Paolo Bonzini [Wed, 13 Mar 2013 14:14:15 +0000 (15:14 +0100)]
dataplane: fix hang introduced by AioContext transition
The bug is that the EventNotifiers do have a NULL io_flush callback.
Because _none_ of the callbacks on the dataplane AioContext have such a
callback, aio_poll will simply do nothing. Fixed by adding the callbacks:
the ioeventfd will always be polled (this can change in the future to
pause/resume the processing during live snapshots or similar operations);
the ioqueue will be polled if there are outstanding requests.
I must admit I have screwed up my testing somehow, because commit 2c20e71 does not work even if cherry-picked on top of 1.4.0, and this
patch fixes it there as well.
Stefan Hajnoczi [Thu, 7 Mar 2013 12:41:50 +0000 (13:41 +0100)]
coroutine: use AioContext for CoQueue BH
CoQueue uses a BH to awake coroutines that were made ready to run again
using qemu_co_queue_next() or qemu_co_queue_restart_all(). The BH
currently runs in the iothread AioContext and would break coroutines
that run in a different AioContext.
This is a slightly tricky problem because the lifetime of the BH exceeds
that of the CoQueue. This means coroutines can be awoken after CoQueue
itself has been freed. Also, there is no qemu_co_queue_destroy()
function which we could use to handle freeing resources.
Introducing qemu_co_queue_destroy() has a ripple effect of requiring us
to also add qemu_co_mutex_destroy() and qemu_co_rwlock_destroy(), as
well as updating all callers. Avoid doing that.
We also cannot switch from BH to GIdle function because aio_poll() does
not dispatch GIdle functions. (GIdle functions make memory management
slightly easier because they free themselves.)
Finally, I don't want to move unlock_queue and unlock_bh into
AioContext. That would break encapsulation - AioContext isn't supposed
to know about CoQueue.
This patch implements a different solution: each qemu_co_queue_next() or
qemu_co_queue_restart_all() call creates a new BH and list of coroutines
to wake up. Callers tend to invoke qemu_co_queue_next() and
qemu_co_queue_restart_all() occasionally after blocking I/O, so creating
a new BH for each call shouldn't be massively inefficient.
Note that this patch does not add an interface for specifying the
AioContext. That is left to future patches which will convert CoQueue,
CoMutex, and CoRwlock to expose AioContext.
Stefan Hajnoczi [Thu, 7 Mar 2013 12:41:49 +0000 (13:41 +0100)]
threadpool: drop global thread pool
Now that each AioContext has a ThreadPool and the main loop AioContext
can be fetched with bdrv_get_aio_context(), we can eliminate the concept
of a global thread pool from thread-pool.c.
The submit functions must take a ThreadPool* argument.
block/raw-posix.c and block/raw-win32.c use
aio_get_thread_pool(bdrv_get_aio_context(bs)) to fetch the main loop's
ThreadPool.
tests/test-thread-pool.c must be updated to reflect the new
thread_pool_submit() function prototypes.
Stefan Hajnoczi [Thu, 7 Mar 2013 12:41:47 +0000 (13:41 +0100)]
aio: add a ThreadPool instance to AioContext
This patch adds a ThreadPool to AioContext. It's possible that some
AioContext instances will never use the ThreadPool, so defer creation
until aio_get_thread_pool().
The reason why AioContext should have the ThreadPool is because the
ThreadPool is bound to a AioContext instance where the work item's
callback function is invoked. It doesn't make sense to keep the
ThreadPool pointer anywhere other than AioContext. For example,
block/raw-posix.c can get its AioContext's ThreadPool and submit work.
Special note about headers: I used struct ThreadPool in aio.h because
there is a circular dependency if aio.h includes thread-pool.h.
Stefan Hajnoczi [Thu, 7 Mar 2013 12:41:46 +0000 (13:41 +0100)]
threadpool: add thread_pool_new() and thread_pool_free()
ThreadPool is tied to an AioContext through its event notifier, which
dictates in which AioContext the work item's callback function will be
invoked.
In order to support multiple AioContexts we need to support multiple
ThreadPool instances.
