Newer
Older
safe since the new features are only enabled by this option (Safer
but not 100% safe since it does not protect from typos and other
dumb errors) (2014-8-6).
* arch/sim/src/Makefile, up_idle.c, up_internal.h, and up_tickless.c:
Add support for a simulated interval timer support verification of
the tickless OS (2014-8-6).
* include/nuttx/arch.h: Added another tickless OS interface to get
the timer remaining on the interval timer (2014-8-6).
* include/nuttx/clock.h, sched/clock_initialize.c and clock_internal.h:
If CONFIG_SCHED_TICKLESS is defined, then the global variable
g_system_timer does not exist (2014-8-6).
* sched/Makefile: Don't build in sched_processtimer.c if
CONFIG_SCHED_TICKLESS is selected (2014-8-6).
* sched/clock_gettime.c and clock_settime.c: Use the clock_systimer()
macro, do access the g_system_timer() global directly (2014-8-6).
* sched/clock_systimer.c: Uses interval timer interfaces to get the
time if CONFIG_SCHED_TICKLESS is selected (2014-8-6).
* sched/sched_timerexpiration.c, Makefile, os_internal.h,
sched_addreadytorun.c, sched_processtimer.c, sched_unlock.c,
wd_cancel.c, wd_internal.h, and wd_start.c: Implements the
tickless OS (2014-8-7).
* Many files: Remove CONFIG_DISABLE_CLOCK. Why? This option is not
really very useful and was only selected in a few unusable, mostly
unusable configurations anyway (m9s12/ostest, kl25z/minnsh,
z180/ostest, 8051, sh1, z80, and z8). The only real loss is probably
to the kl25z/minsh configuration . And with the addition of the
tickless OS, I just wanted to make the pain go away (2014-8-7).
* Many files: Change CONFIG_MSEC_PER_TICK to CONFIG_USEC_PER_TICK.
This gives more options for system timers in general, but more
importantly, let's us realize higher resolution for the case of
CONFIG_SCHED_TICKLESS=y -- of course, at the risk of some new
integer overflow problems 2014-8-7).
* Many files: Change all time conversions. Yech. New timer units
in microseconds breaks all existing logic that used milliseconds
in the conversions. Something likely got broken doing this,
probably because I confused a MSEC2TICK conversion with a TICK2MSEC
conversion (2014-8-7).
* arch/sim/src/up_tickless.c and sched/sched_timerexpiration.c: The
tickless OS now appears fully functional and passes the OS test
on the simulator with no errors (2014-8-7).
* sched/mqueue: Move POSIX message queue files from sched/ to
sched/mqueue (2014-8-8).
* sched/signal: Move signal-related files from sched/ to sched/signal
(2014-8-8).
* sched/pthread: Move all pthread files from sched/ to sched/pthread
(2014-8-8).
* sched/semaphore: Move POSIX counting semaphore files from sched/ to
sched/semaphore (2014-8-8).
* sched/environ: Move environment files from sched/ to sched/environ
(2014-8-8).
* sched/group: Move group logic from sched/ to sched/group (2014-8-8).
* sched/wdog: Move watchdog functions from sched/ to sched/wdog
(2014-8-8).
* sched/irq: Move interrupt dispatch logic from sched/ to sched/irq
(2014-8-8).
* sched/clock: Move clock functions from sched/ to sched/clock
(2014-8-8).
* sched/timer: Move POSIX timer files from sched/ to sched/timer
(2014-8-8).
* sched/paging: Move page fill sources from sched/ to sched/paging
(2014-8-8).
* sched/init: Move initialization functions from sched/ to sched/init
(2014-8-8).
* sched/errno: Move errno related files from sched/ to sched/errno
(2014-8-8).
* sched/task: Move task control files from sched/ to sched/task
(2014-8-8).
* sched/: Move a few files in sched/ that did not seem to have a home to
sched/signal and sched/task (2014-8-8).
* sched/sched: Move scheduler files from sched/ to sched/sched
(2014-8-8).
* (Many files): Replace os_internal.h with sched/sched.h in files that
actually reference something in sched.h (2014-8-8).
* sched/task/exit.c: Move sched/exit.c to sched/task/exit.c (2014-8-8).
* (Many files): Remove os_internal.h it has been replace by several new
header files under sched/. There have been some sneak inclusion paths
via os_internal.h, so expect a few compilation errors for some
architectures (2014-8-8).
