qemu.git
4 years agoMerge remote-tracking branch 'remotes/mst/tags/for_upstream' into staging
Peter Maydell [Thu, 5 Nov 2020 15:16:43 +0000 (15:16 +0000)]
Merge remote-tracking branch 'remotes/mst/tags/for_upstream' into staging

pc,pci,vhost,virtio: fixes

Lots of fixes all over the place.
virtio-mem and virtio-iommu patches are kind of fixes but
it seems better to just make them behave sanely than
try to educate users about the limitations ...

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
# gpg: Signature made Wed 04 Nov 2020 18:40:03 GMT
# gpg:                using RSA key 5D09FD0871C8F85B94CA8A0D281F0DB8D28D5469
# gpg:                issuer "mst@redhat.com"
# gpg: Good signature from "Michael S. Tsirkin <mst@kernel.org>" [full]
# gpg:                 aka "Michael S. Tsirkin <mst@redhat.com>" [full]
# Primary key fingerprint: 0270 606B 6F3C DF3D 0B17  0970 C350 3912 AFBE 8E67
#      Subkey fingerprint: 5D09 FD08 71C8 F85B 94CA  8A0D 281F 0DB8 D28D 5469

* remotes/mst/tags/for_upstream: (31 commits)
  contrib/vhost-user-blk: fix get_config() information leak
  block/export: fix vhost-user-blk get_config() information leak
  block/export: make vhost-user-blk config space little-endian
  configure: introduce --enable-vhost-user-blk-server
  libvhost-user: follow QEMU comment style
  vhost-blk: set features before setting inflight feature
  Revert "vhost-blk: set features before setting inflight feature"
  net: Add vhost-vdpa in show_netdevs()
  vhost-vdpa: Add qemu_close in vhost_vdpa_cleanup
  vfio: Don't issue full 2^64 unmap
  virtio-iommu: Set supported page size mask
  vfio: Set IOMMU page size as per host supported page size
  memory: Add interface to set iommu page size mask
  virtio-iommu: Add notify_flag_changed() memory region callback
  virtio-iommu: Add replay() memory region callback
  virtio-iommu: Call memory notifiers in attach/detach
  virtio-iommu: Add memory notifiers for map/unmap
  virtio-iommu: Store memory region in endpoint struct
  virtio-iommu: Fix virtio_iommu_mr()
  hw/smbios: Fix leaked fd in save_opt_one() error path
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
4 years agoMerge remote-tracking branch 'remotes/bonzini-gitlab/tags/for-upstream' into staging
Peter Maydell [Thu, 5 Nov 2020 13:30:05 +0000 (13:30 +0000)]
Merge remote-tracking branch 'remotes/bonzini-gitlab/tags/for-upstream' into staging

Doc and bug fixes

# gpg: Signature made Wed 04 Nov 2020 17:01:29 GMT
# gpg:                using RSA key F13338574B662389866C7682BFFBD25F78C7AE83
# gpg:                issuer "pbonzini@redhat.com"
# gpg: Good signature from "Paolo Bonzini <bonzini@gnu.org>" [full]
# gpg:                 aka "Paolo Bonzini <pbonzini@redhat.com>" [full]
# Primary key fingerprint: 46F5 9FBD 57D6 12E7 BFD4  E2F7 7E15 100C CD36 69B1
#      Subkey fingerprint: F133 3857 4B66 2389 866C  7682 BFFB D25F 78C7 AE83

* remotes/bonzini-gitlab/tags/for-upstream:
  qapi, qemu-options: make all parsing visitors parse boolean options the same
  qtest: escape device name in device-introspect-test
  ivshmem-test: do not use short-form boolean option
  semihosting: fix order of initialization functions
  fuzz: fuzz offsets within pio/mmio regions
  fuzz: check the MR in the DMA callback
  fuzz: fix writing DMA patterns
  tests/qtest: Fix potential NULL pointer dereference in qos_build_main_args()
  configure: fix gio_libs reference
  meson: fix warning for bad sphinx-build
  tests/qtest/libqos/ahci.c: Avoid NULL dereference in ahci_exec()
  tests/qtest/libqtest.c: Check for setsockopt() failure
  meson: vhost-user-gpu/virtiofsd: use absolute path
  meson: use b_staticpic=false for meson >=0.56.0
  qtest: add a reproducer for LP#1878642
  hw/isa/lpc_ich9: Ignore reserved/invalid SCI IRQ
  scripts/oss-fuzz: rename bin/qemu-fuzz-i386
  exec: Remove dead code (CID 1432876)
  docs: expand sourceset documentation
  cutils: replace strdup with g_strdup

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
4 years agoMerge remote-tracking branch 'remotes/kraxel/tags/fixes-20201104-pull-request' into...
Peter Maydell [Wed, 4 Nov 2020 22:13:02 +0000 (22:13 +0000)]
Merge remote-tracking branch 'remotes/kraxel/tags/fixes-20201104-pull-request' into staging

misc bugfixes for 5.2

# gpg: Signature made Wed 04 Nov 2020 15:46:33 GMT
# gpg:                using RSA key 4CB6D8EED3E87138
# gpg: Good signature from "Gerd Hoffmann (work) <kraxel@redhat.com>" [full]
# gpg:                 aka "Gerd Hoffmann <gerd@kraxel.org>" [full]
# gpg:                 aka "Gerd Hoffmann (private) <kraxel@gmail.com>" [full]
# Primary key fingerprint: A032 8CFF B93A 17A7 9901  FE7D 4CB6 D8EE D3E8 7138

* remotes/kraxel/tags/fixes-20201104-pull-request:
  roms/Makefile: Add qboot to .PHONY list
  ati: check x y display parameter values
  vnc: fix resource leak when websocket channel error

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
4 years agoqapi, qemu-options: make all parsing visitors parse boolean options the same
Paolo Bonzini [Tue, 3 Nov 2020 16:13:39 +0000 (11:13 -0500)]
qapi, qemu-options: make all parsing visitors parse boolean options the same

OptsVisitor, StringInputVisitor and the keyval visitor have
three different ideas of how a human could write the value of
a boolean option.  Pay homage to the backwards-compatibility
gods and make the new common helper accept all four sets (on/off,
true/false, y/n and yes/no), but remove case-insensitivity.

Since OptsVisitor is supposed to match qemu-options, adjust
it as well.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20201103161339.447118-1-pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoqtest: escape device name in device-introspect-test
Paolo Bonzini [Tue, 3 Nov 2020 11:57:46 +0000 (06:57 -0500)]
qtest: escape device name in device-introspect-test

device-introspect-test uses HMP, so it should escape the device name
properly.  Because of this, a few devices that had commas in their
names were escaping testing.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoivshmem-test: do not use short-form boolean option
Paolo Bonzini [Tue, 3 Nov 2020 12:37:46 +0000 (07:37 -0500)]
ivshmem-test: do not use short-form boolean option

This QemuOpts idiom will be deprecated, so get rid of it in the tests.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoMerge remote-tracking branch 'remotes/kraxel/tags/ui-20201104-pull-request' into...
Peter Maydell [Wed, 4 Nov 2020 16:52:17 +0000 (16:52 +0000)]
Merge remote-tracking branch 'remotes/kraxel/tags/ui-20201104-pull-request' into staging

ui: run screendump in coroutine

# gpg: Signature made Wed 04 Nov 2020 13:53:50 GMT
# gpg:                using RSA key 4CB6D8EED3E87138
# gpg: Good signature from "Gerd Hoffmann (work) <kraxel@redhat.com>" [full]
# gpg:                 aka "Gerd Hoffmann <gerd@kraxel.org>" [full]
# gpg:                 aka "Gerd Hoffmann (private) <kraxel@gmail.com>" [full]
# Primary key fingerprint: A032 8CFF B93A 17A7 9901  FE7D 4CB6 D8EE D3E8 7138

* remotes/kraxel/tags/ui-20201104-pull-request:
  console: make QMP/HMP screendump run in coroutine
  console: modify ppm_save to take a pixman image ref
  coroutine: let CoQueue wake up outside a coroutine

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
4 years agoMerge remote-tracking branch 'remotes/kraxel/tags/usb-20201104-pull-request' into...
Peter Maydell [Wed, 4 Nov 2020 14:13:07 +0000 (14:13 +0000)]
Merge remote-tracking branch 'remotes/kraxel/tags/usb-20201104-pull-request' into staging

usb: bugfixes for usb-serial

# gpg: Signature made Wed 04 Nov 2020 12:13:05 GMT
# gpg:                using RSA key 4CB6D8EED3E87138
# gpg: Good signature from "Gerd Hoffmann (work) <kraxel@redhat.com>" [full]
# gpg:                 aka "Gerd Hoffmann <gerd@kraxel.org>" [full]
# gpg:                 aka "Gerd Hoffmann (private) <kraxel@gmail.com>" [full]
# Primary key fingerprint: A032 8CFF B93A 17A7 9901  FE7D 4CB6 D8EE D3E8 7138

