OSDN Git Service

qmiga/qemu.git
7 years agoram: Print block stats also in the complete case
Juan Quintela [Thu, 4 May 2017 08:21:46 +0000 (10:21 +0200)]
ram: Print block stats also in the complete case

Once there, create populate_disk_info.

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Laurent Vivier <lvivier@redhat.com>
--

- create populate_disk_info instead of "abusing" populate_ram_info

7 years agomigration: Don't try to set *errp directly
Eduardo Habkost [Thu, 8 Jun 2017 13:39:05 +0000 (10:39 -0300)]
migration: Don't try to set *errp directly

Assigning directly to *errp is not valid, as errp may be NULL,
&error_fatal, or &error_abort.  Use error_propagate() instead.

Cc: Juan Quintela <quintela@redhat.com>
Cc: "Dr. David Alan Gilbert" <dgilbert@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
7 years agomigration: isolate return path on src
Peter Xu [Wed, 31 May 2017 10:35:34 +0000 (18:35 +0800)]
migration: isolate return path on src

There are some places that binded "return path" with postcopy. Let's be
prepared for its usage even without postcopy. This patch mainly did this
on source side.

Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
7 years agoMerge remote-tracking branch 'remotes/elmarco/tags/char-pull-request' into staging
Peter Maydell [Mon, 12 Jun 2017 18:26:49 +0000 (19:26 +0100)]
Merge remote-tracking branch 'remotes/elmarco/tags/char-pull-request' into staging

# gpg: Signature made Thu 08 Jun 2017 15:12:11 BST
# 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/char-pull-request:
  test-char: start a /char/serial test
  chardev: don't use alias names in parse_compat()
  char: fix alias devices regression

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agoMerge remote-tracking branch 'remotes/stefanha/tags/tracing-pull-request' into staging
Peter Maydell [Mon, 12 Jun 2017 13:51:30 +0000 (14:51 +0100)]
Merge remote-tracking branch 'remotes/stefanha/tags/tracing-pull-request' into staging

# gpg: Signature made Wed 07 Jun 2017 19:55:32 BST
# 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:
  simpletrace: Improve the error message if event is not declared

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agoMerge remote-tracking branch 'remotes/stefanha/tags/block-pull-request' into staging
Peter Maydell [Mon, 12 Jun 2017 13:14:42 +0000 (14:14 +0100)]
Merge remote-tracking branch 'remotes/stefanha/tags/block-pull-request' into staging

# gpg: Signature made Wed 07 Jun 2017 19:06:51 BST
# 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:
  configure: split c and cxx extra flags
  coroutine-lock: do not touch coroutine after another one has been entered
  .gdbinit: load QEMU sub-commands when gdb starts
  coccinelle: fix typo in comment
  oslib: strip trailing '\n' from error_setg() string argument

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agoMerge remote-tracking branch 'remotes/kevin/tags/for-upstream' into staging
Peter Maydell [Mon, 12 Jun 2017 09:43:32 +0000 (10:43 +0100)]
Merge remote-tracking branch 'remotes/kevin/tags/for-upstream' into staging

Block layer patches

# gpg: Signature made Fri 09 Jun 2017 12:47:31 BST
# gpg:                using RSA key 0x7F09B272C88F2FD6
# gpg: Good signature from "Kevin Wolf <kwolf@redhat.com>"
# Primary key fingerprint: DC3D EB15 9A9A F95D 3D74  56FE 7F09 B272 C88F 2FD6

* remotes/kevin/tags/for-upstream:
  block: fix external snapshot abort permission error
  block/qcow.c: Fix memory leak in qcow_create()
  qemu-iotests: Test automatic commit job cancel on hot unplug
  commit: Fix use after free in completion
  qemu-iotests: Block migration test
  migration/block: Clean up BBs in block_save_complete()
  migration: Inactivate images after .save_live_complete_precopy()
  block: Fix anonymous BBs in blk_root_inactivate()

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agoblock: fix external snapshot abort permission error
Jeff Cody [Wed, 7 Jun 2017 13:55:22 +0000 (09:55 -0400)]
block: fix external snapshot abort permission error

In external_snapshot_abort(), we try to undo what was done in
external_snapshot_prepare() calling bdrv_replace_node() to swap the
nodes back.  However, we receive a permissions error as writers are
blocked on the old node, which is now the new node backing file.

An easy fix (initially suggested by Kevin Wolf) is to call
bdrv_set_backing_hd() on the new node, to set the backing node to NULL.

Signed-off-by: Jeff Cody <jcody@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
7 years agoblock/qcow.c: Fix memory leak in qcow_create()
Peter Maydell [Mon, 5 Jun 2017 13:55:54 +0000 (14:55 +0100)]
block/qcow.c: Fix memory leak in qcow_create()

Coverity points out that the code path in qcow_create() for
the magic "fat:" backing file name leaks the memory used to
store the filename (CID 1307771). Free the memory before
we overwrite the pointer.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
7 years agoqemu-iotests: Test automatic commit job cancel on hot unplug
Kevin Wolf [Fri, 2 Jun 2017 21:10:10 +0000 (23:10 +0200)]
qemu-iotests: Test automatic commit job cancel on hot unplug

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
7 years agocommit: Fix use after free in completion
Kevin Wolf [Fri, 2 Jun 2017 21:04:55 +0000 (23:04 +0200)]
commit: Fix use after free in completion

The final bdrv_set_backing_hd() could be working on already freed nodes
because the commit job drops its references (through BlockBackends) to
both overlay_bs and top already a bit earlier.

One way to trigger the bug is hot unplugging a disk for which
blockdev_mark_auto_del() cancels the block job.

Fix this by taking BDS-level references while we're still using the
nodes.

Cc: qemu-stable@nongnu.org
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
7 years agoqemu-iotests: Block migration test
Kevin Wolf [Tue, 23 May 2017 12:53:10 +0000 (14:53 +0200)]
qemu-iotests: Block migration test

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Jeff Cody <jcody@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
7 years agomigration/block: Clean up BBs in block_save_complete()
Kevin Wolf [Mon, 22 May 2017 15:17:49 +0000 (17:17 +0200)]
migration/block: Clean up BBs in block_save_complete()

We need to release any block migrations BlockBackends on the source
before successfully completing the migration because otherwise
inactivating the images will fail (inactivation only tolerates device
BBs).

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Jeff Cody <jcody@redhat.com>
7 years agomigration: Inactivate images after .save_live_complete_precopy()
Kevin Wolf [Mon, 22 May 2017 15:10:38 +0000 (17:10 +0200)]
migration: Inactivate images after .save_live_complete_precopy()

Block migration may still access the image during its
.save_live_complete_precopy() implementation, so we should only
inactivate the image afterwards.

Another reason for the change is that inactivating an image fails when
there is still a non-device BlockBackend using it, which includes the
BBs used by block migration. We want to give block migration a chance to
release the BBs before trying to inactivate the image (this will be done
in another patch).

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Jeff Cody <jcody@redhat.com>
7 years agoblock: Fix anonymous BBs in blk_root_inactivate()
Kevin Wolf [Mon, 22 May 2017 15:03:39 +0000 (17:03 +0200)]
block: Fix anonymous BBs in blk_root_inactivate()

blk->name isn't an array, but a pointer that can be NULL. Checking for
an anonymous BB must involve a NULL check first, otherwise we get
crashes.

Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Jeff Cody <jcody@redhat.com>
7 years agotest-char: start a /char/serial test
Marc-André Lureau [Wed, 7 Jun 2017 18:26:38 +0000 (22:26 +0400)]
test-char: start a /char/serial test

Quite limited test, to check that the chardev can be created with a
path and with the tty alias.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
7 years agochardev: don't use alias names in parse_compat()
Marc-André Lureau [Wed, 7 Jun 2017 17:41:15 +0000 (21:41 +0400)]
chardev: don't use alias names in parse_compat()

"parport" is considered "old" since commit 88a946d32d, when "parallel"
was added. Similarly for "tty" in commit d59044ef74d.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
7 years agochar: fix alias devices regression
Marc-André Lureau [Wed, 7 Jun 2017 17:33:29 +0000 (21:33 +0400)]
char: fix alias devices regression

Fix regression from commit 4d43a603c71, where the serial and parallel
headers got removed from char.c, which broke the alias table.

Move the HAVE_CHARDEV_SERIAL/HAVE_CHARDEV_PARPORT to osdep.h instead
of being in separate headers.

Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
7 years agoMerge remote-tracking branch 'remotes/bonzini/tags/for-upstream' into staging
Peter Maydell [Wed, 7 Jun 2017 17:24:08 +0000 (18:24 +0100)]
Merge remote-tracking branch 'remotes/bonzini/tags/for-upstream' into staging

* virtio-scsi use-after-free fix (Fam)
* SMM fixes and improvements for TCG (myself, Mihail)
* irqchip and AddressSpaceDispatch cleanups and fixes (Peter)
* Coverity fix (Stefano)
* NBD cleanups and fixes (Vladimir, Eric, myself)
* RTC accuracy improvements and code cleanups (Guangrong+Yunfang)
* socket error reporting improvement (Daniel)
* GDB XML description for SSE registers (Abdallah)
* kvmclock update fix (Denis)
* SMM memory savings (Gonglei)
* -cpu 486 fix (myself)
* various bugfixes (Roman, Peter, myself, Thomas)
* rtc-test improvement (Guangrong)
* migration throttling fix (Felipe)
* create docs/ subdirectories (myself)

# gpg: Signature made Wed 07 Jun 2017 17:22:07 BST
# gpg:                using RSA key 0xBFFBD25F78C7AE83
# gpg: Good signature from "Paolo Bonzini <bonzini@gnu.org>"
# gpg:                 aka "Paolo Bonzini <pbonzini@redhat.com>"
# Primary key fingerprint: 46F5 9FBD 57D6 12E7 BFD4  E2F7 7E15 100C CD36 69B1
#      Subkey fingerprint: F133 3857 4B66 2389 866C  7682 BFFB D25F 78C7 AE83

* remotes/bonzini/tags/for-upstream: (31 commits)
  docs: create config/, devel/ and spin/ subdirectories
  cpus: reset throttle_thread_scheduled after sleep
  kvm: don't register smram_listener when smm is off
  nbd: make it thread-safe, fix qcow2 over nbd
  target/i386: Add GDB XML description for SSE registers
  i386/kvm: do not zero out segment flags if segment is unusable or not present
  edu: fix memory leak on msi_broken platforms
  linuxboot_dma: compile for i486
  kvmclock: update system_time_msr address forcibly
  nbd: Fully initialize client in case of failed negotiation
  sockets: improve error reporting if UNIX socket path is too long
  i386: fix read/write cr with icount option
  target/i386: use multiple CPU AddressSpaces
  target/i386: enable A20 automatically in system management mode
  virtio-scsi: Unset hotplug handler when unrealize
  exec: simplify phys_page_find() params
  nbd/client.c: use errp instead of LOG
  nbd: add errp to read_sync, write_sync and drop_sync
  nbd: add errp parameter to nbd_wr_syncv()
  nbd: read_sync and friends: return 0 on success
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agodocs: create config/, devel/ and spin/ subdirectories
Paolo Bonzini [Tue, 6 Jun 2017 14:46:26 +0000 (16:46 +0200)]
docs: create config/, devel/ and spin/ subdirectories

Developer documentation should be its own manual.  As a start, move all
developer-oriented files to a separate directory.

Also move non-text files to their own directories: docs/config/ for
QEMU -readconfig input, and docs/spin/ for formal models to be used
with the SPIN model checker.

Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
7 years agocpus: reset throttle_thread_scheduled after sleep
Felipe Franciosi [Fri, 19 May 2017 21:29:50 +0000 (22:29 +0100)]
cpus: reset throttle_thread_scheduled after sleep

Currently, the throttle_thread_scheduled flag is reset back to 0 before
sleeping (as part of the throttling logic). Given that throttle_timer
(well, any timer) may tick with a slight delay, it so happens that under
heavy throttling (ie. close or on CPU_THROTTLE_PCT_MAX) the tick may
schedule a further cpu_throttle_thread() work item after the flag reset,
but before the previous sleep completed. This results on the vCPU thread
sleeping continuously for potentially several seconds in a row.

The chances of that happening can be drastically minimised by resetting
the flag after the sleep.

Signed-off-by: Felipe Franciosi <felipe@nutanix.com>
Signed-off-by: Malcolm Crossley <malcolm@nutanix.com>
Message-Id: <1495229390-18909-1-git-send-email-felipe@nutanix.com>
Acked-by: Jason J. Herne <jjherne@linux.vnet.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
7 years agokvm: don't register smram_listener when smm is off
Gonglei [Thu, 1 Jun 2017 11:35:15 +0000 (19:35 +0800)]
kvm: don't register smram_listener when smm is off

If the user set disable smm by '-machine smm=off', we
should not register smram_listener so that we can
avoid waster memory in kvm since the added sencond
address space.

Meanwhile we should assign value of the global kvm_state
before invoking the kvm_arch_init(), because
pc_machine_is_smm_enabled() may use it by kvm_has_mm().

Signed-off-by: Gonglei <arei.gonglei@huawei.com>
Message-Id: <1496316915-121196-1-git-send-email-arei.gonglei@huawei.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
7 years agonbd: make it thread-safe, fix qcow2 over nbd
Paolo Bonzini [Thu, 1 Jun 2017 10:44:56 +0000 (12:44 +0200)]
nbd: make it thread-safe, fix qcow2 over nbd

NBD is not thread safe, because it accesses s->in_flight without
a CoMutex.  Fixing this will be required for multiqueue.
CoQueue doesn't have spurious wakeups but, when another coroutine can
run between qemu_co_queue_next's wakeup and qemu_co_queue_wait's
re-locking of the mutex, the wait condition can become false and
a loop is necessary.

In fact, it turns out that the loop is necessary even without this
multi-threaded scenario.  A particular sequence of coroutine wakeups
is happening ~80% of the time when starting a guest with qcow2 image
served over NBD (i.e. qemu-nbd --format=raw, and QEMU's -drive option
has -format=qcow2).  This patch fixes that issue too.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
7 years agotarget/i386: Add GDB XML description for SSE registers
Abdallah Bouassida [Thu, 1 Jun 2017 09:33:15 +0000 (11:33 +0200)]
target/i386: Add GDB XML description for SSE registers

Add an XML description for SSE registers (XMM+MXCSR) for both X86
and X86-64 architectures in the GDB stub:
- configure: Define gdb_xml_files for the X86 targets (32 and 64bit).
- gdb-xml/i386-32bit-sse.xml & gdb-xml/i386-64bit-sse.xml: The XML files
that contain a description of the XMM + MXCSR registers.
- gdb-xml/i386-32bit.xml & gdb-xml/i386-64bit.xml: wrappers that include
the XML file of the core registers and the other XML file of the SSE registers.
- target/i386/cpu.c: Modify the gdb_core_xml_file to the new XML wrapper,
  modify the gdb_num_core_regs to fit the registers number defined in each
  XML file.

