Signed-off-by: Tomasz Grabiec <tgrabiec@cloudius-systems.com>
Signed-off-by: Avi Kivity <avi@cloudius-systems.com>

Dynamically loaded modules use __tls_get_addr() to locate thread local
symbols. Symbol is identified by module index and offset in module's
TLS area. Module index and offset are filled in by dynamic linker when
DSO is loaded.

TLS area for given DSO is allocated dynamically, on-demand. OSv keeps
TLS areas in a vector indexed by module index, inside per-thread
vector.

TLS area of core module should be handled differently than that of
dynamically loaded modules. The TLS offsets for thread local symbols
defined in core module are known at link time and code inside core
module can use these offsets directly. The offsets are relative to TCB
pointer (fs register on x86).

The problem was that __tls_get_addr() was treating core module as a
dynamically loaded module and returned pointer inside dynamically
allocated TLS area instead of a pointer inside core module's TLS. As a
result code inside core module was reading value from different
location than code inside DSO has written value to.

The offending thread local varaible was __once_call. It was set by
call_once() defined in DSO (inlined from a definition inside header)
and read by __once_proxy() defined in core module.

Fixes #125.

Signed-off-by: Tomasz Grabiec <tgrabiec@cloudius-systems.com>
Signed-off-by: Avi Kivity <avi@cloudius-systems.com>

In order to have unform naming, ulong is used in several places.

Signed-off-by: Tomasz Grabiec <tgrabiec@cloudius-systems.com>
Signed-off-by: Avi Kivity <avi@cloudius-systems.com>

Signed-off-by: Tomasz Grabiec <tgrabiec@cloudius-systems.com>
Signed-off-by: Avi Kivity <avi@cloudius-systems.com>

Clean up virtio-blk.cc by using 'auto' type specifier where possible.

Reviewed-by: Dor Laor <dor@cloudius-systems.com>
Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

Move the x86-64 PTE definitions to a new arch specific arch-mmu.hh
header file to make core/mmu.cc smaller and more portable.

Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

If iov->iov_len == 0, it will loop forever.
uiomove() will take care of the zero iov len case.

Signed-off-by: Asias He <asias@cloudius-systems.com>
Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

Signed-off-by: Avi Kivity <avi@cloudius-systems.com>
Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

Signed-off-by: Avi Kivity <avi@cloudius-systems.com>
Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

The bsd specific ones are nenumbered to avoid clashes.

Signed-off-by: Avi Kivity <avi@cloudius-systems.com>
Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

Signed-off-by: Avi Kivity <avi@cloudius-systems.com>
Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

Since most SOL_* macros are equivalent to the IPPROTO_* defines, only one
define needs to be changed.

Signed-off-by: Avi Kivity <avi@cloudius-systems.com>
Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

Move them to a common include file.

Since they're defined externally, there is no real conflict.

Signed-off-by: Avi Kivity <avi@cloudius-systems.com>
Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

Unused.

Signed-off-by: Avi Kivity <avi@cloudius-systems.com>
Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

Switched the virtio-net driver to use if_transmit() instead of legacy
if_start(). This saves us at least 2 additional lock/unlock sequences
per-each mbuf since IF_ENQUEUE() and IF_DEQUEUE() take lock when
pushing/removing the mbuf from the queue if ifnet in a legacy mode.

Signed-off-by: Vlad Zolotarov <vladz@cloudius-systems.com>
Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

Now that backtrace() doesn't use libunwind, we can remove it.

Signed-off-by: Avi Kivity <avi@cloudius-systems.com>

This patch changes backtrace() to use the _Unwind_* facilities provided
by the GCC runtime (libgcc_eh.a), instead of the separate libunwind.a.

After this patch, we don't use libunwind.a in OSv any more, and it can
be removed (see issue #83).

Signed-off-by: Nadav Har'El <nyh@cloudius-systems.com>
Signed-off-by: Avi Kivity <avi@cloudius-systems.com>

thread::destroy() had a "FIXME" comment:
// FIXME: we have a problem in case of a race between join() and the
// thread's completion. Here we can see _joiner==0 and not notify
// anyone, but at the same time join() decided to go to sleep (because
// status is not yet status::terminated) and we'll never wake it.

