tcg-mips: Split out tcg_out_call

Signed-off-by: Richard Henderson <rth@twiddle.net>

tcg-sparc: Create tcg_out_call

Rename the existing tcg_out_calli to tcg_out_call_nodelay.

Signed-off-by: Richard Henderson <rth@twiddle.net>

tcg-ppc64: Rename tcg_out_calli to tcg_out_call

Merge the existing tcg_out_call into tcg_out_op.

Signed-off-by: Richard Henderson <rth@twiddle.net>

tcg-ppc: Split out tcg_out_call

Signed-off-by: Richard Henderson <rth@twiddle.net>

tcg-s390: Rename tgen_calli to tcg_out_call

Signed-off-by: Richard Henderson <rth@twiddle.net>

tcg-i386: Rename tcg_out_calli to tcg_out_call

Signed-off-by: Richard Henderson <rth@twiddle.net>

tcg: Require TCG_TARGET_INSN_UNIT_SIZE

Now that all backends do define TCG_TARGET_INSN_UNIT_SIZE,
remove the fallback definition.

Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <rth@twiddle.net>

tci: Define TCG_TARGET_INSN_UNIT_SIZE

And use tcg pointer differencing functions as appropriate.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <rth@twiddle.net>

tcg-mips: Define TCG_TARGET_INSN_UNIT_SIZE

And use tcg pointer differencing functions as appropriate.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <rth@twiddle.net>

tcg-ia64: Define TCG_TARGET_INSN_UNIT_SIZE

Using a 16-byte aligned structure achieves best results, both for code
cleanliness and compiled code size.  However, this means that we can't
use the trick of encoding the slot number into the low 2 bits.

Thankfully, we only ever use slot2, so make that explicit in the names
of the relocation functions, and drop the code for other slots.

Signed-off-by: Richard Henderson <rth@twiddle.net>

tcg-s390: Define TCG_TARGET_INSN_UNIT_SIZE

And use tcg pointer differencing functions as appropriate.

Signed-off-by: Richard Henderson <rth@twiddle.net>

tcg-aarch64: Define TCG_TARGET_INSN_UNIT_SIZE

And use tcg pointer differencing functions as appropriate.

Acked-by: Claudio Fontana <claudio.fontana@huawei.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <rth@twiddle.net>

tcg-arm: Define TCG_TARGET_INSN_UNIT_SIZE

And use tcg pointer differencing functions as appropriate.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <rth@twiddle.net>

tcg-sparc: Define TCG_TARGET_INSN_UNIT_SIZE

And use tcg pointer differencing functions as appropriate.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <rth@twiddle.net>

tcg-ppc: Define TCG_TARGET_INSN_UNIT_SIZE

And use tcg pointer differencing functions as appropriate.

Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <rth@twiddle.net>

tcg-ppc64: Define TCG_TARGET_INSN_UNIT_SIZE

And use tcg pointer differencing functions as appropriate.

Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <rth@twiddle.net>

tcg-i386: Define TCG_TARGET_INSN_UNIT_SIZE

And use tcg pointer differencing functions as appropriate.

Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <rth@twiddle.net>

tcg: Define tcg_insn_unit for code pointers

To be defined by the tcg backend based on the elemental unit of the ISA.
During the transition, allow TCG_TARGET_INSN_UNIT_SIZE to be undefined,
which allows us to default tcg_insn_unit to the current uint8_t.

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <rth@twiddle.net>

tcg: Introduce byte pointer arithmetic helpers

Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <rth@twiddle.net>

tcg: Avoid undefined behaviour patching code at unaligned addresses

To avoid C undefined behaviour when patching generated code,
provide wrappers tcg_patch8/16/32/64 which use the usual memcpy
trick, and use them in the i386 backend.

Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <rth@twiddle.net>

tcg: Avoid stores to unaligned addresses

Avoid stores to unaligned addresses in TCG code generation, by using the
usual memcpy() approach. (Using bswap.h would drag a lot of QEMU baggage
into TCG, so it's simpler just to do direct memcpy() here.)

Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <rth@twiddle.net>

exec-all.h: Use stl_p to avoid undefined behaviour patching x86 jumps

The code which patches x86 jump instructions assumes it can do an
unaligned write of a uint32_t. This is actually safe on x86, but it's
still undefined behaviour. We have infrastructure for doing efficient
unaligned accesses which doesn't engage in undefined behaviour, so
use it.