* remotes/kraxel/tags/usb-20201104-pull-request:
  dev-serial: store flow control and xon/xoff characters
  dev-serial: add support for setting data_bits in QEMUSerialSetParams
  dev-serial: add always-plugged property to ensure USB device is always attached
  dev-serial: replace DeviceOutVendor/DeviceInVendor with equivalent macros from usb.h
  dev-serial: add trace-events for baud rate and data parameters
  dev-serial: convert from DPRINTF to trace-events
  dev-serial: use USB_SERIAL QOM macro for USBSerialState assignments
  dev-serial: style changes to improve readability and checkpatch fixes

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
4 years agoMerge remote-tracking branch 'remotes/philmd-gitlab/tags/mips-fixes-20201103' into...
Peter Maydell [Wed, 4 Nov 2020 10:28:31 +0000 (10:28 +0000)]
Merge remote-tracking branch 'remotes/philmd-gitlab/tags/mips-fixes-20201103' into staging

MIPS patches queue

- Removal of the 'r4k' machine (deprecated before 5.0)
- Fix LGPL license text (Chetan Pant)
- Support unaligned accesses on Loongson-3 (Huacai Chen)
- Fix out-of-bound access in Loongson-3 embedded I/O interrupt
  controller (Alex Chen)

CI jobs results:
. https://cirrus-ci.com/build/6324890389184512
. https://gitlab.com/philmd/qemu/-/pipelines/211275262
. https://travis-ci.org/github/philmd/qemu/builds/741188958

# gpg: Signature made Tue 03 Nov 2020 17:30:30 GMT
# gpg:                using RSA key FAABE75E12917221DCFD6BB2E3E32C2CDEADC0DE
# gpg: Good signature from "Philippe Mathieu-Daudé (F4BUG) <f4bug@amsat.org>" [full]
# Primary key fingerprint: FAAB E75E 1291 7221 DCFD  6BB2 E3E3 2C2C DEAD C0DE

* remotes/philmd-gitlab/tags/mips-fixes-20201103:
  target/mips: Add unaligned access support for MIPS64R6 and Loongson-3
  target/mips: Fix Lesser GPL version number
  hw/intc/loongson: Fix incorrect 'core' calculation in liointc_read/write
  hw/mips/boston: Fix Lesser GPL version number
  hw/mips: Fix Lesser GPL version number
  hw/mips: Remove the 'r4k' machine

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
4 years agoroms/Makefile: Add qboot to .PHONY list
Bruce Rogers [Tue, 20 Oct 2020 15:25:12 +0000 (09:25 -0600)]
roms/Makefile: Add qboot to .PHONY list

Adding qboot to the .PHONY directive will allow a
make -C roms qboot invocation to work as expected

Signed-off-by: Bruce Rogers <brogers@suse.com>
Message-id: 20201020152512.837769-1-brogers@suse.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
4 years agoati: check x y display parameter values
Prasad J Pandit [Wed, 21 Oct 2020 10:38:18 +0000 (16:08 +0530)]
ati: check x y display parameter values

The source and destination x,y display parameters in ati_2d_blt()
may run off the vga limits if either of s->regs.[src|dst]_[xy] is
zero. Check the parameter values to avoid potential crash.

Reported-by: Gaoning Pan <pgn@zju.edu.cn>
Signed-off-by: Prasad J Pandit <pjp@fedoraproject.org>
Message-id: 20201021103818.1704030-1-ppandit@redhat.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
4 years agovnc: fix resource leak when websocket channel error
Ding Hui [Thu, 29 Oct 2020 03:22:41 +0000 (11:22 +0800)]
vnc: fix resource leak when websocket channel error

When we connect to vnc by websocket channel, and disconnect
(maybe by some network exception) before handshake,
qemu will left CLOSE_WAIT socket and never close it

After 04d2529da2 ("ui: convert VNC server to use QIOChannelSocket")
and dd154c4d9f ("io: fix handling of EOF / error conditions in websock GSource"),
the vnc call qio_channel_add_watch only care about G_IO_IN,
but mising G_IO_HUP and G_IO_ERR.
When the websocket channel get EOF or error, it cannot callback,
because the caller ignore the event, that leads to resource leak

We need handle G_IO_HUP and G_IO_ERR event, then cleanup the channel

Fixes: 04d2529da2 ("ui: convert VNC server to use QIOChannelSocket")
Fixes: dd154c4d9f ("io: fix handling of EOF / error conditions in websock GSource")
Cc: qemu-stable@nongnu.org
Signed-off-by: Ding Hui <dinghui@sangfor.com.cn>
Message-id: 20201029032241.11040-1-dinghui@sangfor.com.cn
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
4 years agoconsole: make QMP/HMP screendump run in coroutine
Marc-André Lureau [Tue, 27 Oct 2020 13:36:02 +0000 (17:36 +0400)]
console: make QMP/HMP screendump run in coroutine

Thanks to the monitors' coroutine support (merge commit b7092cda1b3),
the screendump handler can trigger a graphic_hw_update(), yield and let
the main loop run until update is done. Then the handler is resumed, and
ppm_save() will write the screen image to disk in the coroutine context.

The IO is still blocking though, as the file is set blocking so far,
this could be addressed by some future change (with other caveats).

Related to:
https://bugzilla.redhat.com/show_bug.cgi?id=1230527

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Message-id: 20201027133602.3038018-4-marcandre.lureau@redhat.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
4 years agoconsole: modify ppm_save to take a pixman image ref
Marc-André Lureau [Tue, 27 Oct 2020 13:36:01 +0000 (17:36 +0400)]
console: modify ppm_save to take a pixman image ref

The function is going to be called from a coroutine, and may yield.
Let's ensure our image reference doesn't change over time (due to resize
etc) by keeping a ref.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Gerd Hoffmann <kraxel@redhat.com>
Message-id: 20201027133602.3038018-3-marcandre.lureau@redhat.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
4 years agocoroutine: let CoQueue wake up outside a coroutine
Marc-André Lureau [Tue, 27 Oct 2020 13:36:00 +0000 (17:36 +0400)]
coroutine: let CoQueue wake up outside a coroutine

The assert() was added in commit b681a1c73e15 ("block: Repair the
throttling code."), when the qemu_co_queue_do_restart() function
required to be running in a coroutine. It was later made unnecessary in
commit a9d9235567e7 ("coroutine-lock: reschedule coroutine on the
AioContext it was running on").

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Gerd Hoffmann <kraxel@redhat.com>
Message-id: 20201027133602.3038018-2-marcandre.lureau@redhat.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
4 years agodev-serial: store flow control and xon/xoff characters
Mark Cave-Ayland [Tue, 27 Oct 2020 15:04:56 +0000 (15:04 +0000)]
dev-serial: store flow control and xon/xoff characters

Note that whilst the device does not do anything with these values, they are
logged with trace events and stored to allow future implementation.

The default flow control is set to none at reset as documented in the Linux
ftdi_sio.h header file.

Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: Samuel Thibault <samuel.thibault@ens-lyon.org>
Message-id: 20201027150456.24606-9-mark.cave-ayland@ilande.co.uk
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
4 years agodev-serial: add support for setting data_bits in QEMUSerialSetParams
Mark Cave-Ayland [Tue, 27 Oct 2020 15:04:55 +0000 (15:04 +0000)]
dev-serial: add support for setting data_bits in QEMUSerialSetParams

Also implement the behaviour reported in Linux's ftdi_sio.c whereby if an invalid
data_bits value is provided then the hardware defaults to using 8.

Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: Samuel Thibault <samuel.thibault@ens-lyon.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-id: 20201027150456.24606-8-mark.cave-ayland@ilande.co.uk
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
4 years agodev-serial: add always-plugged property to ensure USB device is always attached
Mark Cave-Ayland [Tue, 27 Oct 2020 15:04:54 +0000 (15:04 +0000)]
dev-serial: add always-plugged property to ensure USB device is always attached

Some operating systems will generate a new device ID when a USB device is unplugged
and then replugged into the USB. If this is done whilst switching between multiple
applications over a virtual serial port, the change of device ID requires going
back into the OS/application to locate the new device accordingly.

Add a new always-plugged property that if specified will ensure that the device
always remains attached to the USB regardless of the state of the backend
chardev.

Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: Samuel Thibault <samuel.thibault@ens-lyon.org>
Message-id: 20201027150456.24606-7-mark.cave-ayland@ilande.co.uk
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
4 years agodev-serial: replace DeviceOutVendor/DeviceInVendor with equivalent macros from usb.h
Mark Cave-Ayland [Tue, 27 Oct 2020 15:04:53 +0000 (15:04 +0000)]
dev-serial: replace DeviceOutVendor/DeviceInVendor with equivalent macros from usb.h

The DeviceOutVendor and DeviceInVendor macros can be replaced with their
equivalent VendorDeviceOutRequest and VendorDeviceRequest macros from usb.h.

Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: Samuel Thibault <samuel.thibault@ens-lyon.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-id: 20201027150456.24606-6-mark.cave-ayland@ilande.co.uk
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
4 years agodev-serial: add trace-events for baud rate and data parameters
Mark Cave-Ayland [Tue, 27 Oct 2020 15:04:52 +0000 (15:04 +0000)]
dev-serial: add trace-events for baud rate and data parameters

Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: Samuel Thibault <samuel.thibault@ens-lyon.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-id: 20201027150456.24606-5-mark.cave-ayland@ilande.co.uk
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
4 years agodev-serial: convert from DPRINTF to trace-events
Mark Cave-Ayland [Tue, 27 Oct 2020 15:04:51 +0000 (15:04 +0000)]
dev-serial: convert from DPRINTF to trace-events

Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: Samuel Thibault <samuel.thibault@ens-lyon.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-id: 20201027150456.24606-4-mark.cave-ayland@ilande.co.uk
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
4 years agodev-serial: use USB_SERIAL QOM macro for USBSerialState assignments
Mark Cave-Ayland [Tue, 27 Oct 2020 15:04:50 +0000 (15:04 +0000)]
dev-serial: use USB_SERIAL QOM macro for USBSerialState assignments

Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: Samuel Thibault <samuel.thibault@ens-lyon.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-id: 20201027150456.24606-3-mark.cave-ayland@ilande.co.uk
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
4 years agodev-serial: style changes to improve readability and checkpatch fixes
Mark Cave-Ayland [Tue, 27 Oct 2020 15:04:49 +0000 (15:04 +0000)]
dev-serial: style changes to improve readability and checkpatch fixes

Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Reviewed-by: Samuel Thibault <samuel.thibault@ens-lyon.org>
Message-id: 20201027150456.24606-2-mark.cave-ayland@ilande.co.uk
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
4 years agocontrib/vhost-user-blk: fix get_config() information leak
Stefan Hajnoczi [Tue, 27 Oct 2020 17:35:21 +0000 (17:35 +0000)]
contrib/vhost-user-blk: fix get_config() information leak

Refuse get_config() in excess of sizeof(struct virtio_blk_config).

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20201027173528.213464-6-stefanha@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
4 years agoblock/export: fix vhost-user-blk get_config() information leak
Stefan Hajnoczi [Tue, 27 Oct 2020 17:35:20 +0000 (17:35 +0000)]
block/export: fix vhost-user-blk get_config() information leak

Refuse get_config() requests in excess of sizeof(struct virtio_blk_config).

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20201027173528.213464-5-stefanha@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
4 years agoblock/export: make vhost-user-blk config space little-endian
Stefan Hajnoczi [Tue, 27 Oct 2020 17:35:19 +0000 (17:35 +0000)]
block/export: make vhost-user-blk config space little-endian

VIRTIO 1.0 devices have little-endian configuration space. The
vhost-user-blk-server.c code already uses little-endian for virtqueue
processing but not for the configuration space fields. Fix this so the
vhost-user-blk export works on big-endian hosts.

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20201027173528.213464-4-stefanha@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
4 years agoconfigure: introduce --enable-vhost-user-blk-server
Stefan Hajnoczi [Tue, 27 Oct 2020 17:35:18 +0000 (17:35 +0000)]
configure: introduce --enable-vhost-user-blk-server

Make it possible to compile out the vhost-user-blk server. It is enabled
by default on Linux.

Note that vhost-user-server.c depends on libvhost-user, which requires
CONFIG_LINUX. The CONFIG_VHOST_USER dependency was erroneous since that
option controls vhost-user frontends (previously known as "master") and
not device backends (previously known as "slave").

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20201027173528.213464-3-stefanha@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
4 years agolibvhost-user: follow QEMU comment style
Stefan Hajnoczi [Tue, 27 Oct 2020 17:35:17 +0000 (17:35 +0000)]
libvhost-user: follow QEMU comment style

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20201027173528.213464-2-stefanha@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
4 years agovhost-blk: set features before setting inflight feature
Jin Yu [Tue, 3 Nov 2020 12:36:17 +0000 (20:36 +0800)]
vhost-blk: set features before setting inflight feature

Virtqueue has split and packed, so before setting inflight,
you need to inform the back-end virtqueue format.

Signed-off-by: Jin Yu <jin.yu@intel.com>
Acked-by: Raphael Norwitz <raphael.norwitz@nutanix.com>
Message-Id: <20201103123617.28256-1-jin.yu@intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
4 years agoRevert "vhost-blk: set features before setting inflight feature"
Stefan Hajnoczi [Mon, 2 Nov 2020 16:57:09 +0000 (16:57 +0000)]
Revert "vhost-blk: set features before setting inflight feature"

This reverts commit adb29c027341ba095a3ef4beef6aaef86d3a520e.

The commit broke -device vhost-user-blk-pci because the
vhost_dev_prepare_inflight() function it introduced segfaults in
vhost_dev_set_features() when attempting to access struct vhost_dev's
vdev pointer before it has been assigned.

To reproduce the segfault simply launch a vhost-user-blk device with the
contrib vhost-user-blk device backend:

  $ build/contrib/vhost-user-blk/vhost-user-blk -s /tmp/vhost-user-blk.sock -r -b /var/tmp/foo.img
  $ build/qemu-system-x86_64 \
        -device vhost-user-blk-pci,id=drv0,chardev=char1,addr=4.0 \
        -object memory-backend-memfd,id=mem,size=1G,share=on \
        -M memory-backend=mem,accel=kvm \
        -chardev socket,id=char1,path=/tmp/vhost-user-blk.sock
  Segmentation fault (core dumped)

Cc: Jin Yu <jin.yu@intel.com>
Cc: Raphael Norwitz <raphael.norwitz@nutanix.com>
Cc: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20201102165709.232180-1-stefanha@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
4 years agonet: Add vhost-vdpa in show_netdevs()
Cindy Lu [Fri, 16 Oct 2020 03:09:09 +0000 (11:09 +0800)]
net: Add vhost-vdpa in show_netdevs()

Fix the bug that while Check qemu supported netdev,
there is no vhost-vdpa

Signed-off-by: Cindy Lu <lulu@redhat.com>
Message-Id: <20201016030909.9522-2-lulu@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
4 years agovhost-vdpa: Add qemu_close in vhost_vdpa_cleanup
Cindy Lu [Fri, 16 Oct 2020 03:09:08 +0000 (11:09 +0800)]
vhost-vdpa: Add qemu_close in vhost_vdpa_cleanup

fix the bug that fd will still open after the cleanup

Signed-off-by: Cindy Lu <lulu@redhat.com>
Message-Id: <20201016030909.9522-1-lulu@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
4 years agovfio: Don't issue full 2^64 unmap
Jean-Philippe Brucker [Fri, 30 Oct 2020 18:05:10 +0000 (19:05 +0100)]
vfio: Don't issue full 2^64 unmap

IOMMUs may declare memory regions spanning from 0 to UINT64_MAX. When
attempting to deal with such region, vfio_listener_region_del() passes a
size of 2^64 to int128_get64() which throws an assertion failure.  Even
ignoring this, the VFIO_IOMMU_DMA_MAP ioctl cannot handle this size
since the size field is 64-bit. Split the request in two.

Acked-by: Alex Williamson <alex.williamson@redhat.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Jean-Philippe Brucker <jean-philippe@linaro.org>
Message-Id: <20201030180510.747225-11-jean-philippe@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
4 years agoUpdate version for v5.2.0-rc0 release
Peter Maydell [Tue, 3 Nov 2020 21:11:57 +0000 (21:11 +0000)]
Update version for v5.2.0-rc0 release

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
4 years agosemihosting: fix order of initialization functions
Paolo Bonzini [Wed, 21 Oct 2020 20:08:53 +0000 (16:08 -0400)]
semihosting: fix order of initialization functions

qemu_semihosting_console_init uses semihosting.chardev which is set
by qemu_semihosting_connect_chardevs.  Thus qemu_semihosting_connect_chardevs
has to be called first.

Both have to be called after processing -serial and friends though, so
that the semihosting console can connect to a multiplexer as in
"-serial mon:stdio -semihosting-config chardev=serial0"

Suggested-by: Alex Bennée <alex.bennee@linaro.org>
Fixes: 619985e937 ("semihosting: defer connect_chardevs a little more to use serialx", 2020-07-27)
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agofuzz: fuzz offsets within pio/mmio regions
Alexander Bulekov [Thu, 29 Oct 2020 17:29:00 +0000 (13:29 -0400)]
fuzz: fuzz offsets within pio/mmio regions

The code did not add offsets to FlatRange bases, so we did not fuzz
offsets within device MemoryRegions.

Signed-off-by: Alexander Bulekov <alxndr@bu.edu>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Message-Id: <20201029172901.534442-4-alxndr@bu.edu>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agofuzz: check the MR in the DMA callback
Alexander Bulekov [Thu, 29 Oct 2020 17:28:59 +0000 (13:28 -0400)]
fuzz: check the MR in the DMA callback

We should be checking that the device is trying to read from RAM, before
filling the region with data. Otherwise, we will try to populate
nonsensical addresses in RAM for callbacks on PIO/MMIO reads. We did
this originally, however the final version I sent had the line commented
out..

Signed-off-by: Alexander Bulekov <alxndr@bu.edu>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Message-Id: <20201029172901.534442-3-alxndr@bu.edu>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agofuzz: fix writing DMA patterns
Alexander Bulekov [Thu, 29 Oct 2020 17:28:58 +0000 (13:28 -0400)]
fuzz: fix writing DMA patterns

This code had all sorts of issues. We used a loop similar to
address_space_write_rom, but I did not remove a "break" that only made
sense in the context of the switch statement in the original code. Then,
after the loop, we did a separate qtest_memwrite over the entire DMA
access range, defeating the purpose of the loop. Additionally, we
increment the buf pointer, and then try to g_free() it. Fix these
problems.