This is indeed a bug, which Glauber discovered was hanging the
tst-threadcomplete.so test once in a while - the test sometimes hangs
with one thread in the "terminated" state (waiting for someone to join
it), and a second thread waiting in join() but missed the other thread's
termination event.

The solution works like this:

join() uses a CAS to set itself as the _joiner. If it succeeded, it
waits like before for the status to become "terminated". But if the CAS
failed, it means a concurrent destroy() call beat us at the race, and we
can just return from join().

destroy() checks (with a CAS) if _joiner was already set - if so we need
to wake this thread just like in the original code. But if _joiner was
not yet set, either there is no-one doing join(), or there's a concurrent
join() call that will soon return (this is what the joiner does when it
loses the CAS race). In this case, all we need to do is to set the status
to "terminated" - and we must do it through a _detached_state we saved
earlier, because if join() already returned the thread may already be
deleted).

Signed-off-by: Nadav Har'El <nyh@cloudius-systems.com>
Signed-off-by: Avi Kivity <avi@cloudius-systems.com>

wake_with(action) was implemented using thread_handle, as the following:

thread_handle h(handle());
action();
h.wake();

This implementation is wrong: It only takes the RCU lock (which prevents
the destruction of _detached_state) during h.wake(), meaning that if the
thread is not sleeping, and action() causes it to exit, _detached_state
may also be destructed, and h.wake() will crash.

thread_handle is simply not needed for wake_with(), and was designed
with a completely different use case in mind (long-term holding of a
thread handler). We just need to use, in-line, the appropriate rcu
lock which keeps _detached_state alive. The resulting code is even
simpler, and nicely parallels the existing code of wake().

This patch fixes a real bug, but unfortunately we don't have a concrete
test-case which it is known to fix.

Signed-off-by: Nadav Har'El <nyh@cloudius-systems.com>
Signed-off-by: Avi Kivity <avi@cloudius-systems.com>

Add a new lock, "rcu_read_lock_in_preempt_disabled", which is exactly
like rcu_read_lock but assuming that preemption is already disabled.
Because all our rcu_read_lock does is to disable preemption, the new
lock type currently does absolutely nothing - but in some future
implementation of RCU it might need to do something.

We'll use the new lock type in the following patch, as an optimization
over the regular rcu_read_lock.

Signed-off-by: Nadav Har'El <nyh@cloudius-systems.com>
Signed-off-by: Avi Kivity <avi@cloudius-systems.com>

Context: going to wait with irqs_disabled is a call for disaster.  While it is
true that not every time we call wait we actually end up waiting, that should
be an invalid call, due to the times we may wait. Because of that, it would
be good to express that nonsense in an assertion.

There is however, places we sleep with irqs disabled currently. Although they
are technically safe, because we implicitly enable interrupts, they end up
reaching wait() in a non-safe state. That happens in the page fault handler.
Explicitly enabling interrupts will allow us to test for valid / invalid wait
status.

With this test applied, all tests in our whitelist still passes.

Signed-off-by: Glauber Costa <glommer@cloudius-systems.com>
Signed-off-by: Avi Kivity <avi@cloudius-systems.com>

Signed-off-by: Raphael S. Carvalho <raphaelsc@cloudius-systems.com>
Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

Signed-off-by: Raphael S. Carvalho <raphaelsc@cloudius-systems.com>
Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

Signed-off-by: Raphael S. Carvalho <raphaelsc@cloudius-systems.com>
Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

Fix a unicode character in the author's name.

Signed-off-by: Avi Kivity <avi@cloudius-systems.com>

Make sure that the address range passed to munmap() is actually mapped.

Reviewed-by: Raphael S. Carvalho <raphaelsc@cloudius-systems.com>
Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

Simplify mmap() by converting flags and permissions in one place.