This is technically fractionally less efficient, at least with gcc 4.6;
instead of one instruction:
 7b2:   89 3e                   mov    %edi,(%rsi)
we get an extra spurious store to the stack slot:
 7b2:   89 7c 24 64             mov    %edi,0x64(%rsp)
 7b6:   89 3e                   mov    %edi,(%rsi)

Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <rth@twiddle.net>

Merge remote-tracking branch 'remotes/qmp-unstable/queue/qmp' into staging

* remotes/qmp-unstable/queue/qmp: (38 commits)
  Revert "qapi: Clean up superfluous null check in qapi_dealloc_type_str()"
  qapi: Document optional arguments' backwards compatibility
  qmp: use valid JSON in transaction example
  qmp: Don't use error_is_set() to suppress additional errors
  dump: Drop pointless error_is_set(), DumpState member errp
  qemu-option: Clean up fragile use of error_is_set()
  qga: Drop superfluous error_is_set()
  qga: Clean up fragile use of error_is_set()
  qapi: Clean up fragile use of error_is_set()
  tests/qapi-schema: Drop superfluous error_is_set()
  qapi: Drop redundant, unclean error_is_set()
  hmp: Guard against misuse of hmp_handle_error()
  qga: Use return values instead of error_is_set(errp)
  error: Consistently name Error ** objects errp, and not err
  qmp: Consistently name Error ** objects errp, and not err
  qga: Consistently name Error ** objects errp, and not err
  qmp hmp: Consistently name Error * objects err, and not errp
  pci-assign: assigned_initfn(): set monitor error in common error handler
  pci-assign: propagate errors from assign_intx()
  pci-assign: propagate errors from assign_device()
  ...

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

Revert "qapi: Clean up superfluous null check in qapi_dealloc_type_str()"

This reverts commit 25a7017555f1b4aeb543b5d323ff4afb8f9c5437.

Turns out the argument *can* be null: QEMU now segfaults if it
receives an invalid parameter via a qmp command instead of throwing an
error.

For example:
{ "execute": "blockdev-add",
     "arguments": { "options" : { "driver": "invalid-driver" } } }

CC: qemu-stable@nongnu.org
Signed-off-by: Peter Lieven <pl@kamp.de>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

qapi: Document optional arguments' backwards compatibility

Signed-off-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Fam Zheng <famz@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

qmp: use valid JSON in transaction example

Our example should use the correct quotes to match what someone
could actually pass over the wire.

* qmp-commands.hx: Use correct JSON quotes.

Signed-off-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

qmp: Don't use error_is_set() to suppress additional errors

Using error_is_set(errp) that way can sweep programming errors under
the carpet when we get called incorrectly with an error set.

encrypted_bdrv_it() does it, because there's no way to make
bdrv_iterate() break its loop.  Actually safe, because qmp_cont()
clears the error before the loop.  Clean it up anyway: replace
bdrv_iterate() by bdrv_next(), break the loop on error.

Replace both occurrences, for consistency.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

dump: Drop pointless error_is_set(), DumpState member errp

In qmp_dump_guest_memory(), the error must be clear on entry, and we
always bail out after setting it, directly or via dump_init().
Therefore, both error_is_set() are always false.  Drop them.

DumpState member errp is now write-only.  Drop it, too.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

qemu-option: Clean up fragile use of error_is_set()

Using error_is_set(ERRP) to find out whether to bail out due to
previous error is either wrong, fragile, or unnecessarily opaque.
It's wrong when ERRP may be null, because errors go undetected when it
is.  It's fragile when proving ERRP non-null involves a non-local
argument.  Else, it's unnecessarily opaque (see commit 84d18f0).

The error_is_set(state->errp) in qemu_opts_from_qdict_1() is merely
fragile, because the callers never pass state argument with null
state->errp.

Make the code more robust and more obviously correct: test
*state->errp directly.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

qga: Drop superfluous error_is_set()

acquire_privilege(), execute_async() and check_suspend_mode() do
nothing when called with an error set.  Callers shouldn't do that, and
no caller does.  Drop the superfluous tests.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Michael Roth <mdroth@linux.vnet.ibm.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

qga: Clean up fragile use of error_is_set()

Using error_is_set(ERRP) to find out whether a function failed is
either wrong, fragile, or unnecessarily opaque.  It's wrong when ERRP
may be null, because errors go undetected when it is.  It's fragile
when proving ERRP non-null involves a non-local argument.  Else, it's
unnecessarily opaque (see commit 84d18f0).