Reported-by: OSS-Fuzz (Issue 26725)
Signed-off-by: Alexander Bulekov <alxndr@bu.edu>
Reported-by: OSS-Fuzz (Issue 26691)
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Message-Id: <20201029172901.534442-2-alxndr@bu.edu>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoMerge remote-tracking branch 'remotes/alistair/tags/pull-riscv-to-apply-20201103...
Peter Maydell [Tue, 3 Nov 2020 16:53:20 +0000 (16:53 +0000)]
Merge remote-tracking branch 'remotes/alistair/tags/pull-riscv-to-apply-20201103' into staging

This series adds support for migration to RISC-V QEMU and expands the
Microchip PFSoC to allow unmodified HSS and Linux boots.

# gpg: Signature made Tue 03 Nov 2020 15:19:45 GMT
# gpg:                using RSA key F6C4AC46D4934868D3B8CE8F21E10D29DF977054
# gpg: Good signature from "Alistair Francis <alistair@alistair23.me>" [full]
# Primary key fingerprint: F6C4 AC46 D493 4868 D3B8  CE8F 21E1 0D29 DF97 7054

* remotes/alistair/tags/pull-riscv-to-apply-20201103:
  target/riscv/csr.c : add space before the open parenthesis '('
  hw/riscv: microchip_pfsoc: Hook the I2C1 controller
  hw/riscv: microchip_pfsoc: Correct DDR memory map
  hw/riscv: microchip_pfsoc: Map the reserved memory at address 0
  hw/riscv: microchip_pfsoc: Connect the SYSREG module
  hw/misc: Add Microchip PolarFire SoC SYSREG module support
  hw/riscv: microchip_pfsoc: Connect the IOSCB module
  hw/misc: Add Microchip PolarFire SoC IOSCB module support
  hw/riscv: microchip_pfsoc: Connect DDR memory controller modules
  hw/misc: Add Microchip PolarFire SoC DDR Memory Controller support
  hw/riscv: microchip_pfsoc: Document where to look at the SoC memory maps
  target/riscv: Add sifive_plic vmstate
  target/riscv: Add V extension state description
  target/riscv: Add H extension state description
  target/riscv: Add PMP state description
  target/riscv: Add basic vmstate description of CPU
  target/riscv: Merge m/vsstatus and m/vsstatush into one uint64_t unit
  hw/riscv: virt: Allow passing custom DTB
  hw/riscv: sifive_u: Allow passing custom DTB

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
4 years agoMerge remote-tracking branch 'remotes/kevin/tags/for-upstream' into staging
Peter Maydell [Tue, 3 Nov 2020 15:59:44 +0000 (15:59 +0000)]
Merge remote-tracking branch 'remotes/kevin/tags/for-upstream' into staging

Block layer patches:

- iotests: Fix pylint/mypy warnings with Python 3.9
- qmp: fix aio_poll() assertion failure on Windows
- Some minor fixes

# gpg: Signature made Tue 03 Nov 2020 15:25:01 GMT
# gpg:                using RSA key DC3DEB159A9AF95D3D7456FE7F09B272C88F2FD6
# gpg:                issuer "kwolf@redhat.com"
# gpg: Good signature from "Kevin Wolf <kwolf@redhat.com>" [full]
# Primary key fingerprint: DC3D EB15 9A9A F95D 3D74  56FE 7F09 B272 C88F 2FD6

* remotes/kevin/tags/for-upstream:
  block/vvfat: Fix bad printf format specifiers
  iotests: Use Python 3 style super()
  iotests: Disable unsubscriptable-object in pylint
  iotests.py: Fix type check errors in wait_migration()
  qemu-img convert: Free @sn_opts in all error cases
  qmp: fix aio_poll() assertion failure on Windows

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
4 years agotarget/mips: Add unaligned access support for MIPS64R6 and Loongson-3
Huacai Chen [Wed, 28 Oct 2020 04:18:00 +0000 (12:18 +0800)]
target/mips: Add unaligned access support for MIPS64R6 and Loongson-3

MIPSR6 (not only MIPS32R6) processors support unaligned access in
hardware, so set MO_UNALN in their default_tcg_memop_mask. Btw, new
Loongson-3 (such as Loongson-3A4000) also support unaligned access,
since both old and new Loongson-3 use the same binaries, we can simply
set MO_UNALN for all Loongson-3 processors.

Signed-off-by: Huacai Chen <chenhc@lemote.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <1604053541-27822-3-git-send-email-chenhc@lemote.com>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
4 years agotarget/mips: Fix Lesser GPL version number
Chetan Pant [Fri, 16 Oct 2020 14:35:09 +0000 (14:35 +0000)]
target/mips: Fix Lesser GPL version number

There is no "version 2" of the "Lesser" General Public License.
It is either "GPL version 2.0" or "Lesser GPL version 2.1".
This patch replaces all occurrences of "Lesser GPL version 2" with
"Lesser GPL version 2.1" in comment section.

Signed-off-by: Chetan Pant <chetan4windows@gmail.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20201016143509.26692-1-chetan4windows@gmail.com>
[PMD: Split hw/ vs target/]
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
4 years agohw/intc/loongson: Fix incorrect 'core' calculation in liointc_read/write
AlexChen [Tue, 3 Nov 2020 09:32:01 +0000 (17:32 +0800)]
hw/intc/loongson: Fix incorrect 'core' calculation in liointc_read/write

According to the loongson spec
(http://www.loongson.cn/uploadfile/cpu/3B1500/Loongson_3B1500_cpu_user_1.pdf)
and the macro definition(#define R_PERCORE_ISR(x) (0x40 + 0x8 * x)), we know
that the ISR size per CORE is 8, so here we need to divide
(addr - R_PERCORE_ISR(0)) by 8, not 4.

Reported-by: Euler Robot <euler.robot@huawei.com>
Signed-off-by: Alex Chen <alex.chen@huawei.com>
Reviewed-by: Jiaxun Yang <jiaxun.yang@flygoat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <5FA12391.8090400@huawei.com>
[PMD: Shortened subject]
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
4 years agohw/mips/boston: Fix Lesser GPL version number
Chetan Pant [Fri, 23 Oct 2020 12:26:33 +0000 (12:26 +0000)]
hw/mips/boston: Fix Lesser GPL version number

There is no "version 2" of the "Lesser" General Public License.
It is either "GPL version 2.0" or "Lesser GPL version 2.1".
This patch replaces all occurrences of "Lesser GPL version 2" with
"Lesser GPL version 2.1" in comment section.

Signed-off-by: Chetan Pant <chetan4windows@gmail.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20201023122633.19466-1-chetan4windows@gmail.com>
[PMD: Added hw/mips/ prefix in subject]
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
4 years agohw/mips: Fix Lesser GPL version number
Chetan Pant [Fri, 16 Oct 2020 14:35:09 +0000 (14:35 +0000)]
hw/mips: Fix Lesser GPL version number

There is no "version 2" of the "Lesser" General Public License.
It is either "GPL version 2.0" or "Lesser GPL version 2.1".
This patch replaces all occurrences of "Lesser GPL version 2" with
"Lesser GPL version 2.1" in comment section.

Signed-off-by: Chetan Pant <chetan4windows@gmail.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20201016143509.26692-1-chetan4windows@gmail.com>
[PMD: Split hw/ vs target/]
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
4 years agohw/mips: Remove the 'r4k' machine
Philippe Mathieu-Daudé [Mon, 2 Nov 2020 10:14:33 +0000 (11:14 +0100)]
hw/mips: Remove the 'r4k' machine

We deprecated the support for the 'r4k' machine for the 5.0 release
(commit d32dc61421), which means that our deprecation policy allows
us to drop it in release 5.2. Remove the code.

To repeat the rationale from the deprecation note:
- this virtual machine has no specification
- the Linux kernel dropped support for it 10 years ago

Users are recommended to use the Malta board instead.

Acked-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
ACKed-by: Peter Krempa <pkrempa@redhat.com>
Message-Id: <20201102201311.2220005-1-f4bug@amsat.org>

4 years agoblock/vvfat: Fix bad printf format specifiers
AlexChen [Tue, 3 Nov 2020 09:42:56 +0000 (17:42 +0800)]
block/vvfat: Fix bad printf format specifiers

We should use printf format specifier "%u" instead of "%d" for
argument of type "unsigned int".
In addition, fix two error format problems found by checkpatch.pl:
ERROR: space required after that ',' (ctx:VxV)
+        fprintf(stderr,"%s attributes=0x%02x begin=%u size=%d\n",
                       ^
ERROR: line over 90 characters
+        fprintf(stderr, "%d, %s (%u, %d)\n", i, commit->path ? commit->path : "(null)", commit->param.rename.cluster, commit->action);

Reported-by: Euler Robot <euler.robot@huawei.com>
Signed-off-by: Alex Chen <alex.chen@huawei.com>
Message-Id: <5FA12620.6030705@huawei.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
4 years agoiotests: Use Python 3 style super()
Kevin Wolf [Tue, 27 Oct 2020 16:38:06 +0000 (17:38 +0100)]
iotests: Use Python 3 style super()

pylint complains about the use of super with the current class and
instance as arguments in VM.__init__():

iotests.py:546:8: R1725: Consider using Python 3 style super() without arguments (super-with-arguments)

No reason not to follow the advice and make it happy, so let's do this.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20201027163806.290960-4-kwolf@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
4 years agoiotests: Disable unsubscriptable-object in pylint
Kevin Wolf [Tue, 27 Oct 2020 16:38:05 +0000 (17:38 +0100)]
iotests: Disable unsubscriptable-object in pylint

When run with Python 3.9, pylint incorrectly warns about things like
Optional[foo] because it doesn't recognise Optional as unsubscriptable.
This is a known pylint bug:

    https://github.com/PyCQA/pylint/issues/3882

Just disable this check to get rid of the warnings.

