OSDN Git Service

qmiga/qemu.git
7 years agovfio-pci: Fix GTT wrap-around for Skylake+ IGD
Alex Williamson [Fri, 10 Feb 2017 20:12:03 +0000 (13:12 -0700)]
vfio-pci: Fix GTT wrap-around for Skylake+ IGD

Previous IGD, up through Broadwell, only seem to write GTT values into
the first 1MB of space allocated for the BDSM, but clearly the GTT
can be multiple MB in size.  Our test in vfio_igd_quirk_data_write()
correctly filters out indexes beyond 1MB, but given the 1MB mask we're
using, we re-apply writes only to the first 1MB of the guest allocated
BDSM.

We can't assume either the host or guest BDSM is naturally aligned, so
we can't simply apply a different mask.  Instead, save the host BDSM
and do the arithmetic to subtract the host value to get the BDSM
offset and add it to the guest allocated BDSM.

Reported-by: Alexander Indenbaum <alexander.indenbaum@gmail.com>
Tested-by: Alexander Indenbaum <alexander.indenbaum@gmail.com>
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
7 years agoMerge remote-tracking branch 'remotes/jnsnow/tags/ide-pull-request' into staging
Peter Maydell [Fri, 10 Feb 2017 18:07:02 +0000 (18:07 +0000)]
Merge remote-tracking branch 'remotes/jnsnow/tags/ide-pull-request' into staging

# gpg: Signature made Fri 10 Feb 2017 16:47:54 GMT
# gpg:                using RSA key 0x7DEF8106AAFC390E
# gpg: Good signature from "John Snow (John Huston) <jsnow@redhat.com>"
# 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/ide-pull-request:
  ahci: advertise HOST_CAP_64

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agoahci: advertise HOST_CAP_64
Ladi Prosek [Fri, 10 Feb 2017 16:47:11 +0000 (11:47 -0500)]
ahci: advertise HOST_CAP_64

The AHCI emulation code supports 64-bit addressing and should advertise this
fact in the Host Capabilities register. Both Linux and Windows drivers test
this bit to decide if the upper 32 bits of various registers may be written
to, and at least some versions of Windows have a bug where DMA is attempted
with an address above 4GB but, in the absence of HOST_CAP_64, the upper 32
bits are left unititialized which leads to a memory corruption.

[Maintainer edit:

This fixes https://bugzilla.redhat.com/show_bug.cgi?id=1411105,
which affects Windows Server 2008 SP2 in some cases.]

Signed-off-by: Ladi Prosek <lprosek@redhat.com>
Message-id: 1484305370-6220-1-git-send-email-lprosek@redhat.com
[Amended commit message --js]
Signed-off-by: John Snow <jsnow@redhat.com>
7 years agotests/prom-env: Ease time-out problems on slow hosts
Thomas Huth [Fri, 10 Feb 2017 15:14:59 +0000 (16:14 +0100)]
tests/prom-env: Ease time-out problems on slow hosts

Peter Maydell recently ran into time-out problems with the
prom-env test on a rather slow ARM board. To tackle this issue,
we can speed up the test by running QEMU with "-nodefaults" for
the pseries machine, so that SLOF has less devices to scan during
boot, and by using the "nvramrc" environment variable instead of
"boot-command", since this variable is evaluated earlier in the
boot process.
And to be really sure that we do not face such time out problems
again, let's also increase the time out value from 100s to 120s
instead.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-id: 1486739699-1076-1-git-send-email-thuth@redhat.com
Tested-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agoMerge remote-tracking branch 'remotes/stsquad/tags/pull-travis-10022017-1' into staging
Peter Maydell [Fri, 10 Feb 2017 15:05:36 +0000 (15:05 +0000)]
Merge remote-tracking branch 'remotes/stsquad/tags/pull-travis-10022017-1' into staging

One minor fix and a build split to reduce timeouts.

# gpg: Signature made Fri 10 Feb 2017 14:46:52 GMT
# gpg:                using RSA key 0xFBD0DB095A9E2A44
# gpg: Good signature from "Alex Bennée (Master Work Key) <alex.bennee@linaro.org>"
# Primary key fingerprint: 6685 AE99 E751 67BC AFC8  DF35 FBD0 DB09 5A9E 2A44

* remotes/stsquad/tags/pull-travis-10022017-1:
  .travis.yml: split VM based builds
  .travis.yml: don't specify CONFIG twice

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years ago.travis.yml: split VM based builds
Alex Bennée [Fri, 20 Jan 2017 17:11:49 +0000 (17:11 +0000)]
.travis.yml: split VM based builds

The Trusty based builds run a little slower than the main container
based ones. This is also true for the latest version of Clang. The
builds are getting very close (and occasionally run over) the 50 minute
timeout. Rather than partitioning by target I just split them into
linux-user and system builds.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
7 years ago.travis.yml: don't specify CONFIG twice
Alex Bennée [Fri, 20 Jan 2017 17:00:51 +0000 (17:00 +0000)]
.travis.yml: don't specify CONFIG twice

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>
7 years agoMerge remote-tracking branch 'remotes/kraxel/tags/pull-ui-20170209-2' into staging
Peter Maydell [Thu, 9 Feb 2017 16:58:39 +0000 (16:58 +0000)]
Merge remote-tracking branch 'remotes/kraxel/tags/pull-ui-20170209-2' into staging

vnc: add support for multiple listening sockets.
vnc: misc fixes and cleanups.

# gpg: Signature made Thu 09 Feb 2017 16:45:02 GMT
# gpg:                using RSA key 0x4CB6D8EED3E87138
# gpg: Good signature from "Gerd Hoffmann (work) <kraxel@redhat.com>"
# gpg:                 aka "Gerd Hoffmann <gerd@kraxel.org>"
# gpg:                 aka "Gerd Hoffmann (private) <kraxel@gmail.com>"
# Primary key fingerprint: A032 8CFF B93A 17A7 9901  FE7D 4CB6 D8EE D3E8 7138

* remotes/kraxel/tags/pull-ui-20170209-2:
  ui: add ability to specify multiple VNC listen addresses
  util: add iterators for QemuOpts values
  ui: let VNC server listen on all resolved IP addresses
  ui: extract code to connect/listen from vnc_display_open
  ui: refactor code for populating SocketAddress from vnc_display_open
  ui: refactor VncDisplay to allow multiple listening sockets
  ui: fix reporting of VNC auth in query-vnc-servers
  ui: fix regression handling bare 'websocket' option to -vnc
  vnc: do not disconnect on EAGAIN
  ui/vnc: Drop unused vnc_has_job() and vnc_jobs_clear()

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agoui: add ability to specify multiple VNC listen addresses
Daniel P. Berrange [Fri, 3 Feb 2017 12:06:49 +0000 (12:06 +0000)]
ui: add ability to specify multiple VNC listen addresses

This change allows the listen address and websocket address
options for -vnc to be repeated. This causes the VNC server
to listen on multiple addresses. e.g.

 $ $QEMU -vnc vnc=localhost:1,vnc=unix:/tmp/vnc,\
              websocket=127.0.0.1:8080,websocket=[::]:8081

results in listening on

127.0.0.1:5901, 127.0.0.1:8080, ::1:5901, :::8081 & /tmp/vnc

Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Message-id: 20170203120649.15637-9-berrange@redhat.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
7 years agoutil: add iterators for QemuOpts values
Daniel P. Berrange [Fri, 3 Feb 2017 12:06:48 +0000 (12:06 +0000)]
util: add iterators for QemuOpts values

To iterate over all QemuOpts currently requires using a callback
function which is inconvenient for control flow. Add support for
using iterator functions more directly

  QemuOptsIter iter;
  QemuOpt *opt;

  qemu_opts_iter_init(&iter, opts, "repeated-key");
  while ((opt = qemu_opts_iter_next(&iter)) != NULL) {
      ....do something...
  }

Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Message-id: 20170203120649.15637-8-berrange@redhat.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
7 years agoui: let VNC server listen on all resolved IP addresses
Daniel P. Berrange [Fri, 3 Feb 2017 12:06:47 +0000 (12:06 +0000)]
ui: let VNC server listen on all resolved IP addresses

Remove the limitation that the VNC server can only listen on
a single resolved IP address. This uses the new DNS resolver
API to resolve a SocketAddress struct into an array of
SocketAddress structs containing raw IP addresses. The VNC
server will then attempt to listen on all resolved IP addresses.
The server must successfully listen on at least one of the
resolved IP addresses, otherwise an error will be reported.

Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Message-id: 20170203120649.15637-7-berrange@redhat.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
7 years agoui: extract code to connect/listen from vnc_display_open
Daniel P. Berrange [Fri, 3 Feb 2017 12:06:46 +0000 (12:06 +0000)]
ui: extract code to connect/listen from vnc_display_open

The code which takes a SocketAddress and connects/listens on the
network is going to get more complicated to deal with multiple
listeners. Pull it out into a separate method to avoid making the
vnc_display_open method even more complex.

Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Message-id: 20170203120649.15637-6-berrange@redhat.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
7 years agoui: refactor code for populating SocketAddress from vnc_display_open
Daniel P. Berrange [Fri, 3 Feb 2017 12:06:45 +0000 (12:06 +0000)]
ui: refactor code for populating SocketAddress from vnc_display_open

The code which interprets the CLI args to populate the SocketAddress
objects for plain & websockets VNC is quite complex already and will
need further enhancements shortly. Refactor it into separate methods
to avoid vnc_display_open getting even larger. As a side effect of
the refactoring, it is now possible to specify a listen address for
the websocket server explicitly. e.g,

  -vnc localhost:5900,websockets=0.0.0.0:8080

will listen on localhost for the plain VNC server, but expose the
websockets VNC server on the public interface. This refactoring
also removes the restriction that prevents enabling websockets
when the plain VNC server is listening on a UNIX socket.

Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Message-id: 20170203120649.15637-5-berrange@redhat.com

[ kraxel: squashed clang build fix ]

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
7 years agoui: refactor VncDisplay to allow multiple listening sockets
Daniel P. Berrange [Fri, 3 Feb 2017 12:06:44 +0000 (12:06 +0000)]
ui: refactor VncDisplay to allow multiple listening sockets

Currently there is only a single listener for plain VNC and
a single listener for websockets VNC. This means that if
getaddrinfo() returns multiple IP addresses, for a hostname,
the VNC server can only listen on one of them. This is
just bearable if listening on wildcard interface, or if
the host only has a single network interface to listen on,
but if there are multiple NICs and the VNC server needs
to listen on 2 or more specific IP addresses, it can't be
done.

This refactors the VncDisplay state so that it holds an
array of listening sockets, but still only listens on
one socket.

Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Message-id: 20170203120649.15637-4-berrange@redhat.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
7 years agoui: fix reporting of VNC auth in query-vnc-servers
Daniel P. Berrange [Fri, 3 Feb 2017 12:06:43 +0000 (12:06 +0000)]
ui: fix reporting of VNC auth in query-vnc-servers

Currently the VNC authentication info is emitted at the
top level of the query-vnc-servers data. This is wrong
because the authentication scheme differs between plain
and websockets when TLS is enabled. We should instead
report auth against the individual servers. e.g.

(QEMU) query-vnc-servers
{
    "return": [
        {
            "clients": [],
            "id": "default",
            "auth": "vencrypt",
            "vencrypt": "x509-vnc",
            "server": [
                {
                    "host": "127.0.0.1"
                    "service": "5901",
                    "websocket": false,
                    "family": "ipv4",
                    "auth": "vencrypt",
                    "vencrypt": "x509-vnc"
                },
                {
                    "host": "127.0.0.1",
                    "service": "5902",
                    "websocket": true,
                    "family": "ipv4",
                    "auth": "vnc"
                }
            ]
        }
    ]
}

This also future proofs the QMP schema so that we can
cope with multiple VNC server instances, listening on
different interfaces or ports, with different auth
setup.

Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Message-id: 20170203120649.15637-3-berrange@redhat.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
7 years agoui: fix regression handling bare 'websocket' option to -vnc
Daniel P. Berrange [Fri, 3 Feb 2017 12:06:42 +0000 (12:06 +0000)]
ui: fix regression handling bare 'websocket' option to -vnc

The -vnc argument is documented as accepting two syntaxes for
the 'websocket' option, either a bare option name, or a port
number. If using the bare option name, it is supposed to apply
the display number as an offset to base port 5700. e.g.

  -vnc localhost:3,websocket

should listen on port 5703, however, this was broken in 2.3.0 since

  commit 4db14629c38611061fc19ec6927405923de84f08
  Author: Gerd Hoffmann <kraxel@redhat.com>
  Date:   Tue Sep 16 12:33:03 2014 +0200

    vnc: switch to QemuOpts, allow multiple servers

instead qemu tries to listen on port "on" which gets looked up in
/etc/services and fails.

Fixes bug: #1455912

Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Message-id: 20170203120649.15637-2-berrange@redhat.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
7 years agovnc: do not disconnect on EAGAIN
Michael Tokarev [Fri, 3 Feb 2017 09:52:29 +0000 (12:52 +0300)]
vnc: do not disconnect on EAGAIN

When qemu vnc server is trying to send large update to clients,
there might be a situation when system responds with something
like EAGAIN, indicating that there's no system memory to send
that much data (depending on the network speed, client and server
and what is happening).  In this case, something like this happens
on qemu side (from strace):

sendmsg(16, {msg_name(0)=NULL,
        msg_iov(1)=[{"\244\"..., 729186}],
        msg_controllen=0, msg_flags=0}, 0) = 103950
sendmsg(16, {msg_name(0)=NULL,
        msg_iov(1)=[{"lz\346"..., 1559618}],
        msg_controllen=0, msg_flags=0}, 0) = -1 EAGAIN
sendmsg(-1, {msg_name(0)=NULL,
        msg_iov(1)=[{"lz\346"..., 1559618}],
        msg_controllen=0, msg_flags=0}, 0) = -1 EBADF

qemu closes the socket before the retry, and obviously it gets EBADF
when trying to send to -1.

This is because there WAS a special handling for EAGAIN, but now it doesn't
work anymore, after commit 04d2529da27db512dcbd5e99d0e26d333f16efcc, because
now in all error-like cases we initiate vnc disconnect.

This change were introduced in qemu 2.6, and caused numerous grief for many
people, resulting in their vnc clients reporting sporadic random disconnects
from vnc server.

Fix that by doing the disconnect only when necessary, i.e. omitting this
very case of EAGAIN.

Hopefully the existing condition (comparing with QIO_CHANNEL_ERR_BLOCK)
is sufficient, as the original code (before the above commit) were
checking for other errno values too.

Apparently there's another (semi?)bug exist somewhere here, since the
code tries to write to fd# -1, it probably should check if the connection
is open before. But this isn't important.

Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>
Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
Message-id: 1486115549-9398-1-git-send-email-mjt@msgid.tls.msk.ru
Fixes: 04d2529da27db512dcbd5e99d0e26d333f16efcc
Cc: Daniel P. Berrange <berrange@redhat.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: qemu-stable@nongnu.org
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
7 years agoui/vnc: Drop unused vnc_has_job() and vnc_jobs_clear()
Peter Maydell [Fri, 3 Feb 2017 18:24:20 +0000 (18:24 +0000)]
ui/vnc: Drop unused vnc_has_job() and vnc_jobs_clear()

The functions vnc_has_job() and vnc_jobs_clear() are
never used; remove them.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Gonglei <arei.gonglei@huawei.com>
Message-id: 1486146260-8092-1-git-send-email-peter.maydell@linaro.org
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
7 years agoMerge remote-tracking branch 'remotes/pmaydell/tags/pull-target-arm-20170207-1' into...
Peter Maydell [Tue, 7 Feb 2017 19:21:30 +0000 (19:21 +0000)]
Merge remote-tracking branch 'remotes/pmaydell/tags/pull-target-arm-20170207-1' into staging

target-arm:
 * new "unimplemented" device for stubbing out devices in a
   system model so accesses can be logged
 * stellaris: document the SoC memory map
 * arm: create instruction syndromes for AArch32 data aborts
 * arm: Correctly handle watchpoints for BE32 CPUs
 * Fix Thumb-1 BE32 execution and disassembly
 * arm: Add cfgend parameter for ARM CPU selection
 * sd: sdhci: check data length during dma_memory_read
 * aspeed: add a watchdog controller
 * integratorcp: adding vmstate for save/restore

# gpg: Signature made Tue 07 Feb 2017 19:20:19 GMT
# gpg:                using RSA key 0x3C2525ED14360CDE
# gpg: Good signature from "Peter Maydell <peter.maydell@linaro.org>"
# gpg:                 aka "Peter Maydell <pmaydell@gmail.com>"
# gpg:                 aka "Peter Maydell <pmaydell@chiark.greenend.org.uk>"
# Primary key fingerprint: E1A5 C593 CD41 9DE2 8E83  15CF 3C25 25ED 1436 0CDE

* remotes/pmaydell/tags/pull-target-arm-20170207-1:
  stellaris: Use the 'unimplemented' device for parts we don't implement
  hw/misc: New "unimplemented" sysbus device
  stellaris: Document memory map and which SoC devices are unimplemented
  target/arm: A32, T32: Create Instruction Syndromes for Data Aborts
  target/arm: Abstract out pbit/wbit tests in ARM ldr/str decode
  arm: Correctly handle watchpoints for BE32 CPUs
  Fix Thumb-1 BE32 execution and disassembly.
  target/arm: Add cfgend parameter for ARM CPU selection.
  hw/arm/integratorcp: Support specifying features via -cpu
  sd: sdhci: check data length during dma_memory_read
  aspeed: add a watchdog controller
  wdt: Add Aspeed watchdog device model
  integratorcp: adding vmstate for save/restore

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agostellaris: Use the 'unimplemented' device for parts we don't implement
Peter Maydell [Tue, 7 Feb 2017 18:30:00 +0000 (18:30 +0000)]
stellaris: Use the 'unimplemented' device for parts we don't implement

Use the 'unimplemented' dummy device to cover regions of the
SoC device memory map which we don't have proper device
implementations for yet.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-id: 1484247815-15279-4-git-send-email-peter.maydell@linaro.org

7 years agohw/misc: New "unimplemented" sysbus device
Peter Maydell [Tue, 7 Feb 2017 18:30:00 +0000 (18:30 +0000)]
hw/misc: New "unimplemented" sysbus device

Create a new "unimplemented" sysbus device, which simply accepts
all read and write accesses, and implements them as read-as-zero,
write-ignored, with logging of the access as LOG_UNIMP.

This is useful for stubbing out bits of an SoC or board model
which haven't been written yet.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-id: 1484247815-15279-3-git-send-email-peter.maydell@linaro.org

7 years agostellaris: Document memory map and which SoC devices are unimplemented
Peter Maydell [Tue, 7 Feb 2017 18:30:00 +0000 (18:30 +0000)]
stellaris: Document memory map and which SoC devices are unimplemented

Add a comment documenting the memory map of the SoC devices and which
are not implemented.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1484247815-15279-2-git-send-email-peter.maydell@linaro.org

7 years agotarget/arm: A32, T32: Create Instruction Syndromes for Data Aborts
Peter Maydell [Tue, 7 Feb 2017 18:30:00 +0000 (18:30 +0000)]
target/arm: A32, T32: Create Instruction Syndromes for Data Aborts

Add support for generating the ISS (Instruction Specific Syndrome)
for Data Abort exceptions taken from AArch32. These syndromes are
used by hypervisors for example to trap and emulate memory accesses.

