tb_lock has to be taken inside the mmap_lock (example:
tb_invalidate_phys_range is called by target_mmap), but
tb_link_page is taking the mmap_lock and it is called
with the tb_lock held.
To fix this, take the mmap_lock in tb_find_slow, not
in tb_link_page.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
There is some iffy lock hierarchy going on in translate-all.c. To
fix it, we need to take the mmap_lock in cpu-exec.c. Make the
functions globally available.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
page_find is reading the radix tree outside all locks, so it has to
use the RCU primitives. It does not need RCU critical sections
because the PageDescs are never removed, so there is never a need
to wait for the end of code sections that use a PageDesc.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
This just removes spinlock as it is not used anymore.
Signed-off-by: KONRAD Frederic <fred.konrad@greensocs.com>
Message-Id: <1439220437-23957-6-git-send-email-fred.konrad@greensocs.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
spinlock is only used in two cases:
* cpu-exec.c: to protect TranslationBlock
* mem_helper.c: for lock helper in target-i386 (which seems broken).
It's a pthread_mutex_t in user-mode, so we can use QemuMutex directly,
with an #ifdef. The #ifdef will be removed when multithreaded TCG
will need the mutex as well.
Signed-off-by: KONRAD Frederic <fred.konrad@greensocs.com>
Message-Id: <1439220437-23957-5-git-send-email-fred.konrad@greensocs.com>
Signed-off-by: Emilio G. Cota <cota@braap.org>
[Merge Emilio G. Cota's patch to remove volatile. - Paolo]
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
This hides the tcg_halt_cond and tcg_cpu_thread global variables
inside qemu_tcg_init_vcpu. Multi-threaded TCG will need one
QemuCond and one QemuThread per virtual cpu, so it's preferrable
to use cpu->halt_cond and cpu->thread.
Signed-off-by: KONRAD Frederic <fred.konrad@greensocs.com>
Message-Id: <1439220437-23957-9-git-send-email-fred.konrad@greensocs.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Protect the list of queued work items with something other than
the BQL, as a preparation for running the work items outside it.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: KONRAD Frederic <fred.konrad@greensocs.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
commit 9b8424d573
"exec: split length -> used_length/max_length"
changed field names in struct RAMBlock
It turns out that scripts/dump-guest-memory.py was
poking at this field, update it accordingly.
Cc: qemu-stable@nongnu.org
Cc: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Message-Id: <1440666378-3152-1-git-send-email-mst@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
This adds "--enable-jemalloc" and "--disable-jemalloc" to allow linking
to jemalloc memory allocator.
We have already tcmalloc support,
but it seem to not working well with a lot of iothreads/disks.
The main problem is that tcmalloc use a shared thread cache of 16MB
by default.
With more threads, this cache is shared, and some bad garbage collections
can occur if the cache is too low.
It's possible to tcmalloc cache increase it with a env var:
TCMALLOC_MAX_TOTAL_THREAD_CACHE_BYTES=256MB
With default 16MB, performances are really bad with more than 2 disks.
Increasing to 256MB, it's helping but still have problem with 16 disks/iothreads.
Jemalloc don't have performance problem with default configuration.
Here the benchmark results in iops of 1 qemu vm randread 4K iodepth=32,
with rbd block backend (librbd is doing a lot of memory allocation),
1 iothread by disk
glibc malloc
------------
1 disk 29052
2 disks 55878
4 disks 127899
8 disks 240566
15 disks 269976
jemalloc
--------
1 disk 41278
2 disks 75781
4 disks 195351
8 disks 294241
15 disks 298199
tcmalloc 2.2.1 default 16M cache
--------------------------------
1 disk 37911
2 disks 67698
4 disks 41076
8 disks 43312
15 disks 37569
tcmalloc : 256M cache
---------------------------
1 disk 33914
2 disks 58839
4 disks 148205
8 disks 213298
15 disks 218383
Signed-off-by: Alexandre Derumier <aderumier@odiso.com>
Message-Id: <1434711418-20429-1-git-send-email-aderumier@odiso.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Coccinelle chokes on some idioms from compiler.h and queue.h.
Extract those in a macro file, to be used with "--macro-file
scripts/cocci-macro-file.h".
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Every arch adds its disas configury to both its own config as well
config_disas_all. Make a small function do to both at once.
Signed-off-by: Peter Crosthwaite <crosthwaite.peter@gmail.com>
Message-Id: <1440844439-19391-1-git-send-email-crosthwaite.peter@gmail.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Fully removing Sparse support requires more invasive changes. Only
remove the really kernel-specific parts such as address space names.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Mostly change severity levels, but some tests can also be adjusted to refer
to QEMU APIs or data structures.
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Mixed declarations do come in handy at the top of #ifdef blocks.
Reluctantly allow this particular usage and suggest an alternative.
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Carlos L. Torres <carlos.torres@rackspace.com>
Message-Id: <11ac63e95d88551f1c2c9b1216b15d3cb8ba4468.1437346779.git.carlos.torres@rackspace.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Add wrapper for strtoull() function. Include unit tests.
