So far, the code was always assuming SuperSpeed for all devices,
which seemed to work OK with QEMU ... but let's better play safe
instead and use the speed from the port status instead.
Signed-off-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
To support USB3 devices attached to a hub, we've got to set
the so-called "route string" in the slot context, which is
a concatenation of the hub port numbers that need to be
traversed between the root hub port and the destination device.
Signed-off-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Nikunj A Dadhania <nikunj@linux.vnet.ibm.com>
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
USB3 devices need to be initialized with usb3_dev_init(), so that
they get assigned a proper slot ID. And for USB3 devices that are
attached to a (non-root) hub, we also need to keep track of the
hub topology, so a new field called "hub" is added to the struct
usb_dev which references the hub devices where the current USB
devices is attached to. The hub topology will be used later to
build the so-called "route string" for the USB3 devices.
Signed-off-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Nikunj A Dadhania <nikunj@linux.vnet.ibm.com>
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
SLOF should be able to support more than just four devices on
the four root hub ports. So we've got to increase the amount of
possible slot IDs. Since QEMU supports up to 44 devices (four
devices at the root ports + 5 tiers of hubs with 8 ports each),
this seems to be a good new value instead. And to make sure that
we do not accidentially get a buffer overflow when accessing the
xhcd->xdevs[slot_id] array, this patch also add a sanity check at
the beginning of the xhci_alloc_dev() function here.
Signed-off-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Nikunj A Dadhania <nikunj@linux.vnet.ibm.com>
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Arrays should not be declared in a header (to avoid that it gets
declared multiple times when the header is included more than once),
so the ps_array_usbX settings should reside in the .c file instead.
Signed-off-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Nikunj A Dadhania <nikunj@linux.vnet.ibm.com>
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Current handling of the keyboard polling was very slow and
keys were getting dropped. Done following for fixing this:
* Use multiple buffers per TRB
* Allocate buffers in xhci according to the number of TRBS.
This reduces the delay of key? processing by getting rid of wait in
the polling routine.
Reported-by: Dinar Valeev <k0da@opensuse.org>
Signed-off-by: Nikunj A Dadhania <nikunj@linux.vnet.ibm.com>
Tested-by: Dinar Valeev <k0da@opensuse.org>
Tested-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
A memory barrier was missing after updating the trb details.
Signed-off-by: Nikunj A Dadhania <nikunj@linux.vnet.ibm.com>
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
When scanning hubs, the code in libusb can be called recursively, for
example usb_hub_init() calls setup_new_device(), which then calls
slof_usb_handle() to execute Forth code for the next device.
If that next device is a hub, we end up recursively in usb_hub_init()
again.
Since stack space is limited in SLOF, we can optimize here a little
bit by splitting up the setup_new_device() function into the part
that retrieves the descriptors (which takes most of the stack space
in this code path since the descriptors are placed on the stack),
and the part that populates the the device tree node of the new
device (which is responsible for the recursion).
Signed-off-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Add support for xhci interrupt pipes needed for keyboard handling
Signed-off-by: Nikunj A Dadhania <nikunj@linux.vnet.ibm.com>
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
The keyboard events being async, need to prepare the link when last but
one trb is queued.
Signed-off-by: Nikunj A Dadhania <nikunj@linux.vnet.ibm.com>
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Current code scanned only the super speed ports. Add support for
scanning high speed ports as well.
Also re-org code to reduce duplication of code.
Signed-off-by: Nikunj A Dadhania <nikunj@linux.vnet.ibm.com>
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Since the controller was being used by single user, event poll loop
works without verifying what event returned. Now with usb-keyboard
addition where we will get keyboard events, this would not work.
Transfer bulk would only look for response from the trb queued and
ignore rest of the events. Moreover, while bulk transfer is going on,
there are no keyboard events in booting use case.
Signed-off-by: Nikunj A Dadhania <nikunj@linux.vnet.ibm.com>
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
QEMU implementation of XHCI doesn't implement halt properly. There might
be ongoing activities and active DMAs, introduce 50ms delay during
shutdown path.
Signed-off-by: Nikunj A Dadhania <nikunj@linux.vnet.ibm.com>
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
commit 706c69e4 "xhci: fix port assignment" partially fixed the
usb port numbering.
Adding parsing of xhci extended capabilities, we can parse the
overlapped ports accordingly and have proper port numbering.
Signed-off-by: Nikunj A Dadhania <nikunj@linux.vnet.ibm.com>
A bit operation bug left the controller in the running state causing
PCI EEH in the host when using pci-passthru for USB3 device.
Signed-off-by: Nikunj A Dadhania <nikunj@linux.vnet.ibm.com>
While doing cleanup of the allocated memory, make sure addresses being
unmapped/free were really allocated. During error conditions, some
address would not have been.
Signed-off-by: Nikunj A Dadhania <nikunj@linux.vnet.ibm.com>
* Fix freeing of memory in error path
* Boolean and int were used interchangebly at various points,
consolidate it to use boolean
Signed-off-by: Nikunj A Dadhania <nikunj@linux.vnet.ibm.com>
Thomas Huth