This patch adds the new/free functions. The free function deserves
special attention because it quiesces remaining worker threads. This
requires a new condition variable and a "stopping" flag to let workers
know they should terminate once idle.
We never needed to do this before since the global threadpool was not
explicitly destroyed until process termination.
Also stash the AioContext pointer in ThreadPool so that we can call
aio_set_event_notifier() in thread_pool_free(). We didn't need to hold
onto AioContext previously since there was no free function.
Stefan Hajnoczi [Thu, 7 Mar 2013 12:41:45 +0000 (13:41 +0100)]
threadpool: move globals into struct ThreadPool
Move global variables into a struct so multiple thread pools can be
supported in the future.
This patch does not change thread-pool.h interfaces. There is still a
global thread pool and it is not yet possible to create/destroy
individual thread pools. Moving the variables into a struct first makes
later patches easier to review.
Stefan Hajnoczi [Thu, 7 Mar 2013 12:41:44 +0000 (13:41 +0100)]
main-loop: add qemu_get_aio_context()
It is very useful to get the main loop AioContext, which is a static
variable in main-loop.c.
I'm not sure whether qemu_get_aio_context() will be necessary in the
future once devices focus on using their own AioContext instead of the
main loop AioContext, but for now it allows us to refactor code to
support multiple AioContext while actually passing the main loop
AioContext.
MORITA Kazutaka [Tue, 12 Mar 2013 07:05:42 +0000 (16:05 +0900)]
sheepdog: use non-blocking fd in coroutine context
Using a blocking socket in the coroutine context reduces the chance of
switching to other work. This patch makes the sheepdog driver use a
non-blocking fd always.
Stefan Hajnoczi [Mon, 4 Mar 2013 14:02:34 +0000 (15:02 +0100)]
qcow2: drop flush in update_cluster_refcount()
The update_cluster_refcount() function increments/decrements a cluster's
refcount and then returns the new refcount value.
There is no need to flush since both update_cluster_refcount() callers
already take care of this:
1. qcow2_alloc_bytes() calls update_cluster_refcount() when compressed
sectors will be appended to an existing cluster with enough free
space. qcow2_alloc_bytes() already flushes so there is no need to do
so in update_cluster_refcount().
2. qcow2_update_snapshot_refcount() sets a cache dependency on refcounts
if it needs to update L2 entries. It also flushes before completing.
Removing this flush significantly speeds up qcow2 snapshot creation:
$ qemu-img create -f qcow2 test.qcow2 -o size=50G,preallocation=metadata
$ time qemu-img snapshot -c new test.qcow2
Time drops from more than 3 minutes to under 1 second.
Stefan Hajnoczi [Mon, 4 Mar 2013 14:02:33 +0000 (15:02 +0100)]
qcow2: flush in qcow2_update_snapshot_refcount()
Users of qcow2_update_snapshot_refcount() do not flush consistently.
qcow2_snapshot_create() flushes but qcow2_snapshot_goto() and
qcow2_snapshot_delete() do not.
Solve this by moving the bdrv_flush() into
qcow2_update_snapshot_refcount().
Stefan Hajnoczi [Mon, 4 Mar 2013 14:02:32 +0000 (15:02 +0100)]
qcow2: set L2 cache dependency in qcow2_alloc_bytes()
Compressed writes use qcow2_alloc_bytes() to allocate space with byte
granularity. The affected clusters' refcounts will be incremented but
we do not need to flush yet.
Set a L2 cache dependency on the refcount block cache, so that the
refcounts get written out before the L2 updates.
Stefan Hajnoczi [Mon, 4 Mar 2013 14:02:31 +0000 (15:02 +0100)]
qcow2: flush refcount cache correctly in qcow2_write_snapshots()
Since qcow2 metadata is cached we need to flush the caches, not just the
underlying file. Use bdrv_flush(bs) instead of bdrv_flush(bs->file).
Also add the error return path when bdrv_flush() fails and move the
flush after checking for qcow2_alloc_clusters() failure so that the
qcow2_alloc_clusters() error return value takes precedence.
