More centralization of handling of segment bases.
Also fixes the note about 16-bit wrap around not fully handled.
Signed-off-by: NRichard Henderson <rth@twiddle.net>
Message-Id: <1450379966-28198-7-git-send-email-rth@twiddle.net>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Having segs[].base as a register significantly improves code
generation for real and protected modes, particularly for TBs
that have multiple memory references where the segment base
can be held in a hard register through the TB.
Reviewed-by: NPeter Maydell <peter.maydell@linaro.org>
Signed-off-by: NRichard Henderson <rth@twiddle.net>
Message-Id: <1450379966-28198-6-git-send-email-rth@twiddle.net>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

I.e. gen_push_v, gen_pop_T0, gen_stack_A0.
More centralization of handling of segment bases.
Signed-off-by: NRichard Henderson <rth@twiddle.net>
Message-Id: <1450379966-28198-5-git-send-email-rth@twiddle.net>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Centralize handling of segment bases.
Signed-off-by: NRichard Henderson <rth@twiddle.net>
Message-Id: <1450379966-28198-4-git-send-email-rth@twiddle.net>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Centralize computation of a MO_SIZE for the stack pointer.
Signed-off-by: NRichard Henderson <rth@twiddle.net>
Message-Id: <1450379966-28198-3-git-send-email-rth@twiddle.net>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Add forgotten zero-extension in the TARGET_X86_64, !CODE64, ss32 case;
use this new function to implement gen_string_movl_A0_EDI,
gen_string_movl_A0_ESI, gen_add_A0_ds_seg.
Signed-off-by: NRichard Henderson <rth@twiddle.net>
Message-Id: <1450379966-28198-2-git-send-email-rth@twiddle.net>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

In previous commit:

  commit f2001a7e
  Author: Daniel P. Berrange <berrange@redhat.com>
  Date:   Tue Jan 19 11:14:30 2016 +0000

    char: don't assume telnet initialization will not block

The code which writes the telnet initialization sequence moved
to an event loop callback. If the TCP chardev is opened as a
server in blocking mode (ie -serial telnet:0.0.0.0:3000,server,wait)
this results in a state where the TCP chardev is connected, but not
yet ready to send/recv data when virtual hardware is created.

When the virtual hardware initialization registers its chardev
callbacks, it triggers tcp_chr_update_read_handler, which will
add I/O watches to the connection.

When the telnet initialization finally runs, it will then call
tcp_chr_connect to finish the connection setup. This will in
turn add I/O watches to the connection too.

There are now two sets of I/O watches registered on the same
connection. This ultimately causes data loss on the connection,
for example, when typing into the telnet console only every
second byte is echoed back to the client.

The same flaw can affect channels running with TLS encryption
too, since they also have delayed connection setup completion.

The fix is to update tcp_chr_update_read_handler so that it
avoids registering watches if the connection is not fully
setup yet.
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>
Message-Id: <1454939707-10869-1-git-send-email-berrange@redhat.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Signed-off-by: NAndrew Jones <drjones@redhat.com>
Message-Id: <1454355464-14999-1-git-send-email-drjones@redhat.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

cpu_to_be64w can't be used to make unaligned stores, but stq_be_p can.
Also, the st?_be_p takes a void* so it is more clearly suited to the
case where you're writing into a byte buffer.

Use the st?_be_p family of functions everywhere in nbd/server.c.
Signed-off-by: NJohn Snow <jsnow@redhat.com>
[Changed to use st?_be_p everywhere. - Paolo]
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

On kernels build without CONFIG_TRACING kvm_stat will bail out even
when traces are not used. This is not very helpful, especially if the
user can't install a new kernel. Instead, we should warn the user and
fall back to debugfs statistics.

These changes check if trace statistics were selected without kernel
support, warn with a small timeout, set the debugfs statistics option
to True and the tracefs one to False.

