The virtio device lifecycle can be observed by looking at the sequence
of set status operations. This is especially important for catching the
reset operation (status value 0), which resets the device and all
virtqueues.
Signed-off-by: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Now that format strings can end in a PRI*64 macro, remove the
workarounds from the trace-events file.
Signed-off-by: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
There is no need to put a newline in trace event format strings. The
backend may use the format string within some context and takes care of
how to display the event. The stderr backend automatically appends "\n"
whereas the ust backend does not want a newline at all.
Signed-off-by: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
This patch makes qemu assign a port when creating the device, not when
attaching it. For most usb devices this isn't a noticable difference
because they are in attached state all the time.
The change affects usb-host devices which live in detached state while
the real device is unplugged from the host. They have a fixed port
assigned all the time now instead of getting grabbing one on attach and
releasing it at detach, i.e. they stop floating around at the usb bus.
The change also allows to simplify usb-hub. It doesn't need the
handle_attach() callback any more to configure the downstream ports.
This can be done at device initialitation time now. The changed
initialization order (first grab upstream port, then register downstream
ports) also fixes some icky corner cases. For example it is not possible
any more to plug the hub into one of its own downstream ports.
The usb host adapters must care too. USBPort->dev being non-NULL
doesn't imply any more the device is in attached state. The host
adapters must additionally check the USBPort->dev->attached flag.
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
This patch adds code to do minimal siTD handling, which is basically
just following the next pointer. This is good enougth to handle the
inactive siTDs used by FreeBSD. Active siTDs are skipped too as we
don't have split transfer support in qemu, additionally a warning is
printed.
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
When configured to pass through a specific host port (using hostbus and
hostport properties), try to claim the port if supported by the kernel.
That will avoid any kernel drivers binding to devices plugged into that
port. It will not stop any userspace apps (such as usb_modeswitch)
access the device via usbfs though.
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Add a bunch of trace points to usb-linux.c Drop a bunch of DPRINTK's in
favor of the trace points. Also cleanup error reporting a bit while being
at it.
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Given that all events with programmatically-controlled state are disabled by
default, we can delete the "disable" property from all events.
Signed-off-by: Lluís Vilanova <vilanova@ac.upc.edu>
Note that this refers to the backend-specific state (whether the output must be
generated), not the event "disabled" property (which always uses the "nop"
backend).
Signed-off-by: Lluís Vilanova <vilanova@ac.upc.edu>
Polarity of external interrupts needs to be handled in the IOAPIC.
Passing it to the APIC is pointless. So remove all these arguments.
Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Can be useful when debugging the device scan phase.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
With this patch, sense data is stored in the generic data structures
for SCSI devices and requests. The SCSI layer takes care of storing
sense data in the SCSIDevice for the subsequent REQUEST SENSE command.
At the same time, get_sense is removed and scsi_req_get_sense can use
an entirely generic implementation.
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
In order to be able to call bdrv_co_readv/writev for drivers that don't
implement the functions natively, add an emulation that uses the AIO functions
to implement them.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Add new block driver callbacks bdrv_co_readv/writev, which work on a
QEMUIOVector like bdrv_aio_*, but don't need a callback. The function may only
be called inside a coroutine, so a block driver implementing this interface can
yield instead of blocking during I/O.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Asynchronous code is becoming very complex. At the same time
synchronous code is growing because it is convenient to write.
Sometimes duplicate code paths are even added, one synchronous and the
other asynchronous. This patch introduces coroutines which allow code
that looks synchronous but is asynchronous under the covers.
A coroutine has its own stack and is therefore able to preserve state
across blocking operations, which traditionally require callback
functions and manual marshalling of parameters.
Creating and starting a coroutine is easy:
coroutine = qemu_coroutine_create(my_coroutine);
qemu_coroutine_enter(coroutine, my_data);
The coroutine then executes until it returns or yields:
void coroutine_fn my_coroutine(void *opaque) {
MyData *my_data = opaque;
/* do some work */
qemu_coroutine_yield();
/* do some more work */
}
Yielding switches control back to the caller of qemu_coroutine_enter().
This is typically used to switch back to the main thread's event loop
after issuing an asynchronous I/O request. The request callback will
then invoke qemu_coroutine_enter() once more to switch back to the
coroutine.
Note that if coroutines are used only from threads which hold the global
mutex they will never execute concurrently. This makes programming with
coroutines easier than with threads. Race conditions cannot occur since
only one coroutine may be active at any time. Other coroutines can only
run across yield.
This coroutines implementation is based on the gtk-vnc implementation
written by Anthony Liguori <anthony@codemonkey.ws> but it has been
significantly rewritten by Kevin Wolf <kwolf@redhat.com> to use
setjmp()/longjmp() instead of the more expensive swapcontext() and by
Paolo Bonzini <pbonzini@redhat.com> for Windows Fibers support.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
"next" is reserved in systemtap thus using this as a
trace parameter name causes trouble when trying to trace
with systemtap.
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
The map cache is a Xen thing, so its API should make this clear.
Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Alexander Graf <agraf@suse.de>
Introduce a new emulated PCI device, specific to fully virtualized Xen
guests. The device is necessary for PV on HVM drivers to work.
Signed-off-by: Steven Smith <ssmith@xensource.com>
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Signed-off-by: Alexander Graf <agraf@suse.de>
This patch introduces phys memory client for Xen.
