Newer
Older
* arch/arm/src/lpc43xx/lpc43_rgu.c: The soft reset logic called from the
beginning of __start seems cause problems. A magic delay seems to improve
the logic some. But I suspect that real fix is to get rid of all of the
soft reset logic. This would also be a critical bugfix if I believed
that it really fixed all of the issues.
* arch/arm/src/lpc43xx/chip/lpc43_cgu.h: Fix a bit mask in the PLL1
control register. Critical bugfix.
* arch/arm/src/lpc43xx/lpc43_clockconfig.c and configs/lpc4330-xplorer/include/board.h:
Implement PLL1 ramp-up logic; Now the LPC43xx is running at 204MHz.
* configs/lpc4330-xplorer/*/defconfig: Re-calibrated delay loops using
the 204MHz clock. The LPC43xx ripping rips! This calibration was performed
with symbols enabled and all optimization disabled. It will need to be
better recalibrated again down the road.
* arch/arm/src/stm32/stm32_exti.c: Renamed to rch/arm/src/stm32/stm32_exti_gpio.c
to make a little room in the file name space.
* arch/arm/src/stm32/stm32_exti_alarm.c: Add initial logic to attached the
RTC alarm EXTI interrupt. This is work be performed mostly by Diego Sanchez.
* arch/arm/src/lpc43xx/lpc43_spifi.c, lpc43_spifi.h, and chip/lpc43_spifi.h: Add
logic to configure and initialize the SPIFI device (does not yet work).
* configs/lpc4330-xplorer/include/board.h: Reduce SPI SCLK value.
* arch/arm/src/lpc43xx/lpc43_spifi.c, lpc43_spifi.h, and chip/lpc43_spifi.h:
Logic completely redesigned. It now creates an MTD driver to access SPIFI...
but the driver still does not work.
patacongo
committed
* arch/arm/src/stm32 and arch/arm/include/stm32: Make name of RTC ALARM interrupt
common on STM32 F1,2,4
* arch/arm/src/stm32 and arch/arm/include/stm32: Add support for the
patacongo
committed
STM32F100x "Value Line" devices. This includes changes to stm32F10xx_rcc.c that
add the ability to run the chip off the internal oscillator. There is no open
board configuration for this part yet (the STM32VLDiscovery would be a candidate).
Contributed by Mike Smith.
* arch/arm/src/stm32: Fixed typos in conditional compilation in the CAN and DMA
and some pin configuration. This would have caused problems for STM32 F107xx.
Typos noted by Mike Smith.
* arch/arm/src/lpc43xx/lpc43_serial.c: Add support for certain RS-485 features
* lib/termios/lib_cfsetispeed.c, lib_cfsetospeed.c, lib_tcflush.c: Add
simple implementations of cfsetispeed(), cfsetospeed(), and tcflush().
* include/sys/str_tty.h, lib/lib_setspeed.c, lib_getspeed.c, and lib_resetspeed.c:
Add APIs to support setting non-standard BAUD values not supported by POSIX
termios. These are non-standard interfaces but have a precedence: There are
similar interfaces in AIX.
* include/sys/str_tty.h, lib/lib_setspeed.c, lib_getspeed.c, and lib_resetspeed.c:
Sigh... removed. We don't need any more almost standard interfaces! (SVN
revision 4968 if you want the short-lived code).
* include/termios.h and lib/termios/*: Open the existing, standard termios
interfaces to permit some non-standard baud settings. The new termios definitions
still supports the POSIX standard except that it does not strictly enforce
baud rate settings, permitting some non-portable, but useful baud rate settings
(this is what the short-lived AIX-like interfaces would have accomplished as well).
* include/termios.h and lib/termios/*: Redesigned yet again (this is getting
painful. NuttX now supports the BOTHER baud setting just as Linux does. termios
Bxxx definitions are again encoded; cf[set|get][o|i]speed now deal with only the
encoded values. If the encoded baud is set to BOTHER, then the values in the (non-
standard) c_ispeed and c_ospeed baud values may be accessed directly.
* arch/arm/src/stm32/stm32_serial.c: Add minimal termios support for the STM32
(BOTHER style baud settings only). Contributed by Mike Smith.
* configs/lpc4343-xplorer/src: Clean up SPIFI-library based build to that it
actually works.
* arch/arm/src/lpc43xx/lpc43_spifi.c: Add support for verification to writes.
Add debug option to dump buffers. Several bugfixes... almost works.