Disabling this shouldn't make us miss any real bug because mypy also
has a similar check ("... is not indexable").

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20201027163806.290960-3-kwolf@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
4 years agoiotests.py: Fix type check errors in wait_migration()
Kevin Wolf [Tue, 27 Oct 2020 16:38:04 +0000 (17:38 +0100)]
iotests.py: Fix type check errors in wait_migration()

Commit 1847a4a8c20 clarified that event_wait() can return None (though
only with timeout=0) and commit f12a282ff47 annotated it as returning
Optional[QMPMessage].

Type checks in wait_migration() fail because of the unexpected optional
return type:

iotests.py:750: error: Value of type variable "Msg" of "log" cannot be "Optional[Dict[str, Any]]"
iotests.py:751: error: Value of type "Optional[Dict[str, Any]]" is not indexable
iotests.py:754: error: Value of type "Optional[Dict[str, Any]]" is not indexable

Fortunately, the non-zero default timeout is used in the event_wait()
call, so we can make mypy happy by just asserting this.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20201027163806.290960-2-kwolf@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
4 years agoqemu-img convert: Free @sn_opts in all error cases
Tuguoyi [Mon, 2 Nov 2020 09:04:57 +0000 (09:04 +0000)]
qemu-img convert: Free @sn_opts in all error cases

@sn_opts is initialized at the beginning, so it should be deleted
after jumping to the lable 'fail_getopt'

Signed-off-by: Guoyi Tu <tu.guoyi@h3c.com>
Message-Id: <6ff1c5d372944494be3932274f75485d@h3c.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
4 years agoqmp: fix aio_poll() assertion failure on Windows
Volker Rümelin [Wed, 21 Oct 2020 06:40:33 +0000 (08:40 +0200)]
qmp: fix aio_poll() assertion failure on Windows

Commit 9ce44e2ce2 "qmp: Move dispatcher to a coroutine" modified
aio_poll() in util/aio-posix.c to avoid an assertion failure. This
change is missing in util/aio-win32.c.

Apply the changes to util/aio-posix.c to util/aio-win32.c too.
This fixes an assertion failure on Windows whenever QEMU exits.

$ ./qemu-system-x86_64.exe -machine pc,accel=tcg -display gtk
**
ERROR:../qemu/util/aio-win32.c:337:aio_poll: assertion failed:
(in_aio_context_home_thread(ctx))
Bail out! ERROR:../qemu/util/aio-win32.c:337:aio_poll: assertion
failed: (in_aio_context_home_thread(ctx))

Fixes: 9ce44e2ce2 ("qmp: Move dispatcher to a coroutine")
Signed-off-by: Volker Rümelin <vr_qemu@t-online.de>
Message-Id: <20201021064033.8600-1-vr_qemu@t-online.de>
Tested-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
4 years agotarget/riscv/csr.c : add space before the open parenthesis '('
Xinhao Zhang [Fri, 30 Oct 2020 00:48:15 +0000 (08:48 +0800)]
target/riscv/csr.c : add space before the open parenthesis '('

Fix code style. Space required before the open parenthesis '('.

Signed-off-by: Xinhao Zhang <zhangxinhao1@huawei.com>
Signed-off-by: Kai Deng <dengkai1@huawei.com>
Reported-by: Euler Robot <euler.robot@huawei.com>
Reviewed-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-id: 20201030004815.4172849-1-zhangxinhao1@huawei.com
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
4 years agohw/riscv: microchip_pfsoc: Hook the I2C1 controller
Bin Meng [Wed, 28 Oct 2020 05:30:10 +0000 (13:30 +0800)]
hw/riscv: microchip_pfsoc: Hook the I2C1 controller

The latest SD card image [1] released by Microchip ships a Linux
kernel with built-in PolarFire SoC I2C driver support. The device
tree file includes the description for the I2C1 node hence kernel
tries to probe the I2C1 device during boot.

It is enough to create an unimplemented device for I2C1 to allow
the kernel to continue booting to the shell.

[1] ftp://ftpsoc.microsemi.com/outgoing/core-image-minimal-dev-icicle-kit-es-sd-20201009141623.rootfs.wic.gz

Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-id: 1603863010-15807-11-git-send-email-bmeng.cn@gmail.com
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
4 years agohw/riscv: microchip_pfsoc: Correct DDR memory map
Bin Meng [Sun, 1 Nov 2020 17:05:38 +0000 (01:05 +0800)]
hw/riscv: microchip_pfsoc: Correct DDR memory map

When system memory is larger than 1 GiB (high memory), PolarFire SoC
maps it at address 0x10_0000_0000. Address 0xC000_0000 and above is
aliased to the same 1 GiB low memory with different cache attributes.

At present QEMU maps the system memory contiguously from 0x8000_0000.
This corrects the wrong QEMU logic. Note address 0x14_0000_0000 is
the alias to the high memory, and even physical memory is only 1 GiB,
the HSS codes still tries to probe the high memory alias address.
It seems there is no issue on the real hardware, so we will have to
take that into the consideration in our emulation. Due to this, we
we increase the default system memory size to 1537 MiB (the minimum
required high memory size by HSS) so that user gets notified an error
when less than 1537 MiB is specified.

Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-id: 20201101170538.3732-1-bmeng.cn@gmail.com
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
4 years agohw/riscv: microchip_pfsoc: Map the reserved memory at address 0
Bin Meng [Wed, 28 Oct 2020 05:30:08 +0000 (13:30 +0800)]
hw/riscv: microchip_pfsoc: Map the reserved memory at address 0

Somehow HSS needs to access address 0 [1] for the DDR calibration data
which is in the chipset's reserved memory. Let's map it.

[1] See the config_copy() calls in various places in ddr_setup() in
    the HSS source codes.

Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-id: 1603863010-15807-9-git-send-email-bmeng.cn@gmail.com
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
4 years agohw/riscv: microchip_pfsoc: Connect the SYSREG module
Bin Meng [Wed, 28 Oct 2020 05:30:07 +0000 (13:30 +0800)]
hw/riscv: microchip_pfsoc: Connect the SYSREG module

Previously SYSREG was created as an unimplemented device. Now that
we have a simple SYSREG module, connect it.

Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-id: 1603863010-15807-8-git-send-email-bmeng.cn@gmail.com
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
4 years agohw/misc: Add Microchip PolarFire SoC SYSREG module support
Bin Meng [Wed, 28 Oct 2020 05:30:06 +0000 (13:30 +0800)]
hw/misc: Add Microchip PolarFire SoC SYSREG module support

This creates a minimum model for Microchip PolarFire SoC SYSREG
module. It only implements the ENVM_CR register to tell guest
software that eNVM is running at the configured divider rate.

Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-id: 1603863010-15807-7-git-send-email-bmeng.cn@gmail.com
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
4 years agohw/riscv: microchip_pfsoc: Connect the IOSCB module
Bin Meng [Wed, 28 Oct 2020 05:30:05 +0000 (13:30 +0800)]
hw/riscv: microchip_pfsoc: Connect the IOSCB module

Previously IOSCB_CFG was created as an unimplemented device. With
the new IOSCB model, its memory range is already covered by the
IOSCB hence remove the previous unimplemented device creation in
the SoC codes.

Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-id: 1603863010-15807-6-git-send-email-bmeng.cn@gmail.com
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
4 years agohw/misc: Add Microchip PolarFire SoC IOSCB module support
Bin Meng [Wed, 28 Oct 2020 05:30:04 +0000 (13:30 +0800)]
hw/misc: Add Microchip PolarFire SoC IOSCB module support

This creates a model for PolarFire SoC IOSCB [1] module. It actually
contains lots of sub-modules like various PLLs to control different
peripherals. Only the mininum capabilities are emulated to make the
HSS DDR memory initialization codes happy. Lots of sub-modules are
created as an unimplemented devices.

[1] PF_SoC_RegMap_V1_1/MPFS250T/mpfs250t_ioscb_memmap_dri.htm in
    https://www.microsemi.com/document-portal/doc_download/1244581-polarfire-soc-register-map

Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-id: 1603863010-15807-5-git-send-email-bmeng.cn@gmail.com
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
4 years agohw/riscv: microchip_pfsoc: Connect DDR memory controller modules
Bin Meng [Wed, 28 Oct 2020 05:30:03 +0000 (13:30 +0800)]
hw/riscv: microchip_pfsoc: Connect DDR memory controller modules

Connect DDR SGMII PHY module and CFG module to the PolarFire SoC.

Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-id: 1603863010-15807-4-git-send-email-bmeng.cn@gmail.com
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
4 years agohw/misc: Add Microchip PolarFire SoC DDR Memory Controller support
Bin Meng [Wed, 28 Oct 2020 05:30:02 +0000 (13:30 +0800)]
hw/misc: Add Microchip PolarFire SoC DDR Memory Controller support

The PolarFire SoC DDR Memory Controller mainly includes 2 modules,
called SGMII PHY module and the CFG module, as documented in the
chipset datasheet.

This creates a single file that groups these 2 modules, providing
the minimum functionalities that make the HSS DDR initialization
codes happy.

Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-id: 1603863010-15807-3-git-send-email-bmeng.cn@gmail.com
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
4 years agohw/riscv: microchip_pfsoc: Document where to look at the SoC memory maps
Bin Meng [Wed, 28 Oct 2020 05:30:01 +0000 (13:30 +0800)]
hw/riscv: microchip_pfsoc: Document where to look at the SoC memory maps

It is not easy to find out the memory map for a specific component
in the PolarFire SoC as the information is scattered in different
documents. Add some comments so that people can know where to get
such information from the Microchip website.

Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-id: 1603863010-15807-2-git-send-email-bmeng.cn@gmail.com
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
4 years agotarget/riscv: Add sifive_plic vmstate
Yifei Jiang [Mon, 26 Oct 2020 11:55:30 +0000 (19:55 +0800)]
target/riscv: Add sifive_plic vmstate

Add sifive_plic vmstate for supporting sifive_plic migration.
Current vmstate framework only supports one structure parameter
as num field to describe variable length arrays, so introduce
num_enables.

Signed-off-by: Yifei Jiang <jiangyifei@huawei.com>
Signed-off-by: Yipeng Yin <yinyipeng1@huawei.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-id: 20201026115530.304-7-jiangyifei@huawei.com
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
4 years agotarget/riscv: Add V extension state description
Yifei Jiang [Mon, 26 Oct 2020 11:55:29 +0000 (19:55 +0800)]
target/riscv: Add V extension state description

In the case of supporting V extension, add V extension description
to vmstate_riscv_cpu.

Signed-off-by: Yifei Jiang <jiangyifei@huawei.com>
Signed-off-by: Yipeng Yin <yinyipeng1@huawei.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-id: 20201026115530.304-6-jiangyifei@huawei.com
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
4 years agotarget/riscv: Add H extension state description
Yifei Jiang [Mon, 26 Oct 2020 11:55:28 +0000 (19:55 +0800)]
target/riscv: Add H extension state description

In the case of supporting H extension, add H extension description
to vmstate_riscv_cpu.

Signed-off-by: Yifei Jiang <jiangyifei@huawei.com>
Signed-off-by: Yipeng Yin <yinyipeng1@huawei.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-id: 20201026115530.304-5-jiangyifei@huawei.com
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
4 years agotarget/riscv: Add PMP state description
Yifei Jiang [Mon, 26 Oct 2020 11:55:27 +0000 (19:55 +0800)]
target/riscv: Add PMP state description

In the case of supporting PMP feature, add PMP state description
to vmstate_riscv_cpu.

'vmstate_pmp_addr' and 'num_rules' could be regenerated by
pmp_update_rule(). But there exists the problem of updating
num_rules repeatedly in pmp_update_rule(). So here extracts
pmp_update_rule_addr() and pmp_update_rule_nums() to update
'vmstate_pmp_addr' and 'num_rules' respectively.

Signed-off-by: Yifei Jiang <jiangyifei@huawei.com>
Signed-off-by: Yipeng Yin <yinyipeng1@huawei.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-id: 20201026115530.304-4-jiangyifei@huawei.com
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
4 years agotarget/riscv: Add basic vmstate description of CPU
Yifei Jiang [Mon, 26 Oct 2020 11:55:26 +0000 (19:55 +0800)]
target/riscv: Add basic vmstate description of CPU

Add basic CPU state description to the newly created machine.c

Signed-off-by: Yifei Jiang <jiangyifei@huawei.com>
Signed-off-by: Yipeng Yin <yinyipeng1@huawei.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-id: 20201026115530.304-3-jiangyifei@huawei.com
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
4 years agotarget/riscv: Merge m/vsstatus and m/vsstatush into one uint64_t unit
Yifei Jiang [Mon, 26 Oct 2020 11:55:25 +0000 (19:55 +0800)]
target/riscv: Merge m/vsstatus and m/vsstatush into one uint64_t unit

mstatus/mstatush and vsstatus/vsstatush are two halved for RISCV32.
This patch expands mstatus and vsstatus to uint64_t instead of
target_ulong so that it can be saved as one unit and reduce some
ifdefs in the code.

Signed-off-by: Yifei Jiang <jiangyifei@huawei.com>
Signed-off-by: Yipeng Yin <yinyipeng1@huawei.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-id: 20201026115530.304-2-jiangyifei@huawei.com

4 years agohw/riscv: virt: Allow passing custom DTB
Anup Patel [Thu, 22 Oct 2020 05:32:25 +0000 (11:02 +0530)]
hw/riscv: virt: Allow passing custom DTB

Extend virt machine to allow passing custom DTB using "-dtb"
command-line parameter. This will help users pass modified DTB
to virt machine.

Signed-off-by: Anup Patel <anup.patel@wdc.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-id: 20201022053225.2596110-2-anup.patel@wdc.com
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
4 years agohw/riscv: sifive_u: Allow passing custom DTB
Anup Patel [Thu, 22 Oct 2020 05:32:24 +0000 (11:02 +0530)]
hw/riscv: sifive_u: Allow passing custom DTB

Extend sifive_u machine to allow passing custom DTB using "-dtb"
command-line parameter. This will help users pass modified DTB
or Linux SiFive DTB to sifive_u machine.

Signed-off-by: Anup Patel <anup.patel@wdc.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-id: 20201022053225.2596110-1-anup.patel@wdc.com
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
4 years agotests/qtest: Fix potential NULL pointer dereference in qos_build_main_args()
AlexChen [Tue, 3 Nov 2020 14:53:09 +0000 (22:53 +0800)]
tests/qtest: Fix potential NULL pointer dereference in qos_build_main_args()

In qos_build_main_args(), the pointer 'path' is dereferenced before
checking it is valid, which may lead to NULL pointer dereference.
So move the assignment to 'cmd_line' after checking 'path' is valid.

Reported-by: Euler Robot <euler.robot@huawei.com>
Signed-off-by: Alex Chen <alex.chen@huawei.com>
Message-Id: <5FA16ED5.4000203@huawei.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoconfigure: fix gio_libs reference
Bruce Rogers [Tue, 3 Nov 2020 14:51:21 +0000 (07:51 -0700)]
configure: fix gio_libs reference

The gio library detection code no longer works, due to a missing $ in
front of the gio_libs reference. Make the string be $gio_libs.

Fixes: 76346b6264a ("configure: Test that gio libs from pkg-config
work")

Signed-off-by: Bruce Rogers <brogers@suse.com>
Message-Id: <20201103145121.668865-1-brogers@suse.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agomeson: fix warning for bad sphinx-build
Paolo Bonzini [Tue, 3 Nov 2020 13:59:42 +0000 (08:59 -0500)]
meson: fix warning for bad sphinx-build

The warning was printing an empty string if the bad sphinx-build
was not passed on the command line.  Instead, always use the
path that was returned by find_program.

Reported-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agotests/qtest/libqos/ahci.c: Avoid NULL dereference in ahci_exec()
Peter Maydell [Tue, 3 Nov 2020 11:52:57 +0000 (11:52 +0000)]
tests/qtest/libqos/ahci.c: Avoid NULL dereference in ahci_exec()

In ahci_exec() we attempt to permit the caller to pass a NULL pointer
for opts_in (in which case we use a default set of options).  However
although we check for NULL when setting up the opts variable at the
top of the function, we unconditionally dereference opts_in at the
end of the function as part of freeing the opts->buffer.

Switch to checking whether the final buffer is the same as the
buffer we started with, instead of assuming the value we started
with is always opts_in->buffer.

At the moment all the callers pass a non-NULL opts argument, so
we never saw any crashes in practice.

Fixes: Coverity CID 1432302
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Message-Id: <20201103115257.23623-1-peter.maydell@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agotests/qtest/libqtest.c: Check for setsockopt() failure
Peter Maydell [Tue, 3 Nov 2020 11:51:12 +0000 (11:51 +0000)]
tests/qtest/libqtest.c: Check for setsockopt() failure

In socket_accept() we use setsockopt() to set SO_RCVTIMEO,
but we don't check the return value for failure. Do so.

Fixes: Coverity CID 1432321
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Message-Id: <20201103115112.19211-1-peter.maydell@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agomeson: vhost-user-gpu/virtiofsd: use absolute path
Marc Hartmayer [Tue, 3 Nov 2020 11:23:33 +0000 (12:23 +0100)]
meson: vhost-user-gpu/virtiofsd: use absolute path

The option `libexecdir` is relative to `prefix` (see
https://mesonbuild.com/Builtin-options.html), so we have to be aware
of this when creating 50-qemu-gpu.json and
50-qemu-virtiofsd.json. Otherwise, tools like libvirt will not be able
to find the executable.