Signed-off-by: Abdallah Bouassida <abdallah.bouassida@lauterbach.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
7 years agoi386/kvm: do not zero out segment flags if segment is unusable or not present
Roman Pen [Thu, 1 Jun 2017 08:56:04 +0000 (10:56 +0200)]
i386/kvm: do not zero out segment flags if segment is unusable or not present

This is a fix for the problem [1], where VMCB.CPL was set to 0 and interrupt
was taken on userspace stack.  The root cause lies in the specific AMD CPU
behaviour which manifests itself as unusable segment attributes on SYSRET[2].

Here in this patch flags are not touched even segment is unusable or is not
present, therefore CPL (which is stored in DPL field) should not be lost and
will be successfully restored on kvm/svm kernel side.

Also current patch should not break desired behavior described in this commit:

4cae9c97967a ("target-i386: kvm: clear unusable segments' flags in migration")

since present bit will be dropped if segment is unusable or is not present.

This is the second part of the whole fix of the corresponding problem [1],
first part is related to kvm/svm kernel side and does exactly the same:
segment attributes are not zeroed out.

[1] Message id: CAJrWOzD6Xq==b-zYCDdFLgSRMPM-NkNuTSDFEtX=7MreT45i7Q@mail.gmail.com
[2] Message id: 5d120f358612d73fc909f5bfa47e7bd082db0af0.1429841474.git.luto@kernel.org

Signed-off-by: Roman Pen <roman.penyaev@profitbricks.com>
Signed-off-by: Mikhail Sennikovskii <mikhail.sennikovskii@profitbricks.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Radim Krčmář <rkrcmar@redhat.com>
Cc: Michael Chapman <mike@very.puzzling.org>
Cc: qemu-devel@nongnu.org
Message-Id: <20170601085604.12980-1-roman.penyaev@profitbricks.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
7 years agoedu: fix memory leak on msi_broken platforms
Paolo Bonzini [Wed, 31 May 2017 12:56:37 +0000 (14:56 +0200)]
edu: fix memory leak on msi_broken platforms

If msi_init fails, the thread has already been created and the
mutex/condvar are not destroyed.  Initialize everything only
after the point where pci_edu_realize cannot fail.

Reported-by: Markus Armbruster <armbru@redhat.com>
Cc: Peter Xu <peterx@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
7 years agolinuxboot_dma: compile for i486
Paolo Bonzini [Wed, 31 May 2017 12:37:15 +0000 (14:37 +0200)]
linuxboot_dma: compile for i486

The ROM uses the cmovne instruction, which is new in Pentium Pro and does not
work when running QEMU with "-cpu 486".  Avoid producing that instruction.

Suggested-by: Richard W.M. Jones <rjones@redhat.com>
Suggested-by: Thomas Huth <thuth@redhat.com>
Reported-by: Rob Landley <rob@landley.net>
Cc: qemu-stable@nongnu.org
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
7 years agokvmclock: update system_time_msr address forcibly
Denis Plotnikov [Mon, 29 May 2017 10:49:04 +0000 (13:49 +0300)]
kvmclock: update system_time_msr address forcibly

Do an update of system_time_msr address every time before reading
the value of tsc_timestamp from guest's kvmclock page.

There is no other code paths which ensure that qemu has an up-to-date
value of system_time_msr. So, force this update on guest's tsc_timestamp
reading.

This bug causes effect on those nested setups which turn off TPR access
interception for L2 guests and that access being intercepted by L0 doesn't
show up in L1.
Linux bootstrap initiate kvmclock before APIC initializing causing TPR access.
That's why on L1 guests, having TPR interception turned on for L2, the effect
of the bug is not revealed.

This patch fixes this problem by making sure it knows the correct
system_time_msr address every time it is needed.

Signed-off-by: Denis Plotnikov <dplotnikov@virtuozzo.com>
Message-Id: <1496054944-25623-1-git-send-email-dplotnikov@virtuozzo.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
7 years agonbd: Fully initialize client in case of failed negotiation
Eric Blake [Sat, 27 May 2017 03:04:21 +0000 (22:04 -0500)]
nbd: Fully initialize client in case of failed negotiation

If a non-NBD client connects to qemu-nbd, we would end up with
a SIGSEGV in nbd_client_put() because we were trying to
unregister the client's association to the export, even though
we skipped inserting the client into that list.  Easy trigger
in two terminals:

$ qemu-nbd -p 30001 --format=raw file
$ nmap 127.0.0.1 -p 30001

nmap claims that it thinks it connected to a pago-services1
server (which probably means nmap could be updated to learn the
NBD protocol and give a more accurate diagnosis of the open
port - but that's not our problem), then terminates immediately,
so our call to nbd_negotiate() fails.  The fix is to reorder
nbd_co_client_start() to ensure that all initialization occurs
before we ever try talking to a client in nbd_negotiate(), so
that the teardown sequence on negotiation failure doesn't fault
while dereferencing a half-initialized object.

While debugging this, I also noticed that nbd_update_server_watch()
called by nbd_client_closed() was still adding a channel to accept
the next client, even when the state was no longer RUNNING.  That
is fixed by making nbd_can_accept() pay attention to the current
state.

Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1451614

Signed-off-by: Eric Blake <eblake@redhat.com>
Message-Id: <20170527030421.28366-1-eblake@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
7 years agosockets: improve error reporting if UNIX socket path is too long
Daniel P. Berrange [Thu, 25 May 2017 15:53:00 +0000 (16:53 +0100)]
sockets: improve error reporting if UNIX socket path is too long

The 'struct sockaddr_un' only allows 108 bytes for the socket
path.

If the user supplies a path, QEMU uses snprintf() to silently
truncate it when too long. This is undesirable because the user
will then be unable to connect to the path they asked for.

If the user doesn't supply a path, QEMU builds one based on
TMPDIR, but if that leads to an overlong path, it mistakenly
uses error_setg_errno() with a stale errno value, because
snprintf() does not set errno on truncation.

In solving this the code needed some refactoring to ensure we
don't pass 'un.sun_path' directly to any APIs which expect
NUL-terminated strings, because the path is not required to
be terminated.

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Message-Id: <20170525155300.22743-1-berrange@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
7 years agoi386: fix read/write cr with icount option
Mihail Abakumov [Fri, 19 May 2017 09:36:15 +0000 (12:36 +0300)]
i386: fix read/write cr with icount option

Running Windows with icount causes a crash in instruction of write cr.
This patch fixes it.

Reading and writing cr cause an icount read because there are called
cpu_get_apic_tpr and cpu_set_apic_tpr functions. So, there is need
gen_io_start()/gen_io_end() calls.