Fixes: 7aa4ee5a ('scripts/kvm/kvm_stat: Improve debugfs access checking')
Signed-off-by: NJanosch Frank <frankja@linux.vnet.ibm.com>
Message-Id: <1454485291-43849-2-git-send-email-frankja@linux.vnet.ibm.com>
[Exit if -t is passed explicitly. - Paolo]
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Capitalise the first letter of sentences (and reword for grammar) the
options section of qemu-nbd.texi.
Signed-off-by: NSitsofe Wheeler <sitsofe@yahoo.com>
Message-Id: <1451979212-25479-4-git-send-email-sitsofe@yahoo.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

- Change some spacing.
- Add disconnect usage to synopsis.
- Highlight the command and its options in the synopsis.
- Fix up the grammar in the description.
- Move filename variable description out of the option table.
- Add a description of the dev variable.
- Remove duplicate entry for --format.
- Reword --discard documentation.
- Add --detect-zeroes documentation.
- Add reference to qemu man page to see also section.
Signed-off-by: NSitsofe Wheeler <sitsofe@yahoo.com>
Message-Id: <1451979212-25479-3-git-send-email-sitsofe@yahoo.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Indented lines in the texi meant the perlpod produced interpreted the
paragraph as being verbatim (thus formatting codes were not
interpreted). Fix this by un-indenting problem lines.
Signed-off-by: NSitsofe Wheeler <sitsofe@yahoo.com>
Message-Id: <1451979212-25479-2-git-send-email-sitsofe@yahoo.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

This adds the SAS1068 device, a SAS disk controller used in VMware that
is oldish but widely supported and has decent performance.  Unlike
megasas, it presents itself as a SAS controller and not as a RAID
controller.  The device corresponds to the mptsas kernel driver in
Linux.

A few small things in the device setup are based on Don Slutz's old
patch, but the device emulation was written from scratch based on Don's
SeaBIOS patch and on the FreeBSD and Linux drivers.  It is 2400 lines
shorter than Don's patch (and roughly the same size as MegaSAS---also
because it doesn't support the similar SPI controller), implements SCSI
task management functions (with asynchronous cancellation), supports
big-endian hosts, has complete support for migration and follows the
QEMU coding standards much more closely.

To write the driver, I first split Don's patch in two parts, with
the configuration bits in one file and the rest in a separate file.
I first left mptconfig.c in place and rewrote the rest, then deleted
mptconfig.c as well.  The configuration pages are still based mostly on
VirtualBox's, though not exactly the same.  However, the implementation
is completely different.  The contents of the pages themselves should
not be copyrightable.
Signed-off-by: NDon Slutz <Don@CloudSwitch.com>
Message-Id: <1347382813-5662-1-git-send-email-Don@CloudSwitch.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

This lets a SAS adapter expose them through its own configuration
mechanism.
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

SAS adapters need to access them in order to publish the SAS addresses
of the end devices connected to them.
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

The __atomic primitives have been available since GCC 4.7 and provide
a richer interface for describing memory ordering requirements. As a
bonus by using the primitives instead of hand-rolled functions we can
use tools such as the ThreadSanitizer which need the use of well
defined APIs for its analysis.

If we have __ATOMIC defines we exclusively use the __atomic primitives
for all our atomic access. Otherwise we fall back to the mixture of
__sync and hand-rolled barrier cases.
Signed-off-by: NAlex Bennée <alex.bennee@linaro.org>
Message-Id: <1453976119-24372-4-git-send-email-alex.bennee@linaro.org>
[Use __ATOMIC_SEQ_CST for atomic_mb_read/atomic_mb_set on !POWER. - Paolo]
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Developers on 64-bit machines will often try to perform a
32-bit build of QEMU by running

  ./configure --extra-cflags="-m32"

Unfortunately if PKG_CONFIG_LIBDIR is not set to point to
the location of the 32-bit pkg-config files, then configure
will silently pick up the 64-bit pkg-config files and still
succeed.

