Merge tag 'usb-serial-4.10-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/johan/usb-serial into usb-linus

Johan writes:

USB-serial fixes for v4.10-rc3

These fixes address a number of long-standing issues in various
USB-serial drivers which would lead to crashes should a malicious device
lack the expected endpoints.

Included are also a few related fixes, and a couple of unrelated ones
that were found during my survey (e.g. a memleak and a
sleep-while-atomic).

A compiler warning revealed an error-handling issue in the new f81534
driver which is also fixed.
Signed-off-by: Johan Hovold <johan@kernel.org>

Fix NULL-pointer dereference in open() should a malicious device lack
the expected endpoints:

Unable to handle kernel NULL pointer dereference at virtual address 00000030
..
[<bf06a6b0>] (ti_open [ti_usb_3410_5052]) from [<bf02e118>] (serial_port_activate+0x68/0x98 [usbserial])

Fixes: 1da177e4 ("Linux-2.6.12-rc2")
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Johan Hovold <johan@kernel.org>

Fix NULL-pointer dereference in open() should the device lack the
expected endpoints:

Unable to handle kernel NULL pointer dereference at virtual address 00000030
...
PC is at spcp8x5_open+0x30/0xd0 [spcp8x5]

Fixes: 619a6f1d ("USB: add usb-serial spcp8x5 driver")
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Johan Hovold <johan@kernel.org>

The write URB was being killed using the synchronous interface while
holding a spin lock in close().

Simply drop the lock and busy-flag update, something which would have
been taken care of by the completion handler if the URB was in flight.

Fixes: f7a33e60 ("USB: serial: add quatech2 usb to serial driver")
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Johan Hovold <johan@kernel.org>

Fix NULL-pointer dereference in open() should a type-0 or type-1 device
lack the expected endpoints:

Unable to handle kernel NULL pointer dereference at virtual address 00000030
...
PC is at pl2303_open+0x38/0xec [pl2303]

Note that a missing interrupt-in endpoint would have caused open() to
fail.

Fixes: 1da177e4 ("Linux-2.6.12-rc2")
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Johan Hovold <johan@kernel.org>

Fix NULL-pointer dereference in open() should the device lack the
expected endpoints:

Unable to handle kernel NULL pointer dereference at virtual address 00000030
...
PC is at oti6858_open+0x30/0x1d0 [oti6858]

Note that a missing interrupt-in endpoint would have caused open() to
fail.

Fixes: 49cdee0e ("USB: oti6858 usb-serial driver (in Nokia CA-42
cable)")
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Johan Hovold <johan@kernel.org>

Fix NULL-pointer dereferences at open() and disconnect() should the
device lack the expected bulk-out endpoints:

Unable to handle kernel NULL pointer dereference at virtual address 000000b4
...
[c0170ff0>] (__lock_acquire) from [<c0172f00>] (lock_acquire+0x108/0x264)
[<c0172f00>] (lock_acquire) from [<c06a5090>] (_raw_spin_lock_irqsave+0x58/0x6c)
[<c06a5090>] (_raw_spin_lock_irqsave) from [<c0470684>] (tty_port_tty_set+0x28/0xa4)
[<c0470684>] (tty_port_tty_set) from [<bf08d384>] (omninet_open+0x30/0x40 [omninet])
[<bf08d384>] (omninet_open [omninet]) from [<bf07c118>] (serial_port_activate+0x68/0x98 [usbserial])

Unable to handle kernel NULL pointer dereference at virtual address 00000234
...
[<bf01f418>] (omninet_disconnect [omninet]) from [<bf0016c0>] (usb_serial_disconnect+0xe4/0x100 [usbserial])

Fixes: 1da177e4 ("Linux-2.6.12-rc2")
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Johan Hovold <johan@kernel.org>

The interrupt URB is killed at final port close since commit
0de9a702 ("USB: overhaul of mos7840 driver").

