Speeds up mgmt null build considerably

Add a comment to condvar explaining that it makes a guarantee that
POSIX Threads' condition variables do not - that there are no spurious
wakeups.

The comment goes on to explain that at least in one place (semaphore.cc)
we make use of this added guarantee, so if the condition variable
implementation is ever rewritten, we'll need to keep this guarantee.

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

Add "-version" option to RunJava, and therefore to java.so and the "java"
CLI command.

java -version now shows:

  java version "1.7.0_25"
  OpenJDK Runtime Environment (1.7.0_25-mockbuild_2013_07_27_13_36-b00)
  OpenJDK 64-Bit Server VM (build 23.7-b01, mixed mode)

I can't explain why the version on the second line is different than what
"java -version" on Fedora 18 shows for the same libjvm.so:

  java version "1.7.0_25"
  OpenJDK Runtime Environment (fedora-2.3.10.4.fc18-x86_64)
  OpenJDK 64-Bit Server VM (build 23.7-b01, mixed mode)

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

  1. MSI-X printous fixed
  2. Duplicate EOLs removed

Signed-off-by: Dmitry Fleytman <dmitry@daynix.com>

Change the name of the argument of wake_with from "Pred" to "Action".
This argument is a function to run, *not* a predicate (it's not supposed
to return a boolean value), so it doesn't make sense to call it Pred.

The implementation of wake_with already used the name "Action" - this
patch fixes the prototype too.

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

  1. Netperf patch removed from repository
  2. Documentation updated

Signed-off-by: Dmitry Fleytman <dmitry@daynix.com>
Signed-off-by: Nadav Har'El <nyh@cloudius-systems.com>

This patch adds support for Linux's alarm(2) system call.

alarm() is needed in some ancient Unix software, like netperf :(

This implementation is fully compatible with Linux's alarm(), but please
note that what alarm() does when the alarm times out is a kill(SIGALRM) -
so all the caveats mentioned in the previous patch regarding kill(),
will also apply here.

Alarm() is implemented by running one "alarm thread", started on the
first alarm() call. This alarm-thread waits for the alarm to expire,
or for instructions to change the alarm. When an alarm expires the
alarm thread does a kill(0,SIGALRM).

tst-kill.cc (from the previous patch) includes a test to see that
alarm() really sends a SIGALRM signal on time, and also that we can
cancel a pending alarm and not receive a signal.

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

Some fixups to --trace-backtrace, plus symbol resolution in the log output.

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

This patch adds support for the Linux kill(2) function.
The next patch will add alarm(2) support, which uses kill(2).

To be honest, we sort-of implement kill(). This implementation is
compatible with the API, but the semantics are somewhat different:
While in Linux kill() causes the signal handler to run on one of the
existing threads, in this implementation, the signal handler is run in a
*new* thread.

Implementing the exact Linux semantics in OSv would require tracking when
OSv runs kernel code (i.e., code in the main executable, not a shared
object) so we can delay running the signal handler until returning to the
user code. Moreover, we'll need to be able to interrupt sleeping kernel
code. This is complicated and adds overhead even if signals aren't used
(and they aren't used in most modern code).

I expect that this code will be "good enough" in many use cases.
This code will *not* be good in enough in programs that expect one of the
following:

1. A program that by using Posix Thread's "signal masks" tried to ensure
   that the signal is delivered to one specific thread, and not to an
   arbitrary thread.

2. A program that used kill() or alarm() not intending to run a signal
   handler, but rather intending to interrupt a sleeping system call
   like sleep() or read(). Our kill() does not interrupt sleeping OSv
   function calls, which will continue to sleep on the thread they run
   on.

The support in this patch (and see next patch, for alarm()) is good
enough for netperf's use of alarm().

P.S. kill() can be used only to send a signal to the current process, the
only process we have in OSv (you can also use pid=0 and pid=-1 to achieve
the same results).

This patch also adds a test for kill() and alarm(). The alarm() test
will fail until the next patch :-)

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

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

Looking up symbols is expensive, store them in a cache.

Speeds up dumping the trace log when a backtrace is taken.

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

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

Also fix other review comments related to 1f161695.

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

The correct range is elf_start:(elf_start + elf_size - 1)

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

task_conv was wrongly converting "/.." and produced a path with unexpected
characters after /
Consequently, repeated calls (Java) to 'new File("/..").exists()' returned
ambigous results, not to mention other properties like isDirectory.

This confused the shell's completion mechanism (and me).

In more detail, in this (/..) situation 'len' was more than 2 (in line 105) which
caused 'tgt' to be pointing before the begining of 'full', from that point
whatever stopped the backward search for '/' (lines 109-112) was unexpected.
I guess this might occur with other combinations.

