Fixed wacom emulation

- for absolute mode, scale coordinates to the real device maximum values,
since some drivers (on Haiku and Linux at least) need them as such,
and the HID descriptor is boggus on some models anyway,
- keep the coordinates even when no button is pressed, on real tablet
the pen is sensed on the surface even without direct contact,
and drivers expect this,
- map left button to pressure according to what the Haiku driver wants,
- map the right button to the pen button,
- map the middle button to the eraser,
- use asynchronous reporting as the hid code does, stops the Haiku driver
(and probably others) from spending 50% cpu polling for changes.

Signed-off-by: François Revol <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

rom loader: also try filename as-is.

In case qemu_find_file fails try to open the file as-is.

Patchworks-ID: 35263
Signed-off-by: Gerd Hoffmann <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

rom loader: fix sparc -kernel boot.

Changes:
(1) register pstrcpy_targphys() in rom list, it is used for kernel
command lines by a number of architectures.
(2) add rom_ptr() function to get a pointer for applying changes
to loaded images.  Needed for example to tell the linux kernel
where it finds the initrd image by updating the header.
(3) make sparc use rom_ptr for initrd setup.

booting sparc-test works now, and 'info roms' shows this:

(qemu) info roms
addr=0000000000000000 size=0x2a3828 mem=ram name="phdr #0: vmlinux-2.6.11+tcx"
addr=00000000007ff000 size=0x00000e mem=ram name="cmdline"
addr=0000000000800000 size=0x400000 mem=ram name="/root/qemu-test/sparc-test/linux.img"
addr=0000000070000000 size=0x0e4000 mem=rom name="phdr #0: /home/kraxel/projects/qemu/build-zfull/pc-bios/openbios-sparc32"

reboot via 'system_reset' works too.

Patchworks-ID: 35262
Signed-off-by: Gerd Hoffmann <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

unlock iothread mutex before running kvm ioctl

Without this, kvm will hold the mutex while it issues its run ioctl,
and never be able to step out of it, causing a deadlock.

Patchworks-ID: 35359
Signed-off-by: Glauber Costa <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

TARGET_I386 is always defined if TARGET_X86_64 is defined

Patchworks-ID: 35378
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

slirp: fix !CONFIG_SLIRP compilation

This moves the code that depens on slirp under CONFIG_SLIRP again.

Patchworks-ID: 35372
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

ide: BMDMAState don't need a pci_dev field anymore

Patchworks-ID: 35306
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

ide: cmd646 ->unit has just the value that we want

Patchworks-ID: 35307
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

ide: cmd646 we can get the pci device with container_of

Patchworks-ID: 35305
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

ide: 'secondary' field is only used by cmd646

Patchworks-ID: 35303
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

ide: PCIIDEState type field is not needed anymore

We have split the functions that needed it for cmd646

Patchworks-ID: 35302
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

ide: split cmd646 and piix from pci.c

This patch splits cmd646 specific code from pci.c.
This patch splits piix4 specific code from pci.c.
And compile new piix.o and cmd646.o when they are needed.
The only change that is not code movemet is removal of cmd646 specific parts
in bmdma_readb/writeb for piix.

Patchworks-ID: 35301
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

ide: export needed ide-pci functions for split

Patchworks-ID: 35300
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

ide: create ide/pci.h for common ide pci definitions

Patchworks-ID: 35299
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

ide: remove uselsess casts from void *

Patchworks-ID: 35298
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

ide: Remove duplicated definitions

Patchworks-ID: 35297
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

ide: Remove cast in pci_register_bar

We already have a PCIDevice at that point

Patchworks-ID: 35296
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

ide: change cast to DO_UPCAST

Patchworks-ID: 35293
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Fix bswap in comment

Replace bsawp -> bswap.

Signed-off-by: Stefan Weil <[email protected]>
Signed-off-by: Aurelien Jarno <[email protected]>

Move dependency generation falgs out of configure

Signed-off-by: malc <[email protected]>

configure: clean up temporary dependency files

Signed-off-by: malc <[email protected]>

winwave: pause/restore playing upon seeing VOICE_DISABLE/ENABLE

Signed-off-by: malc <[email protected]>

winwave: close event handle and delete wait object after closing HWAVEOUT

To avoid possibly being called back and thus racing.

Signed-off-by: malc <[email protected]>

winwave: remove wait object when finalizing DAC voice

Signed-off-by: malc <[email protected]>

CRIS: Segmented addressing only for kernel mode.

Segmented translation through the CRIS MMU is only done for
accesses in kernel mode. In user-mode, all accesses are treated
as paged regardless of the mode config in RW_MM_CFG.

