qemu.git
4 months agoqxl: check release info object
Prasad J Pandit [Thu, 25 Apr 2019 06:35:34 +0000 (12:05 +0530)] 
qxl: check release info object

When releasing spice resources in release_resource() routine,
if release info object 'ext.info' is null, it leads to null
pointer dereference. Add check to avoid it.

Reported-by: Bugs SysSec <bugs-syssec@rub.de>
Signed-off-by: Prasad J Pandit <pjp@fedoraproject.org>
Message-id: 20190425063534.32747-1-ppandit@redhat.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
4 months agoMerge remote-tracking branch 'remotes/ehabkost/tags/python-next-pull-request' into...
Peter Maydell [Fri, 3 May 2019 14:26:09 +0000 (15:26 +0100)] 
Merge remote-tracking branch 'remotes/ehabkost/tags/python-next-pull-request' into staging

Python queue, 2019-05-02

* configure: automatically pick python3 is available
  (Daniel P. Berrangé)

* tests/acceptance (Cleber Rosa, Philippe Mathieu-Daudé):
  * Multi-architecture test support
  * Multiple arch-specific boot_linux_console test cases
  * Increase verbosity of avocado by default
  * docstring improvements

# gpg: Signature made Fri 03 May 2019 01:40:06 BST
# gpg:                using RSA key 2807936F984DC5A6
# gpg: Good signature from "Eduardo Habkost <ehabkost@redhat.com>" [full]
# Primary key fingerprint: 5A32 2FD5 ABC4 D3DB ACCF  D1AA 2807 936F 984D C5A6

* remotes/ehabkost/tags/python-next-pull-request:
  configure: automatically pick python3 is available
  tests/boot_linux_console: add a test for alpha + clipper
  tests/boot_linux_console: add a test for s390x + s390-ccw-virtio
  tests/boot_linux_console: add a test for arm + virt
  tests/boot_linux_console: add a test for aarch64 + virt
  tests/boot_linux_console: add a test for mips64el + malta
  tests/boot_linux_console: add a test for mips + malta
  scripts/qemu.py: support adding a console with the default serial device
  tests/boot_linux_console: refactor the console watcher into utility method
  tests/boot_linux_console: increase timeout
  tests/boot_linux_console: add common kernel command line options
  tests/boot_linux_console: update the x86_64 kernel
  tests/boot_linux_console: rename the x86_64 after the arch and machine
  tests/acceptance: look for target architecture in test tags first
  tests/acceptance: use "arch:" tag to filter target specific tests
  tests/acceptance: introduce arch parameter and attribute
  tests/acceptance: fix doc reference to avocado_qemu directory
  tests/acceptance: improve docstring on pick_default_qemu_bin()
  tests/acceptance: show avocado test execution by default

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
# Conflicts:
# configure

4 months agoMerge remote-tracking branch 'remotes/lersek/tags/smbios_lp_1821884_20190503' into...
Peter Maydell [Fri, 3 May 2019 13:57:35 +0000 (14:57 +0100)] 
Merge remote-tracking branch 'remotes/lersek/tags/smbios_lp_1821884_20190503' into staging

Fix <https://bugs.launchpad.net/qemu/+bug/1821884>:
"Extend uefi-test-tools to report SMBIOS location".

# gpg: Signature made Fri 03 May 2019 10:10:31 BST
# gpg:                using RSA key D39DA71E0D496CFA
# gpg: Good signature from "Laszlo Ersek <lersek@redhat.com>" [marginal]
# gpg: WARNING: This key is not certified with sufficiently trusted signatures!
# gpg:          It is not certain that the signature belongs to the owner.
# Primary key fingerprint: F5D9 660F 1BA5 F310 A95A  C5E0 466A EAE0 6125 3988
#      Subkey fingerprint: B3A5 5D3F 88A8 90ED 2E63  3E8D D39D A71E 0D49 6CFA

* remotes/lersek/tags/smbios_lp_1821884_20190503:
  tests/uefi-boot-images: report the SMBIOS entry point structures
  tests/uefi-test-tools: report the SMBIOS entry point structures

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
4 months agoMerge remote-tracking branch 'remotes/kraxel/tags/usb-20190503-v2-pull-request' into...
Peter Maydell [Fri, 3 May 2019 12:57:51 +0000 (13:57 +0100)] 
Merge remote-tracking branch 'remotes/kraxel/tags/usb-20190503-v2-pull-request' into staging

usb: bugfixes for mtp and xhci, split ohci-pci.

# gpg: Signature made Fri 03 May 2019 07:59:39 BST
# 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-20190503-v2-pull-request:
  hw/usb: avoid format truncation warning when formatting port name
  hw/usb/hcd-ohci: Move PCI-related code into a separate file
  hw/usb/hcd-ohci: Do not use PCI functions with sysbus devices in ohci_die()
  usb/xhci: avoid trigger assertion if guest write wrong epid
  usb-mtp: change default to success for usb_mtp_update_object
  usb-mtp: fix alignment of access of ObjectInfo filename field
  usb-mtp: fix string length for filename when writing metadata

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
4 months agogitmodules: use qemu.org git mirrors
Stefan Hajnoczi [Thu, 25 Apr 2019 14:54:20 +0000 (15:54 +0100)] 
gitmodules: use qemu.org git mirrors

qemu.org hosts git repository mirrors of all submodules.  Update
.gitmodules to use the mirrors and not the upstream repositories.

Mirroring upstream repositories ensures that QEMU continues to build
even when upstream repositories are deleted or temporarily offline.

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-id: 20190425145420.8888-1-stefanha@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
4 months agoMerge remote-tracking branch 'remotes/thibault/tags/samuel-thibault' into staging
Peter Maydell [Fri, 3 May 2019 11:51:02 +0000 (12:51 +0100)] 
Merge remote-tracking branch 'remotes/thibault/tags/samuel-thibault' into staging

slirp: move slirp as git submodule project

Marc-André Lureau (2):
  build-sys: pass CFLAGS & LDFLAGS to subdir-slirp
  build-sys: move slirp as git submodule project

# gpg: Signature made Thu 02 May 2019 23:20:52 BST
# gpg:                using RSA key E61DBB15D4172BDEC97E92D9DB550E89F0FA54F3
# gpg: Good signature from "Samuel Thibault <samuel.thibault@aquilenet.fr>" [unknown]
# gpg:                 aka "Samuel Thibault <sthibault@debian.org>" [marginal]
# gpg:                 aka "Samuel Thibault <samuel.thibault@gnu.org>" [unknown]
# gpg:                 aka "Samuel Thibault <samuel.thibault@inria.fr>" [marginal]
# gpg:                 aka "Samuel Thibault <samuel.thibault@labri.fr>" [marginal]
# gpg:                 aka "Samuel Thibault <samuel.thibault@ens-lyon.org>" [marginal]
# gpg:                 aka "Samuel Thibault <samuel.thibault@u-bordeaux.fr>" [unknown]
# gpg: WARNING: This key is not certified with sufficiently trusted signatures!
# gpg:          It is not certain that the signature belongs to the owner.
# Primary key fingerprint: 900C B024 B679 31D4 0F82  304B D017 8C76 7D06 9EE6
#      Subkey fingerprint: E61D BB15 D417 2BDE C97E  92D9 DB55 0E89 F0FA 54F3

* remotes/thibault/tags/samuel-thibault:
  build-sys: move slirp as git submodule project
  build-sys: pass CFLAGS & LDFLAGS to subdir-slirp

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
4 months agotests/uefi-boot-images: report the SMBIOS entry point structures
Laszlo Ersek [Thu, 25 Apr 2019 09:54:51 +0000 (11:54 +0200)] 
tests/uefi-boot-images: report the SMBIOS entry point structures

Rebuild the "bios-tables-test" UEFI boot images with the SMBIOS entry
point reporting that has been added in the previous patch.

Cc: "Philippe Mathieu-Daudé" <philmd@redhat.com>
Cc: Igor Mammedov <imammedo@redhat.com>
Launchpad: https://bugs.launchpad.net/qemu/+bug/1821884
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Igor Mammedov <imammedo@redhat.com>
4 months agotests/uefi-test-tools: report the SMBIOS entry point structures
Laszlo Ersek [Thu, 25 Apr 2019 09:44:15 +0000 (11:44 +0200)] 
tests/uefi-test-tools: report the SMBIOS entry point structures

On UEFI systems, the SMBIOS entry point (a.k.a. anchor) structures are
found similarly to the ACPI RSD PTR table(s): by scanning the
ConfigurationTable array in the EFI system table for well-known GUIDs.

Locate the SMBIOS 2.1 (32-bit) and 3.0 (64-bit) anchors in the
BiosTablesTest UEFI application, and report the addresses in new fields
appended to the BIOS_TABLES_TEST structure.