The error_is_set(errp) in the guest agent command handler functions
are merely fragile, because all chall chains (do_qmp_dispatch() via
the generated marshalling functions) pass a non-null errp argument.

Make the code more robust and more obviously correct: receive the
error in a local variable, then propagate it through the parameter.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Michael Roth <mdroth@linux.vnet.ibm.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

qapi: Clean up fragile use of error_is_set()

Using error_is_set(ERRP) to find out whether a function failed is
either wrong, fragile, or unnecessarily opaque.  It's wrong when ERRP
may be null, because errors go undetected when it is.  It's fragile
when proving ERRP non-null involves a non-local argument.  Else, it's
unnecessarily opaque (see commit 84d18f0).

The error_is_set(errp) in do_qmp_dispatch() is merely fragile, because
the caller never passes a null errp argument.

Make the code more robust and more obviously correct: receive the
error in a local variable, then propagate it through the parameter.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Michael Roth <mdroth@linux.vnet.ibm.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

tests/qapi-schema: Drop superfluous error_is_set()

visit_type_TestStruct() does nothing when called with an error set.
Callers shouldn't do that, and no caller does.  Drop the superfluous
test.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Michael Roth <mdroth@linux.vnet.ibm.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

qapi: Drop redundant, unclean error_is_set()

do_qmp_dispatch()'s test for qmp_dispatch_check_obj() failure examines
both the return value and the error object.  The latter part is
unclean; it works only when do_qmp_dispatch()'s caller passes a
non-null errp argument.  That's the case, but it's not locally
obvious.  Unclean.

Cleanup would be easy enough, but since the unclean code is also
redundant, let's just drop it.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Michael Roth <mdroth@linux.vnet.ibm.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

hmp: Guard against misuse of hmp_handle_error()

Null errp argument makes no sense.  Assert it's not null, to make this
explicit, and guard against misuse.  All current callers pass non-null
errp.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

qga: Use return values instead of error_is_set(errp)

Using error_is_set(errp) to check whether a function call failed is
fragile: it breaks when errp is null.  ga_get_fd_handle() and
guest_file_handle_add() don't return a useful value when they fail,
but that's just stupid.  Fix that, and check them instead.  As far
as I can tell, errp can't be null there, but this is more robust and
more obviously correct.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Michael Roth <mdroth@linux.vnet.ibm.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

error: Consistently name Error ** objects errp, and not err

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

qmp: Consistently name Error ** objects errp, and not err

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

qga: Consistently name Error ** objects errp, and not err

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

qmp hmp: Consistently name Error * objects err, and not errp

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

pci-assign: assigned_initfn(): set monitor error in common error handler

Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

pci-assign: propagate errors from assign_intx()

Among the callers, only assigned_initfn() should set the  monitor's stored
error. Other callers may run in contexts where the monitor's stored error
makes no sense. For example:

assigned_dev_pci_write_config()
  assigned_dev_update_msix()
    assign_intx()

Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

pci-assign: propagate errors from assign_device()

Also, change the return type to "void"; the function is static (with a
sole caller) and the negative errno values are not distinguished from each
other.

Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

pci-assign: propagate errors from assigned_dev_register_regions()

Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

pci-assign: propagate errors from assigned_dev_register_msix_mmio()

The return type is also changed from "int" to "void", because it was used
in a success vs. failure sense only (the caller didn't distinguish error
codes from each other, and even assigned_dev_register_msix_mmio() masked
mmap()'s errno values with a common -EFAULT).

Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

pci-assign: propagate errors from assigned_device_pci_cap_init()

Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

pci-assign: propagate errors from get_real_device()

Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

pci-assign: assignment should fail if we can't read config space

assigned_initfn()
  get_real_device()
    read()

Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

pci-assign: accept Error from pci_add_capability2()

Propagate any errors while adding PCI capabilities to
assigned_device_pci_cap_init(). We'll continue the propagation upwards
when assigned_device_pci_cap_init() becomes a leaf itself (when none of
its callees will report errors internally any longer when detecting and
returning them).

Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

pci: add Error-propagating pci_add_capability2()

... and rebase pci_add_capability() to it.

Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

pci-assign: propagate Error from check_irqchip_in_kernel()

Rename check_irqchip_in_kernel() to verify_irqchip_in_kernel(), so that
the name reflects our expectation better. Rather than returning a bool,
make it do nothing or set an Error.

Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

pci-assign: propagate errors from get_real_id()

get_real_id() has two thin wrappers (and no other callers),
get_real_vendor_id() and get_real_device_id(); it's easiest to convert
them in one fell swoop.

Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

pci-assign: make assign_failed_examine() just format the cause

This allows us to report the entire error with one error_report() call,
easing future error propagation.

Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

pci-assign: accept Error from monitor_handle_fd_param2()

Propagate any errors in monitor fd handling up to get_real_device(), and
report them there. We'll continue the propagation upwards when
get_real_device() becomes a leaf itself (when none of its callees will
report errors internally any longer when detecting and returning an
error).

Signed-off-by: Laszlo Ersek <lersek@redhat.com>
eviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

monitor: add Error-propagating monitor_handle_fd_param2()

and rebase monitor_handle_fd_param() to it. (Note that this will slightly
change the behavior when the qemu_parse_fd() branch is selected and it
fails: we now report (and in case of QMP, set) the error immediately,
rather than allowing the caller to set its own error message (if any)).

Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

cutils: tighten qemu_parse_fd()

qemu_parse_fd() used to handle at least the following strings incorrectly:
o "-2":         simply let through
o "2147483648": returned as LONG_MAX==INT_MAX on ILP32 (with ERANGE
                ignored); implementation-defined behavior on LP64

Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

qapi: treat all negative return of strtosz_suffix() as error

strtosz_suffix() might return negative error, this patch fixes
the error handling.

This patch also changes to handle error in the if statement
rather than handle success specially, this will make this use
of strtosz_suffix consistent with all other uses.

Signed-off-by: Amos Kong <akong@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

qapi: Add a primitive to include other files from a QAPI schema file

The primitive uses JSON syntax, and include paths are relative to the file using the directive:

  { 'include': 'path/to/file.json' }

Signed-off-by: Lluís Vilanova <vilanova@ac.upc.edu>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

qapi: Use an explicit input file

Use an explicit input file on the command-line instead of reading from standard
input.

It also outputs the proper file name when there's an error.

Signed-off-by: Lluís Vilanova <vilanova@ac.upc.edu>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

qapi: [trivial] Do not catch unknown exceptions in "test-qapi.py"

Signed-off-by: Lluís Vilanova <vilanova@ac.upc.edu>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

qapi: [trivial] Break long command lines

Signed-off-by: Lluís Vilanova <vilanova@ac.upc.edu>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Luiz Capitulino <lcapitulino@redhat.com>

Merge remote-tracking branch 'remotes/stefanha/tags/tracing-pull-request' into staging

Tracing pull request

# gpg: Signature made Wed 07 May 2014 18:14:02 BST using RSA key ID 81AB73C8
# gpg: Good signature from "Stefan Hajnoczi <stefanha@redhat.com>"
# gpg:                 aka "Stefan Hajnoczi <stefanha@gmail.com>"

* remotes/stefanha/tags/tracing-pull-request:
  configure: Show trace output file conditionally
  trace: [tracetool] Minimize the amount of per-backend code
  trace: [simple] Bump up log version number
  trace: [tracetool] Change format docs to point to the generated file
  trace: [tracetool] Show list of frontends and backends sorted by name
  trace: [tracetool] Cosmetic changes
  trace: [tracetool] Spacing changes
  trace: [tracetool] Add methods 'Event.copy' and 'Arguments.copy'
  trace: [tracetool] Add method 'Event.api' to build event names

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

Merge remote-tracking branch 'remotes/spice/tags/pull-spice-7' into staging

spice: small fixes

# gpg: Signature made Thu 08 May 2014 09:50:21 BST using RSA key ID D3E87138
# 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>"

* remotes/spice/tags/pull-spice-7:
  spice: fix libvirt snapshots
  spice: fix "info spice"

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

spice: fix libvirt snapshots

Only notify spice-server about migration events in case we got
target host information beforehand.  So we kick the seamless spice
client migration only in case a actual live migration happens, not
when libvirt uses live-migration-to-file for snapshotting.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>

spice: fix "info spice"

In case no listening address was specified, "info spice" reports
"0.0.0.0" as address.  Which is incorrect in case spice is listening
on ipv6.  Replace it by a wildcard "*" to indicate it is not limited
to a specific address.

Note:  Being more specific is not possible without extending the
spice-server api.  The socket is handled by spice-server not
qemu, so qemu can't easily figure the actual socket address.

