When the enable bits for controller / async schedule / periodic schedule
change just make sure we kick the frame timer and let
ehci_advance_periodic_state and ehci_advance_async_state handle the
controller state changes.

This will make ehci set USBSTS_HALT when the controller shutdown is
actually done, once both schedules are in inactive state and the
USBSTS_PSS and USBSTS_ASS bits are clear.
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

Update the status register in the ehci_set_state function, to make sure
the guest-visible register is in sync with our internal schedule state.
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

Add helper functions to query whenever the async / periodic schedule
is enabled or not.  Put them into use too.
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

Check for the reset bit first when processing USBCMD register writes.
Also break out of the switch, there is no need to check the other bits.
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

Kick async schedule when we get a wakeup
notification from a usb device.
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

When a packet completes which happens to be part of the async schedule
kick the async bottom half for processing,
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

This way we can kick the async schedule independant from the
periodic frame timer.
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

Keep track whenever a EHCIQueue is part of the async or periodic
schedule.  This way we don't have to pass around the async flag
everywhere but can look it up from the EHCIQueue struct when needed.
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

Little tweak for the queue initialization, set the QH address in the
allocation function.
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

Add packet queuing.  Follow the qTD chain to see if there are more
packets we can submit.  Improves performance on larger transfers,
especially with usb-host, as we don't have to wait for a packet to
finish before sending the next one to the host for processing.
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

Move ehci_flush_qh() function up in the source code.
No code change.
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

Keep a USBDevice pointer in EHCIQueue so we don't have to lookup the
device on each usb packet submission.
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

This way it is possible to use ehci_execute to submit others than the
first EHCIPacket of the EHCIQueue.
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

Add a separate EHCIPacket struct and move fields over from EHCIQueue.
Preparing for supporting multiple packets per queue being in flight at
the same time.  No functional changes yet.

Fix some codestyle issues along the way.
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

Properly register reset function via the device class.
Signed-off-by: NJan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

With all scsi migration support bits in place the
final step is pretty simple ;)
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

Repace the running buffer pointer (scsi_buf) with a buffer offset
field (scsi_off).  The later is alot easier to live-migrate.
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

Factor out packet completion to a separate function which
cares to get the MSDState->packet update right.
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

We have the field twice, once in MSDState directly and one in the status
word struct.  Drop one.
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

The multifunction ich9 ehci controller with uhci companions uses a
different interrupt pin for each function.  The three uhci devices
get pins A, B and C, whereas ehci uses pin D.  This way the guest
can assign different IRQ lines to each controller.
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

Cancel transactions before saving vmstate is pretty pointless and just
causes disruptions.  We need to cancel them before *loading* vmstate,
but in that case uhci_reset() handles it already and no special action
is needed.
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

Add a property for the uhci bandwidth.  Can be used to make uhci
emulation run faster than real hardware.
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

Schedule bottom half on completion of async packets instead of calling
uhci_process_frame directly.  This way we run uhci_process_frame only
once in case multiple packets finish in a row.  Also check whenever
there is bandwidth left before scheduling uhci_process_frame.
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

uhci_process_frame() can be invoked multiple times per frame, so
accounting usb bandwith in a local variable doesn't fly, use a variable
in UHCIState instead.  Also check the limit more frequently.
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

Callers are changed to use qerror_report_err() to keep their QError
semantics.
Signed-off-by: NLuiz Capitulino <lcapitulino@redhat.com>
Reviewed-By: NLaszlo Ersek <lersek@redhat.com>

The only backend that really uses it is the socket one, which calls
monitor_get_fd(). But it can use 'cur_mon' instead.
Signed-off-by: NLuiz Capitulino <lcapitulino@redhat.com>
Reviewed-By: NLaszlo Ersek <lersek@redhat.com>

This commit converts qemu_opts_create() from qerror_report() to
error_set().

Currently, most calls to qemu_opts_create() can't fail, so most
callers don't need any changes.

The two cases where code checks for qemu_opts_create() erros are:

 1. Initialization code in vl.c. All of them print their own
    error messages directly to stderr, no need to pass the Error
    object

 2. The functions opts_parse(), qemu_opts_from_qdict() and
    qemu_chr_parse_compat() make use of the error information and
    they can be called from HMP or QMP. In this case, to allow for
    incremental conversion, we propagate the error up using
    qerror_report_err(), which keeps the QError semantics
Signed-off-by: NLuiz Capitulino <lcapitulino@redhat.com>
Reviewed-By: NLaszlo Ersek <lersek@redhat.com>

Most important here is to update our internal endpoint state so we know
the endpoint isn't in halted state any more.  Without this usb-host
tries to clear halt again with the next data transfer submitted.  Doing
this twice is (a) not correct and (b) confuses some usb devices,
rendering them non-functional in the guest.
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

These were identified using: http://github.com/lyda/misspell-check
and run like this to create a bourne shell script using GNU sed's
-i option:

git ls-files|grep -vF .bin | misspellings -f - |grep -v '^ERROR:' |perl \
-pe 's/^(.*?)\[(\d+)\]: (\w+) -> "(.*?)"$/sed -i '\''${2}s!$3!$4!'\'' $1/'

Manually eliding the FP, "rela->real" and resolving "addres" to
address (not "adders") we get this:

  sed -i '450s!thru!through!' Changelog
  sed -i '260s!neccessary!necessary!' coroutine-sigaltstack.c
  sed -i '54s!miniscule!minuscule!' disas.c
  sed -i '1094s!thru!through!' hw/usb/hcd-ehci.c
  sed -i '1095s!thru!through!' hw/usb/hcd-ehci.c
  sed -i '21s!unecessary!unnecessary!' qapi-schema-guest.json
  sed -i '307s!explictly!explicitly!' qemu-ga.c
  sed -i '490s!preceeding!preceding!' qga/commands-posix.c
  sed -i '792s!addres!address!' qga/commands-posix.c
  sed -i '6s!beeing!being!' tests/tcg/test-mmap.c

Also, manually fix "arithmentic", spotted by Peter Maydell:

  sed -i 's!arithmentic!arithmetic!' coroutine-sigaltstack.c
Signed-off-by: NJim Meyering <meyering@redhat.com>
Reviewed-by: NPeter Maydell <peter.maydell@linaro.org>

uhci_reset() clears irq mask and irq status registers, but doesn't
update the irq line.  Which may result in suspious IRQs after uhci
reset.  Fix it.
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

This patch adds a function which creates unique serial numbers for usb
devices and puts it into use.  Windows guests tend to become unhappy if
they find two identical usb devices in the system.  Effects range from
non-functional devices (with yellow exclamation mark in device manager)
to BSODs.  Handing out unique serial numbers to devices fixes this.

With this patch applied almost all emulated devices get a generated,
unique serial number.  There are two exceptions:

 * usb-storage devices will prefer a user-specified serial number
   and will only get a generated number in case the serial property
   is unset.
 * usb-hid devices keep the fixed serial number "42" as it is used
   to signal "remote wakeup actually works".
   See commit 7b074a22Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

We clear our pending async urb list on device disconnect and we may still
receive "packet complete" packets from our peer after this, which will then
refer to packet ids no longer in our list.
Signed-off-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>

Without this disconnected devices look like the last redirected device
in the monitor in "info usb".
Signed-off-by: NHans de Goede <hdegoede@redhat.com>
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>