This causes a problem for glib because it means QEMU will
be pulling in /usr/lib64/glib-2.0/include/glibconfig.h
instead of /usr/lib/glib-2.0/include/glibconfig.h

This causes problems because the 'gsize' type (defined as
'unsigned long') will no longer be fully compatible with
the 'size_t' type (defined as 'unsigned int'). Although
both are the same size, the compiler refuses to allow
casts from 'unsigned long *' to 'unsigned int *' as they
are different pointer types. This results in non-obvious
compiler errors when building QEMU eg

qga/commands-posix.c: In function ‘qmp_guest_set_user_password’:
qga/commands-posix.c:1912:55: error: passing argument 2 of ‘g_base64_decode’ from incompatible pointer type [-Werror=incompatible-pointer-types]
     rawpasswddata = (char *)g_base64_decode(password, &rawpasswdlen);
                                                            ^
In file included from /usr/include/glib-2.0/glib.h:35:0,
                 from qga/commands-posix.c:14:
/usr/include/glib-2.0/glib/gbase64.h:52:9: note: expected ‘gsize * {aka long unsigned int *}’ but argument is of type ‘size_t * {aka unsigned int *}’
 guchar *g_base64_decode         (const gchar  *text,
         ^
cc1: all warnings being treated as errors

To detect this problem, add a check to configure that
verifies that GLIB_SIZEOF_SIZE_T matches sizeof(size_t).
If this fails print a warning suggesting that the dev
probably needs to set PKG_CONFIG_LIBDIR.

On Fedora x86_64 it passes with any of:

 # ./configure
 # PKG_CONFIG_LIBDIR=/usr/lib/pkgconfig ./configure --extra-cflags="-m32"
 # PKG_CONFIG_LIBDIR=/usr/lib64/pkgconfig ./configure --extra-cflags="-m64"

And fails with a mis-match

 # PKG_CONFIG_LIBDIR=/usr/lib64/pkgconfig ./configure --extra-cflags="-m32"
 # PKG_CONFIG_LIBDIR=/usr/lib/pkgconfig ./configure --extra-cflags="-m64"

ERROR: sizeof(size_t) doesn't match GLIB_SIZEOF_SIZE_T.
       You probably need to set PKG_CONFIG_LIBDIR
       to point to the right pkg-config files for your
       build target
Signed-off-by: NDaniel P. Berrange <berrange@redhat.com>
Message-Id: <1453885245-15562-1-git-send-email-berrange@redhat.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

If a process opens the slave pts device, writes data to it, then
immediately closes it, the data doesn't reliably get delivered to the
emulated serial port. This seems to be because a read of the master
pty device returns EIO on Linux if no process has the pts device open,
even when data is waiting "in the pipe".

A fix seems to be for QEMU to keep the pts file descriptor open until
the pty is closed, as per the below patch.
Signed-off-by: NAshley Jonathan <jonathan.ashley@altran.com>
Message-Id: <AC19797808C8D548ABDE0CA4A97AA30A30DEB409@XMB-DCFR-37.europe.corp.altran.com>
Reviewed-by: NMichael Tokarev <mjt@tls.msk.ru>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Although accesses to ram_list.dirty_memory[] use atomics so multiple
threads can safely dirty the bitmap, the data structure is not fully
thread-safe yet.

This patch handles the RAM hotplug case where ram_list.dirty_memory[] is
grown.  ram_list.dirty_memory[] is change from a regular bitmap to an
RCU array of pointers to fixed-size bitmap blocks.  Threads can continue
accessing bitmap blocks while the array is being extended.  See the
comments in the code for an in-depth explanation of struct
DirtyMemoryBlocks.

I have tested that live migration with virtio-blk dataplane works.
Signed-off-by: NStefan Hajnoczi <stefanha@redhat.com>
Message-Id: <1453728801-5398-2-git-send-email-stefanha@redhat.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

This condition is true in the common case, so we can cut out the body of
the function.  In addition, this makes it easier for the compiler to do
at least partial inlining, even if it decides that fully inlining the
function is unreasonable.
Reviewed-by: NStefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

QAPI patches for 2016-02-09

# gpg: Signature made Tue 09 Feb 2016 10:55:51 GMT using RSA key ID EB918653
# gpg: Good signature from "Markus Armbruster <armbru@redhat.com>"
# gpg:                 aka "Markus Armbruster <armbru@pond.sub.org>"

