This commit repartitions the logic by moving some of the changes from mm_sem.c into task_getpid.c. The logic is equivalent for the case of mm_trysemaphore(), but no has wider impact since it potentially affects all callers of getpid(). Hence, this change may also introduce some other issues that will need to be addressed.
net/sixlowpan: Record and restore d_buf on entry/exit of sixlowpan_input. This avoids an issue where d_buf gets set to NULL by sixlowpan, and then is used by forwarding logic from another network interface before getting set back to the drivers internal b
Approved-by: GregoryN <gnutt@nuttx.org>
Fix lc823450 start
* arch/arm/src/lc823450: Use CONFIG_STACK_COLORATION instead of CONFIG_DEBUG_STACK
Signed-off-by: Masayuki Ishikawa <Masayuki.Ishikawa@jp.sony.com>
* arch/arm/src/lc823450: Fix go_os_start() in lc823450_start.c
I found a bug in go_os_start() that it returns immediately
because it does not compare r1 to zero. This commit fixes
this bug. Also, this commit fixes its description.
Signed-off-by: Masayuki Ishikawa <Masayuki.Ishikawa@jp.sony.com>
Approved-by: GregoryN <gnutt@nuttx.org>
1. Configurable option to use named config items instead of enumerated ID/Instance numbers.
2. Ability to iterate through the existing configdata items in the /dev/config device.
3. Ability to "unset" a configdata item.
4. Ability to perform "flash_eraseall" on the /dev/config device.
This change adds a check to mm_trysemaphore() (the root implementation of both kmm_trysemaphore() and umm_trysemaphore()). It checks if the that task that is apparently executing is marked as RUNNING. If not, how could the non-running task be trying to get the MM semaphore? I think only in the exact scenario that Eunbong Song has described.
So I think the solution should provide the same protection as 91aa26774b but without the horrific consequences to memory usage.
This solution to the problem noted by EunBong Song results in major memory fragmentation and and out-of-memory conditions on the PX4 platform. On that platform the lower priority work queue is very low priority and essentially never runs when the system is busy. As a result, the systems gets slowly starved of memory until failures and bad behaviors begin to occur.
This is an addition patch coming later to result the original problem in a different way that does not have cause memory starvation.
This reverts commit 91aa26774b.
Squashed commit of the following:
arch/arm/src/stm32f0l0: Various changes for a clean compilation. Still does not compile correctly due to missing FLASH latency definitions.
arch/arm/src/stm32f0l0/hardware: Add framework for the STM32 L0. Currently set to same as the STM32F0.
arch/arm/src/stm32f0l0/hardware: Very fragmentary FLASH header register definitions for the STM32 L0.
arch/arm/src/stm32f0l0: Bring in DMA v1. Cannot possibly be functionaly yet due to the limited number for M0 interrupts.
arch/arm/src/stm32f0l0: Add STM32 F0/L0 LSE and backup power domain controls.
arch/arm/src/stm32f0l0/hardware/stm32l0_pwr.h: Add STM32L0 PWR header file.
arch/arm/include/stm32f0l0/chip.h: Clean up WIP chip header file.
arch/arm/include/stm32f0l0/chip.h: WIP.
arm/src/stm32f0l0: Resolve some small differences between F0 and L0 GPIO pin options.
arch/arm/src/stm32f0l0: Better integrate STM32L0 header files.
nuttx/arch/arm/include/stm32f0l0: Add STM32L0 IRQ number definition file.
arch/arm/src/stm32f0l0: Add STM32L0 RCC driver.
arch/arm/src/stm32f0l0/hardware: Adds basic STM32L0 header files.
arch/arm/src/stm32f0l0: Add STM32L0 chip selections.
configs/: Hook new STM32L0 boards into the configuration system.
configs: nucleo boards use as default ST LINK MCO as clock input from MCU and for this HSEBYP must be enabled
configs: add basic support for nucleo-l073rz
configs: add basic support for b-l072z-lrwan1
- Avoids the use of up_aesinitialize() entirely, which resolves dependency problems, because this function does not make sure that an actual hardware aes implementation was made available: each SoC is now responsible to ensure the AES hardware is initialized before first use. This applies to lpc43xx, stm32 and sam34.
