luchianmihai/walnux
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Preparation for NuttX-6.33 release (maybe tomorrow)

Browse source
This commit is contained in:
Gregory Nutt 2014-01-29 14:41:18 -06:00
parent ca807a4ed3
commit 075ba4ff89
2 changed files with 331 additions and 22 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

40
ChangeLog
View file

@ -5710,8 +5710,8 @@
the interface in this case (2013-10-5).
* net/netdev_router.c and net/uip/uip-arp.c: When the target IP address
does not lie on the device's networker when we have a routing table,
looking the the correct router IP address to use in the ARP request.
In that case, we want the MAC address of the router, not of the the
looking the correct router IP address to use in the ARP request.
In that case, we want the MAC address of the router, not of the
target endpoint (2013-10-5).
* net/netdev_rxnotify.c and others: Use the new signature of rxnotify
caused by the routing table. From Max Holtzberg (2013-10-6).
@ -6168,7 +6168,7 @@
the top of the loop not the bottom. Otherwise, we will do a bogus
transfer with the out-of-range block before we test it (2013-12-05).
6.33 2014-xx-xx Gregory Nutt <gnutt@nuttx.org>
6.33 2014-01-30 Gregory Nutt <gnutt@nuttx.org>
* arch/arm/include/a1x and src/a1x: Directory structure to support
the Allwinner A10. Not much there on the initial check-in
@ -6208,7 +6208,7 @@
From Ken Pettit (2013-12-12)
* arch/arm/src/a1x/chip/a1x_timer.h: Timer register bit definitions
(incomplete on initial check-in) (2013-12-12).
* lbc/time/lib_strftime.c: Need null-termination on the the string
* lbc/time/lib_strftime.c: Need null-termination on the string
generated by strftime(). From Max Holtzberg (2013-12-12).
* sched/nanosleep.c and include/time.h: Add nanosleep() (2013-12-12).
* libc/unistd/lib_sleep.c and lib_usleep.c. Move sleep() and
@ -6241,7 +6241,7 @@
* arch/arm/src/armv7-a/arm_head.h: Fix some errors in the cache
invalidation logic (only seem to matter for Cortex-A8) (21-3-12-19).
* Kconfig and all Make.defs files: Add CONFIG_DEBUG_NOOPT. Now
you can indepenently enable/disable debug symbols and optimization
you can independently enable/disable debug symbols and optimization
(2013-12-20).
* configs/README.txt and Documentation/NuttxPortingGuide.html:
Remove documentation of NuttX configuration variables. Since
@ -6261,7 +6261,7 @@
* configs/stm3210e-eval/buttons: Converted to use kconfig-frontends
tools (untested) (2013-12-22).
* configs/Kconfig, configs/viewtool-stm32f107/include/board-stm32f103vct6.h,
and /board-stm32f107vct6.h: The viewtools board can now be configured
and /board-stm32f107vct6.h: The viewtool board can now be configured
to support either the STM32F103VCT6 or the STM32F107VCT6 (2013-12-22).
* configs/Kconfig, README.txt, viewtool-stm32f107/highpri, Kconfig,
README.txt, and src/stm32_highpri.c: This is the initial framework
@ -6280,7 +6280,7 @@
and the name of the common vector handling logic so that the MCU-
independent logic and work with these (2013-12-23).
* configs/viewtool-stm32f107/scripts: Move the RAM vector tables to the
beginning of SRAM. It seems to require this alignament. Also, we
beginning of SRAM. It seems to require this alignment. Also, we
don't need different scripts for the CMNVECTOR case now that the vector
table has a common name.
* arch/arm/src: armv-7/up_exception.S and xxx/xxx_vectors.S where
@ -6312,14 +6312,14 @@
pin definition that would prevent building USB for the connectivity and
performance lines (2013-12-25).
* stm32l15xxx_pinmap.h: Fix a typo in USB pin definitions (2013-12-25).
* configs/viewtools-stm32f107: Fix building of USB for F103 and F107.
* configs/viewtool-stm32f107: Fix building of USB for F103 and F107.
F103 has device only; F107 has OTG FS (2013-12-25).
* arch/arm/src/stm32/stm32f10xxx_rcc.c and chip/stm32f10xxx_rcc.h: Add
clocking support for STM32F107 USB OTG FS (which does not work)
(2013-12-26).
* configs/viewtool-stm32f107: Updates to USB for F103 and USB OTG FS
for F107 (2013-12-16).
* arch/arm/src/stm32/chip/stm32f3xxxx_pinmap.h: Fix pin definitioni names
* arch/arm/src/stm32/chip/stm32f3xxxx_pinmap.h: Fix pin definition names
for SPI2 MOSI and MISO. Noted by Brian Webb (2013-12-27).
