Route PC irqs to ISA bus instead of i8259 directly

A PC has its motherboard IRQ lines connected to both the PIC and IOAPIC.
Currently, qemu routes IRQs to the PIC which then calls the IOAPIC, an
incestuous arrangement.  In order to clean this up, create a new ISA IRQ
abstraction, and have devices raise ISA IRQs (which in turn raise the i8259
IRQs as usual).

Signed-off-by: Avi Kivity <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Makefile: fixed rule TAGS

- still works if the build dir is not the src dir
- use find instead of *.c block/*.c etc...

Signed-off-by: Alexandre Bique <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

QEMU set irq0override in fw_cfg

Hi,

After discussing the issue with Avi, Gleb and a couple others on irq,
we came to the conclusion that it is preferred to have QEMU request
features from the BIOS, rather than notifying the BIOS that it is
running on QEMU or KVM. This way memory ranges can change etc. and
an older BIOS will continue to work on newer QEMU if it receives the
info as a fw_cfg value.

This one also matches what qemu-kvm does for irq0override, except I
haven't made it configurable. I leave that as an exercise for whoever
would be interested in switching off irq0override.

Thanks,
Jes

Set irq0 override in fw_cfg, informing the BIOS that QEMU expects
override on irq0. This matches qemu-kvm, and will help sharing a
single BIOS binary.

Signed-off-by: Jes Sorensen <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

SMART ATA Functionality

For the lulz I implemented basic SMART functionality in ide.c.  smartctl
on linux recognizes it just fine and starting self tests with it
complete successfully.

Signed-off-by: Brian Wheeler <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Add missing linefeed in error message

The error message for an unknown network device given to
monitor command set_link looks better with a terminating
linefeed.

Signed-off-by: Stefan Weil <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Clean up VGA type selection; far too many variables being used to track one state leads to confusion if new variables are added.

Signed-off-by: Zachary Amsden <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

When using stdio monitor and VNC display, one can set or clear a VNC password; this should set or turn off VNC authentication as well.

Signed-off-by: Zachary Amsden <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Don't segfault when changing VNC password on an SDL display.

Signed-off-by: Zachary Amsden <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Fix detached migration with exec.

When trying to do detached migration with exec, I found that
the monitor wouldn't always return in a timely manner.  I
tracked this down to exec_start_outgoing_migration.  It
appeared we were setting the fd to NONBLOCK'ing, but in
point of fact we weren't.

This bugfix should also go onto the stable 0.10 branch

Signed-off-by: Chris Lalancette <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Do not disable autostart for live migration

If the user does not want autostart, they can specify -S.

Signed-off-by: Avi Kivity <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

make vga screen_dump use DisplayState properly

Hi all,
currently the vga screen_dump code doesn't use the DisplayState
interface properly and tries to replace it temporarily while taking the
screenshot.
A better approach is to register a DisplayChangeListener, call
vga_hw_update, and finally write the ppm in the next call from dpy_update.

Testing is appreciated.

Signed-off-by: Stefano Stabellini <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>

Restore consistent formatting

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

es1370: Remove unused indirection of PCIES1370State and ES1370State

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

Fix device name completion for 'eject'

Signed-off-by: Blue Swirl <[email protected]>

Revert my commit c00a9de060124a988bd9847c095e5836488c6f01

was incorrect.

Fix segfault of qemu-system-arm with PXA target

qemu-system-arm (0.10.5) segfaults when invoked with a PXA machine target,
e.g. -M tosa. The reason is fairly obvious:

Signed-off-by: Torsten Duwe <[email protected]>
Signed-off-by: Christoph Hellwig <[email protected]>
Signed-off-by: Andrzej Zaborowski <[email protected]>

sdl.c: support 32 bpp cursors

Hello,
currently when a 32 bpp cursor gets defined the result is all-black in
the areas that are not transparent (you'll get a 32 bpp cursor if you
use my previous patch to allow vmware_vga to use a 32 bpp framebuffer).
This is because the switch in sdl.c lacks a 32 bpp case.
The thing I am unsure about though is which byte is the unused one and
should be skipped, the first or the last - for the black-and-white
cursors I tested it doesn't make a difference...

