trivial patches for 2014-11-02

# gpg: Signature made Sun 02 Nov 2014 11:54:43 GMT using RSA key ID A4C3D7DB
# gpg: Good signature from "Michael Tokarev <mjt@tls.msk.ru>"
# gpg:                 aka "Michael Tokarev <mjt@corpit.ru>"
# gpg:                 aka "Michael Tokarev <mjt@debian.org>"

* remotes/mjt/tags/pull-trivial-patches-2014-11-02: (23 commits)
  vdi: wrapped uuid_unparse() in #ifdef
  tap: fix possible fd leak in net_init_tap
  tap: do not close(fd) in net_init_tap_one
  target-i386: Remove unused model_features_t struct
  tap_int.h: remove repeating NETWORK_SCRIPT defines
  os-posix: reorder parent notification for -daemonize
  pidfile: stop making pidfile error a special case
  os-posix: replace goto again with a proper loop
  os-posix: use global daemon_pipe instead of cryptic fds[1]
  dump: Fix dump-guest-memory termination and use-after-close
  virtio-9p-proxy: improve error messages in connect_namedsocket()
  virtio-9p-proxy: fix error return in proxy_init()
  virtio-9p-proxy: Fix sockfd leak
  target-tricore: check return value before using it
  net/slirp: specify logbase for smbd
  Revert "os-posix: report error message when lock file failed"
  util: Improve os_mem_prealloc error message
  sparse: fix build
  target-arm: A64: remove redundant store
  target-xtensa: mark XtensaConfig structs as unused
  ...
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

The last round of patches for soft freeze.  Includes ivshmem bugfixes,
megasas 2108 emulation, and other small patches here and there.

# gpg: Signature made Fri 31 Oct 2014 17:17:54 GMT using RSA key ID 78C7AE83
# gpg: Good signature from "Paolo Bonzini <bonzini@gnu.org>"
# gpg:                 aka "Paolo Bonzini <pbonzini@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: 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: (35 commits)
  virtio-scsi: fix dataplane
  ivshmem: use error_report
  ivshmem: Fix fd leak on error
  ivshmem: Fix potential OOB r/w access
  ivshmem: validate incoming_posn value from server
  ivshmem: Check ivshmem_read() size argument
  i386: fix breakpoints handling in icount mode
  kvm_stat: Add powerpc support
  kvm_stat: Abstract ioctl numbers
  kvm_stat: Rework platform detection
  kvm_stat: Fix the non-x86 exit reasons
  kvm_stat: Only consider online cpus
  virtio-scsi: Fix num_queue input validation
  scsi: devirtualize unrealize of SCSI devices
  virtio-scsi: Fix memory leak when realize failed
  iscsi: Refuse to open as writable if the LUN is write protected
  kvmvapic: patch_instruction fix
  vl.c: Fix Coverity complaining for vmstate_dump_file
  Add skip_dump flag to ignore memory region during dump
  -machine vmport=off: Allow disabling of VMWare ioport emulation
  ...
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

Commit 9d8bf2d1 moved the softmmu slow path out of line and introduce a
regression at the same time by always calling tcg_out_tlb_load with
is_load=1. This makes impossible to run any significant code under
qemu-system-mips*.

Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: qemu-stable@nongnu.org
Reviewed-by: NRichard Henderson <rth@twiddle.net>
Signed-off-by: NAurelien Jarno <aurelien@aurel32.net>

Wrapped uuid_unparse() in #ifdef to avoid "-Wunused-function"
on clang 3.4 or later.
Signed-off-by: NSeokYeon Hwang <syeon.hwang@samsung.com>
Reviewed-by: NStefan Weil <sw@weilnetz.de>
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>

In hotplugging scenario, taking those true branch, the file
handler do not be closed. Let's close them before return.
Signed-off-by: NGonglei <arei.gonglei@huawei.com>
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>

commit 5193e5fb (tap: factor out common tap initialization)
introduce net_init_tap_one(). But it's inappropriate that
we close fd in net_init_tap_one(), we should lay it in the
caller, becuase some callers needn't to close it if we get
the fd by monitor_handle_fd_param().

