Provide different functions for converting from an LE vs a BE
framebuffer. We cannot rely on the simple cases always being
shared surfaces since cirrus will need to always shadow for
cursor emulation, so we need the full set of functions to
be able to later handle runtime switching.
All the macros used to generate different versions of vga_template.h
are now unnecessary, take them all out and remove the _32 suffix from
most functions.
vga: Start cutting out non-32bpp conversion support
Nowadays, we either share a surface with the host, or we create
a 32bpp ARGB console surface.
So we only need to draw/convert to 32bpp, enabling us to remove
all but one instance of vga_template.h inclusion (to be further
cleaned up), rgb_to_pixel_* etc...
Peter Maydell [Tue, 30 Sep 2014 10:02:06 +0000 (11:02 +0100)]
Merge remote-tracking branch 'remotes/pmaydell/tags/pull-target-arm-20140929' into staging
target-arm:
* more EL2/EL3 preparation work
* don't handle c15_cpar changes via tb_flush()
* fix some unused function warnings in ARM devices
* build the GDB XML for 32 bit CPUs into qemu-*-aarch64
* implement guest breakpoint support
# gpg: Signature made Mon 29 Sep 2014 19:25:37 BST using RSA key ID 14360CDE
# gpg: Good signature from "Peter Maydell <[email protected]>"
* remotes/pmaydell/tags/pull-target-arm-20140929:
target-arm: Add support for VIRQ and VFIQ
target-arm: Add IRQ and FIQ routing to EL2 and 3
target-arm: A64: Emulate the SMC insn
target-arm: Add a Hypervisor Trap exception type
target-arm: A64: Emulate the HVC insn
target-arm: A64: Correct updates to FAR and ESR on exceptions
target-arm: Don't take interrupts targeting lower ELs
target-arm: Break out exception masking to a separate func
target-arm: A64: Refactor aarch64_cpu_do_interrupt
target-arm: Add SCR_EL3
target-arm: Add HCR_EL2
target-arm: Don't handle c15_cpar changes via tb_flush()
hw/input/tsc210x.c: Delete unused array tsc2101_rates
hw/display/pxa2xx_lcd.c: Remove unused function pxa2xx_dma_rdst_set
hw/intc/imx_avic.c: Remove unused function imx_avic_set_prio()
hw/display/blizzard.c: Delete unused function blizzard_rgb2yuv
configure: Build GDB XML for 32 bit ARM CPUs into qemu aarch64 binaries
target-arm: Implement handling of breakpoint firing
target-arm: Implement setting guest breakpoints
Peter Maydell [Mon, 29 Sep 2014 17:48:48 +0000 (18:48 +0100)]
target-arm: Don't handle c15_cpar changes via tb_flush()
At the moment we try to handle c15_cpar with the strategy of:
* emit generated code which makes assumptions about its value
* when the register value changes call tb_flush() to throw
away the now-invalid generated code
This works because XScale CPUs are always uniprocessor, but
it's confusing because it suggests that the same approach can
be taken for other registers. It also means we do a tb_flush()
on CPU reset, which makes multithreaded linux-user binaries
even more likely to fail than would otherwise be the case.
Replace it with a combination of TB flags for the access
checks done on cp0/cp1 for the XScale and iwMMXt instructions,
plus a runtime check for cp2..cp13 coprocessor accesses.
Peter Maydell [Mon, 29 Sep 2014 17:48:47 +0000 (18:48 +0100)]
configure: Build GDB XML for 32 bit ARM CPUs into qemu aarch64 binaries
The qemu-aarch64 and qemu-system-aarch64 binaries include support
for all the 32 bit ARM CPUs as well as the 64 bit ones. This means
we need to build in the GDB XML files for the 32 bit CPUs too.
Otherwise gdb will complain:
warning: while parsing target description (at line 1): Could not load XML document "arm-core.xml"
when you try to connect to our gdbserver to debug a 32 bit CPU
running in a qemu-aarch64 or qemu-system-aarch64 binary.