Signed-off-by: Reimar Döffinger <[email protected]>
Signed-off-by: Andrzej Zaborowski <[email protected]>

Use corect depth from DisplaySurface in vmware_vga.c

Hello,
for what I can tell, there is no way for vmware_vga to work correctly
right now. It assumes that the framebuffer bits-per-pixel and the one
from the DisplaySurface are identical (it uses directly the VRAM from
vga.c), but it always assumes 3 bytes per pixel, which is never possible
with the current version of DisplaySurface.
Attached patch fixes that by using ds_get_bits_per_pixel.

Remove the unnecessary and only global in musicpal.c

Merge with [email protected]:/srv/git/qemu.git

Make musicpal.c use the I2C device and the Marvell 88w8618 audio device

Signed-off-by: Benoit Canet <[email protected]>
Signed-off-by: Andrzej Zaborowski <[email protected]>

Extract the Marvell 88w8618 audio device from musicpal.c

Signed-off-by: Benoit Canet <[email protected]>
Signed-off-by: Andrzej Zaborowski <[email protected]>

Extract musicpal.c I2C bitbanging code and make it gpio aware

Signed-off-by: Benoit Canet <[email protected]>
Signed-off-by: Andrzej Zaborowski <[email protected]>

Musicpal qdev conversion: gpio (except I2C part), keyboard and lcd

Signed-off-by: Benoit Canet <[email protected]>
Signed-off-by: Andrzej Zaborowski <[email protected]>

Sparc32: fix monitor commands 'info pic' and 'info irq'

Signed-off-by: Blue Swirl <[email protected]>

Rearrange to suppress gcc 3.3.5 warning about unused variable

Signed-off-by: Blue Swirl <[email protected]>

def-helper.h allow helpers returning pointers

This change allows to define helpers returning pointers.
This looks like a typo in original commit.

Signed-off-by: [email protected]
Signed-off-by: Blue Swirl <[email protected]>

ESP: implement Transfer Pad

Signed-off-by: Blue Swirl <[email protected]>

ESP: Implement select without ATN, fix comments

Signed-off-by: Blue Swirl <[email protected]>

ARM back-end: Handle all possible immediates for ALU ops

this patch handles all possible constants for immediate operand of ALU ops.
I'm not very satisfied by the implementation.

Laurent

Signed-off-by: Laurent Desnogues <[email protected]>
Signed-off-by: Andrzej Zaborowski <[email protected]>

ARM back-end: Add TCG not

this patch:

 - implements TCG not.

Laurent

Signed-off-by: Laurent Desnogues <[email protected]>
Signed-off-by: Andrzej Zaborowski <[email protected]>

sparc32 remove an unnecessary cpu irq set

Signed-off-by: Artyom Tarasenko <[email protected]>
Signed-off-by: Blue Swirl <[email protected]>

Sparc32/64: fix jmpl followed by branch

Fix a case where 'jmpl' instruction followed by a branch instruction was
handled incorrectly.

Signed-off-by: Blue Swirl <[email protected]>

Remove unneded ac97 indirection accessing its state

Searching for "inspiration" to convert another device to qdev, I got
ac97.  Once I understood a bit of qdev, found that ac97 used a not needed
indirection.  To protect the unaware, just fixed it.

Later, Juan.

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

Sparc32/64: Fix user emulator breakage

Signed-off-by: Blue Swirl <[email protected]>

Fix breakage of alpha, mips64, ppc64 and x86_64 targets on non-amd64 host

Signed-off-by: Blue Swirl <[email protected]>

Replace REGX with PRIx64

Signed-off-by: Blue Swirl <[email protected]>

Replace local ADDRX/PADDRX macros with TARGET_FMT_lx/plx

Signed-off-by: Blue Swirl <[email protected]>

Replace always_inline with inline

We define inline as always_inline.

Signed-off-by: Blue Swirl <[email protected]>

user: compile host-utils.c only once

Signed-off-by: Blue Swirl <[email protected]>

Simplify 5ba6531956b9b6486560cbd13604c2238a3542dd

Thanks to Juan Quintela.

