Age | Commit message (Collapse) | Author | Files | Lines |
|
Fix error queue name, rename input queue for clarity. Split input
and output queue for further work and NAK when there is no interrupt
data available.
|
|
Error queue is implemented, possible receiver and device properties
are filled in (in the header).
Signed-off-by: Peter Wu <peter@lekensteyn.nl>
|
|
Prepare for adding DJ support, which I expect to be huge and unrelated
to the HID code.
Signed-off-by: Peter Wu <peter@lekensteyn.nl>
|
|
bNumEndpoints specify the number of endpoints in the endpoint
descriptor. When the mouse endpoint was tried (EP2), it triggered this
assertion because there is only one endpoint for the interface.
Signed-off-by: Peter Wu <peter@lekensteyn.nl>
|
|
Copied from dev-hid.c (order changed to match interfaces order).
Signed-off-by: Peter Wu <peter@lekensteyn.nl>
|
|
For now, the receiver has always a mouse and keyboard paired.
Set_Idle, Get_Idle, Get_Protocol and Set_Protocol support
interface-specific devices.
Interrupt data for the HID mode are also implemented.
Signed-off-by: Peter Wu <peter@lekensteyn.nl>
|
|
Instead of changing bus.c (next_usb_bus) and possibly breaking
configurations relying on this number, just add one in the dump code
for cosmetic purposes.
Signed-off-by: Peter Wu <peter@lekensteyn.nl>
|
|
There is still a discrepancy (seen 300 ms and 600 ms), but it is better
than completely a difference equal to the uptime.
Signed-off-by: Peter Wu <peter@lekensteyn.nl>
|
|
This is verified to reproduce the same capture as done inside the VM.
The differences are in time, bInterval, URB length, URB ID and the
Copy of transfer flags.
Signed-off-by: Peter Wu <peter@lekensteyn.nl>
|
|
And I was wondering why the interrupt packets in dev-hid.c had a
reversed type marking...
|
|
Non-control transfers have a meaningless setup buffer.
Signed-off-by: Peter Wu <peter@lekensteyn.nl>
|
|
Signed-off-by: Peter Wu <peter@lekensteyn.nl>
|
|
The "length" field stores the expected size of data. "len_cap" stores
the real length of the data in the current packet (the "captured" data).
Also try to set a slightly more sane length value depending on
submission type, but this is much braindead.
Signed-off-by: Peter Wu <peter@lekensteyn.nl>
|
|
Signed-off-by: Peter Wu <peter@lekensteyn.nl>
|
|
Signed-off-by: Peter Wu <peter@lekensteyn.nl>
|
|
This allows for less less checking by the consumer and avoids
unnecessary work if the capture cannot be written anyway.
Signed-off-by: Peter Wu <peter@lekensteyn.nl>
|
|
The timestamp calculation in dump_receive is very unobvious now and
looks completely arbitrary. qemu_clock_get_us already returns micro-
seconds, so use that instead of manual calculating with nanoseconds.
While at it, express the private start_ts field in microseconds too
instead of seconds. This is supposed to yield a more accurate initial
timestamp (based on host clock). NOTE: this patch ignores the RTC of
the guest! Perhaps this part should be reverted.
|
|
To make a tidy ship, remove unused parts that were taken from dev-wacom.
Also enable a handle_data callback (currently empty), handle_reset
(currently sets the device mode to HID instead of DJ).
This is incomplete, the handle_data callback must contain something.
Signed-off-by: Peter Wu <peter@lekensteyn.nl>
|
|
It looks very ugly, perhaps a helper function should moved to desc.c?
Signed-off-by: Peter Wu <peter@lekensteyn.nl>
|
|
Signed-off-by: Peter Wu <peter@lekensteyn.nl>
|
|
USB packets are captured in the pcap format using the Linux usbmon
link-layer type. Right now, this only works with handle_ctrl. Data
packets (isochronous, bulk and interrupt) are not (fully)
implemented yet.