Signed-off-by: Edgar E. Iglesias <[email protected]>

Fix Windows host breakage by 45a50b1668822c23afc2a89f724654e176518bc4 (TeLeMan)

Signed-off-by: malc <[email protected]>

winwave: poll mode

Signed-off-by: malc <[email protected]>

Windows Waveform Audio driver (no ADC support yet)

Signed-off-by: malc <[email protected]>

pcnet: Restart poll timer on pcnet_start

Just like we call into pcnet_poll_timer on stop, we need to call it on
start to trigger the setup of the poll timer.

Patchworks-ID: 35313
Signed-off-by: Jan Kiszka <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

pcnet: Drop unused recv_pos field

This state field was never used, simply remained 0. Drop it from the
PCNetState and update the save/restore code accordingly, keeping
backward compatibility.

Patchworks-ID: 35314
Signed-off-by: Jan Kiszka <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

monitor: Convert do_info_cpus() to QObject

Each CPU information is stored in a QDict and the returned
QObject is a QList of all CPUs.

The QDict contains the following information:

- "CPU": cpu index
- "current": "yes" or "no"
- "pc": current PC
- "halted": "yes" or "no"

The user output in the Monitor should not change and the
future monitor protocol is expected to emit something like:

[ { "CPU": 0, "current": "yes", "pc": 0x..., "halted": "no" },
{ "CPU": 1, "current": "no",  "pc": 0x..., "halted": "yes" } ]

which corresponds to the following user output:

* CPU #0: pc=0x00000000fffffff0
CPU #1: pc=0x00000000fffffff0 (halted)

Patchworks-ID: 35352
Signed-off-by: Luiz Capitulino <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

monitor: Convert do_info_balloon() to QObject

On success return a QInt with the balloon's value.

This also introduces monitor_print_balloon() to print the
balloon information in the user protocol.

Please, note that errors are not being converted yet.

Patchworks-ID: 35351
Signed-off-by: Luiz Capitulino <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

monitor: Convert do_info_version() to QObject

The returned data is always a QString.

Also introduces monitor_print_qobject(), which can be used as
a standard way to print QObjects in the user protocol format.

Patchworks-ID: 35350
Signed-off-by: Luiz Capitulino <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

monitor: Convert do_balloon() to QObject

It is important to note that it never fails, as big refactoring
of the virtio code would be needed to get the proper error code.

Patchworks-ID: 35349
Signed-off-by: Luiz Capitulino <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

monitor: Convert do_cont() to QObject

Appropriate error handling support will be needed to have
encrypted images working under the future machine protocol,
but this initial conversion will work with the current
user protocol.

Patchworks-ID: 35348
Signed-off-by: Luiz Capitulino <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

monitor: Convert do_system_powerdown() to QObject

Patchworks-ID: 35346
Signed-off-by: Luiz Capitulino <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

monitor: Convert do_system_reset() to QObject

Patchworks-ID: 35347
Signed-off-by: Luiz Capitulino <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

monitor: Convert do_stop() to QObject

Patchworks-ID: 35343
Signed-off-by: Luiz Capitulino <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

monitor: Convert do_quit() do QObject

Patchworks-ID: 35345
Signed-off-by: Luiz Capitulino <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

monitor: do_info(): handle new and old info handlers

do_info() is special, its job is to call 'info handlers'.
This is similar to what monitor_handle_command() does,
therefore do_info() also has to distinguish among new and
old style info handlers.

This commit converts do_info() to the new QObject style and
makes the appropriate changes so that it can handle both
info handlers styles.

In the future, when all handlers are converted to QObject's
style, it will be possible to share more code with
monitor_handle_command().

This commit also introduces a new function called
monitor_user_noop(), it should be used by handlers which do
not have data to print.

This is the case of do_info().

Patchworks-ID: 35341
Signed-off-by: Luiz Capitulino <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

monitor: Handle new and old style handlers

This commit changes monitor_handle_command() to support old style
_and_ new style handlers.

New style handlers are protocol independent, they return their
data to the Monitor, which in turn decides how to print them
(ie. user protocol vs. machine protocol).

Converted handlers will use the 'user_print' member of 'mon_cmd_t'
to define its user protocol function, which will be called to print
data in the user protocol format.

Handlers which don't have 'user_print' defined are not converted
and are handled as usual.

Patchworks-ID: 35340
Signed-off-by: Luiz Capitulino <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

monitor: Add user_print() to mon_cmd_t

This new struct member will store a pointer to a function that
should be used to output data in the user protocol format.

