Commit Graph

33077 Commits

Author SHA1 Message Date
Gregory Nutt
41bfe821d0 Documentation/NuttxPortingGuide.html: Update some directory structure information. This document is very out of date. 2018-08-10 16:12:32 -06:00
Gregory Nutt
25fa50d504 fs/proc and sched/environ: Add support for a procfs entry that will permit examining the environment of any task. 2018-08-10 10:16:39 -06:00
Gregory Nutt
672c6a5405 arch/arm/src/kinetis: Remove annoying typedefs. Discouraged but not prohibited by coding standard 2018-08-10 09:22:27 -06:00
Gregory Nutt
0bc84d74b6 configs/freedom-k66f/nsh/defconfig: Disable DMA in this configuration. 2018-08-10 08:10:35 -06:00
David Sidrane
9ef481ce4b Squashed commit of the following:
Author: Gregory Nutt <gnutt@nuttx.org>
    arch/arm/src/kinetis:  Cosmetic changes from review of coding standard.

Author: David Sidrane <david_s5@nscdg.com>
    Merged in david_s5/nuttx/master_kinetis (pull request #710)

    kinetis:Add DMA and use it as RX FIFOs on Serial

    * Add initial implementation of the DMA

    * Add across Kxx family support [WIP]

    * Incorporate the DMA into the serial

        Add polling to serial

        do DMA initialisation at UART startup and fix DMA mux setting.

        Fix circular DMA handling in the serial driver.

    * Adapt defconfig to enable UART1,4 RXDMA

    * [WIP] - refactor this with new CONFIG DMA settings

    * kinetis/Kconfig:Bring inline with upstream stucture

          1. Use the Serial console configuration from drivers/serial/Kconfig
            and friends.
          2. Prefix arch specific featurs as such
             UARTx_RXDMA->KINETIS_UARTx_RXDMA

    * Kinetis:Serial formatting and clean up

    * kinetis:DMA Formating and Cleanup

    * kinetis:serial Use cleaner DMA API

    * kinetis:Make.defs fix duplicate/errant kinetis_dma.o

    * kinetis:serial Fix warning/error on only one uart using DMA

          The driver can support no DMA on any UART, DMA on some
          UARTs or DMA on all UARTs.

          In the case of no DMA we disable the DMA based variables
          and logic. In the case of all DMA we disable the non
          DMA variables and logic, and in the mixed case both
          DMA and non DMA variable and logic are enabled.

    * kinetis:dma config fix formatting

    * kinetis:chip/dma fix formatting

    * kinetis:DMA config assume Unknown

           Assume KINETIS_DMA_VERSION_UKN for all SoC not versioned
           This is as the code was in nuttx prior to this commit

    * kinetis:DMAMUX contain versioning- use HAS pattern define Unknown

          The pattern for chip versioning is to define 'HAS' constants.
          Define KINETIS_DMAMUX_HAS_MONOTONIC_CHCFG to contain the
          version numbers to the include soc header and use that define
          in the chip headers.

          Define the KINETIS_DMAMUX_VERSION_UKN for the default as prior
          this commit addressing was by default monotonic for CHCFG

    * freedom-k66f:Disable Serial RXDMA

          The application has to add calling of the kinetis
          serial dma poll as it is application specific.

    Approved-by: GregoryN <gnutt@nuttx.org>
2018-08-10 08:03:50 -06:00
David Sidrane
2e23ffe9d2 Squashed commit of the following:
* kinetis:PIT add Liftime and Chaining

    * kinetis:flexcan fix dup line and ordering

    * kinetis:kinetis_lowput.c LPUART data format with parity fix

          The 9-bit data mode is typically used with parity to allow
          eight bits of data plus the parity

    * kinetis:pindma fix warning

    * kinetis:lowputc  LPUART_BAUD_INIT has to be defined

           build fails with test case enable LPUART0 and make
           UART1 console

           if HAVE_LPUART_DEVICE is defined then LPUART_BAUD_INIT
           has to be defined even if the lpuart is not the console