Peter Maydell [Mon, 29 Sep 2014 17:48:46 +0000 (18:48 +0100)]
target-arm: Implement handling of breakpoint firing
Implement handling of breakpoint event firing to correctly
inject the debug exception into the guest.
Since the breakpoint and watchpoint control register format is
very similar we adjust wp_matches() to also handle breakpoints
as well rather than using a separate function.
Peter Maydell [Mon, 29 Sep 2014 17:48:46 +0000 (18:48 +0100)]
target-arm: Implement setting guest breakpoints
This patch adds support for setting guest breakpoints
based on values the guest writes to the DBGBVR and DBGBCR
registers. (It doesn't include the code to handle when
these breakpoints fire, so has no guest-visible effect.)
Stefan Hajnoczi [Mon, 29 Sep 2014 15:40:12 +0000 (16:40 +0100)]
libqos: use microseconds instead of iterations for virtio timeout
Some hosts are slow or overloaded so test execution takes a long time.
Test cases use timeouts to protect against an infinite loop stalling the
test forever (especially important in automated test setups).
Commit 6cd14054b67774cc58a51fca6660cfa1d3c08059 ("libqos virtio:
Increase ISR timeout") increased the clock_step() value in an attempt to
lengthen the virtio interrupt wait timeout, but timeout failures are
still occuring on the Travis automated testing platform.
This is because clock_step() only affects the guest's virtual time.
Virtio requests can be bottlenecked on host disk I/O latency - which
cannot be improved by stepping the clock, so the fix was ineffective.
This patch changes the qvirtio_wait_queue_isr() and
qvirtio_wait_config_isr() timeout mechanism from loop iterations to
microseconds. This way the test case can specify an absolute 30 second
timeout. Number of loop iterations is not a reliable timeout mechanism
since the speed depends on many factors including host performance.
Tests should no longer timeout on overloaded Travis instances.
Stefan Hajnoczi [Mon, 29 Sep 2014 15:40:11 +0000 (16:40 +0100)]
libqos: improve event_index test with timeout
The virtio event_index feature lets the device driver tell the device
how many requests to process before raising the next interrupt.
virtio-blk-test.c tries to verify that the device does not raise an
interrupt unnecessarily.
Unfortunately the test has a race condition. It spins checking for an
interrupt up to 100 times and then assumes the request has finished. On
a slow host the I/O request could still be in flight and the test would
fail.
This patch waits for the request to complete, or until a 30-second
timeout is reached. If an interrupt is raised while waiting the test
fails since the device was not supposed to raise interrupts.
Peter Maydell [Mon, 29 Sep 2014 11:26:14 +0000 (12:26 +0100)]
Merge remote-tracking branch 'remotes/kevin/tags/for-upstream' into staging
Block patches
# gpg: Signature made Fri 26 Sep 2014 19:57:52 BST using RSA key ID C88F2FD6
# gpg: Good signature from "Kevin Wolf <[email protected]>"
* remotes/kevin/tags/for-upstream:
qemu-iotests: Fail test if explicit test case number is unknown
block: Validate node-name
vpc: fix beX_to_cpu() and cpu_to_beX() confusion
docs: add blkdebug block driver documentation
block: Catch simultaneous usage of options and their aliases
block: Specify -drive legacy option aliases in array
block: Improve message for device name clashing with node name
qemu-nbd: Destroy the BlockDriverState properly
block: Keep DriveInfo alive until BlockDriverState dies
blockdev: Disentangle BlockDriverState and DriveInfo creation
blkdebug: show an error for invalid event names
Simply switch function pointers when entering/leaving vga mode.
Allows to remove wrapper functions which do nothing but dispatch
calls depending on the current qxl mode.
Add a function to allow display emulations to switch the hwops
function pointers. This is useful for devices which have two
completely different operation modes. Typical case is the vga
compatibility mode vs. native mode in qxl and the upcoming
virtio-vga device.