Fixes: 16bf7a3326d8 ("configure: move directory options from config-host.mak to meson")
Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Message-Id: <20201103112333.24734-1-mhartmay@linux.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agomeson: use b_staticpic=false for meson >=0.56.0
Paolo Bonzini [Mon, 19 Oct 2020 10:56:16 +0000 (06:56 -0400)]
meson: use b_staticpic=false for meson >=0.56.0

Meson 0.56.0 correctly builds non-PIC static libraries with -fPIE if
b_pie=true.  We do not have to pass b_staticpic=true if PIE is requested
if Meson is new-enough, which improves performance.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoqtest: add a reproducer for LP#1878642
Alexander Bulekov [Mon, 2 Nov 2020 16:33:36 +0000 (11:33 -0500)]
qtest: add a reproducer for LP#1878642

https://bugs.launchpad.net/qemu/+bug/1878642

Suggested-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Alexander Bulekov <alxndr@bu.edu>
Message-Id: <20201102163336.115444-1-alxndr@bu.edu>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agohw/isa/lpc_ich9: Ignore reserved/invalid SCI IRQ
Philippe Mathieu-Daudé [Fri, 17 Jul 2020 15:17:05 +0000 (17:17 +0200)]
hw/isa/lpc_ich9: Ignore reserved/invalid SCI IRQ

libFuzzer triggered the following assertion:

  cat << EOF | qemu-system-i386 -M pc-q35-5.0 \
    -nographic -monitor none -serial none \
    -qtest stdio -d guest_errors -trace pci\*
  outl 0xcf8 0x8400f841
  outl 0xcfc 0xebed205d
  outl 0x5d02 0xedf82049
  EOF
  pci_cfg_write ICH9-LPC 31:0 @0x41 <- 0xebed205d
  hw/pci/pci.c:268: int pci_bus_get_irq_level(PCIBus *, int): Assertion `irq_num < bus->nirq' failed.

This is because ich9_lpc_sci_irq() returns -1 for reserved
(illegal) values, but ich9_lpc_pmbase_sci_update() considers
it valid and store it in a 8-bit unsigned type. Then the 255
value is used as GSI IRQ, resulting in a PIRQ value of 247,
more than ICH9_LPC_NB_PIRQS (8).

Fix by simply ignoring the invalid access (and reporting it):

  pci_cfg_write ICH9-LPC 31:0 @0x41 <- 0xebed205d
  ICH9 LPC: SCI IRQ SEL #3 is reserved
  pci_cfg_read mch 00:0 @0x0 -> 0x8086
  pci_cfg_read mch 00:0 @0x0 -> 0x29c08086
  ...

Cc: qemu-stable@nongnu.org
Reported-by: Alexander Bulekov <alxndr@bu.edu>
Fixes: 8f242cb724 ("ich9: implement SCI_IRQ_SEL register")
BugLink: https://bugs.launchpad.net/qemu/+bug/1878642
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20200717151705.18611-1-f4bug@amsat.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoscripts/oss-fuzz: rename bin/qemu-fuzz-i386
Alexander Bulekov [Sun, 1 Nov 2020 21:22:45 +0000 (16:22 -0500)]
scripts/oss-fuzz: rename bin/qemu-fuzz-i386

OSS-Fuzz changed the way it scans for fuzzers in $DEST_DIR. The new code
also scans subdirectories for fuzzers. This means that OSS-Fuzz is
considering bin/qemu-fuzz-i386 as an independent fuzzer (it is not - it
requires a --fuzz-target argument). This has led to coverage-build
failures and false crash reports. To work around this, we take advantage
of OSS-Fuzz' filename extension check - OSS-Fuzz will not run anything
that has an extension that is not ".exe":
https://github.com/google/oss-fuzz/blob/master/infra/utils.py#L115

Reported-by: OSS-Fuzz (Issue 26725)
Reported-by: OSS-Fuzz (Issue 26679)
Signed-off-by: Alexander Bulekov <alxndr@bu.edu>
Message-Id: <20201101212245.185819-1-alxndr@bu.edu>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoexec: Remove dead code (CID 1432876)
Philippe Mathieu-Daudé [Fri, 30 Oct 2020 15:37:52 +0000 (16:37 +0100)]
exec: Remove dead code (CID 1432876)

We removed the global_locking field in commit 4174495408a,
leaving dead code around the 'unlocked' variable. Remove it
to fix the DEADCODE issue reported by Coverity (CID 1432876).

Fixes: 4174495408a ("exec: Remove MemoryRegion::global_locking field")
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20201030153752.1557776-1-philmd@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agodocs: expand sourceset documentation
Paolo Bonzini [Thu, 29 Oct 2020 12:06:09 +0000 (13:06 +0100)]
docs: expand sourceset documentation

Expand on the usage of sourcesets and describe the CONFIG_ALL
symbol.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agocutils: replace strdup with g_strdup
Paolo Bonzini [Mon, 2 Nov 2020 18:19:53 +0000 (13:19 -0500)]
cutils: replace strdup with g_strdup

Memory returned by get_relocated_path must be freed with
free or g_free depending on the path that the function
took; Coverity takes exception to this practice.  The
fix lets caller use g_free as is standard in QEMU.

While at it, mention the requirements on the caller in
the doc comment.

Suggested-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoMerge remote-tracking branch 'remotes/berrange-gitlab/tags/sock-next-pull-request...
Peter Maydell [Tue, 3 Nov 2020 14:40:40 +0000 (14:40 +0000)]
Merge remote-tracking branch 'remotes/berrange-gitlab/tags/sock-next-pull-request' into staging

 - Fix inverted logic in abstract socket QAPI support
 - Only report abstract socket support in QAPI on Linux hosts
 - Expand test coverage
 - Misc other code cleanups

# gpg: Signature made Tue 03 Nov 2020 14:00:53 GMT
# gpg:                using RSA key DAF3A6FDB26B62912D0E8E3FBE86EBB415104FDF
# gpg: Good signature from "Daniel P. Berrange <dan@berrange.com>" [full]
# gpg:                 aka "Daniel P. Berrange <berrange@redhat.com>" [full]
# Primary key fingerprint: DAF3 A6FD B26B 6291 2D0E  8E3F BE86 EBB4 1510 4FDF

* remotes/berrange-gitlab/tags/sock-next-pull-request:
  sockets: Make abstract UnixSocketAddress depend on CONFIG_LINUX
  sockets: Bypass "replace empty @path" for abstract unix sockets
  char-socket: Fix qemu_chr_socket_address() for abstract sockets
  sockets: Fix socket_sockaddr_to_address_unix() for abstract sockets
  sockets: Fix default of UnixSocketAddress member @tight
  test-util-sockets: Test the complete abstract socket matrix
  test-util-sockets: Synchronize properly, don't sleep(1)
  test-util-sockets: Factor out test_socket_unix_abstract_one()
  test-util-sockets: Clean up SocketAddress construction
  test-util-sockets: Correct to set has_abstract, has_tight
  test-util-sockets: Plug file descriptor leak

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
4 years agosockets: Make abstract UnixSocketAddress depend on CONFIG_LINUX
Markus Armbruster [Mon, 2 Nov 2020 09:44:22 +0000 (10:44 +0100)]
sockets: Make abstract UnixSocketAddress depend on CONFIG_LINUX

The abstract socket namespace is a non-portable Linux extension.  An
attempt to use it elsewhere should fail with ENOENT (the abstract
address looks like a "" pathname, which does not resolve).  We report
this failure like

    Failed to connect socket abc: No such file or directory

Tolerable, although ENOTSUP would be better.

However, introspection lies: it has @abstract regardless of host
support.  Easy enough to fix: since Linux provides them since 2.2,
'if': 'defined(CONFIG_LINUX)' should do.

The above failure becomes

    Parameter 'backend.data.addr.data.abstract' is unexpected

I consider this an improvement.

Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
4 years agosockets: Bypass "replace empty @path" for abstract unix sockets
Markus Armbruster [Mon, 2 Nov 2020 09:44:21 +0000 (10:44 +0100)]
sockets: Bypass "replace empty @path" for abstract unix sockets

unix_listen_saddr() replaces empty @path by unique value.  It obtains
the value by creating and deleting a unique temporary file with
mkstemp().  This is racy, as the comment explains.  It's also entirely
undocumented as far as I can tell.  Goes back to commit d247d25f18
"sockets: helper functions for qemu (Gerd Hoffman)", v0.10.0.

Since abstract socket addresses have no connection with filesystem
pathnames, making them up with mkstemp() seems inappropriate.  Bypass
the replacement of empty @path.

Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
4 years agochar-socket: Fix qemu_chr_socket_address() for abstract sockets
Markus Armbruster [Mon, 2 Nov 2020 09:44:20 +0000 (10:44 +0100)]
char-socket: Fix qemu_chr_socket_address() for abstract sockets

Commit 776b97d360 "qemu-sockets: add abstract UNIX domain socket
support" neglected to update qemu_chr_socket_address().  It shows
shows neither @abstract nor @tight.  Fix that.

Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
4 years agosockets: Fix socket_sockaddr_to_address_unix() for abstract sockets
Markus Armbruster [Mon, 2 Nov 2020 09:44:19 +0000 (10:44 +0100)]
sockets: Fix socket_sockaddr_to_address_unix() for abstract sockets

Commit 776b97d360 "qemu-sockets: add abstract UNIX domain socket
support" neglected to update socket_sockaddr_to_address_unix().  The
function returns a non-abstract socket address for abstract
sockets (wrong) with a null @path (also wrong; a non-optional QAPI str
member must never be null).

The null @path is due to confused code going back all the way to
commit 17c55decec "sockets: add helpers for creating SocketAddress
from a socket".

Add the required special case, and simplify the confused code.

Fixes: 776b97d3605ed0fc94443048fdf988c7725e38a9
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
4 years agosockets: Fix default of UnixSocketAddress member @tight
Markus Armbruster [Mon, 2 Nov 2020 09:44:18 +0000 (10:44 +0100)]
sockets: Fix default of UnixSocketAddress member @tight

An optional bool member of a QAPI struct can be false, true, or absent.
The previous commit demonstrated that socket_listen() and
socket_connect() are broken for absent @tight, and indeed QMP chardev-
add also defaults absent member @tight to false instead of true.

In C, QAPI members are represented by two fields, has_MEMBER and MEMBER.
We have:

            has_MEMBER    MEMBER
    false         true     false
    true          true      true
    absent       false  false/ignore

When has_MEMBER is false, MEMBER should be set to false on write, and
ignored on read.

For QMP, the QAPI visitors handle absent @tight by setting both
@has_tight and @tight to false.  unix_listen_saddr() and
unix_connect_saddr() however use @tight only, disregarding @has_tight.
This is wrong and means that absent @tight defaults to false whereas it
should default to true.

The same is true for @has_abstract, though @abstract defaults to
false and therefore has the same behavior for all of QMP, HMP and CLI.
Fix unix_listen_saddr() and unix_connect_saddr() to check
@has_abstract/@has_tight, and to default absent @tight to true.

However, this is only half of the story.  HMP chardev-add and CLI
-chardev so far correctly defaulted @tight to true, but defaults to
false again with the above fix for HMP and CLI.  In fact, the "tight"
and "abstract" options now break completely.

Digging deeper, we find that qemu_chr_parse_socket() also ignores
@has_tight, leaving it false when it sets @tight.  That is also wrong,
but the two wrongs cancelled out.  Fix qemu_chr_parse_socket() to set
@has_tight and @has_abstract; writing testcases for HMP and CLI is left
for another day.

Fixes: 776b97d3605ed0fc94443048fdf988c7725e38a9
Reported-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
4 years agotest-util-sockets: Test the complete abstract socket matrix
Markus Armbruster [Mon, 2 Nov 2020 09:44:17 +0000 (10:44 +0100)]
test-util-sockets: Test the complete abstract socket matrix

The test covers only two out of nine combinations.  Test all nine.
Four turn out to be broken.  Marked /* BUG */.

Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
4 years agotest-util-sockets: Synchronize properly, don't sleep(1)
Markus Armbruster [Mon, 2 Nov 2020 09:44:16 +0000 (10:44 +0100)]
test-util-sockets: Synchronize properly, don't sleep(1)

The abstract sockets test spawns a thread to listen and accept, and a
second one to connect, with a sleep(1) in between to "ensure" the
former is listening when the latter tries to connect.  Review fail.
Risks spurious test failure, say when a heavily loaded machine doesn't
schedule the first thread quickly enough.  It's also slow.

Listen and accept in the main thread, and start the connect thread in
between.  Look ma, no sleep!  Run time drops from 2s wall clock to a
few milliseconds.

Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
4 years agotest-util-sockets: Factor out test_socket_unix_abstract_one()
Markus Armbruster [Mon, 2 Nov 2020 09:44:15 +0000 (10:44 +0100)]
test-util-sockets: Factor out test_socket_unix_abstract_one()

Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
4 years agotest-util-sockets: Clean up SocketAddress construction
Markus Armbruster [Mon, 2 Nov 2020 09:44:14 +0000 (10:44 +0100)]
test-util-sockets: Clean up SocketAddress construction

The thread functions build the SocketAddress from global variable
@abstract_sock_name and the tight flag passed as pointer
argument (either NULL or (gpointer)1).  There is no need for such
hackery; simply pass the SocketAddress instead.

While there, dumb down g_rand_int_range() to g_random_int().

Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
4 years agotest-util-sockets: Correct to set has_abstract, has_tight
Markus Armbruster [Mon, 2 Nov 2020 09:44:13 +0000 (10:44 +0100)]
test-util-sockets: Correct to set has_abstract, has_tight

The code tested doesn't care, which is a bug I will fix shortly.

Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
4 years agotest-util-sockets: Plug file descriptor leak
Markus Armbruster [Mon, 2 Nov 2020 09:44:12 +0000 (10:44 +0100)]
test-util-sockets: Plug file descriptor leak

Fixes: 4d3a329af59ef8acd076f99f05e82531d8129b34
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
4 years agoMerge remote-tracking branch 'remotes/mdroth/tags/qga-pull-2020-10-27-v3-tag' into...
Peter Maydell [Tue, 3 Nov 2020 12:47:58 +0000 (12:47 +0000)]
Merge remote-tracking branch 'remotes/mdroth/tags/qga-pull-2020-10-27-v3-tag' into staging

qemu-ga patch queue for soft-freeze

* add guest-get-disks for w32/linux
* add guest-{add,remove,get}-authorized-keys
* fix API violations and schema documentation inconsistencies with
  recently-added guest-get-devices

v3:
- fix checkpatch errors regarding disallowed usages of g_assert*
  macros and other warnings

v2:
- fix BSD build error due to missing stub for guest_get_disks
- fix clang build error on linux due to unused variable
- disable qga-ssh-test for now due to a memory leak within GLib when
  G_TEST_OPTION_ISOLATE_DIRS is passed to g_test_init() since it
  break Gitlab CI build-oss-fuzz test
- rebased and re-tested on master

# gpg: Signature made Tue 03 Nov 2020 02:30:50 GMT
# gpg:                using RSA key CEACC9E15534EBABB82D3FA03353C9CEF108B584
# gpg:                issuer "michael.roth@amd.com"
# gpg: Good signature from "Michael Roth <flukshun@gmail.com>" [full]
# gpg:                 aka "Michael Roth <mdroth@utexas.edu>" [full]
# gpg:                 aka "Michael Roth <mdroth@linux.vnet.ibm.com>" [full]
# Primary key fingerprint: CEAC C9E1 5534 EBAB B82D  3FA0 3353 C9CE F108 B584

* remotes/mdroth/tags/qga-pull-2020-10-27-v3-tag:
  qga: add ssh-get-authorized-keys
  meson: minor simplification
  qga: add *reset argument to ssh-add-authorized-keys
  qga: add ssh-{add,remove}-authorized-keys
  glib-compat: add g_unix_get_passwd_entry_qemu()
  qga: add implementation of guest-get-disks for Windows
  qga: add implementation of guest-get-disks for Linux
  qga: add command guest-get-disks
  qga: Flatten simple union GuestDeviceId
  qga-win: Fix guest-get-devices error API violations
  qga: Use common time encoding for guest-get-devices 'driver-date'
  qga: Rename guest-get-devices return member 'address' to 'id'

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
4 years agovirtio-iommu: Set supported page size mask
Bharat Bhushan [Fri, 30 Oct 2020 18:05:09 +0000 (19:05 +0100)]
virtio-iommu: Set supported page size mask

The virtio-iommu device can deal with arbitrary page sizes for virtual
endpoints, but for endpoints assigned with VFIO it must follow the page
granule used by the host IOMMU driver.

Implement the interface to set the vIOMMU page size mask, called by VFIO
for each endpoint. We assume that all host IOMMU drivers use the same
page granule (the host page granule). Override the page_size_mask field
in the virtio config space.

Signed-off-by: Bharat Bhushan <bbhushan2@marvell.com>
Signed-off-by: Jean-Philippe Brucker <jean-philippe@linaro.org>
Message-Id: <20201030180510.747225-10-jean-philippe@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
4 years agovfio: Set IOMMU page size as per host supported page size
Bharat Bhushan [Fri, 30 Oct 2020 18:05:08 +0000 (19:05 +0100)]
vfio: Set IOMMU page size as per host supported page size

Set IOMMU supported page size mask same as host Linux supported page
size mask.

Acked-by: Alex Williamson <alex.williamson@redhat.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Bharat Bhushan <bbhushan2@marvell.com>
Signed-off-by: Jean-Philippe Brucker <jean-philippe@linaro.org>
Message-Id: <20201030180510.747225-9-jean-philippe@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
4 years agomemory: Add interface to set iommu page size mask
Bharat Bhushan [Fri, 30 Oct 2020 18:05:07 +0000 (19:05 +0100)]
memory: Add interface to set iommu page size mask

Allow to set the page size mask supported by an iommu memory region.
This enables a vIOMMU to communicate the page size granule supported by
an assigned device, on hosts that use page sizes greater than 4kB.

Acked-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Bharat Bhushan <bbhushan2@marvell.com>
Signed-off-by: Jean-Philippe Brucker <jean-philippe@linaro.org>
Message-Id: <20201030180510.747225-8-jean-philippe@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>