Fixes: 0de9a702 ("USB: overhaul of mos7840 driver")
Signed-off-by: Johan Hovold <johan@kernel.org>

Remove code to manage a write URB that was never allocated.
Signed-off-by: Johan Hovold <johan@kernel.org>

Fix NULL-pointer dereference in open() should the device lack the
expected endpoints:

Unable to handle kernel NULL pointer dereference at virtual address 00000030
...
PC is at mos7840_open+0x88/0x8dc [mos7840]

Note that we continue to treat the interrupt-in endpoint as optional for
now.

Fixes: 3f542974 ("USB: Moschip 7840 USB-Serial Driver")
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Johan Hovold <johan@kernel.org>

Since commit b69578df ("USB: usbserial: mos7720: add support for
parallel port on moschip 7715"), the interrupt urb is no longer
submitted at first port open and the endpoint-address initialisation at
port-probe is no longer used.
Signed-off-by: Johan Hovold <johan@kernel.org>

A static usb-serial-driver structure that is used to initialise the
interrupt URB was modified during probe depending on the currently
probed device type, something which could break a parallel probe of a
device of a different type.

Fix this up by overriding the default completion callback for MCS7715
devices in attach() instead. We may want to use two usb-serial driver
instances for the two types later.

Fixes: fb088e33 ("USB: serial: add support for serial port on the
moschip 7715")
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Johan Hovold <johan@kernel.org>

Do not submit the interrupt URB until after the parport has been
successfully registered to avoid another use-after-free in the
completion handler when accessing the freed parport private data in case
of a racing completion.

Fixes: b69578df ("USB: usbserial: mos7720: add support for parallel
port on moschip 7715")
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Johan Hovold <johan@kernel.org>

The interrupt URB was submitted on probe but never stopped on probe
errors. This can lead to use-after-free issues in the completion
handler when accessing the freed usb-serial struct:

Unable to handle kernel paging request at virtual address 6b6b6be7
...
[<bf052e70>] (mos7715_interrupt_callback [mos7720]) from [<c052a894>] (__usb_hcd_giveback_urb+0x80/0x140)
[<c052a894>] (__usb_hcd_giveback_urb) from [<c052a9a4>] (usb_hcd_giveback_urb+0x50/0x138)
[<c052a9a4>] (usb_hcd_giveback_urb) from [<c0550684>] (musb_giveback+0xc8/0x1cc)

Fixes: b69578df ("USB: usbserial: mos7720: add support for parallel
port on moschip 7715")
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Johan Hovold <johan@kernel.org>

Fix NULL-pointer dereference at port open if a device lacks the expected
bulk in and out endpoints.

Unable to handle kernel NULL pointer dereference at virtual address 00000030
...
[<bf071c20>] (mos7720_open [mos7720]) from [<bf0490e0>] (serial_port_activate+0x68/0x98 [usbserial])
[<bf0490e0>] (serial_port_activate [usbserial]) from [<c0470ca4>] (tty_port_open+0x9c/0xe8)
[<c0470ca4>] (tty_port_open) from [<bf049d98>] (serial_open+0x48/0x6c [usbserial])
[<bf049d98>] (serial_open [usbserial]) from [<c0469178>] (tty_open+0xcc/0x5cc)

Fixes: 0f64478c ("USB: add USB serial mos7720 driver")
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Johan Hovold <johan@kernel.org>

Fix NULL-pointer dereference in write() should the device lack the
expected interrupt-out endpoint:

Unable to handle kernel NULL pointer dereference at virtual address 00000054
...
PC is at kobil_write+0x144/0x2a0 [kobil_sct]

Fixes: 1da177e4 ("Linux-2.6.12-rc2")
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Johan Hovold <johan@kernel.org>

Check for the expected endpoints in attach() and fail loudly if not
present.

Note that failing to do this appears to be benign since da280e34
("USB: keyspan_pda: clean up write-urb busy handling") which prevents a
NULL-pointer dereference in write() by never marking a non-existent
write-urb as free.

