qemu.git
16 months agoMerge remote-tracking branch 'remotes/ericb/tags/pull-bitmaps-2020-08-03' into staging
Peter Maydell [Mon, 3 Aug 2020 14:13:49 +0000 (15:13 +0100)] 
Merge remote-tracking branch 'remotes/ericb/tags/pull-bitmaps-2020-08-03' into staging

bitmaps patches for 2020-08-03

- fix bitmap migration involving read-only bitmap from backing chain

# gpg: Signature made Mon 03 Aug 2020 15:06:51 BST
# gpg:                using RSA key 71C2CC22B1C4602927D2F3AAA7A16B4A2527436A
# gpg: Good signature from "Eric Blake <eblake@redhat.com>" [full]
# gpg:                 aka "Eric Blake (Free Software Programmer) <ebb9@byu.net>" [full]
# gpg:                 aka "[jpeg image of size 6874]" [full]
# Primary key fingerprint: 71C2 CC22 B1C4 6029 27D2  F3AA A7A1 6B4A 2527 436A

* remotes/ericb/tags/pull-bitmaps-2020-08-03:
  iotests/169: Test source cont with backing bmap
  qcow2: Release read-only bitmaps when inactivated

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
16 months agoiotests/169: Test source cont with backing bmap
Max Reitz [Thu, 30 Jul 2020 12:02:34 +0000 (14:02 +0200)] 
iotests/169: Test source cont with backing bmap

Test migrating from a VM with a persistent bitmap in the backing chain,
and then continuing that VM after the migration

Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-Id: <20200730120234.49288-3-mreitz@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
16 months agoqcow2: Release read-only bitmaps when inactivated
Max Reitz [Thu, 30 Jul 2020 12:02:33 +0000 (14:02 +0200)] 
qcow2: Release read-only bitmaps when inactivated

During migration, we release all bitmaps after storing them on disk, as
long as they are (1) stored on disk, (2) not read-only, and (3)
consistent.

(2) seems arbitrary, though.  The reason we do not release them is
because we do not write them, as there is no need to; and then we just
forget about all bitmaps that we have not written to the file.  However,
read-only persistent bitmaps are still in the file and in sync with
their in-memory representation, so we may as well release them just like
any R/W bitmap that we have updated.

It leads to actual problems, too: After migration, letting the source
continue may result in an error if there were any bitmaps on read-only
nodes (such as backing images), because those have not been released by
bdrv_inactive_all(), but bdrv_invalidate_cache_all() attempts to reload
them (which fails, because they are still present in memory).

Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-Id: <20200730120234.49288-2-mreitz@redhat.com>
Tested-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
16 months agoMerge remote-tracking branch 'remotes/armbru/tags/pull-qapi-2020-08-03' into staging
Peter Maydell [Mon, 3 Aug 2020 11:21:57 +0000 (12:21 +0100)] 
Merge remote-tracking branch 'remotes/armbru/tags/pull-qapi-2020-08-03' into staging

QAPI patches patches for 2020-08-03

# gpg: Signature made Mon 03 Aug 2020 10:08:30 BST
# gpg:                using RSA key 354BC8B3D7EB2A6B68674E5F3870B400EB918653
# gpg:                issuer "armbru@redhat.com"
# gpg: Good signature from "Markus Armbruster <armbru@redhat.com>" [full]
# gpg:                 aka "Markus Armbruster <armbru@pond.sub.org>" [full]
# Primary key fingerprint: 354B C8B3 D7EB 2A6B 6867  4E5F 3870 B400 EB91 8653

* remotes/armbru/tags/pull-qapi-2020-08-03:
  schemas: Add vim modeline
  qapi: Delete unwanted indentation of top-level expressions
  qapi/machine.json: Fix missing newline in doc comment

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
16 months agoMerge remote-tracking branch 'remotes/kraxel/tags/seabios-20200803-pull-request'...
Peter Maydell [Mon, 3 Aug 2020 08:54:58 +0000 (09:54 +0100)] 
Merge remote-tracking branch 'remotes/kraxel/tags/seabios-20200803-pull-request' into staging

seabios: update to master snapshot

seabios master branch got a few bugfixes, so update
to a newer snapshot to pick them up for 5.1-rc3.

# gpg: Signature made Mon 03 Aug 2020 06:24:17 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/seabios-20200803-pull-request:
  seabios: update to master snapshot

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
16 months agoschemas: Add vim modeline
Andrea Bolognani [Wed, 29 Jul 2020 18:50:24 +0000 (20:50 +0200)] 
schemas: Add vim modeline

The various schemas included in QEMU use a JSON-based format which
is, however, strictly speaking not valid JSON.

As a consequence, when vim tries to apply syntax highlight rules
for JSON (as guessed from the file name), the result is an unreadable
mess which mostly consist of red markers pointing out supposed errors
in, well, pretty much everything.

Using Python syntax highlighting produces much better results, and
in fact these files already start with specially-formatted comments
that instruct Emacs to process them as if they were Python files.

This commit adds the equivalent special comments for vim.

Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Message-Id: <20200729185024.121766-1-abologna@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
16 months agoqapi: Delete unwanted indentation of top-level expressions
Markus Armbruster [Thu, 30 Jul 2020 09:16:56 +0000 (11:16 +0200)] 
qapi: Delete unwanted indentation of top-level expressions

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20200730091656.2633334-1-armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
[One more line de-indented]

16 months agoqapi/machine.json: Fix missing newline in doc comment
Peter Maydell [Wed, 29 Jul 2020 19:10:19 +0000 (20:10 +0100)] 
qapi/machine.json: Fix missing newline in doc comment

In commit 176d2cda0dee9f4 we added the @die-id field
to the CpuInstanceProperties struct, but in the process
accidentally removed the newline between the doc-comment
lines for @core-id and @thread-id.

Put the newline back in; this fixes a misformatting in the
generated HTML QMP reference manual.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Message-Id: <20200729191019.19168-1-peter.maydell@linaro.org>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
16 months agoseabios: update to master snapshot
Gerd Hoffmann [Mon, 3 Aug 2020 05:00:58 +0000 (07:00 +0200)] 
seabios: update to master snapshot

seabios master branch got a few bugfixes, so update
to a newer snapshot to pick them up for 5.1-rc3.

shortlog
========

Kevin O'Connor (2):
      vgabios: Fix preserve memory flag in handle_1000
      ldnoexec: Add script to remove ET_EXEC flag from intermediate build objects

Paul Menzel (1):
      nvme: Increase `nvme_cmd_readwrite()` message log level from 3 to 5

Stefan Reiter (1):
      virtio-scsi: fix boot prio detection by using correct lun

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
16 months agoMerge remote-tracking branch 'remotes/cohuck/tags/s390x-20200731' into staging
Peter Maydell [Fri, 31 Jul 2020 09:28:12 +0000 (10:28 +0100)] 
Merge remote-tracking branch 'remotes/cohuck/tags/s390x-20200731' into staging

Fix a problem introduced in a recent fix.

# gpg: Signature made Fri 31 Jul 2020 09:50:28 BST
# gpg:                using RSA key C3D0D66DC3624FF6A8C018CEDECF6B93C6F02FAF
# gpg:                issuer "cohuck@redhat.com"
# gpg: Good signature from "Cornelia Huck <conny@cornelia-huck.de>" [marginal]
# gpg:                 aka "Cornelia Huck <huckc@linux.vnet.ibm.com>" [full]
# gpg:                 aka "Cornelia Huck <cornelia.huck@de.ibm.com>" [full]
# gpg:                 aka "Cornelia Huck <cohuck@kernel.org>" [marginal]
# gpg:                 aka "Cornelia Huck <cohuck@redhat.com>" [marginal]
# Primary key fingerprint: C3D0 D66D C362 4FF6 A8C0  18CE DECF 6B93 C6F0 2FAF

* remotes/cohuck/tags/s390x-20200731:
  s390x/s390-virtio-ccw: fix off-by-one in loadparm getter

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
16 months agoMerge remote-tracking branch 'remotes/stefanha/tags/tracing-pull-request' into staging
Peter Maydell [Thu, 30 Jul 2020 19:02:26 +0000 (20:02 +0100)] 
Merge remote-tracking branch 'remotes/stefanha/tags/tracing-pull-request' into staging

Pull request

A build fix and a 'simple' trace backend regression fix.

# gpg: Signature made Thu 30 Jul 2020 19:13:10 BST
# gpg:                using RSA key 8695A8BFD3F97CDAAC35775A9CA4ABB381AB73C8
# 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:
  tracetool: carefully define SDT_USE_VARIADIC
  trace/simple: Allow enabling simple traces from command line

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
16 months agotracetool: carefully define SDT_USE_VARIADIC
Stefan Hajnoczi [Wed, 29 Jul 2020 15:39:26 +0000 (16:39 +0100)] 
tracetool: carefully define SDT_USE_VARIADIC

The dtrace backend defines SDT_USE_VARIADIC as a workaround for a
conflict with a LTTng UST header file, which requires SDT_USE_VARIADIC
to be defined.

LTTng UST <lttng/tracepoint.h> breaks if included after generated dtrace
headers because SDT_USE_VARIADIC will already be defined:

  #ifdef LTTNG_UST_HAVE_SDT_INTEGRATION
  #define SDT_USE_VARIADIC <-- error, it's already defined
  #include <sys/sdt.h>

Be more careful when defining SDT_USE_VARIADIC. This fixes the build
when both the dtrace and ust tracers are enabled at the same time.

Fixes: 27e08bab94f7c6ebe0b75938c98c394c969e3fd8 ("tracetool: work around ust <sys/sdt.h> include conflict")
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20200729153926.127083-1-stefanha@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
16 months agos390x/s390-virtio-ccw: fix off-by-one in loadparm getter
Halil Pasic [Thu, 30 Jul 2020 13:01:56 +0000 (15:01 +0200)] 
s390x/s390-virtio-ccw: fix off-by-one in loadparm getter

As pointed out by Peter, g_memdup(ms->loadparm, sizeof(ms->loadparm) + 1)
reads one past of the end of ms->loadparm, so g_memdup() can not be used
here.

Let's use g_strndup instead!

Fixes: d6645483285f ("s390x/s390-virtio-ccw: fix loadparm property getter")
Fixes: Coverity CID 1431058
Reported-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Halil Pasic <pasic@linux.ibm.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-Id: <20200730130156.35063-1-pasic@linux.ibm.com>
Signed-off-by: Cornelia Huck <cohuck@redhat.com>
16 months agotrace/simple: Allow enabling simple traces from command line
Josh DuBois [Thu, 23 Jul 2020 05:33:59 +0000 (00:33 -0500)] 
trace/simple: Allow enabling simple traces from command line

The simple trace backend is enabled / disabled with a call
to st_set_trace_file_enabled().  When initializing tracing
from the command-line, this must be enabled on startup.
(Prior to db25d56c014aa1a9, command-line initialization of
simple trace worked because every call to st_set_trace_file
enabled tracing.)