This is the equivalent for AArch32 guests of the work done for AArch64
guests in commit aaa1f954d4cab243.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
7 years agotarget/arm: Abstract out pbit/wbit tests in ARM ldr/str decode
Peter Maydell [Tue, 7 Feb 2017 18:29:59 +0000 (18:29 +0000)]
target/arm: Abstract out pbit/wbit tests in ARM ldr/str decode

In the ARM ldr/str decode path, rather than directly testing
"insn & (1 << 21)" and "insn & (1 << 24)", abstract these
bits out into wbit and pbit local flags. (We will want to
do more tests against them to determine whether we need to
provide syndrome information.)

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
7 years agoarm: Correctly handle watchpoints for BE32 CPUs
Julian Brown [Tue, 7 Feb 2017 18:29:59 +0000 (18:29 +0000)]
arm: Correctly handle watchpoints for BE32 CPUs

In BE32 mode, sub-word size watchpoints can fail to trigger because the
address of the access is adjusted in the opcode helpers before being
compared with the watchpoint registers.  This patch reverses the address
adjustment before performing the comparison with the help of a new CPUClass
hook.

This version of the patch augments and tidies up comments a little.

Signed-off-by: Julian Brown <julian@codesourcery.com>
Message-id: caaf64ffc72f6ae183015337b7afdbd4b8989cb6.1484929304.git.julian@codesourcery.com
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agoFix Thumb-1 BE32 execution and disassembly.
Julian Brown [Tue, 7 Feb 2017 18:29:59 +0000 (18:29 +0000)]
Fix Thumb-1 BE32 execution and disassembly.

Thumb-1 code has some issues in BE32 mode (as currently implemented). In
short, since bytes are swapped within words at load time for BE32
executables, this also swaps pairs of adjacent Thumb-1 instructions.

This patch un-swaps those pairs of instructions again, both for execution,
and for disassembly. (The previous version of the patch always read four
bytes in arm_read_memory_func and then extracted the proper two bytes,
in a probably misguided attempt to match the behaviour of actual hardware
as described by e.g. the ARM9TDMI TRM, section 3.3 "Endian effects for
instruction fetches". It's less complicated to just read the correct
two bytes though.)

Signed-off-by: Julian Brown <julian@codesourcery.com>
Message-id: ca20462a044848000370318a8bd41dd0a4ed273f.1484929304.git.julian@codesourcery.com
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agotarget/arm: Add cfgend parameter for ARM CPU selection.
Julian Brown [Tue, 7 Feb 2017 18:29:59 +0000 (18:29 +0000)]
target/arm: Add cfgend parameter for ARM CPU selection.

Add a new "cfgend" property which selects whether the CPU resets into
big-endian mode or not.  This setting affects whether we reset with
SCTLR_B (ARMv6 and earlier) or SCTLR_EE (ARMv7 and later) set.

Signed-off-by: Julian Brown <julian@codesourcery.com>
Message-id: 11420d1c49636c1790e60578ee996e51f0f0b835.1484929304.git.julian@codesourcery.com
[PMM: use error_report_err() rather than error_report();
 move the integratorcp changes to their own patch;
 drop an unnecessary extra #include;
 rephrase commit message accordingly;
 move setting of reset_sctlr above registration of cpregs
 so it actually has an effect]
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agohw/arm/integratorcp: Support specifying features via -cpu
Julian Brown [Tue, 7 Feb 2017 18:29:59 +0000 (18:29 +0000)]
hw/arm/integratorcp: Support specifying features via -cpu

Since the integratorcp board creates the CPU object directly
rather than via cpu_arm_init(), we have to call the CPU
class parse_features() method ourselves if we want to
support the user passing features via the -cpu command
line argument as well as just the cpu name. Do so.

Signed-off-by: Julian Brown <julian@codesourcery.com>
[PMM: split out into its own patch]
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agosd: sdhci: check data length during dma_memory_read
Prasad J Pandit [Tue, 7 Feb 2017 18:29:59 +0000 (18:29 +0000)]
sd: sdhci: check data length during dma_memory_read

While doing multi block SDMA transfer in routine
'sdhci_sdma_transfer_multi_blocks', the 's->fifo_buffer' starting
index 'begin' and data length 's->data_count' could end up to be same.
This could lead to an OOB access issue. Correct transfer data length
to avoid it.

Cc: qemu-stable@nongnu.org
Reported-by: Jiang Xin <jiangxin1@huawei.com>
Signed-off-by: Prasad J Pandit <pjp@fedoraproject.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 20170130064736.9236-1-ppandit@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agoaspeed: add a watchdog controller
Cédric Le Goater [Tue, 7 Feb 2017 18:29:59 +0000 (18:29 +0000)]
aspeed: add a watchdog controller

This enables reboot of a guest from U-Boot and Linux.

Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Joel Stanley <joel@jms.id.au>
Message-id: 1485452251-1593-3-git-send-email-clg@kaod.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agowdt: Add Aspeed watchdog device model
Cédric Le Goater [Tue, 7 Feb 2017 18:29:59 +0000 (18:29 +0000)]
wdt: Add Aspeed watchdog device model

The Aspeed SoC includes a set of watchdog timers using 32-bit
decrement counters, which can be based either on the APB clock or
a 1 MHz clock.

The watchdog timer is designed to prevent system deadlock and, in
general, it should be restarted before timeout. When a timeout occurs,
different types of signals can be generated, ARM reset, SOC reset,
System reset, CPU Interrupt, external signal or boot from alternate
block. The current model only performs the system reset function as
this is used by U-Boot and Linux.

Signed-off-by: Joel Stanley <joel@jms.id.au>
Message-id: 1485452251-1593-2-git-send-email-clg@kaod.org
[clg: - fixed compile breakage
      - fixed io region size
      - added watchdog_perform_action() on timer expiry
      - wrote a commit log
      - merged fixes from Andrew Jeffery to scale the reload value ]
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agointegratorcp: adding vmstate for save/restore
Pavel Dovgalyuk [Tue, 7 Feb 2017 18:29:58 +0000 (18:29 +0000)]
integratorcp: adding vmstate for save/restore

VMState added by this patch preserves correct
loading of the integratorcp device state.

Signed-off-by: Pavel Dovgalyuk <pavel.dovgaluk@ispras.ru>
Message-id: 20170131114310.6768.79416.stgit@PASHA-ISP
[PMM: removed unnecessary minimum_version_id_old lines]
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agoMerge remote-tracking branch 'remotes/juanquintela/tags/migration/20170206' into...
Peter Maydell [Tue, 7 Feb 2017 15:29:26 +0000 (15:29 +0000)]
Merge remote-tracking branch 'remotes/juanquintela/tags/migration/20170206' into staging

migration/next for 20170206

# gpg: Signature made Mon 06 Feb 2017 16:13:26 GMT
# gpg:                using RSA key 0xF487EF185872D723
# gpg: Good signature from "Juan Quintela <quintela@redhat.com>"
# gpg:                 aka "Juan Quintela <quintela@trasno.org>"
# Primary key fingerprint: 1899 FF8E DEBF 58CC EE03  4B82 F487 EF18 5872 D723

* remotes/juanquintela/tags/migration/20170206:
  postcopy: Recover block devices on early failure
  Postcopy: Reset state to avoid cleanup assert
  vmstate registration: check return values
  migration: Check for ID length
  vmstate_register_with_alias_id: Take an Error **
  migration: create Migration Incoming State at init time

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agoMerge remote-tracking branch 'remotes/rth/tags/pull-hppa-20170206' into staging
Peter Maydell [Tue, 7 Feb 2017 14:11:52 +0000 (14:11 +0000)]
Merge remote-tracking branch 'remotes/rth/tags/pull-hppa-20170206' into staging

Misc hppa fixes.

# gpg: Signature made Tue 07 Feb 2017 02:28:40 GMT
# gpg:                using RSA key 0xAD1270CC4DD0279B
# gpg: Good signature from "Richard Henderson <rth7680@gmail.com>"
# gpg:                 aka "Richard Henderson <rth@redhat.com>"
# gpg:                 aka "Richard Henderson <rth@twiddle.net>"
# Primary key fingerprint: 9CB1 8DDA F8E8 49AD 2AFC  16A4 AD12 70CC 4DD0 279B

* remotes/rth/tags/pull-hppa-20170206:
  target/hppa: Fix gdb_write_register
  target/hppa: Tidy do_cbranch
  linux-user: define correct UTS machine name for hppa
  linux-user: fix "apt-get update" on linux-user hppa
  linux-user: add hppa magic numbers in qemu-binfmt-conf.sh

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agorules: don't try to create missing include dirs
Daniel P. Berrange [Mon, 6 Feb 2017 11:29:53 +0000 (11:29 +0000)]
rules: don't try to create missing include dirs

In

  commit ba78db44f6532d66a1e704bd44613e841baa2fc5
  Author: Daniel P. Berrange <berrange@redhat.com>
  Date:   Wed Jan 25 16:14:10 2017 +0000

  make: move top level dir to end of include search path

The dir $(BUILD_DIR)/$(@D) was added to the include
path. This would sometimes point to a non-existant
directory, if the sub-dir in question did not contain
any target-independant files (eg tcg/). To deal with
this the rules.mak attempted to create the directory.

While this was succesful, it also caused accidental
creation of files in the parent of the build dir.
e.g. when building common source files into target
specific output files.

Rather than trying to workaround this, just revert
the code that attempted to mkdir the missing include
directories. Instead just turn off the compiler warning
in question as the missing dir is expected & harmless
in general.

NB: you can clean up a build directory parent that has
been filled with empty directories by commit ba78db44f653
using this GNU find command in that parent directory:
  find audio backends block chardev crypto disas fsdev hw io linux-user \
    migration nbd net qapi qom replay slirp target ui util \
    -type d -empty -delete

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Tested-by: Alberto Garcia <berto@igalia.com>
[PMM: added note about how to clean up a polluted directory]
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agotarget/hppa: Fix gdb_write_register
Richard Henderson [Tue, 7 Feb 2017 02:25:31 +0000 (18:25 -0800)]
target/hppa: Fix gdb_write_register

Add a missing break, detected by Coverity.