* remotes/armbru/tags/pull-qapi-2016-02-09: (31 commits)
  qapi: Add missing JSON files in build dependencies
  qapi: Fix compilation failure on MIPS and SPARC
  qmp: Don't abuse stack to track qmp-output root
  qmp: Fix reference-counting of qnull on empty output visit
  qapi: Drop unused error argument for list and implicit struct
  qapi: Tighten qmp_input_end_list()
  qapi: Drop unused 'kind' for struct/enum visit
  qapi: Swap 'name' in visit_* callbacks to match public API
  qom: Swap 'name' next to visitor in ObjectPropertyAccessor
  qapi: Swap visit_* arguments for consistent 'name' placement
  qom: Use typedef for Visitor
  qapi: Don't cast Enum* to int*
  qapi: Consolidate visitor small integer callbacks
  qapi: Make all visitors supply uint64 callbacks
  qapi: Prefer type_int64 over type_int in visitors
  qapi-visit: Kill unused visit_end_union()
  qapi: Track all failures between visit_start/stop
  qapi: Improve generated event use of qapi visitor
  balloon: Improve use of qapi visitor
  vl: Ensure qapi visitor properly ends struct visit
  ...
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

Queued TCG patches

# gpg: Signature made Mon 08 Feb 2016 23:57:30 GMT using RSA key ID 4DD0279B
# gpg: Good signature from "Richard Henderson <rth7680@gmail.com>"
# gpg:                 aka "Richard Henderson <rth@redhat.com>"
# gpg:                 aka "Richard Henderson <rth@twiddle.net>"

* remotes/rth/tags/pull-tcg-20160209:
  tcg: Introduce temp_load
  tcg: Change temp_save argument to TCGTemp
  tcg: Change temp_sync argument to TCGTemp
  tcg: Change temp_dead argument to TCGTemp
  tcg: Change reg_to_temp to TCGTemp pointer
  tcg: Remove tcg_get_arg_str_i32/64
  tcg: More use of TCGReg where appropriate
  tcg: Work around clang bug wrt enum ranges
  tcg: Tidy temporary allocation
  tcg: Change ts->mem_reg to ts->mem_base
  tcg: Change tcg_global_mem_new_* to take a TCGv_ptr
  tcg: Remove lingering references to gen_opc_buf
  tcg: Respect highwater in tcg_out_tb_finalize
Signed-off-by: NPeter Maydell <peter.maydell@linaro.org>

Unify all of the places that realize a temporary into a register.
Reviewed-by: NAurelien Jarno <aurelien@aurel32.net>
Signed-off-by: NRichard Henderson <rth@twiddle.net>

Reviewed-by: NAurelien Jarno <aurelien@aurel32.net>
Signed-off-by: NRichard Henderson <rth@twiddle.net>

Reviewed-by: NAurelien Jarno <aurelien@aurel32.net>
Signed-off-by: NRichard Henderson <rth@twiddle.net>

Reviewed-by: NAurelien Jarno <aurelien@aurel32.net>
Signed-off-by: NRichard Henderson <rth@twiddle.net>

Reviewed-by: NAurelien Jarno <aurelien@aurel32.net>
Signed-off-by: NRichard Henderson <rth@twiddle.net>

Reviewed-by: NAurelien Jarno <aurelien@aurel32.net>
Signed-off-by: NRichard Henderson <rth@twiddle.net>

Reviewed-by: NAurelien Jarno <aurelien@aurel32.net>
Signed-off-by: NRichard Henderson <rth@twiddle.net>

A subsequent patch patch will change the type of REG from int
to enum TCGReg, which provokes the following bug in clang:

  https://llvm.org/bugs/show_bug.cgi?id=16154Signed-off-by: NRichard Henderson <rth@twiddle.net>

In particular, make sure the memory is memset before use.
Continues the increased use of TCGTemp pointers instead of
integer indices where appropriate.
Reviewed-by: NAurelien Jarno <aurelien@aurel32.net>
Signed-off-by: NRichard Henderson <rth@twiddle.net>

Chain the temporaries together via pointers intstead of indices.
The mem_reg value is now mem_base->reg.  This will be important later.