* Documentation/NuttxPortingGuide.html, arch/sim/src/up_tickless.c, and
include/nuttx/arch.h: Fix errors in documentation and comments related
to the Tickless OS. From Vijay Kumar (2014-9-9).
* arch/arm/src/sama5/sam_tc.c and .h: Can now handle non-constant
BOARD_MCK_FREQUENCY. Also now supports methods to attach user
interrupt handlers (2014-8-9).
* drivers/audio/audio_null.c: Fix an error in a variable name
(2014-8-9).
* arch/arm/src/sama5/sam_oneshot.c, .h, Kconfig, Make.defs, sam_tc.c,
.h, and sam_adc.c: Add support for a one-shot timer wrapper around
the low-level timer/counter logic. This also involved several changes
that rippled into the ADC driver (untested) (2014-8-9).
* arch/arm/src/sama5/sam_freerun.c, .h, Kconfig, and Make.defs: Add
support for a free-running timer wrapper around the low-level
timer/counter logic (2014-8-9).
* arch/arm/Kconfig, src/sama5/sam_tickless.c, Kconfig, Make.defs: Use
the one-shot and free-running timers to implement tickless OS support
for SAMA5 (2014-8-9).
* arch/arm/src/sama5/sam_oneshot.c and sam_tc.c: Fix several bugs in
timer/counter interrupt logic and one-shot timer logic. Comments and
debug output updated in additional files (2014-8-10).
* sched/wdog/wd_start.c: Correct a bug in Tickless OS support: logic to
detect changes in head of timer list was wrong (2014-5-10).
* sched/sched/sched_timerexperation.c, sched.h, and wdog/wd_start.c: In
tickless mode, need to stop the interval timer before inserted a new
delay into the timer list. Otherwise, the time is incorrect on the
first entry of the list (2014-8-11).
* arch/arm/src/sama5/sam_freerun.c and sam_oneshot.c: Correct bad time
conversion, 1000000 not 1000 to convert seconds to microseconds
(2014-8-11).
* sched/sched/sched_timerexpiration.c: Fix inaccurate time conversion.
Remove MSEC_PER_TICK from calculation and convert uint32_t to uin64_t
(2014-8-11).
* libc/time/lib_localtime.c, private.h, tzfile.h and several header and
make-related files: Adds support for localtime. From Max Neklyudov
(2014-8-12)
7.5 2014-xx-xx Gregory Nutt <gnutt@nuttx.org>
* clock/clock_systimespec.c and sched/clock/clock_gettime.c: When
reading the system timer, don't read a struct timespec, convert it
to a fake tick count, then back to a timespec (2014-8-15).
* sched/clock/Make.defs, clock_gettime.c, clock_settime.c, include/time.h,
nuttx/clock.h, configs/vsn/src/sif.c: Remove CLOCK_ACTIVETIME. It is non-
standard, unused, and gets in the way of maintaining clocks (2014-8-15).
* libc/time/lib_localtime.c: Modified to that the coding style is much
closer to NuttX coding style. The local header files private.h and
tzfile.h were removed and incorporated into lib_localtime.c. All
conditional compilation within lib_localtime.c that unconditionally
evaluated to FALSE was removed (2014-8-15).
* drivers/net/Kconfig and include/nuttx/arch.h: Standardize a PHY
interrupt attachment interface (2014-8-16).
* configs/Kconfig, sama5d3-xplained, sama5d3x-ek, and sama5d4-ek:
Convert existing board specific PHY interrupt interfaces to use newly
defined standard interface (2014-8-16).
* include/net/slip.h: Move SLIP prototypes from net.h to slip.h
(2014-8-16).
* include/net/net.h and net/: Fix some missing function headers
(2014-8-16).
* drivers/net/phy_notify.c, include/nuttxarch.h, net/phy.h, and
ioctl.h: Add support for an ioctl that can be used to notify an
application when there is a change in the network status signalled
by a PHY interrupt (2014-8-16).
* drivers/net/phy_notify.c, include/net/if.h, ioctl.h, netdev.h, phy.h,
and net/netdev/netdev_ioctl.c: Finishes the ioctl definition to
subscribe to PHY events. Revamp network ioctl signature to support
arguments other than struct mii_ioctl_data (2014-8-16).
* arch/arm/src/stm32/stm32_eth.c: Modified to support the change to the
network ioctl signature changes. Also add support for new ioctl to
setup PHY event notifications (2014-8-16).