    * kinetis:lpserial fix warning
2018-08-09 09:08:39 -06:00
Daniel P. Carvalho
48a030f7f0 arch/arm/src/stm32l4/chip: Merged stm32l4x2xx and stm32l4x3xx pinmap files. Removed references to stm32l4x2xx_pinmap.h 2018-08-09 07:05:25 -06:00
Juha Niskanen
e402dadf5d arch/arm/src/stm32l4: Avoid using redundant CONFIG_STM32L4_STM32L4X2. This is almost always same as CONFIG_STM32L4_STM32L4X3. Avoid duplication to reduce macro clutter. This patch limits CONFIG_STM32L4_STM32L4X2 usage to dependency tracking and pinmap. Also enable ADC for CONFIG_STM32L4_STM32L4X5 (untested, but same RM). 2018-08-09 06:42:54 -06:00
Gregory Nutt
e88093ff37 configs/stm32f4discovery/scripts/ld.script: Missing EXTERN(_vectors) 2018-08-08 17:19:39 -06:00
Gregory Nutt
3973640fff Update some comments. Remove references to non-existent CONFIG_MAX_TASK_ARG. 2018-08-08 17:06:46 -06:00
Gregory Nutt
5433c9f6d1 Revert "arch/arm/src/stm32, stm32f7, stm32l4 Kconfig files: The option to select the STM32 RTC should not be available if the external RTC option is selected."
While this is a good idea and a proper thing to do, it also creates a circular dependency and errors during configuration.  RTC depends on STM32_RTC which depends on RTC_EXTERNAL with depends on RTC.  No can do.

This reverts commit 714e8c1f9e.
2018-08-08 12:51:40 -06:00
Gregory Nutt
714e8c1f9e arch/arm/src/stm32, stm32f7, stm32l4 Kconfig files: The option to select the STM32 RTC should not be available if the external RTC option is selected. 2018-08-08 12:49:01 -06:00
Gregory Nutt
8d68d9ca43 arch/arm/src/stm32, stm32f7, stm32l4, and related defconfig files: The STM32 RTC driver was being selected by the global CONFIG_RTC option. That is in correct. For example, if you want to disabled the STM32 RTC and use an external RTC you cannot because the external RTC also depends on the global CONFIG_RTC. The solution is to add a new CONFIG_STM32xx_RTC configuration option the permits to you select or deselect the STM32 RTC but still be able to selecte the external RTC. 2018-08-08 12:42:04 -06:00
Juha Niskanen
cd897d71d8 configs/nucleo-l452re: Fix TIMx clock configuration. This is cloned from 391f3715 for nucleo-l432kc. Also fixes DAC build failure. 2018-08-08 06:57:22 -06:00
Juha Niskanen
84f8e01c17 arch/arm/src/stm32l4: Fix ADC and COMP breakage caused by CONFIG_STM32L4_STM32L4X2 separation in 9223123c 2018-08-08 06:55:25 -06:00
Gregory Nutt
a876f58a10 Update a README file. 2018-08-07 17:53:16 -06:00
Daniel Pereira Carvalho
391f3715c1 configs/nucleo-l432kc/include/nucleo-l432kc.h: Fix TIMx clock configuration. Also removes definitions related to timers no available in the STM32L432KC. 2018-08-07 14:44:44 -06:00
Gregory Nutt
a2428db499 configs/olimex-stm32_p407: Add logic to support initialization of the kernel module symbol table installed in the pass2/ directory during the application phase of the build. The kmodule configuration appears to be fully functional. 2018-08-07 10:50:25 -06:00
Gregory Nutt
4becaea8e3 This commit extends support for the two-pass build. Its primary purpose is to incorporate source logic generated by applications into the kernel phase of the build.
In the two pass build, the application logic is built during the first phase, pass 1.  In that phase, the application may generate and install source files in the pass1/directory.  The operating system is built during phase 2 of the build.  At that time, those source files in the pass1/ directory will be built and incorporated into the kernel address space.

The primary purpose of the pass1/ directory is to such application-generated kernel symbol tables.  For an example of the use of this feature, look at apps/examples/module/drivers/Makefile.  Kernel symbol tables are needed to support kernel modules.  Of course, addition board-specific logic in, say, configs/<board>/src would have to be included to make use of the application-installed symbol tables.
2018-08-07 08:31:00 -06:00
Russ Webber
de5e633785 Merged in russkel/nuttx/blackpill (pull request #708)
stm32f103-minimum: add blackpill LED support

Approved-by: GregoryN <gnutt@nuttx.org>
2018-08-07 12:28:46 +00:00
Masayuki Ishikawa
af2876f04f Merged in masayuki2009/nuttx.nuttx/lc823450-xgevk_rndis (pull request #707)
configs/lc823450-xgevk: Update defconfig for krndis

Enable SCHED_CHILD_STATUS and SCHED_HAVE_PARENT
Change IOB related params which are the same as rndis

Signed-off-by: Masayuki Ishikawa <Masayuki.Ishikawa@jp.sony.com>

Approved-by: GregoryN <gnutt@nuttx.org>
2018-08-06 22:22:31 +00:00
Evgeniy Bobkov
67287e89d0 This commit basically reverts my version of the change bbee0d70de in favor of Evgeniy Bobkov's version per request of Juha Niskanen. 2018-08-06 06:22:19 -06:00
Juha Niskanen
1d5e8f6066 arch/arm/src/stm32l4/chip/stm32l4x3xx_pinmap.h: Recent commit removed some of the pin definitions for STM32L451xx, and probably some other chips of the STM32L4x3 subfamily, breaking our builds. Some chips of the subfamily don't have these peripherals, but some do, so that is not a valid reason to remove them. 2018-08-06 06:10:41 -06:00
Masayuki Ishikawa
c46741975d Merged in masayuki2009/nuttx.nuttx/fix_http_streaming_with_lc823450 (pull request #705)
Fix http streaming with lc823450