Fixes: db25d56c014aa1a96319c663e0a60346a223b31e
Signed-off-by: Josh DuBois <josh@joshdubois.com>
Message-id: 20200723053359.256928-1-josh@joshdubois.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
16 months agoUpdate version for v5.1.0-rc2 release v5.1.0-rc2
Peter Maydell [Tue, 28 Jul 2020 20:51:03 +0000 (21:51 +0100)] 
Update version for v5.1.0-rc2 release

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
16 months agoMerge remote-tracking branch 'remotes/ericb/tags/pull-nbd-2020-07-28' into staging
Peter Maydell [Tue, 28 Jul 2020 19:43:03 +0000 (20:43 +0100)] 
Merge remote-tracking branch 'remotes/ericb/tags/pull-nbd-2020-07-28' into staging

nbd patches for 2020-07-28

- fix NBD handling of trim/zero requests larger than 2G
- allow no-op resizes on NBD (in turn fixing qemu-img convert -c into NBD)
- several deadlock fixes when using NBD reconnect

# gpg: Signature made Tue 28 Jul 2020 15:59:42 BST
# gpg:                using RSA key 71C2CC22B1C4602927D2F3AAA7A16B4A2527436A
# gpg: Good signature from "Eric Blake <eblake@redhat.com>" [full]
# gpg:                 aka "Eric Blake (Free Software Programmer) <ebb9@byu.net>" [full]
# gpg:                 aka "[jpeg image of size 6874]" [full]
# Primary key fingerprint: 71C2 CC22 B1C4 6029 27D2  F3AA A7A1 6B4A 2527 436A

* remotes/ericb/tags/pull-nbd-2020-07-28:
  block/nbd: nbd_co_reconnect_loop(): don't sleep if drained
  block/nbd: on shutdown terminate connection attempt
  block/nbd: allow drain during reconnect attempt
  block/nbd: split nbd_establish_connection out of nbd_client_connect
  iotests: Test convert to qcow2 compressed to NBD
  iotests: Add more qemu_img helpers
  iotests: Make qemu_nbd_popen() a contextmanager
  block: nbd: Fix convert qcow2 compressed to nbd
  nbd: Fix large trim/zero requests

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
16 months agoMerge remote-tracking branch 'remotes/elmarco/tags/slirp-pull-request' into staging
Peter Maydell [Tue, 28 Jul 2020 18:44:03 +0000 (19:44 +0100)] 
Merge remote-tracking branch 'remotes/elmarco/tags/slirp-pull-request' into staging

slirp: update to latest stable-4.2 branch

# gpg: Signature made Tue 28 Jul 2020 15:30:09 BST
# gpg:                using RSA key 87A9BD933F87C606D276F62DDAE8E10975969CE5
# gpg:                issuer "marcandre.lureau@redhat.com"
# gpg: Good signature from "Marc-André Lureau <marcandre.lureau@redhat.com>" [full]
# gpg:                 aka "Marc-André Lureau <marcandre.lureau@gmail.com>" [full]
# Primary key fingerprint: 87A9 BD93 3F87 C606 D276  F62D DAE8 E109 7596 9CE5

* remotes/elmarco/tags/slirp-pull-request:
  slirp: update to latest stable-4.2 branch
  test-char: abort on serial test error

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
16 months agoMerge remote-tracking branch 'remotes/pmaydell/tags/pull-target-arm-20200727' into...
Peter Maydell [Tue, 28 Jul 2020 17:43:48 +0000 (18:43 +0100)] 
Merge remote-tracking branch 'remotes/pmaydell/tags/pull-target-arm-20200727' into staging

target-arm queue:
 * ACPI: Assert that we don't run out of the preallocated memory
 * hw/misc/aspeed_sdmc: Fix incorrect memory size
 * target/arm: Always pass cacheattr in S1_ptw_translate
 * docs/system/arm/virt: Document 'mte' machine option
 * hw/arm/boot: Fix PAUTH, MTE for EL3 direct kernel boot
 * target/arm: Improve IMPDEF algorithm for IRG

# gpg: Signature made Mon 27 Jul 2020 16:18:38 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-20200727:
  target/arm: Improve IMPDEF algorithm for IRG
  hw/arm/boot: Fix MTE for EL3 direct kernel boot
  hw/arm/boot: Fix PAUTH for EL3 direct kernel boot
  docs/system/arm/virt: Document 'mte' machine option
  target/arm: Always pass cacheattr in S1_ptw_translate
  hw/misc/aspeed_sdmc: Fix incorrect memory size
  ACPI: Assert that we don't run out of the preallocated memory

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
16 months agoMerge remote-tracking branch 'remotes/maxreitz/tags/pull-block-2020-07-28' into staging
Peter Maydell [Tue, 28 Jul 2020 17:00:21 +0000 (18:00 +0100)] 
Merge remote-tracking branch 'remotes/maxreitz/tags/pull-block-2020-07-28' into staging

Block patches for 5.1.0:
- Fix block I/O for split transfers
- Fix iotest 197 for non-qcow2 formats

# gpg: Signature made Tue 28 Jul 2020 14:45:28 BST
# gpg:                using RSA key 91BEB60A30DB3E8857D11829F407DB0061D5CF40
# gpg:                issuer "mreitz@redhat.com"
# gpg: Good signature from "Max Reitz <mreitz@redhat.com>" [full]
# Primary key fingerprint: 91BE B60A 30DB 3E88 57D1  1829 F407 DB00 61D5 CF40

* remotes/maxreitz/tags/pull-block-2020-07-28:
  iotests/197: Fix for non-qcow2 formats
  iotests/028: Add test for cross-base-EOF reads
  block: Fix bdrv_aligned_p*v() for qiov_offset != 0

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
16 months agoMerge remote-tracking branch 'remotes/vivier2/tags/linux-user-for-5.1-pull-request...
Peter Maydell [Tue, 28 Jul 2020 16:15:09 +0000 (17:15 +0100)] 
Merge remote-tracking branch 'remotes/vivier2/tags/linux-user-for-5.1-pull-request' into staging

linux-user 20200728

Fix "pgb_reserved_va: Assertion `guest_base != 0' failed." error
Fix rt_sigtimedwait() errno
Fix getcwd() errno

# gpg: Signature made Tue 28 Jul 2020 13:34:11 BST
# gpg:                using RSA key CD2F75DDC8E3A4DC2E4F5173F30C38BD3F2FBE3C
# gpg:                issuer "laurent@vivier.eu"
# gpg: Good signature from "Laurent Vivier <lvivier@redhat.com>" [full]
# gpg:                 aka "Laurent Vivier <laurent@vivier.eu>" [full]
# gpg:                 aka "Laurent Vivier (Red Hat) <lvivier@redhat.com>" [full]
# Primary key fingerprint: CD2F 75DD C8E3 A4DC 2E4F  5173 F30C 38BD 3F2F BE3C

* remotes/vivier2/tags/linux-user-for-5.1-pull-request:
  linux-user: Use getcwd syscall directly
  linux-user: Fix syscall rt_sigtimedwait() implementation
  linux-user: Ensure mmap_min_addr is non-zero

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
16 months agoMerge remote-tracking branch 'remotes/jasowang/tags/net-pull-request' into staging
Peter Maydell [Tue, 28 Jul 2020 15:28:22 +0000 (16:28 +0100)] 
Merge remote-tracking branch 'remotes/jasowang/tags/net-pull-request' into staging

Want to send earlier but most patches just come.

- fix vhost-vdpa issues when no peer
- fix virtio-pci queue enabling index value
- forbid reentrant RX

Changes from V1:

- drop the patch that has been merged

# gpg: Signature made Tue 28 Jul 2020 09:59:41 BST
# gpg:                using RSA key EF04965B398D6211
# gpg: Good signature from "Jason Wang (Jason Wang on RedHat) <jasowang@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: 215D 46F4 8246 689E C77F  3562 EF04 965B 398D 6211

* remotes/jasowang/tags/net-pull-request:
  net: forbid the reentrant RX
  virtio-net: check the existence of peer before accessing vDPA config
  virtio-pci: fix wrong index in virtio_pci_queue_enabled

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
16 months agoblock/nbd: nbd_co_reconnect_loop(): don't sleep if drained
Vladimir Sementsov-Ogievskiy [Mon, 27 Jul 2020 18:47:50 +0000 (21:47 +0300)] 
block/nbd: nbd_co_reconnect_loop(): don't sleep if drained

We try to go to wakeable sleep, so that, if drain begins it will break
the sleep. But what if nbd_client_co_drain_begin() already called and
s->drained is already true? We'll go to sleep, and drain will have to
wait for the whole timeout. Let's improve it.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20200727184751.15704-5-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
16 months agoblock/nbd: on shutdown terminate connection attempt
Vladimir Sementsov-Ogievskiy [Mon, 27 Jul 2020 18:47:49 +0000 (21:47 +0300)] 
block/nbd: on shutdown terminate connection attempt

On shutdown nbd driver may be in a connecting state. We should shutdown
it as well, otherwise we may hang in
nbd_teardown_connection, waiting for conneciton_co to finish in
BDRV_POLL_WHILE(bs, s->connection_co) loop if remote server is down.

How to reproduce the dead lock:

1. Create nbd-fault-injector.conf with the following contents:

[inject-error "mega1"]
event=data
io=readwrite
when=before

2. In one terminal run nbd-fault-injector in a loop, like this:

n=1; while true; do
    echo $n; ((n++));
    ./nbd-fault-injector.py 127.0.0.1:10000 nbd-fault-injector.conf;
done

3. In another terminal run qemu-io in a loop, like this:

n=1; while true; do
    echo $n; ((n++));
    ./qemu-io -c 'read 0 512' nbd://127.0.0.1:10000;
done

After some time, qemu-io will hang. Note, that this hang may be
triggered by another bug, so the whole case is fixed only together with
commit "block/nbd: allow drain during reconnect attempt".

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20200727184751.15704-4-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
16 months agoblock/nbd: allow drain during reconnect attempt
Vladimir Sementsov-Ogievskiy [Mon, 27 Jul 2020 18:47:48 +0000 (21:47 +0300)] 
block/nbd: allow drain during reconnect attempt

It should be safe to reenter qio_channel_yield() on io/channel read/write
path, so it's safe to reduce in_flight and allow attaching new aio
context. And no problem to allow drain itself: connection attempt is
not a guest request. Moreover, if remote server is down, we can hang
in negotiation, blocking drain section and provoking a dead lock.