Signed-off-by: Peter Wu <peter@lekensteyn.nl>
|
|
Based on hw/usb/dev-wacom.c
|
|
* remotes/rth/tcg-aarch-6-2:
tcg-aarch64: Introduce tcg_out_insn_3405
tcg-aarch64: Support div, rem
tcg-aarch64: Support muluh, mulsh
tcg-aarch64: Support add2, sub2
tcg-aarch64: Support deposit
tcg-aarch64: Use tcg_out_insn for setcond
tcg-aarch64: Support movcond
tcg-aarch64: Support andc, orc, eqv, not, neg
tcg-aarch64: Handle constant operands to and, or, xor
tcg-aarch64: Handle constant operands to add, sub, and compare
tcg-aarch64: Implement mov with tcg_out_insn
tcg-aarch64: Introduce tcg_out_insn_3401
tcg-aarch64: Convert shift insns to tcg_out_insn
tcg-aarch64: Introduce tcg_out_insn
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
|
|
staging
Block pull request
# gpg: Signature made Fri 14 Mar 2014 16:12:14 GMT using RSA key ID 81AB73C8
# gpg: Good signature from "Stefan Hajnoczi <stefanha@redhat.com>"
# gpg: aka "Stefan Hajnoczi <stefanha@gmail.com>"
# gpg: WARNING: This key is not certified with a trusted signature!
# gpg: There is no indication that the signature belongs to the owner.
# Primary key fingerprint: 8695 A8BF D3F9 7CDA AC35 775A 9CA4 ABB3 81AB 73C8
* remotes/stefanha/tags/block-pull-request:
qemu-iotests: remove 085 and 087 from 'quick' group
qemu-iotests: add 083 NBD client disconnect tests
tests: add nbd-fault-injector.py utility
nbd: close socket if connection breaks
block: Explicitly specify 'unsigned long long' for VHDX 64-bit constants
blockdev: Refuse to open encrypted image unless paused
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
|
|
* remotes/bonzini/scsi-next:
virtio-scsi: actually honor sense_size from configuration space
scsi: Fix migration of scsi sense data
spapr-vscsi: fix CRQ status
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
|
|
Cleaning up the implementation of tcg_out_movi at the same time.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Reviewed-by: Claudio Fontana <claudio.fontana@huawei.com>
Tested-by: Claudio Fontana <claudio.fontana@huawei.com>
|
|
Clean up multiply at the same time.
For remainder, generic code will produce mul+sub,
whereas we can implement with msub.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Reviewed-by: Claudio Fontana <claudio.fontana@huawei.com>
Tested-by: Claudio Fontana <claudio.fontana@huawei.com>
|
|
Signed-off-by: Richard Henderson <rth@twiddle.net>
Reviewed-by: Claudio Fontana <claudio.fontana@huawei.com>
Tested-by: Claudio Fontana <claudio.fontana@huawei.com>
|
|
Signed-off-by: Richard Henderson <rth@twiddle.net>
Reviewed-by: Claudio Fontana <claudio.fontana@huawei.com>
Tested-by: Claudio Fontana <claudio.fontana@huawei.com>
|
|
Also tidy the implementation of ubfm, sbfm, extr in order to share code.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Reviewed-by: Claudio Fontana <claudio.fontana@huawei.com>
Tested-by: Claudio Fontana <claudio.fontana@huawei.com>
|
|
Signed-off-by: Richard Henderson <rth@twiddle.net>
Reviewed-by: Claudio Fontana <claudio.fontana@huawei.com>
Tested-by: Claudio Fontana <claudio.fontana@huawei.com>
|
|
Signed-off-by: Richard Henderson <rth@twiddle.net>
Reviewed-by: Claudio Fontana <claudio.fontana@huawei.com>
Tested-by: Claudio Fontana <claudio.fontana@huawei.com>
|
|
Signed-off-by: Richard Henderson <rth@twiddle.net>
Reviewed-by: Claudio Fontana <claudio.fontana@huawei.com>
Tested-by: Claudio Fontana <claudio.fontana@huawei.com>
|
|
Handle a simplified set of logical immediates for the moment.
The way gcc and binutils do it, with 52k worth of tables, and
a binary search depth of log2(5334) = 13, seems slow for the
most common cases.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Reviewed-by: Claudio Fontana <claudio.fontana@huawei.com>
Tested-by: Claudio Fontana <claudio.fontana@huawei.com>
|
|
Signed-off-by: Richard Henderson <rth@twiddle.net>
Reviewed-by: Claudio Fontana <claudio.fontana@huawei.com>
Tested-by: Claudio Fontana <claudio.fontana@huawei.com>
|
|
Avoid the magic numbers in the current implementation.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Reviewed-by: Claudio Fontana <claudio.fontana@huawei.com>
Tested-by: Claudio Fontana <claudio.fontana@huawei.com>
|
|
This merges the implementation of tcg_out_addi and tcg_out_subi.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Reviewed-by: Claudio Fontana <claudio.fontana@huawei.com>
Tested-by: Claudio Fontana <claudio.fontana@huawei.com>
|
|
Signed-off-by: Richard Henderson <rth@twiddle.net>
Reviewed-by: Claudio Fontana <claudio.fontana@huawei.com>
Tested-by: Claudio Fontana <claudio.fontana@huawei.com>
|
|
Converting the add/sub (3.5.2) and logical shifted (3.5.10) instruction
groups to the new scheme.