It will also serve as a flag to say if a given handler has already
been converted to the new QObject style.

Patchworks-ID: 35339
Signed-off-by: Luiz Capitulino <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

monitor: union for command handlers

This commits adds a new union member to mon_cmd_t for command
handlers and convert monitor_handle_command() and qemu-monitor.hx
to use it.

This improves type safety.

Patchworks-ID: 35337
Signed-off-by: Luiz Capitulino <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

monitor: union for info handlers

This commit adds a union to mon_cmd_t for info handlers and
converts do_info() and info_cmds[] array to use it.

This improves type safety.

Next commit will convert command handlers.

Patchworks-ID: 35336
Signed-off-by: Luiz Capitulino <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

monitor: Convert mon_cmd_t initializations to C99 style

Patchworks-ID: 35335
Signed-off-by: Luiz Capitulino <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Introduce QList unit-tests

This suite contains tests to assure that QList API works as expected.

To execute it you should have check installed and build QEMU with
check support enabled (--enable-check-utests) and then run:

$ ./check-qlist

Patchworks-ID: 35333
Signed-off-by: Luiz Capitulino <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Introduce QList

QList is a high-level data type that can be used to store QObjects
in a singly-linked list.

The following functions are available:

- qlist_new()    Create a new QList
- qlist_append() Append a QObject to the list
- qlist_iter()   Iterate over stored QObjects

Patchworks-ID: 35334
Signed-off-by: Luiz Capitulino <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

QObject: Accept NULL

It is convenient that QDECREF() and QINCREF() accept the QObject
parameter to be NULL, so that we don't duplicate 'if' tests in
the callers.

Patchworks-ID: 35332
Signed-off-by: Luiz Capitulino <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Only compile ssi when one target uses it

Patchworks-ID: 35221
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Only compile ssi-sd when one target uses it

Patchworks-ID: 35220
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Only compile max111x when one target uses it

Patchworks-ID: 35219
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Only compile ads7846 when one target uses it

Patchworks-ID: 35218
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Only compile sd0323 when one target uses it

Patchworks-ID: 35217
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Only compile sd0303 when one target uses it

Patchworks-ID: 35216
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Only compile stellaris_input when one target uses it

Patchworks-ID: 35213
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Only compile tmp105 when one target uses it

Patchworks-ID: 35214
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Only compile lm832x when one target uses it

Patchworks-ID: 35211
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Only compile tsc2005 when one target uses it

Patchworks-ID: 35215
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Only compile twl92230 when one target uses it

Patchworks-ID: 35212
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Only compile wm8750 when one target uses it

Patchworks-ID: 35209
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Only compile max7310 when one target uses it

Patchworks-ID: 35210
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Only compile sd when one target uses it

Patchworks-ID: 35208
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Only compile ptimer when one target uses it

Patchworks-ID: 35207
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Only compile m48t59 when one target uses it

Patchworks-ID: 35204
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Only compile escc when one target uses it

Patchworks-ID: 35205
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Only compile esp when one target uses it

Patchworks-ID: 35206
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Only compile ecc when one target uses it

Patchworks-ID: 35203
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Only compile nand when one target uses it

Patchworks-ID: 35202
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Only compile qdev_addr when one target uses it

Patchworks-ID: 35201
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Only compile isa_mmio when one target uses it

Patchworks-ID: 35200
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Only compile usb_ohci when one target uses it

Patchworks-ID: 35199
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Generate gdbstub-xml.c only when needed

First user of new config-devices.mak

Patchworks-ID: 35198
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Generate config-devices.h

Generate config-devices.h for each target and config-all-devices.h for
common library.  We don't want to name both config-devices.h to avoid
path problems

Patchworks-ID: 35195
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Add new config-devices.mak for each target

We generate config-devices.h from there automatically.
We need to do it in main Makefile, because we are going to need a main
Makefile for them.

Patchworks-ID: 35196
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Move generation of config-target.h to Makefile from configure

Patchworks-ID: 35194
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Rename config.{h, mak} config-target.{h, mak}

Add config.h file that includes config-target.h and config-host.h

Patchworks-ID: 35193
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Remove useless check for config-host.mak

If config-host.mak dont' exist, we have exited in the check at
the beginning of the file.
Once here, move the bits to the else part of the test at the beginning of
the file.

Patchworks-ID: 35191
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Move generation of config-host.h to Makefile from configure

Use timestamp based appreach to avoid not needed recompilation.
Add it to rules.mak

Many thanks to Paolo Bonzini for helpding the design, and the debug.