Signed-off-by: Blue Swirl <[email protected]>

Fix desynchronization of condition code state when a memory access traps

Signed-off-by: Blue Swirl <[email protected]>

PPC: clean up ppc405

Rely on the subpage system instead of the local version.
Make most functions "static".
Fix wrong parameter passed to ppc4xx_pob_reset.

Signed-off-by: Blue Swirl <[email protected]>

Use pstrcpy to avoid OpenBSD linker warnings

Signed-off-by: Blue Swirl <[email protected]>

bsd-user: Add generic env variable handling

Based on 04a6dfebb6b52532a1e0bd637899f1eba14e94c6.

Adds support for qemu to modify target process environment
variables using -E and -U commandline switches. This replaces
eventually the -drop-ld-preload flag.

Signed-off-by: Blue Swirl <[email protected]>

linux-user: compile envlist.c only once

Signed-off-by: Blue Swirl <[email protected]>

user: compile ioport-user.c only for x86

Signed-off-by: Blue Swirl <[email protected]>

user: compile path.c only once

Also merge bsd-user/path.c and linux-user/path.c.

Signed-off-by: Blue Swirl <[email protected]>

Fix migration for ide devices

commit 93c8cfd9e67a62711b86f4c93747566885eb7928
Author: Gleb Natapov <[email protected]>
Date:   Sun Aug 2 11:36:47 2009 +0300

    make windows notice media change

Broke save/restore by loading a new field but not saving it.

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

Unbreak Sparc64

EBus is a sort of ISA bus.

Signed-off-by: Blue Swirl <[email protected]>

Unbreak Sparc32 and PPC

Convert qdev ptr type to chr, allow a NULL pointer.

Signed-off-by: Blue Swirl <[email protected]>

Third attempt to invoke create_config properly

Second attempt failed due to $_ not being standard and as such it's
interpretation by certain shells when they were symlinked to /bin/sh
and invoked as such led to unpredictable results. So instead of trying
to be clever just use /bin/sh directly (That's what direct execution
would have led to anyway)

Hopefully this time nothing will break (Mingw?)

Thanks to Jordan Justen for report and analysis.
[Previous attempt (THISSHELL one) deserves a credit but reporter is
 too humble]

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

Fix typo

Do not use $SHELL to invoke our scripts

http://www.opengroup.org/onlinepubs/7990989775/xbd/envvar.html

<quote>
SHELL
    A pathname of the user's preferred command language
    interpreter. If this interpreter does not conform to the XSI Shell
    Command Language in the XCU specification, Shell Command Language,
    utilities may behave differently from those described in this
    specification set.
</quote>

So using shells for users who prefer csh variants is a no go.

Aesthetics

Reformat to make item borders more visible
Fix cases of stray tabs and vertical misalignments

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

use C99 initializers for all audio/*

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

use C99 initializers for audio_pcm_ops

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

Use C99 initializers for audio_option

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

Use proper struct initializers and remove INIT_FIELD() macro

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

Fix virtio-blk

Reported-by: Christoph Hellwig
Signed-off-by: Anthony Liguori <[email protected]>

Add emacs header to Makefile.target

Signed-off-by: Anthony Liguori <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>
Message-Id:

check for PR_SET_NAME being defined

Depending on what glibc/kernel headers you are compiling against,
PR_SET_NAME may or may not be defined.  Do the right thing if
PR_SET_NAME isn't defined and skip setting the process name.

Signed-off-by: Nathan Froyd <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>
Message-Id:

rename "info qdrv" to "info qdm"

As requested by avi: driver != device model.

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

fix qdev_print_devinfo()

snprintf returns number of bytes needed for the output, not the number
of bytes actually written.  Thus the math is wrong ...

Spotted by Markus Armbruster.

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

qdev/prop: convert xilinx_timer.c to helper macros.

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

qdev/prop: convert xilinx_intc.c to helper macros.

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

qdev/prop: convert xilinx_ethlite.c to helper macros.

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

qdev/prop: convert virtio-pci.c to helper macros.

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

qdev/prop: convert vga.c to helper macros.