This sets 'len' to be 1 when starting an absolute path conversion, to match
the behavior where relative path conversion sets 'len = strlen(cwd)'.

Signed-off-by: Or Cohen <orc@fewbytes.com>

Explain in README that Gcc 4.7 is required, gcc 4.8 is recommended.

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

Check that link returns EEXIST if destination exists.

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

Check that link returns ENOENT if source does not exist.

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

Make the link source and target paths configurable so the test can be
run on Linux:

  $ g++ -std=c++11 tst-fs-link.cc && ./a.out foo foo2

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

We can safely use the same function exit points for both successful and
error cases.

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

Call the dentry that refers to the newpath parent directory "newdirdp"
to make the code more explicit.

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

Filesystems are expected to wire up ->vop_link so it can never be NULL.

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

Add a new test to see if the content from one link would be the same in
another.

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

Add a basic test case for filesystem hard link support. You can run it as
follows:

  ./scripts/run.py -e "tests/tst-fs-link.so"

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

Make vn_stat() return inode number and link count in st_ino and
st_nlink, respectively.

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

This patch adds the link() system call.

The system call uses the filesystem specific VOP_LINK vnode operation to
do the actual work.  ZFS is currently the only available file system
that supports hard links and the other filesystems return EPERM.

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

Update ->va_nlink() in zfs_getattr() in preparation for sys_link().

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

Wire up the VOP_LINK vnode operation for ZFS in preparation for
sys_link().

Signed-off-by: Raphael S. Carvalho <raphael.scarv@gmail.com>
[ penberg: drop FIGNORE, cleanup, split ]
Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

Wire up the VOP_LINK vnode operation for ramfs in preparation for
sys_link().

Signed-off-by: Raphael S. Carvalho <raphael.scarv@gmail.com>
[ penberg: split to separate commit ]
Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

Wire up the VOP_LINK vnode operation for devfs in preparation for
sys_link().

Signed-off-by: Raphael S. Carvalho <raphael.scarv@gmail.com>
[ penberg: split to separate commit ]
Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

Add VOP_LINK to the VFS interface in preparation for sys_link().

Signed-off-by: Raphael S. Carvalho <raphael.scarv@gmail.com>
[ penberg: split to separate commit ]
Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

Use ->rn_namelen as the destination buffer size for strlcpy().

Spotted by Coverity.

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

The ramfs_remove_node() will free vp->v_data so we cannot look it up in
ramfs_remove().

Spotted by Coverity.

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

signal() verified that its argument is not >nsigs, but it should be >=nsigs
because nsigs itself is already illegal (0...nsigs-1 are valid signals).

Discovered by Coverity.

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

GCC implements static destructors in shared-objects (DSO) by registering a
single finalization function, __cxa_finalize(), to run when the DSO is
unloaded, and each constructor calls __cxa_atexit() to register a destructor
which __cxa_finalize() should call.

This patch implements the missing __cxa_finalize() and __cxa_atexit()
functions. The do-nothing stubs we had never registered, and never ran,
destructors. The implementation in this patch is amazingly simple to
write in C++11 :-)

Before this fix, we had a crash when running tst-tracepoint.so twice:
The first run ran the tracepoints' constructors, which added to the global
linked list of tracepoints pointers into the mapped DSO. When the DSO
was later unmapped but destructors not run, the linked list of tracepoints
was left with pointers to unmapped memory. When we ran the test again,
and it wanted to construct more tracepoints, it accessed the broken linked
list and crashed.

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

ELF allows specifying initializers - functions to be run after loading a
a shared object, in DT_INIT_ARRAY, and also finalizers - functions to be
run before unloading a shared objects, in DT_FINI_ARRAY. The existing code
ran the initializers, but forgot to run the finalizers, and this patch
fixes this oversight.

This fix is necessary for destructors of static objects defined in the
shared object. But this fix is not sufficient for C++ destructors - see
also the next patch.

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

Add to libc/user.cc stub setresgid(), setresgid(), which as usual in
OSv only allow user and group 0. These functions are used by Rogue ;-)

Also move getuid() and friends from runtime.cc to libc/user.cc, where
they feel more at home.

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

The "cscope" make target piped a list of files into "cscope -bq", but
this command doesn't read its input. The command should be "cscope -bq -i-"
which reads the list of files from stdin.

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

Signed-off-by: Dmitry Fleytman <dmitry@daynix.com>

Coverity thinks we have a use-before-assigned of r2 here, despite the
join(), so let's just move the test inside the thread to make this
warning go away.

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