Peter Maydell [Fri, 26 Sep 2014 17:44:25 +0000 (18:44 +0100)]
Merge remote-tracking branch 'remotes/mjt/tags/trivial-patches-2014-09-26' into staging
trivial patches for 2014-09-26
# gpg: Signature made Fri 26 Sep 2014 18:33:53 BST using RSA key ID A4C3D7DB
# gpg: Good signature from "Michael Tokarev <[email protected]>"
# gpg: aka "Michael Tokarev <[email protected]>"
# gpg: aka "Michael Tokarev <[email protected]>"
# 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-2014-09-26:
os-posix: report error message when lock file failed
os-posix: remove confused errno
os-posix: change tab to space avoid violating coding style
qapi: Update docs given recent event, spacing fixes
qapi: Ignore files created during make check
qapi: Consistent whitespace in tests/Makefile
vmxcap: Update according to SDM of September 2014
.travis.yml: remove "make check" from main matrix
.travis.yml: pre-seed sub-modules for speed
.travis.yml: make the make slightly more parallel
.travis.yml: add more linux-user to the build matrix
tests: avoid running duplicate qom-tests
The QERR_ macros are leftovers from the days of "rich" error objects.
They're used with error_set() and qerror_report(), and expand into the
first *two* arguments. This trickiness has become pointless. Clean
up.
os-posix: report error message when lock file failed
It will cause that create vm failed When manager
tool is killed forcibly (kill -9 libvirtd_pid),
the file not was unlink, and unlock. It's better
that report the error message for users.
If we get inside the 'else if (status == 1)' conditional,
then we know that read() succeeded, and therefore errno is
unspecified. Printing strerror(errno) on a random value
is not helpful.
Eric Blake [Fri, 26 Sep 2014 15:20:33 +0000 (09:20 -0600)]
qapi: Update docs given recent event, spacing fixes
Commit 21cd70d added event support but didn't document what the
generated code looks like. Commit 05dfb26 removed some unwanted
spaces in the generated code, but didn't reflect those changes
into the documentation. Finally, the docs start with a big
disclaimer about QMP not using QAPI yet, which feels rather stale.
Eric Blake [Fri, 26 Sep 2014 15:20:31 +0000 (09:20 -0600)]
qapi: Consistent whitespace in tests/Makefile
tests/Makefile had a mix of TAB vs. 8-space indentation; given
that it is a Makefile, TAB is more idiomatic even though in these
particular cases the choice of whitespace didn't matter.
Michael Roth [Thu, 18 Sep 2014 20:36:42 +0000 (15:36 -0500)]
tests: add QMP input visitor test for unions with no discriminator
This is more of an exercise of the dealloc visitor, where it may
erroneously use an uninitialized discriminator field as indication
that union fields corresponding to that discriminator field/type are
present, which can lead to attempts to free random chunks of heap
memory.
Michael Roth [Thu, 18 Sep 2014 20:36:40 +0000 (15:36 -0500)]
qapi: add visit_start_union and visit_end_union
In some cases an input visitor might bail out on filling out a
struct for various reasons, such as missing fields when running
in strict mode. In the case of a QAPI Union type, this may lead
to cases where the .kind field which encodes the union type
is uninitialized. Subsequently, other visitors, such as the
dealloc visitor, may use this .kind value as if it were
initialized, leading to assumptions about the union type which
in this case may lead to segfaults. For example, freeing an
integer value.
However, we can generally rely on the fact that the always-present
.data void * field that we generate for these union types will
always be NULL in cases where .kind is uninitialized (at least,
there shouldn't be a reason where we'd do this purposefully).
So pass this information on to Visitor implementation via these
optional start_union/end_union interfaces so this information
can be used to guard against the situation above. We will make
use of this information in a subsequent patch for the dealloc
visitor.
virtio-balloon: fix integer overflow in memory stats feature
When a QMP client changes the polling interval time by setting
the guest-stats-polling-interval property, the interval value
is stored and manipulated as an int64_t variable.