Signed-off-by: Carlos L. Torres <carlos.torres@rackspace.com>
Message-Id: <e0f0f611c9a81f3c29f451d0b17d755dfab1e90a.1437346779.git.carlos.torres@rackspace.com>
[Use uint64_t in prototype. - Paolo]
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Add wrapper for strtoll() function. Include unit tests.
Signed-off-by: Carlos L. Torres <carlos.torres@rackspace.com>
Message-Id: <7454a6bb9ec03b629e8beb4f109dd30dc2c9804c.1437346779.git.carlos.torres@rackspace.com>
[Use int64_t in prototype, since that's what QEMU uses. - Paolo]
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Add wrapper for strtoul() function. Include unit tests.
Signed-off-by: Carlos L. Torres <carlos.torres@rackspace.com>
Message-Id: <9621b4ae8e35fded31c715c2ae2a98f904f07ad0.1437346779.git.carlos.torres@rackspace.com>
[Fix tests for 32-bit build. - Paolo]
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Add wrapper for strtol() function. Include unit tests.
Signed-off-by: Carlos L. Torres <carlos.torres@rackspace.com>
Message-Id: <07199f1c0ff3892790c6322123aee1e92f580550.1437346779.git.carlos.torres@rackspace.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
l1_map is based on physical addresses in full-system mode, as pointed
out in an earlier comment. Said comment also mentions that virtual
addresses are only used in l1_map in user-only mode.
Signed-off-by: Emilio G. Cota <cota@braap.org>
Message-Id: <1440375847-17603-11-git-send-email-cota@braap.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
We were unlocking this lock after fork, which is wrong since
only the thread that holds a mutex is allowed to unlock it.
Signed-off-by: Emilio G. Cota <cota@braap.org>
Message-Id: <1440375847-17603-9-git-send-email-cota@braap.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Using ccache with CCACHE_BASEDIR set to $(SRC_PATH) or a parent will
rewrite all absolute paths to relative paths. This interacts poorly with
QEMU's two-level build directory scheme. For example, lets say
BUILD_DIR=$(SRC_PATH)/build so build/blockdev.d will contain:
blockdev.o: ../blockdev.c ../include/sysemu/block-backend.h \
Now the target build under build/x86_64-softmmu or similar will depend
on ../blockdev.o which in turn will get make to source ../blockdev.d to
check its dependencies. Since make always considers paths relative to
the current working directory rather than the makefile the path appeared
in the relative path to ../blockdev.c is useless.
This change simply adds the top level build directory to vpath so paths
relative to the source directory, top build directory, and target build
directory all work just fine.
Signed-off-by: Michael Marineau <michael.marineau@coreos.com>
Message-Id: <1439103775-11836-1-git-send-email-michael.marineau@coreos.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <1439547914-18249-1-git-send-email-dgilbert@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signals are slow and do not exist on Win32. The previous patches
have done most of the legwork to introduce memory barriers (some
of them were even there already for the sake of Windows!) and
we can now set the flags directly in the iothread.
qemu_cpu_kick_thread is not used anymore on TCG, since the TCG thread is
never outside usermode while the CPU is running (not halted). Instead run
the content of the signal handler (now in qemu_cpu_kick_no_halt) directly.
qemu_cpu_kick_no_halt is also used in qemu_mutex_lock_iothread to avoid
the overhead of qemu_cond_broadcast.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Use the same API to trigger interruption of a CPU, no matter if
under TCG or KVM. There is no difference: these calls come from
the CPU thread, so the qemu_cpu_kick calls will send a signal
to the running thread and it will be processed synchronously,
just like a call to cpu_exit. The only difference is in the
overhead, but neither call to cpu_exit (now qemu_cpu_kick)
is in a hot path.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Synchronize the remaining pair of accesses in cpu_signal. These should
be necessary on Windows as well, at least in theory. Probably
SuspendProcess and ResumeProcess introduce some implicit memory
barrier.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
TCG has not been reading cpu->current_tb from signal handlers for years.
The code that synchronized cpu_exec with the signal handler is not
needed anymore.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
TLS is now required on all platforms, so DECLARE_TLS/DEFINE_TLS is not
needed anymore. Removing it does not break Windows because of the
previous patch.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
This is already useful on Windows in order to remove tls.h, because
accesses to current_cpu are done from a different thread on that
platform. It will be used on POSIX platforms as soon TCG stops using
signals to interrupt the execution of translated code.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
This is unused. cpu_exit now is almost exclusively an internal function
to the CPU execution loop. In a few patches, we'll change the remaining
occurrences to qemu_cpu_kick, making it truly internal.
Reviewed-by: Richard henderson <rth@twiddle.net>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
The i8257 DMA controller uses an idle bottom half, which by default
does not cause the main loop to exit. Therefore, the DMA_schedule
function is there to ensure that the CPU relinquishes the iothread
mutex to the iothread.
However, this is not enough since the iothread will call
aio_compute_timeout() and go to sleep again. In the iothread
world, forcing execution of the idle bottom half is much simpler,
and only requires a call to qemu_notify_event(). Do it, removing
the need for the "cpu_request_exit" pseudo-irq. The next patch
will remove it.