Fixes: 1da177e4 ("Linux-2.6.12-rc2")
Cc: stable <stable@vger.kernel.org>	# < v3.3
Signed-off-by: Johan Hovold <johan@kernel.org>

Fix NULL-pointer dereference at open should the device lack a bulk-in or
bulk-out endpoint:

Unable to handle kernel NULL pointer dereference at virtual address 00000030
...
PC is at iuu_open+0x78/0x59c [iuu_phoenix]

Fixes: 07c3b1a1 ("USB: remove broken usb-serial num_endpoints
check")
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Johan Hovold <johan@kernel.org>

Bind to the interface, but do not register any ports, after having
downloaded the firmware. The device will still disconnect and
re-enumerate, but this way we avoid an error messages from being logged
as part of the process:

io_ti: probe of 1-1.3:1.0 failed with error -5
Signed-off-by: Johan Hovold <johan@kernel.org>

Cancel the heartbeat work on driver unbind in order to avoid I/O after
disconnect in case the port is held open.

Note that the cancel in release() is still needed to stop the heartbeat
after late probe errors.

Fixes: 26c78daa ("USB: io_ti: Add heartbeat to keep idle EP/416
ports from disconnecting")
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Johan Hovold <johan@kernel.org>

In case a device is left in "boot-mode" we must not register any port
devices in order to avoid a NULL-pointer dereference on open due to
missing endpoints. This could be used by a malicious device to trigger
an OOPS:

Unable to handle kernel NULL pointer dereference at virtual address 00000030
...
[<bf0caa84>] (edge_open [io_ti]) from [<bf0b0118>] (serial_port_activate+0x68/0x98 [usbserial])
[<bf0b0118>] (serial_port_activate [usbserial]) from [<c0470ca4>] (tty_port_open+0x9c/0xe8)
[<c0470ca4>] (tty_port_open) from [<bf0b0da0>] (serial_open+0x48/0x6c [usbserial])
[<bf0b0da0>] (serial_open [usbserial]) from [<c0469178>] (tty_open+0xcc/0x5cc)

Fixes: 1da177e4 ("Linux-2.6.12-rc2")
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Johan Hovold <johan@kernel.org>

Fix NULL-pointer dereference when clearing halt at open should a
malicious device lack the expected endpoints when in download mode.

Unable to handle kernel NULL pointer dereference at virtual address 00000030
...
[<bf011ed8>] (edge_open [io_ti]) from [<bf000118>] (serial_port_activate+0x68/0x98 [usbserial])
[<bf000118>] (serial_port_activate [usbserial]) from [<c0470ca4>] (tty_port_open+0x9c/0xe8)
[<c0470ca4>] (tty_port_open) from [<bf000da0>] (serial_open+0x48/0x6c [usbserial])
[<bf000da0>] (serial_open [usbserial]) from [<c0469178>] (tty_open+0xcc/0x5cc)

Fixes: 1da177e4 ("Linux-2.6.12-rc2")
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Johan Hovold <johan@kernel.org>

Fix NULL-pointer dereference when initialising URBs at open should a
non-EPIC device lack a bulk-in or interrupt-in endpoint.

Unable to handle kernel NULL pointer dereference at virtual address 00000028
...
PC is at edge_open+0x24c/0x3e8 [io_edgeport]

Note that the EPIC-device probe path has the required sanity checks so
this makes those checks partially redundant.

Fixes: 1da177e4 ("Linux-2.6.12-rc2")
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Johan Hovold <johan@kernel.org>

Make sure to free the URB transfer buffer in case submission fails (e.g.
due to a disconnect).

Fixes: 1da177e4 ("Linux-2.6.12-rc2")
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Johan Hovold <johan@kernel.org>

Fix NULL-pointer dereference when clearing halt at open should the device
lack a bulk-out endpoint.

Unable to handle kernel NULL pointer dereference at virtual address 00000030
...
PC is at cyberjack_open+0x40/0x9c [cyberjack]

Fixes: 1da177e4 ("Linux-2.6.12-rc2")
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Johan Hovold <johan@kernel.org>

Stop Endpoint command can come at any point and we
have no control of that. We should make sure to
handle COMP_STOP on SETUP phase as well, otherwise
urb->actual_length might be set to negative values
in some occasions such as below:

 urb->length = 4;
 build_control_transfer_td_for(urb, ep);

 					stop_endpoint(ep);

COMP_STOP:
	[...]
	urb->actual_length = urb->length - trb->length;

trb->length is 8 for SETUP stage (8 control request
bytes), so actual_length would be set to -4 in this
case.

