skb: reduce stack size to be able to run in QEMU

Signed-off-by: Roni Häcki <roni.haecki@inf.ethz.ch>

Make hwmodel work somewhat

Signed-off-by: Lukas Humbel <lukas.humbel@inf.ethz.ch>

skb: increae memory to 128M

Signed-off-by: Reto Achermann <reto.achermann@inf.ethz.ch>

skb: measure the execution of the queries

Signed-off-by: Reto Achermann <reto.achermann@inf.ethz.ch>

libc: removing newlib, adding BSD libc
posixcompat: moving gettimeofday to libc
harness: adding a simple math test
lwip, lwip2: moving includes to /include, compiling with BSD headers

Signed-off-by: Adam Turowski <adam.turowski@inf.ethz.ch>

skb: switching from dlmalloc to libc malloc

Signed-off-by: Adam Turowski <adam.turowski@inf.ethz.ch>

Adjustments of IPC to match new convention (static buffers).
(Flounder) Fixing a token passing in 32-bit architectures.

Signed-off-by: Adam Turowski <adam.turowski@inf.ethz.ch>

skb: Update copyright notices.

Signed-off-by: Moritz Hoffmann <moritz.hoffmann@inf.ethz.ch>

skb: improve debugging

Signed-off-by: Moritz Hoffmann <moritz.hoffmann@inf.ethz.ch>

posixcompat (dlfcn): improve C++ compatibility and document how to use.

Note: a minor API change was necessary to fix an incompatibility between C and
C++, which requires array sizes to be fixed:
void bar(struct foo (*f)[], int size);

This results in the following error:
error: parameter 'f' includes pointer to array of unknown bound 'foo []'
error: cannot convert 'foo (*)[3]' to 'foo (*)[]' for argument '1' to 'void bar(foo (*)[], int)'

Instead we simply pass the array as a pointer to its first element:
void bar(struct foo *f, int size);

SKB compiles for ARM. However ec_init fails with weird errors,
most likely due to a bug in the memory implementation for ARM.

last renaming changes

--HG--
branch : octopus.renaming

More renaming

--HG--
branch : octopus.renaming

Renaming header files.

--HG--
branch : octopus.renaming

Re-enable GC in SKB.

Removed some debug output in code and made small fixes.

Fixed some issues in code. Now ACPI/IOAPIC/PCI daemons seem to
work on real HW.

--HG--
branch : octopus.acpi-service

Remove pci_db loading.

Changed trigger code a bit to support persistent triggers
(will used for device manager).

--HG--
branch : octopus.add-device-mngr

Added some very basic functionality that adds devices and looks them up in the device manager.

--HG--
branch : octopus.add-device-mngr

Clean-up in skb initialization code. Removed some debug output.

--HG--
branch : octopus.chips-remove

This is currently work in progress. Started to remove dependencies from chips.
Now everything boots somewhat but for some reason event thread cannot be started
yet.

--HG--
branch : octopus.chips-remove

Added benchmark to measure throughput.

Added bitfield index for publish / subscribe + some changes with
ID registration on server.

Some changed eclipse settings for variable and add records.

Final configuration for benchmarking get/set API on server.

Worked on index for publish subscribe - objects3.pl now seems pretty much ok
for benchmarking.

Getset functioning again with triggers. Trying to get pubsub to work with attribute indexes.

Working on attribute index written in C using skip lists. This should be a lot faster than
the implementation in prolog using ordset.

Adjusted prolog implementation with constant access time (avoids copying).

some changes for benchmarks

some changes for benchmarking

New implementation for record storage in eclipse using hashtable.
Small adjustments for new implementation in code generation/test file.

Remove debug output, add cycle counter for eclipse engine only

Boots correctly for both freec and newlib

Trigger Support

Added trigger, can get rid of exists_not call, can make working barrier implementation.
Need to make sure everything is thread safe since we have a 2nd threads for events
now.

Refactoring

Current state does not compile! Using 2nd Thread for asynchronous events,
remove dist_event and move in dist2 interface. Even tough we still use
2 bindings this gets rid of some boiler plate code.

Worked on Locking, Barriers, Sequential Objects

Moved the Sequential set code into Prolog. I still have
problems loading the regex library, however split is needed
to read the sequence of existing objects so I added
the necessary predicated manually in the eclipse namespace.

Initial Barrier code written but untested. In addition I
modified the exists and exists_not handler to use the
new set_watch() code instead of their own custom versions.

Fixed ec_resume() handling for SKB Queries

There is one Query for the skb (the one in the pci code using [bridge_page]),
which has multiple results. In that case they must either be handled using
ec_resume() to iterate through them or cut using ec_cut_to_chp(), otherwise
subsequent ec_resume() calls provide wrong results. The current
fix just discards multiple results using ec_cut_to_chp().

trying to find eclipse ec_resume() bug (not returning pfail)

exist exist_not handler

dist2 refactoring done

--HG--
rename : lib/dist2/parser/ast.h => include/dist2/parser/ast.h
rename : lib/dist2/server/service.h => include/dist2_server/service.h

reworked mem management in skb server

--HG--
rename : usr/skb/skb_debug.h => usr/skb/include/skb_debug.h

basic publish/subscribe functionality

added dist2 library, test and parser in skb

Some initial prolog programs for MA

Merge from ASPLOS tree.

Initial version of public stable barrelfish repository.