How to reproduce the dead lock:

1. Create nbd-fault-injector.conf with the following contents:

[inject-error "mega1"]
event=data
io=readwrite
when=before

2. In one terminal run nbd-fault-injector in a loop, like this:

n=1; while true; do
    echo $n; ((n++));
    ./nbd-fault-injector.py 127.0.0.1:10000 nbd-fault-injector.conf;
done

3. In another terminal run qemu-io in a loop, like this:

n=1; while true; do
    echo $n; ((n++));
    ./qemu-io -c 'read 0 512' nbd://127.0.0.1:10000;
done

After some time, qemu-io will hang trying to drain, for example, like
this:

 #3 aio_poll (ctx=0x55f006bdd890, blocking=true) at
    util/aio-posix.c:600
 #4 bdrv_do_drained_begin (bs=0x55f006bea710, recursive=false,
    parent=0x0, ignore_bds_parents=false, poll=true) at block/io.c:427
 #5 bdrv_drained_begin (bs=0x55f006bea710) at block/io.c:433
 #6 blk_drain (blk=0x55f006befc80) at block/block-backend.c:1710
 #7 blk_unref (blk=0x55f006befc80) at block/block-backend.c:498
 #8 bdrv_open_inherit (filename=0x7fffba1563bc
    "nbd+tcp://127.0.0.1:10000", reference=0x0, options=0x55f006be86d0,
    flags=24578, parent=0x0, child_class=0x0, child_role=0,
    errp=0x7fffba154620) at block.c:3491
 #9 bdrv_open (filename=0x7fffba1563bc "nbd+tcp://127.0.0.1:10000",
    reference=0x0, options=0x0, flags=16386, errp=0x7fffba154620) at
    block.c:3513
 #10 blk_new_open (filename=0x7fffba1563bc "nbd+tcp://127.0.0.1:10000",
    reference=0x0, options=0x0, flags=16386, errp=0x7fffba154620) at
    block/block-backend.c:421

And connection_co stack like this:

 #0 qemu_coroutine_switch (from_=0x55f006bf2650, to_=0x7fe96e07d918,
    action=COROUTINE_YIELD) at util/coroutine-ucontext.c:302
 #1 qemu_coroutine_yield () at util/qemu-coroutine.c:193
 #2 qio_channel_yield (ioc=0x55f006bb3c20, condition=G_IO_IN) at
    io/channel.c:472
 #3 qio_channel_readv_all_eof (ioc=0x55f006bb3c20, iov=0x7fe96d729bf0,
    niov=1, errp=0x7fe96d729eb0) at io/channel.c:110
 #4 qio_channel_readv_all (ioc=0x55f006bb3c20, iov=0x7fe96d729bf0,
    niov=1, errp=0x7fe96d729eb0) at io/channel.c:143
 #5 qio_channel_read_all (ioc=0x55f006bb3c20, buf=0x7fe96d729d28
    "\300.\366\004\360U", buflen=8, errp=0x7fe96d729eb0) at
    io/channel.c:247
 #6 nbd_read (ioc=0x55f006bb3c20, buffer=0x7fe96d729d28, size=8,
    desc=0x55f004f69644 "initial magic", errp=0x7fe96d729eb0) at
    /work/src/qemu/master/include/block/nbd.h:365
 #7 nbd_read64 (ioc=0x55f006bb3c20, val=0x7fe96d729d28,
    desc=0x55f004f69644 "initial magic", errp=0x7fe96d729eb0) at
    /work/src/qemu/master/include/block/nbd.h:391
 #8 nbd_start_negotiate (aio_context=0x55f006bdd890,
    ioc=0x55f006bb3c20, tlscreds=0x0, hostname=0x0,
    outioc=0x55f006bf19f8, structured_reply=true,
    zeroes=0x7fe96d729dca, errp=0x7fe96d729eb0) at nbd/client.c:904
 #9 nbd_receive_negotiate (aio_context=0x55f006bdd890,
    ioc=0x55f006bb3c20, tlscreds=0x0, hostname=0x0,
    outioc=0x55f006bf19f8, info=0x55f006bf1a00, errp=0x7fe96d729eb0) at
    nbd/client.c:1032
 #10 nbd_client_connect (bs=0x55f006bea710, errp=0x7fe96d729eb0) at
    block/nbd.c:1460
 #11 nbd_reconnect_attempt (s=0x55f006bf19f0) at block/nbd.c:287
 #12 nbd_co_reconnect_loop (s=0x55f006bf19f0) at block/nbd.c:309
 #13 nbd_connection_entry (opaque=0x55f006bf19f0) at block/nbd.c:360
 #14 coroutine_trampoline (i0=113190480, i1=22000) at
    util/coroutine-ucontext.c:173

Note, that the hang may be
triggered by another bug, so the whole case is fixed only together with
commit "block/nbd: on shutdown terminate connection attempt".

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20200727184751.15704-3-vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
16 months agoblock/nbd: split nbd_establish_connection out of nbd_client_connect
Vladimir Sementsov-Ogievskiy [Mon, 27 Jul 2020 18:47:47 +0000 (21:47 +0300)] 
block/nbd: split nbd_establish_connection out of nbd_client_connect

We are going to implement non-blocking version of
nbd_establish_connection, which for a while will be used only for
nbd_reconnect_attempt, not for nbd_open, so we need to call it
separately.

Refactor nbd_reconnect_attempt in a way which makes next commit
simpler.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20200727184751.15704-2-vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
16 months agoiotests: Test convert to qcow2 compressed to NBD
Nir Soffer [Mon, 27 Jul 2020 21:58:46 +0000 (00:58 +0300)] 
iotests: Test convert to qcow2 compressed to NBD

Add test for "qemu-img convert -O qcow2 -c" to NBD target. The tests    
create a OVA file and write compressed qcow2 disk content directly into
the OVA file via qemu-nbd.

Signed-off-by: Nir Soffer <nsoffer@redhat.com>
Message-Id: <20200727215846.395443-5-nsoffer@redhat.com>
Tested-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
16 months agoiotests: Add more qemu_img helpers
Nir Soffer [Mon, 27 Jul 2020 21:58:45 +0000 (00:58 +0300)] 
iotests: Add more qemu_img helpers

Add 2 helpers for measuring and checking images:
- qemu_img_measure()
- qemu_img_check()

Both use --output-json and parse the returned json to make easy to use
in other tests. I'm going to use them in a new test, and I hope they
will be useful in may other tests.

Signed-off-by: Nir Soffer <nsoffer@redhat.com>
Message-Id: <20200727215846.395443-4-nsoffer@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
16 months agoiotests: Make qemu_nbd_popen() a contextmanager
Nir Soffer [Mon, 27 Jul 2020 21:58:44 +0000 (00:58 +0300)] 
iotests: Make qemu_nbd_popen() a contextmanager

Instead of duplicating the code to wait until the server is ready and
remember to terminate the server and wait for it, make it possible to
use like this:

    with qemu_nbd_popen('-k', sock, image):
        # Access image via qemu-nbd socket...

Only test 264 used this helper, but I had to modify the output since it
did not consistently when starting and stopping qemu-nbd.

Signed-off-by: Nir Soffer <nsoffer@redhat.com>
Message-Id: <20200727215846.395443-3-nsoffer@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
16 months agoblock: nbd: Fix convert qcow2 compressed to nbd
Nir Soffer [Mon, 27 Jul 2020 21:58:43 +0000 (00:58 +0300)] 
block: nbd: Fix convert qcow2 compressed to nbd

When converting to qcow2 compressed format, the last step is a special
zero length compressed write, ending in a call to bdrv_co_truncate(). This
call always fails for the nbd driver since it does not implement
bdrv_co_truncate().

For block devices, which have the same limits, the call succeeds since
the file driver implements bdrv_co_truncate(). If the caller asked to
truncate to the same or smaller size with exact=false, the truncate
succeeds. Implement the same logic for nbd.

Example failing without this change:

In one shell start qemu-nbd:

$ truncate -s 1g test.tar
$ qemu-nbd --socket=/tmp/nbd.sock --persistent --format=raw --offset 1536 test.tar

In another shell convert an image to qcow2 compressed via NBD:

$ echo "disk data" > disk.raw
$ truncate -s 1g disk.raw
$ qemu-img convert -f raw -O qcow2 -c disk1.raw nbd+unix:///?socket=/tmp/nbd.sock; echo $?
1

qemu-img failed, but the conversion was successful:

$ qemu-img info nbd+unix:///?socket=/tmp/nbd.sock
image: nbd+unix://?socket=/tmp/nbd.sock
file format: qcow2
virtual size: 1 GiB (1073741824 bytes)
...

$ qemu-img check nbd+unix:///?socket=/tmp/nbd.sock
No errors were found on the image.
1/16384 = 0.01% allocated, 100.00% fragmented, 100.00% compressed clusters
Image end offset: 393216

$ qemu-img compare disk.raw nbd+unix:///?socket=/tmp/nbd.sock
Images are identical.

Fixes: https://bugzilla.redhat.com/1860627
Signed-off-by: Nir Soffer <nsoffer@redhat.com>
Message-Id: <20200727215846.395443-2-nsoffer@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
[eblake: typo fixes]
Signed-off-by: Eric Blake <eblake@redhat.com>
16 months agoslirp: update to latest stable-4.2 branch
Marc-André Lureau [Sun, 26 Jul 2020 15:41:00 +0000 (19:41 +0400)] 
slirp: update to latest stable-4.2 branch

Dr. David Alan Gilbert (1):
      ip_stripoptions use memmove

Jindrich Novy (4):
      Fix possible infinite loops and use-after-free
      Use secure string copy to avoid overflow
      Be sure to initialize sockaddr structure
      Check lseek() for failure

Marc-André Lureau (2):
      util: do not silently truncate
      Merge branch 'stable-4.2' into 'stable-4.2'

Philippe Mathieu-Daudé (3):
      Fix win32 builds by using the SLIRP_PACKED definition
      Fix constness warnings
      Remove unnecessary break

Ralf Haferkamp (2):
      Drop bogus IPv6 messages
      Fix MTU check

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
16 months agotest-char: abort on serial test error
Marc-André Lureau [Sun, 26 Jul 2020 15:30:23 +0000 (19:30 +0400)] 
test-char: abort on serial test error

We are having issues debugging and bisecting this issue that happen
mostly on patchew. Let's make it abort where it failed to gather some
new informations.

Suggested-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
16 months agoMerge remote-tracking branch 'remotes/mdroth/tags/qga-pull-2020-07-27-tag' into staging
Peter Maydell [Tue, 28 Jul 2020 14:24:31 +0000 (15:24 +0100)] 
Merge remote-tracking branch 'remotes/mdroth/tags/qga-pull-2020-07-27-tag' into staging

qemu-ga patch queue for hard-freeze

* document use of -1 when pci_controller field can't be retrieved for
  guest-get-fsinfo
* fix incorrect filesystem type reporting on w32 for guest-get-fsinfo
  when a volume is not mounted

# gpg: Signature made Tue 28 Jul 2020 00:16:50 BST
# gpg:                using RSA key CEACC9E15534EBABB82D3FA03353C9CEF108B584
# gpg:                issuer "mdroth@linux.vnet.ibm.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-2020-07-27-tag:
  qga/qapi-schema: Document -1 for invalid PCI address fields
  qga-win: fix "guest-get-fsinfo" wrong filesystem type

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
16 months agonbd: Fix large trim/zero requests
Eric Blake [Wed, 22 Jul 2020 21:22:31 +0000 (16:22 -0500)] 
nbd: Fix large trim/zero requests

Although qemu as NBD client limits requests to <2G, the NBD protocol
allows clients to send requests almost all the way up to 4G.  But
because our block layer is not yet 64-bit clean, we accidentally wrap
such requests into a negative size, and fail with EIO instead of
performing the intended operation.

