Call xxx_timer_initialize from clock subsystem to make timer ready for use as soon as possiblei and revert the workaround:
commit 0863e771a9
Author: Gregory Nutt <gnutt@nuttx.org>
Date: Fri Apr 26 07:24:57 2019 -0600
Revert "sched/clock/clock_initialize.c: clock_inittime() needs to be done with CONFIG_SCHED_TICKLESS and clock_initialize should skip clock_inittime() for external RTC case since the RTC isn't ready yet."
This reverts commit 2bc709d4b9.
Commit 2bc709d4b9 was intended to handle the case where up_timer_gettime may not start from zero case. However, this change has the side-effect of breaking every implementation of tickless mode: After this change the tickless timer structures are used before they are initialized in clock_inittime(). Initialization happens later when up_initialize is called() when arm_timer_initialize().
Since the tickless mode timer is very special, one solution might be to
1. Rename xxx_timer_initialize to up_timer_initialize
2 Move up_timer_initialize to include/nuttx/arch.h
3. Call it from clock subsystem instead up_initialize
Basically, this change make timer initialization almost same as rtc initialization(up_rtc_initialize).
For now, however, we just need to revert the change.
* Serial Fixed interrupt storm
The target would randomly hang in the serial isr.
The priv->ie and the hardware were inconsistent.
The isr used the priv->ie to gate offloading
the RX data. Bang! Hung.
imxrt_disableuartint(priv, &ie);
ret = imxrt_setup(dev);
/* Restore the interrupt state */
imxrt_restoreuartint(priv, ie);
interrupt-> Of no return
priv->ie = ie;
On a fast cpu with FIFO, this will not work
with out proper protections.
* Serial: Conditionally enable 9 bit mode
* armv7-mi/mpu.hi: Restructure API
Preserve the existing API and enabled better granualriy on
setting.
* Enable MPU for non protected builds to set cache
* mpuinit use symbolic values for addresses
* Allow DTCM on HEAP
* allocateheap Fix Coding style
Author: Xiang Xiao <xiaoxiang@xiaomi.com>
Fix warning: selects SDIO_DMA or SDIO_BLOCKSETUP which has unmet direct dependencies (MMCSD)
Fix warning: selects NETDEV_TELNET which has unmet direct dependencies (NETDEVICES && NET && NET_TCP)
Fix warning: selects ARCH_PHY_INTERRUPT which has unmet direct dependencies (NETDEVICES && ARCH_HAVE_PHY)
Fix warning: selects USBHOST_HAVE_ASYNCH which has unmet direct dependencies (USBHOST)
Signed-off-by: Xiang Xiao <xiaoxiang@xiaomi.com>
Author: Alan Carvalho de Assis <acassis@gmail.com>
Run nxstyle to check .c and .h files and fix reported issues
Author: liuzhao <happypapa@yeah.net>
Add Quectel EC20 4G LTE Module USB CDC/ACM support
Run all .c and .h files in last PR through nxstyle.
Author: 脚蹬破拖鞋 <happypapa@yeah.net>
Add Imxrt usdhc insert or remove detection (#43)
* Add usdhc card insertion and removal detection using interrupt for imxrt.
* Add interrupt handle when usdhc insert or remove.
Author: patacongo <spudarnia@yahoo.com>
arch/arm/src/imxrt/imxrt_enet.c: Fix some warnings found in build testing.
Also fix coding standard problems reported by nxstyle. (#26)
Co-authored-by: Gregory Nutt <gnutt@nuttx.org>
* Simplify EINTR/ECANCEL error handling
1. Add semaphore uninterruptible wait function
2 .Replace semaphore wait loop with a single uninterruptible wait
3. Replace all sem_xxx to nxsem_xxx
* Unify the void cast usage
1. Remove void cast for function because many place ignore the returned value witout cast
2. Replace void cast for variable with UNUSED macro
Squashed commit of the following:
Author: Gregory Nutt <gnutt@nuttx.org>
Ran nxstyle against many of the affected files. But this job was too big for today. Many of the network drivers under arch are highly non-compiant and generate many, many faults from nxstyle. Those will have to be visited again another day.