However, the balloon_stats_change_timer() function, which is
used to set the actual timer with the interval value, takes
an int instead, causing an overflow for big interval values.
This commit fix this bug by changing balloon_stats_change_timer()
to take an int64_t and also it limits the polling interval value
to UINT_MAX to avoid other kinds of overflow.
Commit cdaa86a54 ("Add G_IO_HUP handler for socket chardev") exposed a bug in
the way the HMP monitor handles its command buffer. When a client closes the
connection to the monitor, tcp_chr_read() will detect the G_IO_HUP condition
and call tcp_chr_disconnect() to close the server-side connection too. Due to
the fact that monitor reads 1 byte at a time (for each tcp_chr_read()), the
monitor readline state / buffers might contain junk (i.e. a half-finished
command). Thus, without calling readline_restart() on mon->rs in
CHR_EVENT_OPEN, future HMP commands will fail.
Alex Bennée [Wed, 24 Sep 2014 10:13:56 +0000 (11:13 +0100)]
.travis.yml: remove "make check" from main matrix
There are problems with unreliability in "make check" which still need
to be tracked down. As the tests are broadly the same for all targets if
added one explicit target to the matrix to run it. However this does
build all softmmu targets to ensure they at least "run"
Alex Bennée [Wed, 24 Sep 2014 10:13:55 +0000 (11:13 +0100)]
.travis.yml: pre-seed sub-modules for speed
A significant portion of the build time is spent initialising all the
sub-modules we use in the source tree. Often this is almost as long as
the build itself. By pre-seeding the .git/modules tree this will
hopefully improve things.
Alex Bennée [Wed, 24 Sep 2014 10:13:53 +0000 (11:13 +0100)]
.travis.yml: add more linux-user to the build matrix
At the same time I've grouped the $ARCH-linux-user and $ARCH-softmmu
builds together (hoping FS cache helps) and grouped all $ARCH-softmmu
only builds into one target. This reduces the build matrix slightly
which will hopefully help with build times.
Michael Roth [Thu, 18 Sep 2014 22:59:33 +0000 (17:59 -0500)]
tests: avoid running duplicate qom-tests
Since 3687d532 we've been unconditionally adding qom-test to our qtests
for every arch. However, some archs inherit their tests from Makefile
variables for other archs, such as i386/x86_64,
microblaze/microblazeel, and xtensa/xtensaeb. Since these are evaluated
in a lazy manner, we ultimately end up adding qom-test twice.
In the case x86_64, where we have a large number of machine types that
we rerun qom-test for, this has lead to a fairly noticeable increase
in the overall run-time of `make check` (78s vs. 42s on my machine).
Similar speed-ups are visible for other such archs, but not nearly as
significant.
Fix this by only adding qom-test to an arch's test list if it's not
already present.
Peter Maydell [Fri, 26 Sep 2014 14:41:50 +0000 (15:41 +0100)]
Merge remote-tracking branch 'remotes/bonzini/tags/for-upstream' into staging
Usual mix of patches, the most important being Alex and Marcelo's
kvmclock fix. This was reverted last minute for 2.1, but it is now back
with the problematic case fixed.
Note: I will soon switch to a subkey for signing purposes. To verify
future signed pull requests from me, please update my key with
"gpg --recv-keys 9B4D86F2". You should see 3 new subkeys---the
one for signing will be a 2048-bit RSA key, 4E6B09D7.
