xHCI host controllers can have both USB 3.1 and 3.0 extended speed
protocol lists. If the USB3.1 speed is parsed first and 3.0 second then
the minor revision supported will be overwritten by the 3.0 speeds and
the USB3 roothub will only show support for USB 3.0 speeds.
This was the case with a xhci controller with the supported protocol
capability listed below.
In xhci-mem.c, the USB 3.1 speed is parsed first, the min_rev of usb3_rhub
is set as 0x10. And then USB 3.0 is parsed. However, the min_rev of
usb3_rhub will be changed to 0x00. If USB 3.1 device is connected behind
this host controller, the speed of USB 3.1 device just reports 5G speed
using lsusb.
00 01 02 03 04 05 06 07 08 09 0A 0B 0C 0D 0E 0F
00 01 08 00 00 00 00 00 40 00 00 00 00 00 00 00 00
10 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
20 02 08 10 03 55 53 42 20 01 02 00 00 00 00 00 00 //USB 3.1
30 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
40 02 08 00 03 55 53 42 20 03 06 00 00 00 00 00 00 //USB 3.0
50 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
60 02 08 00 02 55 53 42 20 09 0E 19 00 00 00 00 00 //USB 2.0
70 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
This patch fixes the issue by only owerwriting the minor revision if
it is higher than the existing one.
[reword commit message -Mathias]
Cc: <stable@vger.kernel.org>
Signed-off-by: YD Tseng <yd_tseng@asmedia.com.tw>
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Felipe writes:
usb: fixes for v4.12-rc5
Alan Stern fixed a GPF in gadgetfs found by the kernel fuzzying project
composite.c learned that if it deactivates a function during bind, it
must reactivate it during unbind.
Network devices can allocate reasources and private memory using
netdev_ops->ndo_init(). However, the release of these resources
can occur in one of two different places.
Either netdev_ops->ndo_uninit() or netdev->destructor().
The decision of which operation frees the resources depends upon
whether it is necessary for all netdev refs to be released before it
is safe to perform the freeing.
netdev_ops->ndo_uninit() presumably can occur right after the
NETDEV_UNREGISTER notifier completes and the unicast and multicast
address lists are flushed.
netdev->destructor(), on the other hand, does not run until the
netdev references all go away.
Further complicating the situation is that netdev->destructor()
almost universally does also a free_netdev().
This creates a problem for the logic in register_netdevice().
Because all callers of register_netdevice() manage the freeing
of the netdev, and invoke free_netdev(dev) if register_netdevice()
fails.
If netdev_ops->ndo_init() succeeds, but something else fails inside
of register_netdevice(), it does call ndo_ops->ndo_uninit(). But
it is not able to invoke netdev->destructor().
This is because netdev->destructor() will do a free_netdev() and
then the caller of register_netdevice() will do the same.
However, this means that the resources that would normally be released
by netdev->destructor() will not be.
Over the years drivers have added local hacks to deal with this, by
invoking their destructor parts by hand when register_netdevice()
fails.
Many drivers do not try to deal with this, and instead we have leaks.
Let's close this hole by formalizing the distinction between what
private things need to be freed up by netdev->destructor() and whether
the driver needs unregister_netdevice() to perform the free_netdev().
netdev->priv_destructor() performs all actions to free up the private
resources that used to be freed by netdev->destructor(), except for
free_netdev().
netdev->needs_free_netdev is a boolean that indicates whether
free_netdev() should be done at the end of unregister_netdevice().
Now, register_netdevice() can sanely release all resources after
ndo_ops->ndo_init() succeeds, by invoking both ndo_ops->ndo_uninit()
and netdev->priv_destructor().
And at the end of unregister_netdevice(), we invoke
netdev->priv_destructor() and optionally call free_netdev().
Signed-off-by: David S. Miller <davem@davemloft.net>
If a function sets bind_deactivated flag, upon removal we will be left
with an unbalanced deactivation. Let's make sure that we conditionally
call usb_function_activate() from usb_remove_function() and make sure
usb_remove_function() is called from remove_config().
Signed-off-by: Felipe Balbi <felipe.balbi@linux.intel.com>
This driver is no longer needed and can be removed. The reason why
it's safe to remove this driver is that most omap devices don't have a
USB low-speed or full-speed compatible PHY installed and configured
with drivers/mfd/omap-usb-host.c. This means that devices like
beagleboard and pandaboard need to use a high-speed USB hub in order
to use devices like keyboard and mice.
Currently the only known configured for a full-speed PHY is the
mdm6600 modem on droid 4 and I've verified it works just fine with
ohci-platform.