Cc: "Philippe Mathieu-Daudé" <philmd@redhat.com>
Cc: Igor Mammedov <imammedo@redhat.com>
Launchpad: https://bugs.launchpad.net/qemu/+bug/1821884
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Igor Mammedov <imammedo@redhat.com>
4 months agohw/usb: avoid format truncation warning when formatting port name
Daniel P. Berrangé [Fri, 12 Apr 2019 12:16:25 +0000 (13:16 +0100)] 
hw/usb: avoid format truncation warning when formatting port name

hw/usb/hcd-xhci.c: In function ‘usb_xhci_realize’:
hw/usb/hcd-xhci.c:3339:66: warning: ‘%d’ directive output may be truncated writing between 1 and 10 bytes into a region of size 5 [-Wformat-trunca\
tion=]
 3339 |             snprintf(port->name, sizeof(port->name), "usb2 port #%d", i+1);
      |                                                                  ^~
hw/usb/hcd-xhci.c:3339:54: note: directive argument in the range [1, 2147483647]
 3339 |             snprintf(port->name, sizeof(port->name), "usb2 port #%d", i+1);
      |                                                      ^~~~~~~~~~~~~~~

The xhci code formats the port name into a fixed length
buffer which is only large enough to hold port numbers
upto 5 digits in decimal representation. We're never
going to have a port number that large, so aserting the
port number is sensible is sufficient to tell GCC the
formatted string won't be truncated.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20190412121626.19829-5-berrange@redhat.com>

[ kraxel: also s/int/unsigned int/ to tell gcc they can't
          go negative. ]

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
4 months agoconfigure: automatically pick python3 is available
Daniel P. Berrangé [Wed, 27 Mar 2019 17:07:01 +0000 (17:07 +0000)] 
configure: automatically pick python3 is available

Unless overridden via an env var or configure arg, QEMU will only look
for the 'python' binary in $PATH. This is unhelpful on distros which
are only shipping Python 3.x (eg Fedora) in their default install as,
if they comply with PEP 394, the bare 'python' binary won't exist.

This changes configure so that by default it will search for all three
common python binaries, preferring to find Python 3.x versions.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20190327170701.23798-1-berrange@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
4 months agotests/boot_linux_console: add a test for alpha + clipper
Cleber Rosa [Tue, 12 Mar 2019 17:18:24 +0000 (13:18 -0400)] 
tests/boot_linux_console: add a test for alpha + clipper

Similar to the x86_64 + pc test, it boots a Linux kernel on a Malta
board and verify the serial is working.  One extra command added to
the QEMU command line is '-vga std', because the kernel used is
known to crash without it.

If alpha is a target being built, "make check-acceptance" will
automatically include this test by the use of the "arch:alpha" tags.

Alternatively, this test can be run using:

    $ avocado run -t arch:alpha tests/acceptance
    $ avocado run -t machine:clipper tests/acceptance

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Cleber Rosa <crosa@redhat.com>
Reviewed-by: Caio Carrara <ccarrara@redhat.com>
Message-Id: <20190312171824.5134-21-crosa@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
4 months agotests/boot_linux_console: add a test for s390x + s390-ccw-virtio
Cleber Rosa [Tue, 12 Mar 2019 17:18:23 +0000 (13:18 -0400)] 
tests/boot_linux_console: add a test for s390x + s390-ccw-virtio

Just like the previous tests, boots a Linux kernel on a s390x target
using the s390-ccw-virtio machine.

Because it's not possible to have multiple VT220 consoles,
'-nodefaults' is used, so that the one set with set_console() works
correctly.

Signed-off-by: Cleber Rosa <crosa@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Reviewed-by: Caio Carrara <ccarrara@redhat.com>
Message-Id: <20190312171824.5134-20-crosa@redhat.com>
[ehabkost: Updated kernel URL to point to fedoraproject.org]
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
4 months agotests/boot_linux_console: add a test for arm + virt
Cleber Rosa [Tue, 12 Mar 2019 17:18:22 +0000 (13:18 -0400)] 
tests/boot_linux_console: add a test for arm + virt

Just like the previous tests, boots a Linux kernel on an arm target
using the virt machine.

Signed-off-by: Cleber Rosa <crosa@redhat.com>
Reviewed-by: Caio Carrara <ccarrara@redhat.com>
Message-Id: <20190312171824.5134-19-crosa@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
4 months agotests/boot_linux_console: add a test for aarch64 + virt
Cleber Rosa [Tue, 12 Mar 2019 17:18:21 +0000 (13:18 -0400)] 
tests/boot_linux_console: add a test for aarch64 + virt

Just like the previous tests, boots a Linux kernel on a aarch64 target
using the virt machine.

One special option added is the CPU type, given that the kernel
selected fails to boot on the virt machine's default CPU (cortex-a15).

Signed-off-by: Cleber Rosa <crosa@redhat.com>
Reviewed-by: Caio Carrara <ccarrara@redhat.com>
Message-Id: <20190312171824.5134-18-crosa@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
4 months agotests/boot_linux_console: add a test for mips64el + malta
Cleber Rosa [Tue, 12 Mar 2019 17:18:18 +0000 (13:18 -0400)] 
tests/boot_linux_console: add a test for mips64el + malta

Similar to the x86_64 + pc test, it boots a Linux kernel on a Malta
board and verify the serial is working.

If mips64el is a target being built, "make check-acceptance" will
automatically include this test by the use of the "arch:mips64el"
tags.

Alternatively, this test can be run using:

    $ avocado run -t arch:mips64el tests/acceptance
    $ avocado run -t machine:malta tests/acceptance

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Cleber Rosa <crosa@redhat.com>
Reviewed-by: Aleksandar Markovic <amarkovic@wavecomp.com>
Message-Id: <20190312171824.5134-15-crosa@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
4 months agotests/boot_linux_console: add a test for mips + malta
Philippe Mathieu-Daudé [Tue, 12 Mar 2019 17:18:17 +0000 (13:18 -0400)] 
tests/boot_linux_console: add a test for mips + malta

Similar to the x86_64 + pc test, it boots a Linux kernel on a Malta
board and verify the serial is working.  Also, it relies on the serial
device set by the machine itself.

If mips is a target being built, "make check-acceptance" will
automatically include this test by the use of the "arch:mips" tags.

Alternatively, this test can be run using:

    $ avocado run -t arch:mips tests/acceptance
    $ avocado run -t machine:malta tests/acceptance
    $ avocado run -t endian:big tests/acceptance

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Cleber Rosa <crosa@redhat.com>
Reviewed-by: Aleksandar Markovic <amarkovic@wavecomp.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20190312171824.5134-14-crosa@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
4 months agoscripts/qemu.py: support adding a console with the default serial device
Cleber Rosa [Tue, 12 Mar 2019 17:18:16 +0000 (13:18 -0400)] 
scripts/qemu.py: support adding a console with the default serial device

The set_console() utility function either adds a device based on the
explicitly given device type, or adds a known good type of device
based on the machine type.

But, for a number of machine types, it may be impossible or
inconvenient to add the devices by means of "-device" command line
options, and then it may better to just use the "-serial" option and
let QEMU itself, based on the machine type, set the device
accordingly.

To achieve that, the behavior of set_console() now flags the intention
to add a console device on launch(), and if no explicit device type is
given the "-serial" option is going to be added to the QEMU command
line, instead of raising exceptions.

Based on testing with different machine types, the CONSOLE_DEV_TYPES
is not necessary anymore, so it's being removed, as is the logic to
use it.

Signed-off-by: Cleber Rosa <crosa@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <20190312171824.5134-13-crosa@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
4 months agotests/boot_linux_console: refactor the console watcher into utility method
Cleber Rosa [Tue, 12 Mar 2019 17:18:15 +0000 (13:18 -0400)] 
tests/boot_linux_console: refactor the console watcher into utility method

This introduces a utility method that monitors the console device and
looks for either a message that signals the test success or failure.

Signed-off-by: Cleber Rosa <crosa@redhat.com>
Reviewed-by: Caio Carrara <ccarrara@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <20190312171824.5134-12-crosa@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
4 months agotests/boot_linux_console: increase timeout
Cleber Rosa [Tue, 12 Mar 2019 17:18:14 +0000 (13:18 -0400)] 
tests/boot_linux_console: increase timeout

When running on very low powered environments, some tests may time out
causing false negatives.  As a conservative change, and for
considering that human time (investigating false negatives) is worth
more than some extra machine cycles (and time), let's increase the
overall timeout.

CC: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Cleber Rosa <crosa@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <20190312171824.5134-11-crosa@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
4 months agotests/boot_linux_console: add common kernel command line options
Cleber Rosa [Tue, 12 Mar 2019 17:18:13 +0000 (13:18 -0400)] 
tests/boot_linux_console: add common kernel command line options

The 'printk.time=0' option makes it easier to parse the console
output.  Let's set it as a default, and reusable, kernel command line
options for this and future similar tests.

Signed-off-by: Cleber Rosa <crosa@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <20190312171824.5134-10-crosa@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
4 months agotests/boot_linux_console: update the x86_64 kernel
Cleber Rosa [Tue, 12 Mar 2019 17:18:12 +0000 (13:18 -0400)] 
tests/boot_linux_console: update the x86_64 kernel

Update to the stock Fedora 29 kernel, from the Fedora 28.  New tests
will be added using the 29 kernel, so for consistency, let's also
update it here.