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
There are pieces of guest panic handling code
that can be shared in one generic function.
These code replaced by call qemu_system_guest_panicked().
Signed-off-by: Andrey Smetanin <asmetanin@virtuozzo.com>
Signed-off-by: Denis V. Lunev <den@openvz.org>
CC: Paolo Bonzini <pbonzini@redhat.com>
CC: Andreas Färber <afaerber@suse.de>
Message-Id: <1435924905-8926-10-git-send-email-den@openvz.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
It has been reported that at least tgtd returns a block size of 0
for LUN 0. To avoid running into divide by zero later on and protect
against other problematic block sizes validate the block size right
at connection time.
Cc: qemu-stable@nongnu.org
Reported-by: Andrey Korolyov <andrey@xdel.ru>
Signed-off-by: Peter Lieven <pl@kamp.de>
Message-Id: <1439552016-8557-1-git-send-email-pl@kamp.de>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
iohandler.c shares the same interface with aio, but with duplicated
code. It's better to rebase iohandler, also because that aio is a
more friendly interface to multi-threads.
Create a global AioContext instance and let its GSource handle the
iohandler events.
Signed-off-by: Fam Zheng <famz@redhat.com>
Message-Id: <1441596538-4412-1-git-send-email-famz@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
When QEMU starts the RCU thread executes qemu_mutex_lock_thread
causing error "qemu:qemu_cpu_kick_thread: No such process" and exits.
This isn't occur frequently but in glibc the thread id can exist and
this not guarantee that the thread is on active/running state. If is
inserted a sleep(1) after newthread assignment [1] the issue appears.
So not make assumption that thread exist if first_cpu->thread is set
then change the validation of cpu to created that is set into cpu
threads (kvm, tcg, dummy).
[1] https://sourceware.org/git/?p=glibc.git;a=blob;f=nptl/pthread_create.c;h=d10f4ea8004e1d8f3a268b95cc0f8d93b8d89867;hb=HEAD#l621
Cc: qemu-stable@nongnu.org
Signed-off-by: Aníbal Limón <anibal.limon@linux.intel.com>
Message-Id: <1441313313-3040-1-git-send-email-anibal.limon@linux.intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
With this change we make sure that the compiler will not
optimise the read of the sequence number in any way.
Signed-off-by: Emilio G. Cota <cota@braap.org>
Message-Id: <1440375847-17603-8-git-send-email-cota@braap.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Emilio G. Cota <cota@braap.org>
Message-Id: <1440375847-17603-7-git-send-email-cota@braap.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
guest_base must be used only in linux-user mode.
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
Message-id: 1440757421-9674-1-git-send-email-laurent@vivier.eu
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Fix a capitalization error in the OS X build instructions;
this was picked up in review of commit b352153f5f and intended to be
corrected before I applied it, but I accidentally didn't include it.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
qemu-doc.texi: Add information on compiling source code on Mac OS X
Add information to the documentation on how to build QEMU
on Mac OS X.
Signed-off-by: John Arbuckle <programmingkidx@gmail.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
[PMM: fixed a minor capitalization error]
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1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=C7Zf
-----END PGP SIGNATURE-----
Merge remote-tracking branch 'remotes/weil/tags/pull-tci-20150826' into staging
tci patch queue
# gpg: Signature made Wed 26 Aug 2015 19:51:07 BST using RSA key ID 677450AD
# gpg: Good signature from "Stefan Weil <sw@weilnetz.de>"
# gpg: aka "Stefan Weil <stefan.weil@weilnetz.de>"
# gpg: aka "Stefan Weil <stefan.weil@bib.uni-mannheim.de>"
# 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: 4923 6FEA 75C9 5D69 8EC2 B78A E08C 21D5 6774 50AD
* remotes/weil/tags/pull-tci-20150826:
exec-all: Translate TCI return addresses backwards too
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
This subtraction of return addresses applies directly to TCI as well as
host-TCG. This fixes Linux boots for at least Microblaze, CRIS, ARM and
SH4 when using TCI.
[sw: Removed indentation for preprocessor statement]
[sw: The patch also fixes Linux boot for x86_64]
Reviewed-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Peter Crosthwaite <crosthwaite.peter@gmail.com>
The _cmp_bytes variable added by commit "bea60dd ui/vnc: fix potential
memory corruption issues" can become negative. Result is (possibly
exploitable) memory corruption. Reason for that is it uses the stride
instead of bytes per scanline to apply limits.
For the server surface is is actually fine. vnc creates that itself,
there is never any padding and thus scanline length always equals stride.
For the guest surface scanline length and stride are typically identical
too, but it doesn't has to be that way. So add and use a new variable
(guest_ll) for the guest scanline length. Also rename min_stride to
line_bytes to make more clear what it actually is. Finally sprinkle
in an assert() to make sure we never use a negative _cmp_bytes again.
Reported-by: 范祚至(库特) <zuozhi.fzz@alibaba-inc.com>
Reviewed-by: P J P <ppandit@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