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Acked-by: Roger Quadros <rogerq@ti.com>
Reviewed-by: Sebastian Reichel <sebastian.reichel@collabora.co.uk>
Signed-off-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
User space applications in some cases have the need to enforce a
specific port type(DFP/UFP/DRP). This change allows userspace to
attempt setting the desired port type. Low level drivers can
however reject the request if the specific port type is not supported.
Signed-off-by: Badhri Jagan Sridharan <Badhri@google.com>
Reviewed-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
If USB PD contract is established after creation of the
partner, the power delivery support attribute of the partner
needs to be updated separately. This can be done in
typec_set_pwr_opmode() by checking if the port has already
partner and updating the value if it does.
Signed-off-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>
Reviewed-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Add xhci_get_hw_deq() helper to retrieve the hardware dequeue pointer an
endpoint or stream stopped on.
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The values for the new dequeue segment, new dequeue pointer and new cycle
state are needed for manually moving the xHC ring dequeue pointer.
These are conveniently stored in a xhci_dequeue_state structure.
stream support was added later and stream_id was carried
as a function parameter.
Move the stream_id to the xhci_dequeue_state structure instead.
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When xHC is asked to stop an endpoint it will save the position it
stopped on in the endpoint or stream context.
xhci driver needs to know if the controller stopped on the exact same
TRB that the driver was asked to cancel as it then needs to move past
the TD instead of turning the TD to no-op TRBs.
xhci driver used to get the stopped position from a "stopped" transfer
event before the stop endpoint command completed, but if the ring
is already stopped, or in a halted or error state this event is missing.
Get the stopped position from the endpoint or stream context instead
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This optimization significantly reduces xhci driver load time.
In ACPI tables the acpi companion port devices are children of
the hub device. The port devices are identified by their port number
returned by the ACPI _ADR method.
_ADR 0 is reserved for the root hub device.
The current implementation to find a acpi companion port device
loops through all acpi port devices under that parent hub, evaluating
their _ADR method each time a new port device is added.
for a xHC controller with 25 ports under its roothub it
will end up invoking ACPI bytecode 625 times before all ports
are ready, making it really slow.
The _ADR values are already read and cached earler. So instead of
running the bytecode again we can check the cached _ADR value first,
and then fall back to the old way.
As one of the more significant changes, the xhci load time on
Intel kabylake reduced by 70%, (28ms) from
initcall xhci_pci_init+0x0/0x49 returned 0 after 39537 usecs
to
initcall xhci_pci_init+0x0/0x49 returned 0 after 11270 usecs
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
We no longer keep track of where we stopped in a stopped_td pointer.
We get the ring dequeue pointer from the endpoint or stream context
Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Felipe writes:
usb: fixes for v4.12-rc4
A fix to a really old synchronization bug on mass storage gadget.
Support for Meson8 SoCs on dwc2
Synchronization fixes on renesas USB driver.
With the new API to perform the in-kernel buffer copy, we can get rid
of set_fs() usage in this driver, finally.
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Reviewed-by: Takashi Sakamoto <o-takashi@sakamocchi.jp>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
The PN_INT_ENA register should be used after usb3_pn_change() is called.
So, this patch moves the access from renesas_usb3_stop_controller() to
usb3_disable_pipe_n().
Fixes: 746bfe63bb ("usb: gadget: renesas_usb3: add support for Renesas USB3.0 peripheral controller")
Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Signed-off-by: Felipe Balbi <felipe.balbi@linux.intel.com>
This controller disallows to change the PIPE until reading/writing
a packet finishes. However. the previous code is not enough to hold
the lock in some functions. So, this patch fixes it.
Fixes: 746bfe63bb ("usb: gadget: renesas_usb3: add support for Renesas USB3.0 peripheral controller")
Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Signed-off-by: Felipe Balbi <felipe.balbi@linux.intel.com>
This patch fixes an issue that this driver is possible to cause
deadlock by double-spinclocked in renesas_usb3_stop_controller().
So, this patch removes spinlock API calling in renesas_usb3_stop().
(In other words, the previous code had a redundant lock.)
Fixes: 746bfe63bb ("usb: gadget: renesas_usb3: add support for Renesas USB3.0 peripheral controller")
Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Signed-off-by: Felipe Balbi <felipe.balbi@linux.intel.com>
This patch fixes an issue that this driver is possible to access
the registers before pm_runtime_get_sync() if a gadget driver is
installed first. After that, oops happens on R-Car Gen3 environment.
To avoid it, this patch changes the pm_runtime call timing from
probe/remove to udc_start/udc_stop.