Signed-off-by: Cleber Rosa <crosa@redhat.com>
Reviewed-by: Caio Carrara <ccarrara@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Acked-by: Cornelia Huck <cohuck@redhat.com>
CC: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20190312171824.5134-9-crosa@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
4 months agotests/boot_linux_console: rename the x86_64 after the arch and machine
Cleber Rosa [Tue, 12 Mar 2019 17:18:11 +0000 (13:18 -0400)] 
tests/boot_linux_console: rename the x86_64 after the arch and machine

Given that the test is specific to x86_64 and pc, and new tests are
going to be added to the same class, let's rename it accordingly.
Also, let's make the class documentation not architecture specific.

Signed-off-by: Cleber Rosa <crosa@redhat.com>
Reviewed-by: Caio Carrara <ccarrara@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <20190312171824.5134-8-crosa@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
4 months agotests/acceptance: look for target architecture in test tags first
Cleber Rosa [Tue, 12 Mar 2019 17:18:10 +0000 (13:18 -0400)] 
tests/acceptance: look for target architecture in test tags first

A test can, optionally, be tagged for one or many architectures.  If a
test has been tagged for a single architecture, there's a high chance
that the test won't run on other architectures.  This changes the
default order of choosing a default target architecture to use based
on the 'arch' tag value first.

The precedence order is for choosing a QEMU binary to use for a test
is now:

 * qemu_bin parameter
 * arch parameter
 * arch tag value (for example, x86_64 if ":avocado: tags=arch:x86_64
   is used)

This means that if one runs:

 $ avocado run -p qemu_bin=/usr/bin/qemu-system-x86_64 test.py

No arch parameter or tag will influence the selection of the QEMU
target binary.  If one runs:

 $ avocado run -p arch=ppc64 test.py

The target binary selection mechanism will attempt to find a binary
such as "ppc64-softmmu/qemu-system-ppc64".  And finally, if one runs
a test that is tagged (in its docstring) with "arch:aarch64":

 $ avocado run aarch64.py

The target binary selection mechanism will attempt to find a binary
such as "aarch64-softmmu/qemu-system-aarch64".

At this time, no provision is made to cancel the execution of tests if
the arch parameter given (manually) does not match the test "arch"
tag, but it may be a useful default behavior to be added in the
future.

Signed-off-by: Cleber Rosa <crosa@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <20190312171824.5134-7-crosa@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
4 months agotests/acceptance: use "arch:" tag to filter target specific tests
Cleber Rosa [Tue, 12 Mar 2019 17:18:09 +0000 (13:18 -0400)] 
tests/acceptance: use "arch:" tag to filter target specific tests

Currently, some tests contains target architecture information, in the
form of a "x86_64" tag.  But that tag is not respected in the default
execution, that is, "make check-acceptance" doesn't do anything with
it.

That said, even the target architecture handling currently present in
the "avocado_qemu.Test" class is pretty limited.  For instance, by
default, it chooses a target based on the host architecture.

Because the original implementation of the tags feature in Avocado did
not include any time of namespace or "key:val" mechanism, no tag has
relation to another tag.  The new implementation of the tags feature
from version 67.0 onwards, allows "key:val" tags, and because of that,
a test can be classified with a tag in a given key.  For instance, the
new proposed version of the "boot_linux_console.py" test, which
downloads and attempts to run a x86_64 kernel, is now tagged as:

  :avocado: tags=arch:x86_64

This means that it can be filtered (out) when no x86_64 target is
available.  At the same time, tests that don't have a "arch:" tag,
will not be filtered out.

Signed-off-by: Cleber Rosa <crosa@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <20190312171824.5134-6-crosa@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
4 months agotests/acceptance: introduce arch parameter and attribute
Cleber Rosa [Tue, 12 Mar 2019 17:18:08 +0000 (13:18 -0400)] 
tests/acceptance: introduce arch parameter and attribute

It's useful to define the architecture that should be used in
situations such as:
 * the intended target of the QEMU binary to be used on tests
 * the architecture of code to be run within the QEMU binary, such
   as a kernel image or a full blown guest OS image

This commit introduces both a test parameter and a test instance
attribute, that will contain such a value.

Now, when the "arch" test parameter is given, it will influence the
selection of the default QEMU binary, if one is not given explicitly
by means of the "qemu_img" parameter.

Signed-off-by: Cleber Rosa <crosa@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <20190312171824.5134-5-crosa@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
4 months agotests/acceptance: fix doc reference to avocado_qemu directory
Cleber Rosa [Tue, 12 Mar 2019 17:18:07 +0000 (13:18 -0400)] 
tests/acceptance: fix doc reference to avocado_qemu directory

The "this directory" reference is misleading and confusing, it's a
leftover from when this text was proposed in a README file inside
the "tests/acceptance/avocado_qemu" directory.

When that text was moved to the top level docs directory, the
reference was not updated.

Signed-off-by: Cleber Rosa <crosa@redhat.com>
Reviewed-by: Caio Carrara <ccarrara@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <20190312171824.5134-4-crosa@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
4 months agotests/acceptance: improve docstring on pick_default_qemu_bin()
Cleber Rosa [Tue, 12 Mar 2019 17:18:06 +0000 (13:18 -0400)] 
tests/acceptance: improve docstring on pick_default_qemu_bin()

Making it clear what is returned by this utility function.

Signed-off-by: Cleber Rosa <crosa@redhat.com>
Reviewed-by: Caio Carrara <ccarrara@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <20190312171824.5134-3-crosa@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
4 months agotests/acceptance: show avocado test execution by default
Cleber Rosa [Tue, 12 Mar 2019 17:18:05 +0000 (13:18 -0400)] 
tests/acceptance: show avocado test execution by default

The current version of the "check-acceptance" target will only show
one line for execution of all tests.  That's probably OK if the tests
to be run are quick enough and they're always the same.

But, there's already one test alone that takes on average ~5 seconds
to run, we intend to adapt the list of tests to match the user's build
environment (among other choices).

Because of that, let's present the default Avocado UI by default.
Users can always choose a different output by setting the AVOCADO_SHOW
variable.

Signed-off-by: Cleber Rosa <crosa@redhat.com>
Reviewed-by: Caio Carrara <ccarrara@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <20190312171824.5134-2-crosa@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
4 months agobuild-sys: move slirp as git submodule project
Marc-André Lureau [Wed, 24 Apr 2019 11:00:41 +0000 (13:00 +0200)] 
build-sys: move slirp as git submodule project

The slirp project is now hosted on freedesktop at:
https://gitlab.freedesktop.org/slirp.

The libslirp source was extracted from qemu/slirp filtered through
clang-format (available in project tree). The qemu slirp directory can
be swapped by a git submodule.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20190424110041.8175-3-marcandre.lureau@redhat.com>
Signed-off-by: Samuel Thibault <samuel.thibault@ens-lyon.org>
4 months agobuild-sys: pass CFLAGS & LDFLAGS to subdir-slirp
Marc-André Lureau [Wed, 24 Apr 2019 11:00:40 +0000 (13:00 +0200)] 
build-sys: pass CFLAGS & LDFLAGS to subdir-slirp

CFLAGS/LDFLAGS have debug and sanitizers flags, which should be passed
to slirp compilation.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20190424110041.8175-2-marcandre.lureau@redhat.com>
Signed-off-by: Samuel Thibault <samuel.thibault@ens-lyon.org>
4 months agoMerge remote-tracking branch 'remotes/huth-gitlab/tags/pull-request-2019-05-02' into...
Peter Maydell [Thu, 2 May 2019 16:17:09 +0000 (17:17 +0100)] 
Merge remote-tracking branch 'remotes/huth-gitlab/tags/pull-request-2019-05-02' into staging

- Move qtest accel code to accel/qtest.c, get rid of AccelClass->available
- Test TCG interpreter in gitlab-ci
- Small improvements to the configure script
- Use object_initialize_child in hw/pci-host

# gpg: Signature made Thu 02 May 2019 17:07:34 BST
# gpg:                using RSA key 2ED9D774FE702DB5
# gpg: Good signature from "Thomas Huth <th.huth@gmx.de>" [full]
# gpg:                 aka "Thomas Huth <thuth@redhat.com>" [full]
# gpg:                 aka "Thomas Huth <huth@tuxfamily.org>" [full]
# gpg:                 aka "Thomas Huth <th.huth@posteo.de>" [unknown]
# Primary key fingerprint: 27B8 8847 EEE0 2501 18F3  EAB9 2ED9 D774 FE70 2DB5

* remotes/huth-gitlab/tags/pull-request-2019-05-02:
  hw/pci-host: Use object_initialize_child for correct reference counting
  configure: Relax check for libseccomp
  configure: Remove old *-config-devices.mak.d files when running configure
  configure: Add -Wno-typedef-redefinition to CFLAGS (for Clang)
  accel: Remove unused AccelClass::available field
  qtest: Don't compile qtest accel on non-POSIX systems
  qtest: Move accel code to accel/qtest.c
  gitlab-ci.yml: Test the TCG interpreter in a CI pipeline

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
4 months agohw/pci-host: Use object_initialize_child for correct reference counting
Thomas Huth [Tue, 30 Apr 2019 19:15:52 +0000 (21:15 +0200)] 
hw/pci-host: Use object_initialize_child for correct reference counting

Both functions, object_initialize() and object_property_add_child() increase
the reference counter of the new object, so one of the references has to be
dropped afterwards to get the reference counting right. Otherwise the child
object might not be properly cleaned up when the parent gets destroyed.
Some functions of the pci-host devices miss to drop one of the references.
Fix it by using object_initialize_child() instead, which takes care of
calling object_initialize(), object_property_add_child() and object_unref()
in the right order.