* include/termios.h, lib/termios/*, and arch/arm/src/stm32/stm32_serial.c: :
BOTHER is gone again.
* arch/arm/src/stm32/stm32_sdio.c and chip/stm32f20xx_pinmap.h: STM32 F2 SDIO
fixes from Gary Teravskis and Scott Rondestvedt.
* include/termios.h and lib/termios/*: Replace cfsetispeed and cfsetospeed with
cfsetspeed (with definitions for the input/outputs in termios.h).
* configs/stm32f4discovery/src and configs/stm32f4discovery/pm: Add a power
management configuration for the STM32F4Discovery and supporting logic. This
check-in also includes some fixes for the F4 RTC alarm logic.
* drivers/input/pga11x.c and include/nuttx/input/pga11x.h: Add support for the
TI PGA112/3/6/7 amplifier/multiplexer parts.
* configs/mirtoo/README.txt, nsh/defconfig, and nxffs/defconfig: Add support
for the PGA117 on the Mirtoo module.
* drivers/analog/pga11x.c and include/nuttx/analog/pga11x.h: These belong in
the analog subdirectories, not input.
* configs/compal_e99/src/ssd1783.c and /ssd1783.h: Drivers for the SSD1783
LCD found in the Motorola C155 telephone. The driver is specific to the C155
because it uses the uwire transport. Contributed by Denis Carilki and
Alan Carvalho de Assis.
* drivers/power/pm_changestate.c. Correct a case where interrupts were not
being re-enabled. Found by Diego Sanchez.
patacongo
committed
* configs/mirtoo/nxffs/defconfig: This Mirtoo NXFFS configuration now uses the
open Pinguino toolchain by default. This is necessary because the free C32
toolchain does not support any optimization and the unoptimized NXFFS image
hits the PIC32MX2 FLASH size (128K). There is plenty of room to grow using
the Pinguino toolchain with -O2 optimization.
* configs/mirtoo/src/up_adc.c. This is just a stub for now, but this is
where Mirtoo ADC logic will eventually need to go.
* arch/mips/src/pic32mx/pic32mx-gpio.c: Now supports the PIC32MX1/2 ANSEL
IOPORT register.
* lib/string/lib_memchr.c: Add support for memchr() (contributed by Mike Smith)
* lib/string/lib_memccpy.c: Add support for memccpy()
* arch/arm/src/lpc17xx/lpc17_serial.c: Now supports ioctl commands to change
the baud using tcsetattr() (contributed by Chris Taglia).
* arch/*/src/*_serial.c: Fix ioctl method return values. These methods
should return a negated errno value; they should not set the errno
variable.
* sched/on_exit.c, sched/task_exithook.c, and include/nuttx/sched.c: Add
support for multiple registered on_exit() functions if CONFIG_SCHED_ONEXIT_MAX
is defined.
patacongo
committed
* drivers/syslog/ramlog.c: Move the RAM SYSLOG device into drivers/syslog
so that it will be in the same directory as some new SYSLOGing devices
in the works.
* include/nuttx/syslog.h and drivers/syslog/ramlog.c: The SYSLOG putc function
now has a common name that is independent of the device that provides the
SYSLOG.
* include/nuttx/syslog.h and drivers/syslog/syslog.c: This is a new, generic
SYSLOG device that can redirect debug output to any character device or file.
So you can log debug output to a file or you can put the console on /dev/ttyS0
and the debug output on /dev/ttyS1.
* arch/arm/src/lpc43xxl/lpc43_spifi.c: Correct an addressing error in the LPC43
SPIFI MTD driver
* drivers/syslog/syslog.c and fs/fs_syslog.c: Moved the generic syslog logic
from drivers/syslog to fs/ where is belongs. Especially after realizing that
the syslog logic is going to have to some internal FS operations in order
to realize a totally thread-independent SYSLOG interface.
patacongo
committed
* arch/arm/src/stm32/stm32*_rcc.c and .h: If CONFIG_PM is defined, add a
function called stm32_clockenable() that can be used by PM logic to re-start
the PLL after re-awakening from deep sleep modes.
* fs/fs_foreachinode.c and fs/fs_foreachmountpoint.c: Add logic to traverse
inodes and mountpoints in the NuttX pseudo-file system.
* fs/fat/fs_fat32.c: Max. filename length reported by statfs() was wrong
if FAT long file names were enabled.