On the other hand, in other exceptional branches fd isn't
closed, so that's incomplete anyway.
Signed-off-by: NGonglei <arei.gonglei@huawei.com>
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>

The struct is not used anymore and can be removed.
Signed-off-by: NEduardo Habkost <ehabkost@redhat.com>
Reviewed-by: NIgor Mammedov <imammedo@redhat.com>
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>

DEFAULT_NETWORK_SCRIPT and DEFAULT_NETWORK_DOWN_SCRIPT
have been defined in net/net.h included in
tap.c, which is the only C file that using those two macro.
Let's remove the repeating macroinstruction.
Signed-off-by: NGonglei <arei.gonglei@huawei.com>
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>

Put "success" parent reporting in os_setup_post() to after
all other initializers which may also fail, to the very end,
so more possible failure cases are reported properly to the
calling process.
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>
Reviewed-by: NGonglei <arei.gonglei@huawei.com>

In case of -daemonize, we write non-zero to the daemon
pipe only if pidfile creation failed, so the parent will
report error about pidfile problem.  There's no need to
make special case for this, since all other errors are
reported by the child just fine.  Let the parent report
error and simplify logic in os_daemonize().

This way, we don't need os_pidfile_error() function, since
it only prints error now, so put the error reporting printf
into the only place where qemu_create_pidfile() is called,
in vl.c.

While at it, fix wrong indentation in os_daemonize().
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>

Eliminiate two fullwrite implementations with goto replacing them with
a proper do..while loop.
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>
Reviewed-by: NGonglei <arei.gonglei@huawei.com>

When asked to -daemonize, we fork a child and setup a pipe between
it and parent to pass exit status.  os-posix.c used global fds[2]
array for that, but actually only the writing side of the pipe is
needed to be global, and this name is really too generic.  Use
just one interger for the writing side of the pipe, and name it
daemon_pipe to be more understandable than cryptic fds[1].
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>
Reviewed-by: NGonglei <arei.gonglei@huawei.com>

dump_iterate() dumps blocks in a loop.  Eventually, get_next_block()
returns "no more".  We then call dump_completed().  But we neglect to
break the loop!  Broken in commit 4c7e251a.

Because of that, we dump the last block again.  This attempts to write
to s->fd, which fails if we're lucky.  The error makes dump_iterate()
return failure.  It's the only way it can ever return.

Theoretical: if we're not so lucky, something else has opened something
for writing and got the same fd.  dump_iterate() then keeps looping,
messing up the something else's output, until a write fails, or the
process mercifully terminates.

The obvious fix is to restore the return lost in commit 4c7e251a.  But
the root cause of the bug is needlessly opaque loop control.  Replace it
by a clean do ... while loop.

This makes the badly chosen return values of get_next_block() more
visible.  Cleaning that up is outside the scope of this bug fix.
Signed-off-by: NGonglei <arei.gonglei@huawei.com>
Signed-off-by: NMarkus Armbruster <armbru@redhat.com>
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>

Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>
Reviewed-by: NGonglei <arei.gonglei@huawei.com>

proxy_init() does not check the return value of connect_namedsocket(),
fix this by rearranging code a little bit.
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>

If connect() in connect_namedsocket() return false, the sockfd will leak.
Plug it.
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>
Signed-off-by: NGonglei <arei.gonglei@huawei.com>

We reference the return value of cpu before checking whether it is NULL,
The checking code is after that which violates code style.

It makes no difference if the cpu is NULL, qemu process will terminate.
But one will be 'Segmentation fault' and the other will report a error
which is what we want.
Signed-off-by: Nzhanghailiang <zhang.zhanghailiang@huawei.com>
Reviewed-by: NBastian Koppelmann <kbastian@mail.uni-paderborn.de>
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>

It looks like smbd always logs to /var/log/samba/log.$progname
even if config file specifies different logfile -- when it needs
to log something before completing reading the config file.  But
if it can't open it for writing, it fails and exits.  Tell smbd
to use our temp dir as logbase (-l option) to avoid that.