The bug is visible in modern systems with something as simple as:

$ qemu-img create -f qcow2 /tmp/image.img 5G
$ sudo qemu-nbd --connect=/dev/nbd0 /tmp/image.img
$ sudo blkdiscard /dev/nbd0

or with user-space only:

$ truncate --size=3G file
$ qemu-nbd -f raw file
$ nbdsh -u nbd://localhost:10809 -c 'h.trim(3*1024*1024*1024,0)'

Although both blk_co_pdiscard and blk_pwrite_zeroes currently return 0
on success, this is also a good time to fix our code to a more robust
paradigm that treats all non-negative values as success.

Alas, our iotests do not currently make it easy to add external
dependencies on blkdiscard or nbdsh, so we have to rely on manual
testing for now.

This patch can be reverted when we later improve the overall block
layer to be 64-bit clean, but for now, a minimal fix was deemed less
risky prior to release.

CC: qemu-stable@nongnu.org
Fixes: 1f4d6d18ed
Fixes: 1c6c4bb7f0
Fixes: https://github.com/systemd/systemd/issues/16242
Signed-off-by: Eric Blake <eblake@redhat.com>
Message-Id: <20200722212231.535072-1-eblake@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
[eblake: rework success tests to use >=0]

16 months agoMerge remote-tracking branch 'remotes/ericb/tags/pull-bitmaps-2020-07-27' into staging
Peter Maydell [Tue, 28 Jul 2020 13:38:17 +0000 (14:38 +0100)] 
Merge remote-tracking branch 'remotes/ericb/tags/pull-bitmaps-2020-07-27' into staging

bitmaps patches for 2020-07-27

- Improve handling of various post-copy bitmap migration scenarios. A lost
bitmap should merely mean that the next backup must be full rather than
incremental, rather than abruptly breaking the entire guest migration.
- Associated iotest improvements

# gpg: Signature made Mon 27 Jul 2020 21:46:17 BST
# gpg:                using RSA key 71C2CC22B1C4602927D2F3AAA7A16B4A2527436A
# gpg: Good signature from "Eric Blake <eblake@redhat.com>" [full]
# gpg:                 aka "Eric Blake (Free Software Programmer) <ebb9@byu.net>" [full]
# gpg:                 aka "[jpeg image of size 6874]" [full]
# Primary key fingerprint: 71C2 CC22 B1C4 6029 27D2  F3AA A7A1 6B4A 2527 436A

* remotes/ericb/tags/pull-bitmaps-2020-07-27: (24 commits)
  migration: Fix typos in bitmap migration comments
  iotests: Adjust which migration tests are quick
  qemu-iotests/199: add source-killed case to bitmaps postcopy
  qemu-iotests/199: add early shutdown case to bitmaps postcopy
  qemu-iotests/199: check persistent bitmaps
  qemu-iotests/199: prepare for new test-cases addition
  migration/savevm: don't worry if bitmap migration postcopy failed
  migration/block-dirty-bitmap: cancel migration on shutdown
  migration/block-dirty-bitmap: relax error handling in incoming part
  migration/block-dirty-bitmap: keep bitmap state for all bitmaps
  migration/block-dirty-bitmap: simplify dirty_bitmap_load_complete
  migration/block-dirty-bitmap: rename finish_lock to just lock
  migration/block-dirty-bitmap: refactor state global variables
  migration/block-dirty-bitmap: move mutex init to dirty_bitmap_mig_init
  migration/block-dirty-bitmap: rename dirty_bitmap_mig_cleanup
  migration/block-dirty-bitmap: rename state structure types
  migration/block-dirty-bitmap: fix dirty_bitmap_mig_before_vm_start
  qemu-iotests/199: increase postcopy period
  qemu-iotests/199: change discard patterns
  qemu-iotests/199: improve performance: set bitmap by discard
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
16 months agoiotests/197: Fix for non-qcow2 formats
Max Reitz [Tue, 28 Jul 2020 13:11:34 +0000 (15:11 +0200)] 
iotests/197: Fix for non-qcow2 formats

While 197 is very much a qcow2 test, and it looks like the partial
cluster case at the end (introduced in b0ddcbbb36a66a6) is specifically
a qcow2 case, the whole test scripts actually marks itself to work with
generic formats (and generic protocols, even).

Said partial cluster case happened to work with non-qcow2 formats as
well (mostly by accident), but 1855536256 broke that, because it sets
the compat option, which does not work for non-qcow2 formats.

So go the whole way and force IMGFMT=qcow2 and IMGPROTO=file, as done in
other places in this test.

Fixes: 1855536256eb0a5708b04b85f744de69559ea323
       ("iotests/197: Fix for compat=0.10")
Reported-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-Id: <20200728131134.902519-1-mreitz@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
16 months agoiotests/028: Add test for cross-base-EOF reads
Max Reitz [Tue, 28 Jul 2020 12:08:05 +0000 (14:08 +0200)] 
iotests/028: Add test for cross-base-EOF reads

Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-Id: <20200728120806.265916-3-mreitz@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Tested-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Tested-by: Claudio Fontana <cfontana@suse.de>
16 months agoblock: Fix bdrv_aligned_p*v() for qiov_offset != 0
Max Reitz [Tue, 28 Jul 2020 12:08:04 +0000 (14:08 +0200)] 
block: Fix bdrv_aligned_p*v() for qiov_offset != 0

Since these functions take a @qiov_offset, they must always take it into
account when working with @qiov.  There are a couple of places where
they do not, but they should.

Fixes: 65cd4424b9df03bb5195351c33e04cbbecc0705c
       ("block/io: bdrv_aligned_preadv: use and support qiov_offset")
Fixes: 28c4da28695bdbe04b336b2c9c463876cc3aaa6d
       ("block/io: bdrv_aligned_pwritev: use and support qiov_offset")
Reported-by: Claudio Fontana <cfontana@suse.de>
Reported-by: Bruce Rogers <brogers@suse.com>
Cc: qemu-stable@nongnu.org
Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-Id: <20200728120806.265916-2-mreitz@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Tested-by: Claudio Fontana <cfontana@suse.de>
Tested-by: Bruce Rogers <brogers@suse.com>
16 months agoMerge remote-tracking branch 'remotes/mcayland/tags/qemu-openbios-20200727' into...
Peter Maydell [Tue, 28 Jul 2020 12:46:31 +0000 (13:46 +0100)] 
Merge remote-tracking branch 'remotes/mcayland/tags/qemu-openbios-20200727' into staging

qemu-openbios queue

# gpg: Signature made Mon 27 Jul 2020 16:09:34 BST
# gpg:                using RSA key CC621AB98E82200D915CC9C45BC2C56FAE0F321F
# gpg:                issuer "mark.cave-ayland@ilande.co.uk"
# gpg: Good signature from "Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>" [full]
# Primary key fingerprint: CC62 1AB9 8E82 200D 915C  C9C4 5BC2 C56F AE0F 321F

* remotes/mcayland/tags/qemu-openbios-20200727:
  Update OpenBIOS images to 7f28286f built from submodule.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
16 months agonet: forbid the reentrant RX
Jason Wang [Wed, 22 Jul 2020 08:57:46 +0000 (16:57 +0800)] 
net: forbid the reentrant RX

The memory API allows DMA into NIC's MMIO area. This means the NIC's
RX routine must be reentrant. Instead of auditing all the NIC, we can
simply detect the reentrancy and return early. The queue->delivering
is set and cleared by qemu_net_queue_deliver() for other queue helpers
to know whether the delivering in on going (NIC's receive is being
called). We can check it and return early in qemu_net_queue_flush() to
forbid reentrant RX.

Signed-off-by: Jason Wang <jasowang@redhat.com>
16 months agovirtio-net: check the existence of peer before accessing vDPA config
Jason Wang [Sat, 25 Jul 2020 00:13:17 +0000 (08:13 +0800)] 
virtio-net: check the existence of peer before accessing vDPA config

We try to check whether a peer is VDPA in order to get config from
there - with no peer, this leads to a NULL
pointer dereference. Add a check before trying to access the peer
type. No peer means not VDPA.

Fixes: 108a64818e69b ("vhost-vdpa: introduce vhost-vdpa backend")
Cc: Cindy Lu <lulu@redhat.com>
Tested-by: Cornelia Huck <cohuck@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
16 months agovirtio-pci: fix wrong index in virtio_pci_queue_enabled
Yuri Benditovich [Mon, 27 Jul 2020 14:38:07 +0000 (17:38 +0300)] 
virtio-pci: fix wrong index in virtio_pci_queue_enabled

We should use the index passed by the caller instead of the queue_sel
when checking the enablement of a specific virtqueue.

This is reported in https://bugzilla.redhat.com/show_bug.cgi?id=1702608

Fixes: f19bcdfedd53 ("virtio-pci: implement queue_enabled method")
Signed-off-by: Yuri Benditovich <yuri.benditovich@daynix.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
16 months agoMerge remote-tracking branch 'remotes/maxreitz/tags/pull-block-2020-07-27' into staging
Peter Maydell [Tue, 28 Jul 2020 08:15:44 +0000 (09:15 +0100)] 
Merge remote-tracking branch 'remotes/maxreitz/tags/pull-block-2020-07-27' into staging

Block patches for 5.1:
- Coverity fix
- iotests fix for rx and avr
- iotests fix for qcow2 -o compat=0.10

# gpg: Signature made Mon 27 Jul 2020 15:36:23 BST
# gpg:                using RSA key 91BEB60A30DB3E8857D11829F407DB0061D5CF40
# gpg:                issuer "mreitz@redhat.com"
# gpg: Good signature from "Max Reitz <mreitz@redhat.com>" [full]
# Primary key fingerprint: 91BE B60A 30DB 3E88 57D1  1829 F407 DB00 61D5 CF40

* remotes/maxreitz/tags/pull-block-2020-07-27:
  iotests/197: Fix for compat=0.10
  iotests: Select a default machine for the rx and avr targets
  block/amend: Check whether the node exists

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
16 months agoqga/qapi-schema: Document -1 for invalid PCI address fields
Thomas Huth [Wed, 22 Jul 2020 04:40:25 +0000 (06:40 +0200)] 
qga/qapi-schema: Document -1 for invalid PCI address fields

The "guest-get-fsinfo" could also be used for non-PCI devices in the
future. And the code in GuestPCIAddress() in qga/commands-win32.c seems
to be using "-1" for fields that it can not determine already. Thus
let's properly document "-1" as value for invalid PCI address fields.

Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Michael Roth <mdroth@linux.vnet.ibm.com>
16 months agoqga-win: fix "guest-get-fsinfo" wrong filesystem type
Basil Salman [Thu, 30 Apr 2020 12:11:40 +0000 (15:11 +0300)] 
qga-win: fix "guest-get-fsinfo" wrong filesystem type

This patch handles the case where unmounted volumes exist,
where in that case GetVolumePathNamesForVolumeName returns
empty path, GetVolumeInformation will use the current working
directory instead.
This patch fixes the issue by opening a handle to the volumes,
and using GetVolumeInformationByHandleW instead.