# gpg: Signature made Fri 26 Sep 2014 15:34:44 BST using RSA key ID 9B4D86F2
# gpg: Good signature from "Paolo Bonzini <[email protected]>"
# gpg: aka "Paolo Bonzini <[email protected]>"
* remotes/bonzini/tags/for-upstream:
kvm/valgrind: don't mark memory as initialized
po: fix conflict with %.mo rule in rules.mak
kvmvapic: fix migration when VM paused and when not running Windows
serial: check if backed by a physical serial port at realize time
serial: reset state at startup
target-i386: update fp status fix
hw/dma/i8257: Silence phony error message
kvmclock: Ensure time in migration never goes backward
kvmclock: Ensure proper env->tsc value for kvmclock_current_nsec calculation
Introduce cpu_clean_all_dirty
pit: fix pit interrupt can't inject into vm after migration
since commit 7dda5dc82a77 ("migration: initialize RAM to zero") the
guest memory is defined zero. No need to call valgrind on guest memory.
This reverts commit 62fe83318d2f ("qemu: Use valgrind annotations to
mark kvm guest memory as defined") thus speeding up kvm start if
<includedir>/valgrind/valgrind.h is available.
Paolo Bonzini [Mon, 22 Sep 2014 14:19:05 +0000 (16:19 +0200)]
po: fix conflict with %.mo rule in rules.mak
po/Makefile includes rules.mak to use the nice quiet-command macro.
However, this also brings in a %.mo rule that breaks "make build".
Put our own rule before the include, so that it has precedence.
Pavel Dovgalyuk [Mon, 22 Sep 2014 12:14:24 +0000 (16:14 +0400)]
kvmvapic: fix migration when VM paused and when not running Windows
This patch fixes migration by extending do_vapic_enable function. This function
called vapic_enable which read cpu number from the guest memory. When cpu
number could not be read, vapic was not enabled while loading the VM state.
This patch adds required code for cpu_number=0 to do_vapic_enable function,
because it is called only when cpu_number=0.
Peter Maydell [Fri, 26 Sep 2014 11:26:07 +0000 (12:26 +0100)]
Merge remote-tracking branch 'remotes/stefanha/tags/tracing-pull-request' into staging
# gpg: Signature made Fri 26 Sep 2014 11:59:34 BST using RSA key ID 81AB73C8
# gpg: Good signature from "Stefan Hajnoczi <[email protected]>"
# gpg: aka "Stefan Hajnoczi <[email protected]>"
* remotes/stefanha/tags/tracing-pull-request:
ohci: drop computed flags from trace events
ohci: Split long traces to smaller ones
scripts/tracetool: don't barf on formats with precision
trace: install trace-events file
trace-events: Fix comments pointing to source files
trace-events: Drop orphaned monitor trace event
trace-events: Drop unused megasas trace event
cleanup-trace-events.pl: Tighten search for trace event call
trace: tighten up trace-events regex to fix bad parse
trace-events: drop orphan iscsi trace events
trace-events: drop orphan usb_mtp_data_out
trace-events: drop orphan virtio_blk_data_plane_complete_request
trace: [hmp] Reimplement "trace-event" and "info trace-events" using QMP
trace: [qmp] Add commands to query and control event tracing state
trace: docs: add trace file description
trace: [ust] Fix format string computation in tcg-enabled events
Peter Maydell [Thu, 25 Sep 2014 15:51:09 +0000 (16:51 +0100)]
main-loop.c: Handle SIGINT, SIGHUP and SIGTERM synchronously
Add the termination signals SIGINT, SIGHUP and SIGTERM to the
list of signals which we handle synchronously via a signalfd.
This avoids a race condition where if we took the SIGTERM
in the middle of qemu_shutdown_requested:
int r = shutdown_requested;
[SIGTERM here...]
shutdown_requested = 0;
then the setting of the shutdown_requested flag by
termsig_handler() would be lost and QEMU would fail to
shut down. This was causing 'make check' to hang occasionally.
Alex Bennée [Thu, 25 Sep 2014 09:38:44 +0000 (10:38 +0100)]
ohci: drop computed flags from trace events
This exceeded the trace argument limit for LTTNG UST and wasn't really
needed as the flags value is stored anyway. Dropping this fixes the
compile failure for UST. It can probably be merged with the previous
trace shortening patch.