Patchworks-ID: 35190
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Don't include config-host.mak from inside config.mak

Include it directly in Makefile.target

Patchworks-ID: 35189
Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

CRIS: Update PR_EDA on TLB faults.

Signed-off-by: Edgar E. Iglesias <[email protected]>

Clean up test for qdev_init() failure

Some callers test for != 0, some for < 0.  Normalize to < 0.

Patchworks-ID: 35171
Signed-off-by: Markus Armbruster <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Warn if value of qdev_init() isn't checked

After qdev_init() fails, the device is gone.  Failure to check runs a
high risk of use-after-free.

Patchworks-ID: 35166
Signed-off-by: Markus Armbruster <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Make isa_create() terminate program on failure

Callers don't check the return value anyway.

Patchworks-ID: 35172
Signed-off-by: Markus Armbruster <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

New qdev_init_nofail()

Like qdev_init(), but terminate program via hw_error() instead of
returning an error value.

Use it instead of qdev_init() where terminating the program on failure
is okay, either because it's during machine construction, or because
we know that failure can't happen.

Because relying in the latter is somewhat unclean, and the former is
not always obvious, it would be nice to go back to qdev_init() in the
not-so-obvious cases, only with proper error handling.  I'm leaving
that for another day, because it involves making sure that error
values are properly checked by all callers.

Patchworks-ID: 35168
Signed-off-by: Markus Armbruster <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Check return value of qdev_init()

But do so only where it may actually fail.  Leave the rest for the
next commit.

Patchworks-ID: 35167
Signed-off-by: Markus Armbruster <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Make qdev_init() destroy the device on failure

Before, every caller had to do this.  Only two actually did.

Patchworks-ID: 35170
Signed-off-by: Markus Armbruster <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Unbreak USB autoconnect filters

Commit 22f84e73 added a qdev_init() missing on the path through
usb_host_device_open(), but that broke the path through
usb_host_auto_scan(), which already had one.  Remove that one.

Patchworks-ID: 35169
Signed-off-by: Markus Armbruster <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

tcg: improve output log

Signed-off-by: Aurelien Jarno <[email protected]>

target-i386: fix ARPL

The arpl implementation in target-i386/translate.c uses cpu_A0
temporary across a brcond op.  This patch fixes that issue.

Signed-off-by: Laurent Desnogues <[email protected]>
Signed-off-by: Aurelien Jarno <[email protected]>

configure and Makefile are not generated in qemu

Just say it to make

Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Don't use implicit rules for Makefile

This remove implicit rules + implicit variables.
Explicit rules like the generation of %.h and %.c from %.hx still works
as expected.

As an added bonus, now the output of make -d is readable.

As another added bonus, time spend on Makefiles is way smaller.
We run make -j3 in a fully compiled tree, and results are:

Before:

$ time make -j3
real 0m1.225s
user 0m1.660s
sys 0m0.253s

After:
$ time make -j3
real 0m0.422s
user 0m0.393s
sys 0m0.248s

Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

optionrom: create .PHONY variable

Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

add build-all to .PHONY rules

Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

optionrom: remove use of implicit RM variable

Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

AIOLIBS is not used anywhere

Signed-off-by: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

qdev: move comment

Move comment back next to main_system_bus to avoid confusion.

Signed-off-by: Gerd Hoffmann <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

qemu/pci: make pci not depend on msix

Making pci device cleanup msix automatically makes pci.c depend on
msix.c, which is IMO messy.  Since devices do msix_init it's easy and
natural for them to also do msix_uninit.

Signed-off-by: Michael S. Tsirkin <[email protected]>
Cc: Gerd Hoffmann <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

qemu/pci: clarify pci config load routine

PCI load routine has to be called with size equal to 256 (otherwise it
will crash in weird ways).  So assert this, making code clearer.
Also avoid dynamically sized array on stack - good for portability.

Signed-off-by: Michael S. Tsirkin <[email protected]>
Cc: Juan Quintela <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Final net cleanup after conversion to QemuOpts

Now that net_client_init() has no users, kill it off and rename
net_client_init_from_opts().

There is no further need for the old code in net_client_parse() either.
We use qemu_opts_parse() 'firstname' facitity for that. Instead, move
the special handling of the 'vmchannel' type there.

Simplify the vl.c code into merely call net_client_parse() for each
-net command line option and then calling net_init_clients() later
to iterate over the options and create the clients.

Signed-off-by: Mark McLoughlin <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Port PCI NIC hotplug to QemuOpts

Signed-off-by: Mark McLoughlin <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>