- Remove definitions of the NEVER used aes_init and aes_update operations. The new AES API will be more suitable.
- Change the unusual naming in stm32 (avoiding possible naming clashes)
- Change the unusual naming in sam34 (avoiding possible naming clashes)
- Add some FAR to pointers and enforce the 80 col limit in stm32 and sam
Change rndis mac
* configs/lc823450-xgevk: Change RNDIS MAC address assignment
In previous implementation, mac[0] was assigned to 0xaa for RNDIS
host to avoid MAC address conflicts with RNDIS device.
However, I noticed that this assignment causes a random MAC address
generation on ubuntu16.04 or later which is inconvenient to set up
network interface.
This new assignment scheme fixes this issue.
Signed-off-by: Masayuki Ishikawa <Masayuki.Ishikawa@jp.sony.com>
* configs/stm32f4discovery: Change RNDIS MAC address assignment
In previous implementation, mac[0] was assigned to 0xaa for RNDIS
host to avoid MAC address conflicts with RNDIS device.
However, I noticed that this assignment causes a random MAC address
generation on ubuntu16.04 or later which is inconvenient to set up
network interface.
This new assignment scheme fixes this issue.
Signed-off-by: Masayuki Ishikawa <Masayuki.Ishikawa@jp.sony.com>
* configs/viewtool-stm32f107: Change RNDIS MAC address assignment
In previous implementation, mac[0] was assigned to 0xaa for RNDIS
host to avoid MAC address conflicts with RNDIS device.
However, I noticed that this assignment causes a random MAC address
generation on ubuntu16.04 or later which is inconvenient to set up
network interface.
This new assignment scheme fixes this issue.
Signed-off-by: Masayuki Ishikawa <Masayuki.Ishikawa@jp.sony.com>
Approved-by: GregoryN <gnutt@nuttx.org>
sixlowpan: Fixes logic surrounding the Universal/Local bit. This bit represents whether the IID is locally/globally administered. The U/L bit is bit 1 of the MSB of the EUI-64. It should only be inverted in cases where there is a full EUI-64. In cases whe
* sixlowpan: Fixes logic surrounding the Universal/Local bit. This bit represents whether the IID is locally/globally administered. The U/L bit is bit 1 of the MSB of the EUI-64. It should only be inverted in cases where there is a full EUI-64. In cases where the IID is derived from say, a short address, this bit should be forced to 0, indicating that it is locally administered.
See:
https://tools.ietf.org/html/rfc4291#section-2.5.1https://tools.ietf.org/html/rfc4944#section-6https://tools.ietf.org/html/rfc2464#section-4
* sixlowpan: Account for endianness with U/L bit.
Approved-by: GregoryN <gnutt@nuttx.org>
Fix lc823450 related
* configs/lc823450-xgevk: Fix IOB params in rndis/defconfig
These prameters work for HTTP audio streaming.
Signed-off-by: Masayuki Ishikawa <Masayuki.Ishikawa@jp.sony.com>
* arch/arm/src/lc823450: Fix up_allocate_heap() in lc823450_allocateheap2.c
This change fixes heap size and also implements up_addregion().
Signed-off-by: Masayuki Ishikawa <Masayuki.Ishikawa@jp.sony.com>
Approved-by: GregoryN <gnutt@nuttx.org>
net/sixlowpan: Fixes decompression of ipaddr from MAC address. The logic used to populate the IP from the radio address should match sixlowpan_ipfromsaddr/sixlowpan_ipfromeaddr
Approved-by: GregoryN <gnutt@nuttx.org>
