
0 is used for that purpose. This is instead the optional interrupt
endpoint that can be contained by a VideoControl interface. There
is also a Class-specific VC Interrupt Endpoint Descriptor and a
SuperSpeed companion descriptor that are also for the VC interface's
interrupt endpoint but are named as though they are for the control
endpoint.
Rename the variables to make that clear.
Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
Link: https://lore.kernel.org/r/20230130105045.120886-2-dan.scally@ideasonboard.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Currently connect/disconnect of USB cable calls afunc_bind and
eventually increments the bNumEndpoints. Performing multiple
plugin/plugout will increment bNumEndpoints incorrectly, and on
the next plug-in it leads to invalid configuration of descriptor
and hence enumeration fails.
Fix this by resetting the value of bNumEndpoints to 1 on every
afunc_bind call.
Fixes: 40c73b3054 ("usb: gadget: f_uac2: add adaptive sync support for capture")
Cc: stable <stable@kernel.org>
Signed-off-by: Pratham Pratap <quic_ppratap@quicinc.com>
Signed-off-by: Prashanth K <quic_prashk@quicinc.com>
Link: https://lore.kernel.org/r/1674631645-28888-1-git-send-email-quic_prashk@quicinc.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
__ffs_ep0_queue_wait executes holding the spinlock of &ffs->ev.waitq.lock
and unlocks it after the assignments to usb_request are done.
However in the code if the request is already NULL we bail out returning
-EINVAL but never unlocked the spinlock.
Fix this by adding spin_unlock_irq &ffs->ev.waitq.lock before returning.
Fixes: 6a19da1110 ("usb: gadget: f_fs: Prevent race during ffs_ep0_queue_wait")
Reviewed-by: John Keeping <john@metanate.com>
Signed-off-by: Udipto Goswami <quic_ugoswami@quicinc.com>
Link: https://lore.kernel.org/r/20230124091149.18647-1-quic_ugoswami@quicinc.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
We need the USB fixes in here and this resolves merge conflicts as
reported in linux-next in the following files:
drivers/usb/host/xhci.c
drivers/usb/host/xhci.h
drivers/usb/typec/ucsi/ucsi.c
Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In Google internal bug 265639009 we've received an (as yet) unreproducible
crash report from an aarch64 GKI 5.10.149-android13 running device.
AFAICT the source code is at:
https://android.googlesource.com/kernel/common/+/refs/tags/ASB-2022-12-05_13-5.10
The call stack is:
ncm_close() -> ncm_notify() -> ncm_do_notify()
with the crash at:
ncm_do_notify+0x98/0x270
Code: 79000d0b b9000a6c f940012a f9400269 (b9405d4b)
Which I believe disassembles to (I don't know ARM assembly, but it looks sane enough to me...):
// halfword (16-bit) store presumably to event->wLength (at offset 6 of struct usb_cdc_notification)
0B 0D 00 79 strh w11, [x8, #6]
// word (32-bit) store presumably to req->Length (at offset 8 of struct usb_request)
6C 0A 00 B9 str w12, [x19, #8]
// x10 (NULL) was read here from offset 0 of valid pointer x9
// IMHO we're reading 'cdev->gadget' and getting NULL
// gadget is indeed at offset 0 of struct usb_composite_dev
2A 01 40 F9 ldr x10, [x9]
// loading req->buf pointer, which is at offset 0 of struct usb_request
69 02 40 F9 ldr x9, [x19]
// x10 is null, crash, appears to be attempt to read cdev->gadget->max_speed
4B 5D 40 B9 ldr w11, [x10, #0x5c]
which seems to line up with ncm_do_notify() case NCM_NOTIFY_SPEED code fragment:
event->wLength = cpu_to_le16(8);
req->length = NCM_STATUS_BYTECOUNT;
/* SPEED_CHANGE data is up/down speeds in bits/sec */
data = req->buf + sizeof *event;
data[0] = cpu_to_le32(ncm_bitrate(cdev->gadget));
My analysis of registers and NULL ptr deref crash offset
(Unable to handle kernel NULL pointer dereference at virtual address 000000000000005c)
heavily suggests that the crash is due to 'cdev->gadget' being NULL when executing:
data[0] = cpu_to_le32(ncm_bitrate(cdev->gadget));
which calls:
ncm_bitrate(NULL)
which then calls:
gadget_is_superspeed(NULL)
which reads
((struct usb_gadget *)NULL)->max_speed
and hits a panic.
AFAICT, if I'm counting right, the offset of max_speed is indeed 0x5C.