Suggested-by: Eduardo Habkost <ehabkost@redhat.com>
Message-Id: <20190430191552.4027-1-thuth@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
4 months agoconfigure: Relax check for libseccomp
Helge Deller [Thu, 4 Apr 2019 18:39:23 +0000 (20:39 +0200)] 
configure: Relax check for libseccomp

All major distributions do support libseccomp version >= 2.3.0, so there
is no need to special-case on various architectures any longer.

Signed-off-by: Helge Deller <deller@gmx.de>
Message-Id: <20190404183923.GA22347@ls3530.dellerweb.de>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Acked-by: Eduardo Otubo <otubo@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
4 months agoconfigure: Remove old *-config-devices.mak.d files when running configure
Thomas Huth [Mon, 11 Mar 2019 10:20:34 +0000 (11:20 +0100)] 
configure: Remove old *-config-devices.mak.d files when running configure

When running "make" in a build directory from the pre-Kconfig merge time,
the build process currently fails with:

 make: *** No rule to make target `.../default-configs/pci.mak',
  needed by `aarch64-softmmu/config-devices.mak'.  Stop.

To make sure that this problem at least goes away when the user runs
"configure" (or "sh config.status") again, we have to make sure that
we re-generate the .mak.d files. Thus remove the old stale files
while running the configure script.

Message-Id: <1552300145-12526-1-git-send-email-thuth@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
4 months agoconfigure: Add -Wno-typedef-redefinition to CFLAGS (for Clang)
Thomas Huth [Wed, 24 Apr 2019 11:05:25 +0000 (13:05 +0200)] 
configure: Add -Wno-typedef-redefinition to CFLAGS (for Clang)

Without the -Wno-typedef-redefinition option, clang complains if a typedef
gets redefined in gnu99 mode (since this is officially a C11 feature). This
used to also happen with older versions of GCC, but since we've bumped our
minimum GCC version to 4.8, all versions of GCC that we support do not seem
to issue this warning in gnu99 mode anymore. So this has become a common
problem for people who only test their code with GCC - they do not notice
the issue until they submit their patches and suddenly patchew or a
maintainer complains.

Now that we do not urgently need to keep the code clean from typedef
redefintions anymore with recent versions of GCC, we can ease the
situation with clang, too, and simply shut these warnings off for good.

Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20190427154539.11336-1-thuth@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
4 months agoaccel: Remove unused AccelClass::available field
Eduardo Habkost [Mon, 22 Apr 2019 21:04:48 +0000 (18:04 -0300)] 
accel: Remove unused AccelClass::available field

The field is not used anymore, we can remove it.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Message-Id: <20190422210448.2488-4-ehabkost@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com> [on mingw64]
Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
4 months agoqtest: Don't compile qtest accel on non-POSIX systems
Eduardo Habkost [Mon, 22 Apr 2019 21:04:47 +0000 (18:04 -0300)] 
qtest: Don't compile qtest accel on non-POSIX systems

qtest_available() will always return 0 on non-POSIX systems.
It's simpler to just not compile the accelerator code on those
systems instead of relying on the AccelClass::available function.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Message-Id: <20190422210448.2488-3-ehabkost@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com> [on mingw64]
Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
4 months agoqtest: Move accel code to accel/qtest.c
Eduardo Habkost [Mon, 22 Apr 2019 21:04:46 +0000 (18:04 -0300)] 
qtest: Move accel code to accel/qtest.c

QTest has two parts: the server (-qtest) and the accelerator
(-machine accel=qtest).  The accelerator depends on CONFIG_POSIX
due to its usage of sigwait(), but the server doesn't.

Move the accel code to accel/qtest.c.  Later we will disable
compilation of accel/qtest.c on non-POSIX systems.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Message-Id: <20190422210448.2488-2-ehabkost@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
[thuth: added fixup for MAINTAINERS file]
Signed-off-by: Thomas Huth <thuth@redhat.com>
4 months agogitlab-ci.yml: Test the TCG interpreter in a CI pipeline
Thomas Huth [Wed, 10 Apr 2019 12:35:50 +0000 (14:35 +0200)] 
gitlab-ci.yml: Test the TCG interpreter in a CI pipeline

So far we do not have any test coverage for TCI (the TCG interpreter) yet.
Thus let's add a CI pipeline that runs at least some basic TCG tests with
a TCI build, to make sure that there are no further regressions.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20190410123550.2362-1-thuth@redhat.com>

4 months agoMerge remote-tracking branch 'remotes/jnsnow/tags/bitmaps-pull-request' into staging
Peter Maydell [Thu, 2 May 2019 11:04:51 +0000 (12:04 +0100)] 
Merge remote-tracking branch 'remotes/jnsnow/tags/bitmaps-pull-request' into staging

Pull request

# gpg: Signature made Wed 01 May 2019 21:24:16 BST
# gpg:                using RSA key F9B7ABDBBCACDF95BE76CBD07DEF8106AAFC390E
# gpg: Good signature from "John Snow (John Huston) <jsnow@redhat.com>" [full]
# Primary key fingerprint: FAEB 9711 A12C F475 812F  18F2 88A9 064D 1835 61EB
#      Subkey fingerprint: F9B7 ABDB BCAC DF95 BE76  CBD0 7DEF 8106 AAFC 390E

* remotes/jnsnow/tags/bitmaps-pull-request:
  docs/interop/bitmaps: rewrite and modernize doc
  Makefile: add nit-picky mode to sphinx-build

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
4 months agoMerge remote-tracking branch 'remotes/dgilbert/tags/pull-hmp-20190501' into staging
Peter Maydell [Thu, 2 May 2019 10:21:02 +0000 (11:21 +0100)] 
Merge remote-tracking branch 'remotes/dgilbert/tags/pull-hmp-20190501' into staging

HMP pull

New gva2gpa command
delvm now uses hmp_handle_error so gets Error: prefix in messages

# gpg: Signature made Wed 01 May 2019 11:57:43 BST
# gpg:                using RSA key 45F5C71B4A0CB7FB977A9FA90516331EBC5BFDE7
# gpg: Good signature from "Dr. David Alan Gilbert (RH2) <dgilbert@redhat.com>" [full]
# Primary key fingerprint: 45F5 C71B 4A0C B7FB 977A  9FA9 0516 331E BC5B FDE7

* remotes/dgilbert/tags/pull-hmp-20190501:
  hmp: gva2gpa debug command
  hmp: delvm: use hmp_handle_error

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
4 months agohw/usb/hcd-ohci: Move PCI-related code into a separate file
Thomas Huth [Fri, 19 Apr 2019 07:56:25 +0000 (09:56 +0200)] 
hw/usb/hcd-ohci: Move PCI-related code into a separate file

Some machines (like the pxa2xx-based ARM machines) only have a sysbus
OHCI controller, but no PCI. With the new Kconfig-style build system,
it will soon be possible to create QEMU binaries that only contain
such PCI-less machines. However, the two OHCI controllers, for sysbus
and for PCI, are currently both located in one file, so the PCI code
is still required for linking here. Move the OHCI-PCI device code
into a separate file, so that it is possible to use the sysbus OHCI
device also without the PCI dependency.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20190419075625.24251-3-thuth@redhat.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
4 months agohw/usb/hcd-ohci: Do not use PCI functions with sysbus devices in ohci_die()
Thomas Huth [Fri, 19 Apr 2019 07:56:24 +0000 (09:56 +0200)] 
hw/usb/hcd-ohci: Do not use PCI functions with sysbus devices in ohci_die()

The ohci_die() function always assumes to be running with a PCI OHCI
controller and calls the PCI-specific functions pci_set_word(). However,
this function might also get called for the sysbus OHCI devices, so it
likely fails in that case. To fix this issue, change the code now, so that
there are two implementations now, one for sysbus and one for PCI, and
use the right function via a function pointer in the OHCIState structure.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20190419075625.24251-2-thuth@redhat.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
4 months agousb/xhci: avoid trigger assertion if guest write wrong epid
Longpeng [Tue, 30 Apr 2019 06:21:41 +0000 (14:21 +0800)] 
usb/xhci: avoid trigger assertion if guest write wrong epid

we found the following core in our environment:
0  0x00007fc6b06c2237 in raise ()
1  0x00007fc6b06c3928 in abort ()
2  0x00007fc6b06bb056 in __assert_fail_base ()
3  0x00007fc6b06bb102 in __assert_fail ()
4  0x0000000000702e36 in xhci_kick_ep (...)
5  0x000000000047897a in memory_region_write_accessor (...)
6  0x000000000047767f in access_with_adjusted_size (...)
7  0x000000000047944d in memory_region_dispatch_write (...)
(mr=mr@entry=0x7fc6a0138df0, addr=addr@entry=156, data=1648892416,
size=size@entry=4, attrs=attrs@entry=...)
8  0x000000000042df17 in address_space_write_continue (...)
10 0x000000000043084d in address_space_rw (...)
11 0x000000000047451b in kvm_cpu_exec (cpu=cpu@entry=0x1ab11b0)
12 0x000000000045dcf5 in qemu_kvm_cpu_thread_fn (arg=0x1ab11b0)
13 0x0000000000870631 in qemu_thread_start (args=args@entry=0x1acfb50)
14 0x00000000008959a7 in thread_entry_for_hotfix (pthread_cb=<optimized out>)
15 0x00007fc6b0a60dd5 in start_thread ()
16 0x00007fc6b078a59d in clone ()

(gdb) f 5
5  0x000000000047897a in memory_region_write_accessor (...)
529     mr->ops->write(mr->opaque, addr, tmp, size);
(gdb) p /x tmp
$9 = 0x62481a00 <-- last byte 0x00 is @epid

xhci_doorbell_write() already check the upper bound of @slotid an @epid,
it also need to check the lower bound.