* lib/stdio/lib_libvsprintf.c: Fieldwidth and justification were not
supported for the %s format. As a result, %s, %12s, and %-12s all
produced the same output.
* lib/stdio/lib_libdtoa.c: Fix several issues with presenting floating
point numbers (conversions are fine, but presentation was bad). This
is a critical bug fix if you use printf or sprintf to deal with floating
point numbers.
* lib/stdio/lib_libdtoa.c and lib_libvsprintf.c: Correct some floating
point options.
patacongo
committed
* arch/arm/lpc43xx/lpc32_usb0dev.c: Add framework for development of
an USB0, device-side driver for the LPC43XX. The initial check-in,
however, is simply for the LPC31xx driver with name changes. The
LPC31xx has the same USB IP, but will require some additional initialization
(and lots of testing) before it can be used with the LPC43xx.
* nuttx/Documentation/NuttShell.html: Added a section covering ways to
* arch/arm/src/stm32/chip/stm32f1*_pinmap.h: STM32 CAN TX/RX pins reversed;
inconsistent conditional compilation. Reported by Max Holtzberg.
* arch/arm/*/stm32: Add support for STM32 F107 "Connectivity Line"
Ethernet (contributed by Max Holtzberg).
* configs/olimex-stm32-p107: Add board support for the Olimiex STM32-P107
board (contributed by Max Holtzberg).
* arch/arm/src/stm32/stm32f2xx_dma.c, stm32f4xx_dma.c, stm32_serial.c, and
stm32_spic.c: DMA priority was getting zeroed by STM32 F2/F4 DMA drivers
so that all DMAs ran at the lowest priority.
patacongo
committed
* configs/stm3240g-eval/include/board.h and configs/stm3220: Drop SD card
frequency from 24 to 16 MHz. Apparently 24 MHz is too fast for the board.
This (plus the change to the STM32 DMA (above) fixes SDIO DMA on the
STM3240G-EVAL (and probably STM3220G-EVAL -- untested).
* arch/arm/src/stm32/stm32f2xx_dma.c and stm32f4xx_dma.c: Backed out the
DMA priority change just above. The reduced SD card frequency was
necessary and sufficient to resolve the problem.
patacongo
committed
* drivers/serial/serial.c: open, read, write, and poll methods may now
abort return EINTR (or a short transfer size) if a signal is received
while waiting to receive or send serial data. This behavior is required
by POSIX.
* include/sys/types.h: Define NULL to be (0) if __cplusplus is defined.
(contributed by Mike Smith)
* include/ctype.h: Remove a stray semi-colon in a definitions (Thanks
Mike Smith).
* configs/.../Make.defs. Fix C++ include path set-up in Make.defs file
for all 8-bit AVR platforms (Thanks Richard Cochran).
* lib/stdio/lib_*stream.c: Revised to handle new error return values from
serial.c.
* arch/arm/src/stm32/stm32_spi.c: SPI driver can now service re-
initialization (Mike Smith).
* tools/mkconfig.c: If CONFIG_DRAM_END is not specified, this tool
will provide default definition of (CONFIG_DRAM_START + CONFIG_DRAM_SIZE)
* arch/arm/src/stm32/stm32_otgfshost.c: Renamed from stm32_usbhost.c.
This is nearly code complete and, with any luck, will be available
* configs/*/defconfig: Update all defconfig files to remove syntax
that is incompatible with the mconf configuration tool.
* arch/arm/src/stm32/stm32_otgfshost.c: This driver now appears to be
6.22 2012-09-29 Gregory Nutt <gnutt@nuttx.org>
* include/semaphore.h, sched/sem_holders.c, and lib/semaphore/sem_init.c:
Fix some strange (and probably wrong) list handling when
CONFIG_PRIORITY_INHERITANCE and CONFIG_SEM_PREALLOCHOLDERS are defined.
This list handling was probably causing errors reported by Mike Smith
* sched/sched_waitpid.c: Fix a possible issue with logic logic that
should be brought into a critical section (suggested by Mike Smith)
* sched/sched_setuptaskfiles.c: Should be 'struct socket' not
'struct sockets'. How did this compile before? (found by Kate)
* syscall/syscall.csv: Fix prototype for usleep() and prctl() (also
from Kate).
* arch/arm/src/lpc17xx/lpc17_ethernet.c: Conditionally elide setting PHY
speed/duplex. This does not work for certain PHYs. Still some unresolved
issues (also from Kate).