Reported-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/hppa: Tidy do_cbranch
Richard Henderson [Thu, 26 Jan 2017 21:35:05 +0000 (13:35 -0800)]
target/hppa: Tidy do_cbranch

Removes some dead code detected by Covarity.

Reported-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agolinux-user: define correct UTS machine name for hppa
Laurent Vivier [Thu, 26 Jan 2017 08:04:49 +0000 (09:04 +0100)]
linux-user: define correct UTS machine name for hppa

the correct UTS machine name (as expected by systemd) is "parisc",
not "hppa".

Signed-off-by: Laurent Vivier <laurent@vivier.eu>
Message-Id: <20170126080449.28255-4-laurent@vivier.eu>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agolinux-user: fix "apt-get update" on linux-user hppa
Laurent Vivier [Thu, 26 Jan 2017 08:04:48 +0000 (09:04 +0100)]
linux-user: fix "apt-get update" on linux-user hppa

apt-get was hanging on linux-user hppa.

strace has shown the netlink data stream was not correctly byte swapped.

It appears the fd translator function is unregistered just after it
has been registered, so the translator function is not called.

This patch removes the fd_trans_unregister() after the do_socket()
in the TARGET_NR_socket case.

This fd_trans_unregister() was added by commit
    e36800c linux-user: add signalfd/signalfd4 syscalls
when do_socket() was not registering any fd translator.
And as now it is, we must remove this fd_trans_unregister() to keep them.

Reported-by: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
Tested-by: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>
Message-Id: <20170126080449.28255-3-laurent@vivier.eu>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agolinux-user: add hppa magic numbers in qemu-binfmt-conf.sh
Laurent Vivier [Thu, 26 Jan 2017 08:04:47 +0000 (09:04 +0100)]
linux-user: add hppa magic numbers in qemu-binfmt-conf.sh

As we have now a linux-user HPPA target, we can add it to the list of
supported targets in qemu-binfmt-conf.sh

Signed-off-by: Laurent Vivier <laurent@vivier.eu>
Reviewed-by: Richard Henderson <rth@twiddle.net>
Message-Id: <20170126080449.28255-2-laurent@vivier.eu>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agopostcopy: Recover block devices on early failure
Dr. David Alan Gilbert [Thu, 2 Feb 2017 15:59:09 +0000 (15:59 +0000)]
postcopy: Recover block devices on early failure

An early postcopy failure can be recovered from as long as we know
we haven't sent the command to run the destination.
We have to undo the bdrv_inactivate_all by calling
bdrv_invalidate_cache_all

Note that I'm not using ms->block_inactive because once we've
sent the postcopy package we dont want anything else to try
and recover the block storage on the source; the destination
might have started writing to it.

Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <20170202155909.31784-3-dgilbert@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
7 years agoPostcopy: Reset state to avoid cleanup assert
Dr. David Alan Gilbert [Thu, 2 Feb 2017 15:59:08 +0000 (15:59 +0000)]
Postcopy: Reset state to avoid cleanup assert

On a destination host with no userfault support an incoming
postcopy would cause the state to enter ADVISE before
it realised there was no support, and because it was in ADVISE
state it would perform a cleanup at the end.  Since there
was no support the cleanup function should be unreachable,
but ends up being called and asserting.

Reset the state when we realise we have no support, thus the
cleanup doesn't happen.

Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <20170202155909.31784-2-dgilbert@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
7 years agovmstate registration: check return values
Dr. David Alan Gilbert [Thu, 2 Feb 2017 12:59:56 +0000 (12:59 +0000)]
vmstate registration: check return values

Check qdev's call to vmstate_register_with_alias_id; that gets
most of the common uses; there's hundreds of calls via vmstate_register
which could get fixed over time.

Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Message-Id: <20170202125956.21942-4-dgilbert@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
7 years agomigration: Check for ID length
Dr. David Alan Gilbert [Thu, 2 Feb 2017 12:59:55 +0000 (12:59 +0000)]
migration: Check for ID length

The qdev id of a device can be huge if it's on the end of a chain
of bridges; in reality such chains shouldn't occur but they can
be made to by chaining PCIe bridges together.

The migration format has a number of 256 character long format
limits; check we don't hit them (we already use pstrcat/cpy but
that just protects us from buffer overruns, we fairly quickly
hit an assert).

Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Message-Id: <20170202125956.21942-3-dgilbert@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
7 years agovmstate_register_with_alias_id: Take an Error **
Dr. David Alan Gilbert [Thu, 2 Feb 2017 12:59:54 +0000 (12:59 +0000)]
vmstate_register_with_alias_id: Take an Error **

I'll be adding an error to it in a subsequent patch.

Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Message-Id: <20170202125956.21942-2-dgilbert@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
7 years agomigration: create Migration Incoming State at init time
Juan Quintela [Mon, 23 Jan 2017 21:32:06 +0000 (22:32 +0100)]
migration: create Migration Incoming State at init time

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <1485207141-1941-3-git-send-email-quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
7 years agoMerge remote-tracking branch 'remotes/kraxel/tags/pull-usb-20170206-1' into staging
Peter Maydell [Mon, 6 Feb 2017 11:44:08 +0000 (11:44 +0000)]
Merge remote-tracking branch 'remotes/kraxel/tags/pull-usb-20170206-1' into staging

usb: various bugfixes, mostly xhci.

# gpg: Signature made Mon 06 Feb 2017 11:26:35 GMT
# gpg:                using RSA key 0x4CB6D8EED3E87138
# gpg: Good signature from "Gerd Hoffmann (work) <kraxel@redhat.com>"
# gpg:                 aka "Gerd Hoffmann <gerd@kraxel.org>"
# gpg:                 aka "Gerd Hoffmann (private) <kraxel@gmail.com>"
# Primary key fingerprint: A032 8CFF B93A 17A7 9901  FE7D 4CB6 D8EE D3E8 7138

* remotes/kraxel/tags/pull-usb-20170206-1:
  xhci: fix event queue IRQ handling
  usb: ccid: check ccid apdu length
  xhci: guard xhci_kick_epctx against recursive calls
  xhci: don't kick in xhci_submit and xhci_fire_ctl_transfer
  xhci: rename xhci_complete_packet to xhci_try_complete_packet
  xhci: only free completed transfers
  usb: accept usb3 control requests
  usb/uas: more verbose error message
  hw/usb/dev-hid: Improve guest compatibility of usb-tablet

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agoxhci: fix event queue IRQ handling
Gerd Hoffmann [Fri, 3 Feb 2017 06:51:45 +0000 (07:51 +0100)]
xhci: fix event queue IRQ handling

The qemu xhci emulation doesn't handle the ERDP_EHB flag correctly.

When the host adapter queues a new event the ERDP_EHB flag is set.  The
flag is cleared (via w1c) by the guest when it updates the ERDP (event
ring dequeue pointer) register to notify the host adapter which events
it has fetched.

An IRQ must be raised in case the ERDP_EHB flag flips from clear to set.
If the flag is set already (which implies there are events queued up
which are not yet processed by the guest) xhci must *not* raise a IRQ.

Qemu got that wrong and raised an IRQ on every event, thereby generating
spurious interrupts in case we've queued events faster than the guest
processed them.  This patch fixes that.

With that change in place we also have to check ERDP updates, to see
whenever the guest has fetched all queued events.  In case there are
still pending events set ERDP_EHB and raise an IRQ again, to make sure
the events don't linger unseen forever.

The linux kernel driver and the microsoft windows driver (shipped with
win8+) can deal with the spurious interrupts without problems.  The
renesas windows driver (v2.1.39) which can be used on older windows
versions is quite upset though.  It does spurious ERDP updates now and
then (not every time, seems we must hit a race window for this to
happen), which in turn makes the qemu xhci emulation think the event
ring is full.  Things go south from here ...

tl;dr: This is the "fix xhci on win7" patch.

Cc: M.Cerveny@computer.org
Cc: 1373228@bugs.launchpad.net
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Message-id: 1486104705-13761-1-git-send-email-kraxel@redhat.com

7 years agoMerge remote-tracking branch 'remotes/dgibson/tags/isa-cleanup-20170206' into staging
Peter Maydell [Mon, 6 Feb 2017 10:29:11 +0000 (10:29 +0000)]
Merge remote-tracking branch 'remotes/dgibson/tags/isa-cleanup-20170206' into staging

Allow ISA to be disabled on some platforms (v3)

This makes some cleanups that are a start on allowing ISA to be
compiled out for platforms which don't use it.

I posted this series last November, and it collected a number of R-bs
and no apparent objections.  So, I've now rebased it (trivially) and
am sending a pull request in the hopes of merge.  A lot of the pieces
here don't have a clear maintainer, so I'm sending it directly to
Peter.

Notes:
  * Patch 3/3 triggers a style warning, but that's just because I'm
    moving a C++ // comment verbatim from one file to another

Changes since v2:
  * Trivial rebase

Changes since v1:
  * Fixed some silly compile errors in 3/3 exposed by some
    changes in other headers

# gpg: Signature made Mon 06 Feb 2017 01:37:50 GMT
# gpg:                using RSA key 0x6C38CACA20D9B392
# gpg: Good signature from "David Gibson <david@gibson.dropbear.id.au>"
# gpg:                 aka "David Gibson (Red Hat) <dgibson@redhat.com>"
# gpg:                 aka "David Gibson (ozlabs.org) <dgibson@ozlabs.org>"
# gpg:                 aka "David Gibson (kernel.org) <dwg@kernel.org>"
# Primary key fingerprint: 75F4 6586 AE61 A66C C44E  87DC 6C38 CACA 20D9 B392

* remotes/dgibson/tags/isa-cleanup-20170206:
  Split ISA and sysbus versions of m48t59 device
  Allow ISA bus to be configured out
  Split serial-isa into its own config option

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agousb: ccid: check ccid apdu length
Prasad J Pandit [Thu, 2 Feb 2017 19:22:28 +0000 (00:52 +0530)]
usb: ccid: check ccid apdu length

CCID device emulator uses Application Protocol Data Units(APDU)
to exchange command and responses to and from the host.
The length in these units couldn't be greater than 65536. Add
check to ensure the same. It'd also avoid potential integer
overflow in emulated_apdu_from_guest.