Stefan Hajnoczi [Mon, 4 Mar 2013 14:02:30 +0000 (15:02 +0100)]
qcow2: flush refcount cache correctly in alloc_refcount_block()
update_refcount() affects the refcount cache, it does not write to disk.
Therefore bdrv_flush(bs->file) does nothing. We need to flush the
refcount cache in order to write out the refcount updates!
While we're here also add error returns when qcow2_cache_flush() fails.
Kevin Wolf [Fri, 15 Mar 2013 09:35:05 +0000 (10:35 +0100)]
qemu-option: Add qemu_opts_absorb_qdict()
This adds a function that adds all entries of a QDict to a QemuOpts if
the keys are known, and leaves only the rest in the QDict.
This way a single QDict of -drive options can be processed in multiple
places (generic block layer, block driver, backing file block driver,
etc.), where each part picks the options it knows. If at the end of the
process the QDict isn't empty, the user specified an invalid option.
Kevin Wolf [Fri, 15 Mar 2013 09:35:04 +0000 (10:35 +0100)]
block: Add options QDict to bdrv_open_common()
The options are passed down to the block drivers, which are supposed to
remove all options they have processed. Anything that is left over in
the end is an unknown option and results in an error.
Anthony Liguori [Thu, 14 Mar 2013 19:54:37 +0000 (14:54 -0500)]
Merge remote-tracking branch 'kraxel/chardev.5' into staging
* kraxel/chardev.5:
spice-qemu-char: Remove dead debugging code
spice-qemu-char: Fix name parameter issues after qapi-ifying
qemu-char.c: fix waiting for telnet connection message
Revert "hmp: Disable chardev-add and chardev-remove"
chardev: add udp support to qapi
chardev: add memory (ringbuf) support to qapi
chardev: add vc support to qapi
chardev: add spice support to qapi
chardev: add pipe support to qapi
chardev: add console support to qapi
chardev: switch pty init to qapi
chardev: switch parallel init to qapi
chardev: switch serial/tty init to qapi
chardev: add stdio support to qapi
chardev: switch file init to qapi
chardev: add braille support to qapi
chardev: add msmouse support to qapi
chardev: switch null init to qapi
chardev: add mux chardev support to qapi
chardev: add support for qapi-based chardev initialization
Anthony Liguori [Thu, 14 Mar 2013 19:50:58 +0000 (14:50 -0500)]
Merge remote-tracking branch 'afaerber/qom-cpu' into staging
# By Andreas Färber (16) and Igor Mammedov (1)
# Via Andreas Färber
* afaerber/qom-cpu:
target-lm32: Update VMStateDescription to LM32CPU
target-arm: Override do_interrupt for ARMv7-M profile
cpu: Replace do_interrupt() by CPUClass::do_interrupt method
cpu: Pass CPUState to cpu_interrupt()
exec: Pass CPUState to cpu_reset_interrupt()
cpu: Move halted and interrupt_request fields to CPUState
target-cris/helper.c: Update Coding Style
target-i386: Update VMStateDescription to X86CPU
cpu: Introduce cpu_class_set_vmsd()
cpu: Register VMStateDescription through CPUState
stubs: Add a vmstate_dummy struct for CONFIG_USER_ONLY
vmstate: Make vmstate_register() static inline
target-sh4: Move PVR/PRR/CVR into SuperHCPUClass
target-sh4: Introduce SuperHCPU subclasses
cpus: Replace open-coded CPU loop in qmp_memsave() with qemu_get_cpu()
monitor: Use qemu_get_cpu() in monitor_set_cpu()
cpu: Fix qemu_get_cpu() to return NULL if CPU not found
Anthony Liguori [Thu, 14 Mar 2013 19:50:30 +0000 (14:50 -0500)]
Merge remote-tracking branch 'mdroth/qga-pull-3-11-2013' into staging
# By Laszlo Ersek (3) and others
# Via Michael Roth
* mdroth/qga-pull-3-11-2013:
qga: implement qmp_guest_set_vcpus() for Linux with sysfs
qga: implement qmp_guest_get_vcpus() for Linux with sysfs
qga: introduce guest-get-vcpus / guest-set-vcpus with stubs
qga: add guest-set-time command
qga: add guest-get-time command
qemu-ga: use key-value store to avoid recycling fd handles after restart
qemu-ga: make guest-sync-delimited available during fsfreeze
qemu-ga: fix confusing GAChannelMethod comparison