* tools/Config.mk, Makefile, configs/*/Make.defs: Add a new Makefile
fragment to de-quote certain strings from the Kconfig logic that
need to be used at path segments (Richard Cochran).
* arch/arm/src/stm32/stm32_usbotghost.c: The STM32 USB host driver only
works with debug turned on. The problem appears to be that with debug
OFF, there are more NAKs occuring in more places than before and this
reveals a variety of errors. This check in improves NAK robustness
for control transfers but does not resolve all of the issues.
* configs/stm3220g-eval/*/defconfig: Calibrated delay loop. It had
* sched/sem_holder.c: Add logic to handler some priority inheritance
cases when sem_post() is called from an interrupt handler. The
logic is clearly wrong, but it is not known if this is the
cause of any known bugs.
* lib/stdio/lib_perror(): Add perror(). Contributed by Kate.
* lib/string/lib_strerror(): Add option CONFIG_LIBC_STRERROR that
is now required to enabled strerror(). Add an option
CONFIG_LIBC_STRERROR_SHORT that can be used to output shortened
strings by strerror().
* arch/arm/src/stm32/stm32_usbotghost.c: Finally... the USB OTG FS
* configs/stm32f4discovery/*: Added and verifed support for USB OTG FS
host on the STM32F4Discovery board.
* configs/*/defconfig: Remove configuration documentation from config
files. It is redundant, error-prone, and difficult to maintain.
Configuration documentation is available in configs/README.txt for
common configurations and in configs/*/README.txt for board and MCU-
specific configurations.
* configs/stm3240g-eval: Add USB host support.
* sched/os_bring.c, configs/*/defconfig, tools/mkconfig.c, and others: Added
configuration variable CONFIG_USER_ENTRYPOINT that may be used to change
the default entry from user_start to some other symbol. Contributed by
Kate. NOTE: This change does introduce a minor backward incompatibility.
For example, if your application uses NSH as its start-up program, then your
build will now fail because it will be unable to find "user_start". The fix
for this link failure is to add the following to your configuration file:
CONFIG_USER_ENTRYPOINT="nsh_main".
* libs/stdio/lib_libfread.c and lib_*flush*.c: Correct a couple of
error cases where the lib semaphore was not be released on error
exits (thanks Ronen Vainish). Also, improved some error reporting:
the generic ERROR was being used instead of the specific errno
value; the errno variable was not always set correctly.
* tools/mkfsdata.pl: The uIP web server CGI image making perl script was
moved from apps/netutils/webserver/makefsdata to nuttx/tools/mkfsdata.pl
(Part of a larger change submitted by Max Holtzberg).
* configs/stm3240g-eval/script/ld.script: All of the identical ld.script
files for the STM3240G-EVAL were replaced by one version in this directory.
* configs/stm3240g-eval/webserver: Configuration submitted by Max Holtzberg
for testing the changes to the uIP web server (see apps/ChangeLog.txt).
* lib/stdio/lib_perror.c: Remove CONFIG_LIBC_PERROR_DEVNAME. What was I
thinking? Arbitrary streams cannot be shared by different tasks.
patacongo
committed
* tools/mksyscall.c, csvparser.c, and csvparser.h: Separate CSV parsing
logic from mksyscall.c into files where it can be shared.
* tools/mksymtab.c: Add a tool that can be used to convert a CSV file
into a NuttX-style symbol table.
* sched/work_cancel.c: Fix a bad assertion (reported by Mike Smith)
* configs/stm3210e-eval/src/up_idle.c: Correct some power management
compilation errors (reported by Diego Sanchez).
* include/nuttx/wqueue.h, sched/work*, and others: Added logic to support
a second, lower priority work queue (CONFIG_SCHED_LPWORK).
* arch/arm/src/stm32/stm32_dma.c, chip/stm32*_memorymap.h: FSMC SRAM is
only 16-bits wide and the SDIO DMA must be set up differently.
* arch/arm/src/stm32/stm32_dma.c: Back out the 16-bit DMA change. It
is incorrect.
* configs/: Make use of UART4/5 vs USART4/5 consistent in all places.
* Kconfig: Serial 2STOP setting must be integer 0/1, not a boolean.
* lib/misc/sendfile.c and include/sys/sendfile.h: Add a Linux style
sendfile() (non-standard!)
* Kconfig: Refactor serial settings (moved from chip to drivers/serial).
AVR "teensy" now builds with Kconfig (contributed by Richard Cochran).
