It's preferable to use equal_range() to do VMA lookups so drop the
unused find_vma() function.

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

Replace magic numbers with constants for CR0 and CR4 control register
values in arch/x86/boot.S.

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

This prevented the script to work on tmpfs.

Signed-off-by: Benoit Canet <benoit@irqsave.net>
Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

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

As of commit 31fee0fa ("build: Build images as QCOW2 files.") we no
longer build raw images.

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

condvar_wait() expects an absolute time, not a duration.

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

The ACPI subsystem uses the memory allocator a lot so enable the debug
allocator before initializing it to detect boot-time problems.

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

Fill newly allocated memory with a well known pattern (byte, byte+1,
byte+2, ...) to catch uses of uninitialized memory.

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

We create a memory_image and assign it to _core before *this is fully
initialized, triggering corruption in program::_module_index_list.

This fixes memory corruption issues seen with the upcoming demand paging
patches.

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

Make use of the python nbd client.

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

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

This fixes 'osv info threads' command.

when $rsp holds 0xffffc0003fdf6fc8, ulong() for gdb.Value representing
it returns 18446673706036981704. This is out of int range. This
causes that thread_context.__exit__() throws "Numeric constant too
large" exception from

  set $rbp = 18446673706036981704

The fix removes the additional conversion as it seems to be
unnecessary. The gdb.Value should format itself properly.

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

Spanning tree causes the bridge port to start in listening mode.
That causes dhcp packet to be ignored. Since we're not fearing of
layer2 loops, we can safely remove the spanning tree config.

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

Add the explanation by Avi Kivity to AcpiOsUnmapMemory() to explain why
it's a no-op.

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

The script supports 3 modes of operation:
  - public release (default) for official releases
  - private release - for internal testing
  - development instance - for development

Public release means AMI created public (visible to everyone) and replicated
over all existing Amazon regions.

Private release means AMI created private (visible to releaser's account only) and
not replicated over regions.

Development instance means no AMI created, script stops after instance creation.

The script's help screen (--help command line parameter) provides detailed information
about script interface, dependencies and configurations required.

It's interesting to note that each public release made by this script costs about 20 USD
mainly because of AMI replication over Amazon regions

Signed-off-by: Dmitry Fleytman <dmitry@daynix.com>
Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

Signed-off-by: Dmitry Fleytman <dmitry@daynix.com>
Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

Due to incorrect mapping logic ACPI tables longer than PAGE_SIZE
mapped partially and to wrong base address. As a result system crashed
on ACPI tables checksum verification.

Found when running on EC2 HVM instances, the problematic table was:
ACPI: DSDT 0xfc002c40 0321F (v02    Xen      HVM 00000000 INTL 20090220)

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

slop made page size by default because this is the most frequent case

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

If we do, we'll have extra xen threads running around for no reason in the !xen
case. is_xen() should already work here since the initializers are run after
cpuid discovery, which is when the information for is_xen is filled up.

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

Fix the following breakage caused by commit c98b928e ("Explicitly use python
version number") when starting GDB:

  Traceback (most recent call last):
    File "/home/penberg/osv/scripts/loader.py", line 833, in <module>
      setup_libstdcxx()
    File "/home/penberg/osv/scripts/loader.py", line 496, in setup_libstdcxx

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

Ironically, my commit 0c77d3c2 broke xen_features initialization.
This patch fixes it.

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

Consolidates generated files, and makes it easier to maintain the eclipse
index.

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

Currently all log messages are silently ignored,
which is probably not what we want.

The default JDK config logs on console.

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

I dismissed the idea to configure the bridge in default JVM
configuration because having a default like this would force
all apps to use particular slf4j-api version.

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

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

Makes this go away:

  Note: io/osv/OsvSystemClassLoader.java uses unchecked or unsafe
  operations.  Note: Recompile with -Xlint:unchecked for details.

Unfortunately javac does not recognize statement level suppressions
like this:

  //noinspection unchecked
  return (T) method.invoke(target, args);

I need to switch to more coarse grained suppression on method level:

  @SuppressWarnings("unchecked")

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

Switch to Markdown syntax in README. I did the minimal amount of changes
to make the rendering on Github look sane. We can prettify it even more
later.

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

It is needed for systems where python points to python3

Signed-off-by: Anatol Pomozov <anatol.pomozov@gmail.com>
Signed-off-by: Pekka Enberg <penberg@cloudius-systems.com>

The code wouldn't work if XENFEAT_NR_SUBMAPS > 1. It's currently 1.
The assignment to xen_features must take i as well as j into consideration.

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

Use the initialization sequence specified in 10.1.2.2 ("ACPICA
Initialization With Early ACPI Table Access") from ACPICA reference
manual.

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

Use mmu:isreadable() to actually check if a memory region is readable in
AcpiOsReadable().

