forked from nuttx/nuttx-update
Updated changelogs and release notes in preparation for the 7.2 release. Still need to do a little more before the release
This commit is contained in:
parent
33ddbeab16
commit
7633031713
2 changed files with 293 additions and 7 deletions
15
ChangeLog
15
ChangeLog
|
@ -7003,7 +7003,7 @@
|
|||
* sched/nanosleep.c: Fix a missing call to re-enable interrupts. From
|
||||
Jason Jiang (2014-3-15).
|
||||
|
||||
7.2 2014-xx-xx Gregory Nutt <gnutt@nuttx.org>
|
||||
7.2 2014-04-29 Gregory Nutt <gnutt@nuttx.org>
|
||||
|
||||
* arch/arm/src/sam34/sam_udp.c and .h: Add a USB full speed device
|
||||
driver (UDP) for the SAM4E (2014-3-17).
|
||||
|
@ -7035,7 +7035,7 @@
|
|||
Add a new MTD driver that can be used to contain another driver and
|
||||
force its apparent sector size to be 512 bytes (2014-3-24).
|
||||
* arch/arm/src/sam34/sam_lowputc.c sam_serial.c: Fix a mysterious
|
||||
multithreading bug that can log up the serial port (2014-3-14).
|
||||
multithreading bug that can lock up the serial port (2014-3-14).
|
||||
* drivers/usbdev/Kconfig, usbmsc.c, usbmsc.h, and usbmsc_scsi.c:
|
||||
Redesign threading module used with the USB MSC driver. It was using
|
||||
pthreads before and these were changed to a kernel thread. The reason
|
||||
|
@ -7062,7 +7062,7 @@
|
|||
would have failed if n=0; one byte was always copying before testing
|
||||
for the end of the copy.
|
||||
* nuttx/configs/sama5d3-xplained/: Add a board support configuration
|
||||
for the Atmel SAMA5D3 Xplaiend board. The initial check in is not
|
||||
for the Atmel SAMA5D3 Xplained board. The initial check in is not
|
||||
finished. For example, it still builds to run out of NOR FLASH but
|
||||
the SAMA5D Xlpained has no NOR FLASH (2013-4-28).
|
||||
* arch/arm/src/sama5/sam_pmc.c and .h: Add functions to calculate
|
||||
|
@ -7146,11 +7146,11 @@
|
|||
I thought it was going to do. Result was the NOR FLASH accesses
|
||||
were very slow (2014-4-3).
|
||||
* arch/arm/src/sama5/sam_udphs.c: Fix a case where received status
|
||||
was not clear, causing OUT SETUP commands to fail (2014-4-6).
|
||||
was not being cleared, causing OUT SETUP commands to fail (2014-4-6).
|
||||
* arch/arm/src/tiva/tiva_syscontrol.c: Add logic for TM4C125GXL based
|
||||
on logic from Daniel Carvalho with modifications. I think the
|
||||
LM4F120 may have broken before as well(?). In any event, the LM4F120
|
||||
also works well with this chanage (2014-4-7).
|
||||
also works well with this change (2014-4-7).
|
||||
* Back out part of 610e2aa0c224e2936fe8009ef4a2351ce607067a: Need to
|
||||
be able to select USB console devices when there is no /dev/console
|
||||
enabled. Noted by Librae (2014-4-8).
|
||||
|
@ -7160,7 +7160,7 @@
|
|||
ramlog.h to include/nutt/syslog (2014-4-10).
|
||||
* sched/errno_getptr.c and getpid.c: Add some checks. If these
|
||||
functions are called early in initialization before the tasking
|
||||
structures are initialized, they will not behavr properly
|
||||
structures are initialized, they will not behave properly
|
||||
(2014-4-10).
|
||||
* arch/arm/src/sama5/sam_twi.c: TWI data sending is fails to increment
|
||||
the number of byes transferred on first byte sent. from David Sidrane
|
||||
|
@ -7188,7 +7188,7 @@
|
|||
introduced in the above change and is Sourceforge patch 39 from Luciano
|
||||
Neri (2014-4-17).
|
||||
* configs/sama5dx-ek/include/board.h and sama5d3-xplained.h: These files
|
||||
referenct type xcpt_t and so must include nuttx/irq.h. This is
|
||||
reference type xcpt_t and so must include nuttx/irq.h. This is
|
||||
Sourceforge bug 25 submitted by Luciano Neri (2014-4-17).
|
||||
* arch/arm/src/stm32/stm32_otgfshost.c: Some fixes related to low-speed
|
||||
devices and interrupt endpoints from Leo (2014-4-17).
|
||||
|
@ -7255,3 +7255,4 @@
|
|||
Ticket #26. up_prioritize_irq should not be called if
|
||||
CONFIG_ARCH_IRQPRIO is not defined (2014-4-25).