qga: cast to int for DWORD type
Anthony Liguori [Thu, 14 Mar 2013 19:50:21 +0000 (14:50 -0500)]
Merge remote-tracking branch 'riku/linux-user-for-upstream' into staging
# By Peter Maydell (5) and others
# Via Riku Voipio
* riku/linux-user-for-upstream:
linux-user/syscall.c: Don't warn about unimplemented get_robust_list
linux-user: Implement accept4
linux-user: Implement sendfile and sendfile64
linux-user: make bogus negative iovec lengths fail EINVAL
linux-user: Fix layout of usage table to account for option text
linux-user: Add more sparc syscall numbers
linux-user: Support setgroups syscall with no groups
linux-user: fix futex strace of FUTEX_CLOCK_REALTIME
linux-user/syscall.c: handle FUTEX_WAIT_BITSET in do_futex
linux-user: improve print_fcntl()
linux-user: Add Alpha socket constants
It is impossible to set the debug parameter, so all the dprintf calls
are essentially nops. Since we've not needed the debug parameter in ages this
is not a problem, if it later turns out we do need some more debugging options
we can add more trace-points.
Hans de Goede [Wed, 13 Mar 2013 09:41:31 +0000 (10:41 +0100)]
spice-qemu-char: Fix name parameter issues after qapi-ifying
The strings passed in through the qapi calls are dynamic memory, since
we want to have them stick around longer then just the call to
qemu_chr_open_spice_* we need to strdup them.
Igor Mitsyanko [Sun, 10 Mar 2013 13:58:05 +0000 (17:58 +0400)]
qemu-char.c: fix waiting for telnet connection message
Current colon position in "waiting for telnet connection" message template
produces messages like:
QEMU waiting for connection on: telnet::127.0.0.16666,server
After moving a colon to the right, we will get a correct messages like:
QEMU waiting for connection on: telnet:127.0.0.1:6666,server
Gerd Hoffmann [Thu, 21 Feb 2013 10:39:12 +0000 (11:39 +0100)]
chardev: add support for qapi-based chardev initialization
This patch add support for a new way to initialize chardev devices.
Instead of calling a initialization function with a QemuOpts we will
now create a (qapi) ChardevBackend, optionally call a function to
fill ChardevBackend from QemuOpts, then go create the chardev using
the new qapi code path which is also used by chardev-add.
Gerd Hoffmann [Fri, 8 Mar 2013 10:42:24 +0000 (11:42 +0100)]
Add search path support for qemu data files.
This patch allows to specify multiple directories where qemu should look
for data files. To implement that the behavior of the -L switch is
slightly different now: Instead of replacing the data directory the
path specified will be appended to the data directory list. So when
specifiying -L multiple times all directories specified will be checked,
in the order they are specified on the command line, instead of just the
last one.
Additionally the default paths are always appended to the directory
data list. This allows to specify a incomplete directory (such as the
seabios out/ directory) via -L. Anything not found there will be loaded
from the default paths, so you don't have to create a symlink farm for
all the rom blobs.
For trouble-shooting a tracepoint has been added, logging which blob
has been loaded from which location.
Stefan Berger [Wed, 27 Feb 2013 17:47:54 +0000 (12:47 -0500)]
Add support for cancelling of a TPM command
This patch adds support for cancelling an executing TPM command.
In Linux for example a user can cancel a command through the TPM's
sysfs 'cancel' entry using
echo "1" > /sysfs/class/misc/tpm0/device/cancel
This patch propagates the cancellation of a command inside a VM
to the host TPM's sysfs entry.
It also uses the possibility to cancel the command before QEMU VM
shutdown or reboot, which helps in preventing QEMU from hanging while
waiting for the completion of the command.
To relieve higher layers or users from having to determine the TPM's
cancel sysfs entry, the driver searches for the entry in well known
locations.