Author: Xiang Xiao <xiaoxiang@xiaomi.com>
This effects all network drivers as well as timing related portions of net/: devif_poll_tcp_timer shouldn't be skipped in the multiple card case. devif_timer will be called multiple time in one period if the multiple card exist, the elapsed time calculated for the first callback is right, but the flowing callback in the same period is wrong(very short) because the global variable g_polltimer is used in the calculation. So let's pass the delay time to devif_timer and remove g_polltimer.
I did suspect that it was just about possible that there's a use for this check (specifically, when you're changing serial the clock immediately after the character is sent) but since we're only testing that the character has left the holding register and not that it's actually been sent to line I don't think it's valid even for this case.
1) Operational issue
TC (Transmission complete) and TDRE (TX Buffer Empty) were transposed in imxrt_serial.c. The end result was that for unoptimised code everything worked fine, but optimised code got itself into a real mess and continually fired interrupts. The patch attached fixes this. This one would have been found much more quickly if this particular board had supported SWO :-/
2) Startup issue
There are a number of chip errata that apply to the 1052 first revision (A-suffix) that don't apply to the second (B-suffix). Those got me for a while and it's important to use an EVK_B_ dev board if you're suffering stability problems with this particular chip. However, even with that resolved with either optimised or unoptimised code when there is no SNVS (Battery Backup) power and the power is switched on the CPU appears to boot but gets stuck with timer interrupts not being generated. The CPU is running and it will execute linear code. I have determined this by putting an 'imxrt_lowputc('A'+irq)' into up_doirq. For the non-running case the output looks like this;
LLLL
this implies 4 0x0b interrupts have been generated, and nothing else. If I hit keys on the keyboard I get 'e' in the flow. 0x0b is the SVC instruction and is the mechanism by which NuttX handles task switching. 'e' is 0x24, which corresponds to the LPUART1 interrupt. Other than these, the system does not respond further but is happily in the idle loop. In this circumstance if you hit SW2 on the EVKB board though the logjam 'unjams' and normal service is established, output now looks like;
LLLLPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPPLPP ...etc.
Where 'P' is 0x0f (the SYSTICK interrupt)....this is what _should_ be happening so; SYSTICK starts up after SW2 has been pressed.
From this point on everything works correctly and you can reboot the CPU, put new code into it or do whatever you wish, **provided power is not cycled off**. If power is cycled off then return to the top and go through the process again.
When there is SNVS power to the CPU then power can be switched on and off as you wish, and the CPU will boot correctly. This has been determined by putting 3V onto J6.
I think we are probably doing something naughty with the way we are starting the clocks to the timers. There are certainly some restrictions on imxrt clock manipulation which we have studiously avoided. I am investigating further and have some potential patches but even if I find the root cause it should not be included in 8.2 as the change could be disruptive.
https://github.com/zephyrproject-rtos/zephyr/pull/8535/commits: "The imxrt1050 is configured to use SYSTICK for the kernel timer, but SYSTICK cannot wake up the soc from low-power modes. Disable low-power modes on this soc until we have support for an alternative timer."
"This fixes k_sleep on the EVKB version of the mimxrt1050_evk board. An earlier version of the board (EVK, not EVKB), had A0 silicon which by default did not enter low-power mode on a wfi."
This patch reverts the WFI enable for further investigation post 8.2. With this patch in place together with the previous one my system is fully stable(*), just drinking a bit more power than I'd like.
(1) unifies the iomux definitions - previously some pins had them, and some didn't. This effectively made it impossible to use the pinmuxes without editing the header file in the standard distribution tree.
(2) unifies the pin definitions so that every pin now has a suffix. This makes it *much* easier to see when a pin is in use in your code, because it will always have a definition in your board.h file.