Buglink: https://bugzilla.redhat.com/show_bug.cgi?id=1746667
Signed-off-by: Basil Salman <bsalman@redhat.com>
Signed-off-by: Basil Salman <basil@daynix.com>
*fix crash when guest_build_fsinfo() sets errp multiple times
*make new error message more distinct from existing ones
Signed-off-by: Michael Roth <mdroth@linux.vnet.ibm.com>
16 months agomigration: Fix typos in bitmap migration comments
Eric Blake [Mon, 27 Jul 2020 20:32:06 +0000 (15:32 -0500)] 
migration: Fix typos in bitmap migration comments

Noticed while reviewing the file for newer patches.

Fixes: b35ebdf076
Signed-off-by: Eric Blake <eblake@redhat.com>
Message-Id: <20200727203206.134996-1-eblake@redhat.com>

16 months agoiotests: Adjust which migration tests are quick
Eric Blake [Mon, 27 Jul 2020 19:51:17 +0000 (14:51 -0500)] 
iotests: Adjust which migration tests are quick

A quick run of './check -qcow2 -g migration' shows that test 169 is
NOT quick, but meanwhile several other tests ARE quick.  Let's adjust
the test designations accordingly.

Signed-off-by: Eric Blake <eblake@redhat.com>
Message-Id: <20200727195117.132151-1-eblake@redhat.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
16 months agoqemu-iotests/199: add source-killed case to bitmaps postcopy
Vladimir Sementsov-Ogievskiy [Mon, 27 Jul 2020 19:42:36 +0000 (22:42 +0300)] 
qemu-iotests/199: add source-killed case to bitmaps postcopy

Previous patches fixes behavior of bitmaps migration, so that errors
are handled by just removing unfinished bitmaps, and not fail or try to
recover postcopy migration. Add corresponding test.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Andrey Shinkevich <andrey.shinkevich@virtuozzo.com>
Tested-by: Eric Blake <eblake@redhat.com>
Message-Id: <20200727194236.19551-22-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
16 months agoqemu-iotests/199: add early shutdown case to bitmaps postcopy
Vladimir Sementsov-Ogievskiy [Mon, 27 Jul 2020 19:42:35 +0000 (22:42 +0300)] 
qemu-iotests/199: add early shutdown case to bitmaps postcopy

Previous patches fixed two crashes which may occur on shutdown prior to
bitmaps postcopy finished. Check that it works now.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Andrey Shinkevich <andrey.shinkevich@virtuozzo.com>
Tested-by: Eric Blake <eblake@redhat.com>
Message-Id: <20200727194236.19551-21-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
16 months agoqemu-iotests/199: check persistent bitmaps
Vladimir Sementsov-Ogievskiy [Mon, 27 Jul 2020 19:42:34 +0000 (22:42 +0300)] 
qemu-iotests/199: check persistent bitmaps

Check that persistent bitmaps are not stored on source and that bitmaps
are persistent on destination.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Andrey Shinkevich <andrey.shinkevich@virtuozzo.com>
Message-Id: <20200727194236.19551-20-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
16 months agoqemu-iotests/199: prepare for new test-cases addition
Vladimir Sementsov-Ogievskiy [Mon, 27 Jul 2020 19:42:33 +0000 (22:42 +0300)] 
qemu-iotests/199: prepare for new test-cases addition

Move future common part to start_postcopy() method. Move checking
number of bitmaps to check_bitmap().

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Andrey Shinkevich <andrey.shinkevich@virtuozzo.com>
Message-Id: <20200727194236.19551-19-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
16 months agomigration/savevm: don't worry if bitmap migration postcopy failed
Vladimir Sementsov-Ogievskiy [Mon, 27 Jul 2020 19:42:32 +0000 (22:42 +0300)] 
migration/savevm: don't worry if bitmap migration postcopy failed

First, if only bitmaps postcopy is enabled (and not ram postcopy)
postcopy_pause_incoming crashes on an assertion
assert(mis->to_src_file).

And anyway, bitmaps postcopy is not prepared to be somehow recovered.
The original idea instead is that if bitmaps postcopy failed, we just
lose some bitmaps, which is not critical. So, on failure we just need
to remove unfinished bitmaps and guest should continue execution on
destination.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Andrey Shinkevich <andrey.shinkevich@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20200727194236.19551-18-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
16 months agomigration/block-dirty-bitmap: cancel migration on shutdown
Vladimir Sementsov-Ogievskiy [Mon, 27 Jul 2020 19:42:31 +0000 (22:42 +0300)] 
migration/block-dirty-bitmap: cancel migration on shutdown

If target is turned off prior to postcopy finished, target crashes
because busy bitmaps are found at shutdown.
Canceling incoming migration helps, as it removes all unfinished (and
therefore busy) bitmaps.

Similarly on source we crash in bdrv_close_all which asserts that all
bdrv states are removed, because bdrv states involved into dirty bitmap
migration are referenced by it. So, we need to cancel outgoing
migration as well.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Andrey Shinkevich <andrey.shinkevich@virtuozzo.com>
Message-Id: <20200727194236.19551-17-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
16 months agomigration/block-dirty-bitmap: relax error handling in incoming part
Vladimir Sementsov-Ogievskiy [Mon, 27 Jul 2020 19:42:30 +0000 (22:42 +0300)] 
migration/block-dirty-bitmap: relax error handling in incoming part

Bitmaps data is not critical, and we should not fail the migration (or
use postcopy recovering) because of dirty-bitmaps migration failure.
Instead we should just lose unfinished bitmaps.

Still we have to report io stream violation errors, as they affect the
whole migration stream.

While touching this, tighten code that was previously blindly calling
malloc on a size read from the migration stream, as a corrupted stream
(perhaps from a malicious user) should not be able to convince us to
allocate an inordinate amount of memory.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20200727194236.19551-16-vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
[eblake: typo fixes, enhance commit message]
Signed-off-by: Eric Blake <eblake@redhat.com>
16 months agomigration/block-dirty-bitmap: keep bitmap state for all bitmaps
Vladimir Sementsov-Ogievskiy [Mon, 27 Jul 2020 19:42:29 +0000 (22:42 +0300)] 
migration/block-dirty-bitmap: keep bitmap state for all bitmaps

Keep bitmap state for disabled bitmaps too. Keep the state until the
end of the process. It's needed for the following commit to implement
bitmap postcopy canceling.

To clean-up the new list the following logic is used:
We need two events to consider bitmap migration finished:
1. chunk with DIRTY_BITMAP_MIG_FLAG_COMPLETE flag should be received
2. dirty_bitmap_mig_before_vm_start should be called
These two events may come in any order, so we understand which one is
last, and on the last of them we remove bitmap migration state from the
list.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Andrey Shinkevich <andrey.shinkevich@virtuozzo.com>
Message-Id: <20200727194236.19551-15-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
16 months agomigration/block-dirty-bitmap: simplify dirty_bitmap_load_complete
Vladimir Sementsov-Ogievskiy [Mon, 27 Jul 2020 19:42:28 +0000 (22:42 +0300)] 
migration/block-dirty-bitmap: simplify dirty_bitmap_load_complete

bdrv_enable_dirty_bitmap_locked() call does nothing, as if we are in
postcopy, bitmap successor must be enabled, and reclaim operation will
enable the bitmap.

So, actually we need just call _reclaim_ in both if branches, and
making differences only to add an assertion seems not really good. The
logic becomes simple: on load complete we do reclaim and that's all.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Andrey Shinkevich <andrey.shinkevich@virtuozzo.com>
Message-Id: <20200727194236.19551-14-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
16 months agomigration/block-dirty-bitmap: rename finish_lock to just lock
Vladimir Sementsov-Ogievskiy [Mon, 27 Jul 2020 19:42:27 +0000 (22:42 +0300)] 
migration/block-dirty-bitmap: rename finish_lock to just lock

finish_lock is bad name, as lock used not only on process end.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Andrey Shinkevich <andrey.shinkevich@virtuozzo.com>
Message-Id: <20200727194236.19551-13-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
16 months agomigration/block-dirty-bitmap: refactor state global variables
Vladimir Sementsov-Ogievskiy [Mon, 27 Jul 2020 19:42:26 +0000 (22:42 +0300)] 
migration/block-dirty-bitmap: refactor state global variables

Move all state variables into one global struct. Reduce global
variable usage, utilizing opaque pointer where possible.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Andrey Shinkevich <andrey.shinkevich@virtuozzo.com>
Message-Id: <20200727194236.19551-12-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
16 months agomigration/block-dirty-bitmap: move mutex init to dirty_bitmap_mig_init
Vladimir Sementsov-Ogievskiy [Mon, 27 Jul 2020 19:42:25 +0000 (22:42 +0300)] 
migration/block-dirty-bitmap: move mutex init to dirty_bitmap_mig_init

No reasons to keep two public init functions.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Andrey Shinkevich <andrey.shinkevich@virtuozzo.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <20200727194236.19551-11-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
16 months agomigration/block-dirty-bitmap: rename dirty_bitmap_mig_cleanup
Vladimir Sementsov-Ogievskiy [Mon, 27 Jul 2020 19:42:24 +0000 (22:42 +0300)] 
migration/block-dirty-bitmap: rename dirty_bitmap_mig_cleanup

Rename dirty_bitmap_mig_cleanup to dirty_bitmap_do_save_cleanup, to
stress that it is on save part.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Andrey Shinkevich <andrey.shinkevich@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20200727194236.19551-10-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
16 months agomigration/block-dirty-bitmap: rename state structure types
Vladimir Sementsov-Ogievskiy [Mon, 27 Jul 2020 19:42:23 +0000 (22:42 +0300)] 
migration/block-dirty-bitmap: rename state structure types

Rename types to be symmetrical for load/save part and shorter.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Andrey Shinkevich <andrey.shinkevich@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20200727194236.19551-9-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
16 months agomigration/block-dirty-bitmap: fix dirty_bitmap_mig_before_vm_start
Vladimir Sementsov-Ogievskiy [Mon, 27 Jul 2020 19:42:22 +0000 (22:42 +0300)] 
migration/block-dirty-bitmap: fix dirty_bitmap_mig_before_vm_start

Using the _locked version of bdrv_enable_dirty_bitmap to bypass locking
is wrong as we do not already own the mutex.  Moreover, the adjacent
call to bdrv_dirty_bitmap_enable_successor grabs the mutex.

Fixes: 58f72b965e9e1q
Cc: qemu-stable@nongnu.org # v3.0
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Andrey Shinkevich <andrey.shinkevich@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20200727194236.19551-8-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
16 months agoqemu-iotests/199: increase postcopy period
Vladimir Sementsov-Ogievskiy [Mon, 27 Jul 2020 19:42:21 +0000 (22:42 +0300)] 
qemu-iotests/199: increase postcopy period

The test wants to force a bitmap postcopy. Still, the resulting
postcopy period is very small. Let's increase it by adding more
bitmaps to migrate. Also, test disabled bitmaps migration.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Andrey Shinkevich <andrey.shinkevich@virtuozzo.com>
Tested-by: Eric Blake <eblake@redhat.com>
Message-Id: <20200727194236.19551-7-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
16 months agoqemu-iotests/199: change discard patterns
Vladimir Sementsov-Ogievskiy [Mon, 27 Jul 2020 19:42:20 +0000 (22:42 +0300)] 
qemu-iotests/199: change discard patterns

iotest 199 works too long because of many discard operations. At the
same time, postcopy period is very short, in spite of all these
efforts.