* graphics/nxmu, nxsu, and nxglib and libc/nx, nxmu, and nxglib: Massive
reshuffling of files with (hopefully) no logic changes. This
@ -6360,7 +6360,7 @@
Thomas Grubler (2014-1-2).
* tools/configure.c and mkdeps.c: Fixes for Windows build issues
from Max Holtzberg (2014-1-4).
* configs/olimex-stm32-p107/nsh/Make.defs: Add native Windoes build
* configs/olimex-stm32-p107/nsh/Make.defs: Add native Windows build
support for the Olimex STM32 P107. From Max Holtzberg (2014-1-4).
* Makefile.win: Changes for native Windows build: fix creation of
a .version file if one does not exist. Make sure that the APPDIR
@ -6431,7 +6431,7 @@
* All of Jason Jiang's TCP write buffering logic is checked in.
Unfortunately, it does not yet work (2013-1-14).
* configs/px4fmu-v2_upstream: Configuration for testing simple
configurations on the the PX4FMU v2. This version is incomplete
configurations on the PX4FMU v2. This version is incomplete
for the PX4 application and is not a replacement for the version
in the PX4 GIT repository.
* fs/fat/fs_fat32.c: A correction to FAT cluster allocation from
@ -6440,7 +6440,7 @@
then the socket-related write buffering information must copied
with the other cloned socket data (2014-1-14).
* net/net_close.c: If CONFIG_NET_TCP_WRITE_BUFFERS is selected,
then it is necessary to free the write buffere callback structure
then it is necessary to free the write buffer callback structure
when the socket is closed (2014-1-14).
* fs/fat/fs_fat32.c: Fix some root directory logic that was
conditionally done only for FAT 32. Apparently this needs to
@ -6450,7 +6450,7 @@
implementations of up_maskack_irq() for all Cortex-M architectures: Do
not disable and enable the IRQ on each interrupt. Because (1)
interrupts are already disabled on interrupt entry, (2) this
interferes with controlling the IRQ interrrupt setting from interrupt
interferes with controlling the IRQ interrupt setting from interrupt
handlers, and (3) up_disable_irq() does not work anyway so that this
has never done anything (2014-1-15).
* All implementations of up_disable_irq() for all Cortex-M3 and M4
@ -6461,7 +6461,7 @@
not work; writing a '0' to the ISER register has no effect. That
means that up_disable_irq() was doing nothing! It turns out that that
is not really important because up_disable_irq() is not really used
for that purpose. But some spurions STM32 ADC interrupts have been
for that purpose. But some spurious STM32 ADC interrupts have been
reported to me and this turned out to be the cause in that case. My
concern now that up_disable_irq() works is that there may now be
unmasked bugs that leave devices in the disabled state? Thanks to
@ -6478,7 +6478,7 @@
source the Kconfig file from the custom board directory (2014-1-16).
* Makefile.unix, Makefile.win, and configs/Kconfig: After thinking a
little harder, I re-implemented the backed-out custom configuration
feature. The new version does not have bad side-effect (1) (but stil
feature. The new version does not have bad side-effect (1) (but still
has bad side-effect (2)) (2014-1-16).
* arch/x86/Kconfig and configs/qemu-i486/nsh/Make.defs and
ostest/Make.defs: Add a configuration option to select the -m32
@ -6486,8 +6486,8 @@
compiler (2014-1-18).
* include/ctype.h: Typo in macro name: iscntrl, not iscontrol (2104-1-17)
* libc/unistd/lib_getopt.c: If there are no arguments (argc == 1), then
getopt() make like the optind variable in an undefined state (2014-1-20).
* configs/olimex-stm32-p107: Failes to build is SPI3 for UEXT is not
getopt() will leave the optind variable in an undefined state (2014-1-20).
* configs/olimex-stm32-p107: Fails to build if SPI3 for UEXT is not
remapped. From Max Holtzberg (2014-1-21).
* Several network related files: Changes from Max Holtzberg to improve
how network status is reported. New controls to manage carrier
@ -6501,7 +6501,7 @@
* /arch/arm/src/Makefile: dependency directory list is now computed
from the VPATH. From Richard Cochran (2014-1-23).
* arch/arm/src/efm32 and include/efm32: Basic support for the EFM3
processor family from Richarc Cochran (2014-1-23).
processor family from Richard Cochran (2014-1-23).
* configs/efm32-dk3650: This is Energy Micro's development kit for
the Leopard Gecko MCU, which is an ARM Cortex-M3 device. From Richard
Cochran (2014-1-23).