The same option is used by samba3 and samba4, so there should
be no incompatible changes.
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>
Reviewed-by: NJan Kiszka <jan.kiszka@siemens.com>
Tested-by: NJan Kiszka <jan.kiszka@siemens.com>

This reverts commit e5048d15.

qemu_create_pidfile() is only created from main(), and there,
if that function returns failure, os_pidfile_error() function
is called, to, guess that, report error (which is done differently
whenever we're daemonizing or not).

qemu_create_pidfile() function has several error returns, this
lockf() failure is one of them, there are others (another shown
in the patch context too).

So this patch makes whole thing inconsistent at least.

If we need to show error message when we're daemonizing, it
looks like we should modify os_pidfile_error() routine to always
report error and only after that check for daemon mode.  This way
all errors will be reported the same way.
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>

Currently, when the preallocating guest memory process fails, a not
so helpful error message is printed out:

    # virsh start migt10
    error: Failed to start domain migt10
    error: internal error: process exited while connecting to monitor:
    os_mem_prealloc: failed to preallocate pages

From the error message it's not clear at the first glance where the
problem lies. However, changing the error message might give users a
clue.
Signed-off-by: NMichal Privoznik <mprivozn@redhat.com>
Reviewed-by: NEric Blake <eblake@redhat.com>
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>

c++ compiler isn't wrapped with cgcc, resulting in gcc complaining about
the sparse compiler flags which it doesn't know in case qemu is built
with --enable-sparse.
Signed-off-by: NGerd Hoffmann <kraxel@redhat.com>
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>

There is not much point storing the same value twice in a row.
Reported-by: NLaurent Desnogues <laurent.desnogues@gmail.com>
Signed-off-by: NAlex Bennée <alex.bennee@linaro.org>
Reviewed-by: NLaurent Desnogues <laurent.desnogues@gmail.com>
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>

The XtensaConfig structs will be defined but not used if they are
for the opposite endianness from that of the binary being built;
keep the compiler from complaining about this by marking them
with the 'unused' attribute.
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>
Acked-by: NMax Filippov <jcmvbkbc@gmail.com>
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>

This will avoid unexpected circular header dependencies in the future.
Signed-off-by: NEduardo Habkost <ehabkost@redhat.com>
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>

This removes the following circular dependency:

bitops.h -> qemu-common.h -> target-i386/cpu.h -> target-i386/cpu-qom.h ->
qom/cpu.h -> qdev-core.h -> bitmap.h -> bitops.h.
Signed-off-by: NEduardo Habkost <ehabkost@redhat.com>
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>

The test code needs osdep.h for the ARRAY_SIZE macro.
Signed-off-by: NEduardo Habkost <ehabkost@redhat.com>
Signed-off-by: NMichael Tokarev <mjt@tls.msk.ru>

Commit 361dcc79 (virtio-scsi: dataplane: fail setup gracefully, 2014-10-15)
actually broke successful dataplane setup in a not-so-graceful manner:

    qemu-system-x86_64: .../util/rfifolock.c:71: rfifolock_unlock: Assertion `r->nesting > 0' failed.

due to a missing return statement.

Fixes: 361dcc79Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Replace all the fprintf(stderr, ...) calls with error_report.
Also make sure exit() consistently uses the error code 1. A few calls
used -1. While at it cleanup some indentation in the printf argument
lists.
Signed-off-by: NAndrew Jones <drjones@redhat.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Reported-by: NStefan Hajnoczi <stefanha@redhat.com>
Cc: qemu-stable@nongnu.org
Signed-off-by: NAndreas Färber <afaerber@suse.de>
Reviewed-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Fix OOB access via malformed incoming_posn parameters
and check that requested memory is actually alloc'ed.
Signed-off-by: NSebastian Krahmer <krahmer@suse.de>
[AF: Rebased, cleanups, avoid fd leak]
Cc: qemu-stable@nongnu.org
Signed-off-by: NAndreas Färber <afaerber@suse.de>
Reviewed-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Check incoming_posn to avoid out-of-bounds array accesses if the ivshmem
server on the host sends invalid values.