* Kconfig: Add configuration settings for the LPC17xx
* Kconfig: Add configuration settings for the LM3S (from Richard Cochran).
* Kconfig: Verify configuration settings for the STM32. This includes
changes in the way that the external SRAM is configured: Define
CONFIG_HEAP2_SIZE (decimal) instead of CONFIG_HEAP2_END (hex).
* tools/configure.sh: Don't append the apps directory path setting
if the correct setting is already in defined in the defconfig file.
* fs/fat/fs_utils.c: Improperly constructed bool expression. This
would cause many unnecessary writes to FLASH (Thanks Ronen Vainish).
* Kconfig: Verify configuration settings for the LPC43xx. This includes
some corrections to configuration variable names and defconfig settings.
* Kconfig: Add and verify configuration settings for the LPC31xx.
patacongo
committed
* arch/arm/src/stm32/stm32_uart.h and stm32_serial.c: Add logic to
re-initialize the console UART as needed to enable DMA on the
console UART (contributed by Mike Smith).
* net/recvfrom.c, net/Kconfig, include/nuttx/net/uipopt.h: Remove delay
after receiving data. That has historical reasons to be there (it
was needed before read-ahead buffering was added), but kills performance.
(Noted by Max Holtzberg).
patacongo
committed
* configs/shenzhou: Add beginnings of a board configuration for the
Shenzhou STM32107 board (see www.armjishu.com). Very little is in
place as of this initial check-in.
* QEMU: Fixes from Richard Cochran to build QEMU with Kconfig files.
* arch/*/src/Makefile: Remove some old logic that was kicked off
when CONFIG_BOOT_RUNFROMFLASH=y. The old logic used to use
objcopy to move sections. Newer logic changes the load position
of sections in the the linker script. As far as I can tell, there
is nothing in the source tree now that depends on the old way of
doing things (if I am wrong, they will need a change to the linker
script).
* configs/fire-stm32v2: Configuration for the M3 Wildfire board. I
don't know very much about this board other than is has an
STM32F103VET6 chip, LCD, touchscreen, and ENC28J60 network. Very
little is in place on the initial check-in.
* configs/shenzhou: Coding for the Shenzhou board port is complete,
but tested has been deferred until I get the right tools.
* arch/arc/include/stm32/chip.h and arch/arm/src/stm32/chip.h:
Add support for the STM32F103VET6.
* fs/fs_fdopen.c: Bad check for failure to allocate memory. (Noted
by Ronen Vainish).
* drivers/mmcsd/mmcsd_sdio.c: If the MMC/SD driver were ever
uninitialized then there would be a double release of memory
(Noted by Ronen Vainish).
* fs/mmap/fs_rammap.c: Fix logic error and errno check (contributed
by Kate).
* arch/avr/src: Fixes from AVR32 build errors that have crept in
over the time; incorporated Kconfig for AVR3 (Richard Cochran).
* fs/fat and include/nuttx/fs/fat.h: The FAT file system allocates
memory for sector I/O buffers used to exchange data with the
configured block driver. In some contexts, the block driver may
require DMA-capable memory. If CONFIG_FAT_DMAMEMORY is defined,
then the FAT FS will use platform-provided DMA memory allocators
to allocate the block driver I/O buffers.
* CONFIG_NET_ENC28J60 renamed CONFIG_ENC28J60 to be consistent
in all places.
* drivers/enc28j60.c, include/nuttx/net/enc28j60.h, and
olimex-strp711/src/up_enc28j60.c: No longer passes IRQ number
as a parameter. Instead now passes a call table to manage
ENC28J60 GPIO interrupts. That is because GPIO interrupts are
handled in different ways by different MCUs and some do not
support IRQ numbers for GPIO interrupts.
* mm/mm_gran* and include/nuttx/gran.h: Add a simple granule-
based allocator. The intent of this allocator is to support
simple allocation of DMA I/O buffers. The initial check-in
is code complete but untested (not event built into the
mm/Makefile yet.
* confgs/fire-stm32v2: The board port is basically functional.
Not all features have been verified. The ENC28J60 network
is not yet functional.
* configs/stm3240g-eval/discover: A configuration for testing
the UDP discovery utility. Contributed by Max Holtzberg.
* mm/README.txt: Add a new README file.
patacongo
committed
* include/nuttx/usb/usb.h, arch/*/src/*usb.c, and arch/*/src/*otg*.c:
Add hooks to to use common, external DMA buffer allocation
implementation.
patacongo
committed
* net/recvfrom.c: Don't block in recvfrom if (1) read-ahead buffering
is enabled and (2) some data was obtained from read-ahead buffers.