* arch/arm/src/sama5/sam_emaca.c, sam_emacb.c, and sam_gmac.c: Implement
all network ioctls, including the new ioctl to setup PHY event
notifications (2014-8-16).
* include/net/if.h: Add some helper macros to make the code that uses
the ioctls a little more compact (2014-8-17).
* libc/stdio/lib_sscanf.c: NuttX libc tried to guess how many characters
to parse, extracted them into a buffer, then ran strtol() on that
buffer. That guess is often wrong. A better approach would be to
call strtol() directly on the input data, using the endptr return
value to determine how many characters to skip after parsing. From
Kosma Moczek (2014-8-18).
* net/arp/arp_dump.c: Add missing configuration option to select
dumping of ARP packet headers. Move ARP dumping logic from
arp_inout.c to its own file (2014-8-18).
* net/arp/arp_out.c: Move arp_out() from arp_inout.c to its own file
(2014-8-18).
* net/arp/arp_arpin.c: Move arp_arpin() from arp_inout.c to its own file
(2014-8-18).
* net/arp/arp_ipin.c: Rename arp_inout.c to arpipin.c because that is
all that is left in the file (2014-8-18).
* net/arp/arp_format.c: Move the logic that formats an ARP packet into a
separate file where it can be re-used. (2014-8-18).
* net/arp/arp_out.c and net/pkt/pkt_send.c: Change how the IFF_NOARP
flag is handled. This should be set only when data is moved into the
buffer and cleared after tested by the ARP logic. Setting it globally
can cause packets to be sent with no valid MAC addresses (2014-8-18).
* net/arp/arp_send.c: Partial implementation of logic to send ARP
requests to assure that an IP address mapping is present in the ARP
table (2014-8-18).
* net/arp/arp_poll.c: Finished the implementation of logic to send ARP
requests. Complete, fully hooked into the networking logic, and ready
for test (but still untested) (2014-8-18).
* net/icmp/icmp_ping.c, net/socket/net_sendfile.c, sendto.c,
net/tcp/tcp_send_buffered.c, and tcp_send_unbuffered.c: If
CONFIG_NET_ARP_SEND is enabled, then all ICMP, TCP, and UDP send
operations will call arp_send() before attempting the real send
operation. arp_send() will check if the the IP address mapping is in
the ARP table and, if not send ARP requests periodically until it is.
This eliminates losing the first outgoing message because there is not
mapping in the ARP table (2014-8-18).
* net/arp/arp_notify.c, arp_send.c, arp_arpin.c, and arp.h: Add
signalling logic so that we do not have to wait so long with the
network responses to ARP requests quickly (2014-8-19).
* include/nuttx/lcd/ili9341.h: \Add a register definition header file
for the ILI9341 LCD (2014-8-19).
* net/arp/Kconfig: ARP request logic is no longer EXPERIMENTAL
(2014-8-20).
* net/arp/arp_send.c: ARP request logic needs to do the right thing if
(1) the address is a broadcast or multicast address, and (2) if the
IP address lies outside of the network. Problem noted by Manuel Stuehn
(2014-8-20).
* configs/sam4e-ek/src/Kconfig and src/sam_ili9341.c: Add ILI9341-based
* configs/sam4e-ek/nxwm: Add an NxWM configuration for the SAM4E-EK
(2014-8-21).
* configs/viewtool-stm32f107: Refresh some configurations (2014-8-20).
* nuttx/sched/wdog: If we run out of pre-allocated watchdog times, the
logic will allocate additional timers from the heap. A reserve of
pre-allocated watchdog timers is maintained so that there will always
be timers available for interrupt handlers (2014-8-21).
* include/nuttx/wdog.h (and many affected files): wdog.h does not
contain any application interface, only internal OS interface.
Further, it is non-standard. Move wdog.h from include/ to
include/nuttx (2014-8-21).
* Documentation/NuttxPortingGuide.h: For the same reason as above, move
the description of the watchdog timer interfaces from the Users Guide
to the Porting Guide (2014-8-21).
* arch/arm/src/stm32/stm32_uart.h: STM32 F401: Correct support for
USART6 on this chip. From Freddie Chopin (2014-8-22).
* nuttx/sched/wdog and include/nuttx/wdog.h: Add support for statically
allocated watchdog timer (also eliminate some unconvential typing)
(2014-8-22).
* configs/p112/ostest and tools/mkdeps.c: Changes to try to get P112 to
compile with laster SDCC (it still does not) (2014-8-22).
