Peter Maydell [Tue, 15 Mar 2016 09:13:06 +0000 (09:13 +0000)]
Merge remote-tracking branch 'remotes/kevin/tags/for-upstream' into staging
Block layer patches
# gpg: Signature made Mon 14 Mar 2016 16:36:52 GMT using RSA key ID C88F2FD6
# gpg: Good signature from "Kevin Wolf <kwolf@redhat.com>"
* remotes/kevin/tags/for-upstream: (40 commits)
iotests: Add test for QMP event rates
monitor: Use QEMU_CLOCK_VIRTUAL for the event queue in qtest mode
monitor: Separate QUORUM_REPORT_BAD events according to the node name
quorum: Fix crash in quorum_aio_cb()
iotests: Correct 081's reference output
block: Remove unused typedef of BlockDriverDirtyHandler
block: Move block dirty bitmap code to separate files
typedefs: Add BdrvDirtyBitmap
block: Include hbitmap.h in block.h
backup: Use Bitmap to replace "s->bitmap"
vpc: Use BB functions in .bdrv_create()
vmdk: Use BB functions in .bdrv_create()
vhdx: Use BB functions in .bdrv_create()
vdi: Use BB functions in .bdrv_create()
sheepdog: Use BB functions in .bdrv_create()
qed: Use BB functions in .bdrv_create()
qcow2: Use BB functions in .bdrv_create()
qcow: Use BB functions in .bdrv_create()
parallels: Use BB functions in .bdrv_create()
block: Introduce blk_set_allow_write_beyond_eof()
...
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
We discriminate here between opcodes that are illegal in the current
cpu mode or with illegal arguments (such as modrm.mod == 3) and
encodings that are unknown (such as an unimplemented isa extension).
Signed-off-by: Richard Henderson <rth@twiddle.net>
Paolo Bonzini [Wed, 2 Mar 2016 15:04:38 +0000 (16:04 +0100)]
target-i386: Fix addr16 prefix
While ADDSEG will only be false in 16-bit mode for LEA, it can be
false even in other cases when 16-bit addresses are obtained via
the 67h prefix in 32-bit mode. In this case, gen_lea_v_seg forgets
to add a nonzero FS or GS base if CS/DS/ES/SS are all zero. This
case is pretty rare but happens when booting Windows 95/98, and
this patch fixes it.
The bug is visible since commit d6a291498, but it was introduced
together with gen_lea_v_seg and it probably could be reproduced
with a "addr16 gs movsb" instruction as early as in commit ca2f29f555805d07fb0b9ebfbbfc4e3656530977.
Reported-by: Hervé Poussineau <hpoussin@reactos.org> Tested-by: Hervé Poussineau <hpoussin@reactos.org> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <1456931078-21635-1-git-send-email-pbonzini@redhat.com> Signed-off-by: Richard Henderson <rth@twiddle.net>
In non-64-bit modes, the instruction always stores 16 bits.
But in 64-bit mode, when the destination is a register, the
instruction can write 32 or 64 bits.
Tested-by: Hervé Poussineau <hpoussin@reactos.org> Signed-off-by: Richard Henderson <rth@twiddle.net>
Paolo Bonzini [Tue, 1 Mar 2016 15:12:25 +0000 (16:12 +0100)]
target-i386: Avoid repeated calls to the bnd_jmp helper
Two flags were tested the wrong way.
Tested-by: Hervé Poussineau <hpoussin@reactos.org> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <1456845145-18891-1-git-send-email-pbonzini@redhat.com> Signed-off-by: Richard Henderson <rth@twiddle.net>
[rth: Fixed enable test as well.]
Kevin Wolf [Mon, 14 Mar 2016 16:36:31 +0000 (17:36 +0100)]
Merge remote-tracking branch 'mreitz/tags/pull-block-for-kevin-2016-03-14-v2' into queue-block
Block patches for pi day, v2.
# gpg: Signature made Mon Mar 14 17:35:29 2016 CET using RSA key ID E838ACAD
# gpg: Good signature from "Max Reitz <mreitz@redhat.com>"
* mreitz/tags/pull-block-for-kevin-2016-03-14-v2:
iotests: Add test for QMP event rates
monitor: Use QEMU_CLOCK_VIRTUAL for the event queue in qtest mode
monitor: Separate QUORUM_REPORT_BAD events according to the node name
quorum: Fix crash in quorum_aio_cb()
iotests: Correct 081's reference output
block: Remove unused typedef of BlockDriverDirtyHandler
block: Move block dirty bitmap code to separate files
typedefs: Add BdrvDirtyBitmap
block: Include hbitmap.h in block.h
backup: Use Bitmap to replace "s->bitmap"
Alberto Garcia [Thu, 10 Mar 2016 11:55:27 +0000 (13:55 +0200)]
iotests: Add test for QMP event rates
This test verifies that the rate-limited QMP events are emitted at a
maximum rate of 1 per second as defined in monitor_qapi_event_conf in
monitor.c
It also checks that QUORUM_REPORT_BAD events generated from different
nodes are kept in separate queues so they don't mask each other.