Reported-by: Li Qiang <liqiang6-s@360.cn>
Signed-off-by: Prasad J Pandit <pjp@fedoraproject.org>
Message-id: 20170202192228.10847-1-ppandit@redhat.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
7 years agoxhci: guard xhci_kick_epctx against recursive calls
Gerd Hoffmann [Thu, 2 Feb 2017 11:36:12 +0000 (12:36 +0100)]
xhci: guard xhci_kick_epctx against recursive calls

Track xhci_kick_epctx processing being active in a variable.  Check the
variable before calling xhci_kick_epctx from xhci_kick_ep.  Add an
assert to make sure we don't call recursively into xhci_kick_epctx.

Cc: 1653384@bugs.launchpad.net
Fixes: 94b037f2a451b3dc855f9f2c346e5049a361bd55
Reported-by: Fabian Lesniak <fabian@lesniak-it.de>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Message-id: 1486035372-3621-1-git-send-email-kraxel@redhat.com
Message-id: 1485790607-31399-5-git-send-email-kraxel@redhat.com

7 years agoxhci: don't kick in xhci_submit and xhci_fire_ctl_transfer
Gerd Hoffmann [Mon, 30 Jan 2017 15:36:46 +0000 (16:36 +0100)]
xhci: don't kick in xhci_submit and xhci_fire_ctl_transfer

xhci_submit and xhci_fire_ctl_transfer are is called from
xhci_kick_epctx processing loop only, so there is no need to call
xhci_kick_epctx make sure processing continues.  Also eecursive calls
into xhci_kick_epctx can cause trouble.

Drop the xhci_kick_epctx calls.

Cc: 1653384@bugs.launchpad.net
Fixes: 94b037f2a451b3dc855f9f2c346e5049a361bd55
Reported-by: Fabian Lesniak <fabian@lesniak-it.de>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Message-id: 1485790607-31399-4-git-send-email-kraxel@redhat.com

7 years agoxhci: rename xhci_complete_packet to xhci_try_complete_packet
Gerd Hoffmann [Mon, 30 Jan 2017 15:36:45 +0000 (16:36 +0100)]
xhci: rename xhci_complete_packet to xhci_try_complete_packet

Make clear that this isn't guaranteed to actually complete the transfer,
the usb packet can still be in flight after calling that function.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Message-id: 1485790607-31399-3-git-send-email-kraxel@redhat.com

7 years agoxhci: only free completed transfers
Gerd Hoffmann [Mon, 30 Jan 2017 15:36:44 +0000 (16:36 +0100)]
xhci: only free completed transfers

Most callsites check already, one was missed.

Cc: 1653384@bugs.launchpad.net
Fixes: 94b037f2a451b3dc855f9f2c346e5049a361bd55
Reported-by: Fabian Lesniak <fabian@lesniak-it.de>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Message-id: 1485790607-31399-2-git-send-email-kraxel@redhat.com

7 years agousb: accept usb3 control requests
Gerd Hoffmann [Tue, 31 Jan 2017 13:52:07 +0000 (14:52 +0100)]
usb: accept usb3 control requests

Windows 10 reportedly sends these, so accept them in case
the device in question is a superspeed (usb3) device.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Message-id: 1485870727-21956-2-git-send-email-kraxel@redhat.com

7 years agousb/uas: more verbose error message
Gerd Hoffmann [Tue, 31 Jan 2017 13:52:06 +0000 (14:52 +0100)]
usb/uas: more verbose error message

Print some more details in case we get a unknown
control request, to ease trouble-shooting.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Message-id: 1485870727-21956-1-git-send-email-kraxel@redhat.com

7 years agohw/usb/dev-hid: Improve guest compatibility of usb-tablet
Phil Dennis-Jordan [Wed, 25 Jan 2017 17:24:35 +0000 (18:24 +0100)]
hw/usb/dev-hid: Improve guest compatibility of usb-tablet

 1. Set bInterfaceProtocol to 0x00 for usb-tablet. This should be
    non-zero for boot protocol devices only, which the usb-tablet is not.
 2. Set the usb-tablet's usage to "mouse" in the report descriptor.

The boot protocol of 0x02 specifically confused OS X/macOS' HID driver
stack, causing it to generate additional bogus HID events with relative
motion in addition to the tablet's absolute coordinate events.

Absolute pointing devices with HID Report Descriptor usage of 0x01
(pointing) are treated by the macOS HID driver as analog sticks, and
absolute coordinates are not directly translated to absolute mouse
cursor positions. Changing it to 0x02 (mouse) fixes the problem, and
does not have any adverse effect in other operating systems and
windowing systems. (VMWare does the same thing.)

Signed-off-by: Phil Dennis-Jordan <phil@philjordan.eu>
Message-id: 1485365075-32702-1-git-send-email-phil@philjordan.eu
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
7 years agoSplit ISA and sysbus versions of m48t59 device
David Gibson [Tue, 8 Nov 2016 06:00:35 +0000 (17:00 +1100)]
Split ISA and sysbus versions of m48t59 device

The m48t59 device supports both ISA and direct sysbus attached versions of
the device in the one .c file.  This can be awkward for some embedded
machine types which need the sysbus M48T59, but don't want to pull in the
ISA bus code and its other dependencies.

Therefore, this patch splits out the code for the ISA attached M48T59 into
its own C file.  It will be built when both CONFIG_M48T59 and
CONFIG_ISA_BUS are enabled.

Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
7 years agoAllow ISA bus to be configured out
David Gibson [Wed, 9 Nov 2016 02:31:22 +0000 (13:31 +1100)]
Allow ISA bus to be configured out

Currently, the code to handle the legacy ISA bus is always included in
qemu.  However there are lots of platforms that don't include ISA legacy
devies, and quite a few that have never used ISA legacy devices at all.

This patch allows the ISA bus code to be disabled in the configuration for
platforms where it doesn't make sense.

For now, the default configs are adjusted to include ISA on all platforms
including PCI: anything with PCI can at least in principle add an i82378
PCI->ISA bridge.  Also, CONFIG_IDE_CORE which is already in pci.mak
requires ISA support.

We also explicitly enable ISA on some other non-PCI platforms which include
ISA devices: moxie, sparc and unicore32.  We may want to pare this down in
future.

The platforms that will lose ISA by default are: cris, lm32, microblazeel,
microblaze, openrisc, s390x, tricore, xtensaeb, xtensa.  As far as I can
tell none of these ever used ISA.

Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
7 years agoSplit serial-isa into its own config option
David Gibson [Mon, 8 Feb 2016 00:36:54 +0000 (10:36 +1000)]
Split serial-isa into its own config option

At present, the core device model code for 8250-like serial ports
(serial.c) and the code for serial ports attached to ISA-style legacy IO
(serial-isa.c) are both controlled by the CONFIG_SERIAL variable.

There are lots and lots of embedded platforms that have 8250-like serial
ports but have never had anything resembling ISA legacy IO.  Therefore,
split serial-isa into its own CONFIG_SERIAL_ISA option so it can be
disabled for platforms where it's not appropriate.

For now, I enabled CONFIG_SERIAL_ISA in every default-config where
CONFIG_SERIAL is enabled, excepting microblaze, or32, and xtensa.  As best
as I can tell, those platforms never used legacy ISA, and also don't
include PCI support (which would allow connection of a PCI->ISA bridge
and/or a southbridge including legacy ISA serial ports).

Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Acked-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
7 years agoMerge remote-tracking branch 'remotes/stefanha/tags/block-pull-request' into staging
Peter Maydell [Fri, 3 Feb 2017 14:41:48 +0000 (14:41 +0000)]
Merge remote-tracking branch 'remotes/stefanha/tags/block-pull-request' into staging

# gpg: Signature made Fri 03 Feb 2017 14:37:45 GMT
# gpg:                using RSA key 0x9CA4ABB381AB73C8
# gpg: Good signature from "Stefan Hajnoczi <stefanha@redhat.com>"
# gpg:                 aka "Stefan Hajnoczi <stefanha@gmail.com>"
# Primary key fingerprint: 8695 A8BF D3F9 7CDA AC35  775A 9CA4 ABB3 81AB 73C8

* remotes/stefanha/tags/block-pull-request:
  iothread: enable AioContext polling by default

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agoiothread: enable AioContext polling by default
Stefan Hajnoczi [Thu, 26 Jan 2017 17:01:19 +0000 (17:01 +0000)]
iothread: enable AioContext polling by default

IOThread AioContexts are likely to consist only of event sources like
virtqueue ioeventfds and LinuxAIO completion eventfds that are pollable
from userspace (without system calls).

We recently merged the AioContext polling feature but didn't enable it
by default yet.  I have gone back over the performance data on the
mailing list and picked a default polling value that gave good results.

Let's enable AioContext polling by default so users don't have another
switch they need to set manually.  If performance regressions are found
we can still disable this for the QEMU 2.9 release.

Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Christian Borntraeger <borntraeger@de.ibm.com>
Cc: Karl Rister <krister@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20170126170119.27876-1-stefanha@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
7 years agoMerge remote-tracking branch 'remotes/sstabellini/tags/xen-20170202' into staging
Peter Maydell [Fri, 3 Feb 2017 12:31:40 +0000 (12:31 +0000)]
Merge remote-tracking branch 'remotes/sstabellini/tags/xen-20170202' into staging

Xen 2017/02/02

# gpg: Signature made Thu 02 Feb 2017 18:26:58 GMT
# gpg:                using RSA key 0x894F8F4870E1AE90
# gpg: Good signature from "Stefano Stabellini <sstabellini@kernel.org>"
# gpg:                 aka "Stefano Stabellini <stefano.stabellini@eu.citrix.com>"
# Primary key fingerprint: D04E 33AB A51F 67BA 07D3  0AEA 894F 8F48 70E1 AE90

* remotes/sstabellini/tags/xen-20170202:
  xen: use qdev_unplug() instead of g_free() in xen_pv_find_xendev()
  MAINTAINERS: Update xen-devel mailing list address
  xen-platform: add missing disk unplug option
  xen-platform: add support for unplugging NVMe disks...
  xen-platform: re-structure unplug_disks

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agotci: Remove invalid assertions
Stefan Weil [Thu, 2 Feb 2017 19:56:01 +0000 (20:56 +0100)]
tci: Remove invalid assertions

tb_jmp_insn_offset and tb_jmp_reset_offset are pointers
and cannot be used with ARRAY_SIZE.

Signed-off-by: Stefan Weil <sw@weilnetz.de>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Message-id: 20170202195601.11286-1-sw@weilnetz.de
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agoMerge remote-tracking branch 'remotes/kraxel/tags/pull-vga-20170202-2' into staging
Peter Maydell [Fri, 3 Feb 2017 09:52:51 +0000 (09:52 +0000)]
Merge remote-tracking branch 'remotes/kraxel/tags/pull-vga-20170202-2' into staging

cirrus: multiple bugfixes, including CVE-2017-2615 fix.

# gpg: Signature made Thu 02 Feb 2017 15:03:35 GMT
# gpg:                using RSA key 0x4CB6D8EED3E87138
# gpg: Good signature from "Gerd Hoffmann (work) <kraxel@redhat.com>"
# gpg:                 aka "Gerd Hoffmann <gerd@kraxel.org>"
# gpg:                 aka "Gerd Hoffmann (private) <kraxel@gmail.com>"
# Primary key fingerprint: A032 8CFF B93A 17A7 9901  FE7D 4CB6 D8EE D3E8 7138

* remotes/kraxel/tags/pull-vga-20170202-2:
  cirrus: fix oob access issue (CVE-2017-2615)
  cirrus: fix blit address mask handling
  cirrus: allow zero source pitch in pattern fill rops
  cirrus: handle negative pitch in cirrus_invalidate_region()

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agoMerge remote-tracking branch 'remotes/dgibson/tags/ppc-for-2.9-20170202' into staging
Peter Maydell [Thu, 2 Feb 2017 18:48:06 +0000 (18:48 +0000)]
Merge remote-tracking branch 'remotes/dgibson/tags/ppc-for-2.9-20170202' into staging

ppc patch queue 2017-02-02

This obsoletes ppc-for-2.9-20170112, which had a MacOS build bug.

This is a long overdue ppc pull request for qemu-2.9.  It's been a
long time coming due to some holidays and inconveniently timed
problems with testing.  So, there's a lot in here:

    * More POWER9 instruction implementations for TCG
    * The simpler parts of my CPU compatibility mode cleanup
        * This changes behaviour to prefer compatibility modes over
          "raW" mode for new machine type versions
    * New "40p" machine type which is essentially a modernized and
      cleaned up "prep".  The intention is that it will replace "prep"
      once it has some more testing and polish.
    * Add pseries-2.9 machine type
    * Implement H_SIGNAL_SYS_RESET hypercall
    * Consolidate the two alternate CPU init paths in pseries by
      making it always go through CPU core objects to initialize CPU
    * A number of bugfixes and cleanups
    * Stop the guest timebase when the guest is stopped under KVM.
      This makes the guest system clock also stop when paused, which
      matches the x86 behaviour.
    * Some preliminary cleanups leading towards implementation of the
      POWER9 MMU.

There are also some changes not strictly related to ppc code, but for
its benefit:

    * Limit the pxi-expander-bridge (PXB) device to x86 guests only
      (it's essentially a hack to work around historical x86
      limitations)
    * Some additions to the 128-bit math in host_utils, necessary for
      some of the new instructions.
    * Revise a number of qtests and enable them for ppc

# gpg: Signature made Thu 02 Feb 2017 01:40:16 GMT
# gpg:                using RSA key 0x6C38CACA20D9B392
# gpg: Good signature from "David Gibson <david@gibson.dropbear.id.au>"
# gpg:                 aka "David Gibson (Red Hat) <dgibson@redhat.com>"
# gpg:                 aka "David Gibson (ozlabs.org) <dgibson@ozlabs.org>"
# gpg:                 aka "David Gibson (kernel.org) <dwg@kernel.org>"
# Primary key fingerprint: 75F4 6586 AE61 A66C C44E  87DC 6C38 CACA 20D9 B392

* remotes/dgibson/tags/ppc-for-2.9-20170202: (107 commits)
  hw/ppc/pnv: Use error_report instead of hw_error if a ROM file can't be found
  ppc/kvm: Handle the "family" CPU via alias instead of registering new types
  target/ppc/mmu_hash64: Fix incorrect shift value in amr calculation
  target/ppc/mmu_hash64: Fix printing unsigned as signed int
  tcg/POWER9: NOOP the cp_abort instruction
  target/ppc/debug: Print LPCR register value if register exists
  target-ppc: Add xststdc[sp, dp, qp] instructions
  target-ppc: Add xvtstdc[sp,dp] instructions
  target-ppc: Add MMU model check for booke machines
  ppc: switch to constants within BUILD_BUG_ON
  target/ppc/cpu-models: Fix/remove bad CPU aliases
  target/ppc: Remove unused POWERPC_FAMILY(POWER)
  spapr: clock should count only if vm is running
  ppc: Remove unused function cpu_ppc601_rtc_init()
  target/ppc: Add pcr_supported to POWER9 cpu class definition
  powerpc/cpu-models: rename ISAv3.00 logical PVR definition
  target-ppc: Add xvcv[hpsp, sphp] instructions
  target-ppc: Add xsmulqp instruction
  target-ppc: Add xsdivqp instruction
  target-ppc: Add xscvsdqp and xscvudqp instructions
  ...

# Conflicts:
# hw/pci-bridge/Makefile.objs

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agoxen: use qdev_unplug() instead of g_free() in xen_pv_find_xendev()
Juergen Gross [Wed, 1 Feb 2017 06:52:02 +0000 (07:52 +0100)]
xen: use qdev_unplug() instead of g_free() in xen_pv_find_xendev()

The error exits of xen_pv_find_xendev() free the new xen-device via
g_free() which is wrong.

As the xen-device has been initialized as qdev it must be removed
via qdev_unplug().

This bug has been introduced with commit 3a6c9172ac5951e6dac2b3f6
("xen: create qdev for each backend device").

Reported-by: Roger Pau Monné <roger.pau@citrix.com>
Tested-by: Roger Pau Monné <roger.pau@citrix.com>
Signed-off-by: Juergen Gross <jgross@suse.com>
Reviewed-by: Stefano Stabellini <sstabellini@kernel.org>
7 years agoMerge remote-tracking branch 'remotes/stefanha/tags/tracing-pull-request' into staging
Peter Maydell [Thu, 2 Feb 2017 16:08:28 +0000 (16:08 +0000)]
Merge remote-tracking branch 'remotes/stefanha/tags/tracing-pull-request' into staging

# gpg: Signature made Wed 01 Feb 2017 13:44:32 GMT
# gpg:                using RSA key 0x9CA4ABB381AB73C8
# gpg: Good signature from "Stefan Hajnoczi <stefanha@redhat.com>"
# gpg:                 aka "Stefan Hajnoczi <stefanha@gmail.com>"
# Primary key fingerprint: 8695 A8BF D3F9 7CDA AC35  775A 9CA4 ABB3 81AB 73C8

* remotes/stefanha/tags/tracing-pull-request:
  trace: clean up trace-events files
  qapi: add missing trace_visit_type_enum() call
  trace: improve error reporting when parsing simpletrace header
  trace: update docs to reflect new code generation approach
  trace: switch to modular code generation for sub-directories
  trace: move setting of group name into Makefiles
  trace: move hw/i386/xen events to correct subdir
  trace: move hw/xen events to correct subdir
  trace: move hw/block/dataplane events to correct subdir
  make: move top level dir to end of include search path

# Conflicts:
# Makefile

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agoMerge remote-tracking branch 'remotes/borntraeger/tags/s390x-20170201' into staging
Peter Maydell [Thu, 2 Feb 2017 15:14:24 +0000 (15:14 +0000)]
Merge remote-tracking branch 'remotes/borntraeger/tags/s390x-20170201' into staging

s390x fixes

- build error with old gcc versions
- race between cmma reset and rom/loader resets
- linux-user vs. cpu model

# gpg: Signature made Wed 01 Feb 2017 08:24:47 GMT
# gpg:                using RSA key 0x117BBC80B5A61C7C
# gpg: Good signature from "Christian Borntraeger (IBM) <borntraeger@de.ibm.com>"
# Primary key fingerprint: F922 9381 A334 08F9 DBAB  FBCA 117B BC80 B5A6 1C7C

* remotes/borntraeger/tags/s390x-20170201:
  target/s390x: use "qemu" cpu model in user mode
  s390x/kvm: fix small race reboot vs. cmma
  s390-pci: fix compilation on older GCC versions

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agocirrus: fix oob access issue (CVE-2017-2615)
Li Qiang [Wed, 1 Feb 2017 08:35:01 +0000 (09:35 +0100)]
cirrus: fix oob access issue (CVE-2017-2615)

When doing bitblt copy in backward mode, we should minus the
blt width first just like the adding in the forward mode. This
can avoid the oob access of the front of vga's vram.