Fixes: 746bfe63bb ("usb: gadget: renesas_usb3: add support for Renesas USB3.0 peripheral controller")
Signed-off-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Signed-off-by: Felipe Balbi <felipe.balbi@linux.intel.com>
f_mass_storage has a memorry barrier issue with the sleep and wake
functions that can cause a deadlock. This results in intermittent hangs
during MSC file transfer. The host will reset the device after receiving
no response to resume the transfer. This issue is seen when dwc3 is
processing 2 transfer-in-progress events at the same time, invoking
completion handlers for CSW and CBW. Also this issue occurs depending on
the system timing and latency.
To increase the chance to hit this issue, you can force dwc3 driver to
wait and process those 2 events at once by adding a small delay (~100us)
in dwc3_check_event_buf() whenever the request is for CSW and read the
event count again. Avoid debugging with printk and ftrace as extra
delays and memory barrier will mask this issue.
Scenario which can lead to failure:
-----------------------------------
1) The main thread sleeps and waits for the next command in
get_next_command().
2) bulk_in_complete() wakes up main thread for CSW.
3) bulk_out_complete() tries to wake up the running main thread for CBW.
4) thread_wakeup_needed is not loaded with correct value in
sleep_thread().
5) Main thread goes to sleep again.
The pattern is shown below. Note the 2 critical variables.
* common->thread_wakeup_needed
* bh->state
CPU 0 (sleep_thread) CPU 1 (wakeup_thread)
============================== ===============================
bh->state = BH_STATE_FULL;
smp_wmb();
thread_wakeup_needed = 0; thread_wakeup_needed = 1;
smp_rmb();
if (bh->state != BH_STATE_FULL)
sleep again ...
As pointed out by Alan Stern, this is an R-pattern issue. The issue can
be seen when there are two wakeups in quick succession. The
thread_wakeup_needed can be overwritten in sleep_thread, and the read of
the bh->state maybe reordered before the write to thread_wakeup_needed.
This patch applies full memory barrier smp_mb() in both sleep_thread()
and wakeup_thread() to ensure the order which the thread_wakeup_needed
and bh->state are written and loaded.
However, a better solution in the future would be to use wait_queue
method that takes care of managing memory barrier between waker and
waiter.
Cc: <stable@vger.kernel.org>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Thinh Nguyen <thinhn@synopsys.com>
Signed-off-by: Felipe Balbi <felipe.balbi@linux.intel.com>
USB support in the Meson8 SoCs is provided by a DWC2 controller which
works with the same settings as Meson8b and GXBB. Using the generic
"snps,dwc2" binding results in an endless stream of "Overcurrent change
detected" messages.
Acked-by: Rob Herring <robh@kernel.org>
Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
Signed-off-by: Felipe Balbi <felipe.balbi@linux.intel.com>
This commit allows a gadget that does not support SuperSpeed to indicate
that it supports LPM. It does this by setting the 'lpm_capable' flag in
the gadget structure.
If a gadget sets this, the composite gadget framework will set the
bcdUSB to 0x0201 to indicate that this supports BOS descriptors, and
also return a USB 2.0 Extension descriptor as part of the BOS descriptor
set.
See USB 2.0 LPM ECN Section 3.
Signed-off-by: John Youn <johnyoun@synopsys.com>
Signed-off-by: Sevak Arakelyan <sevaka@synopsys.com>
Signed-off-by: Felipe Balbi <felipe.balbi@linux.intel.com>
There were individual waitqueues for each epfile but eps_enable
would iterate through all of them, resulting in essentially the
same wakeup time.
The waitqueue represents the function being enabled, so a central
waitqueue in ffs_data makes more sense and is less redundant.
Also use wake_up_interruptible to reflect use of
wait_event_interruptible.
Acked-by: Michal Nazarewicz <mina86@mina86.com>
Signed-off-by: Jerry Zhang <zhangjerry@google.com>
Signed-off-by: Felipe Balbi <felipe.balbi@linux.intel.com>
This allows users to make an ioctl call as the first action on a
connection. Ex, some functions might want to get endpoint size
before making any i/os.
Previously, calling ioctls before read/write would depending on the
timing of endpoints being enabled.
ESHUTDOWN is now a possible return value and ENODEV is not, so change
docs accordingly.
Acked-by: Michal Nazarewicz <mina86@mina86.com>
Signed-off-by: Jerry Zhang <zhangjerry@google.com>
Signed-off-by: Felipe Balbi <felipe.balbi@linux.intel.com>
Instead, we can require caller to pass a buffer for the function to
use. This cleans things quite a bit.
Signed-off-by: Felipe Balbi <felipe.balbi@linux.intel.com>
Instead of going for a 512 byte buffer and using snprintf(), let's
rely on helps __string() and __assign_str() where possible.