Signed-off-by: Alberto Garcia <berto@igalia.com> Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-id: 0dbd3ee88a59a6363042ad81cfb345037bfbf612.1457610443.git.berto@igalia.com
[mreitz@redhat.com: Renamed test from 146 to 148] Signed-off-by: Max Reitz <mreitz@redhat.com>
Alberto Garcia [Thu, 10 Mar 2016 11:55:26 +0000 (13:55 +0200)]
monitor: Use QEMU_CLOCK_VIRTUAL for the event queue in qtest mode
This allows us to perform tests on the monitor queues to verify that
the rate limits are enforced.
Signed-off-by: Alberto Garcia <berto@igalia.com> Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-id: dde511809e954a5c32d5b648bb184c03c89ed5d5.1457610443.git.berto@igalia.com Signed-off-by: Max Reitz <mreitz@redhat.com>
Alberto Garcia [Thu, 10 Mar 2016 11:55:25 +0000 (13:55 +0200)]
monitor: Separate QUORUM_REPORT_BAD events according to the node name
The QUORUM_REPORT_BAD event is emitted whenever there's an I/O error
in a child of a Quorum device. This event is emitted at a maximum rate
of 1 per second. This means that an error in one of the children will
mask errors in the other children if they happen within the same 1
second interval.
This patch modifies qapi_event_throttle_equal() so QUORUM_REPORT_BAD
events are kept separately if they come from different children.
Signed-off-by: Alberto Garcia <berto@igalia.com> Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-id: b989c0cb3755bc4b6696e796fa8ed2ef6c56606a.1457610443.git.berto@igalia.com Reviewed-by: Eric Blake <eblake@redhat.com> Signed-off-by: Max Reitz <mreitz@redhat.com>
Alberto Garcia [Thu, 10 Mar 2016 11:55:24 +0000 (13:55 +0200)]
quorum: Fix crash in quorum_aio_cb()
quorum_aio_cb() emits the QUORUM_REPORT_BAD event if there's
an I/O error in a Quorum child. However sacb->aiocb must be
correctly initialized for this to happen. read_quorum_children() and
read_fifo_child() are not doing this, which results in a QEMU crash.
Signed-off-by: Alberto Garcia <berto@igalia.com> Reviewed-by: Max Reitz <mreitz@redhat.com>
Message-id: 8138570d071ba7e25db3736979234a1fd71dbd05.1457610443.git.berto@igalia.com Signed-off-by: Max Reitz <mreitz@redhat.com>
Max Reitz [Fri, 11 Mar 2016 14:14:47 +0000 (15:14 +0100)]
iotests: Correct 081's reference output
The newly added type parameter for the QUORUM_REPORT_BAD event changed
the output of iotest 081, so the reference should be amended
accordingly.
Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-id: 1457705687-27122-1-git-send-email-mreitz@redhat.com Reviewed-by: Alberto Garcia <berto@igalia.com>
Fam Zheng [Tue, 8 Mar 2016 04:44:52 +0000 (12:44 +0800)]
backup: Use Bitmap to replace "s->bitmap"
"s->bitmap" tracks done sectors, we only check bit states without using any
iterator which HBitmap is good for. Switch to "Bitmap" which is simpler and
more memory efficient.
Meanwhile, rename it to done_bitmap, to reflect the intention.
Signed-off-by: Fam Zheng <famz@redhat.com> Reviewed-by: John Snow <jsnow@redhat.com>
Message-id: 1457412306-18940-2-git-send-email-famz@redhat.com Signed-off-by: Max Reitz <mreitz@redhat.com>
Peter Maydell [Mon, 14 Mar 2016 16:22:17 +0000 (16:22 +0000)]
Merge remote-tracking branch 'remotes/stefanha/tags/tracing-pull-request' into staging
# gpg: Signature made Mon 14 Mar 2016 11:27:01 GMT using RSA key ID 81AB73C8
# gpg: Good signature from "Stefan Hajnoczi <stefanha@redhat.com>"
# gpg: aka "Stefan Hajnoczi <stefanha@gmail.com>"
* remotes/stefanha/tags/tracing-pull-request:
trace: separate MMIO tracepoints from TB-access tracepoints
trace: include CPU index in trace_memory_region_*()
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Kevin Wolf [Tue, 8 Mar 2016 14:57:05 +0000 (15:57 +0100)]
vpc: Use BB functions in .bdrv_create()
All users of the block layers are supposed to go through a BlockBackend.
The .bdrv_create() implementation is one such user, so this patch
converts it.
Kevin Wolf [Tue, 8 Mar 2016 14:57:05 +0000 (15:57 +0100)]
vmdk: Use BB functions in .bdrv_create()
All users of the block layers are supposed to go through a BlockBackend.