So, let's use less discards (and with more interesting patterns) to
reduce test timing. In the next commit we'll increase postcopy period.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Andrey Shinkevich <andrey.shinkevich@virtuozzo.com>
Tested-by: Eric Blake <eblake@redhat.com>
Message-Id: <20200727194236.19551-6-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
16 months agoqemu-iotests/199: improve performance: set bitmap by discard
Vladimir Sementsov-Ogievskiy [Mon, 27 Jul 2020 19:42:19 +0000 (22:42 +0300)] 
qemu-iotests/199: improve performance: set bitmap by discard

Discard dirties dirty-bitmap as well as write, but works faster. Let's
use it instead.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Andrey Shinkevich <andrey.shinkevich@virtuozzo.com>
Tested-by: Eric Blake <eblake@redhat.com>
Message-Id: <20200727194236.19551-5-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
16 months agoqemu-iotests/199: better catch postcopy time
Vladimir Sementsov-Ogievskiy [Mon, 27 Jul 2020 19:42:18 +0000 (22:42 +0300)] 
qemu-iotests/199: better catch postcopy time

The test aims to test _postcopy_ migration, and wants to do some write
operations during postcopy time.

Test considers migrate status=complete event on source as start of
postcopy. This is completely wrong, completion is completion of the
whole migration process. Let's instead consider destination start as
start of postcopy, and use RESUME event for it.

Next, as migration finish, let's use migration status=complete event on
target, as such method is closer to what libvirt or another user will
do, than tracking number of dirty-bitmaps.

Finally, add a possibility to dump events for debug. And if
set debug to True, we see, that actual postcopy period is very small
relatively to the whole test duration time (~0.2 seconds to >40 seconds
for me). This means, that test is very inefficient in what it supposed
to do. Let's improve it in following commits.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Andrey Shinkevich <andrey.shinkevich@virtuozzo.com>
Tested-by: Eric Blake <eblake@redhat.com>
Message-Id: <20200727194236.19551-4-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
16 months agoqemu-iotests/199: drop extra constraints
Vladimir Sementsov-Ogievskiy [Mon, 27 Jul 2020 19:42:17 +0000 (22:42 +0300)] 
qemu-iotests/199: drop extra constraints

We don't need any specific format constraints here. Still keep qcow2
for two reasons:
1. No extra calls of format-unrelated test
2. Add some check around persistent bitmap in future (require qcow2)

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Andrey Shinkevich <andrey.shinkevich@virtuozzo.com>
Tested-by: Eric Blake <eblake@redhat.com>
Message-Id: <20200727194236.19551-3-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
16 months agoqemu-iotests/199: fix style
Vladimir Sementsov-Ogievskiy [Mon, 27 Jul 2020 19:42:16 +0000 (22:42 +0300)] 
qemu-iotests/199: fix style

Mostly, satisfy pep8 complaints.

Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Andrey Shinkevich <andrey.shinkevich@virtuozzo.com>
Tested-by: Eric Blake <eblake@redhat.com>
Message-Id: <20200727194236.19551-2-vsementsov@virtuozzo.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
16 months agoqcow2: Fix capitalization of header extension constant.
Andrey Shinkevich [Fri, 17 Jul 2020 08:14:49 +0000 (11:14 +0300)] 
qcow2: Fix capitalization of header extension constant.

Make the capitalization of the hexadecimal numbers consistent for the
QCOW2 header extension constants in docs/interop/qcow2.txt.

Suggested-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Andrey Shinkevich <andrey.shinkevich@virtuozzo.com>
Reviewed-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <1594973699-781898-2-git-send-email-andrey.shinkevich@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
16 months agolinux-user: Use getcwd syscall directly
Andreas Schwab [Thu, 23 Jul 2020 10:27:13 +0000 (12:27 +0200)] 
linux-user: Use getcwd syscall directly

The glibc getcwd function returns different errors than the getcwd
syscall, which triggers an assertion failure in the glibc getcwd function
when running under the emulation.

When the syscall returns ENAMETOOLONG, the glibc wrapper uses a fallback
implementation that potentially handles an unlimited path length, and
returns with ERANGE if the provided buffer is too small.  The qemu
emulation cannot distinguish the two cases, and thus always returns ERANGE.
This is unexpected by the glibc wrapper.

Signed-off-by: Andreas Schwab <schwab@suse.de>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Message-Id: <mvmmu3qplvi.fsf@suse.de>
[lv: updated description]
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
16 months agolinux-user: Fix syscall rt_sigtimedwait() implementation
Filip Bozuta [Fri, 24 Jul 2020 18:16:51 +0000 (20:16 +0200)] 
linux-user: Fix syscall rt_sigtimedwait() implementation

Implementation of 'rt_sigtimedwait()' in 'syscall.c' uses the
function 'target_to_host_timespec()' to transfer the value of
'struct timespec' from target to host. However, the implementation
doesn't check whether this conversion succeeds and thus can cause
an unaproppriate error instead of the 'EFAULT (Bad address)' which
is supposed to be set if the conversion from target to host fails.

This was confirmed with the LTP test for rt_sigtimedwait:
"/testcases/kernel/syscalls/rt_sigtimedwait/rt_sigtimedwait01.c"
which causes an unapropriate error in test case "test_bad_adress3"
which is run with a bad adress for the 'struct timespec' argument:

FAIL: test_bad_address3 (349): Unexpected failure: EAGAIN/EWOULDBLOCK (11)

The test fails with an unexptected errno 'EAGAIN/EWOULDBLOCK' instead
of the expected EFAULT.

After the changes from this patch, the test case is executed successfully
along with the other LTP test cases for 'rt_sigtimedwait()':

PASS: test_bad_address3 (349): Test passed

Signed-off-by: Filip Bozuta <Filip.Bozuta@syrmia.com>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Message-Id: <20200724181651.167819-1-Filip.Bozuta@syrmia.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
16 months agolinux-user: Ensure mmap_min_addr is non-zero
Richard Henderson [Fri, 24 Jul 2020 21:23:14 +0000 (14:23 -0700)] 
linux-user: Ensure mmap_min_addr is non-zero

When the chroot does not have /proc mounted, we can read neither
/proc/sys/vm/mmap_min_addr nor /proc/sys/maps.

The enforcement of mmap_min_addr in the host kernel is done by
the security module, and so does not apply to processes owned
by root.  Which leads pgd_find_hole_fallback to succeed in probing
a reservation at address 0.  Which confuses pgb_reserved_va to
believe that guest_base has not actually been initialized.

We don't actually want NULL addresses to become accessible, so
make sure that mmap_min_addr is initialized with a non-zero value.

Buglink: https://bugs.launchpad.net/qemu/+bug/1888728
Reported-by: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Tested-by: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>
Acked-by: Laurent Vivier <laurent@vivier.eu>
Message-Id: <20200724212314.545877-1-richard.henderson@linaro.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
16 months agoMerge remote-tracking branch 'remotes/mst/tags/for_upstream' into staging
Peter Maydell [Mon, 27 Jul 2020 20:00:01 +0000 (21:00 +0100)] 
Merge remote-tracking branch 'remotes/mst/tags/for_upstream' into staging

virtio,pci: bugfixes

Minor bugfixes all over the places, including one CVE.

Additionally, a fix for an ancient bug in migration -
one has to wonder how come no one noticed.

The fix is also non-trivial since we dare not break all
existing machine types with pci - we have a work around
in the works, for now we just skip the work-around for
old machine types.

Great job by Hogan Wang noticing, debugging and fixing it,
and thanks to Dr. David Alan Gilbert for reviewing the patches.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
# gpg: Signature made Mon 27 Jul 2020 16:34:58 BST
# 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:
  virtio-pci: fix virtio_pci_queue_enabled()
  MAINTAINERS: Cover the firmware JSON schema
  vhost-vdpa :Fix Coverity CID 1430270 / CID 1420267
  libvhost-user: Report descriptor index on panic
  Fix vhost-user buffer over-read on ram hot-unplug
  hw/pci-host: save/restore pci host config register
  virtio-mem-pci: force virtio version 1

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
16 months agoMerge remote-tracking branch 'remotes/cohuck/tags/s390x-20200727' into staging
Peter Maydell [Mon, 27 Jul 2020 16:25:06 +0000 (17:25 +0100)] 
Merge remote-tracking branch 'remotes/cohuck/tags/s390x-20200727' into staging

fixes for protected virtualization and loadparm handling

# gpg: Signature made Mon 27 Jul 2020 15:01:32 BST
# gpg:                using RSA key C3D0D66DC3624FF6A8C018CEDECF6B93C6F02FAF
# gpg:                issuer "cohuck@redhat.com"
# gpg: Good signature from "Cornelia Huck <conny@cornelia-huck.de>" [marginal]
# gpg:                 aka "Cornelia Huck <huckc@linux.vnet.ibm.com>" [full]
# gpg:                 aka "Cornelia Huck <cornelia.huck@de.ibm.com>" [full]
# gpg:                 aka "Cornelia Huck <cohuck@kernel.org>" [marginal]
# gpg:                 aka "Cornelia Huck <cohuck@redhat.com>" [marginal]
# Primary key fingerprint: C3D0 D66D C362 4FF6 A8C0  18CE DECF 6B93 C6F0 2FAF

* remotes/cohuck/tags/s390x-20200727:
  s390x/s390-virtio-ccw: fix loadparm property getter
  s390x/protvirt: allow to IPL secure guests with -no-reboot

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
16 months agovirtio-pci: fix virtio_pci_queue_enabled()
Laurent Vivier [Mon, 27 Jul 2020 15:33:19 +0000 (17:33 +0200)] 
virtio-pci: fix virtio_pci_queue_enabled()

In legacy mode, virtio_pci_queue_enabled() falls back to
virtio_queue_enabled() to know if the queue is enabled.

But virtio_queue_enabled() calls again virtio_pci_queue_enabled()
if k->queue_enabled is set. This ends in a crash after a stack
overflow.

The problem can be reproduced with
"-device virtio-net-pci,disable-legacy=off,disable-modern=true
 -net tap,vhost=on"

And a look to the backtrace is very explicit:

    ...
    #4  0x000000010029a438 in virtio_queue_enabled ()
    #5  0x0000000100497a9c in virtio_pci_queue_enabled ()
    ...
    #130902 0x000000010029a460 in virtio_queue_enabled ()
    #130903 0x0000000100497a9c in virtio_pci_queue_enabled ()
    #130904 0x000000010029a460 in virtio_queue_enabled ()
    #130905 0x0000000100454a20 in vhost_net_start ()
    ...