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

qdev/prop: convert tcx.c to helper macros.

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

qdev/prop: convert syborg_timer.c to helper macros.

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

qdev/prop: convert syborg_serial.c to helper macros.

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

qdev/prop: convert syborg_pointer.c to helper macros.

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

qdev/prop: convert syborg_keyboard.c to helper macros.

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

qdev/prop: convert syborg_interrupt.c to helper macros.

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

qdev/prop: convert syborg_fb.c to helper macros.

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

qdev/prop: convert sun4u.c to helper macros.

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

qdev/prop: convert sun4m.c to helper macros.

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

qdev/prop: convert sparc32_dma.c to helper macros.

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

qdev/prop: convert smbus_eeprom.c to helper macros.

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

qdev/prop: convert slavio_timer.c to helper macros.

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

qdev/prop: convert slavio_intctl.c to helper macros.

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

qdev/prop: convert pcnet.c to helper macros.

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

qdev/prop: convert m48t59.c to helper macros.

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

qdev/prop: convert iommu.c to helper macros.

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

qdev/prop: convert integratorcp.c to helper macros.

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

qdev/prop: convert i2c.c to helper macros.

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

qdev/prop: convert etraxfs_pic.c to helper macros.

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

qdev/prop: convert escc.c to helper macros.

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

qdev/prop: convert eccmemctl.c to helper macros.

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

qdev/prop: convert armv7m.c to helper macros.

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

qdev/prop: convert arm_sysctl.c to helper macros.

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

qdev/prop: convert pci.c to helper macros.

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

qdev/prop: add CharDriverState property.

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

qdev/prop: macros for creating typechecked properties.

There are DEFINE_PROP_$TYPE("name", struct, field, default) macros for
each property type.  These macros link the qdev_prop_$name struct to the
type used by that property.  typeof(struct->field) is verifyed to be the
correct one for the given property.

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

make windows notice media change

Windows seems to be very stupid about cdrom media change. It polls
cdrom status and if status goes ready->media not present->ready
it assumes that media was changed. If "media not present" step doesn't
happen even if "medium may have changed" was seen it assumes media
haven't changed. Fake "media not present" step.

Filip Navara did a great job debugging this issue in Windows and this is
what he found out:

BINGO! ... The media present notifications were broken ever since
Windows 2000 it seems. The media change is detected properly and it's
passed to ClassSetMediaChangeState function which in turn calls
ClasspInternalSetMediaChangeState. This function is responsible for
changing some internal state of the device object and sending the PnP
events which later result in application notifications. It has this
tiny bit of code (not copied byte for byte):

if (oldMediaState == NewState) {
  // Media is in the same state it was before.
  return;
}

so the end result is that for the case of UNIT NEEDS ATTENTION /
MEDIUM MAY HAVE CHANGED without NOT READY in-between is really broken.
It results in the internal media change counter incremented, so the
media contents are re-read when necessary, instead of relying on the
cache, but the notifications to applications are never sent.

Signed-off-by: Gleb Natapov <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>
Message-Id:

Fix do_commit() behavior

Commit 751c6a17042b5d011013d6963c0505d671cf708e changed the monitor's
'commit' command to this behavior:

1. Any string you type as argument will cause do_commit() to
call bdrv_commit() to all devices

2. If you enter a device name, it will be the only one ignored
by do_commit() :)

The fix is to call bdrv_commit() to the specified device only and
ignore the others (when 'all' is not specified).

Signed-off-by: Luiz Capitulino <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>
Message-Id:

net: Fix do_set_link() return type

do_set_link() returns int, but Monitor handler functions should
always return void.

Signed-off-by: Luiz Capitulino <[email protected]>
Signed-off-by: Anthony Liguori <[email protected]>
Message-Id:

Add/Fix command-line checks for smbios options v2

- One type 4 table is required per cpu. Add a check for this.
- Fix check for smbios file.

Changes from v1:
- static designation of smbios_validate_table, and remove whitespace

Signed-off-by: Beth Kon <[email protected]>
--
Signed-off-by: Anthony Liguori <[email protected]>
Message-Id:

Move libsunpath to use the same style than everything else

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