* sched/group/group_leave.c: Need to release the address environment when
the task group is released (2014-8-22).
* mm/mm_granreserve.c and and mm_granmark.c: gran_reserve(): Add a
new function to reserve unallocatable regions in the granule heap
(2014-8-23).
* include/nuttx/pgalloc.h and mm/mm_pgalloc.c: Add a simple page
allocator based on the existing NuttX granule allocator. I am not
certain if the granule allocator is sufficiently deterministic for
long range use, but it gets get a page allocator in place for testing
very quickly (2014-8-23).
* ARMv7-A: Add skeleton implementation and build support for process
address environments (2014-8-23).
* Many files: Change CONFIG_ADDRENV to CONFIG_ARCH_ADDRENV; change how
it is selected -- the architecture must first declare support
(2014-8-24).
* include/nuttx/addrenv.h, arch/arm/include/arch.h, armv7-a/irq.h, and
other files: Add addrenv.h; First cut at Cortex-A address environment
structures; Add configuration options to setup address environment
(2014-8-24).
* addrenv interface changes: up_addrenv_create() may need to create
.text and .bss/.data separately because of differing access privileges
(read/execute vs read/write). And, as a consequence, up_addrenv_vaddr()
needs to be split into up_addrenv_vtext() and up_addrenv_vdata().
Affects several files (2014-8-24).
* arch/arm/src/armv7-a/arch_coherent_cache.c: New file. These cache
operations are called from the ELF loader in order to flush D-cache
and invalidate I-cache after an ELF module has been loaded into memory.
With this change, ELF modules work correcttly on the SAMA5/Cortex-A
platform (2014-8-24).
* configs/sim: Add build support for ELF modules. Useless at the moment
because there is no x86 relocation logic support (2014-8-25).
* configs/sim: Update some sim/ defconfig files. This now automatically
switches them to x86_64 builds (2014-8-25).
* arch/arm/include/arch.h, armv7-a/irq.h, src/armv7-a/arm_addrenv.c,
armv7-a/arm_mmu.c/.h, sama5/Kconfig, Make.defs, sam_pgalloc.c,
include/nuttx/addrenv.h, arch.h, pgalloc.h, mm/Kconfig, mm_pgalloc.c,
sched/init/os_start.c: Cortex-A/SAMA5 address environment support is
code complete (untested) (2014-8-25).
* arch/arm/include/tiva and src/tive: Add support for the TI CC3200.
From Jim Ewing (2014-8-26).
* configs/cc3200-launchpad: Add support for the TI CC3200 Launchpad.
From Jim Ewing (2014-8-26).
* arch/*/*_blocktask.c, arch/*/*_unblocktask.c, and arch/*/*_exit.c:
Add address environment support to all implementation of up_block_task(),
up_unblock_task(), and _exit() (2014-8-31).
* arch/arm/src/stm32/stm32_flash.c and chip/stm32_flash.h,
include/nuttx/progmem.h, and configs/vsn/src/sif.c: STM32 FLASH fixes:
use size_t instead of uint16_t, make interface more generic. From
Freddie Chopin (2014-8-28).
* Kconfig, Makefile.unix, Makefile.win, syscall/Kconfig,
arch/arm/src/armv6-m/svcall.h, up_svcall.c, armv7-m/svcall.h, and
up_svcall.c: The system call library can now be built with
CONFIG_NUTTX_KERNEL. New selection: CONFIG_LIB_SYSCALL (2014-8-28).
* arch/arm/src/armv7-a/syscall.h: "Add an ARMv7-A system call definition
header file (2014-8-28).
* arch/arm/include/armv7-a/irq.h, arm_syscall.c, syscall/Kconfig,
syscall_funclookup.c, syscall_nparms.c, and syscall_stublookup.c:
ARMv7-A: Add SYSCALL handling logic (2014-8-28).
* nuttx/fs/, drivers/, and sched/: Remove explicit references to errno.
That is a problem from within the kernel for certain configurations
(there are still more references to be fixed) (2014-8-28).
* sched/clock/clock.h: In some configurations, g_system_tmer must be
extern'ed as a private variable in the OS (2014-8-28).
* include/errno.h and nuttx/clock.h: errno and g_system_timer must be
handled in a special way if there are external modules (2014-8-28).
* include/sys/syscall.h: System definitions should depend on
CONFIG_LIB_SYSCALL not CONFIG_NUTTX_KERNEL (2014-8-28).