Stefan Berger [Wed, 27 Feb 2013 17:47:53 +0000 (12:47 -0500)]
Add a TPM Passthrough backend driver implementation
This patch is based of off version 9 of Stefan Berger's patch series
"QEMU Trusted Platform Module (TPM) integration"
and adds a new backend driver for it.
This patch adds a passthrough backend driver for passing commands sent to the
emulated TPM device directly to a TPM device opened on the host machine.
Thus it is possible to use a hardware TPM device in a system running on QEMU,
providing the ability to access a TPM in a special state (e.g. after a Trusted
Boot).
This functionality is being used in the acTvSM Trusted Virtualization Platform
which is available on [1].
Some notes about the host TPM:
The TPM needs to be enabled and activated. If that's not the case one
has to go through the BIOS/UEFI and enable and activate that TPM for TPM
commands to work as expected.
It may be necessary to boot the kernel using tpm_tis.force=1 in the boot
command line or 'modprobe tpm_tis force=1' in case of using it as a module.
Stefan Berger [Wed, 27 Feb 2013 17:47:51 +0000 (12:47 -0500)]
Add a debug register
This patch uses the possibility to add a vendor-specific register and
adds a debug register useful for dumping the TIS's internal state. This
register is only active in a debug build (#define DEBUG_TIS).
Stefan Berger [Wed, 27 Feb 2013 17:47:50 +0000 (12:47 -0500)]
Add TPM (frontend) hardware interface (TPM TIS) to QEMU
This patch adds the main code of the TPM frontend driver, the TPM TIS
interface, to QEMU. The code is largely based on the previous implementation
for Xen but has been significantly extended to meet the standard's
requirements, such as the support for changing of localities and all the
functionality of the available flags.
Communication with the backend (i.e., for Xen or the libtpms-based one)
is cleanly separated through an interface which the backend driver needs
to implement.
Whenever the frontend has collected a complete packet, it will submit
a task to the backend, which then starts processing the command. Once
the result has been returned, the backend invokes a callback function
(tpm_tis_receive_cb()).
Testing the proper functioning of the different flags and localities
cannot be done from user space when running in Linux for example, since
access to the address space of the TPM TIS interface is not possible. Also
the Linux driver itself does not exercise all functionality. So, for
testing there is a fairly extensive test suite as part of the SeaBIOS patches
since from within the BIOS one can have full access to all the TPM's registers.
Stefan Berger [Wed, 27 Feb 2013 17:47:49 +0000 (12:47 -0500)]
Support for TPM command line options
This patch adds support for TPM command line options.
The command line options supported here are
./qemu-... -tpmdev passthrough,path=<path to TPM device>,id=<id>
-device tpm-tis,tpmdev=<id>,id=<other id>
and
./qemu-... -tpmdev help
where the latter works similar to -soundhw help and shows a list of
available TPM backends (for example 'passthrough').
Using the type parameter, the backend is chosen, i.e., 'passthrough' for the
passthrough driver. The interpretation of the other parameters along
with determining whether enough parameters were provided is pushed into
the backend driver, which needs to implement the interface function
'create' and return a TPMDriverOpts structure if the VM can be started or
'NULL' if not enough or bad parameters were provided.
Monitor support for 'info tpm' has been added. It for example prints the
following:
(qemu) info tpm
TPM devices:
tpm0: model=tpm-tis
\ tpm0: type=passthrough,path=/dev/tpm0,cancel-path=/sys/devices/pnp0/00:09/cancel
Andreas Färber [Sat, 2 Feb 2013 11:33:14 +0000 (12:33 +0100)]
target-arm: Override do_interrupt for ARMv7-M profile
Enable ARMCPUInfo to specify a custom class_init functions.
Introduce arm_v7m_class_init() and use it for "cortex-m3" model.
Instead of forwarding from arm_cpu_do_interrupt() to do_interrupt_v7m(),
override CPUClass::do_interrupt with arm_v7m_cpu_do_interrupt()
in arm_v7m_class_init().
Andreas Färber [Fri, 25 Jan 2013 16:37:28 +0000 (17:37 +0100)]
target-cris/helper.c: Update Coding Style
Reindent, add missing braces and drop/adjust whitespace.