Signed-off-by: Li Qiang <liqiang6-s@360.cn>
{ kraxel: with backward blits (negative pitch) addr is the topmost
          address, so check it as-is against vram size ]

Cc: qemu-stable@nongnu.org
Cc: P J P <ppandit@redhat.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Wolfgang Bumiller <w.bumiller@proxmox.com>
Fixes: d3532a0db02296e687711b8cdc7791924efccea0 (CVE-2014-8106)
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Message-id: 1485938101-26602-1-git-send-email-kraxel@redhat.com
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
7 years agoMerge remote-tracking branch 'remotes/cody/tags/block-pull-request' into staging
Peter Maydell [Thu, 2 Feb 2017 14:00:10 +0000 (14:00 +0000)]
Merge remote-tracking branch 'remotes/cody/tags/block-pull-request' into staging

# gpg: Signature made Wed 01 Feb 2017 05:32:23 GMT
# gpg:                using RSA key 0xBDBE7B27C0DE3057
# gpg: Good signature from "Jeffrey Cody <jcody@redhat.com>"
# gpg:                 aka "Jeffrey Cody <jeff@codyprime.org>"
# gpg:                 aka "Jeffrey Cody <codyprime@gmail.com>"
# Primary key fingerprint: 9957 4B4D 3474 90E7 9D98  D624 BDBE 7B27 C0DE 3057

* remotes/cody/tags/block-pull-request:
  sheepdog: reorganize check for overlapping requests
  sheepdog: simplify inflight_aio_head management
  sheepdog: do not use BlockAIOCB
  sheepdog: reorganize coroutine flow
  sheepdog: remove unused cancellation support

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agoMerge remote-tracking branch 'remotes/mst/tags/for_upstream' into staging
Peter Maydell [Thu, 2 Feb 2017 11:03:37 +0000 (11:03 +0000)]
Merge remote-tracking branch 'remotes/mst/tags/for_upstream' into staging

virtio, vhost, pci: fixes, features

generic pci root port support
disable shpc by default
safer version of ARRAY_SIZE and QEMU_BUILD_BUG_ON
fixes and cleanups all over the place

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
# gpg: Signature made Wed 01 Feb 2017 01:38:34 GMT
# gpg:                using RSA key 0x281F0DB8D28D5469
# gpg: Good signature from "Michael S. Tsirkin <mst@kernel.org>"
# gpg:                 aka "Michael S. Tsirkin <mst@redhat.com>"
# 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: (22 commits)
  arm: add trailing ; after MISMATCH_CHECK
  arm: better stub version for MISMATCH_CHECK
  hw/pci: disable pci-bridge's shpc by default
  vhost-user: delete chardev on cleanup
  vhost: skip ROM sections
  virtio: make virtio_should_notify static
  pci: Convert msix_init() to Error and fix callers
  hcd-xhci: check & correct param before using it
  msix: Follow CODING_STYLE
  hw/i386: check if nvdimm is enabled before plugging
  hw/pcie: Introduce Generic PCI Express Root Port
  hw/ioh3420: derive from PCI Express Root Port base class
  hw/pcie: Introduce a base class for PCI Express Root Ports
  intel_iommu: fix and simplify size calculation in process_device_iotlb_desc()
  pci: mark ROMs read-only
  ARRAY_SIZE: check that argument is an array
  compiler: expression version of QEMU_BUILD_BUG_ON
  compiler: rework BUG_ON using a struct
  QEMU_BUILD_BUG_ON: use __COUNTER__
  ppc: switch to constants within BUILD_BUG_ON
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agoMerge remote-tracking branch 'remotes/elmarco/tags/chr-split-pull-request' into staging
Peter Maydell [Thu, 2 Feb 2017 09:50:21 +0000 (09:50 +0000)]
Merge remote-tracking branch 'remotes/elmarco/tags/chr-split-pull-request' into staging

# gpg: Signature made Tue 31 Jan 2017 19:32:40 GMT
# gpg:                using RSA key 0xDAE8E10975969CE5
# gpg: Good signature from "Marc-André Lureau <marcandre.lureau@redhat.com>"
# gpg:                 aka "Marc-André Lureau <marcandre.lureau@gmail.com>"
# 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: 87A9 BD93 3F87 C606 D276  F62D DAE8 E109 7596 9CE5

* remotes/elmarco/tags/chr-split-pull-request: (41 commits)
  char: headers clean-up
  char: move parallel chardev in its own file
  char: move serial chardev to its own file
  char: move pty chardev in its own file
  char: move pipe chardev in its own file
  char: move console in its own file
  char: move stdio in its own file
  char: move file chardev in its own file
  char: move udp chardev in its own file
  char: move socket chardev to its own file
  char: move win-stdio into its own file
  char: move win chardev base class in its own file
  char: move fd chardev in its own file
  char: move QIOChannel-related stuff to char-io.h
  char: remove unused READ_RETRIES
  char: rename and move to header CHR_READ_BUF_LEN
  char: move ringbuf/memory to its own file
  char: move mux to its own file
  char: move null chardev to its own file
  char: make null_chr_write() the default method
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agohw/ppc/pnv: Use error_report instead of hw_error if a ROM file can't be found
Thomas Huth [Wed, 1 Feb 2017 11:03:10 +0000 (12:03 +0100)]
hw/ppc/pnv: Use error_report instead of hw_error if a ROM file can't be found

hw_error() is for CPU related errors only (it dumps the CPU registers
and  calls abort()!), so using error_report() is the better choice
of reporting an error in case we simply did not find a file.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
7 years agoppc/kvm: Handle the "family" CPU via alias instead of registering new types
Thomas Huth [Tue, 31 Jan 2017 13:11:58 +0000 (14:11 +0100)]
ppc/kvm: Handle the "family" CPU via alias instead of registering new types

When running with KVM on POWER, we are registering a "family" CPU
type for the host CPU that we are running on. For example, on all
POWER8-compatible hosts, we register a "POWER8" CPU type, so that
you can always start QEMU with "-cpu POWER8" there, without the
need to know whether you are running on a POWER8, POWER8E or POWER8NVL
host machine.
However, we also have a "POWER8" CPU alias in the ppc_cpu_aliases list
(that is mainly useful for TCG). This leads to two cosmetical drawbacks:
If the user runs QEMU with "-cpu ?", we always claim that POWER8 is an
"alias for POWER8_v2.0" - which is simply not true when running with
KVM on POWER. And when using the 'query-cpu-definitions' QMP call,
there are currently two entries for "POWER8", one for the alias, and
one for the additional registered type.
To solve the two problems, we should rather update the "family" alias
instead of registering a new types. We then only have one "POWER8"
CPU definition around, an alias, which also points to the right
destination.

Buglink: https://bugzilla.redhat.com/show_bug.cgi?id=1396536
Signed-off-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
7 years agotarget/ppc/mmu_hash64: Fix incorrect shift value in amr calculation
Suraj Jitindar Singh [Fri, 13 Jan 2017 06:28:23 +0000 (17:28 +1100)]
target/ppc/mmu_hash64: Fix incorrect shift value in amr calculation

We are calculating the authority mask register key value wrong.

The pte entry contains the key value with the two upper bits and the three
lower bits stored separately. We should use these two portions to get a 5
bit value, not or them together which will only give us a 3 bit value.

Fix this.

Signed-off-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
7 years agotarget/ppc/mmu_hash64: Fix printing unsigned as signed int
Suraj Jitindar Singh [Fri, 13 Jan 2017 06:28:22 +0000 (17:28 +1100)]
target/ppc/mmu_hash64: Fix printing unsigned as signed int

We were printing an unsigned value as a signed value, fix this.

Signed-off-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
7 years agotcg/POWER9: NOOP the cp_abort instruction
Suraj Jitindar Singh [Fri, 13 Jan 2017 06:28:21 +0000 (17:28 +1100)]
tcg/POWER9: NOOP the cp_abort instruction

The cp_abort instruction is used to remove the state of an in progress
copy paste sequence. POWER9 compilers add this in various places, such
as context switches which causes illegal instruction signals since we
don't yet implement this instruction.

Given there is no implementation of the copy paste facility and that we
don't claim to support it, we can just noop this instruction.

Signed-off-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
7 years agotarget/ppc/debug: Print LPCR register value if register exists
Suraj Jitindar Singh [Fri, 13 Jan 2017 06:28:20 +0000 (17:28 +1100)]
target/ppc/debug: Print LPCR register value if register exists

It can be useful when debugging to print the LPCR value.

Thus we add the LPCR to the "info registers" output if the register had
been defined.

Signed-off-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
7 years agotarget-ppc: Add xststdc[sp, dp, qp] instructions
Nikunj A Dadhania [Fri, 13 Jan 2017 09:23:40 +0000 (14:53 +0530)]
target-ppc: Add xststdc[sp, dp, qp] instructions

xststdcsp: VSX Scalar Test Data Class Single-Precision
xststdcdp: VSX Scalar Test Data Class Double-Precision
xststdcqp: VSX Scalar Test Data Class Quad-Precision

Signed-off-by: Nikunj A Dadhania <nikunj@linux.vnet.ibm.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
7 years agotarget-ppc: Add xvtstdc[sp,dp] instructions
Nikunj A Dadhania [Fri, 13 Jan 2017 09:23:39 +0000 (14:53 +0530)]
target-ppc: Add xvtstdc[sp,dp] instructions

xvtstdcsp: VSX Vector Test Data Class Single-Precision
xvtstdcdp: VSX Vector Test Data Class Double-Precision

Signed-off-by: Nikunj A Dadhania <nikunj@linux.vnet.ibm.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
7 years agotarget-ppc: Add MMU model check for booke machines
Valentin Plotkin [Thu, 26 Jan 2017 10:15:59 +0000 (10:15 +0000)]
target-ppc: Add MMU model check for booke machines

Machines bamboo, e500 and virtex-ml507 assume a certain MMU model,
otherwise resulting in unpredictable behavior. Add apropriate checks
into *_init functions.