Signed-off-by: Richard Henderson <rth@twiddle.net>
Reviewed-by: Claudio Fontana <claudio.fontana@huawei.com>
Tested-by: Claudio Fontana <claudio.fontana@huawei.com>
|
|
The 'quick' group in qemu-iotests are not allowed to run QEMU since we
don't know which targets are available. In other words, they may only
use qemu-img, qemu-io, and qemu-nbd.
Drop 085 and 087 from the 'quick' group since they run QEMU. This
makes "make check-block" pass again.
Reported-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
|
|
This new test case uses nbd-fault-injector.py to simulate broken TCP
connections at each stage in the NBD protocol. This way we can exercise
block/nbd-client.c's socket error handling code paths.
In particular, this serves as a regression test to make sure
nbd-client.c doesn't cause an infinite loop by leaving its
nbd_receive_reply() fd handler registered after the connection has been
closed. This bug was fixed in an earlier patch.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
|
|
The nbd-fault-injector.py script is a special kind of NBD server. It
throws away all writes and produces zeroes for reads. Given a list of
fault injection rules, it can simulate NBD protocol errors and is useful
for testing NBD client error handling code paths.
See the patch for documentation. This scripts is modelled after Kevin
Wolf <kwolf@redhat.com>'s blkdebug block driver.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
|
|
nbd_receive_reply() is called by the event loop whenever data is
available or the socket has been closed by the remote side.
This patch closes the socket when an error occurs to prevent the
nbd_receive_reply() handler from being called indefinitely after the
connection has failed.
Note that we were already correctly returning EIO for pending requests
but leaving the nbd_receive_reply() handler registered resulted in high
CPU consumption and a flood of error messages.
Reuse nbd_teardown_connection() to close the socket.
Reported-by: Zhifeng Cai <bluewindow@h3c.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
|
|
On 32-bit hosts, some compilers will warn on too large integer constants
for constants that are 64-bit in length. Explicitly put a 'ULL' suffix
on those defines.
Reported-by: Alexander Graf <agraf@suse.de>
Signed-off-by: Jeff Cody <jcody@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
|
|
Opening an encrypted image takes an additional step: setting the key.
Between open and the key set, the image must not be used.
We have some protection against accidental use in place: you can't
unpause a guest while we're missing keys. You can, however, hot-plug
block devices lacking keys into a running guest just fine, or insert
media lacking keys. In the latter case, notifying the guest of the
insert is delayed until the key is set, which may suffice to protect
at least some guests in common usage.
This patch makes the protection apply in more cases, in a rather
heavy-handed way: it doesn't let you open encrypted images unless
we're in a paused state.
It doesn't extend the protection to users other than the guest (block
jobs?). Use of runstate_check() from block.c is disgusting. Best I
can do right now.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
|
|
We were always truncating the sense size to 96 bytes.
Reviewed-by: Fam Zheng <famz@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
|
|
c5f52875 changed the size of sense array in vmstate_scsi_device by
mistake. This patch restores the old size, and add a subsection for the
remaining part of the buffer size. So that migration is not broken.
Signed-off-by: Fam Zheng <famz@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
|
|
Normally VIOSRP_OK (0) means success and non-zero value means error
except VIOSRP_OK2 (0x99) which is another success code by weird accident.
This uses 0 as success code always as some guests do not cope with
the 0x99 value well. The existing linux driver checks for both VIOSRP_OK
and VIOSRP_OK2 since 2.6.32.
This returns non-zero code (VIOSRP_ADAPTER_FAIL == 0x10) on errors which
can only happen if DMA write failed.
Suggested-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
|
|
Signed-off-by: Anthony Liguori <aliguori@amazon.com>
|
|
into staging
QOM/QTest infrastructure fixes
* QOM cast fix for virtserialport and regression test
* QTest error handling fix
* QTest output cleanup
# gpg: Signature made Thu 13 Mar 2014 20:43:34 GMT using RSA key ID 3E7E013F
# gpg: Good signature from "Andreas Färber <afaerber@suse.de>"
# gpg: aka "Andreas Färber <afaerber@suse.com>"
* remotes/afaerber/tags/qom-devices-for-2.0:
main-loop: Suppress "I/O thread spun" warnings for qtest
qtest: Fix crash if SIGABRT during qtest_init()
virtio-console-test: Test virtserialport as well
virtio-console: Fix VIRTIO_CONSOLE() cast macro
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
|