The .bdrv_create() implementation is one such user, so this patch
converts it.
Kevin Wolf [Tue, 8 Mar 2016 14:57:05 +0000 (15:57 +0100)]
vhdx: Use BB functions in .bdrv_create()
All users of the block layers are supposed to go through a BlockBackend.
The .bdrv_create() implementation is one such user, so this patch
converts it.
Kevin Wolf [Tue, 8 Mar 2016 14:57:05 +0000 (15:57 +0100)]
vdi: Use BB functions in .bdrv_create()
All users of the block layers are supposed to go through a BlockBackend.
The .bdrv_create() implementation is one such user, so this patch
converts it.
Kevin Wolf [Tue, 8 Mar 2016 14:57:05 +0000 (15:57 +0100)]
sheepdog: Use BB functions in .bdrv_create()
All users of the block layers are supposed to go through a BlockBackend.
The .bdrv_create() implementation is one such user, so this patch
converts it.
Kevin Wolf [Tue, 8 Mar 2016 14:57:05 +0000 (15:57 +0100)]
qed: Use BB functions in .bdrv_create()
All users of the block layers are supposed to go through a BlockBackend.
The .bdrv_create() implementation is one such user, so this patch
converts it.
Kevin Wolf [Tue, 8 Mar 2016 14:57:05 +0000 (15:57 +0100)]
qcow2: Use BB functions in .bdrv_create()
All users of the block layers are supposed to go through a BlockBackend.
The .bdrv_create() implementation is one such user, so this patch
converts it.
Kevin Wolf [Tue, 8 Mar 2016 14:57:05 +0000 (15:57 +0100)]
qcow: Use BB functions in .bdrv_create()
All users of the block layers are supposed to go through a BlockBackend.
The .bdrv_create() implementation is one such user, so this patch
converts it.
Kevin Wolf [Tue, 8 Mar 2016 14:57:05 +0000 (15:57 +0100)]
parallels: Use BB functions in .bdrv_create()
All users of the block layers are supposed to go through a BlockBackend.
The .bdrv_create() implementation is one such user, so this patch
converts it.
Kevin Wolf [Tue, 23 Feb 2016 16:33:24 +0000 (17:33 +0100)]
hmp: 'drive_add -n' for creating a node without BB
This patch adds an option to the drive_add HMP command to create only a
BlockDriverState without a BlockBackend on top.
The motivation for this is that libvirt needs to specify options to a
migration target (specifically, detect-zeroes). drive-mirror doesn't
allow specifying options, and the proper way to do this is to create the
target BDS separately with blockdev-add (where you can specify options)
and then use blockdev-mirror to that BDS.
However, libvirt can't use blockdev-add as long as it is still
experimental, and we're expecting that it will still take some time, so
we need to resort to drive_add.
The problem with drive_add is that so far it always created a BB, and
BDSes with a BB can't be used as a mirroring target as long as we don't
support multiple BBs per BDS - and while we're working towards that
goal, it's another thing that will still take some time.
So to achieve the goal, the simplest solution to provide the
functionality now without adding one-off options to the mirror QMP
commands is to extend drive_add to create nodes without BBs.
Fam Zheng [Tue, 8 Mar 2016 08:24:34 +0000 (16:24 +0800)]
vmdk: Switch to heap arrays for vmdk_write_cid
It is only called once for each opened image, so we can do it the easy
way.
Reviewed-by: Peter Xu <peterx@redhat.com> Signed-off-by: Fam Zheng <famz@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Kevin Wolf [Mon, 7 Mar 2016 13:23:04 +0000 (14:23 +0100)]
block: Fix cache mode defaults in bds_tree_init()
Without setting explicit defaults in the options, blockdev-add without
an ID ended up defaulting to writethrough. It should be writeback as
documented.
Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com>
Kevin Wolf [Mon, 7 Mar 2016 12:02:15 +0000 (13:02 +0100)]
block: Fix snapshot=on cache modes
Since commit 91a097e, we end up with a somewhat weird cache mode
configuration with snapshot=on: The commit broke the cache mode
inheritance for the snapshot overlay so that it is opened as
writethrough instead of unsafe now. The following bdrv_append() call to
put it on top of the tree swaps the WCE flag with the snapshot's backing
file (i.e. the originally given file), so what we eventually get is
cache=writeback on the temporary overlay and
cache=writethrough,cache.no-flush=on on the real image file.
This patch changes things so that the temporary overlay gets
cache=unsafe again like it used to, and the real images get whatever the
user specified. This means that cache.direct is now respected even with
snapshot=on, and in the case of committing changes, the final flush is
no longer ignored except explicitly requested by the user.
Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com>
Kevin Wolf [Wed, 2 Mar 2016 11:16:44 +0000 (12:16 +0100)]
blockdev: Snapshotting must not open second instance of old top
Calling bdrv_img_create() with a size of -1 means that it determines the
size automatically by opening the backing file. However, in the case of
live snapshots, the backing file is already opened and we must avoid
opening the same image twice at the same time. Apart from that, just
getting the size from the already existing BDS is a lot less overhead
than opening a new instance.
Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Jeff Cody <jcody@redhat.com>
Cc: Dr. David Alan Gilbert <dgilbert@redhat.com> Cc: Wen Congyang <wency@cn.fujitsu.com> Signed-off-by: Wen Congyang <wency@cn.fujitsu.com> Signed-off-by: Changlong Xie <xiecl.fnst@cn.fujitsu.com> Reviewed-by: Alberto Garcia <berto@igalia.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Changlong Xie [Fri, 26 Feb 2016 01:39:01 +0000 (09:39 +0800)]
qmp event: Refactor QUORUM_REPORT_BAD
Introduce QuorumOpType, and make QUORUM_REPORT_BAD compatible
with it.
Cc: Dr. David Alan Gilbert <dgilbert@redhat.com> Cc: Wen Congyang <wency@cn.fujitsu.com> Signed-off-by: Wen Congyang <wency@cn.fujitsu.com> Signed-off-by: Changlong Xie <xiecl.fnst@cn.fujitsu.com> Reviewed-by: Alberto Garcia <berto@igalia.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Changlong Xie [Fri, 26 Feb 2016 01:39:00 +0000 (09:39 +0800)]
docs: fix invalid node name in qmp event
Cc: Dr. David Alan Gilbert <dgilbert@redhat.com> Cc: Wen Congyang <wency@cn.fujitsu.com> Signed-off-by: Wen Congyang <wency@cn.fujitsu.com> Signed-off-by: Changlong Xie <xiecl.fnst@cn.fujitsu.com> Reviewed-by: Alberto Garcia <berto@igalia.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Jeff Cody [Thu, 25 Feb 2016 17:27:29 +0000 (12:27 -0500)]
block/vpc: give option to force the current_size field in .bdrv_create
When QEMU creates a VHD image, it goes by the original spec,
calculating the current_size based on the nearest CHS geometry (with an
exception for disks > 127GB).
Apparently, Azure will only allow images that are sized to the nearest
MB, and the current_size as calculated from CHS cannot guarantee that.
Allow QEMU to create images similar to how Hyper-V creates images, by
setting current_size to the specified virtual disk size. This
introduces an option, force_size, to be passed to the vpc format during
image creation, e.g.:
qemu-img convert -f raw -o force_size -O vpc test.img test.vhd
When using the "force_size" option, the creator app field used by
QEMU will be "qem2" instead of "qemu", to indicate the difference.
In light of this, we also add parsing of the "qem2" field during
vpc_open.
Jeff Cody [Thu, 25 Feb 2016 17:27:28 +0000 (12:27 -0500)]
block/vpc: tests for auto-detecting VPC and Hyper-V VHD images
This tests auto-detection, and overrides, of VHD image sizes created
by Virtual PC, Hyper-V, and Disk2vhd.
This adds three sample images:
hyperv2012r2-dynamic.vhd.bz2 - dynamic VHD image created with Hyper-V
virtualpc-dynamic.vhd.bz2 - dynamic VHD image created with Virtual PC
d2v-zerofilled.vhd.bz2 - dynamic VHD image created with Disk2vhd
Signed-off-by: Jeff Cody <jcody@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Jeff Cody [Thu, 25 Feb 2016 17:27:27 +0000 (12:27 -0500)]
block/vpc: choose size calculation method based on creator_app field
The VHD file format is used by both Virtual PC, and Hyper-V. However,
how the virtual disk size is calculated varies between the two.
Virtual PC uses the CHS drive parameters to determine the drive size.
Hyper-V, on the other hand, uses the current_size field in the footer
when determining image size.
This is problematic for a few reasons:
* VHD images from Hyper-V, using CHS calculations, will likely be
trunctated.
* If we just rely always on current_size, then QEMU may have data
compatibility issues with Virtual PC (we may write too much data
into a VHD file to be used by Virtual PC, for instance).
* Existing VHD images created by QEMU have used the CHS calculations,
except for images exceeding the 127GB limit. We want to remain
compatible with our own generated images.
Luckily, the VHD specification defines a 'Creator App' field, that is
used to indicate what software created the VHD file.
This patch does two things:
1. Uses the 'Creator App' field to help determine how to calculate
size, and
2. Adds a VPC format option 'force_size_calc', so that the user can
override the 'Creator App' auto-detection, in case there exist
VHD images with unknown or contradictory 'Creator App' entries.
N.B.: We currently use the maximum CHS value as an indication to use the
current_size field. This patch does not change that, even with the
'force_size_calc' option.
Signed-off-by: Jeff Cody <jcody@redhat.com> Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Kevin Wolf [Fri, 26 Feb 2016 20:16:38 +0000 (21:16 +0100)]
block/qapi: Include empty drives in query-blockstats
Since commit 5ec18f8c, query-blockstats didn't return the statistics of
drives without media any more because such drives have only a BB now,
but not a BDS any more.