Reported-by: David Jaša <djasa@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>

Merge remote-tracking branch 'remotes/mjt/tags/trivial-patches-2014-05-07' into staging

trivial patches for 2014-05-07

# gpg: Signature made Wed 07 May 2014 18:01:15 BST 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>"
# gpg: WARNING: This key is not certified with a trusted signature!
# gpg:          There is no indication that the signature belongs to the owner.
# Primary key fingerprint: 6EE1 95D1 886E 8FFB 810D  4324 457C E0A0 8044 65C5
#      Subkey fingerprint: 6F67 E18E 7C91 C5B1 5514  66A7 BEE5 9D74 A4C3 D7DB

* remotes/mjt/tags/trivial-patches-2014-05-07: (21 commits)
  libcacard: remove unnecessary EOL from debug prints
  docs/memory.txt: Fix document on MMIO operations
  readline: Sort completions before printing them.
  readline: use g_strndup instead of open-coding it
  qmp: report path ambiguity error
  libcacard: replace pstrcpy() with memcpy()
  glib: move g_poll() replacement into glib-compat.h
  do not call g_thread_init() for glib >= 2.31
  hw/9pfs: Add include file for exported symbol
  xen: remove unused global, xen_xcg
  hw: Add missing 'static' attributes
  qemu-timer: Add missing 'static' attribute
  ui: Add missing 'static' attribute
  monitor: Add missing 'static' attribute
  hw/s390x: Add missing 'static' attribute
  hw/mips: Add missing 'static' and 'const' attributes
  hw/9pfs: Add missing 'static' attributes
  arch_init: Be sure of only one exit entry with DPRINTF() for ram_load()
  tests/tcg: Fix compilation of test_path
  qga: Fix typo (plural) in comment
  ...

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

Merge remote-tracking branch 'remotes/sstabellini/xen-140507-2' into staging

* remotes/sstabellini/xen-140507-2:
  xen_disk: add discard support
  pass an inclusive address range to xc_domain_pin_memory_cacheattr
  xen: factor out common functions
  xen: move Xen HVM files under hw/i386/xen
  xen: move Xen PV machine files to hw/xenpv
  qemu-xen: free all the pirqs for msi/msix when driver unload
  exec: Limit translation limiting in address_space_translate to xen

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

configure: Show trace output file conditionally

It is only used with the simple trace backend.

Signed-off-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>

trace: [tracetool] Minimize the amount of per-backend code

Backends now only contain the essential backend-specific code, and most of the work is moved to frontend code.

Signed-off-by: Lluís Vilanova <vilanova@ac.upc.edu>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>

trace: [simple] Bump up log version number

The following tracetool cleanup changes the event numbering policy.

Signed-off-by: Lluís Vilanova <vilanova@ac.upc.edu>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>

trace: [tracetool] Change format docs to point to the generated file

Signed-off-by: Lluís Vilanova <vilanova@ac.upc.edu>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>

trace: [tracetool] Show list of frontends and backends sorted by name

Signed-off-by: Lluís Vilanova <vilanova@ac.upc.edu>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>

trace: [tracetool] Cosmetic changes

Signed-off-by: Lluís Vilanova <vilanova@ac.upc.edu>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>

trace: [tracetool] Spacing changes

Signed-off-by: Lluís Vilanova <vilanova@ac.upc.edu>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>

trace: [tracetool] Add methods 'Event.copy' and 'Arguments.copy'

Signed-off-by: Lluís Vilanova <vilanova@ac.upc.edu>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>

trace: [tracetool] Add method 'Event.api' to build event names

Makes it easier to ensure proper naming across the different frontends and backends.

Signed-off-by: Lluís Vilanova <vilanova@ac.upc.edu>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>

libcacard: remove unnecessary EOL from debug prints

Signed-off-by: Alon Levy <alevy@redhat.com>
Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>

docs/memory.txt: Fix document on MMIO operations

.impl.valid should be .impl.unaligned and the description needs some
fixes.

.old_portio is removed since commit b40acf99b (ioport: Switch
dispatching to memory core layer).

Signed-off-by: Fam Zheng <famz@redhat.com>
Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>

readline: Sort completions before printing them.

Signed-off-by: Hani Benhabiles <hani@linux.com>
Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>

readline: use g_strndup instead of open-coding it

Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>
Reviewed-by: Andreas Färber <afaerber@suse.de>

qmp: report path ambiguity error

Without this, ambiguous path is reported to the user as
"not found", which is confusing at least.

Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>

libcacard: replace pstrcpy() with memcpy()

Commit 2e679780ae86c6ca8 replaced strncpy() with pstrcpy()
in one place in libcacard.  This is a qemu-specific function,
while libcacard is a stand-alone library (or tries to be).
But since we know the exact length of the string to copy,
and know that it definitely will fit in the destination
buffer, use memcpy() instead, and null-terminate the string
after that.

An alternative is to use g_strlcpy() or strncpy(), but memcpy()
is more than adequate in this place.

Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>
Cc: qemu-trivial@nongnu.org
Cc: Alon Levy <alevy@redhat.com>

glib: move g_poll() replacement into glib-compat.h

We have a dedicated header file for wrappers to smooth over glib version
differences.  Move the g_poll() definition into glib-compat.h for
consistency.

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>
Cc: qemu-trivial@nongnu.org

do not call g_thread_init() for glib >= 2.31

glib >= 2.31 always enables thread support and g_thread_supported()
is #defined to 1, there's no need to call g_thread_init() anymore,
and it definitely does not need to report error which never happens.
Keep code for old < 2.31 glibc anyway for now, just #ifdef it
differently.

Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Cc: qemu-trivial@nongnu.org

hw/9pfs: Add include file for exported symbol

Signed-off-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>

xen: remove unused global, xen_xcg

Reviewed-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Jim Meyering <meyering@redhat.com>
Acked-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>

hw: Add missing 'static' attributes

This fixes warnings from the static code analysis (smatch).

Signed-off-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>

qemu-timer: Add missing 'static' attribute

This fixes a warning from the static code analysis (smatch).

Signed-off-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>

ui: Add missing 'static' attribute

There was already a forward declaration using 'static',
but the attribute was missing in the implementation.

This fixes a warning from the static code analysis (smatch).

Signed-off-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>

monitor: Add missing 'static' attribute

This fixes a warning from the static code analysis (smatch).

Signed-off-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>

hw/s390x: Add missing 'static' attribute

Signed-off-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>

hw/mips: Add missing 'static' and 'const' attributes

This fixes a warning from the static code analysis (smatch).

Signed-off-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>
Acked-by: Aurelien Jarno <aurelien@aurel32.net>

hw/9pfs: Add missing 'static' attributes

Signed-off-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>

arch_init: Be sure of only one exit entry with DPRINTF() for ram_load()

When DPRINTF() has effect, the original author wants to print all
ram_load() calling results. So need use 'goto' instead of 'return'
within ram_load(), just like other areas have done.

Signed-off-by: Chen Gang <gang.chen.5i5j@gmail.com>
Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>

tests/tcg: Fix compilation of test_path

The test_path binary is (unlike the other test binaries in tests/tcg)
actually intended to be compiled with the same compiler used to build
the main QEMU executables. It actually #includes a number of the
QEMU source files in an attempt to unit-test the util/path.c functions,
and so if it is not compiled with the same compiler used by configure
to set CONFIG_ settings then it is liable to fail to build.
Fix the makefile to build it with the default C compiler rules, not
CC_I386, and fix the test itself not to include a lot of unnecessary
trace related source files which cause the build to fail if the trace
backend is anything other than 'simple'.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>

qga: Fix typo (plural) in comment

Signed-off-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>

configure: remove bashism

Commit e26110cfc67d48 added a check for shacmd to create a hash
for modules.  This check in configure is using bash construct &>
to redirect both stdout and stderr, which does fun things on some
shells.  Get rid of it, use standard redirection instead.

Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>
Reviewed-by: Fam Zheng <famz@redhat.com>

xen_disk: add discard support

Implement discard support for xen_disk. It makes use of the existing
discard code in qemu.

The discard support is enabled unconditionally. The tool stack may
provide a property "discard-enable" in the backend node to optionally
disable discard support.  This is helpful in case the backing file was
intentionally created non-sparse to avoid fragmentation.

Signed-off-by: Olaf Hering <olaf@aepfle.de>
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>

pass an inclusive address range to xc_domain_pin_memory_cacheattr

xc_domain_pin_memory_cacheattr expects an inclusive address range:
adjust the parameters.

Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>

xen: factor out common functions

So common functions used by both HVM and PV are factored out from
xen-all.c to xen-common.c.

Finally rename xen-all.c to xen-hvm.c, as those functions are only
useful to HVM guest.

Create *-stub files and modify Makefile.target to reflect the changes.

Signed-off-by: Wei Liu <wei.liu2@citrix.com>
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>