This patch fixes the problem by introducing a new function
for the legacy case and calls it from virtio_pci_queue_enabled().
It also calls it from virtio_queue_enabled() to avoid code duplication.

Fixes: f19bcdfedd53 ("virtio-pci: implement queue_enabled method")
Cc: Jason Wang <jasowang@redhat.com>
Cc: Cindy Lu <lulu@redhat.com>
CC: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Laurent Vivier <lvivier@redhat.com>
Message-Id: <20200727153319.43716-1-lvivier@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
16 months agotarget/arm: Improve IMPDEF algorithm for IRG
Richard Henderson [Mon, 27 Jul 2020 15:12:11 +0000 (16:12 +0100)] 
target/arm: Improve IMPDEF algorithm for IRG

When GCR_EL1.RRND==1, the choosing of the random value is IMPDEF,
and the kernel is not expected to have set RGSR_EL1.  Force a
non-zero value into SEED, so that we do not continually return
the same tag.

Reported-by: Vincenzo Frascino <vincenzo.frascino@arm.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20200724163853.504655-4-richard.henderson@linaro.org
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
16 months agohw/arm/boot: Fix MTE for EL3 direct kernel boot
Richard Henderson [Mon, 27 Jul 2020 15:12:10 +0000 (16:12 +0100)] 
hw/arm/boot: Fix MTE for EL3 direct kernel boot

When booting an EL3 cpu with -kernel, we set up EL3 and then
drop down to EL2.  We need to enable access to v8.5-MemTag
tag allocation at EL3 before doing so.

Reported-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20200724163853.504655-3-richard.henderson@linaro.org
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
16 months agohw/arm/boot: Fix PAUTH for EL3 direct kernel boot
Richard Henderson [Mon, 27 Jul 2020 15:12:10 +0000 (16:12 +0100)] 
hw/arm/boot: Fix PAUTH for EL3 direct kernel boot

When booting an EL3 cpu with -kernel, we set up EL3 and then
drop down to EL2.  We need to enable access to v8.3-PAuth
keys and instructions at EL3 before doing so.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20200724163853.504655-2-richard.henderson@linaro.org
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
16 months agodocs/system/arm/virt: Document 'mte' machine option
Peter Maydell [Mon, 27 Jul 2020 15:12:10 +0000 (16:12 +0100)] 
docs/system/arm/virt: Document 'mte' machine option

Commit 6a0b7505f1fd6769c which added documentation of the virt board
crossed in the post with commit 6f4e1405b91da0d0 which added a new
'mte' machine option. Update the docs to include the new option.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
16 months agotarget/arm: Always pass cacheattr in S1_ptw_translate
Richard Henderson [Mon, 27 Jul 2020 15:12:10 +0000 (16:12 +0100)] 
target/arm: Always pass cacheattr in S1_ptw_translate

When we changed the interface of get_phys_addr_lpae to require
the cacheattr parameter, this spot was missed.  The compiler is
unable to detect the use of NULL vs the nonnull attribute here.

Fixes: 7e98e21c098
Reported-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Tested-by: Jan Kiszka <jan.kiskza@siemens.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
16 months agohw/misc/aspeed_sdmc: Fix incorrect memory size
Philippe Mathieu-Daudé [Mon, 27 Jul 2020 15:12:10 +0000 (16:12 +0100)] 
hw/misc/aspeed_sdmc: Fix incorrect memory size

The SDRAM Memory Controller has a 32-bit address bus, thus
supports up to 4 GiB of DRAM. There is a signed to unsigned
conversion error with the AST2600 maximum memory size:

  (uint64_t)(2048 << 20) = (uint64_t)(-2147483648)
                         = 0xffffffff40000000
                         = 16 EiB - 2 GiB

Fix by using the IEC suffixes which are usually safer, and add
an assertion check to verify the memory is valid. This would have
caught this bug:

  $ qemu-system-arm -M ast2600-evb
  qemu-system-arm: hw/misc/aspeed_sdmc.c:258: aspeed_sdmc_realize: Assertion `asc->max_ram_size < 4 * GiB' failed.
  Aborted (core dumped)

Fixes: 1550d72679 ("aspeed/sdmc: Add AST2600 support")
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
16 months agoACPI: Assert that we don't run out of the preallocated memory
Dongjiu Geng [Mon, 27 Jul 2020 15:12:10 +0000 (16:12 +0100)] 
ACPI: Assert that we don't run out of the preallocated memory

data_length is a constant value, so we use assert instead of
condition check.

Signed-off-by: Dongjiu Geng <gengdongjiu@huawei.com>
Message-id: 20200622113146.33421-1-gengdongjiu@huawei.com
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
16 months agoUpdate OpenBIOS images to 7f28286f built from submodule.
Mark Cave-Ayland [Mon, 27 Jul 2020 08:15:42 +0000 (09:15 +0100)] 
Update OpenBIOS images to 7f28286f built from submodule.

Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Cc: qemu-stable@nongnu.org
16 months agoMerge remote-tracking branch 'remotes/stsquad/tags/pull-fixes-for-rc2-270720-1' into...
Peter Maydell [Mon, 27 Jul 2020 14:55:56 +0000 (15:55 +0100)] 
Merge remote-tracking branch 'remotes/stsquad/tags/pull-fixes-for-rc2-270720-1' into staging

Various fixes for rc2:

  - get shippable working again
  - semihosting bug fixes
  - tweak tb-size handling for low memory machines
  - i386 compound literal float fix
  - linux-user MAP_FIXED->MAP_NOREPLACE on fallback
  - docker binfmt_misc fixes
  - linux-user nanosleep fix
  - tests/vm drain console fixes

# gpg: Signature made Mon 27 Jul 2020 09:45:31 BST
# gpg:                using RSA key 6685AE99E75167BCAFC8DF35FBD0DB095A9E2A44
# gpg: Good signature from "Alex Bennée (Master Work Key) <alex.bennee@linaro.org>" [full]
# Primary key fingerprint: 6685 AE99 E751 67BC AFC8  DF35 FBD0 DB09 5A9E 2A44

* remotes/stsquad/tags/pull-fixes-for-rc2-270720-1:
  tests/vm: add shutdown timeout in basevm.py
  python/qemu: Change ConsoleSocket to optionally drain socket.
  python/qemu: Cleanup changes to ConsoleSocket
  linux-user, ppc: fix clock_nanosleep() for linux-user-ppc
  linux-user: fix clock_nanosleep()
  tests/docker: add support for DEB_KEYRING
  tests/docker: fix binfmt_misc image building
  tests/docker: fix update command due to python3 str/bytes distinction
  linux-user: don't use MAP_FIXED in pgd_find_hole_fallback
  target/i386: floatx80: avoid compound literals in static initializers
  accel/tcg: better handle memory constrained systems
  util/oslib-win32: add qemu_get_host_physmem implementation
  util: add qemu_get_host_physmem utility function
  semihosting: don't send the trailing '\0'
  semihosting: defer connect_chardevs a little more to use serialx
  shippable: add one more qemu to registry url

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
16 months agoiotests/197: Fix for compat=0.10
Max Reitz [Mon, 27 Jul 2020 13:52:37 +0000 (15:52 +0200)] 
iotests/197: Fix for compat=0.10

Writing zeroes to a qcow2 v2 images without a backing file results in an
unallocated cluster as of 61b3043965.  197 has a test for COR-ing a
cluster on an image without a backing file, which means that the data
will be zero, so now on a v2 image that cluster will just stay
unallocated, and so the test fails.  Just force compat=1.1 for that
particular case to enforce the cluster to get allocated.

Fixes: 61b3043965fe3552ee2684a97e7cc809ca7a71b3
Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-Id: <20200727135237.1096841-1-mreitz@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
16 months agoMAINTAINERS: Cover the firmware JSON schema
Philippe Mathieu-Daudé [Fri, 3 Jul 2020 18:34:50 +0000 (20:34 +0200)] 
MAINTAINERS: Cover the firmware JSON schema

Add an entry to cover firmware.json (see commit 3a0adfc9bf:
schema that describes the different uses and properties of
virtual machine firmware).

Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: Michael S. Tsirkin <mst@redhat.com>
Cc: Kashyap Chamarthy <kchamart@redhat.com>
Cc: Daniel P. Berrange <berrange@redhat.com>
Suggested-by: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20200703183450.32398-1-philmd@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Kashyap Chamarthy <kchamart@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
16 months agovhost-vdpa :Fix Coverity CID 1430270 / CID 1420267
Cindy Lu [Fri, 10 Jul 2020 06:46:42 +0000 (14:46 +0800)] 
vhost-vdpa :Fix Coverity CID 1430270 / CID 1420267

In the function vhost_vdpa_dma_map/unmap, The struct msg was not initialized all its fields.

Signed-off-by: Cindy Lu <lulu@redhat.com>
Message-Id: <20200710064642.24505-1-lulu@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Acked-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Li Qiang <liq3ea@gmail.com>
Cc: qemu-stable@nongnu.org
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
16 months agolibvhost-user: Report descriptor index on panic
Philippe Mathieu-Daudé [Thu, 23 Jul 2020 17:19:35 +0000 (19:19 +0200)] 
libvhost-user: Report descriptor index on panic

We want to report the index of the descriptor,
not its pointer.

Fixes: 7b2e5c65f4 ("contrib: add libvhost-user")
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20200723171935.18535-1-philmd@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Raphael Norwitz <raphael.norwitz@nutanix.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Cc: qemu-stable@nongnu.org
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
16 months agoFix vhost-user buffer over-read on ram hot-unplug
Raphael Norwitz [Fri, 17 Jul 2020 04:21:30 +0000 (04:21 +0000)] 
Fix vhost-user buffer over-read on ram hot-unplug

The VHOST_USER_PROTOCOL_F_CONFIGURE_MEM_SLOTS vhost-user protocol
feature introduced a shadow-table, used by the backend to dynamically
determine how a vdev's memory regions have changed since the last
vhost_user_set_mem_table() call. On hot-remove, a memmove() operation
is used to overwrite the removed shadow region descriptor(s). The size
parameter of this memmove was off by 1 such that if a VM with a backend
supporting the VHOST_USER_PROTOCOL_F_CONFIGURE_MEM_SLOTS filled it's
shadow-table (by performing the maximum number of supported hot-add
operatons) and attempted to remove the last region, Qemu would read an
out of bounds value and potentially crash.

This change fixes the memmove() bounds such that this erroneous read can
never happen.

Signed-off-by: Peter Turschmid <peter.turschm@nutanix.com>
Signed-off-by: Raphael Norwitz <raphael.norwitz@nutanix.com>
Message-Id: <1594799958-31356-1-git-send-email-raphael.norwitz@nutanix.com>
Fixes: f1aeb14b0809 ("Transmit vhost-user memory regions individually")
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Cc: qemu-stable@nongnu.org
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
16 months agohw/pci-host: save/restore pci host config register
Hogan Wang [Mon, 27 Jul 2020 08:46:20 +0000 (16:46 +0800)] 
hw/pci-host: save/restore pci host config register

The pci host config register is used to save PCI address for
read/write config data. If guest writes a value to config register,
and then QEMU pauses the vcpu to migrate, after the migration, the guest
will continue to write pci config data, and the write data will be ignored
because of new qemu process losing the config register state.

