Newer
Older
* arch/arm/src/kinetis: Added header files defining all Kinetis registers
and bit fields within all Kinetis registers.
* configs/twr-k60n512: Add support for the Kinetis K60 Tower board
(TWR-K60N512).
* drivers/can.c: Fixed a semaphore overflow problem in the CAN driver
(reported by Li Zhouy (Lzzy)).
* 8/18/2011: The basic port to the FreeScale Kinetics TWR-K60N512 board is
now functional.
* confgs/twr-k60n512: Add Kinetics TWR-K60N512 NSH configuration.
* drivers/analog and include/nuttx/analog: Add ADC driver infrastructure
and TI ADS1255 driver developed and submitted by Li Zhouy (Lzzy)).
* arch/arm/stm32/stm32_sdio.h and drivers/mmcsd/mmcsd.c: Add logic to
multiplex usage of the GPIO pins (contributed by Uros Platise).
* configs/twr-k60n512/nsh: Added and verified a NuttShell (NSH)
configuration for the Freescale, Kinetis TWR-K60N512 board.
* arch/arm/src/kinetis/kinetis_sdhc.c: SDHC driver for Kinetis parts.
Initially check-in is just a crude port of the STM32 SDIO driver.
* graphics/, include/nuttx/nx: Add new NX interfaces for drawing
circles -- both circular outlines and filled circles.
* graphic/nxglib/nxglib_spitline.c: Add a "fudge factor" that eliminates
some problems for rendering nearly horizontal, wide lines. Hmm...
but I suspect this fudge factor also leads to new problems rendering
very thin, nearly horizontal lines. More tuning is needed.
* drivers/analog, include/nuttx/analog, arch/arch/src/lpcxx: (1) Add
updates to the ADS1255 driver, (2) fix errors from my last merge (sorry),
(3) Add DAC infrastructure, (4) add AD5410 DAC driver, and (5) add
LPC17xx ADC and DAC drivers. All contributed by Li Zhuoyi (Lzyy).
* tools/mkexport.sh: Extended the script that implements the top-level
'make export' logic. The script now also finds and bundles up all of
the architecture-specific header files as well.
* drivers/arch/arm/src/stm32/stm32_i2c.c: Add a reset to the I2C
initialization logic to prevent spurious interrupts when the I2C
interrupts are enabled (submitted by Uros Platise).
* Scripts/makefiles/documents. Several adjustments, corrections and
typo fixes so that NuttX will build correctly on FreeBSD using the
ASH shell (submitted by Kurt Lidl).
* drivers/mtd/flash_eraseall.c: Add a callable function that accepts
the path to a block driver and then erases the underlying FLASH memory
(assuming that the block driver is an MTD driver wrapped in the FTL
layer). Hmmm... this is probably not the best long term solution;
flash_eraseall() should be a user-callable function that operates
one driver interfaces; not an internal, OS function that operates
on directly on block drivers.
* drivers/bch: Fixed some important bugs in the BCH driver (noted by
Li Zhuoyi (Lzyy)). This would have effected any large reads or writes
(larger than the hardware sector size).
* arch/*/src/Makefile: Use of -print-libgcc-file-name to get path to
libgcc.a may select the wrong libgcc.a if a multilib toolchain (like
CodeSourcery) is used. This can be a serious problem and can cause
crashes on Cortex-M3 if the ARM libgcc is used, for example. The fix
is to include ARCHCPUFLAGS on the gcc command line when asking it to
-print-libgcc-file-name.
* lib/time/lib_gmtimer.c: Correct several calculations that could lead
to errors in dates.
* drivers/pm: Add the beginnings of a NuttX power management sub-system.
* arch/arm/src/stm32/stm32_irq.c: Fix a error introduced in 6.8.
Timeout calculation uses clock_settime() instead of clock_gettime().
Pretty gross error, but actually it works with the side effect of setting
a bad time.
* drivers/mtd/at24xx.c: Driver for I2C-based at24cxx EEPROM submitted by
Li Zhuoyi (Lzyy).
* arch/arm/src/lpc17xx/lpc17_i2c.c: I2C driver for the NXP LPC17xx family
submitted by Li Zhuoyi (Lzyy)
* arch/arm/src/stm32_i2c.c: Correct two issues with the STM32 I2C driver:
(1) Clocking needs to be based on PCLK1, not HCLK and fast speed settings
need some additional bits; and (2) Correct a hang that will occur on
I2C1 if FSMC is also enabled.
* drivers/sensors/lm75.c and include/nuttx/sensors/lm75.h: Add an LM-75
temperature sensor driver.
* configs/stm3210e-eval/src/up_lm75.c: Add support for the LM-75 on the
STMicro STM3210E-EVAL board.
* sched/clock_gettime.c: Correct an error in the tv_nsec calculation
that happens only config CONFIG_RTC is enabled.
* arch/arm/src/stm32/stm32_i2c.c: Correct some bugs related to waiting
6.10 2011-xx-xx Gregory Nutt <spudmonkey@racsa.co.cr>
* lib/stdio/lib_fopen.c: Fix an error in fopen(); the file pointer was not
being positioned at the end of the file when the "a" and "a+" modes are
used. There are other issues with the "a+" modes (see the top-level TODO
list).
* drivers/usbdev/cdc_serial.c and include/nuttx/usb/cdc.h and cdc_serial.h:
Add support for the CDC ACM serial device class.
* fs/fat/fs_fat32.c: Fix a critical bug in the write logic: It a tiny write
cross a sector boundary, then two sector writes will occur. The first part
in the first sector may be written to the wrong sector number.
* fs/fat/fs_fat32util.c: Fix a stray write into the FAT (always sector 964 on
FAT32). This bug will cause some lost chains. I'm sure this bug could
corrupt files but at present, the only thing I have seen is that before
fixing this bug, the Windows chkdsk utility would report these lost chains.
patacongo
committed
* arch/arm/src/stm32/stm32_i2c.c: Driver can now operate in a faster polled
mode (at the expense of using more cpu cycles).
patacongo
committed
* arch/arm/src/stm32/stm32_i2c.c: Add trace debug capability. Enabled with
CONFIG_I2C_TRACE.
* arch/arm/src/stm32/stm32_i2c.c: Fix another bug where I2C conflicts with FSMC
being enabled. That time at the tail end of the transaction where there is
an unfinished stop condition.
* sched/mq_timedreceive.c and sched/mq_timedsend.c: The count of threads
waiting on the message queues was not being decremented after a timeout.
This would cause the accounting logic to become out of sync and the, perhaps,
an assertion to be triggered. This is an important bug and fixes a
potential crash when using mq_timedreceived() and mq_timedsend().
* sched/mq_sndinternal.c: Related to the above, the send logic was incrementing
the wrong counter when it waited as part of the mq_timedsend.c logic.
* fs/fat: Fix an error in the long file name logic: If the long file name
is an even multiple of 13 bytes in length, then it should not include a
NULL terminating character. Fix contributed by Kaushal Parikh.