Prepares for CPUArchState-to-CPUState field movements in
cpu_cris_handle_mmu_fault(), do_interruptv10() and do_interrupt().
The remaining functions were so minor that they can be fixed in one go.
Andreas Färber [Sun, 20 Jan 2013 19:23:22 +0000 (20:23 +0100)]
cpu: Register VMStateDescription through CPUState
In comparison to DeviceClass::vmsd, CPU VMState is split in two,
"cpu_common" and "cpu", and uses cpu_index as instance_id instead of -1.
Therefore add a CPU-specific CPUClass::vmsd field.
Unlike the legacy CPUArchState registration, rather register CPUState.
Andreas Färber [Fri, 15 Feb 2013 14:41:49 +0000 (15:41 +0100)]
cpus: Replace open-coded CPU loop in qmp_memsave() with qemu_get_cpu()
No functional change, just less usages of first_cpu and next_cpu fields.
env is passed to cpu_memory_rw_debug(), which in turn passes it to
target-specific cpu_get_phys_page_debug(). Changing both would be a
larger refactoring, so defer that by using env_ptr for now.
Andreas Färber [Fri, 15 Feb 2013 16:01:09 +0000 (17:01 +0100)]
monitor: Use qemu_get_cpu() in monitor_set_cpu()
No functional change, just a reduction of CPU loops.
The mon_cpu field is left untouched for now since changing that requires
a number of larger prerequisites, including cpu_synchronize_state() and
mon_get_cpu().
Igor Mammedov [Thu, 7 Mar 2013 18:12:43 +0000 (19:12 +0100)]
cpu: Fix qemu_get_cpu() to return NULL if CPU not found
Commit 55e5c2850 breaks CPU not found return value, and returns
CPU corresponding to the last non NULL env.
Fix it by returning CPU only if env is not NULL, otherwise CPU is
not found and function should return NULL.
Michael Roth [Fri, 1 Mar 2013 17:40:27 +0000 (11:40 -0600)]
qemu-ga: use key-value store to avoid recycling fd handles after restart
Hosts hold on to handles provided by guest-file-open for periods that can
span beyond the life of the qemu-ga process that issued them. Since these
are issued starting from 0 on every restart, we run the risk of issuing
duplicate handles after restarts/reboots.
As a result, users with a stale copy of these handles may end up
reading/writing corrupted data due to their existing handles effectively
being re-assigned to an unexpected file or offset.
We unfortunately do not issue handles as strings, but as integers, so a
solution such as using UUIDs can't be implemented without introducing a
new interface.
As a workaround, we fix this by implementing a persistent key-value store
that will be used to track the value of the last handle that was issued
across restarts/reboots to avoid issuing duplicates.
The store is automatically written to the same directory we currently
set via --statedir to track fsfreeze state, and so should be applicable
for stable releases where this flag is supported.
A follow-up can use this same store for handling fsfreeze state, but
that change is cosmetic and left out for now.
Signed-off-by: Michael Roth <[email protected]> Cc: [email protected]
* fixed guest_file_handle_add() return value from uint64_t to int64_t
Michael Roth [Fri, 1 Mar 2013 17:49:38 +0000 (11:49 -0600)]
qemu-ga: make guest-sync-delimited available during fsfreeze
We currently maintain a whitelist of commands that are safe during
fsfreeze. During fsfreeze, we disable all commands that aren't part of
that whitelist.
guest-sync-delimited meets the criteria for being whitelisted, and is
also required for qemu-ga clients that rely on guest-sync-delimited for
re-syncing the channel after a timeout.
Stefan Hajnoczi [Tue, 19 Feb 2013 14:12:34 +0000 (15:12 +0100)]
qemu-ga: fix confusing GAChannelMethod comparison
In commit 7868e26e5930f49ca942311885776b938dcf3b77
("qemu-ga: add initial win32 support") support was added for qemu-ga on
Windows using virtio-serial. Other channel methods (ISA serial and UNIX
domain socket) are not supported on Windows.