Blocking is a bad idea in that case because there is no timeout!
(submitted by Max Holtzberg).
* configs/stm3240g-eval/xmlrpc: An example configuration for the
Embeddable Lightweight XML-RPC Server at apps/examples/xmlrpc.
See http://www.drdobbs.com/web-development/\
an-embeddable-lightweight-xml-rpc-server/184405364 for more info.
Contributed by Max Holtzberg.
* configs/*/nxwm/defconfig and sched/task_exithook.c: Fixes for
bugs that crept in during recent changes. (Submitted by Max
Holtzberg).
* arch/arm/include/armv7-m/irq.h: Fix a critical bug in irqsave().
It looks like sometimes the compile will re-order some instructions
inapproapriately. This end result is that interrupts will get
* drivers/mtd/w25.c: Beginning of a driver for the Windbond SPI
FLASH family (W25x16, W25x32, and W25x64). The initial check-in
is basically just the SST25 driver with some name changes.
* arch/arm/include/armv7-m/irq.h and arch/arm/src/stm32/stm32_spi.c:
Back out the last change in irq.h. It is (most likely) fine the
way it was. The really interrupt related problem was in stm32_spi.c:
When SPI3 is not enabled, then the irqrestore() falls in the
else clause.
* include/nuttx/compiler.h and other files: Moved always_inline
and noinline __attributes__ here. Also replaced all occurrences
of explicit __atributes__ in other files with definitions from
this header file.
* drivers/mtd/w25.c: The Windbond SPI FLASH W25 FLASH driver is
code complete (but still untested).
* arch/arm/src/stm32/stm32_i2c.c: I2C improvements from Mike Smith.
Unified configuration logic; dynamic timeout calculations;
I2C reset logic to recover from locked devices on the bus.
* configs/*/*/Make.defs, tools/Config.mk, Makefile: Refactor all
common make definitions from the various Make.defs files into
the common tools/Config.mk. Add support for a verbosity options:
Specify V=1 on the make command line in order to see the exact
commands used in the build (Contributed by Richard Cochran).
* drivers/net/enc28j60.c: The ENC28J60 Ethernet driver is
now functional.
* configs/fire-stm32v2: Add support or the fire-stm32v3 board as
well (untested because I do not have a v3 board).
* lib/stdio/lib_sscanf.c: Add %n psuedo-format (from Kate).
* lib/stdio/lib_sscanf.c: There is an issue of handling input
when (1) no fieldwidth is provided and (2) there is no space
seperating the input values. No solutions is in place for this
case now (either space or a fieldwidth must be provided). But
at least some of the bad logic that attempted to handle this
case has been removed (noted by Kate).
* arch/arm/src/stm32/stm32_eth.c: DMA buffer sizes must be an
even multiple of 4, 8, or 16 bytes.
* arch/arm/src/stm32/stm32_idle.c: Fixes STM32F107 DMA issues:
We cannot go into sleep mode while Ethernet is actively DMAing.
* configs/shenzhou/src/up_ssd1289.c: Add infrastructure to support
SSD1289 LCD. Initial checkin is just a clone of the
STM32F4Discovery's FSMC-based LCD interface. The Shenzhou
will need a completely need bit-banging interface; this
initial check-in is only for the framework.
* configs/shenzhou/src/up_ssd1289.c: Bit-banging driver is
code complete.
* configs/shenzhou/src/up_lcd.c: Oops. Shenzhou LCD does not
have an SSD1289 controller. Its an ILI93xx. Ported the
STM3240G-EVAL ILI93xx driver to work on the Shenzhou board.
* configs/shenzhou/nxwm: Added an NxWM configuration for the
Shenzhou board. This is untested on initial check-in. It will
be used to verify the Shenzhou LCD driver (and eventually the
touchscreen driver).
* configs/shenzhou/src/up_touchscreen.c: Add ADS7843E touchscreen
support for the Shenzhou board. The initial check-in is untested
and basically a clone of the the touchscreen support for the SAM-3U.
* tools/cfgparser.c: There are some NxWidget configuration
settings that must be de-quoted.
* arch/arm/src/stm32/Kconfig: There is no SPI4. Some platforms
* nuttx/configs/shenzhou: Various fixes to build new NxWM
configuration.