Signed-off-by: Mihail Abakumov <mikhail.abakumov@ispras.ru>
Message-Id: <ffb376034ff184f2fcbe93d5317d9e76@ispras.ru>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
7 years agotarget/i386: use multiple CPU AddressSpaces
Paolo Bonzini [Wed, 1 Mar 2017 09:34:48 +0000 (10:34 +0100)]
target/i386: use multiple CPU AddressSpaces

This speeds up SMM switches.  Later on it may remove the need to take
the BQL, and it may also allow to reuse code between TCG and KVM.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
7 years agotarget/i386: enable A20 automatically in system management mode
Paolo Bonzini [Thu, 11 May 2017 11:35:28 +0000 (13:35 +0200)]
target/i386: enable A20 automatically in system management mode

Ignore env->a20_mask when running in system management mode.

Reported-by: Anthony Xu <anthony.xu@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <1494502528-12670-1-git-send-email-pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
7 years agoarm_gicv3: Fix ICC_BPR1 reset value when EL3 not implemented
Peter Maydell [Wed, 7 Jun 2017 15:29:29 +0000 (16:29 +0100)]
arm_gicv3: Fix ICC_BPR1 reset value when EL3 not implemented

If EL3 is not implemented (ie only one security state) then the
one and only ICC_BPR1 register behaves like the Non-secure
ICC_BPR1 in an EL3-present configuration. In particular, its
reset value is GIC_MIN_BPR_NS, not GIC_MIN_BPR.

Correct the erroneous reset value; this fixes a problem where
we might hit the assert added in commit a89ff39ee901.

Reported-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Message-id: 1496849369-30282-1-git-send-email-peter.maydell@linaro.org

7 years agoconfigure: split c and cxx extra flags
Bruno Dominguez [Tue, 6 Jun 2017 13:07:47 +0000 (14:07 +0100)]
configure: split c and cxx extra flags

There was no possibility to add specific cxx flags using the configure
file. So A new entrance has been created to support it.

Duplication of information in configure and rules.mak. Taking
QEMU_CFLAGS and add them to QEMU_CXXFLAGS, now the value of
QEMU_CXXFLAGS is stored in config-host.mak, so there is no need for
it.

The makefile for libvixl was adding flags for QEMU_CXXFLAGS in
QEMU_CFLAGS because of the addition in rules.mak. That was removed, so
adding them where it should be.

Signed-off-by: Bruno Dominguez <bru.dominguez@gmail.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 1496754467-20893-1-git-send-email-bru.dominguez@gmail.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
7 years agoMerge remote-tracking branch 'remotes/juanquintela/tags/migration/20170607' into...
Peter Maydell [Wed, 7 Jun 2017 14:06:42 +0000 (15:06 +0100)]
Merge remote-tracking branch 'remotes/juanquintela/tags/migration/20170607' into staging

migration/next for 20170607

# gpg: Signature made Wed 07 Jun 2017 10:02:01 BST
# 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/20170607:
  qemu/migration: fix the double free problem on from_src_file
  ram: Make RAMState dynamic
  ram: Use MigrationStats for statistics
  ram: Move ZERO_TARGET_PAGE inside XBZRLE
  ram: Call migration_page_queue_free() at ram_migration_cleanup()
  ram: We only print throttling information sometimes
  ram: Unfold get_xbzrle_cache_stats() into populate_ram_info()

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agocoroutine-lock: do not touch coroutine after another one has been entered
Roman Pen [Thu, 1 Jun 2017 16:08:47 +0000 (18:08 +0200)]
coroutine-lock: do not touch coroutine after another one has been entered

Submission of requests on linux aio is a bit tricky and can lead to
requests completions on submission path:

44713c9e8547 ("linux-aio: Handle io_submit() failure gracefully")
0ed93d84edab ("linux-aio: process completions from ioq_submit()")

That means that any coroutine which has been yielded in order to wait
for completion can be resumed from submission path and be eventually
terminated (freed).

The following use-after-free crash was observed when IO throttling
was enabled:

 Program received signal SIGSEGV, Segmentation fault.
 [Switching to Thread 0x7f5813dff700 (LWP 56417)]
 virtqueue_unmap_sg (elem=0x7f5804009a30, len=1, vq=<optimized out>) at virtio.c:252
 (gdb) bt
 #0  virtqueue_unmap_sg (elem=0x7f5804009a30, len=1, vq=<optimized out>) at virtio.c:252
                              ^^^^^^^^^^^^^^
                              remember the address

 #1  virtqueue_fill (vq=0x5598b20d21b0, elem=0x7f5804009a30, len=1, idx=0) at virtio.c:282
 #2  virtqueue_push (vq=0x5598b20d21b0, elem=elem@entry=0x7f5804009a30, len=<optimized out>) at virtio.c:308
 #3  virtio_blk_req_complete (req=req@entry=0x7f5804009a30, status=status@entry=0 '\000') at virtio-blk.c:61
 #4  virtio_blk_rw_complete (opaque=<optimized out>, ret=0) at virtio-blk.c:126
 #5  blk_aio_complete (acb=0x7f58040068d0) at block-backend.c:923
 #6  coroutine_trampoline (i0=<optimized out>, i1=<optimized out>) at coroutine-ucontext.c:78

 (gdb) p * elem
 $8 = {index = 77, out_num = 2, in_num = 1,
       in_addr = 0x7f5804009ad8, out_addr = 0x7f5804009ae0,
       in_sg = 0x0, out_sg = 0x7f5804009a50}
       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
       'in_sg' and 'out_sg' are invalid.
       e.g. it is impossible that 'in_sg' is zero,
       instead its value must be equal to:

       (gdb) p/x 0x7f5804009ad8 + sizeof(elem->in_addr[0]) + 2 * sizeof(elem->out_addr[0])
       $26 = 0x7f5804009af0

Seems 'elem' was corrupted.  Meanwhile another thread raised an abort:

 Thread 12 (Thread 0x7f57f2ffd700 (LWP 56426)):
 #0  raise () from /lib/x86_64-linux-gnu/libc.so.6
 #1  abort () from /lib/x86_64-linux-gnu/libc.so.6
 #2  qemu_coroutine_enter (co=0x7f5804009af0) at qemu-coroutine.c:113
 #3  qemu_co_queue_run_restart (co=0x7f5804009a30) at qemu-coroutine-lock.c:60
 #4  qemu_coroutine_enter (co=0x7f5804009a30) at qemu-coroutine.c:119
                           ^^^^^^^^^^^^^^^^^^
                           WTF?? this is equal to elem from crashed thread

 #5  qemu_co_queue_run_restart (co=0x7f57e7f16ae0) at qemu-coroutine-lock.c:60
 #6  qemu_coroutine_enter (co=0x7f57e7f16ae0) at qemu-coroutine.c:119
 #7  qemu_co_queue_run_restart (co=0x7f5807e112a0) at qemu-coroutine-lock.c:60
 #8  qemu_coroutine_enter (co=0x7f5807e112a0) at qemu-coroutine.c:119
 #9  qemu_co_queue_run_restart (co=0x7f5807f17820) at qemu-coroutine-lock.c:60
 #10 qemu_coroutine_enter (co=0x7f5807f17820) at qemu-coroutine.c:119
 #11 qemu_co_queue_run_restart (co=0x7f57e7f18e10) at qemu-coroutine-lock.c:60
 #12 qemu_coroutine_enter (co=0x7f57e7f18e10) at qemu-coroutine.c:119
 #13 qemu_co_enter_next (queue=queue@entry=0x5598b1e742d0) at qemu-coroutine-lock.c:106
 #14 timer_cb (blk=0x5598b1e74280, is_write=<optimized out>) at throttle-groups.c:419

Crash can be explained by access of 'co' object from the loop inside
qemu_co_queue_run_restart():

  while ((next = QSIMPLEQ_FIRST(&co->co_queue_wakeup))) {
      QSIMPLEQ_REMOVE_HEAD(&co->co_queue_wakeup, co_queue_next);
                           ^^^^^^^^^^^^^^^^^^^^
                           on each iteration 'co' is accessed,
                           but 'co' can be already freed

      qemu_coroutine_enter(next);
  }

When 'next' coroutine is resumed (entered) it can in its turn resume
'co', and eventually free it.  That's why we see 'co' (which was freed)
has the same address as 'elem' from the first backtrace.