|
||||
|
||||
7.3 2014-xx-xx Gregory Nutt <gnutt@nuttx.org>
|
||||
|
|
285
ReleaseNotes
285
ReleaseNotes
|
@ -6065,3 +6065,288 @@ Bugfixes (see the ChangeLog for details). Some of these are very important:
|
|||
prevents the NSH prompt from be in the same line as the final line
|
||||
of the file in the case where there is no newline at the end of the
|
||||
file.
|
||||
|
||||
NuttX-7.2
|
||||
---------
|
||||
|
||||
The 102nd release of NuttX, Version 7.2, was made on April 29, 2014,
|
||||
and is available for download from the SourceForge website. Note
|
||||
that release consists of two tarballs: nuttx-7.2.tar.gz and
|
||||
apps-7.2.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:
|
||||
|
||||
- Add support for CLOCK_REALTIME. From Macs N.
|
||||
|
||||
* Network Utilities:
|
||||
|
||||
- Add a very primitive NTP client.
|
||||
- DNS resolver: Long needed major clean up for coding style and
|
||||
unification of naming conventions (resolv vs dns).
|
||||
|
||||
* Common Drivers:
|
||||
|
||||
- Add a new MTD driver that can be used to contain another driver and
|
||||
force its apparent sector size to be 512 bytes.
|
||||
- If DCD can support queuing of stall requests, then the USB MSC stall
|
||||
work around delays are not necessary.
|
||||
- Redesign threading module used with the USB MSC driver. It was using
|
||||
pthreads before and these were changed to a kernel thread. The reason
|
||||
for this has to do with task grouping: A pthread is a memory of the
|
||||
group of the task that started it. A kernel thread is independent of
|
||||
the task that started in (other than knowing it as the parent). This
|
||||
allows me to remove so kludge logic to "deparent" the pthread on
|
||||
startup.
|
||||
- Update the MIO183QT-9A LCD driver to support reading from the LCD.
|
||||
From Toby Duckworth.
|
||||
- Timer "upper half" driver and l"lower half" interface definition from
|
||||
Bob Doiron.
|
||||
|
||||
* Atmel SAM4E/4S Board Support:
|
||||
|
||||
- All SAM4E configurations updated to run with 120MHz CPU speed and
|
||||
with the CMCC enabled.
|
||||
- Support for the SAM4S Xplained Pro board from Bob Doiron.
|
||||
- Add watchdog driver. From Bob Doiron.
|
||||
|
||||
* Atmel SAM4E/4S Drivers:
|
||||
|
||||
- Add a USB full speed device driver (UDP) for the SAM4E.
|
||||
- Add support in HSMCI and SPI drivers to invalidate cached memory if
|
||||
the CMCC is enabled.
|
||||
- Add logic to handle SPI word widths > 8
|
||||
- Port RTC driver from SAMA5 to SAM3/4. From Bob Doiron.
|
||||
- HSMCI: Extensions from Bob Doiron needed to support PDC-based HSMCI
|
||||
DMA transfers. Verified on the SAM4S Xplained Pro.
|
||||
|
||||
* Atmel SAMA5:
|
||||
|
||||
- Add support for the SAMA5D36 chip.
|
||||
- Add functions to calculate PLLACK, PCK, and MCK frequencies given
|
||||
the main clock frequency.
|
||||
- Implemented RAM functions. When booting directly into NOR flash,
|
||||
reconfiguring the PLL while executing out of NOR FLASH causes
|
||||
crashes. This was fixed by David Sidrane by implementing RAM
|
||||
functions. The killer code is copied and executed from ISRAM and
|
||||
the crash is avoided.
|
||||
|
||||
* Atmel SAMA5 Drivers:
|
||||
|
||||
- Add support for the SAMA5D3 DBGU.
|
||||
|
||||
* Atmel SAMA5 Board Support:
|
||||
|
||||
- Add a board support configuration for the Atmel SAMA5D3 Xplained
|
||||
board running out of SDRAM. When booting from SDRAM, start-up logic
|
||||
needd to query the PMC registers to determine the MCK, PCK, etc.
|
||||
- For all SAMA5 boards, added support for 528MHz clocking.
|
||||
|
||||
* TI Tiva/Stellaris:
|
||||
|
||||
- Modify the logic to disables the interrupts and sets the default
|
||||
interrupt priority so that it uses the ICTR to get the number of
|
||||
interrupt lines/registers. This is instead of using some fixed
|
||||
number of initializations based a priori knowledge of the number
|
||||
of interrupt lines in the MCU.
|
||||
|
||||
* STMicro STM32
|
||||
|
||||
- Add support for the STM401RE. From Frank Bennett.