Lei Li [Mon, 28 Jan 2013 04:49:09 +0000 (12:49 +0800)]
qga: cast to int for DWORD type
This patch fixes a compiler warning when cross-build:
qga/service-win32.c: In function 'printf_win_error':
qga/service-win32.c:32:5: warning: format '%d' expects argument of type 'int',
but argument 3 has type 'DWORD' [-Wformat]
Anthony Liguori [Mon, 11 Mar 2013 13:30:34 +0000 (08:30 -0500)]
Merge remote-tracking branch 'quintela/migration.next' into staging
# By Paolo Bonzini (40) and others
# Via Juan Quintela
* quintela/migration.next: (46 commits)
page_cache: dup memory on insert
page_cache: fix memory leak
Fix cache_resize to keep old entry age
Fix page_cache leak in cache_resize
migration: inline migrate_fd_close
migration: eliminate s->migration_file
migration: move contents of migration_close to migrate_fd_cleanup
migration: move rate limiting to QEMUFile
migration: small changes around rate-limiting
migration: use qemu_ftell to compute bandwidth
migration: use QEMUFile for writing outgoing migration data
migration: use QEMUFile for migration channel lifetime
qemu-file: simplify and export qemu_ftell
qemu-file: add writable socket QEMUFile
qemu-file: check exit status when closing a pipe QEMUFile
qemu-file: fsync a writable stdio QEMUFile
migration: merge qemu_popen_cmd with qemu_popen
migration: use qemu_file_rate_limit consistently
migration: remove useless qemu_file_get_error check
migration: detect error before sleeping
...
Paolo Bonzini [Mon, 11 Mar 2013 10:31:26 +0000 (11:31 +0100)]
arm: fix compilation with CONFIG_FDT
A conflict was resolved the wrong way when merging commit 320ba5f (build:
always link device_tree.o into emulators if libfdt available, 2013-02-05).
This causes a build failure for the arm-softmmu target due to multiply
defined symbol.
Peter Maydell [Fri, 8 Feb 2013 04:34:54 +0000 (04:34 +0000)]
linux-user/syscall.c: Don't warn about unimplemented get_robust_list
The nature of the kernel ABI for the get_robust_list and set_robust_list
syscalls means we cannot implement them in QEMU. Make get_robust_list
silently return ENOSYS rather than using the default "print message and
then fail ENOSYS" code path, in the same way we already do for
set_robust_list, and add a comment documenting why we do this.
This silences warnings which were being produced for emulating
even trivial programs like 'ls' in x86-64-on-x86-64.
Peter Maydell [Fri, 8 Feb 2013 07:58:41 +0000 (07:58 +0000)]
linux-user: make bogus negative iovec lengths fail EINVAL
If the guest passes us a bogus negative length for an iovec, fail
EINVAL rather than proceeding blindly forward. This fixes some of
the error cases tests for readv and writev in the LTP.
Peter Maydell [Thu, 14 Feb 2013 08:46:43 +0000 (08:46 +0000)]
linux-user: Fix layout of usage table to account for option text
The linux-user usage message attempts to line up the columns in
its table by calculating the maximum width of any item in them.
However for the 'Argument' column it was only accounting for the
length of the option switch (eg "-d"), not the additional example
text (eg "item[,...]"). This currently has no adverse effects
because the widest item in the column happens to be the argumentless
"-singlestep" option, but improving the "-d" option help to read
"-d item[,...]" exceeds that limit.
Fix this by correctly calculating maxarglen as the width of the
first column text including a possible option argument, and
adjusting its uses to match.
Peter Lieven [Mon, 25 Feb 2013 17:12:04 +0000 (19:12 +0200)]
page_cache: dup memory on insert
The page cache frees all data on finish, on resize and
if there is collision on insert. So it should be the caches
responsibility to dup the data that is stored in the cache.
Peter Lieven [Mon, 25 Feb 2013 17:12:03 +0000 (19:12 +0200)]
page_cache: fix memory leak
XBZRLE encoded migration introduced a MRU page cache
meachnism. Unfortunately, cached items where never freed in
case of a collision in the page cache on cache_insert().
This lead to out of memory conditions during XBZRLE migration
if the page cache was small and there where a lot of collisions
in the cache.
projects / qemu.git / log