The fix is obvious: use temporary queue and do not touch coroutine after
first qemu_coroutine_enter() is invoked.

The issue is quite rare and happens every ~12 hours on very high IO
and CPU load (building linux kernel with -j512 inside guest) when IO
throttling is enabled.  With the fix applied guest is running ~35 hours
and is still alive so far.

Signed-off-by: Roman Pen <roman.penyaev@profitbricks.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 20170601160847.23720-1-roman.penyaev@profitbricks.com
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Fam Zheng <famz@redhat.com>
Cc: Stefan Hajnoczi <stefanha@redhat.com>
Cc: Kevin Wolf <kwolf@redhat.com>
Cc: qemu-devel@nongnu.org
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
7 years ago.gdbinit: load QEMU sub-commands when gdb starts
Stefan Hajnoczi [Wed, 17 May 2017 12:40:42 +0000 (13:40 +0100)]
.gdbinit: load QEMU sub-commands when gdb starts

The scripts/qemu-gdb.py file is not easily discoverable.  Add a .gdbinit
file so GDB either loads qemu-gdb.py automatically or prints a message
informing the user how to enable them (some systems disable ./.gdbinit
loading for security reasons).

Symlink .gdbinit and the scripts directory in order to make out-of-tree
builds work.  The scripts directory is used to find the qemu-gdb.py file
specified by a relative path in .gdbinit.

Suggested-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Tested-by: Eric Blake <eblake@redhat.com>
Message-id: 20170517124042.1430-1-stefanha@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
7 years agococcinelle: fix typo in comment
Philippe Mathieu-Daudé [Tue, 16 May 2017 00:11:50 +0000 (21:11 -0300)]
coccinelle: fix typo in comment

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
7 years agooslib: strip trailing '\n' from error_setg() string argument
Philippe Mathieu-Daudé [Tue, 16 May 2017 00:11:49 +0000 (21:11 -0300)]
oslib: strip trailing '\n' from error_setg() string argument

spotted by Coccinelle script scripts/coccinelle/err-bad-newline.cocci

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Laurent Vivier <lvivier@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
7 years agosimpletrace: Improve the error message if event is not declared
Jose Ricardo Ziviani [Mon, 29 May 2017 16:30:04 +0000 (13:30 -0300)]
simpletrace: Improve the error message if event is not declared

Today, if we use a trace-event file which does not declare an event
existing in the log file we'll get the following error:

$ scripts/simpletrace.py trace-events trace-68508
Traceback (most recent call last):
  File "scripts/simpletrace.py", line 242, in <module>
    run(Formatter())
  File "scripts/simpletrace.py", line 217, in run
    process(events, sys.argv[2], analyzer, read_header=read_header)
  File "scripts/simpletrace.py", line 192, in process
    for rec in read_trace_records(edict, log):
  File "scripts/simpletrace.py", line 107, in read_trace_records
    rec = read_record(edict, idtoname, fobj)
  File "scripts/simpletrace.py", line 71, in read_record
    return get_record(edict, idtoname, rechdr, fobj)
  File "scripts/simpletrace.py", line 45, in get_record
    event = edict[name]
KeyError: 'qemu_mutex_locked'

This patch improves this error by adding a hint instead of just that
KeyError log:

$ scripts/simpletrace.py trace-events trace-68508
'qemu_mutex_locked' event is logged but is not declared in the trace
events file, try using trace-events-all instead.

Signed-off-by: Jose Ricardo Ziviani <joserz@linux.vnet.ibm.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-id: 1496075404-8845-1-git-send-email-joserz@linux.vnet.ibm.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
7 years agoMerge remote-tracking branch 'remotes/vivier/tags/m68k-for-2.10-pull-request' into...
Peter Maydell [Wed, 7 Jun 2017 10:56:00 +0000 (11:56 +0100)]
Merge remote-tracking branch 'remotes/vivier/tags/m68k-for-2.10-pull-request' into staging

# gpg: Signature made Wed 07 Jun 2017 10:29:50 BST
# gpg:                using RSA key 0xF30C38BD3F2FBE3C
# gpg: Good signature from "Laurent Vivier <lvivier@redhat.com>"
# gpg:                 aka "Laurent Vivier <laurent@vivier.eu>"
# gpg:                 aka "Laurent Vivier (Red Hat) <lvivier@redhat.com>"
# Primary key fingerprint: CD2F 75DD C8E3 A4DC 2E4F  5173 F30C 38BD 3F2F BE3C

* remotes/vivier/tags/m68k-for-2.10-pull-request:
  target/m68k: implement rtd

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agoMerge remote-tracking branch 'remotes/jasowang/tags/net-pull-request' into staging
Peter Maydell [Wed, 7 Jun 2017 10:16:22 +0000 (11:16 +0100)]
Merge remote-tracking branch 'remotes/jasowang/tags/net-pull-request' into staging

# gpg: Signature made Wed 07 Jun 2017 04:29:20 BST
# gpg:                using RSA key 0xEF04965B398D6211
# gpg: Good signature from "Jason Wang (Jason Wang on RedHat) <jasowang@redhat.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: 215D 46F4 8246 689E C77F  3562 EF04 965B 398D 6211

* remotes/jasowang/tags/net-pull-request:
  Revert "Change net/socket.c to use socket_*() functions" again
  net/rocker: Cleanup the useless return value check

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agotarget/m68k: implement rtd
Laurent Vivier [Mon, 5 Jun 2017 10:00:14 +0000 (12:00 +0200)]
target/m68k: implement rtd

Add "Return and Deallocate" (rtd) instruction.