For anyone who is already using this CPU, a couple of small changes are needed to existing code;
In respect of (1) add IOMUX decorators to your pin definitions. You will find defaults in haardware/imxrt_iomux.h. Every pin should have a IOMUX decorator and in general the defaults should be OK. So while previously in your board.h file you might have had;
You will now have;
In respect of (2) you will need to add selectors for any function that previously only had one pinning option. There aren't many of those but LPUART1 is a good example. That just means adding into your board.h something like;
Most of the bluetooth and wifi chips appear to need external firmware, and the 43438 is no exception. Fortunately, since Cypress got involved, these are much more straightforward to obtain and are shipped as part of their SDK, which is downloadable from their website. Those firmwares are already provided as C arrays, so their names just need updating to;
const unsigned char bt_firmware_hcd -> The bt firmware array.
const int bt_firmware_len = sizeof(bt_firmware_hcd);
Testing of the WiFi is minimal so far but functionality is proven. I'm specifically not happy that the driver doesn't recover elegantly from a DMA data checksum failure, but that is an issue that can be dealt with in due course ... I'm trying to get the rest of the interfaces fleshed out and the hardware proven so it can go for pre-production build. I _think_ there's only Bluetooth and USB-device left to implement now.
libs/: Remove references to CONFIG_DISABLE_SIGNALS. Signals can no longer be disabled.
syscall/: Remove references to CONFIG_DISABLE_SIGNALS. Signals can no longer be disabled.
wireless/: Remove references to CONFIG_DISABLE_SIGNALS. Signals can no longer be disabled.
Documentation/: Remove references to CONFIG_DISABLE_SIGNALS. Signals can no longer be disabled.
include/: Remove references to CONFIG_DISABLE_SIGNALS. Signals can no longer be disabled.
drivers/: Remove references to CONFIG_DISABLE_SIGNALS. Signals can no longer be disabled.
sched/: Remove references to CONFIG_DISABLE_SIGNALS. Signals can no longer be disabled.
configs: Remove references to CONFIG_DISABLE_SIGNALS. Signals can no longer be disabled.
arch/xtensa: Remove references to CONFIG_DISABLE_SIGNALS. Signals can no longer be disabled.
arch/z80: Remove references to CONFIG_DISABLE_SIGNALS. Signals can no longer be disabled.
arch/x86: Remove references to CONFIG_DISABLE_SIGNALS. Signals can no longer be disabled.
arch/renesas and arch/risc-v: Remove references to CONFIG_DISABLE_SIGNALS. Signals can no longer be disabled.
arch/or1k: Remove all references to CONFIG_DISABLE_SIGNALS. Signals are always enabled.
arch/misoc: Remove all references to CONFIG_DISABLE_SIGNALS. Signals are always enabled.
arch/mips: Remove all references to CONFIG_DISABLE_SIGNALS. Signals are always enabled.
arch/avr: Remove all references to CONFIG_DISABLE_SIGNALS. Signals are always enabled.
arch/arm: Remove all references to CONFIG_DISABLE_SIGNALS. Signals are always enabled.
Author: Gregory Nutt <gnutt@nuttx.org>
arch/arm/src/imxrt/: Fix complaints from tools/nxstyle in new LCD files.
Use tools/lowhex to convert hexadecimal constants to lower case.
Use tools/rmcr to review white space at the end of lines.
Author: Johannes <nivus.entwicklung@gmail.com>
arch/arm/src/imxrt and configs/imxrt1050-evk: Adds framebuffer support for thje i.MXRT 1050. Includdd DCD initialization for SDRAM to provide space for the framebufer. External code I used/ported is from NXP and is BSD 3-clause license.
sched/init/nx_bringup.c: Fix a naming collision.
sched/init: Rename os_start() to nx_start()
sched/init: Rename os_smp* to nx_smp*
sched/init: Rename os_bringup to nx_bringup
sched/init: rename all internal static functions to begin with nx_ vs os_