While doing that, also make sure to use TRB_TYPE
field of the actual TRB instead of matching pointers
to figure out in which stage of the control transfer
we got our completion event.

Cc: <stable@vger.kernel.org>
Signed-off-by: Felipe Balbi <felipe.balbi@linux.intel.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Intel Apollo Lake also requires XHCI_PME_STUCK_QUIRK.
Adding its PCI ID to quirk.

Cc: <stable@vger.kernel.org>
Signed-off-by: Wan Ahmad Zainie <wan.ahmad.zainie.wan.mohamad@intel.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Current abort operation has race.

    xhci_handle_command_timeout()
      xhci_abort_cmd_ring()
        xhci_write_64(CMD_RING_ABORT)
        xhci_handshake(5s)
	  do {
	    check CMD_RING_RUNNING
            udelay(1)
					 ...
					 COMP_CMD_ABORT event
					 COMP_CMD_STOP event
					 xhci_handle_stopped_cmd_ring()
					   restart cmd_ring
                                           CMD_RING_RUNNING become 1 again
	  } while ()
          return -ETIMEDOUT
        xhci_write_64(CMD_RING_ABORT)
        /* can abort random command */

To do abort operation correctly, we have to wait both of COMP_CMD_STOP
event and negation of CMD_RING_RUNNING.

But like above, while timeout handler is waiting negation of
CMD_RING_RUNNING, event handler can restart cmd_ring. So timeout
handler never be notice negation of CMD_RING_RUNNING, and retry of
CMD_RING_ABORT can abort random command (BTW, I guess retry of
CMD_RING_ABORT was workaround of this race).

To fix this race, this moves xhci_handle_stopped_cmd_ring() to
xhci_abort_cmd_ring().  And timeout handler waits COMP_CMD_STOP event.

At this point, timeout handler is owner of cmd_ring, and safely
restart cmd_ring by using xhci_handle_stopped_cmd_ring().

[FWIW, as bonus, this way would be easily extend to add CMD_RING_PAUSE
operation]

[locks edited as patch is rebased on other locking fixes -Mathias]
Signed-off-by: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

This is preparation to fix abort operation race (See "xhci: Fix race
related to abort operation"). To make timeout sleepable, use
delayed_work instead of timer.

[change a newly added pending timer fix to pending work -Mathias]
Signed-off-by: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

In command timer function, xhci_handle_command_timeout(), xhci->lock
is unlocked before call into xhci_abort_cmd_ring(). This might cause
race between the timer function and the event handler.

The xhci_abort_cmd_ring() function sets the CMD_RING_ABORT bit in the
command register and polling it until the setting takes effect. A stop
command ring event might be handled between writing the abort bit and
polling for it. The event handler will restart the command ring, which
causes the failure of polling, and we ever believed that we failed to
stop it.

As a bonus, this also fixes some issues of calling functions without
locking in xhci_handle_command_timeout().

Cc: <stable@vger.kernel.org> # 3.7+
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

If we get a command completion event at the same time as the command
timeout work starts on another cpu we might end up aborting the wrong
command.

If the command completion takes the xhci lock before the timeout work, it
will handle the command, pick the next command, mark it as current_cmd, and
re-queue the timeout work. When the timeout work finally gets the lock
It will start aborting the wrong command.

This case can be resolved by checking if the timeout work is pending inside
the timeout function itself. A new timeout work can only be pending if the
command completed and a new command was queued.

If there are no more commands pending then command completion will set
the current_cmd to NULL, which is already handled in the timeout work.

Cc: <stable@vger.kernel.org>
Reported-by: Baolin Wang <baolin.wang@linaro.org>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

When current command was supposed to be aborted, host will free the command
in handle_cmd_completion() function. But it might be still referenced by
xhci->current_cmd, which need to set NULL.

Cc: <stable@vger.kernel.org>
Signed-off-by: Baolin Wang <baolin.wang@linaro.org>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

handle_cmd_completion() frees a command structure which might be still
referenced by xhci->current_cmd.
This might cause problem when xhci->current_cmd is accessed after that.