This patch fixes the regression so that query-blockstats iterates over
BBs by default and empty drives are displayed again.
Signed-off-by: Kevin Wolf <kwolf@redhat.com> Reviewed-by: Max Reitz <mreitz@redhat.com>
Peter Maydell [Mon, 14 Mar 2016 15:11:39 +0000 (15:11 +0000)]
Merge remote-tracking branch 'remotes/awilliam/tags/vfio-update-20160311.0' into staging
VFIO updates 2016-03-11
- Allow devices to be specified via sysfs path (Alex Williamson)
- vfio region helpers and generalization for future device specific regions
(Alex Williamson)
- Automatic ROM device ID and checksum fixup (Alex Williamson)
- Split VGA setup to allow enabling VGA from quirks (Alex Williamson)
- Remove fixed string limit for ROM MemoryRegion name (Neo Jia)
- MAINTAINERS update (Thomas Huth)
# gpg: Signature made Fri 11 Mar 2016 15:55:31 GMT using RSA key ID 3BB08B22
# gpg: Good signature from "Alex Williamson <alex.williamson@redhat.com>"
# gpg: aka "Alex Williamson <alex@shazbot.org>"
# gpg: aka "Alex Williamson <alwillia@redhat.com>"
# gpg: aka "Alex Williamson <alex.l.williamson@gmail.com>"
* remotes/awilliam/tags/vfio-update-20160311.0:
MAINTAINERS: Add entry for the include/hw/vfio/ folder
vfio/pci: replace fixed string limit by g_strdup_printf
vfio/pci: Split out VGA setup
vfio/pci: Fixup PCI option ROMs
vfio/pci: Convert all MemoryRegion to dynamic alloc and consistent functions
vfio: Generalize region support
vfio: Wrap VFIO_DEVICE_GET_REGION_INFO
vfio: Add sysfsdev property for pci & platform
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Peter Maydell [Mon, 14 Mar 2016 13:51:21 +0000 (13:51 +0000)]
Merge remote-tracking branch 'remotes/amit-migration/tags/migration-for-2.6-7' into staging
migration:
- postcopy is no longer experimental
- fix a use-after-free in postcopy
- fix a compile warning
# gpg: Signature made Fri 11 Mar 2016 12:29:33 GMT using RSA key ID 854083B6
# gpg: Good signature from "Amit Shah <amit@amitshah.net>"
# gpg: aka "Amit Shah <amit@kernel.org>"
# gpg: aka "Amit Shah <amitshah@gmx.net>"
* remotes/amit-migration/tags/migration-for-2.6-7:
postcopy: Remove the x-
postcopy: listen thread is never joined
migration: fix use-after-free in loadvm_postcopy_handle_run_bh
migration: fix warning for source_return_path_thread
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Peter Maydell [Mon, 14 Mar 2016 11:49:32 +0000 (11:49 +0000)]
Merge remote-tracking branch 'remotes/berrange/tags/pull-io-win32-2016-03-11-1' into staging
Merge I/O fixes for win32
# gpg: Signature made Fri 11 Mar 2016 10:03:20 GMT using RSA key ID 15104FDF
# gpg: Good signature from "Daniel P. Berrange <dan@berrange.com>"
# gpg: aka "Daniel P. Berrange <berrange@redhat.com>"
* remotes/berrange/tags/pull-io-win32-2016-03-11-1:
osdep: remove use of socket_error() from all code
osdep: add wrappers for socket functions
char: remove qemu_chr_open_socket_fd method
char: remove socket_try_connect method
char: remove qemu_chr_finish_socket_connection method
io: implement socket watch for win32 using WSAEventSelect+select
io: remove checking of EWOULDBLOCK
io: use qemu_accept to ensure SOCK_CLOEXEC is set
io: introduce qio_channel_create_socket_watch
io: pass HANDLE to g_source_add_poll on Win32
io: fix copy+paste mistake in socket error message
io: assert errors before asserting content in I/O test
io: set correct error object in background reader test thread
io: wait for incoming client in socket test
io: bind to socket before creating QIOChannelSocket
io: initialize sockets in test program
io: use bind() to check for IPv4/6 availability
osdep: fix socket_error() to work with Mingw64
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Peter Maydell [Mon, 14 Mar 2016 11:13:11 +0000 (11:13 +0000)]
Merge remote-tracking branch 'remotes/cohuck/tags/s390x-20160311' into staging
CPU hotplug via cpu-add for s390x, cleanup of the s390x machine
compat code and a bugfix in the s390-ccw bios.