Cc: Gonglei <arei.gonglei@huawei.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Longpeng <longpeng2@huawei.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 1556605301-44112-1-git-send-email-longpeng2@huawei.com

[ kraxel: fixed typo in subject line ]

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
4 months agousb-mtp: change default to success for usb_mtp_update_object
Bandan Das [Mon, 15 Apr 2019 16:47:12 +0000 (12:47 -0400)] 
usb-mtp: change default to success for usb_mtp_update_object

Commit c5ead51f90cf (usb-mtp: return incomplete transfer on a lstat
failure) checks if lstat succeeded when updating attributes of a
file. However, it also changed behavior to return an error by
default. This is incorrect because for smaller file sizes, Qemu
will attempt to write the file in one go and there won't be
an object for it.

Fixes: c5ead51f90cf
Signed-off-by: Bandan Das <bsd@redhat.com>
Message-id: jpgwojv9pwv.fsf@linux.bootlegged.copy
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
4 months agousb-mtp: fix alignment of access of ObjectInfo filename field
Daniel P. Berrangé [Mon, 15 Apr 2019 15:45:03 +0000 (16:45 +0100)] 
usb-mtp: fix alignment of access of ObjectInfo filename field

The ObjectInfo struct's "filename" field is following a uint8_t
field in a packed struct and thus has bad alignment for a 16-bit
field. Switch the field to to uint8_t and use the helper function
for accessing unaligned 16-bit data.

Note that although the MTP spec specifies big endian, when transported
over the USB protocol, data is little endian.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-id: 20190415154503.6758-4-berrange@redhat.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
4 months agousb-mtp: fix string length for filename when writing metadata
Daniel P. Berrangé [Mon, 15 Apr 2019 15:45:01 +0000 (16:45 +0100)] 
usb-mtp: fix string length for filename when writing metadata

The ObjectInfo 'length' field provides the length of the
wide character string filename. This is then converted to
a multi-byte character string. This may have a different
byte count to the wide character string. We should use the
C string length of the multi-byte string instead.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-id: 20190415154503.6758-2-berrange@redhat.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
4 months agodocs/interop/bitmaps: rewrite and modernize doc
John Snow [Fri, 26 Apr 2019 22:15:28 +0000 (18:15 -0400)] 
docs/interop/bitmaps: rewrite and modernize doc

This just about rewrites the entirety of the bitmaps.rst document to
make it consistent with the 4.0 release. I have added new features seen
in the 4.0 release, as well as tried to clarify some points that keep
coming up when discussing this feature both in-house and upstream.

It does not yet cover pull backups or migration details, but I intend to
keep extending this document to cover those cases.

Signed-off-by: John Snow <jsnow@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-id: 20190426221528.30293-3-jsnow@redhat.com
[Adjusted commit message. --js]
Signed-off-by: John Snow <jsnow@redhat.com>
4 months agoMakefile: add nit-picky mode to sphinx-build
John Snow [Fri, 26 Apr 2019 22:15:27 +0000 (18:15 -0400)] 
Makefile: add nit-picky mode to sphinx-build

If we add references that don't resolve (or accidentally remove them),
it will be helpful to have warning messages alerting us to that.

Further, turn those warnings into errors so we can be alerted to these
problems sooner rather than later.

Signed-off-by: John Snow <jsnow@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-id: 20190426221528.30293-2-jsnow@redhat.com
[adjusted commit message. --js]
Signed-off-by: John Snow <jsnow@redhat.com>
4 months agohmp: gva2gpa debug command
Dr. David Alan Gilbert [Fri, 12 Apr 2019 15:26:52 +0000 (16:26 +0100)] 
hmp: gva2gpa debug command

Add a gva2gpa command purely for debug which performs
address translation on the gva, the existing gpa2hva
command can then also be used to find it in the qemu
userspace; e.g.

(qemu) info registers
.... RSP=ffffffff81c03e98
....
(qemu) gva2gpa 0xffffffff81c03e98
gpa: 0x1c03e98
(qemu) gpa2hva 0x1c03e98
Host virtual address for 0x1c03e98 (pc.ram) is 0x7f0599a03e98
(qemu) x/10x 0xffffffff81c03e98
ffffffff81c03e98: 0x81c03eb8 0xffffffff 0x8101ea3f 0xffffffff
ffffffff81c03ea8: 0x81d27b00 0xffffffff 0x00000000 0x00000000
ffffffff81c03eb8: 0x81c03ec8 0xffffffff

gdb -p ...qemu...
(gdb) x/10x 0x7f0599a03e98
0x7f0599a03e98: 0x81c03eb8 0xffffffff 0x8101ea3f 0xffffffff
0x7f0599a03ea8: 0x81d27b00 0xffffffff 0x00000000 0x00000000
0x7f0599a03eb8: 0x81c03ec8 0xffffffff

Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <20190412152652.827-1-dgilbert@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
4 months agohmp: delvm: use hmp_handle_error
Cole Robinson [Wed, 24 Apr 2019 21:49:08 +0000 (17:49 -0400)] 
hmp: delvm: use hmp_handle_error

This gives us the consistent 'Error:' prefix added in 66363e9a43f,
which helps users like libvirt who still need to scrape hmp error
messages to detect failure.

Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Cole Robinson <crobinso@redhat.com>
Message-Id: <984634226fa14316641f31f84e7dd0bb770bd94c.1556141033.git.crobinso@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
4 months agoMerge remote-tracking branch 'remotes/kevin/tags/for-upstream' into staging
Peter Maydell [Tue, 30 Apr 2019 16:06:57 +0000 (17:06 +0100)] 
Merge remote-tracking branch 'remotes/kevin/tags/for-upstream' into staging

Block layer patches:

- iotests: Fix output of qemu-io related tests
- Don't ignore bdrv_set_aio_context() for nodes with bs->drv = NUL
- vmdk: Set vmdk parent backing_format to vmdk
- qcow2: Preallocation fixes (especially for external data files)
- Add linear-buffer-based APIs (as wrappers around qiov-based ones)
- Various code cleanups and small corner case fixes

# gpg: Signature made Tue 30 Apr 2019 16:35:09 BST
# gpg:                using RSA key 7F09B272C88F2FD6
# 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: (27 commits)
  block/qed: add missed coroutine_fn markers
  iotests: Check that images are in read-only mode after block-commit
  commit: Make base read-only if there is an early failure
  qemu-img: use buffer-based io
  block/stream: use buffer-based io
  block/commit: use buffer-based io
  block/backup: use buffer-based io
  block/parallels: use buffer-based io
  block/qed: use buffer-based io
  block/qcow: use buffer-based io
  block/qcow2: use buffer-based io
  block: introduce byte-based io helpers
  qcow2: Fix error handling in the compression code
  qcow2: Fix qcow2_make_empty() with external data file
  qemu-img: Make create hint at protocol options
  iotests: Perform the correct test in 082
  qcow2: Fix full preallocation with external data file
  qcow2: Add errp to preallocate_co()
  qcow2: Avoid COW during metadata preallocation
  qemu-img: Saner printing of large file sizes
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
4 months agoblock/qed: add missed coroutine_fn markers
Vladimir Sementsov-Ogievskiy [Tue, 30 Apr 2019 12:36:11 +0000 (15:36 +0300)] 
block/qed: add missed coroutine_fn markers

qed_read_table and qed_write_table use coroutine-only interfaces but
are not marked coroutine_fn. Happily, they are called only from
coroutine context, so we only need to add missed markers.

Reported-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
4 months agoiotests: Check that images are in read-only mode after block-commit
Alberto Garcia [Mon, 29 Apr 2019 13:51:09 +0000 (15:51 +0200)] 
iotests: Check that images are in read-only mode after block-commit

This tests the fix from the previous patch.

Signed-off-by: Alberto Garcia <berto@igalia.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
4 months agocommit: Make base read-only if there is an early failure
Alberto Garcia [Mon, 29 Apr 2019 13:51:08 +0000 (15:51 +0200)] 
commit: Make base read-only if there is an early failure

You can reproduce this by passing an invalid filter-node-name (like
"1234") to block-commit. In this case the base image is put in
read-write mode but is never reset back to read-only.