This does require that the frame pointer have a global temporary
allocated for it.  This is simple bar the existing reserved_regs check.
Reviewed-by: NAurelien Jarno <aurelien@aurel32.net>
Signed-off-by: NRichard Henderson <rth@twiddle.net>

Thus, use cpu_env as the parameter, not TCG_AREG0 directly.
Update all uses in the translators.
Reviewed-by: NAurelien Jarno <aurelien@aurel32.net>
Signed-off-by: NRichard Henderson <rth@twiddle.net>

Three in comments and one in code in the stub tcg_liveness_analysis.
Signed-off-by: NRichard Henderson <rth@twiddle.net>

Undo the workaround at b17a6d33.

If there are lots of memory operations in a TB, the slow path code
can exceed the highwater reservation.  Add a check within the loop.
Tested-by: NAurelien Jarno <aurelien@aurel32.net>
Reviewed-by: NAurelien Jarno <aurelien@aurel32.net>
Signed-off-by: NRichard Henderson <rth@twiddle.net>

Forgotten in commit 1dde0f48 (trace.json) and commit fafa4d50
(rocker.json).
Signed-off-by: NLluís Vilanova <vilanova@ac.upc.edu>
Message-Id: <145461055662.15201.2702170180078718114.stgit@localhost>
Reviewed-by: NEric Blake <eblake@redhat.com>
[Commit message tweaked]
Signed-off-by: NMarkus Armbruster <armbru@redhat.com>

Commit 86f4b687 broke compilation on MIPS and SPARC, which have a
preprocessor pollution of '#define mips 1' and '#define sparc 1',
respectively.  Treat it the same way as we do for the pollution with
'unix', so that QMP remains backwards compatible and only the C code
needs to use the alternative 'q_mips', 'q_sparc' spelling.

CC: James Hogan <james.hogan@imgtec.com>
CC: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: NEric Blake <eblake@redhat.com>
Tested-by: NJames Hogan <james.hogan@imgtec.com>
Signed-off-by: NMarkus Armbruster <armbru@redhat.com>

The previous commit documented an inconsistency in how we are
using the stack of qmp-output-visitor.  Normally, pushing a
single top-level object puts the object on the stack twice:
once as the root, and once as the current container being
appended to; but popping that struct only pops once.  However,
qmp_ouput_add() was trying to either set up the added object
as the new root (works if you parse two top-level scalars in a
row: the second replaces the first as the root) or as a member
of the current container (works as long as you have an open
container on the stack; but if you have popped the first
top-level container, it then resolves to the root and still
tries to add into that existing container).

Fix the stupidity by not tracking two separate things in the
stack.  Drop the now-useless qmp_output_first() and
qmp_output_last() while at it.

Saved for a later patch: we still are rather sloppy in that
qmp_output_get_object() can be called in the middle of a parse,
rather than requiring that a visit is complete.
Signed-off-by: NEric Blake <eblake@redhat.com>
Message-Id: <1454075341-13658-26-git-send-email-eblake@redhat.com>
Signed-off-by: NMarkus Armbruster <armbru@redhat.com>

Commit 6c2f9a15 ensured that we would not return NULL when the
caller used an output visitor but had nothing to visit. But
in doing so, it added a FIXME about a reference count leak
that could abort qemu in the (unlikely) case of SIZE_MAX such
visits (more plausible on 32-bit).  (Although that commit
suggested we might fix it in time for 2.5, we ran out of time;
fortunately, it is unlikely enough to bite that it was not
worth worrying about during the 2.5 release.)

This fixes things by documenting the internal contracts, and
explaining why the internal function can return NULL and only
the public facing interface needs to worry about qnull(),
thus avoiding over-referencing the qnull_ global object.

It does not, however, fix the stupidity of the stack mixing
up two separate pieces of information; add a FIXME to explain
that issue, which will be fixed shortly in a future patch.
Signed-off-by: NEric Blake <eblake@redhat.com>
Cc: qemu-stable@nongnu.org
Message-Id: <1454075341-13658-25-git-send-email-eblake@redhat.com>
Signed-off-by: NMarkus Armbruster <armbru@redhat.com>