(remember there's a GKI KABI reservation of 16 bytes in struct work_struct)
It's not at all clear to me how this is all supposed to work...
but returning 0 seems much better than panic-ing...
Cc: Felipe Balbi <balbi@kernel.org>
Cc: Lorenzo Colitti <lorenzo@google.com>
Cc: Carlos Llamas <cmllamas@google.com>
Cc: stable@vger.kernel.org
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Cc: stable <stable@kernel.org>
Link: https://lore.kernel.org/r/20230117131839.1138208-1-maze@google.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
As per the documentation, function usb_ep_free_request guarantees
the request will not be queued or no longer be re-queued (or
otherwise used). However, with the current implementation it
doesn't make sure that the request in ep0 isn't reused.
Fix this by dequeuing the ep0req on functionfs_unbind before
freeing the request to align with the definition.
Fixes: ddf8abd259 ("USB: f_fs: the FunctionFS driver")
Signed-off-by: Udipto Goswami <quic_ugoswami@quicinc.com>
Tested-by: Krishna Kurapati <quic_kriskura@quicinc.com>
Link: https://lore.kernel.org/r/20221215052906.8993-3-quic_ugoswami@quicinc.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
While performing fast composition switch, there is a possibility that the
process of ffs_ep0_write/ffs_ep0_read get into a race condition
due to ep0req being freed up from functionfs_unbind.
Consider the scenario that the ffs_ep0_write calls the ffs_ep0_queue_wait
by taking a lock &ffs->ev.waitq.lock. However, the functionfs_unbind isn't
bounded so it can go ahead and mark the ep0req to NULL, and since there
is no NULL check in ffs_ep0_queue_wait we will end up in use-after-free.
Fix this by making a serialized execution between the two functions using
a mutex_lock(ffs->mutex).
Fixes: ddf8abd259 ("USB: f_fs: the FunctionFS driver")
Signed-off-by: Udipto Goswami <quic_ugoswami@quicinc.com>
Tested-by: Krishna Kurapati <quic_kriskura@quicinc.com>
Link: https://lore.kernel.org/r/20221215052906.8993-2-quic_ugoswami@quicinc.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Commit fb1f16d74e ("usb: gadget: f_fs: change ep->status safe in
ffs_epfile_io()") added a new ffs_io_data::status field to fix lifetime
issues in synchronous requests.
While there are no similar lifetime issues for asynchronous requests
(the separate ep member in ffs_io_data avoids them) using the status
field means the USB request can be freed earlier and that there is more
consistency between the synchronous and asynchronous I/O paths.
Cc: Linyu Yuan <quic_linyyuan@quicinc.com>
Signed-off-by: John Keeping <john@metanate.com>
Reviewed-by: Linyu Yuan <quic_linyyuan@quicinc.com>
Link: https://lore.kernel.org/r/20221124170430.3998755-1-john@metanate.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In the specification documents for the Uncompressed and MJPEG USB
Video Payloads, the field name is bmInterlaceFlags - it has been
misnamed within the kernel.
Although renaming the field does break the kernel's interface to
userspace it should be low-risk in this instance. The field is read
only and hardcoded to 0, so there was never any value in anyone
reading it. A search of the uvc-gadget application and all the
forks that I could find for it did not reveal any users either.
Fixes: cdda479f15 ("USB gadget: video class function driver")
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Kieran Bingham <kieran.bingham@ideasonboard.com>
Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
Link: https://lore.kernel.org/r/20221206161203.1562827-1-dan.scally@ideasonboard.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The embedded struct cdev does not have its lifetime correctly tied to
the enclosing struct f_hidg, so there is a use-after-free if /dev/hidgN
is held open while the gadget is deleted.
This can readily be replicated with libusbgx's example programs (for
conciseness - operating directly via configfs is equivalent):
gadget-hid
exec 3<> /dev/hidg0
gadget-vid-pid-remove
exec 3<&-
Pull the existing device up in to struct f_hidg and make use of the
cdev_device_{add,del}() helpers. This changes the lifetime of the
device object to match struct f_hidg, but note that it is still added
and deleted at the same time.
Fixes: 71adf11894 ("USB: gadget: add HID gadget driver")
Tested-by: Lee Jones <lee@kernel.org>
Reviewed-by: Andrzej Pietrasiewicz <andrzej.p@collabora.com>
Reviewed-by: Lee Jones <lee@kernel.org>
Signed-off-by: John Keeping <john@metanate.com>
Link: https://lore.kernel.org/r/20221122123523.3068034-2-john@metanate.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