# gpg: Signature made Fri 11 Mar 2016 09:48:02 GMT using RSA key ID C6F02FAF
# gpg: Good signature from "Cornelia Huck <huckc@linux.vnet.ibm.com>"
# gpg: aka "Cornelia Huck <cornelia.huck@de.ibm.com>"
* remotes/cohuck/tags/s390x-20160311:
s390x/cpu: use g_new0
s390x: Introduce S390MachineClass
s390x: Introduce machine definition macros
pc-bios/s390-ccw: fix old bug in ptr increment
s390x/cpu: Allow hotplug of CPUs
s390x/cpu: Add error handling to cpu creation
s390x/cpu: Add CPU property links
s390x/cpu: Tolerate max_cpus
s390x/cpu: Get rid of side effects when creating a vcpu
s390x/cpu: Set initial CPU state in common routine
s390x/cpu: Cleanup init in preparation for hotplug
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
trace: separate MMIO tracepoints from TB-access tracepoints
Memory accesses to code which has previously been translated into a TB show up
in the MMIO path, so that they may invalidate the TB. It's extremely confusing
to mix those in with device MMIOs, so split them into their own tracepoint.
Signed-off-by: Hollis Blanchard <hollis_blanchard@mentor.com> Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 1456949575-1633-2-git-send-email-hollis_blanchard@mentor.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
trace: include CPU index in trace_memory_region_*()
Knowing which CPU performed an action is essential for understanding SMP guest
behavior.
However, cpu_physical_memory_rw() may be executed by a machine init function,
before any VCPUs are running, when there is no CPU running ('current_cpu' is
NULL). In this case, store -1 in the trace record as the CPU index. Trace
analysis tools may need to be aware of this special case.
Signed-off-by: Hollis Blanchard <hollis_blanchard@mentor.com>
Message-id: 1456949575-1633-1-git-send-email-hollis_blanchard@mentor.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Postcopy seems to have survived a cycle with only a few fixes,
and Jiri has the current libvirt wired up and working
( https://www.redhat.com/archives/libvir-list/2016-March/msg00080.html )
so remove the experimental tag.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <1457690016-9070-3-git-send-email-dgilbert@redhat.com> Signed-off-by: Amit Shah <amit.shah@redhat.com>
We don't join the listen thread, it does its own cleanup.
Mark as detached not joinable.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com> Reported-by: Paolo Bonzini <pbonzini@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <1457690016-9070-2-git-send-email-dgilbert@redhat.com> Signed-off-by: Amit Shah <amit.shah@redhat.com>
Denis V. Lunev [Wed, 9 Mar 2016 15:35:08 +0000 (18:35 +0300)]
migration: fix use-after-free in loadvm_postcopy_handle_run_bh
MigrationState is destroyed before we can come into bottom half.
Signed-off-by: Denis V. Lunev <den@openvz.org> Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com> CC: Juan Quintela <quintela@redhat.com> CC: Amit Shah <amit.shah@redhat.com> CC: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <1457537708-8622-1-git-send-email-den@openvz.org> Signed-off-by: Amit Shah <amit.shah@redhat.com>
Peter Xu [Wed, 9 Mar 2016 06:12:12 +0000 (14:12 +0800)]
migration: fix warning for source_return_path_thread
max_len is not necessary, while it brings a warning during compilation
when specify "-Wstack-usage=1000000". Replacing using sizeof().
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <1457503932-31763-1-git-send-email-peterx@redhat.com> Signed-off-by: Amit Shah <amit.shah@redhat.com>
Alex Williamson [Thu, 10 Mar 2016 16:39:08 +0000 (09:39 -0700)]
vfio/pci: Fixup PCI option ROMs
Devices like Intel graphics are known to not only have bad checksums,
but also the wrong device ID. This is not so surprising given that
the video BIOS is typically part of the system firmware image rather
that embedded into the device and needs to support any IGD device
installed into the system.
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Alex Williamson [Thu, 10 Mar 2016 16:39:08 +0000 (09:39 -0700)]
vfio/pci: Convert all MemoryRegion to dynamic alloc and consistent functions
Match common vfio code with setup, exit, and finalize functions for
BAR, quirk, and VGA management. VGA is also changed to dynamic
allocation to match the other MemoryRegions.
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Alex Williamson [Thu, 10 Mar 2016 16:39:07 +0000 (09:39 -0700)]
vfio: Generalize region support
Both platform and PCI vfio drivers create a "slow", I/O memory region
with one or more mmap memory regions overlayed when supported by the
device. Generalize this to a set of common helpers in the core that
pulls the region info from vfio, fills the region data, configures
slow mapping, and adds helpers for comleting the mmap, enable/disable,
and teardown. This can be immediately used by the PCI MSI-X code,
which needs to mmap around the MSI-X vector table.
This also changes VFIORegion.mem to be dynamically allocated because
otherwise we don't know how the caller has allocated VFIORegion and
therefore don't know whether to unreference it to destroy the
MemoryRegion or not.