Signed-off-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
4 months agoqemu-img: use buffer-based io
Vladimir Sementsov-Ogievskiy [Mon, 22 Apr 2019 14:58:38 +0000 (17:58 +0300)] 
qemu-img: use buffer-based io

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
4 months agoblock/stream: use buffer-based io
Vladimir Sementsov-Ogievskiy [Mon, 22 Apr 2019 14:58:37 +0000 (17:58 +0300)] 
block/stream: use buffer-based io

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
4 months agoblock/commit: use buffer-based io
Vladimir Sementsov-Ogievskiy [Mon, 22 Apr 2019 14:58:36 +0000 (17:58 +0300)] 
block/commit: use buffer-based io

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
4 months agoblock/backup: use buffer-based io
Vladimir Sementsov-Ogievskiy [Mon, 22 Apr 2019 14:58:35 +0000 (17:58 +0300)] 
block/backup: use buffer-based io

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
4 months agoblock/parallels: use buffer-based io
Vladimir Sementsov-Ogievskiy [Mon, 22 Apr 2019 14:58:34 +0000 (17:58 +0300)] 
block/parallels: use buffer-based io

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
4 months agoblock/qed: use buffer-based io
Vladimir Sementsov-Ogievskiy [Mon, 22 Apr 2019 14:58:33 +0000 (17:58 +0300)] 
block/qed: use buffer-based io

Move to _co_ versions of io functions qed_read_table() and
qed_write_table(), as we use qemu_co_mutex_unlock()
anyway.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
4 months agoblock/qcow: use buffer-based io
Vladimir Sementsov-Ogievskiy [Mon, 22 Apr 2019 14:58:32 +0000 (17:58 +0300)] 
block/qcow: use buffer-based io

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
4 months agoblock/qcow2: use buffer-based io
Vladimir Sementsov-Ogievskiy [Mon, 22 Apr 2019 14:58:31 +0000 (17:58 +0300)] 
block/qcow2: use buffer-based io

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
4 months agoblock: introduce byte-based io helpers
Vladimir Sementsov-Ogievskiy [Mon, 22 Apr 2019 14:58:30 +0000 (17:58 +0300)] 
block: introduce byte-based io helpers

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
4 months agoqcow2: Fix error handling in the compression code
Alberto Garcia [Tue, 30 Apr 2019 10:08:02 +0000 (13:08 +0300)] 
qcow2: Fix error handling in the compression code

This patch fixes a few things in the way error codes are handled in
the qcow2 compression code:

a) qcow2_co_pwritev_compressed() expects qcow2_co_compress() to only
   return -1 or -2 on failure, but this is not correct. Since the
   change from qcow2_compress() to qcow2_co_compress() in commit
   ceb029cd6feccf9f7607 the new code can also return -EINVAL (although
   there does not seem to exist any code path that would cause that
   error in the current implementation).

b) -1 and -2 are ad-hoc error codes defined in qcow2_compress().
   This patch replaces them with standard constants from errno.h.

c) Both qcow2_compress() and qcow2_co_do_compress() return a negative
   value on failure, but qcow2_co_pwritev_compressed() stores the
   value in an unsigned data type.

Signed-off-by: Alberto Garcia <berto@igalia.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
4 months agoqcow2: Fix qcow2_make_empty() with external data file
Kevin Wolf [Mon, 29 Apr 2019 10:52:21 +0000 (12:52 +0200)] 
qcow2: Fix qcow2_make_empty() with external data file

make_completely_empty() is an optimisated path for bdrv_make_empty()
where completely new metadata is created inside the image file instead
of going through all clusters and discarding them. For an external data
file, however, we actually need to do discard operations on the data
file; just overwriting the qcow2 file doesn't get rid of the data.

The necessary slow path with an explicit discard operation already
exists for other cases. Use it for external data files, too.

Cc: qemu-stable@nongnu.org
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
4 months agoqemu-img: Make create hint at protocol options
Max Reitz [Sat, 13 Apr 2019 15:20:37 +0000 (17:20 +0200)] 
qemu-img: Make create hint at protocol options

qemu-img create allows giving just a format and "-o help" to get a list
of the options supported by that format.  Users may not realize that the
protocol level may offer even more options, which they only get to see
by specifying a filename.

This patch adds a note to hint at that fact.

Signed-off-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
4 months agoiotests: Perform the correct test in 082
Max Reitz [Sat, 13 Apr 2019 15:20:36 +0000 (17:20 +0200)] 
iotests: Perform the correct test in 082

In the "amend" section of 082, we perform a single "convert" test
(namely "convert -o help").  That does not make sense, especially
because we have done exactly that "convert" test earlier in 082 already.

Replacing "convert" by "amend" yields an error, which is correct because
there is no point in "amend" having a default format.  The user has to
either specify the format, or give a file for qemu-img to probe.

Signed-off-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
4 months agoqcow2: Fix full preallocation with external data file
Kevin Wolf [Mon, 15 Apr 2019 14:34:30 +0000 (16:34 +0200)] 
qcow2: Fix full preallocation with external data file

preallocate_co() already gave the data file the full size without
forwarding the requested preallocation mode to the protocol. When
bdrv_co_truncate() was called later with the preallocation mode, the
file didn't actually grow any more, so the data file stayed unallocated
even if full preallocation was requested.

Pass the right preallocation mode to preallocate_co() and remove the
second bdrv_co_truncate() to fix this. As a side effect, the ugly
one-byte write in preallocate_co() is replaced with a truncate call,
now leaving the last block unallocated on the protocol level as it
should be.

Cc: qemu-stable@nongnu.org
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
4 months agoqcow2: Add errp to preallocate_co()
Kevin Wolf [Mon, 15 Apr 2019 14:56:07 +0000 (16:56 +0200)] 
qcow2: Add errp to preallocate_co()

We'll add a bdrv_co_truncate() call in the next patch which can return
an Error that we don't want to discard. So add an errp parameter to
preallocate_co().

Cc: qemu-stable@nongnu.org
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
4 months agoqcow2: Avoid COW during metadata preallocation
Kevin Wolf [Mon, 15 Apr 2019 14:25:01 +0000 (16:25 +0200)] 
qcow2: Avoid COW during metadata preallocation

Limiting the allocation to INT_MAX bytes isn't particularly clever
because it means that the final cluster will be a partial cluster which
will be completed through a COW operation. This results in unnecessary
data read and write requests which lead to an unwanted non-sparse
filesystem block for metadata preallocation.

Align the maximum allocation size down to the cluster size to avoid this
situation.

Cc: qemu-stable@nongnu.org
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
4 months agoqemu-img: Saner printing of large file sizes
Eric Blake [Wed, 17 Apr 2019 17:11:01 +0000 (12:11 -0500)] 
qemu-img: Saner printing of large file sizes

Disk sizes close to INT64_MAX cause overflow, for some pretty
ridiculous output:

  $ ./nbdkit -U - memory size=$((2**63 - 512)) --run 'qemu-img info $nbd'
  image: nbd+unix://?socket=/tmp/nbdkitHSAzNz/socket
  file format: raw
  virtual size: -8388607T (9223372036854775296 bytes)
  disk size: unavailable

But there's no reason to have two separate implementations of integer
to human-readable abbreviation, where one has overflow and stops at
'T', while the other avoids overflow and goes all the way to 'E'. With
this patch, the output now claims 8EiB instead of -8388607T, which
really is the correct rounding of largest file size supported by qemu
(we could go 511 bytes larger if we used byte-accurate sizing instead
of rounding up to the next sector boundary, but that wouldn't change
the human-readable result).

Quite a few iotests need updates to expected output to match.

Reported-by: Richard W.M. Jones <rjones@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
Tested-by: Richard W.M. Jones <rjones@redhat.com>
Reviewed-by: Alberto Garcia <berto@igalia.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Tested-by: Max Reitz <mreitz@redhat.com>
4 months agocutils: Fix size_to_str() on 32-bit platforms
Eric Blake [Wed, 17 Apr 2019 17:11:00 +0000 (12:11 -0500)] 
cutils: Fix size_to_str() on 32-bit platforms

When extracting a human-readable size formatter, we changed 'uint64_t
div' pre-patch to 'unsigned long div' post-patch. Which breaks on
32-bit platforms, resulting in 'inf' instead of intended values larger
than 999GB.

Fixes: 22951aaa
CC: qemu-stable@nongnu.org
Reported-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
4 months agoblock/vhdx: Use IEC binary prefixes for size constants
Stefano Garzarella [Wed, 27 Mar 2019 09:56:34 +0000 (10:56 +0100)] 
block/vhdx: Use IEC binary prefixes for size constants

Using IEC binary prefixes in order to make the code more readable,
with the exception of DEFAULT_LOG_SIZE because it's passed to
stringify().

Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: John Snow <jsnow@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
4 months agoblock/vhdx: Remove redundant IEC binary prefixes definition
Stefano Garzarella [Wed, 27 Mar 2019 09:56:33 +0000 (10:56 +0100)] 
block/vhdx: Remove redundant IEC binary prefixes definition

IEC binary prefixes are already defined in "qemu/units.h",
so we can remove redundant definitions in "block/vhdx.h".

Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: John Snow <jsnow@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
4 months agovmdk: Set vmdk parent backing_format to vmdk
Sam Eiderman [Tue, 26 Mar 2019 19:58:37 +0000 (21:58 +0200)] 
vmdk: Set vmdk parent backing_format to vmdk

Commit b69864e5a ("vmdk: Support version=3 in VMDK descriptor files")
fixed the probe function to correctly guess vmdk descriptors with
version=3.

This solves the issue where vmdk snapshot with parent vmdk descriptor
containing "version=3" would be treated as raw instead vmdk.