  RTD #d

    (SP) -> PC
    SP + 4 + d -> SP

Signed-off-by: Laurent Vivier <laurent@vivier.eu>
Reviewed-by: Richard Henderson <rth@twiddle.net>
Tested-By: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>
Message-Id: <20170605100014.22981-1-laurent@vivier.eu>

7 years agoMerge remote-tracking branch 'remotes/rth/tags/pull-s390-20170606' into staging
Peter Maydell [Wed, 7 Jun 2017 09:14:54 +0000 (10:14 +0100)]
Merge remote-tracking branch 'remotes/rth/tags/pull-s390-20170606' into staging

Queued s390 patches

# gpg: Signature made Wed 07 Jun 2017 01:18:29 BST
# 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-s390-20170606: (70 commits)
  target/s390x: addressing exceptions are suppressing
  target/s390x: mark ETF2 and ETF2-ENH facilities as available
  target/s390x: check alignment in CDSG in the !CONFIG_ATOMIC128 case
  target/s390x: implement STORE PAIR TO QUADWORD
  target/s390x: implement LOAD PAIR FROM QUADWORD
  target/s390x: implement TRANSLATE ONE/TWO TO ONE/TWO
  target/s390x: implement TEST DECIMAL
  target/s390x: implement UNPACK UNICODE
  target/s390x: implement UNPACK ASCII
  target/s390x: implement PACK UNICODE
  target/s390x: implement PACK ASCII
  target/s390x: implement MOVE LONG UNICODE
  target/s390x: implement COMPARE LOGICAL LONG UNICODE
  target/s390x: improve MOVE LONG and MOVE LONG EXTENDED
  target/s390x: fix adj_len_to_page
  target/s390x: implement COMPARE LOGICAL LONG
  target/s390x: fix COMPARE LOGICAL LONG EXTENDED
  target/s390x: improve 24-bit and 31-bit lengths read/write
  target/s390x: improve 24-bit and 31-bit addresses write
  target/s390x: improve 24-bit and 31-bit addresses read
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
7 years agoqemu/migration: fix the double free problem on from_src_file
QingFeng Hao [Tue, 6 Jun 2017 05:24:38 +0000 (07:24 +0200)]
qemu/migration: fix the double free problem on from_src_file

In load_snapshot, mis->from_src_file is freed twice, the first free is by
qemu_fclose, the second is by migration_incoming_state_destroy and
it causes Illegal instruction exception. The fix is just to remove the
first free.

This problem is found by qemu-iotests case 068 since commit
"660819b migration: shut src return path unconditionally". The error is:
068 1s ... - output mismatch (see 068.out.bad)
    --- tests/qemu-iotests/068.out 2017-05-06 01:00:26.417270437 +0200
    +++ 068.out.bad 2017-06-03 13:59:55.360274640 +0200
    @@ -6,6 +6,8 @@
     QEMU X.Y.Z monitor - type 'help' for more information
     (qemu) savevm 0
     (qemu) quit
    +./common.config: line 107: 242472 Illegal instruction     (core dumped) ( if [ -n "${QEMU_NEED_PID}" ]; then
    +    echo $BASHPID > "${QEMU_TEST_DIR}/qemu-${_QEMU_HANDLE}.pid";
    +fi; exec "$QEMU_PROG" $QEMU_OPTIONS "$@" )
     QEMU X.Y.Z monitor - type 'help' for more information
    -(qemu) quit
    -*** done
    +(qemu) *** done

Signed-off-by: QingFeng Hao <haoqf@linux.vnet.ibm.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
7 years agoram: Make RAMState dynamic
Juan Quintela [Thu, 4 May 2017 09:46:24 +0000 (11:46 +0200)]
ram: Make RAMState dynamic

We create the variable while we are at migration and we remove it
after migration.

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
7 years agoram: Use MigrationStats for statistics
Juan Quintela [Tue, 6 Jun 2017 17:49:03 +0000 (19:49 +0200)]
ram: Use MigrationStats for statistics

RAM Statistics need to survive migration to make info migrate work, so we
need to store them outside of RAMState.  As we already have an struct
with those fields, just used them. (MigrationStats and XBZRLECacheStats).

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
7 years agoram: Move ZERO_TARGET_PAGE inside XBZRLE
Juan Quintela [Tue, 9 May 2017 14:22:01 +0000 (16:22 +0200)]
ram: Move ZERO_TARGET_PAGE inside XBZRLE

It was only used by XBZRLE anyways.

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
7 years agoram: Call migration_page_queue_free() at ram_migration_cleanup()
Juan Quintela [Thu, 4 May 2017 09:45:01 +0000 (11:45 +0200)]
ram: Call migration_page_queue_free() at ram_migration_cleanup()

We shouldn't be using memory later than that.

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
7 years agoram: We only print throttling information sometimes
Juan Quintela [Wed, 3 May 2017 11:16:38 +0000 (13:16 +0200)]
ram: We only print throttling information sometimes

Change it to be consistent with everything else.

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
7 years agoram: Unfold get_xbzrle_cache_stats() into populate_ram_info()
Juan Quintela [Thu, 4 May 2017 08:09:21 +0000 (10:09 +0200)]
ram: Unfold get_xbzrle_cache_stats() into populate_ram_info()

They were called consecutively always.

Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
7 years agoRevert "Change net/socket.c to use socket_*() functions" again
Daniel P. Berrange [Fri, 5 May 2017 16:23:05 +0000 (17:23 +0100)]
Revert "Change net/socket.c to use socket_*() functions" again

This reverts commit 883e4f7624e10b98d16d9adaffb8b1795664d899.

This code changed net/socket.c from using socket()+connect(),
to using socket_connect(). In theory this is great, but in
practice this has completely broken the ability to connect
the frontend and backend:

  $ ./x86_64-softmmu/qemu-system-x86_64 \
       -device e1000,id=e0,netdev=hn0,mac=DE:AD:BE:EF:AF:05 \
       -netdev socket,id=hn0,connect=localhost:1234
  qemu-system-x86_64: -device e1000,id=e0,netdev=hn0,mac=DE:AD:BE:EF:AF:05: Property 'e1000.netdev' can't find value 'hn0'

The old code would call net_socket_fd_init() synchronously,
while letting the connect() complete in the backgorund. The
new code moved net_socket_fd_init() so that it is only called
after connect() completes in the background.

Thus at the time we initialize the NIC frontend, the backend
does not exist.

The socket_connect() conversion as done is a bad fit for the
current code, since it did not try to change the way it deals
with async connection completion. Rather than try to fix this,
just revert the socket_connect() conversion entirely.

The code is about to be converted to use QIOChannel which
will let the problem be solved in a cleaner manner. This
revert is more suitable for stable branches in the meantime.

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
7 years agonet/rocker: Cleanup the useless return value check
Mao Zhongyi [Wed, 24 May 2017 02:57:18 +0000 (10:57 +0800)]
net/rocker: Cleanup the useless return value check

None of pci_dma_read()'s callers check the return value except
rocker. There is no need to check it because it always return
0. So the check work is useless. Remove it entirely.

Suggested-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Mao Zhongyi <maozy.fnst@cn.fujitsu.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
7 years agotarget/s390x: addressing exceptions are suppressing
David Hildenbrand [Mon, 29 May 2017 12:12:28 +0000 (14:12 +0200)]
target/s390x: addressing exceptions are suppressing

We have to make the address in the old PSW point at the next
instruction, as addressing exceptions are suppressing and not
nullifying.

I assume that there are a lot of other broken cases (as most instructions
we care about are suppressing) - all trigger_pgm_exception() specifying
and explicit number or ILEN_LATER look suspicious, however this is another
story that might require bigger changes (and I have to understand when
the address might already have been incremented first).

This is needed to make an upcoming kvm-unit-test work.

Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20170529121228.2789-1-david@redhat.com>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: mark ETF2 and ETF2-ENH facilities as available
Aurelien Jarno [Wed, 31 May 2017 22:01:28 +0000 (00:01 +0200)]
target/s390x: mark ETF2 and ETF2-ENH facilities as available

Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Message-Id: <20170531220129.27724-30-aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: check alignment in CDSG in the !CONFIG_ATOMIC128 case
Aurelien Jarno [Sun, 4 Jun 2017 20:20:34 +0000 (22:20 +0200)]
target/s390x: check alignment in CDSG in the !CONFIG_ATOMIC128 case

The CDSG instruction requires a 16-byte alignement, as expressed in
the MO_ALIGN_16 passed to helper_atomic_cmpxchgo_be_mmu. In the non
parallel case, use check_alignment to enforce this.

Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Message-Id: <20170604202034.16615-4-aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: implement STORE PAIR TO QUADWORD
Aurelien Jarno [Sun, 4 Jun 2017 20:20:33 +0000 (22:20 +0200)]
target/s390x: implement STORE PAIR TO QUADWORD

Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Message-Id: <20170604202034.16615-3-aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: implement LOAD PAIR FROM QUADWORD
Aurelien Jarno [Sun, 4 Jun 2017 20:20:32 +0000 (22:20 +0200)]
target/s390x: implement LOAD PAIR FROM QUADWORD

Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Message-Id: <20170604202034.16615-2-aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: implement TRANSLATE ONE/TWO TO ONE/TWO
Aurelien Jarno [Wed, 31 May 2017 22:01:27 +0000 (00:01 +0200)]
target/s390x: implement TRANSLATE ONE/TWO TO ONE/TWO

Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Message-Id: <20170531220129.27724-29-aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: implement TEST DECIMAL
Aurelien Jarno [Wed, 31 May 2017 22:01:26 +0000 (00:01 +0200)]
target/s390x: implement TEST DECIMAL

Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Message-Id: <20170531220129.27724-28-aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: implement UNPACK UNICODE
Aurelien Jarno [Wed, 31 May 2017 22:01:25 +0000 (00:01 +0200)]
target/s390x: implement UNPACK UNICODE

Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Message-Id: <20170531220129.27724-27-aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: implement UNPACK ASCII
Aurelien Jarno [Wed, 31 May 2017 22:01:24 +0000 (00:01 +0200)]
target/s390x: implement UNPACK ASCII

Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Message-Id: <20170531220129.27724-26-aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: implement PACK UNICODE
Aurelien Jarno [Wed, 31 May 2017 22:01:23 +0000 (00:01 +0200)]
target/s390x: implement PACK UNICODE

Use a common helper with PACK ASCII as the differences are limited to
the stride of the source operand.

Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Message-Id: <20170531220129.27724-25-aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: implement PACK ASCII
Aurelien Jarno [Wed, 31 May 2017 22:01:22 +0000 (00:01 +0200)]
target/s390x: implement PACK ASCII

Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Message-Id: <20170531220129.27724-24-aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: implement MOVE LONG UNICODE
Aurelien Jarno [Wed, 31 May 2017 22:01:21 +0000 (00:01 +0200)]
target/s390x: implement MOVE LONG UNICODE

Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Message-Id: <20170531220129.27724-23-aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: implement COMPARE LOGICAL LONG UNICODE
Aurelien Jarno [Wed, 31 May 2017 22:01:20 +0000 (00:01 +0200)]
target/s390x: implement COMPARE LOGICAL LONG UNICODE

For that we need to make program_interrupt available to qemu-user.
Fortunately there is almost nothing to change as both kvm_enabled and
CONFIG_KVM evaluate to false in that case.

Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Message-Id: <20170531220129.27724-22-aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: improve MOVE LONG and MOVE LONG EXTENDED
Aurelien Jarno [Wed, 31 May 2017 22:01:19 +0000 (00:01 +0200)]
target/s390x: improve MOVE LONG and MOVE LONG EXTENDED

As MVCL and MVCLE only differ by their operands, use a common
do_mvcl helper. Optimize it calling fast_memmove and fast_memset.
Correctly write back addresses. Check that r1 and r2/r3 registers
are even.

Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Message-Id: <20170531220129.27724-21-aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: fix adj_len_to_page
Aurelien Jarno [Wed, 31 May 2017 22:01:18 +0000 (00:01 +0200)]
target/s390x: fix adj_len_to_page

adj_len_to_page doesn't return the correct result when the address
is already page aligned and the length is bigger than a page. Fix that.

Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Message-Id: <20170531220129.27724-20-aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: implement COMPARE LOGICAL LONG
Aurelien Jarno [Wed, 31 May 2017 22:01:17 +0000 (00:01 +0200)]
target/s390x: implement COMPARE LOGICAL LONG

As CLCL and CLCLE mostly differ by their operands, use a common do_clcl
helper. Another difference is that CLCL is not interruptible.

Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Message-Id: <20170531220129.27724-19-aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: fix COMPARE LOGICAL LONG EXTENDED
Aurelien Jarno [Wed, 31 May 2017 22:01:16 +0000 (00:01 +0200)]
target/s390x: fix COMPARE LOGICAL LONG EXTENDED

There are multiple issues with the COMPARE LOGICAL LONG EXTENDED
instruction:
- The test between the two operands is inverted, leading to an inversion
  of the cc values 1 and 2.
- The address and length of an operand continue to be decreased after
  reaching the end of this operand. These values are then wrong write
  back to the registers.
- We should limit the amount of bytes to process, so that interrupts can
  be served correctly.

At the same time rename dest into src1 and src into src3 to match the
operand names and make the code less confusing.

Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Message-Id: <20170531220129.27724-18-aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: improve 24-bit and 31-bit lengths read/write
Aurelien Jarno [Wed, 31 May 2017 22:01:15 +0000 (00:01 +0200)]
target/s390x: improve 24-bit and 31-bit lengths read/write

Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Message-Id: <20170531220129.27724-17-aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: improve 24-bit and 31-bit addresses write
Aurelien Jarno [Wed, 31 May 2017 22:01:14 +0000 (00:01 +0200)]
target/s390x: improve 24-bit and 31-bit addresses write

Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Message-Id: <20170531220129.27724-16-aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: improve 24-bit and 31-bit addresses read
Aurelien Jarno [Wed, 31 May 2017 22:01:13 +0000 (00:01 +0200)]
target/s390x: improve 24-bit and 31-bit addresses read

Improve fix_address to also handle the 24-bit mode. Rename fix_address
to wrap_address to better explain what is changed.

Replace the calls to get_address with x2 = 0 and b2 = 0 by
call to wrap_address, leading to the removal of this function. Rename
get_address_31fix into get_address.

Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Message-Id: <20170531220129.27724-15-aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: implement MOVE ZONES
Aurelien Jarno [Wed, 31 May 2017 22:01:12 +0000 (00:01 +0200)]
target/s390x: implement MOVE ZONES

Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Message-Id: <20170531220129.27724-14-aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: implement MOVE WITH OFFSET
Aurelien Jarno [Wed, 31 May 2017 22:01:11 +0000 (00:01 +0200)]
target/s390x: implement MOVE WITH OFFSET

Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Message-Id: <20170531220129.27724-13-aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: implement MOVE NUMERICS
Aurelien Jarno [Wed, 31 May 2017 22:01:10 +0000 (00:01 +0200)]
target/s390x: implement MOVE NUMERICS

Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Message-Id: <20170531220129.27724-12-aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: implement MOVE INVERSE
Aurelien Jarno [Wed, 31 May 2017 22:01:09 +0000 (00:01 +0200)]
target/s390x: implement MOVE INVERSE

Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Message-Id: <20170531220129.27724-11-aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: implement COMPARE AND SIGNAL
Aurelien Jarno [Wed, 31 May 2017 22:01:08 +0000 (00:01 +0200)]
target/s390x: implement COMPARE AND SIGNAL

These functions differ from COMPARE by generating an exception for a
QNaN input. Use the non quiet version of floatXX_compare.

Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Message-Id: <20170531220129.27724-10-aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: implement PACK
Aurelien Jarno [Wed, 31 May 2017 22:01:05 +0000 (00:01 +0200)]
target/s390x: implement PACK

Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Message-Id: <20170531220129.27724-7-aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: implement TEST ADDRESSING MODE
Aurelien Jarno [Wed, 31 May 2017 22:01:04 +0000 (00:01 +0200)]
target/s390x: implement TEST ADDRESSING MODE

Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Message-Id: <20170531220129.27724-6-aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: implement TEST AND SET
Aurelien Jarno [Wed, 31 May 2017 22:01:03 +0000 (00:01 +0200)]
target/s390x: implement TEST AND SET

Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Message-Id: <20170531220129.27724-5-aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: implement local-TLB-clearing in IPTE
Aurelien Jarno [Wed, 31 May 2017 22:01:02 +0000 (00:01 +0200)]
target/s390x: implement local-TLB-clearing in IPTE

And at the same time make IPTE SMP aware.

Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Message-Id: <20170531220129.27724-4-aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: remove some Linux assumptions from IPTE
Aurelien Jarno [Wed, 31 May 2017 22:01:01 +0000 (00:01 +0200)]
target/s390x: remove some Linux assumptions from IPTE

Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Message-Id: <20170531220129.27724-3-aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: remove dead code in translate.c
Aurelien Jarno [Wed, 31 May 2017 22:01:00 +0000 (00:01 +0200)]
target/s390x: remove dead code in translate.c

Reviewed-by: Richard Henderson <rth@twiddle.net>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Message-Id: <20170531220129.27724-2-aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x/cpu_models: Allow some additional feature bits for the "qemu" CPU
Thomas Huth [Thu, 25 May 2017 09:22:12 +0000 (11:22 +0200)]
target/s390x/cpu_models: Allow some additional feature bits for the "qemu" CPU

Currently we only present the plain z900 feature bits to the guest,
but QEMU already emulates some additional features (but not all of
the next CPU generation, so we can not use the next CPU level as
default yet). Since newer Linux kernels are checking the feature bits
and refuse to work if a required feature is missing, it would be nice
to have a way to present more of the supported features when we are
running with the "qemu" CPU.
This patch now adds the supported features to the "full_feat" bitmap,
so that additional features can be enabled on the command line now,
for example with:

 qemu-system-s390x -cpu qemu,stfle=true,ldisp=true,eimm=true,stckf=true

Acked-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <1495704132-5675-1-git-send-email-thuth@redhat.com>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: Re-implement a few EXECUTE target insns directly
Richard Henderson [Wed, 24 May 2017 21:34:10 +0000 (14:34 -0700)]
target/s390x: Re-implement a few EXECUTE target insns directly

While the previous patch is required for proper conformance,
the vast majority of target insns are MVC and XC for implementing
memmove and memset respectively.  The next most common are CLC,
TR, and SVC.

Implementing these (and a few others for which we already have
an implementation) directly is faster than going through full
translation to a TB.

Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: Implement EXECUTE via new TranslationBlock
Richard Henderson [Sun, 21 May 2017 16:50:00 +0000 (09:50 -0700)]
target/s390x: Implement EXECUTE via new TranslationBlock

Previously, helper_ex would construct the insn and then implement
the insn via direct calls other helpers.  This was sufficient to
boot Linux but that is all.

It is easy enough to go the whole nine yards by stashing state for
EXECUTE within the cpu, and then rely on a new TB to be created
that properly and completely interprets the insn.

Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: End the TB after EXECUTE
Richard Henderson [Wed, 24 May 2017 20:00:16 +0000 (13:00 -0700)]
target/s390x: End the TB after EXECUTE

This split will be required for implementing EXECUTE properly.
Do this now as a separate step to aid comparison of before and
after TB listings.

Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: Save current ilen during translation
Richard Henderson [Wed, 24 May 2017 19:27:36 +0000 (12:27 -0700)]
target/s390x: Save current ilen during translation

Use this saved value instead of recomputing from next_pc difference.

Reviewed-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: Implement CSPG
Richard Henderson [Tue, 23 May 2017 01:58:42 +0000 (18:58 -0700)]
target/s390x: Implement CSPG

Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: Use atomic operations for COMPARE SWAP PURGE
Richard Henderson [Tue, 23 May 2017 01:34:42 +0000 (18:34 -0700)]
target/s390x: Use atomic operations for COMPARE SWAP PURGE

Also provide the cross-cpu tlb flushing required by the PoO.

Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: Fix EXECUTE with R1==0
Richard Henderson [Wed, 24 May 2017 18:56:15 +0000 (11:56 -0700)]
target/s390x: Fix EXECUTE with R1==0

The PoO specifies that when R1==0, no ORing into the insn
loaded from storage takes place.  Load a zero for this case.

Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: Fix some helper_ex problems
Richard Henderson [Wed, 24 May 2017 18:49:53 +0000 (11:49 -0700)]
target/s390x: Fix some helper_ex problems

(1) The OR of the low bits or R1 into INSN were not being done
consistently; it was forgotten along all but the SVC path.
(2) The setting of ILEN was wrong on SVC path for EXRL.
(3) The data load for ICM read too much.

Fix these by consolidating data load at the beginning, using
get_ilen to control the number of bytes loaded, and ORing in
the byte from R1.  Use extract64 from the full aligned insn
to extract arguments.

Pass in ILEN rather than RET as the more natural way to give
the required data along the SVC path.

Modify ENV->CC_OP directly rather than include it in the
functional interface.

Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: Use unwind data for helper_mvcs/mvcp
Richard Henderson [Tue, 23 May 2017 02:45:43 +0000 (19:45 -0700)]
target/s390x: Use unwind data for helper_mvcs/mvcp

Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: Use unwind data for helper_lra
Richard Henderson [Mon, 22 May 2017 20:59:02 +0000 (13:59 -0700)]
target/s390x: Use unwind data for helper_lra

Fix saving exception_index around mmu_translate; eliminate a dead store.

Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: Use unwind data for helper_tprot
Richard Henderson [Mon, 22 May 2017 19:54:35 +0000 (12:54 -0700)]
target/s390x: Use unwind data for helper_tprot

Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: Use unwind data for helper_testblock
Richard Henderson [Mon, 22 May 2017 19:53:32 +0000 (12:53 -0700)]
target/s390x: Use unwind data for helper_testblock

Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: Use unwind data for helper_stctl
Richard Henderson [Mon, 22 May 2017 19:43:23 +0000 (12:43 -0700)]
target/s390x: Use unwind data for helper_stctl

Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>
7 years agotarget/s390x: Use unwind data for helper_lctl
Richard Henderson [Mon, 22 May 2017 19:18:42 +0000 (12:18 -0700)]
target/s390x: Use unwind data for helper_lctl

Reviewed-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Aurelien Jarno <aurelien@aurel32.net>
Signed-off-by: Richard Henderson <rth@twiddle.net>