* arch/arm/src/armv7-a: Rename syscall.h to svcall.h to work around
some include path name collisions; fix some compilation errors in
SYSCALL logic when debug is enabled (2014-8-29).
* arch/arm/include/arm/syscall.h and armv7-a/syscall.h: Fix ARM7/9
and Cortex-A SYSCALLs: For threads in SVC mode, the SVC instructions
clobbers R14. This must be taken account in the inline assembly
(2014-8029).
* arch/arm/src/armv7-a/arm_syscall.c: Fix hard coded values in
dispatch_syscall() inline assembly. Back out/corect part of last
change; that was going the wrong direction. With this change, the
apps/examples/elf test works with address environments and call
gates! (2014-8029).
* Many, many files: Renamed CONFIG_NUTTX_KERNEL to CONFIG_BUILD_PROTECTED.
Added configuration CONFIG_BUILD_KERNEL that will be used to generate
a kernel (only) with full process support and will execute programs
only from a file system. This new configuration is defined but only
partially integrated (2014-8-29).
* tools/refresh.sh: Add a tool to make refreshing configurations easier
when you wnat to do a lot of them (2014-8-29).
* tools/mksyscall.c: Build syscalls that do not need header files (there
are none, however) (2014-8-30).
* include/nuttx/errno.h syscall/syscall.csv and syscall_funclookup.c:
Yet more issues with errno access via syscalls fixed (2014-8-30).
* Makefile.unix, Makefile.win, Directories.mk, FlatLibs.mk, ProtectedLibs.mk,
and KernelLibs.mk: Separate the common directory selection logic
into Directories.mk; add support for kernel build. Move library
selection logic in common *.mk files, one each for flat-, protected-,
and kernel- mode builds (2014-8-30).
* include/nuttx/syscall.h syscall/syscall.csv: clock_systimer needs the
same fix as for get/set_errno when used with system calls (2014-8-30).
* sched/Kconfig sched/init/os_bringup.c tools/cfgdefine.c: Add
configuration options to start the system from a program on a file
system (2014-8-30).
* mm/ and include/nuttx/mm.h: Remove CONFIG_MM_MULTIHEAP. Non-multiheap
operation is no longer supported (2014-8-31).
* configs/sama5d4-ek/kernel: Add a configuration for testing the
kernel build configuration (still a work in progress) (2014-8-31).
* include/nuttx/arch.h, kmalloc.h, mm.h, sched/init/os_start.c,
sched/sched_free.c, and arch/arm/src/sama5/sam_allocateheap.c: Clean
up some kernel build heap allocation issues. The Cortex-A kernel
build now compiles without errors (but cannot link until brk() and
sbrk() are implemented) (2014-8-31).
* mm/*mm_brkaddr.c, *mm_extend.c, Makefile, and include/nuttx/mm.h: Add
low-level memory management hooks that will be needed to support brk()
and sbrk() (2014-8-31).
* Rename krealloc() to kmm_realloc() and kmemalign to kmm_memalign() for
* Rename kurealloc() to kumm_realloc(), kumemalign to kumm_memalign(),
and kufree to kumm_free for consistency with the rest of the naming
* Rename kumalloc() to kumm_malloc() and kuzalloc to kumm_zalloc() for
consistency with the rest of the naming (2014-8-31).
* Rename kmalloc() to kmm_malloc(), kzalloc() to kmm_zalloc(), and
kfree() to kmm_free() and kuzalloc to kumm_zalloc() for consistency
with the rest of the naming (2014-8-31).
* mm/mm_sbrk.c, kmm_sbrk.c, umm_sbrk.c, and include/unistd.h: Initial
* implementation of sbrk() (2014-9-1).
* arch/arm/src/armv7-a/arm_pgalloc.c, mm/umm_malloc.c, and mm_zalloc.c:
Completes the implementation of sbrk() (untested) (2014-9-1).
* configs/cc3200-launchpad: Updates to the CC3200-Launchpad configuration
for better compatibility. From Jim Ewing (2014-9-1).
* Removed all support from the 8051 from the NuttX source tree. The
obsoleted code along with the removal patch can now be found at
misc/Obsoleted/ (2014-9-1).
* sched/sched/sched_releasetcb.c and task_setup.c, include/nuttx/arch.h
and sched.h, and arch/*/*/*/Make.defs: There used to be two ways to
pass parameters to new tasks, depending upon the configuration: Either
(1) argv[] as created as an array with each string strdup'ed. Or (1)
argv[] array and strings were created on the stack before the new tas
was started. Now, there is only one way, way (1). Way (2) might be
slightly more compact, but this is not worth carry the complexity of
two different ways of doing the same thing (2014-9-1).