Signed-off-by: Valentin Plotkin <caliborn@sdf.org>
[regarding virtex parts]
Reviewed-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Tested-by: Edgar E. Iglesias <edgar.iglesias@xilinx.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
7 years agocirrus: fix blit address mask handling
Gerd Hoffmann [Wed, 25 Jan 2017 10:09:56 +0000 (11:09 +0100)]
cirrus: fix blit address mask handling

Apply the cirrus_addr_mask to cirrus_blt_dstaddr and cirrus_blt_srcaddr
right after assigning them, in cirrus_bitblt_start(), instead of having
this all over the place in the cirrus code, and missing a few places.

Reported-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Message-id: 1485338996-17095-1-git-send-email-kraxel@redhat.com

7 years agocirrus: allow zero source pitch in pattern fill rops
Wolfgang Bumiller [Tue, 24 Jan 2017 15:35:38 +0000 (16:35 +0100)]
cirrus: allow zero source pitch in pattern fill rops

The rops used by cirrus_bitblt_common_patterncopy only use
the destination pitch, so the source pitch shoul allowed to
be zero and the blit with used for the range check around the
source address.

Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
Message-id: 1485272138-23249-1-git-send-email-w.bumiller@proxmox.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
7 years agocirrus: handle negative pitch in cirrus_invalidate_region()
Wolfgang Bumiller [Wed, 25 Jan 2017 13:48:57 +0000 (14:48 +0100)]
cirrus: handle negative pitch in cirrus_invalidate_region()

cirrus_invalidate_region() calls memory_region_set_dirty()
on a per-line basis, always ranging from off_begin to
off_begin+bytesperline. With a negative pitch off_begin
marks the top most used address and thus we need to do an
initial shift backwards by a line for negative pitches of
backward blits, otherwise the first iteration covers the
line going from the start offset forwards instead of
backwards.
Additionally since the start address is inclusive, if we
shift by a full `bytesperline` we move to the first address
*not* included in the blit, so we only shift by one less
than bytesperline.

Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
Message-id: 1485352137-29367-1-git-send-email-w.bumiller@proxmox.com

[ kraxel: codestyle fixes ]

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
7 years agotarget/s390x: use "qemu" cpu model in user mode
David Hildenbrand [Mon, 30 Jan 2017 14:50:25 +0000 (15:50 +0100)]
target/s390x: use "qemu" cpu model in user mode

"any" does not exist, therefore resulting in a misleading error message.

Reported-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20170130145025.26475-1-david@redhat.com>
Reviewed-by: Stefan Weil <sw@weilnetz.de>
Reviewed-by: Alexander Graf <agraf@suse.de>
Cc: qemu-stable@nongnu.org
7 years agos390x/kvm: fix small race reboot vs. cmma
Christian Borntraeger [Tue, 24 Jan 2017 21:17:47 +0000 (22:17 +0100)]
s390x/kvm: fix small race reboot vs. cmma

Right now we reset all devices before we reset the cmma states.  This
can result in the host kernel discarding guest pages that were
previously in the unused state but already contain a bios or a -kernel
file before the cmma reset has finished.  This race results in random
guest crashes or hangs during very early reboot.

Fixes: 1cd4e0f6f0a6 ("s390x/cmma: clean up cmma reset")
Cc: qemu-stable@nongnu.org
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
7 years agos390-pci: fix compilation on older GCC versions
Paolo Bonzini [Fri, 27 Jan 2017 13:20:52 +0000 (14:20 +0100)]
s390-pci: fix compilation on older GCC versions

S390PCIBusDevice is typedef'ed earlier in the file, before the hunks
that this patch modifies.  The double typedef causes old versions of
GCC to complain.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <1485523252-88288-1-git-send-email-pbonzini@redhat.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
7 years agosheepdog: reorganize check for overlapping requests
Paolo Bonzini [Tue, 29 Nov 2016 11:32:45 +0000 (12:32 +0100)]
sheepdog: reorganize check for overlapping requests

Wrap the code that was copied repeatedly in the two functions,
sd_aio_setup and sd_aio_complete.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 20161129113245.32724-6-pbonzini@redhat.com
Signed-off-by: Jeff Cody <jcody@redhat.com>
7 years agosheepdog: simplify inflight_aio_head management
Paolo Bonzini [Tue, 29 Nov 2016 11:32:44 +0000 (12:32 +0100)]
sheepdog: simplify inflight_aio_head management

Add to the list in add_aio_request and, indirectly, resend_aioreq.  Inline
free_aio_req in the caller, it does not simply undo alloc_aio_req's job.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 20161129113245.32724-5-pbonzini@redhat.com
Signed-off-by: Jeff Cody <jcody@redhat.com>
7 years agosheepdog: do not use BlockAIOCB
Paolo Bonzini [Tue, 29 Nov 2016 11:32:43 +0000 (12:32 +0100)]
sheepdog: do not use BlockAIOCB

Sheepdog's AIOCB are completely internal entities for a group of
requests and do not need dynamic allocation.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 20161129113245.32724-4-pbonzini@redhat.com
Signed-off-by: Jeff Cody <jcody@redhat.com>
7 years agosheepdog: reorganize coroutine flow
Paolo Bonzini [Tue, 29 Nov 2016 11:32:42 +0000 (12:32 +0100)]
sheepdog: reorganize coroutine flow

Delimit co_recv's lifetime clearly in aio_read_response.

Do a simple qemu_coroutine_enter in aio_read_response, letting
sd_co_writev call sd_write_done.

Handle nr_pending in the same way in sd_co_rw_vector,
sd_write_done and sd_co_flush_to_disk.

Remove sd_co_rw_vector's return value; just leave with no
pending requests.

[Jeff: added missing 'return' back, spotted by Paolo after
       series was applied.]

Signed-off-by: Jeff Cody <jcody@redhat.com>
7 years agosheepdog: remove unused cancellation support
Paolo Bonzini [Tue, 29 Nov 2016 11:32:41 +0000 (12:32 +0100)]
sheepdog: remove unused cancellation support

SheepdogAIOCB is internal to sheepdog.c, hence it is never canceled.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 20161129113245.32724-2-pbonzini@redhat.com
Signed-off-by: Jeff Cody <jcody@redhat.com>
7 years agoarm: add trailing ; after MISMATCH_CHECK
Michael S. Tsirkin [Fri, 27 Jan 2017 16:20:07 +0000 (18:20 +0200)]
arm: add trailing ; after MISMATCH_CHECK

Macro calls without a trailing ; look weird in C, this works as a side
effect of how QEMU_BUILD_BUG_ON is implemented. Fix this up.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
7 years agoarm: better stub version for MISMATCH_CHECK
Michael S. Tsirkin [Fri, 27 Jan 2017 16:17:25 +0000 (18:17 +0200)]
arm: better stub version for MISMATCH_CHECK

stub version of MISMATCH_CHECK is empty so it's easy to misuse for
people not building kvm on arm.  Use QEMU_BUILD_BUG_ON similar to the
non-stub version to make it easier to catch bugs.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
7 years agohw/pci: disable pci-bridge's shpc by default
Marcel Apfelbaum [Wed, 25 Jan 2017 08:44:46 +0000 (10:44 +0200)]
hw/pci: disable pci-bridge's shpc by default

The shpc component is optional while  ACPI hotplug is used
for hot-plugging PCI devices into a PCI-PCI bridge.
Disabling the shpc by default will make slot 0 usable at boot time
and not only for hot-plug, without loosing any functionality.
Older machines will have shpc enabled for compatibility reasons.

Signed-off-by: Marcel Apfelbaum <marcel@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
7 years agovhost-user: delete chardev on cleanup
Marc-André Lureau [Tue, 24 Jan 2017 19:02:58 +0000 (23:02 +0400)]
vhost-user: delete chardev on cleanup

Remove the chardev implicitly when cleaning up the netdev. This
prevents from reusing the chardev since it would be in an incorrect
state with the slave.

Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=1256618

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
7 years agovhost: skip ROM sections
Michael S. Tsirkin [Tue, 24 Jan 2017 17:03:40 +0000 (19:03 +0200)]
vhost: skip ROM sections

vhost does not support RO protections on memory at the moment - adding
ROMs would mean that e.g. a buggy guest might change them in-memory - a
condition from which guest reset does not recover. Not nice.

We also definitely don't want to try logging writes into ROMs -
in particular guests set very high addresses for ROM BARs
so logging these writes would waste a lot of memory.

Maybe ROMs could be supported with the iotlb variant -
not sure, but there seems to be no good reason for virtio
to try to do DMA from ROM. So let's just skip ROM memory.

Suggested-by: Laurent Vivier <lvivier@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Laurent Vivier <lvivier@redhat.com>
Tested-by: Laurent Vivier <lvivier@redhat.com>
7 years agovirtio: make virtio_should_notify static
Paolo Bonzini [Fri, 20 Jan 2017 17:07:51 +0000 (18:07 +0100)]
virtio: make virtio_should_notify static

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
7 years agopci: Convert msix_init() to Error and fix callers
Cao jin [Tue, 17 Jan 2017 06:18:48 +0000 (14:18 +0800)]
pci: Convert msix_init() to Error and fix callers

msix_init() reports errors with error_report(), which is wrong when
it's used in realize().  The same issue was fixed for msi_init() in
commit 1108b2f. In order to make the API change as small as possible,
leave the return value check to later patch.

For some devices(like e1000e, vmxnet3, nvme) who won't fail because of
msix_init's failure, suppress the error report by passing NULL error
object.

Bonus: add comment for msix_init.

CC: Jiri Pirko <jiri@resnulli.us>
CC: Gerd Hoffmann <kraxel@redhat.com>
CC: Dmitry Fleytman <dmitry@daynix.com>
CC: Jason Wang <jasowang@redhat.com>
CC: Michael S. Tsirkin <mst@redhat.com>
CC: Hannes Reinecke <hare@suse.de>
CC: Paolo Bonzini <pbonzini@redhat.com>
CC: Alex Williamson <alex.williamson@redhat.com>
CC: Markus Armbruster <armbru@redhat.com>
CC: Marcel Apfelbaum <marcel@redhat.com>
Signed-off-by: Cao jin <caoj.fnst@cn.fujitsu.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>