@ -6522,7 +6522,7 @@
to fix 'Error: .size expression for idle_stack does not evaluate to
a constant" (2014-1-25).
* arch/arm/include/efm32, arch/arm/src/efm32, and configs/efm32-dk360:
Removed all EFM32 support. Not yet readyto be fielded (2014-1-27).
Removed all EFM32 support. Not yet ready to be fielded (2014-1-27).
* arch/arm/src/armv7-a/arm_head.S, sama5/sam_boot.c, and several other
files: Now supports execution from NOR FLASH with .data and .bss in
SDRAM. This was not possible prior to this because .bss and .data
@ -6533,3 +6533,5 @@
original, efficient logic violated the ARM C ABI. It was okay when
called from boot logic, but not when called from C logic. This fixes
bugginess reported on 2014-1-28. (2014-1-29).
6.34 2014-xx-xx Gregory Nutt <gnutt@nuttx.org>

313
ReleaseNotes
View file

@ -5346,7 +5346,7 @@ Additional new features and extended functionality:
- Many new drivers including CAN, PWM, SSC/I2S
- PCK: Add support for programmable clock outputs.
- NAND: Driver with hardware ECC and DMA support
* Atmel AT91 SAMA5D3x Boards:
- SAMA5D3x-EK: Add support for the apps/examples/i2schar test.
@ -5478,7 +5478,7 @@ Bugfixes (see the ChangeLog for details). Some of these are very important:
* MTD Drivers
- MTD Partitions: Fix erase block vs page block confusion. From Ken
Pettit
Pettit
- SST25 Serial Flash: Improved write performance by fixing a bug that
prevented operation in the faster write mode. From David Sidrane.
@ -5487,7 +5487,7 @@ Bugfixes (see the ChangeLog for details). Some of these are very important:
- USB MSC Device: pthread_join() does not work if called from a
different task group than the pthread. This is correct behavior,
but still a problem. The correct solution would be configure the
USB MSC thread to a task, however, this workaround from David
USB MSC thread to a task, however, this workaround from David
Sidrane plugs the hole for now.
- USB CDC/ACM, USB MSC, and Composite Device Classes: Un-initialization
logic caused re-use of a stale pointer. Changed to a two pass
@ -5519,3 +5519,310 @@ Bugfixes (see the ChangeLog for details). Some of these are very important:
correctly.
- apps/system/composite/composite_main.c: The wrong handle was getting
nullified. From David Sidrane.
NuttX-6.32
----------
The 100th release of NuttX, Version 6.33, was made on January 30, 2014,
and is available for download from the SourceForge website. Note
that release consists of two tarballs: nuttx-6.33.tar.gz and
apps-6.33.tar.gz. Both may be needed (see the top-level nuttx/README.txt
file for build information).
Additional new features and extended functionality:
* Core OS Interfaces
- nanosleep() added. sleep() and usleep() are not longer core OS
interfaces. These have been moved into the library and re-implemented
as simple wrappers around nanosleep().
* File Systems
- procfs restructure by Ken Pettit. Added files for MTD status.
- procfs extended to show uptime and task group status.
- procfs now shows stack information.
* Networking
- Add support for the SO_LINGER socket option. Extended from logic
provided by Jason Jiang. Enabled with CONFIG_NET_SOLINGER.
- TCP write buffering support added. From Jason Jiang.
- Changes from Max Holtzberg to improve how network status is
reported. New controls to manage carrier detect.
* Graphics
- Massive reshuffling of files with little or no logic changes. This
reshuffling was necessary to build graphics applications with the
kernel builds where the graphics application lie in user space and
the core graphic server likes in kernel space. Moved much logic out
of nuttx/graphics to a new user library, libnx.
* Cortex-A5/A8/A9
- Existing Cortex-A5 support updated to include Cortex-A8/9.
- Add the syscall.h header file needed for the Cortex-A architecture.
- Start-up logic now supports execution from FLASH with .data and
.bss in SDRAM. This was not possible prior to this because .bss
and .data were initialized before SDRAM was configured.
* Allwinner A10:
- Basic support for the Allwinner (A10 Cortex-A8) is in place.
* Allwinner A10 boards
- Support for the pcDuino v1 board has been added. This support
is not yet very mature and still lacks many drivers. It is a
good starting point, however, if anyone wants to develop A10
support.
* Atmel SAMA5D3X
- Start-up logic now supports execution from FLASH with .data and
.bss in SDRAM. This was not possible prior to this because .bss
and .data were initialized before SDRAM was configured.