Recent traces rework introduced 2 tracepoints with 13 and 20
arguments. When dtrace backend is selected
(--enable-trace-backend=dtrace), compile fails as
sys/sdt.h defines DTRACE_PROBE up to DTRACE_PROBE12 only.
Stefan Hajnoczi [Tue, 23 Sep 2014 15:29:35 +0000 (16:29 +0100)]
trace: install trace-events file
Install the ./trace-events file into the data directory. This file
contains the list of trace events that were built into QEMU at
compile-time.
The file is a handy reference for the set of trace events that the QEMU
binary was built with. It is also needed by the simpletrace.py tool
that parses binary trace data either emitted from QEMU when built with
--enable-trace-backend=simple or by the SystemTap simpletrace script
that QEMU provides.
cleanup-trace-events.pl: Tighten search for trace event call
The script can get fooled too easily. For instance, it finds
trace_megasas_io_read_start when looking for trace_megasas_io_read,
and incorrectly concludes that event megasas_io_read is used.
Stefan Hajnoczi [Tue, 23 Sep 2014 10:37:06 +0000 (11:37 +0100)]
trace: tighten up trace-events regex to fix bad parse
Use \w for properties and trace event names since they are both drawn
from [a-zA-Z0-9_] character sets.
The .* for matching properties was too aggressive and caused the
following failure with foo(int rc) "(this is a test)":
Traceback (most recent call last):
File "scripts/tracetool.py", line 139, in <module>
main(sys.argv)
File "scripts/tracetool.py", line 134, in main
binary=binary, probe_prefix=probe_prefix)
File "scripts/tracetool/__init__.py", line 334, in generate
events = _read_events(fevents)
File "scripts/tracetool/__init__.py", line 262, in _read_events
res.append(Event.build(line))
File "scripts/tracetool/__init__.py", line 225, in build
return Event(name, props, fmt, args, arg_fmts)
File "scripts/tracetool/__init__.py", line 185, in __init__
% ", ".join(unknown_props))
ValueError: Unknown properties: foo(int, rc)
Stefan Hajnoczi [Mon, 22 Sep 2014 14:03:15 +0000 (15:03 +0100)]
trace-events: drop orphan iscsi trace events
iscsi_aio_write16_cb, iscsi_aio_writev, iscsi_aio_read16_cb, and
iscsi_aio_readv have not not been in use since commit 063c3378a9e3c25cc0afac3c72e4823d0621e352 ("block/iscsi: introduce
bdrv_co_{readv, writev, flush_to_disk}").
These were the only trace events in block/iscsi.c so drop the the
trace.h include.
Chen Fan [Fri, 11 Jul 2014 03:24:37 +0000 (11:24 +0800)]
trace: docs: add trace file description
When user used the trace print command from docs/tracing.txt:
./scripts/simpletrace.py trace-events trace-*
the user maybe be misled by the "trace-*", because if user
directly copy the comand line to run, there alway print the
bored message:
"usage: ./scripts/simpletrace.py <trace-events> <trace-file>"
then we should describe that the "trace-*" represented.
The signal is currently checked by 10 targets, but only actually
raised by Sparc and ARM. For the sake of one test-and-branch,
we can handle this generic bit generically.
Note that the code that was within the "exit" ifdef block
was identical to the cpu_compute_eflags inline, so make that
simplification at the same time.
Around the cpu_exec_enter/exit hooks contain many empty
ifdef blocks. Delete all of these to highlight those
targets for which we actually need to do work.
Peter Maydell [Thu, 25 Sep 2014 15:58:04 +0000 (16:58 +0100)]
Merge remote-tracking branch 'remotes/mcayland/tags/qemu-openbios-signed' into staging
Update OpenBIOS images
# gpg: Signature made Thu 25 Sep 2014 13:35:55 BST using RSA key ID AE0F321F
# gpg: Good signature from "Mark Cave-Ayland <[email protected]>"
# 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: CC62 1AB9 8E82 200D 915C C9C4 5BC2 C56F AE0F 321F
projects / qemu.git / log