|
||||
|
||||
* STMicro STM32 Boards:
|
||||
|
||||
- STM32F4Discovery: Add support to register keyboard and mouse
|
||||
classes if so configured. From Leo.
|
||||
- Add support for the Nucleo-F401RE board. From Frank Bennett.
|
||||
|
||||
* STMicro STM32 Drivers:
|
||||
|
||||
- Add STM32 Ethernet IGMP hashing support. From Manuel Stühn.
|
||||
- OTG FS Host: USB host tracing added by Leo.
|
||||
|
||||
* Freescale Kinetis
|
||||
|
||||
- Modify the logic to disables the interrupts and sets the default
|
||||
interrupt priority so that it uses the ICTR to get the number of
|
||||
interrupt lines/registers. This is instead of using some fixed
|
||||
number of initializations based a priori knowledge of the number
|
||||
of interrupt lines in the MCU.
|
||||
|
||||
* NXP LPC17xx
|
||||
|
||||
- Modify the logic to disables the interrupts and sets the default
|
||||
interrupt priority so that it uses the ICTR to get the number of
|
||||
interrupt lines/registers. This is instead of using some fixed
|
||||
number of initializations based a priori knowledge of the number
|
||||
of interrupt lines in the MCU.
|
||||
|
||||
* ZiLOG ZNEO Drivers
|
||||
|
||||
- Added ZNEO ESPI driver.
|
||||
|
||||
* Library:
|
||||
|
||||
- Finish incomplete support for ferror(). From Macs N.
|
||||
|
||||
* Documentation:
|
||||
|
||||
- UsbTrace.html: Add some discussion of the USB monitor.
|
||||
|
||||
* Configuration/Build System:
|
||||
|
||||
- compiler.h: Add macro UNUSED() that can be used to eliminate
|
||||
warnings about variables that are set to values that are not
|
||||
used.
|
||||
|
||||
* Debugging Support:
|
||||
|
||||
- Add a configuration option to enable dumping of the USB trace buffer
|
||||
on assertion.
|
||||
|
||||
* Applications:
|
||||
|
||||
- IGMP Example: Fill out Kconfig file; update makefile so that the
|
||||
IGMP example can be used as an NSH built-in application. From Manuel
|
||||
Stühn.
|
||||
- lcdrw Example: Fill out empty Kconfig file.
|
||||
- touchscreen Example: Add a configuration option to indicate if
|
||||
architecture-specific initialized is required, yes or no.
|
||||
- System usbmsc: Back out the USBMSC daemon... it is no longer needed
|
||||
with the recent fix to the USBMSC class driver.
|
||||
- NSH: malloc/free IOBUFFER for 'cat' and 'hexdump' commands instead of
|
||||
using the stack. From Bob Doiron.
|
||||
- cpuhog, serialblaster, and serialrx Examples: Stress test examples
|
||||
added by Bob Doiron.
|
||||
|
||||
Bugfixes (see the ChangeLog for details). Some of these are very important:
|
||||
|
||||
* Core OS:
|
||||
|
||||
- errno_get_errno_ptr() and getpid(): Add some checks. If these
|
||||
functions are called very early in initialization before the tasking
|
||||
structures are initialized, they would not behave properly.
|
||||
|
||||
* Networking:
|
||||
|
||||
- IGMP-related bug fixes from Manuel Stühn.
|
||||
- In netdev_findbyaddr(), skip network devices that are in the "down"
|
||||
state. It does not make sense to consider an address match with a
|
||||
"down" device. From Brennan Ashton.
|
||||
|
||||
* Network Utilities:
|
||||
|
||||
- DHCPD: The correct value of the PAD option. Should be zero, not
|
||||
one. From Brennan Ashton.
|
||||
- Corrections to the NTP client from Manuel Stuehn.
|
||||
|
||||
* Common Drivers:
|
||||
|
||||
- usbhost.h: Fix some USB HID-related conditional compilation. From Leo.
|
||||
- USB HID mouse driver: Don't include NAKs from the device in the error
|
||||
count. From Leo.
|
||||
|
||||
* TI Tiva/Stellaris:
|
||||
|
||||
- Add corrected control logic for TM4C125GXL based on logic from Daniel
|
||||
Carvalho with modifications. The LM4F120 also works well with this
|
||||
change.
|
||||
|
||||
* STMicro STM32 Drivers:
|
||||
|
||||
- STM32 F2 SPI: Correct a mapping for SPI MOSI pin. From dlsitzer.
|
||||
- OTG FS Host: Fixes related to low-speed devices and interrupt
|
||||
endpoints from Leo. Necessary in order to use a USB HID mouse
|
||||
device.