* sched/init/os_start.c: Fix error in initialize of IDLE TCB due to last
change. The argument list is no long in the TCB, but in the stack.
But not for the IDLE task. It needs its own mini-argv[] list so that
it looks more like other threads (2014-9-1).
* arch/arm/src/armv7-a/arm_addrenv.c, include/nuttx/addrenv.h, and
mm/umm_*.c: Space at the beginning of the process data space is now
reserved for user heap management structures. In the kernel build
mode, these heap structures are shared between the kernel and use
code in order to allocate user-specific data (2014-9-2).
* include/unistd.h, nuttx/mm.h, and kmalloc.h, sched/init/os_start.c
and sched/sched/sched_free.c, mm/kmm_sbrk.c and mm_sbrk.c, and
arch/arm/src/sama5/sam_allocateheap.c: Restructuring of build to
allow use of use-space allocators by kernel logic in the kernel
build (2014-9-2).
* arch/arm/include/armv7-a/irq.h, arm_syscall.c, crt0.c,
include/nuttx/addrenv.h, and mm/umm_*.c: Add support for delivery
of use-mode signals in the kernel build (2014-9-2).
* mm/mm_graninit.c: Granule allocator initialization uses wrong
allocator to setting aside kernel memory (2014-903).
* sched/init/os_start.c: IDLE TCB setup needs to indicate the the
IDLE thread is a privileged, kernel thread (2014-903).
* include/nuttx/sched.h: Add a flag to group structure: If the group
is created by a kernel thread, then all resources in the group must
be privileged (2014-903).
* sched/group/group_malloc.c, group_zalloc.c, group_free.c,
group_create.c, group.h: Group creation logic must use new group flag
when allocating resources so that the privileges on the resource are
set correctly (2014-903).
* sched/Kconfig and sched/init/os_bringup.c: Add capability to perform
initial board initialization on a separate worker thread (2014-9-3).
* tools/mkexport.sh: Fix 'make export'. Adding subdirectories to the
sched/ directory broke the header file collection logic (2014-9-4).
* Makefile.unix, Makefile.win, FlatLibs.mk, KernelLibs.mk, and
ProtectedLibs.mk: In the kernel or protected builds, only the user
libraries should be exported (2014-9-4).
* Makefile.unix, Makefile.win, and tools/mkexport.sh: In 'make export',
do not copy internal header files or build scripts if this is a
kernel or protected build (014-9-4).
* tools/mkexport.sh, arch/*/src/Makefile, arm/src/a1x/Make.defs,
sama5/Make.defs, armv7-a/crt0.c: The 'make export' target needs to
bundle up the user C startup file (crt0), not the kernel head object
for the kernel and protected builds (2014-9-4).
* tools/mkexport.sh: Add .config file to export package (2014-9-5).
* configs/16z: Support for this board has been removed from the NuttX
source tree (but still can be found in the misc/Obsoleted directory).
This port is not ready for usage but may return to the NuttX tree
at some point in the future (2014-9-5).
* tools/Makefile.export and mkexport.sh: The apps/ import build requires
that we extract more information from the platform Make.defs file
(2014-9-5).
* tools/Makefile.export and mkexport.sh: Need to export the setting of
WINTOOL as well (2014-9-5).
* SAMA5D4-EK kernel configuration now builds NSH instead of examples/elf
(2014-9-6).
* libc/pthread/pthread_attrinit.c, sched/pthread/pthread_create.c, and
include/nuttx/pthread.h: Fix some conditional compilation in kernel
build mode. And, while we are touching this logic, make
g_default_pthread_attr const (2014-9-7).
* arch/arm/src/armv7-a/arm_addrenv.c: Correct size comparison (pages vs.
sections) (2014-9-7).
* configs/sama5d4-ek/src/sam_bringup.c and sama5d4-ek.h: Need to call
elf_initialize() on bring-up before attempting to load ELF modules
(2014-9-7).
* arch/arm/src/armv7-a/arm_addrenv.c: Fix loop counter... was overrunning
a table on larger ELF files (2104-9-7).
* arch/arm/src/sama5/sam_memories.c, sam_pgalloc.c, and sam_pgalloc.h: In
kernel build with address environment, need logic to map user virtual
addresses to physical addresses, and vice versa (2104-9-7).