Reviewed-by: Raphael S. Carvalho <raphaelsc@cloudius-systems.com>
Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

Move mincore() to libc/mman.cc where all other memory mapping libc
functions are.

Reviewed-by: Raphael S. Carvalho <raphaelsc@cloudius-systems.com>
Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

Add a mmu::is_page_aligned() helper function and use it to get rid of
open-coded checks.

Reviewed-by: Raphael S. Carvalho <raphaelsc@cloudius-systems.com>
Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

 - It's compiled out when mode=release.
 - Uses an assert() for issuing the assert.
 - Has a printf-like semantics.

Signed-off-by: Vlad Zolotarov <vladz@cloudius-systems.com>
Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

 - Add -DNDEBUG to the compiler flags when mode!=debug.
 - Prevent assert() from compiling out in kernel when mode=release

Signed-off-by: Vlad Zolotarov <vladz@cloudius-systems.com>
Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

Fix some compilation errors that would arrise when NDEBUG is defined:
 - tests/misc-tcp.cc: missing #include<iostream> that would come from
   include/boost/assert.hpp: line 81 when NDEBUG is not defined.
 - bsd/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_dataset.c:
   when NDEBUG is not defined assert() is defined to __assert_fail,
   which has "noreturn" attribute, which satisfies the compiler.
   When NDEBUG is defined and assert() completly compiles out, the
   compiler start complaining about the missing "return" statment.

Signed-off-by: Vlad Zolotarov <vladz@cloudius-systems.com>
Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

Add the missing #infdef X #define X protection to include/api/assert.h

Signed-off-by: Vlad Zolotarov <vladz@cloudius-systems.com>
Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

This patch adds sorting as an option (-s). Examples:

 Not sorted:

   gdb$ osv trace duration

 Sorted, narrowed down to one function:

   gdb$ osv trace duration -s vfs_pwritev

Not sorting allows us to start printing traces right away. There's
also no need to keep them in memory, which makes the command a bit
faster.

Signed-off-by: Tomasz Grabiec <tgrabiec@cloudius-systems.com>
Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

Make blacklisted tests visible during test.py run to attract attention
to them and hopefully get them fixed:

  TEST tst-strerror_r.so         OK  (1.083 s)
  TEST tst-threadcomplete.so     SKIPPED
  TEST tst-tracepoint.so         OK  (1.120 s)

Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

The test hangs at times and while there's a fix brewing, it's making
'make check' less useful.

Lets add it back when it works all the time.

Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

Previously, the variables image_file, cmd_args and opt_path were used
globally along the run.py script, I did not consider this convenient and
thus, opted to convert these global variables, into one variable that is
passed as a parameter.

This variables was renamed to options, with the mindset that, if
desired, this variable could come from a configuration file, thus making
the passing of command line arguments optional. So, if this
functionality for a configuration file... Or perhaps default values, be
added to the script, there would not be much need for refactoring (or
renaming) as there would with these global variables.

Reviewed-by: Tomasz Grabiec <tgrabiec@gmail.com>
Signed-off-by: Juan Antonio Osorio Robles <jaosorior@gmail.com>
Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

The '-nographic' option is deprecated and no longer imposes
the behavior which we want since this commit:

qemu.git 02c4bdf1d2ca8c02a9bae16398f260b5c08d08bf

We should use 'signal' option of the chardev instead, which
works with current qemu master as well as on 1.4.0

This patch renames '-g' option to '-s' as the former no longer
has adequate name.

Reported-by: Juan Antonio Osorio Robles <jaosorior@gmail.com>
Signed-off-by: Tomasz Grabiec <tgrabiec@cloudius-systems.com>
Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

After commit dc40b49e, the place of
the _cpu and _status fields of a thread has moved (into a structure
pointed by the _detached_thread field). So fix loader.py to use
these new locations, so "osv info threads" will work again.

Signed-off-by: Nadav Har'El <nyh@cloudius-systems.com>
Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

There's no reason to run a locale test as part of OSv boot sequence...

Avi writes:

  It used to be a major stumbling block but is of no interest now.

Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>