We have _KERNEL defines scattered throughout the code, which makes
understanding it difficult.

Define it just once, and adjust the source to build.

We define it in an overridable variable, so that non-kernel imported code
can undo it.

Imported from FreeBSD 245655, no changes.

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

Import FreeBSD files, changeset 245655.

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

Imported with no change from FreeBSD 245655.

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

Imported with no change from FreeBSD 245655.

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

vfs mount/unmount related functionality.

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

Iterate over all inodes, droping directory entries and vnodes belonging
to the mount point.

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

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

In order to unmount, we need to drop all vnodes on a mount point, so we
need to find all directory entries pointing to them.

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

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

Similar to /etc/mnttab, but without going through the file system.

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

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

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

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

nmount() is a mount variant that works with key/value pairs to provide
options.  Implement it for osv.

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

zfs interprets the device parameter as a pool name, so don't insist on opening
it as a device.

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

Partition table fixes.

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

The partition offset can be > 32 bits, so allow for it in the calculation.

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

head=254 is legal, and the check also assumes that heads=255, which is not
a given.  Remove the check and trust the LBA.

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

Partitions table entries with cyl=0 or head=0 are legal, avoid false
positives.

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

Some vfs fixes needed for mkfs; already posted as part of the mkfs patchset.

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

Needed by libzfs.

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

open()'s prototype is traditionally open(const char*, int, ...) since
the third mode_t parameter is optional, and only needed if O_CREAT is
specified.

Change the definition to conform.

This allows the LFS64() macro to be used to define open64() with a
similar prototype.

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

Make vget() accept a const char*, and adjust various other functions that
it calls in the same way.  This makes it easier to convert vfs_mount.c to
C++.

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

For zfs.

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

We store the flags in a different format than fcntl() expects; perform the
needed conversions.

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

The vfs converts the read/write flags from 0/1/2 (read-only/write-only/rw) to
bit fields which are more easily testable.  But in doing so it corrupts the
other flags.

Fix FFLAGS and OFLAGS to preserve those other flags.  The test for a file
that is not readable or writable is dropped, since it is impossible to fail.

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

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

Switch to size_t type for ->bio_bcount in struct bio and clean up
virtio-blk signedness casts.

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

command line option -b/--bridge added

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

Some trivial changes to tst-loadbalance.so. In particular, this benchmark
assumes it is run on two cpus, so don't do anything if that isn't true.

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

We had a different Linux version compiled into uname() (3.7.0) than
we had compiled into other code (via <linux/version.h>).

This patch makes them both pretend to be 3.7.0 - arbitrarily chosen
as the current Linux version at the time OSv was created. The new code
verifies with a static assertion that both files contain the same version,
so if they diverge, uname.c will fail compilation.

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

This patch adds a "osv runqueue" command to gdb, which shows the (sorted)
run queue of each CPU. Note that the run queue does not include the thread
currently running - just those which are ready to run.

This command is useful for debugging the scheduler and load balancer, but
is less generally useful than the "osv info threads" command, which lists
all threads, specially marking those which are runnable as well as those
which are running.

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

This patch begins to document OSv's native (non-Linux-compatible) C++ APIs
using Doxygen. With doxygen, functions, classes, and other APIs are
documented using comments inside the source code, similarly to the way
Javadoc is used in the Java world.

After this patch, running "doxygen" generates the documentation in HTML
form (and also ugly ones in LaTeX...) in the "doxyout" directory.
After running "doxygen" in the OSv top-level directory, open your browser to
file://.../doxyout/html/index.html

 to view this documentation.

This patch adds a doxygen configuration file, "Doxyfile", and documents
three functions in the osv namespace: osv::halt(), osv::poweroff(),
osv::halt().

The format of the resulting documentation is not perfect, but it's not
as bad as I feared, and also the documentation doesn't uglify the code
too much (it mainly impacts the header files), so I now believe it will
better for us than writing man pages (an avenue which I explored
previously).

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

This patch adds interrupt (by default ctrl-C) and quit (by default ctrl-\)
handling to the console driver, when the ISIG termios flag is enabled (and
this patch enables this flag by default). The application can change these
characters or disable ISIG with ioctls, exactly as in Unix.

The ^C character sends a SIGINT to the (only) process, and ^\ sends SIGQUIT.
By default both abort OSv, but when running Java we get a much more
interesting SIGQUIT handler - it shows a trace of all running threads
and other information.

Note that our line-discipline code currently sits in the console driver
(we do not have ptys yet), so it only applies to the console, not to
SSH connections or the likes.

Also note that by default (before the patch to issue #49), qemu will catch
the ^C and ^\ and won't even pass them to OSv, so apply the patch to #49
(or change the interrupt or quit characters) to see this feature working.

This patch relates to issue #53.

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

Fix memory_page_alloc tracepoint if untracked_alloc_page() uses the
early page allocator.

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

This patch adds page allocator support to memory analyzer post
processing script.

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

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

The tst-mmap.so test case noticed the following breakage in msync():

  Running mmap tests
  Assertion failed: msync(buf, 4096*9, MS_ASYNC) == 0
  (../../tests/tst-mmap.cc: main: 207)
  Aborted

The problem is that msync() uses the contains function which checks if
[start, end) is in a vma range but we really are interested in whether
the vma intersects with [start, end).

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

This Close issue #8 ""Make Java InetAddress.getHostName() works".

Signed-off-by: Benoit Canet <benoit@irqsave.net>
Signed-off-by: Avi Kivity <avi@cloudius-systems.com>