To trigger the bug:
1. guest is booting in seabios.
2. guest enables the SMRAM in seabios:piix4_apmc_smm_setup, and then
   expects to disable the SMRAM by pci_config_writeb.
3. after guest writes the pci host config register, QEMU pauses vcpu
   to finish migration.
4. guest write of config data(0x0A) fails to disable the SMRAM because
   the config register state is lost.
5. guest continues to boot and crashes in ipxe option ROM due to SMRAM
   in enabled state.

Example Reproducer:

step 1. Make modifications to seabios and qemu for increase reproduction
efficiency, write 0xf0 to 0x402 port notify qemu to stop vcpu after
0x0cf8 port wrote i440 configure register. qemu stop vcpu when catch
0x402 port wrote 0xf0.

seabios:/src/hw/pci.c
@@ -52,6 +52,11 @@ void pci_config_writeb(u16 bdf, u32 addr, u8 val)
         writeb(mmconfig_addr(bdf, addr), val);
     } else {
         outl(ioconfig_cmd(bdf, addr), PORT_PCI_CMD);
+       if (bdf == 0 && addr == 0x72 && val == 0xa) {
+            dprintf(1, "stop vcpu\n");
+            outb(0xf0, 0x402); // notify qemu to stop vcpu
+            dprintf(1, "resume vcpu\n");
+        }
         outb(val, PORT_PCI_DATA + (addr & 3));
     }
 }

qemu:hw/char/debugcon.c
@@ -60,6 +61,9 @@ static void debugcon_ioport_write(void *opaque, hwaddr addr, uint64_t val,
     printf(" [debugcon: write addr=0x%04" HWADDR_PRIx " val=0x%02" PRIx64 "]\n", addr, val);
 #endif

+    if (ch == 0xf0) {
+        vm_stop(RUN_STATE_PAUSED);
+    }
     /* XXX this blocks entire thread. Rewrite to use
      * qemu_chr_fe_write and background I/O callbacks */
     qemu_chr_fe_write_all(&s->chr, &ch, 1);

step 2. start vm1 by the following command line, and then vm stopped.
$ qemu-system-x86_64 -machine pc-i440fx-5.0,accel=kvm\
 -netdev tap,ifname=tap-test,id=hostnet0,vhost=on,downscript=no,script=no\
 -device virtio-net-pci,netdev=hostnet0,id=net0,bus=pci.0,addr=0x13,bootindex=3\
 -device cirrus-vga,id=video0,vgamem_mb=16,bus=pci.0,addr=0x2\
 -chardev file,id=seabios,path=/var/log/test.seabios,append=on\
 -device isa-debugcon,iobase=0x402,chardev=seabios\
 -monitor stdio

step 3. start vm2 to accept vm1 state.
$ qemu-system-x86_64 -machine pc-i440fx-5.0,accel=kvm\
 -netdev tap,ifname=tap-test1,id=hostnet0,vhost=on,downscript=no,script=no\
 -device virtio-net-pci,netdev=hostnet0,id=net0,bus=pci.0,addr=0x13,bootindex=3\
 -device cirrus-vga,id=video0,vgamem_mb=16,bus=pci.0,addr=0x2\
 -chardev file,id=seabios,path=/var/log/test.seabios,append=on\
 -device isa-debugcon,iobase=0x402,chardev=seabios\
 -monitor stdio \
 -incoming tcp:127.0.0.1:8000

step 4. execute the following qmp command in vm1 to migrate.
(qemu) migrate tcp:127.0.0.1:8000

step 5. execute the following qmp command in vm2 to resume vcpu.
(qemu) cont
Before this patch, we get KVM "emulation failure" error on vm2.
This patch fixes it.

Cc: qemu-stable@nongnu.org
Signed-off-by: Hogan Wang <hogan.wang@huawei.com>
Message-Id: <20200727084621.3279-1-hogan.wang@huawei.com>
Reported-by: "Dr. David Alan Gilbert" <dgilbert@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
16 months agovirtio-mem-pci: force virtio version 1
David Hildenbrand [Mon, 27 Jul 2020 11:59:05 +0000 (13:59 +0200)] 
virtio-mem-pci: force virtio version 1

Trying to run simple virtio-mem-pci examples currently fails with
    qemu-system-x86_64: -device virtio-mem-pci,id=vm0,memdev=mem0,node=0,
    requested-size=300M: device is modern-only, use disable-legacy=on
due to the added safety checks in 9b3a35ec8236 ("virtio: verify that legacy
support is not accidentally on").

As noted by Conny, we have to force virtio version 1. While at it, use
qdev_realize() to set the parent bus and realize - like most other
virtio-*-pci implementations.

Fixes: 0b9a2443a48b ("virtio-pci: Proxy for virtio-mem")
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Cc: Cornelia Huck <cohuck@redhat.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20200727115905.129397-1-david@redhat.com>
Cc: qemu-stable@nongnu.org
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
16 months agoiotests: Select a default machine for the rx and avr targets
Thomas Huth [Wed, 22 Jul 2020 16:19:08 +0000 (18:19 +0200)] 
iotests: Select a default machine for the rx and avr targets

If you are building only with either the new rx-softmmu or avr-softmmu
target, "make check-block" fails a couple of tests since there is no
default machine defined in these new targets. We have to select a machine
in the "check" script for these, just like we already do for the arm- and
tricore-softmmu targets.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20200722161908.25383-1-thuth@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
16 months agoblock/amend: Check whether the node exists
Max Reitz [Fri, 10 Jul 2020 09:50:37 +0000 (11:50 +0200)] 
block/amend: Check whether the node exists

We should check whether the user-specified node-name actually refers to
a node.  The simplest way to do that is to use bdrv_lookup_bs() instead
of bdrv_find_node() (the former wraps the latter, and produces an error
message if necessary).

Reported-by: Coverity (CID 1430268)
Fixes: ced914d0ab9fb2c900f873f6349a0b8eecd1fdbe
Signed-off-by: Max Reitz <mreitz@redhat.com>
Message-Id: <20200710095037.10885-1-mreitz@redhat.com>
Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
16 months agotests/vm: add shutdown timeout in basevm.py
Robert Foley [Fri, 24 Jul 2020 06:45:09 +0000 (07:45 +0100)] 
tests/vm: add shutdown timeout in basevm.py

We are adding the shutdown timeout to solve an issue
we now see where the aarch64 VMs timeout on shutdown
under TCG.

There is a new 3 second timeout in machine.py,
which we override in basevm.py when shutting down.

Signed-off-by: Robert Foley <robert.foley@linaro.org>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20200717203041.9867-4-robert.foley@linaro.org>
Message-Id: <20200724064509.331-17-alex.bennee@linaro.org>

16 months agopython/qemu: Change ConsoleSocket to optionally drain socket.
Robert Foley [Fri, 24 Jul 2020 06:45:08 +0000 (07:45 +0100)] 
python/qemu: Change ConsoleSocket to optionally drain socket.

The primary purpose of this change is to clean up
machine.py's console_socket property to return a single type,
a ConsoleSocket.

ConsoleSocket now derives from a socket, which means that
in the default case (of not draining), machine.py
will see the same behavior as it did prior to ConsoleSocket.

Signed-off-by: Robert Foley <robert.foley@linaro.org>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20200717203041.9867-3-robert.foley@linaro.org>
Message-Id: <20200724064509.331-16-alex.bennee@linaro.org>

16 months agopython/qemu: Cleanup changes to ConsoleSocket
Robert Foley [Fri, 24 Jul 2020 06:45:07 +0000 (07:45 +0100)] 
python/qemu: Cleanup changes to ConsoleSocket

The changes to console_socket.py and machine.py are to
cleanup for pylint and flake8.

Signed-off-by: Robert Foley <robert.foley@linaro.org>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20200717203041.9867-2-robert.foley@linaro.org>
Message-Id: <20200724064509.331-15-alex.bennee@linaro.org>

16 months agolinux-user, ppc: fix clock_nanosleep() for linux-user-ppc
Laurent Vivier [Fri, 24 Jul 2020 06:45:06 +0000 (07:45 +0100)] 
linux-user, ppc: fix clock_nanosleep() for linux-user-ppc

Our safe_clock_nanosleep() returns -1 and updates errno.

We don't need to update the CRF bit in syscall.c because it will
be updated in ppc/cpu_loop.c as the return value is negative.

Signed-off-by: Laurent Vivier <laurent@vivier.eu>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20200722174612.2917566-3-laurent@vivier.eu>
Message-Id: <20200724064509.331-14-alex.bennee@linaro.org>

16 months agolinux-user: fix clock_nanosleep()
Laurent Vivier [Fri, 24 Jul 2020 06:45:05 +0000 (07:45 +0100)] 
linux-user: fix clock_nanosleep()

If the call is interrupted by a signal handler, it fails with error EINTR
and if "remain" is not NULL and "flags" is not TIMER_ABSTIME, it returns
the remaining unslept time in "remain".

Update linux-user to not overwrite the "remain" structure if there is no
error.

Found with "make check-tcg", linux-test fails on nanosleep test:

  TEST    linux-test on x86_64
.../tests/tcg/multiarch/linux-test.c:242: nanosleep

Reported-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20200722174612.2917566-2-laurent@vivier.eu>
Message-Id: <20200724064509.331-13-alex.bennee@linaro.org>

16 months agotests/docker: add support for DEB_KEYRING
Alex Bennée [Fri, 24 Jul 2020 06:45:04 +0000 (07:45 +0100)] 
tests/docker: add support for DEB_KEYRING

For installing stuff from sid or ports you may need to manually
specify the location of the keyring. You can even import keys into
your personal keyring and point it there, e.g.:

  gpg --keyserver keyring.debian.org --recv-keys 84C573CD4E1AFD6C
  make docker-binfmt-image-debian-sid-hppa DEB_TYPE=sid DEB_ARCH=hppa \
      DEB_URL=http://ftp.ports.debian.org/debian-ports/ \
      EXECUTABLE=./hppa-linux-user/qemu-hppa V=1 \
      DEB_KEYRING=${HOME}/.gnupg/pubring.kbx

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20200724064509.331-12-alex.bennee@linaro.org>

16 months agotests/docker: fix binfmt_misc image building
Alex Bennée [Fri, 24 Jul 2020 06:45:03 +0000 (07:45 +0100)] 
tests/docker: fix binfmt_misc image building

When we updated the arguments for docker.py we missed a bit.

Fixes: dfae628459 ("docker.py/build: support -t and -f arguments")
Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20200724064509.331-11-alex.bennee@linaro.org>

16 months agotests/docker: fix update command due to python3 str/bytes distinction
Alex Bennée [Fri, 24 Jul 2020 06:45:02 +0000 (07:45 +0100)] 
tests/docker: fix update command due to python3 str/bytes distinction

Does this seem convoluted to you? It feels a little complicated to me.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20200724064509.331-10-alex.bennee@linaro.org>