In the future case where a new vmdk version is introduced, we will again
experience this issue, even if the user will provide "-f vmdk" it will
only apply to the tip image and not to the underlying "misprobed" parent
image.

The code in vmdk.c already assumes that the backing file of vmdk must be
vmdk (see vmdk_is_cid_valid which returns 0 if backing file is not
vmdk).

So let's make it official by supplying the backing_format as vmdk.

Reviewed-by: Mark Kanda <mark.kanda@oracle.com>
Reviewed-By: Liran Alon <liran.alon@oracle.com>
Reviewed-by: Arbel Moshe <arbel.moshe@oracle.com>
Signed-off-by: Shmuel Eiderman <shmuel.eiderman@oracle.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Fam Zheng <fam@euphon.net>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
4 months agovpc: unlock Coroutine lock to make IO submit Concurrently
Zhengui li [Fri, 15 Mar 2019 14:04:38 +0000 (22:04 +0800)] 
vpc: unlock Coroutine lock to make IO submit Concurrently

Concurrent IO becomes serial IO because of the qemu Coroutine lock,
which reduce IO performance severely.

So unlock Coroutine lock before bdrv_co_pwritev and
bdrv_co_preadv to fix it.

Signed-off-by: Zhengui li <lizhengui@huawei.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
4 months agoblock: Fix AioContext switch for bs->drv == NULL
Kevin Wolf [Wed, 17 Apr 2019 15:15:25 +0000 (17:15 +0200)] 
block: Fix AioContext switch for bs->drv == NULL

Even for block nodes with bs->drv == NULL, we can't just ignore a
bdrv_set_aio_context() call. Leaving the node in its old context can
mean that it's still in an iothread context in bdrv_close_all() during
shutdown, resulting in an attempted unlock of the AioContext lock which
we don't hold.

This is an example stack trace of a related crash:

 #0  0x00007ffff59da57f in raise () at /lib64/libc.so.6
 #1  0x00007ffff59c4895 in abort () at /lib64/libc.so.6
 #2  0x0000555555b97b1e in error_exit (err=<optimized out>, msg=msg@entry=0x555555d386d0 <__func__.19059> "qemu_mutex_unlock_impl") at util/qemu-thread-posix.c:36
 #3  0x0000555555b97f7f in qemu_mutex_unlock_impl (mutex=mutex@entry=0x5555568002f0, file=file@entry=0x555555d378df "util/async.c", line=line@entry=507) at util/qemu-thread-posix.c:97
 #4  0x0000555555b92f55 in aio_context_release (ctx=ctx@entry=0x555556800290) at util/async.c:507
 #5  0x0000555555b05cf8 in bdrv_prwv_co (child=child@entry=0x7fffc80012f0, offset=offset@entry=131072, qiov=qiov@entry=0x7fffffffd4f0, is_write=is_write@entry=true, flags=flags@entry=0)
         at block/io.c:833
 #6  0x0000555555b060a9 in bdrv_pwritev (qiov=0x7fffffffd4f0, offset=131072, child=0x7fffc80012f0) at block/io.c:990
 #7  0x0000555555b060a9 in bdrv_pwrite (child=0x7fffc80012f0, offset=131072, buf=<optimized out>, bytes=<optimized out>) at block/io.c:990
 #8  0x0000555555ae172b in qcow2_cache_entry_flush (bs=bs@entry=0x555556810680, c=c@entry=0x5555568cc740, i=i@entry=0) at block/qcow2-cache.c:51
 #9  0x0000555555ae18dd in qcow2_cache_write (bs=bs@entry=0x555556810680, c=0x5555568cc740) at block/qcow2-cache.c:248
 #10 0x0000555555ae15de in qcow2_cache_flush (bs=0x555556810680, c=<optimized out>) at block/qcow2-cache.c:259
 #11 0x0000555555ae16b1 in qcow2_cache_flush_dependency (c=0x5555568a1700, c=0x5555568a1700, bs=0x555556810680) at block/qcow2-cache.c:194
 #12 0x0000555555ae16b1 in qcow2_cache_entry_flush (bs=bs@entry=0x555556810680, c=c@entry=0x5555568a1700, i=i@entry=0) at block/qcow2-cache.c:194
 #13 0x0000555555ae18dd in qcow2_cache_write (bs=bs@entry=0x555556810680, c=0x5555568a1700) at block/qcow2-cache.c:248
 #14 0x0000555555ae15de in qcow2_cache_flush (bs=bs@entry=0x555556810680, c=<optimized out>) at block/qcow2-cache.c:259
 #15 0x0000555555ad242c in qcow2_inactivate (bs=bs@entry=0x555556810680) at block/qcow2.c:2124
 #16 0x0000555555ad2590 in qcow2_close (bs=0x555556810680) at block/qcow2.c:2153
 #17 0x0000555555ab0c62 in bdrv_close (bs=0x555556810680) at block.c:3358
 #18 0x0000555555ab0c62 in bdrv_delete (bs=0x555556810680) at block.c:3542
 #19 0x0000555555ab0c62 in bdrv_unref (bs=0x555556810680) at block.c:4598
 #20 0x0000555555af4d72 in blk_remove_bs (blk=blk@entry=0x5555568103d0) at block/block-backend.c:785
 #21 0x0000555555af4dbb in blk_remove_all_bs () at block/block-backend.c:483
 #22 0x0000555555aae02f in bdrv_close_all () at block.c:3412
 #23 0x00005555557f9796 in main (argc=<optimized out>, argv=<optimized out>, envp=<optimized out>) at vl.c:4776

The reproducer I used is a qcow2 image on gluster volume, where the
virtual disk size (4 GB) is larger than the gluster volume size (64M),
so we can easily trigger an ENOSPC. This backend is assigned to a
virtio-blk device using an iothread, and then from the guest a
'dd if=/dev/zero of=/dev/vda bs=1G count=1' causes the VM to stop
because of an I/O error. qemu_gluster_co_flush_to_disk() sets
bs->drv = NULL on error, so when virtio-blk stops the dataplane, the
block nodes stay in the iothread AioContext. A 'quit' monitor command
issued from this paused state crashes the process.

Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1631227
Cc: qemu-stable@nongnu.org
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
4 months agotests/qemu-iotests: Fix output of qemu-io related tests
Thomas Huth [Sun, 28 Apr 2019 15:54:44 +0000 (17:54 +0200)] 
tests/qemu-iotests: Fix output of qemu-io related tests

One of the recent commits changed the way qemu-io prints out its
errors and warnings - they are now prefixed with the program name.
We've got to adapt the iotests accordingly to prevent that they
are failing.

Fixes: 99e98d7c9fc1a1639fad ("qemu-io: Use error_[gs]et_progname()")
Signed-off-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
4 months agoMerge remote-tracking branch 'remotes/rth/tags/pull-hppa-20190429' into staging
Peter Maydell [Tue, 30 Apr 2019 09:20:22 +0000 (10:20 +0100)] 
Merge remote-tracking branch 'remotes/rth/tags/pull-hppa-20190429' into staging

Implement fast tlb insert insns.
Fix data tlb exception for pa 1.1.

# gpg: Signature made Mon 29 Apr 2019 22:54:08 BST
# gpg:                using RSA key 7A481E78868B4DB6A85A05C064DF38E8AF7E215F
# gpg:                issuer "richard.henderson@linaro.org"
# gpg: Good signature from "Richard Henderson <richard.henderson@linaro.org>" [full]
# Primary key fingerprint: 7A48 1E78 868B 4DB6 A85A  05C0 64DF 38E8 AF7E 215F

* remotes/rth/tags/pull-hppa-20190429:
  target/hppa: Always return EXCP_DMAR for protection id trap
  target/hppa: Implement Fast TLB Insert instructions

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
4 months agotarget/hppa: Always return EXCP_DMAR for protection id trap
Nick Hudson [Tue, 23 Apr 2019 06:36:21 +0000 (07:36 +0100)] 
target/hppa: Always return EXCP_DMAR for protection id trap

The EXCP_DMP trap is considered legacy.