* configs/shenzhou: Oops. The Shenzhou LCD is and SSD1289,
not an ILI93xx.
* configs/shenzhou/src/up_ssd1289.c: The LCD is basically functional
on the Shenzhou board.
* graphics/nxmu: Correct some bad parameter checking that caused
failures when DEBUG was enabled.
patacongo
committed
* arch/arm/src/armv7-m/nvic.h: Add bit definitions for the AIRCR
register.
* drivers/input/ads7843.c: Need semaphore protection in logic
that samples the position.
* drivers/lcd/ssd1289.c: On some platforms we are unable to
read the device ID -- reason unknown; workaround in place.
* drivers/input/ads7843.c: Add thresholding options and an
option to swap X and Y positions. Fix some logic errors in
the SPI locking/selecting logic.
* arch/arm/src/armv7-m/up_systemreset.c: Add logic to reset
the Cortex-Mx using the AIRCR register. Contributed by Darcy
Gong.
* arch/arm/src/stm32/up_eth.c: Add logic specifically for the
DM9161 PHY. If the DM9161 failed to initialize, then use the
up_sysemreset() logic to reset the MCU. Contributed by Darcy
Gong.
* arch/arm/src/stm32/stm32_gpio.c: Add missing logic to set bit
for SPI3 remap. This fixes the XPT2046 touchscreen driver using
drivers/input/ads7843.c
patacongo
committed
* configs/shenzhou/src/up_ssd1289.c: Fix naming error in
conditional compilation.
* configs/shenzhou/nxwm/defconfig: Disable reading from the LCD.
This does not work. The hardware and the driver support the
capability, but there is some bug that causes memory corruption.
The work around for now: Just disable reading from the LCD.
* drivers/lcd/ssd1289.c: Add some logic to reduce the amount of
output when CONFIG_DEBUG_LCD is enabled.
* configs/shenzhou/nxwm/defconfig: Bug found and fixed... The
original configuration had too much stuff turned on. Reducing
stack sizes, some features, and buffer sizes made the
configuration reliable (Reading from the LCD is still disabled).
patacongo
committed
* net/uip/uip_icmpping.c: Fix problem that prevented ping from
going outside of local network. Submitted by Darcy Gong
6.23 2012-09-29 Gregory Nutt <gnutt@nuttx.org>
* arch/arm/src/stm32/stm32_rng.c, chip/stm32_rng.h, and other files:
Implementation of /dev/random using the STM32 Random Number
Generator (RNG).
patacongo
committed
* board.h file for shenzhou, fire-stm32v2, and olimex-stm32-p107:
Add frequencies for HSE, HSI, LSE, and LSI. These are needed
by the STM32 watchdog driver.
patacongo
committed
* CONFIG_EXAMPLES_*: To make things consistent, changed all occurrences
of CONFIG_EXAMPLE_* to CONFIG_EXAMPLES_*.
* drivers/mtd/w25.c and configs/*/src/up_w25.c: Several fixes for the
W25 SPI FLASH.
* configs/*/Make.defs: All buildroot tools now use the extension
xxx-nuttx-elf- vs. xxx-elf-
* configs/shenzhou/*/Make.defs: Now uses the new buildroot 4.6.3
EABI toolchain.
* lib/stdio/lib_libdtoa.c: Another dtoa() fix from Mike Smith.
* configs/shenzhou/src/up_adc.c: Add ADC support for the Shenzhou
board (Darcy Gong).
* configs/shenzhou/thttpd: Add a THTTPD configuration for the
Shenzhou board (Darcy Gong).
* include/termios.h and lib/termios/libcf*speed.c: The non-standard,
"hidden" c_speed cannot be type const or else static instantiations
of termios will be required to initialize it (Mike Smith).
* drivers/input/max11802.c/h, and include/nuttx/input max11802.h: Adds
support for the Maxim MAX11802 touchscreen controller (contributed by
Petteri Aimonen).
* graphics/nxtk/nxtk_events.c: Missing implementatin of the blocked
method. This is a critical bugfix for graphics support (contributed
by Petteri Aimonen).
* drivers/usbdev/pl2303.c, drivers/usbdev/usbmsc.h, and
include/nuttx/usb/cdcacm.h: USB_CONFIG_ATTR_SELFPOWER vs.
USB_CONFIG_ATT_SELFPOWER (contributed by Petteri Aimonen).