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Now that QEMU wraps the Win32 sockets methods to automatically
set errno upon failure, there is no reason for callers to use
the socket_error() method. They can rely on accessing errno
even on Win32. Remove all use of socket_error() from general
code, leaving it as a static method in oslib-win32.c only.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
The windows socket functions look identical to the normal POSIX
sockets functions, but instead of setting errno, the caller needs
to call WSAGetLastError(). QEMU has tried to deal with this
incompatibility by defining a socket_error() method that callers
must use that abstracts the difference between WSAGetLastError()
and errno.
This approach is somewhat error prone though - many callers of
the sockets functions are just using errno directly because it
is easy to forget the need use a QEMU specific wrapper. It is
not always immediately obvious that a particular function will
in fact call into Windows sockets functions, so the dev may not
even realize they need to use socket_error().
This introduces an alternative approach to portability inspired
by the way GNULIB fixes portability problems. We use a macro to
redefine the original socket function names to refer to a QEMU
wrapper function. The wrapper function calls the original Win32
sockets method and then sets errno from the WSAGetLastError()
value.
Thus all code can simply call the normal POSIX sockets APIs are
have standard errno reporting on error, even on Windows. This
makes the socket_error() method obsolete.
We also bring closesocket & ioctlsocket into this approach. Even
though they are non-standard Win32 names, we can't wrap the normal
close/ioctl methods since there's no reliable way to distinguish
between a file descriptor and HANDLE in Win32.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
The qemu_chr_open_socket_fd method takes care of either doing a
synchronous socket connect, or creating a listener socket. Part
of the work when creating the listener socket is to register a
watch for incoming clients. The caller of qemu_chr_open_socket_fd
may not want this watch created, as it might be doing a synchronous
wait for the first client. Rather than passing yet more parameters
into qemu_chr_open_socket_fd to let it handle this, just remove
the qemu_chr_open_socket_fd method an inline its functionality
into the caller. This allows for a clearer control flow and shorter
code.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
The qemu_chr_open_socket_fd() method multiplexes three different
actions into one method. The socket_try_connect() method is one
of its callers, but it only ever want one specific action
performed. By inlining that action into socket_try_connect()
we see that there is not in fact any failure scenario, so there
is not even any reason for socket_try_connect to exist. Just
inline the asynchronous connection attempts directly at the
places that need them. This shortens & clarifies the code.
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
The qemu_chr_finish_socket_connection method is multiplexing two
different actions into one method. Each caller of it though, only
wants one specific action. The code is shorter & clearer if we
thus remove the method and just inline the specific actions
where needed.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Paolo Bonzini [Mon, 7 Mar 2016 11:12:36 +0000 (12:12 +0100)]
io: implement socket watch for win32 using WSAEventSelect+select
On Win32 we cannot directly poll on socket handles. Instead we
create a Win32 event object and associate the socket handle with
the event. When the event signals readyness we then have to
use select to determine which events are ready. Creating Win32
events is moderately heavyweight, so we don't want todo it
every time we create a GSource, so this associates a single
event with a QIOChannel.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Paolo Bonzini [Mon, 7 Mar 2016 10:16:39 +0000 (11:16 +0100)]
io: introduce qio_channel_create_socket_watch
Sockets are not in the same namespace as file descriptors on Windows.
As an initial step, introduce separate APIs for file descriptor and
socket watches.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
io: assert errors before asserting content in I/O test
When checking the results of an I/O operation test, assert that
the error objects are NULL before asserting on the content. This
is found to give more useful indication of the problem when
diagnosing test failures.
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
io: set correct error object in background reader test thread
The reader thread was accidentally setting the error pointer
intended for the writer thread. If both threads set errors
this would result in QEMU abort'ing due to the error already
being set.
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
io: bind to socket before creating QIOChannelSocket
In the QIOChannelSocket test we create a socket file
descriptor and then try to create a QIOChannelSocket.
This works on Linux, but fails on Win32 because it is
not valid to call getsockname() on an unbound socket.
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Currently the test-io-channel-socket.c test uses getifaddrs
to see if an IPv4/6 address is present on any host NIC, as
a way to determine if IPv4/6 sockets can be used. This is
problematic because getifaddrs is not available on Win32.
Rather than testing indirectly via getifaddrs, just create
a socket and try to bind() to the loopback address instead.
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Historically QEMU has had a socket_error() macro that was
defined to map to WSASocketError(). The os-win32.h header
file would define errno constants that mapped to the
WSA error constants. This worked fine with Mingw32 since
its header files never defined any errno values, nor did
it even provide an errno.h. So callers of socket_error()
could match on traditional Exxxx constants and it would
all "just work".
With Mingw64 though, things work rather differently. First
there is an errno.h file which defines all the traditional
errno constants you'd expect from a UNIX platform. There
is then a winerror.h which defined the WSA error constants.
Crucially the WSAExxxx errno values in winerror.h do not
match the Exxxx errno values in error.h.
If QEMU had only imported winerror.h it would still work,
but the qemu/osdep.h file unconditionally imports errno.h.
So callers of socket_error() will get now WSAExxxx values
back and compare them to the Exxx constants. This will
always fail silently at runtime.