A real-life case could be like this. The host takes a very long time to
respond to a command, and the command timer is fired at the same time
when the command completion event arrives. The command completion
handler frees xhci->current_cmd before the timer function can grab
xhci->lock. Afterward, timer function grabs the lock and go ahead with
checking and setting members of xhci->current_cmd.

Cc: <stable@vger.kernel.org> # v3.16+
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

In function xhci_mtk_probe(), variable ret takes the return value. Its
value should be negative on failures. However, when the call to function
platform_get_irq() fails, it does not set the error code, and 0 will be
returned. 0 indicates no error. As a result, the callers of function
xhci_mtk_probe() will not be able to detect the error. This patch fixes
the bug by assigning the return value of platform_get_irq() to variable
ret if it fails.

CC: <stable@vger.kernel.org>
Signed-off-by: Pan Bian <bianpan2016@163.com>
Reviewed-by: Matthias Brugger <matthias.bgg@gmail.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

xhci_setup_device() should return failure with correct error number
when xhci host has died, removed or halted.

During usb device enumeration, if usb host is not accessible (died,
removed or halted), the hc_driver->address_device() should return
a corresponding error code to usb core. But current xhci driver just
returns success. This misleads usb core to continue the enumeration
by reading the device descriptor, which will result in failure, and
users will get a misleading message like "device descriptor read/8,
error -110".

Cc: <stable@vger.kernel.org> # v4.3+
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

the tt_info provided by a HS hub might be in use to by a child device
Make sure we free the devices in the correct order.

This is needed in special cases such as when xhci controller is
reset when resuming from hibernate, and all virt_devices are freed.

Also free the virt_devices starting from max slot_id as children
more commonly have higher slot_id than parent.

CC: <stable@vger.kernel.org>
Reported-by: Guenter Roeck <groeck@chromium.org>
Tested-by: Guenter Roeck <groeck@chromium.org>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Merge tag 'fixes-for-v4.10-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/balbi/usb into usb-linus

Felipe writes:

usb: fixes for v4.10-rc3

The first set of fixes for v4.10-rc cycle. The most
important of which is a big regression on dwc3-pci
which prevents it from probing altogether.

There's also a fix to avoid Overflow events on DWC3
and another to make sure we don't starve DMA
resources.

Dummy HCD got some love after a long hiatus and DWC2
got a couple fixes related to DMA usage. Other than
these, we have a set of minor fixes here and there.

This fixes an error message that was probably copied and pasted. The same
message is used for both the in and out endpoints, so it makes it impossible
to know which one actually failed because both cases say "IN".

Make the out endpoint error message say "OUT".
Signed-off-by: David Lechner <david@lechnology.com>
Signed-off-by: Felipe Balbi <felipe.balbi@linux.intel.com>

DCFG.DEVSPD == 0x3 is not valid and we need to set
DCFG.DEVSPD to 0x1 for full speed mode. Same goes for
DSTS.CONNECTSPD.

Old databooks had 0x3 for full speed in 48MHz mode for
USB1.1 transceivers which was never supported. Newer databooks
don't mention 0x3 at all.

Cc: John Youn <John.Youn@synopsys.com>
Signed-off-by: Roger Quadros <rogerq@ti.com>
Signed-off-by: Felipe Balbi <felipe.balbi@linux.intel.com>

With gcc 4.1.2:

    drivers/usb/serial/f81534.c: In function ‘f81534_port_probe’:
    drivers/usb/serial/f81534.c:1250: warning: comparison is always false due to limited range of data type

f81534_logic_to_phy_port() may return a negative error value, which is
ignored by assigning it to u8 f81534_port_private.phy_num.

Use an intermediate variable of type int to fix this.
While at it, forward the actual error code instead of converting it to
-ENODEV, and drop the useless check for F81534_NUM_PORT, as the callee
always returns a valid port number in case of success.

Fixes: 0c9bd600 ("USB: serial: add Fintek F81532/534 driver")
Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: Johan Hovold <johan@kernel.org>