* Atmel SAMA5D3X Boards
- Partial logic in place to support the OV2640 camera. Still needs
quite a bit of additional logic to be useful.
* ARMv7-M
- Add support for high priority, nested interrupts. This change
effects not only core ARMv-7M logic, but all ARMv7-M MCU logic:
LM3S, LM4F, LPC17xx, LPC43xx, SAM3, SAM4, and STM32. See
http://www.nuttx.org/doku.php?id=wiki:nxinternal:highperfints
* STMicro STM32
- Add clocking support for STM32F107 USB OTG FS (which does not yet
work)
- Single-wire UART support for the F1 series from Thomas Grubler.
* STMicro STM32 Boards
- The Viewtool board can now be configured to support either the
STM32F103VCT6 or the STM32F107VCT6.
- The Viewtool STM32F107VCT6 now supports networking with the
DP83848C module installed.
- Add support for Viewtool SSD1289-based LCD on the Viewtool
STM32F103VCT6 board (untested)
- Add support for the XPT2046 touchscreen controller on the Viewtool
LCD module connected to the Viewtool STM32F103 board.
- Added kernel mode build support for the STM3240G-EVAL board. Added
a new kernel mode NxWM build configuration
- C++ static constructor logic from the nuttx/configs/stm3240g-eval/
directory to the apps/platform/stm3240g-eval/up_cxxinitialize.c
where is belongs so that it is available in user-space in the kernel
mode build.
- Add native Windows build support for the Olimex STM32 P107. From
Max Holtzberg
- Support for the STM32VL-Discovery board. Contributed by Alan
Carvalho de Assis
- Added a configuration for testing simple configurations on the
PX4FMU v2.
* ZiLOG ZNEO Boards
- Add an NSH configuration for the Z16F2800100ZCOG ZNEO board.
- Added support for the Toyaga 16Z ZNEO board. The 16Z board is
based on the ZiLOG ZNEOZ16F2811AL20EG part. See
https://github.com/toyaga/16z for further information
- Developed a patch to work around a compiler error that is revealed
when building the NSH configuration.
* Architecture-Independent Drivers
- OV2640 Camera Driver.
- Support for the Sharp Memory LCD from Librae.
* Library
- Pattern matching logic extended to handle set of characters and
ranges of character values. From Ken Pettit.
- sleep() and usleep() move library. These are not longer core OS
interfaces; they are simple wrappers for the nanosleep().
- Add an implementation of gets_s().
- Extend fopen() to include support for C11 exclusive open ("x")
open mode.
- Add an implementation of stpcpy()
- strtol(), strtoll, strtoul(), strtoull(), and strtod() from
libc/string to libc/stdlib where they belong.
* Header Files
- Add rsize_t to include/sys/types.h
- Add the C11 thread.h header file
* Configuration/Build System
- Top level Makefile now supports qconfig and/or gconfig targets that
may be used for configuration. These may be used if you build
kconfig-frontends with support for kconfig-qconf and/or
kconfig-gconf.
- Add CONFIG_DEBUG_NOOPT. Now you can independently enable/disable
debug symbols and optimization
- File system related header files moved to include/nuttx/fs
- Video related header files moved to include/nuttx/video
- Changes for native Windows build: fix creation of a .version file
if one does not exist. Make sure that the APPDIR environment
variable is set before configuring. From Max Holtzberg.
- Board configuration sub-directory can now be specified. The default
need not be used. This is only useful when CONFIG_ARCH_BOARD_CUSTOM
is selected and there is no meaningful default sub-directory.
- Many functions renamed to better conform with the naming standard:
up_buttoninit() renamed to board_button_initialize(), up_buttons()
renamed to board_buttons(), up_irqbutton() renamed to
board_button_irq(), up_ledinit() renamed to board_led_intialize(),
up_ledon() renamed to board_led_on(), and up_ledoff() renamed to
board_led_off(). All prototypes removed from board.h header files.
Now ONLY prototypes in include/nuttx/arch.h.
* Applications
- NSH: Refactor. Separate NSH command handling and command execution
from NSH line parsing.
- NSH: Will now support multiple commands on a command line, each
separated with a semi-colon.
- NSH: Add support of commands enclosed in back quotes as command
arguments.
- NSH: Can now handle arguments that are concatenations of constant
strings, command return data, application return data, and
environment variables.
- NSH: Add true and false commands.
- NSH: Add support for while-do-done and until-do-done loops. These
only work when executing a script file because they depend on the
ability to seek in the file to implement the looping behaviors.
Can be conditionally compiled out to reduce footprint.