"In PA-RISC 1.1 (Second Edition) and later revisions, processors must use
traps 26, 27,and 28 which provide equivalent functionality"

Signed-off-by: Nick Hudson <skrll@netbsd.org>
Message-Id: <20190423063621.8203-3-nick.hudson@gmx.co.uk>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
4 months agotarget/hppa: Implement Fast TLB Insert instructions
Nick Hudson [Tue, 23 Apr 2019 06:36:20 +0000 (07:36 +0100)] 
target/hppa: Implement Fast TLB Insert instructions

These instructions are present on pcxl and pcxl2 machines,
and are used by NetBSD and OpenBSD.  See

     https://parisc.wiki.kernel.org/images-parisc/a/a9/Pcxl2_ers.pdf
     page 13-9 (195/206)

Signed-off-by: Nick Hudson <skrll@netbsd.org>
Message-Id: <20190423063621.8203-2-nick.hudson@gmx.co.uk>
[rth: Use extending loads, locally managed temporaries.]
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
4 months agoMerge remote-tracking branch 'remotes/stefanha/tags/tracing-pull-request' into staging
Peter Maydell [Mon, 29 Apr 2019 18:11:15 +0000 (19:11 +0100)] 
Merge remote-tracking branch 'remotes/stefanha/tags/tracing-pull-request' into staging

Pull request

# gpg: Signature made Mon 29 Apr 2019 17:10:40 BST
# gpg:                using RSA key 9CA4ABB381AB73C8
# gpg: Good signature from "Stefan Hajnoczi <stefanha@redhat.com>" [full]
# gpg:                 aka "Stefan Hajnoczi <stefanha@gmail.com>" [full]
# Primary key fingerprint: 8695 A8BF D3F9 7CDA AC35  775A 9CA4 ABB3 81AB 73C8

* remotes/stefanha/tags/tracing-pull-request:
  trace: fix runstate tracing

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
4 months agoMerge remote-tracking branch 'remotes/pmaydell/tags/pull-target-arm-20190429' into...
Peter Maydell [Mon, 29 Apr 2019 17:05:56 +0000 (18:05 +0100)] 
Merge remote-tracking branch 'remotes/pmaydell/tags/pull-target-arm-20190429' into staging

target-arm queue:
 * remove "bag of random stuff" hw/devices.h header
 * implement FPU for Cortex-M and enable it for Cortex-M4 and -M33
 * hw/dma: Compile the bcm2835_dma device as common object
 * configure: Remove --source-path option
 * hw/ssi/xilinx_spips: Avoid variable length array
 * hw/arm/smmuv3: Remove SMMUNotifierNode

# gpg: Signature made Mon 29 Apr 2019 17:58:57 BST
# gpg:                using RSA key E1A5C593CD419DE28E8315CF3C2525ED14360CDE
# gpg:                issuer "peter.maydell@linaro.org"
# gpg: Good signature from "Peter Maydell <peter.maydell@linaro.org>" [ultimate]
# gpg:                 aka "Peter Maydell <pmaydell@gmail.com>" [ultimate]
# gpg:                 aka "Peter Maydell <pmaydell@chiark.greenend.org.uk>" [ultimate]
# Primary key fingerprint: E1A5 C593 CD41 9DE2 8E83  15CF 3C25 25ED 1436 0CDE

* remotes/pmaydell/tags/pull-target-arm-20190429: (42 commits)
  hw/devices: Move SMSC 91C111 declaration into a new header
  hw/net/lan9118: Export TYPE_LAN9118 and use it instead of hardcoded string
  hw/net/ne2000-isa: Add guards to the header
  hw/devices: Move LAN9118 declarations into a new header
  hw/devices: Move TI touchscreen declarations into a new header
  hw/devices: Move Gamepad declarations into a new header
  hw/devices: Move CBus declarations into a new header
  hw/devices: Move Blizzard declarations into a new header
  hw/devices: Move TC6393XB declarations into a new header
  hw/display/tc6393xb: Remove unused functions
  hw/arm/nseries: Use TYPE_TMP105 instead of hardcoded string
  hw/arm/aspeed: Use TYPE_TMP105/TYPE_PCA9552 instead of hardcoded string
  hw/dma: Compile the bcm2835_dma device as common object
  target/arm: Enable FPU for Cortex-M4 and Cortex-M33
  target/arm: Implement VLLDM for v7M CPUs with an FPU
  target/arm: Implement VLSTM for v7M CPUs with an FPU
  target/arm: Implement M-profile lazy FP state preservation
  target/arm: Add lazy-FP-stacking support to v7m_stack_write()
  target/arm: New function armv7m_nvic_set_pending_lazyfp()
  target/arm: New helper function arm_v7m_mmu_idx_all()
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
4 months agohw/devices: Move SMSC 91C111 declaration into a new header
Philippe Mathieu-Daudé [Fri, 12 Apr 2019 16:54:16 +0000 (18:54 +0200)] 
hw/devices: Move SMSC 91C111 declaration into a new header

This commit finally deletes "hw/devices.h".

Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20190412165416.7977-13-philmd@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
4 months agohw/net/lan9118: Export TYPE_LAN9118 and use it instead of hardcoded string
Philippe Mathieu-Daudé [Fri, 12 Apr 2019 16:54:15 +0000 (18:54 +0200)] 
hw/net/lan9118: Export TYPE_LAN9118 and use it instead of hardcoded string

Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20190412165416.7977-12-philmd@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
4 months agohw/net/ne2000-isa: Add guards to the header
Philippe Mathieu-Daudé [Fri, 12 Apr 2019 16:54:14 +0000 (18:54 +0200)] 
hw/net/ne2000-isa: Add guards to the header

Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20190412165416.7977-11-philmd@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
4 months agohw/devices: Move LAN9118 declarations into a new header
Philippe Mathieu-Daudé [Fri, 12 Apr 2019 16:54:13 +0000 (18:54 +0200)] 
hw/devices: Move LAN9118 declarations into a new header

Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20190412165416.7977-10-philmd@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
4 months agohw/devices: Move TI touchscreen declarations into a new header
Philippe Mathieu-Daudé [Fri, 12 Apr 2019 16:54:12 +0000 (18:54 +0200)] 
hw/devices: Move TI touchscreen declarations into a new header

Since uWireSlave is only used in this new header, there is no
need to expose it via "qemu/typedefs.h".

Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20190412165416.7977-9-philmd@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
4 months agohw/devices: Move Gamepad declarations into a new header
Philippe Mathieu-Daudé [Fri, 12 Apr 2019 16:54:11 +0000 (18:54 +0200)] 
hw/devices: Move Gamepad declarations into a new header

Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20190412165416.7977-8-philmd@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
4 months agohw/devices: Move CBus declarations into a new header
Philippe Mathieu-Daudé [Fri, 12 Apr 2019 16:54:10 +0000 (18:54 +0200)] 
hw/devices: Move CBus declarations into a new header

Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20190412165416.7977-7-philmd@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
4 months agohw/devices: Move Blizzard declarations into a new header
Philippe Mathieu-Daudé [Fri, 12 Apr 2019 16:54:09 +0000 (18:54 +0200)] 
hw/devices: Move Blizzard declarations into a new header

Add an entries the Blizzard device in MAINTAINERS.

Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20190412165416.7977-6-philmd@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
4 months agohw/devices: Move TC6393XB declarations into a new header
Philippe Mathieu-Daudé [Fri, 12 Apr 2019 16:54:08 +0000 (18:54 +0200)] 
hw/devices: Move TC6393XB declarations into a new header

Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20190412165416.7977-5-philmd@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
4 months agohw/display/tc6393xb: Remove unused functions
Philippe Mathieu-Daudé [Fri, 12 Apr 2019 16:54:07 +0000 (18:54 +0200)] 
hw/display/tc6393xb: Remove unused functions

No code used the tc6393xb_gpio_in_get() and tc6393xb_gpio_out_set()
functions since their introduction in commit 88d2c950b002. Time to
remove them.

Suggested-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20190412165416.7977-4-philmd@redhat.com
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
4 months agohw/arm/nseries: Use TYPE_TMP105 instead of hardcoded string
Philippe Mathieu-Daudé [Fri, 12 Apr 2019 16:54:06 +0000 (18:54 +0200)] 
hw/arm/nseries: Use TYPE_TMP105 instead of hardcoded string

Suggested-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20190412165416.7977-3-philmd@redhat.com
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
4 months agohw/arm/aspeed: Use TYPE_TMP105/TYPE_PCA9552 instead of hardcoded string
Philippe Mathieu-Daudé [Fri, 12 Apr 2019 16:54:05 +0000 (18:54 +0200)] 
hw/arm/aspeed: Use TYPE_TMP105/TYPE_PCA9552 instead of hardcoded string

Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20190412165416.7977-2-philmd@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
4 months agohw/dma: Compile the bcm2835_dma device as common object
Philippe Mathieu-Daudé [Mon, 29 Apr 2019 16:36:03 +0000 (17:36 +0100)] 
hw/dma: Compile the bcm2835_dma device as common object

This device is used by both ARM (BCM2836, for raspi2) and AArch64
(BCM2837, for raspi3) targets, and is not CPU-specific.
Move it to common object, so we build it once for all targets.

Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20190427133028.12874-1-philmd@redhat.com
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
4 months agotarget/arm: Enable FPU for Cortex-M4 and Cortex-M33
Peter Maydell [Mon, 29 Apr 2019 16:36:03 +0000 (17:36 +0100)] 
target/arm: Enable FPU for Cortex-M4 and Cortex-M33

Enable the FPU by default for the Cortex-M4 and Cortex-M33.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20190416125744.27770-27-peter.maydell@linaro.org

4 months agotarget/arm: Implement VLLDM for v7M CPUs with an FPU
Peter Maydell [Mon, 29 Apr 2019 16:36:03 +0000 (17:36 +0100)] 
target/arm: Implement VLLDM for v7M CPUs with an FPU

Implement the VLLDM instruction for v7M for the FPU present cas.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20190416125744.27770-26-peter.maydell@linaro.org

4 months agotarget/arm: Implement VLSTM for v7M CPUs with an FPU
Peter Maydell [Mon, 29 Apr 2019 16:36:03 +0000 (17:36 +0100)] 
target/arm: Implement VLSTM for v7M CPUs with an FPU

Implement the VLSTM instruction for v7M for the FPU present case.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20190416125744.27770-25-peter.maydell@linaro.org