Cc: Cam Macdonell <cam@cs.ualberta.ca>
Reported-by: NSebastian Krahmer <krahmer@suse.de>
Signed-off-by: NStefan Hajnoczi <stefanha@redhat.com>
[AF: Tighten upper bound check for posn in close_guest_eventfds()]
Cc: qemu-stable@nongnu.org
Signed-off-by: NAndreas Färber <afaerber@suse.de>
Reviewed-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

The third argument to the fd_read() callback implemented by
ivshmem_read() is the number of bytes, not a flags field.  Fix this and
check we received enough bytes before accessing the buffer pointer.

Cc: Cam Macdonell <cam@cs.ualberta.ca>
Reported-by: NSebastian Krahmer <krahmer@suse.de>
Signed-off-by: NStefan Hajnoczi <stefanha@redhat.com>
[AF: Handle partial reads via FIFO]
Reported-by: NPeter Maydell <peter.maydell@linaro.org>
Cc: qemu-stable@nongnu.org
Signed-off-by: NAndreas Färber <afaerber@suse.de>
Reviewed-by: NMichael S. Tsirkin <mst@redhat.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

This patch fixes instructions counting when execution is stopped on
breakpoint (e.g. set from gdb). Without a patch extra instruction is translated
and icount is incremented by invalid value (which equals to number of
executed instructions + 1).
Signed-off-by: NPavel Dovgalyuk <pavel.dovgaluk@ispras.ru>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>
Signed-off-by: NPavel Dovgalyuk <Pavel.Dovgaluk@ispras.ru>

Add support for powerpc platforms. We use uname -m, which allows us to
detect ppc, ppc64 and ppc64le/el.
Signed-off-by: NMichael Ellerman <mpe@ellerman.id.au>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Unfortunately ioctl numbers are platform specific, so abstract them out
of the code so they can be overridden. As it happens x86 and s390 share
the same values, so nothing needs to change yet.
Signed-off-by: NMichael Ellerman <mpe@ellerman.id.au>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

The current platform detection is a little bit messy. We look for lines
in /proc/cpuinfo starting with 'flags' OR 'vendor-id', and scan both
for values we know will only occur in one or the other. We also keep
scanning once we've found a value, which could be a feature, but isn't
in this case.

We'd also like to add another platform, powerpc, which will just make it
worse. So clean it up in preparation.
Signed-off-by: NMichael Ellerman <mpe@ellerman.id.au>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

In kvm_stat we have a dictionary of exit reasons for s390. Firstly these
are not s390 specific, they are the generic exit reasons. So rename the
dictionary to reflect that, and add it separately to filters[].

Secondly, the values are defined using hex, but in the kernel header
they are decimal. That means values above 9 in kvm_stat are incorrect.

While we're there, fix the whitespace to match the rest of the file.
Signed-off-by: NMichael Ellerman <mpe@ellerman.id.au>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

In kvm_stat we grovel through /sys to find out how many cpus are in the
system. However if a cpu is offline it will still be present in /sys,
and the perf_event_open() will fail.

Modify the logic to only return online cpus. We need to be careful on
systems which don't support cpu hotplug, the online file will not be
present at all.
Signed-off-by: NMichael Ellerman <mpe@ellerman.id.au>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

We need to count the ctrlq and eventq, and also cleanup before
returning. Besides, the format string should be unsigned.

The number could never be less than zero.
Signed-off-by: NFam Zheng <famz@redhat.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

All implementations are the same.
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>