- NSH: Loosen up if-then-else-fi syntax to allow a command to be on
the same line as the then or else tokens like:
"if true; then echo true; else echo false; fi". More like bash.
- NSH: Add a break command that can be executed with a loop to
terminate the loop immediately.
- system/inifile: A simple .INI file parser.
- system/vi: Add support for a tiny, VI work-alike editor. This is
a minimal full screen editor that words with a terminal emulator
that using VT100 commands.
Efforts In Progress. The following are features that are partially
implemented but present in this release. They are not likely be be
completed soon.
* kconfig-fronted Configuration:
- Conversion of old configurations to use the kconfig-frontends
tool is an ongoing effort that will continue for some time.
At this time, only 53% of the configurations have been converted
to use the kconfig-frontends tools.
Bugfixes (see the ChangeLog for details). Some of these are very important:
* Core OS
- Added missing sem_timedwait() system call.
* Networking
- Prevent tcp_connect callback from being double freed. From Max
Holtzberg
- uiplib: Support new definitions and state passing for network
device status. From Maz Holtzberg.
* File Systems
- FAT: (1) A correction to FAT cluster allocation, and (2) Fix to some
root directory logic that was conditionally done only for FAT 32.
Apparently this needs to done for all FAT types. Both from Andrew
"Tridge" Tridgell via Lorenz Meier.
* Binary Formats
- Fix a small memory leak when attempting to load a program from a file.
* Cortex-A8/9
- Fix some errors in the cache invalidation logic (only seem to matter
for Cortex-A8).
- Add more nop's after enabling the MMU. The cortex-a8 seems to need
these
* Graphics
- Fix a typo that caused a compilation error when
CONFIG_NXCONSOLE_BPP < 8. From Librae
* ARMv7-M
- Correct alignment of RAM vector table.
- Interrupt handling: Do not disable and enable the IRQ on each
entry, (2) this interferes with controlling the IRQ interrupt
setting from interrupt handlers, and (3) up_disable_irq() does
not work anyway so that this has never done anything.
- Fix all implementations of up_disable_irq() for all Cortex-M3 and
M4 architectures: To enable an interrupt on the Cortex-M3/4 CPU,
you need to set a bit in the ISER register. To disable the
interrupt, you need to set a bit in the ICER register. Existing
logic was trying to disable interrupts by clearing the bit in the
ISER register. That will not work; writing a '0' to the ISER
register has no effect. That means that up_disable_irq() was
doing nothing! It turns out that that is not really important
because up_disable_irq() is not really used for that purpose. But
some spurions STM32 ADC interrupts have been reported to me and this
turned out to be the cause in that case. Thanks to Manuel Stühn for
the tip.
* STMicro STM32
- Fix configuration and pin definitions that would prevent building
USB for the connectivity and performance lines.
- STM32L15xx: Fix a typo in USB pin definitions
- Fix pin definition names for SPI2 MOSI and MISO. Noted by Brian Webb
- Various fixes for STM32F103ZE SPI3 pin mapping definitions. From
Steve Redler IV
* STMicro STM32 Boards
- Olimex STM32-P107: Failed to build if SPI3 for UEXT is not remapped.
From Max Holtzberg
* x86 Boards
- Add a configuration option to select the -m32 compiler option when
building for a 32-bit target on a native 64-bit compiler. So far,
only used for the qemu/i486 configuration.
- Patch from Matt Campbell to fix 'Error: .size expression for idle_stack
does not evaluate to a constant".
* Library
- strftime(): Needs null termination on the generated string. From
Max Holtzberg.
- Typo in ctype.h macro name: iscntrl(), not iscontrol().
- If there are no arguments (argc == 1), then getopt() will leave the
optind variable in an undefined state (2014-1-20).
- ARMv7-M memcpy(): Assembler changes with the gcc-47 distribution
from ARM mean that we need to be explicit about branch sizes; one
or more of the wide branch opcodes results in bad table branching.
From Mike Smith.
* Configuration/Build System
- Fixes for some tools/ for Windows build issues from Max Holtzberg.
- Config.mk: 'cypath' must be called if we are using a Windows native
toolchain with the Cygwin 'make' to convert paths to proper
Windows paths. From Richard Cochran.
- ARM Makefile: Dependency directory list is now computed from the
VPATH. From Richard Cochran.
- builtin/Makefile: Fixes for native Windows build from Max Holtzberg.
- Refactor some configuration dependencies: NSH networking features
should depend on netutil selections; netutil selections, in turn,
should depend on networking selections.
* Applications
- NSH: Modified the df -h logic to eliminate truncating numbers in
conversion (like 7900 -> 7M). From Ken Pettit.