To solve this QEMU needs to stop assuming the WSAExxxx
constant values match the Exxx constant values. Thus the
socket_error() macro is turned into a small function that
re-maps WSAExxxx values into Exxx.
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Alex Williamson [Thu, 10 Mar 2016 16:39:07 +0000 (09:39 -0700)]
vfio: Add sysfsdev property for pci & platform
vfio-pci currently requires a host= parameter, which comes in the
form of a PCI address in [domain:]<bus:slot.function> notation. We
expect to find a matching entry in sysfs for that under
/sys/bus/pci/devices/. vfio-platform takes a similar approach, but
defines the host= parameter to be a string, which can be matched
directly under /sys/bus/platform/devices/. On the PCI side, we have
some interest in using vfio to expose vGPU devices. These are not
actual discrete PCI devices, so they don't have a compatible host PCI
bus address or a device link where QEMU wants to look for it. There's
also really no requirement that vfio can only be used to expose
physical devices, a new vfio bus and iommu driver could expose a
completely emulated device. To fit within the vfio framework, it
would need a kernel struct device and associated IOMMU group, but
those are easy constraints to manage.
To support such devices, which would include vGPUs, that honor the
VFIO PCI programming API, but are not necessarily backed by a unique
PCI address, add support for specifying any device in sysfs. The
vfio API already has support for probing the device type to ensure
compatibility with either vfio-pci or vfio-platform.
With this, a vfio-pci device could either be specified as:
The same change is made for vfio-platform, specifying sysfsdev has
precedence over the old host option.
Tested-by: Eric Auger <eric.auger@linaro.org> Reviewed-by: Eric Auger <eric.auger@linaro.org> Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Janosch Frank [Mon, 7 Mar 2016 08:04:21 +0000 (09:04 +0100)]
s390x: Introduce S390MachineClass
As we now have the new machine definitions, that let us disable/enable
machine options more easily, we need a way to save them and make them
publicly available.
The new s390-virtio-ccw.h header exports the s390 ccw machine state
and class, so they can be easily used in other C files.
Janosch Frank [Thu, 3 Mar 2016 11:48:34 +0000 (12:48 +0100)]
s390x: Introduce machine definition macros
Most of the machine definition code looks the same between different
machine versions. The new DEFINE_CCW_MACHINE macro makes defining a
new machine easier by inserting standard machine version
definitions. This also makes it possible to propagate values between
machine versions.
Matthew Rosato [Fri, 4 Mar 2016 17:34:32 +0000 (12:34 -0500)]
s390x/cpu: Tolerate max_cpus
Once hotplug is enabled, interrupts may come in for CPUs
with an address > smp_cpus. Allocate for this and allow
search routines to look beyond smp_cpus.
Signed-off-by: Matthew Rosato <mjrosato@linux.vnet.ibm.com>
Message-Id: <1457112875-5209-5-git-send-email-mjrosato@linux.vnet.ibm.com> Reviewed-by: David Hildenbrand <dahi@linux.vnet.ibm.com> Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Matthew Rosato [Fri, 4 Mar 2016 17:34:29 +0000 (12:34 -0500)]
s390x/cpu: Cleanup init in preparation for hotplug
Ensure a valid cpu_model is set upfront by setting the
default value directly into the MachineState when none is
specified. This is needed to ensure hotplugged CPUs share
the same cpu_model.
Signed-off-by: Matthew Rosato <mjrosato@linux.vnet.ibm.com> Reviewed-by: David Hildenbrand <dahi@linux.vnet.ibm.com>
Message-Id: <1457112875-5209-2-git-send-email-mjrosato@linux.vnet.ibm.com> Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Peter Maydell [Thu, 10 Mar 2016 02:51:14 +0000 (02:51 +0000)]
Merge remote-tracking branch 'remotes/kraxel/tags/pull-ui-20160309-1' into staging
add linux evdev support, vnc and console fixes.
# gpg: Signature made Wed 09 Mar 2016 09:02:47 GMT using RSA key ID D3E87138
# gpg: Good signature from "Gerd Hoffmann (work) <kraxel@redhat.com>"
# gpg: aka "Gerd Hoffmann <gerd@kraxel.org>"
# gpg: aka "Gerd Hoffmann (private) <kraxel@gmail.com>"
* remotes/kraxel/tags/pull-ui-20160309-1:
ui/console: add escape sequence \e[5, 6n
input-linux: add switch to enable auto-repeat events
input-linux: add option to toggle grab on all devices
input: linux evdev support
vnc: send cursor when a new client is connecting
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Ren Kimura [Tue, 8 Mar 2016 19:51:21 +0000 (04:51 +0900)]
ui/console: add escape sequence \e[5, 6n
Add support of escape sequence "\e[5n" and "\e[6n" to console.
"\e[5n" reports status of console and it always succeed
in virtual console.
"\e[6n" reports now cursor position in console.
projects / thirdparty / qemu.git / log