* arch/arm/src/lc823450: Change C-Buffer under-level control

    The under-level setting was changed from 1KB to 55KB.
    In previous implementation, the setting can ben changed
    by the tx threshold but it is fixed at startup.

    Also, check write size and adjust alignment if needed.

    Signed-off-by: Masayuki Ishikawa <Masayuki.Ishikawa@jp.sony.com>

* configs/lc823450-xgevk: Change IOB related params for rndis.

    Since TCP flow control scheme was changed, HTTP audio streaming
    has not been working. These IOB params are not optimized but
    HTTP audio streaming now works.

    Signed-off-by: Masayuki Ishikawa <Masayuki.Ishikawa@jp.sony.com>

Approved-by: GregoryN <gnutt@nuttx.org>
2018-08-06 03:39:45 +00:00
Masayuki Ishikawa
651f5e2622 Merged in masayuki2009/nuttx.nuttx/fix_loadable_elf_with_smp (pull request #706)
sched/group: Fix a deadlock when loading an ELF

Signed-off-by: Masayuki Ishikawa <Masayuki.Ishikawa@jp.sony.com>

Approved-by: GregoryN <gnutt@nuttx.org>
2018-08-06 03:36:14 +00:00
Gregory Nutt
4e5cf1229c drivers/mmcsd/Kconfig: three configuration settings were within ifdef/endif and led to warning: xxx selects xxxx which has unmet direct dependencies. Fix by moving settings to drivers/Kconfig and outside of the ifdef-endif. Also renamed CONFIG_SDIO_PREFLIGHT to CONFIG_ARCH_HAVE_SDIO_PREFLIGHT to follow naming of similar hidden architecture capability configurations. 2018-08-05 17:58:17 -06:00
Gregory Nutt
c6adc4ba8e sched/pthread/pthread_initialize.c: Fix a warning found in build testing (that is actually and bug and would cause an incorrect value to be returned in many cases. 2018-08-05 16:23:07 -06:00
Gregory Nutt
9726352d7a configs/olimex-stm32-p407/kmodule: Add a configuration that can be used for testing installation of kernel modules in the protected build. This configuration is useless in its present state, not due to any problem with the OS itself, but rather with the current apps/ build system. The symbol table is built by apps/examples/module/drivers/Makefile in user space. The problem with that is that the kernel module does reference internal kernel symbols which are not available in the user space build context. A mechanism is needed in the build system to build the symbol table in the context of the kernel. There is currently no way to do that. 2018-08-05 15:42:11 -06:00
Gregory Nutt
e79772f6d8 tools/showsize.sh needs to be executable. 2018-08-05 15:34:25 -06:00
Gregory Nutt
c2267a57a0 Kconfig files: Fix several errors noted by Alex Denisov in Bitbucket issue 115. 2018-08-05 10:48:02 -06:00
Gregory Nutt
b4b531ef1b include/nuttx/binfmt/binfmt.h: Remove fields from struct binary_s that are not longer used after commit 20a86dfc1b. 2018-08-05 08:50:43 -06:00
Gregory Nutt
4fa546c5d8 Refresh a configuration: Fix an assertion that was messed up in the previous commit. 2018-08-05 08:41:32 -06:00
Gregory Nutt
20a86dfc1b binfmt/ and sched/group: Re-architect the way that loadable ELF or NXFLAT modules are unloaded. Memory resources must be recovered when the task loaded into memory exits. The originmal implementatino used the death-of-child SIGCHLD signal to perform the unload. There are several problems with this: It is overly complex, it requires that the parent task stay resident while the loaded task runs, and it has fatal logic flaws in the protected and kernel model builds because the user signal handler attempts to run in the kernel address space. This commit corrects with using a mindlessly simply BINFMT callback when the task exits. 2018-08-05 08:09:54 -06:00
Gregory Nutt
3e814eb160 Update a README file. 2018-08-04 16:39:10 -06:00
Gregory Nutt
d68b0c35f9 Squashed commit of the following:
configs/olimex-stm32-p407:  Various changes as necessary to get the new kelf configuration working.
    configs/olimex-stm32-p407:  Add kernel ELF (kelf) configuration.
2018-08-04 16:18:51 -06:00
Gregory Nutt
5996d70883 Revise commit 09ccd43d61: That change had the subtle side-effect of unconditionally enabling interrupts in the primask. That may be what we want in most cases, but certainly not all. This does increse the size of the inline function by about 48-bits per instantiation. 2018-08-04 07:37:31 -06:00
GregoryN
1185ff1fc6 Merged loadable_app into master 2018-08-03 21:37:08 -06:00
Gregory Nutt
679d2875f2 Update some comments. 2018-08-03 16:20:42 -06:00
Gregory Nutt
9b7e5ede0a configs/metro-m4: Add an option to build the Metro M4 image to run out of SRAM. This ought to be a safer and quicker way to do the initial bring-up (having bricked the first Metro M4 due to a bad FLASH image). 2018-08-03 15:26:44 -06:00
Gregory Nutt
bbee0d70de arch/arm/src/stm32f7: Fix an error introduced by the patch by Evgeniy Bobkov in Issue #114. That patch brok all STM32 F7 builds. 2018-08-03 14:26:29 -06:00
Gregory Nutt
056d704cf9 This commit brings in a fragmentary, experimental implementation of NETLINK sockets. There is not too much to that socket support on this initial commit, just the netlink socketer framework. However, I decided to bring it into master because there is a enough that I would not want to lose what is in place. And since it is dependent on CONFIG_EXPERIMENATL, its presence on master should be innocuous.
Squashed commit of the following:

    net/netlink:  Mark netlink support as EXPERIMENTAL.
    net/netlink/netlink_sockif.c:  Add netlink_getpeername to the socket interface.
    net:  Add getpeeername() support for netlink sockets.
    include/netpacket/netlink.h:  Add a few more definitions and structures used at the NetLink interface.  Still missing many.
    net/netlink:  Add basic framework for Netlink socket support.
    include/:  Add basic Netlink definitions.
2018-08-03 13:22:36 -06:00
Evgeniy Bobkov
caa7fa7327 This commit ports the stm32f20xxf40xx_flash.c driver to the STM32F7 platform and fixes some relevant problems in the header files. 2018-08-03 11:02:59 -06:00
Gregory Nutt
bfc8f9dfdb arch/arm/src/imxrt/imxrt_enet.c: Fix a warning about imxrt_calcethcrc() being defined, but not used. 2018-08-03 10:53:37 -06:00
Daniel P. Carvalho
9574b0f29b configs/nucleo-l432kc: Fix stm32_userleds and some comments. After change LD2 to LD3 in board.h stm32_userleds.c was broken. 2018-08-03 09:31:40 -06:00
Gregory Nutt
2fd4925965 Update some comments 2018-08-03 09:30:31 -06:00
Gregory Nutt
10069067c1 include/nuttx/i2c/i2c_master.h: Rename I2C_M_NORESTART to I2C_M_NOSTART since it may be used in other contexts than a repeated start. Add comments to clarilfy setup for repeated start. 2018-08-03 08:51:55 -06:00
Gregory Nutt
ee28cd9aeb include/nuttx/i2c/i2c_master.h: Add a definition to distinguish a new START of messages from a repeated start. No lower-half I2C drivers actually implement this new flag bit, however. drivers/i2c/i2c_writeread.c: Use new repeated START definition where appopriated. Other: Some cosmetic changes, updates to README files, etc. 2018-08-03 07:43:57 -06:00
Masayuki Ishikawa
ac5b2ea049 Merged in masayuki2009/nuttx.nuttx/fix_tcp_statistics (pull request #703)
net/tcp: Remove g_netstats.tcp.syndrop++ from tcp_data_event()

Signed-off-by: Masayuki Ishikawa <Masayuki.Ishikawa@jp.sony.com>

Approved-by: GregoryN <gnutt@nuttx.org>
2018-08-03 01:25:53 +00:00
Gregory Nutt
09ccd43d61 arch/arm/include/armv7-m, arch/arm/include/armv7-m, arch/arm/include/stm32f7: Add a configuration to enable workaround for r0p1 Errata 837070: Increasing priority usingwrite to BASEPRI does not take effect immediately.
This update is required to be serialized to the instruction stream meaning that after this update completes, it takes effect immediately and no exceptions of lower priority than the new boosted priority can pre-empt execution.  Because of this erratum, the priority boosting does not take place immediately, allowing the instruction after the MSR to be interrupted by an exception of lower priority than the new boosted priority. This effect is only limited to the next instruction. Subsequent instructions are guaranteed to see the new boosted priority.

This was raised in Bitbucket issue 113 from Vadzim Dambrouski.
2018-08-02 16:51:58 -06:00
Gregory Nutt
f649db2405 Costmetic changes during review of MAX3421E logic. 2018-08-02 13:58:26 -06:00