|
||||
- STM32 SDIO: The dmaflight method must be conditioned on
|
||||
CONFIG_SDIO_PREFLIGHT. Noted by Pelle Windestam.
|
||||
- Sourceforge Ticket #26. up_prioritize_irq should not be called
|
||||
if CONFIG_ARCH_IRQPRIO is not defined.
|
||||
|
||||
* Atmel SAM3/4:
|
||||
|
||||
- Fix some backward conditional compilation in the start-up logic.
|
||||
- Fix initialization of the default interrupt priorities. This is
|
||||
a critical bug fix!
|
||||
- SAM4S: Fix error in macros that disable peripheral clocking. From
|
||||
Bob Doiron.
|
||||
|
||||
* Atmel SAMA5:
|
||||
|
||||
- Fix double mapping of SDRAM when executing out of SDRAM. In this
|
||||
case, the SDRAM was already mapping in arm_head.S.
|
||||
- In clock configuration, the BMS bit handling was fixed to match
|
||||
what the HW really does. From David Sidrane.
|
||||
- TLB Bufferable bit did not do what I thought it was going to do.
|
||||
I though I was selecting a write-through cache, but the result was
|
||||
that all caching was disabled and NOR FLASH accesses were very slow.
|
||||
- Move the un-definitions of __ramfuncs__ from the clock configuration
|
||||
.c file to the .h file. Otherwise, the attribute will be applied
|
||||
differently for the prototype and the function definition.
|
||||
Sourceforge patch 38 from Luciano Neri.
|
||||
|
||||
* Atmel SAMA5 Drivers:
|
||||
|
||||
- Fix a case in the UDPHS driver where received status was not being
|
||||
cleared, causing OUT SETUP commands to fail.
|
||||
- TWI data sending fails to increment the number of byes transferred
|
||||
on first byte sent. from David Sidrane.
|
||||
- If running from SDRAM, then BOARD_MCK_FREQUENCY is not a constant
|
||||
and cannot be used in conditional compilation. All drivers fixed
|
||||
EXCEPT for the PWM and Timer/Counter drivers which will need some
|
||||
more extensive redesign.
|
||||
|
||||
* Atmel SAMA5 Boards:
|
||||
|
||||
- SAMA5D3x-EK: The red LED is controlled by PE24 which is also the
|
||||
camera/ISI interface reset line. So if the a camera is installed,
|
||||
then we must avoid controlling the red LED or we inadvertently
|
||||
also reset the camera module. Noted by David Sidrane.
|
||||
- Header files reference type xcpt_t in certain configurations and
|
||||
so must include nuttx/irq.h. This is Sourceforge bug 25 submitted
|
||||
by Luciano Neri.
|
||||
|
||||
* NXP LPC17xx Drivers
|
||||
|
||||
- arm/src/lpc17xx/Kconfig and lpc17_ethernet.c: The default interrupt
|
||||
priority should be default, not the highest priority (2014-4-19).
|
||||
|
||||
* ZiLOG ZNEO
|
||||
|
||||
- Fix calculation of timer reload and prescaler. The timer frequency
|
||||
was way too fast.
|
||||
|
||||
* Library:
|
||||
|
||||
- strncpy(): Change ordering of test and copy. Logic would have
|
||||
failed if n=0; one byte was always copying before testing
|
||||
for the end of the copy.
|
||||
- Fix a counting error in the return value from sscanf(). Noted by
|
||||
kfrolov. Also, sscanf() should return EOF if no values were
|
||||
converted.
|
||||
- Fix a logic error in ftell(). It was simply using the file offset
|
||||
and did not take into account data buffered in memory. From Macs N.
|
||||
- sccanf(): Use stroul() vs strol() with %u format otherwise, range
|
||||
of values is restricted because of sign bit. From kfrolov.
|
||||
|
||||
* Configuration/Build System:
|
||||
|
||||
- Fix a build error that occurs when only USB device tracing is enabled.
|
||||
|
||||
* NuttShell (NSH):
|
||||
|
||||
- Fix NuttShell version number display in the NSH welcome message.
|
||||
version.h was not being included.
|
||||
|
||||
* Other Applications:
|
||||
|
||||
- sysinfo: CONFIG_VERSION_BUILD is a string an needs to be
|
||||
printed with %s, not %d. Noted by Librae.
|
||||
- nxplayer: Compilation failure in one configuration reported by
|
||||
Manuel Stuhn.
|
||||
- System sdcard: Remove an STM32 dependency. From Bob Doiron.
|
||||
- telnetd Example: Naming is confused. In someplaces 'telnetd',
|
||||
and in others 'shell.' All changes to telnetd. Noted by Pelle
|
||||
Windestam.
|
||||
|
|
Loading…
Reference in a new issue