Not canceling the I2C timeout watch dog immediately after finishing
all I2C transactions in interrupt context can lead to a race condition
due to nxsem_wait(&priv->wait) in lpc17_40_i2c_start() not resuming
execution fast enough (this can be easily triggered if another task /
thread is using a lot of cpu time).
Falling to cancel the watchdog up to time will cause the priv->wait
semaphore to be incremented twice (first by lpc17_40_i2c_stop() then
by lpc17_40_i2c_timeout()), so all I2C transactions after that will
return immediately and priv->msgs will hold pointers to memory it
doesn't own anymore.
Canceling the priv->timeout watch dog in lpc17_40_i2c_stop() prevents
this as it is executed from the I2C interrupt handler.
arch/arm/src/lpc17xx_40xx/lpc17_40_i2c.c: Fix timeout calculation
For each byte received / transmitted, an acknowledge bit is also
transmitted / received, requiring effectively 9 bits for each byte.
* imxrt: GPIO make tables const
* imxrt: Call out to board to set up FlexRAM
* imxrt: Add Knob for adding the ROM bootloaders 40Kib of RAM to heap
* imxrt: imxrt1060-evk:Add the ability to run from OCRAM
* arch/arm/src/imxrt/imxrt_usbdev.c: Add USB Device support for i.MX RT (USB OTG1)
Based on the LPC43xx USB Device driver.
* imxrt:usbotg Nxstyle fixes
Co-authored-by: thomasactia <61285689+thomasactia@users.noreply.github.com>
There are cases that USE_SERIALDRIVER is defined but USE_EARLYSERIALINIT not defined in some configs. So use ifdef USE_EARLYSERIALINIT to include up_earlyserialinit anyway.
An error was introduced from:
commit f982ee43db683b7e34b154f5bc8c60a58b956820
Author: Xiang Xiao <xiaoxiang@xiaomi.com>
Date: Tue Feb 18 09:55:04 2020 +0800
drivers/serial: Remove the lowconsole driver
Replace with the syslog console driver which has more capability than lowconsole
* arch/arm/src/tiva/Make.defs: Compile only when enabled by configuration CONFIG_ARM_SEMIHOSTING_HOSTFS
* arch/arm/src/common/up_hostfs.c: Remove the ifdef conditional because it's redundant with the make logic.
Largely copy-and-paste from stm32f4discovery.
Also arch/arm/src/armv7-m/mpu.h: Ensure RBAR alignment
Fix crashes on init task startup I observed on qemu-system-arm -M lm3s6965evb.
And remove syslog_init_e because all initialization is later now and we don't
distinguish the initialition phase anymore after ramlog don't need special
initialize.
Because we can get the same function by CONSOLE_SYSLOG/syslog_console_init.
BTW, it isn't a good choice to use g_ramlogfops as /dev/console since nsh
will read back what it send out which will surprise most people.
it doesn't make sense that iob initialization is in up_initialize
but other memory components initialization is called in nx_start
Change-Id: Id43aeaa995f340c5943f59a0067a483ff3ac34a2
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
To ensure size_t same as toolchain definition in the first place and rename CXX_NEWLONG to ARCH_SIZET_LONG. The change also check whether __SIZE_TYPE__ exist before CONFIG_ARCH_SIZET_LONG so our definition can align with toolchain(gcc/clang) definition automatically.
Currently up_allocate_heap() assumes that g_idle_topstack points
top of the heap memory. However, g_idle_topstack pointed incorrect
address in SMP mode which resulted in heap corruption. This PR
moves g_idle_topstack at the end of .noinit to avoid this issue.
Signed-off-by: Masayuki Ishikawa <Masayuki.Ishikawa@jp.sony.com>