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

Fix the following issues with the debug version of realloc():

  - Make realloc with NULL ptr behave like malloc()

  - Make realloc with non-NULL pointer and zero size behave like free()

  - Avoid malloc() if we have enough space in original memory block

  - Return NULL if the new allocation fails but don't deallocate the
    original memory block

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

Switch to console::write_ll() in AcpiOsPrintf() and AcpiOsvPrintf() to
make ACPI boot-time output readable:

  [penberg@localhost osv]$ ./scripts/run.py
  ACPI: RSDP 0xfd8d0 00014 (v00 BOCHS )
  ACPI: RSDT 0x3fffe380 00034 (v01 BOCHS  BXPCRSDT 00000001 BXPC 00000001)
  ACPI: FACP 0x3fffff80 00074 (v01 BOCHS  BXPCFACP 00000001 BXPC 00000001)
  ACPI: DSDT 0x3fffe3c0 011A9 (v01   BXPC   BXDSDT 00000001 INTL 20100528)
  ACPI: FACS 0x3fffff40 00040
  ACPI: SSDT 0x3ffff6e0 00858 (v01 BOCHS  BXPCSSDT 00000001 BXPC 00000001)
  ACPI: APIC 0x3ffff5b0 00090 (v01 BOCHS  BXPCAPIC 00000001 BXPC 00000001)
  ACPI: HPET 0x3ffff570 00038 (v01 BOCHS  BXPCHPET 00000001 BXPC 00000001)
  acpi 0 apic 0
  acpi 1 apic 1
  acpi 2 apic 2
  acpi 3 apic 3
  APIC base fee00000
  OSv v0.02-64-ga556eab Copyright 2013 Cloudius Systems
  locale works
  ACPI: RSDP 0xfd8d0 00014 (v00 BOCHS )
  ACPI: RSDT 0x3fffe380 00034 (v01 BOCHS  BXPCRSDT 00000001 BXPC 00000001)
  ACPI: FACP 0x3fffff80 00074 (v01 BOCHS  BXPCFACP 00000001 BXPC 00000001)
  ACPI: DSDT 0x3fffe3c0 011A9 (v01   BXPC   BXDSDT 00000001 INTL 20100528)
  ACPI: FACS 0x3fffff40 00040
  ACPI: SSDT 0x3ffff6e0 00858 (v01 BOCHS  BXPCSSDT 00000001 BXPC 00000001)
  ACPI: APIC 0x3ffff5b0 00090 (v01 BOCHS  BXPCAPIC 00000001 BXPC 00000001)
  ACPI: HPET 0x3ffff570 00038 (v01 BOCHS  BXPCHPET 00000001 BXPC 00000001)
  ACPI: All ACPI Tables successfully acquired

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

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

This patch simplifies the host-side work in the new /usr zfs filesystem
build process.

Previously, we copied the files to a temporary directory, used "cpio"
to archive them and sent its output to the guest with "netcat".

With this patch, we no longer have a temporary directory, and do not
need either cpio or netcat on the build machine.

Rather, mkzfs.py itself, using python (instead of a separate "nc" process),
connects to the guest and sends it the files - still using the CPIO format.

Rather than arbitrarily sleep for 3 seconds before the host tries to
connect to the guest (which might not be enough for some, or a waste
of time for others), with this patch the host looks at the guest's output
and connects when it sees the message "Waiting for connection".

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

We want to return 0, not ETIMEDWAIT, if a wake_one/all() was consumed.
In one racy case, we returned the wrong one (we received a wakeup, but
still returned ETIMEDWAIT).

This probably doesn't matter at all in practice...

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

Add Doxygen comments to the condvar class. Only the C++ interface
(condvar's methods) is documented, not the alternative C interface
(condvar_* functions).

A reminder: run "doxygen" and point your browser to doxyout/html/index.html
to see the API documentation we have so far.

A lot can still be added to this condvar documentation, including a
good introduction to how to use condition variables, why they have
a mutex, etc. But it's at least a start.

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

The tst-zfs-simple.so test case has serverd its purpose for bringup.  As
OSv build relies on working ZFS now, there's no need to run the tests.

Furthermore, we have the full ztest stress test in the tree:

  bsd/cddl/contrib/opensolaris/cmd/ztest/ztest.c

which we can use if needed.

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

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

The TLS segment is a little wierd in that it grows backwards from percpu_base
instead of forwards.  This causes the alignment code to calculate wrong
offsets when the segment size is 8 (mod 16).  A failure was seen where
::percpu_base was set at offset 0xfffffffffffffa08 in code that was in the
same translation unit as ::percpu_base, and 0xfffffffffffffa10 elsewhere.
This caused all dynamic_percpu instances to crash.

Fix by aligning the segment size.  For good measure, align also the segment
base, both to a cacheline boundary.

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