* arch/arm/src/armv7-m/up_memcpy.S: An optimized memcpy() function for
the ARMv7-M family contributed by Mike Smith.
* lib/strings/lib_vikmemcpy.c: As an option, the larger but faster
implemementation of memcpy from Daniel Vik is now available (this is
from http://www.danielvik.com/2010/02/fast-memcpy-in-c.html).
* lib/strings/lib_memset.c: CONFIG_MEMSET_OPTSPEED will select a
version of memset() optimized for speed. By default, memset() is
optimized for size.
* lib/strings/lib_memset.c: CONFIG_MEMSET_64BIT will perform 64-bit
aligned memset() operations.
* arch/arm/src/stm32/stm32_adc.c: Need to put the ADC back into the
initial reset in the open/setup logic. Opening the ADC driver works
the first time, but not the second because the device is left in a
powered down state on the last close.
* configs/olimex-lpc1766stck/scripts: Replace all of the identical
ld.script files with the common one in this directory.
* configs/stm3220g-eval/scripts: Replace all of the identical
ld.script files with the common one in this directory.
* configs/hymini-stm32v/scripts: Replace all of the identical
ld.script files with the common one in this directory.
* configs/lpcxpresso-lpc1768/scripts: Replace all of the identical
ld.script files with the common one in this directory.
* binfmt/elf.c, binfmt/libelf, include/elf.h, include/nuttx/elf.h: Add
basic framework for loadable ELF module support. The initial check-
in is non-functional and is simply the framework for ELF support.
* include/nuttx/binfmt.h, nxflat.h, elf.h, and symtab.h: Moved to
include/nuttx/binfmt/.
* arch/sim/src/up_elf.c and arch/x86/src/common/up_elf.c: Add
for ELF modules.
* arch/arm/include/elf.h: Added ARM ELF header file.
* include/elf32.h: Renamed elf.h to elf32.h.
* configs/stm32f4discovery/ostest: Converted to use the new
Kconfig-based configuration system.
* configs/stm32f4discovery/elf and configs/stm32f4discovery/scripts/gnu-elf.ld
Add a configuration for testing the ARM ELF loader.
* binfmt/libelf: Can't use fstat(). NuttX does not yet support it. Damn!
* binfmt/libelf: The basic ELF module execution appears fully functional.
* configs/shenzhou/src/up_relays.c: Add support for relays from the
Shenzhou board. Contributed by Darcy Gong.
* lib/fixedmath: Moved the old lib/math to lib/fixedmath to make room for
the math library from the Rhombus OS
* lib/math: Now contains the math library from the Rhombus OS by Nick Johnson
(submitted by Darcy Gong).
* include/float.h: Add a first cut at the float.h header file. This
really should be an architecture/toolchain-specific header file. It
is only used if CONFIG_ARCH_FLOAT_H is defined.
* lib/math: Files now conform to coding standards. Separated float,
double, and long double versions of code into separate files so that
they don't draw in so much un-necessary code when doing a dumb link.
* binfmt/libelf: The ELF loader is working correctly with C++ static
constructors and destructors and all.
* Documentation/NuttXBinfmt.html: Add documentionof the binary loader.
* configs/sim/ostest: Converted to use the mconfig configuration tool.
* configs/sim/cxxtest: New test that will be used to verify the uClibc++
port (eventually).
* include/nuttx/fs/fs.h, lib/stdio/lib_libfread.c, lib_ferror.c,
lib_feof.c, and lib_clearerr.c: Add support for ferror(), feof(),
and clearerror(). ferror() support is bogus at the moment (it
is equivalent to !feof()); the others should be good.
* configs/stm32f4discovery/include/board.h: Correct timer 2-7
base frequency (provided by Freddie Chopin).
* include/nuttx/sched.h, sched/atexit.c, and sched/task_deletehook.c:
If both atexit() and on_exit() are enabled, then implement atexit()
as just a special caseof on_exit(). This assumes that the ABI can
handle receipt of more call parameters than the receiving function
expects. That is usually the case if parameters are passed in
registers.
* libxx/libxx_cxa_atexit(): Implements __cxa_atexit()
* configs/stm32f4discovery/cxxtest: New test that will be used to
verify the uClibc++ port (eventually). The sim platform turned not
to be a good platform for testing uClibc++. The sim example will not
run because the simulator will attempt to execute the static
constructors before main() starts. BUT... NuttX is not initialized
and this results in a crash. On the STM324Discovery, I will have
better control over when the static constructors run.