Signed-off-by: Felipe Balbi <felipe.balbi@linux.intel.com>
Instead of printing out enqueue and dequeue pointer value as a header
to the output, let's mark the TRBs in question with 'E' and 'D'. The
output looks slightly easier to read.
Signed-off-by: Felipe Balbi <felipe.balbi@linux.intel.com>
%p will leak kernel pointers, so let's not expose the information on
dmesg and instead use %pK. %pK will only show the actual addresses if
explicitly enabled under /proc/sys/kernel/kptr_restrict.
Cc: <stable@vger.kernel.org>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Felipe Balbi <felipe.balbi@linux.intel.com>
Some local constants don't change from call to call and are good
candidates to become static. This will prevent copying of these
constants to stack during runtime.
Signed-off-by: Maksim Salau <maksim.salau@gmail.com>
Signed-off-by: Johan Hovold <johan@kernel.org>
Currently VBUS is turned off while a usb device is detached, and turned
on again by the polling routine. This short period VBUS loss prevents
usb modem to switch mode.
VBUS should be constantly on for host-only mode, so this changes the
driver to not turn off VBUS for host-only mode.
Fixes: 2f3fd2c5bd ("usb: musb: Prepare dsps glue layer for PM runtime support")
Cc: stable@vger.kernel.org #v4.11
Reported-by: Moreno Bartalucci <moreno.bartalucci@tecnorama.it>
Acked-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Bin Liu <b-liu@ti.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The combo of list_empty() and list_first_entry() can be replaced with
list_first_entry_or_null().
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The Aspeed 2400/2500 families have a variant of UHCI which requires
some quirks to the driver to work:
- The register offsets are different. We add a remapping helper.
- All accesses have to be done via 32-bit loads and stores. We
force all accessors to use readl/writel. This is of no consequence
for reads as we never read "in the middle" of a register. For writes
it also works fine as the registers only actually implement the bits
we try to write (16-bit for the registers accessed with writew and
8-bit for the register accessed with writeb), so always using a
32-bit write will have no negative effect. We never do partial writes.
- The resume detect interrupt is broken
- The number of ports is (optionally) provided via the device-tree
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
--
v2. Remove the bulk of the #ifdef's
drivers/usb/host/Kconfig | 6 ++++-
drivers/usb/host/uhci-hcd.c | 17 +++++++++++---
drivers/usb/host/uhci-hcd.h | 51 ++++++++++++++++++++++++++++++++++++++++
drivers/usb/host/uhci-platform.c | 22 ++++++++++++++++-
4 files changed, 91 insertions(+), 5 deletions(-)
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
ci_role BUGs when the role is >= CI_ROLE_END.
This is the case while the role is changing.
Signed-off-by: Michael Thalmeier <michael.thalmeier@hale.at>
Signed-off-by: Peter Chen <peter.chen@nxp.com>
Johan writes:
USB-serial fixes for v4.12-rc2
Here's a fix for a long-standing issue in the ftdi_sio driver that
prevented unprivileged users from updating the low-latency flag,
something which became apparent after a recent change that restored the
older setting of not using low-latency mode by default.
A run of sparse revealed a couple of endianness issues that are now
fixed, and addressed is also a user-triggerable division-by-zero in
io_ti when debugging is enabled.
Finally there are some new device ids, including a simplification of how
we deal with a couple of older Olimex JTAG adapters.
All have been in linux-next with no reported issues.
Signed-off-by: Johan Hovold <johan@kernel.org>
In the ene_usb6250 sub-driver for usb-storage, the INQUIRY data
returned by the driver indicates that the device has removable media.
While this is technically correct (memory cards can be removed from
the reader), it is not useful because the device automatically
disconnects itself from the USB bus when no media is present.
In addition, the driver does not support the PREVENT-ALLOW MEDIUM
REMOVAL and START STOP UNIT commands, and this can cause
user-interface frameworks to get confused when the user asks for the
card to be removed or ejected.
This patch fixes the problem by changing the INQUIRY data to specify
non-removable media; in practice this value works much better.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Reported-and-tested-by: Andreas Hartmann <andihartmann@01019freenet.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In the ene_usb6250 sub-driver for usb-storage, the sd_scsi_inquiry()
and ms_scsi_inquiry() subroutines (one meant for use with SD memory
cards and the other for use with MS memory cards) are exact
duplicates. This patch removes the duplication by creating a single
do_scsi_inquiry() command and using it instead of the other two.
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Tested-by: Andreas Hartmann <andihartmann@01019freenet.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