Only sync dirty_bitmap and set_memory are actually implemented.
migration_log will stay empty for the moment.
Xen can only log one range for bit change, so only the range in the
first call will be synced.
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
Signed-off-by: Alexander Graf <agraf@suse.de>
This patch fixes a bunch of issues in the itd descriptor handling.
Most important fix is to handle transfers which cross page borders
correctly by looking up the address of the next page. Luckily the
linux uses physically contigous memory so the data used to hits the
correct location even with this bug instead of corrupting guest
memory. Also the transfer length updates for outgoing transfers wasn't
correct.
While being at it DPRINTFs have been replaced by tracepoints.
The isoch_pause logic has been disabled. Not clear to me which propose
this serves and I think it is incorrect too as we just skip processing
itds. Even when no xfer happens we have to clear the active bit.
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
This patch adds support for keeping multiple queues going at the same
time. One slow device will not affect other devices any more.
The patch adds code to manage EHCIQueue structs. It also does a number
of changes to the state machine:
* The state machine will never ever stop in EXECUTING any more.
Instead it will continue with the next queue (aka HORIZONTALQH) when
the usb device returns USB_RET_ASYNC.
* The state machine will stop processing when it figures it walks in
circles (easy to figure now that we have a EHCIQueue struct for each
QH we've processed). The bailout logic should not be needed any
more. For now it is still in, but will assert() in case it triggers.
* The state machine will just skip queues with a async USBPacket in
flight.
* The state machine will resume processing as soon as the async
USBPacket is finished.
The patch also takes care to flush the QH struct back to guest memory
when needed, so we don't get stale data when (re-)loading it from guest
memory in FETCHQH state.
It also makes the writeback code to not touch the first three dwords of
the QH struct as the EHCI must not write them. This actually fixes a
bug where QH chaining changes (next ptr) by the linux ehci driver where
overwritten by the emulated EHCI.
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Add a separate tracepoint to log how register values change in response
to a mmio write. Especially useful for registers which have read-only
or clear-on-write bits in them.
No change in behavior.
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Trace usb port operations (attach, detach, reset),
drop a few obsolete DPRINTF's.
No change in behavior.
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Add functions to get and set the current state of the state machine,
add tracepoints there to trace state transitions. Add support for
traceing the queue heads and transfer descriptors as we look at them.
Drop a few DPRINTFs and all DPRINTF_ST lines, they are obsolete now.
No change in behavior.
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
This patch starts adding trace support to ehci. It traces
updates of the status register (USBSTS), mmio access and
controller reset.
It also adds functions to set and clear status register bits
and puts them in use everywhere.
Some DPRINTF's are dropped in favor of the new tracepoints.
No change in behavior.
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
The simple backend only supports a maximum of 6 arguments. Split the
scsi_req_parsed event in two parts to cope with the limit.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
This abstracts calling the command_complete callback, reducing churn
in the following patches.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
One strategy to limit the startup delay of consistency check when
opening image files is to ensure that the file is marked dirty for as
little time as possible.
QED currently marks the image dirty when the first allocating write
request is issued and clears the dirty bit again when the image is
cleanly closed. In practice that means the image is marked dirty for
most of a guest's lifetime and prone to being in a dirty state upon
crash or power failure.
It is safe to clear the dirty bit after all allocating write requests
have completed and a flush has been performed. This patch adds a timer
after the last allocating write request completes. When the timer fires
it will flush and then clear the dirty bit. The timer is set to 5
seconds and is cancelled upon arrival of a new allocating write request.
Signed-off-by: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
This function allows to unlock a ram_ptr give by qemu_get_ram_ptr. After
a call to qemu_put_ram_ptr, the pointer may be unmap from QEMU when
used with Xen.
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
Acked-by: Alexander Graf <agraf@suse.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
On IA32 host or IA32 PAE host, at present, generally, we can't create
an HVM guest with more than 2G memory, because generally it's almost
impossible for Qemu to find a large enough and consecutive virtual
address space to map an HVM guest's whole physical address space.
The attached patch fixes this issue using dynamic mapping based on
little blocks of memory.
Each call to qemu_get_ram_ptr makes a call to qemu_map_cache with the
lock option, so mapcache will not unmap these ram_ptr.
Blocks that do not belong to the RAM, but usually to a device ROM or to
a framebuffer, are handled in a separate function. So the whole RAMBlock
can be map.
Signed-off-by: Jun Nakajima <jun.nakajima@intel.com>
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Signed-off-by: Alexander Graf <agraf@suse.de>
This patch adds support for Milkymist's minimal Ethernet MAC v2. It
superseds minimac1.
Signed-off-by: Michael Walle <michael@walle.cc>
Signed-off-by: Edgar E. Iglesias <edgar.iglesias@gmail.com>
Trace events cannot use %s in their format strings because trace
backends vary in how they can deference pointers (if at all). Recording
const char * values is not meaningful if their contents are not recorded
too.
Change grlib trace events that rely on strings so that they communicate
similar information without using strings.
A follow-up patch explains this limitation and updates docs/tracing.txt.
Signed-off-by: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
It can be handy to know when the guest locks/unlocks the CD-ROM tray.
This trace event makes that possible.
Signed-off-by: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
