[PATCH rtems-lwip v2 0/7] lwIP port for STM32F4 BSP

2022-09-06 Thread Duc Doan
This patch set aims to port RTEMS lwIP for STM32F4 BSP. It also contains generic drivers for STM32 chips in general. It is tested with a TCP echo server application on STM32F407 Discovery Board. Prerequisite: this patch set requires my STM32F4 patches to be applied because it uses STM32 HAL. v2:

[PATCH rtems-lwip v2 1/7] lwip.py: Change arch and bsp check method

2022-09-06 Thread Duc Doan
--- lwip.py | 30 +- 1 file changed, 17 insertions(+), 13 deletions(-) diff --git a/lwip.py b/lwip.py index 84eef2c..9db5348 100644 --- a/lwip.py +++ b/lwip.py @@ -99,6 +99,8 @@ def build(bld): drv_incl = [] arch_lib_path = rtems.arch_bsp_lib_path(bld.env.RTE

[PATCH rtems-lwip v2 2/7] lwip.py: Use ant_glob instead of os.walk()

2022-09-06 Thread Duc Doan
Updates #4714 --- lwip.py | 10 +- 1 file changed, 1 insertion(+), 9 deletions(-) diff --git a/lwip.py b/lwip.py index 9db5348..9425dc8 100644 --- a/lwip.py +++ b/lwip.py @@ -110,15 +110,7 @@ def build(bld): source_files.extend(common_source_files) def walk_sources(path): -

[PATCH rtems-lwip v2 3/7] Add STM32 Ethernet source

2022-09-06 Thread Duc Doan
This patch adds ST's Ethernet and lwIP port and DP83848 driver. The files are generated using STM32CubeIDE with STM32F4 Cube FW v1.27.1, under no RTOS mode. --- stm32/driver/dp83848.c | 664 + stm32/driver/dp83848.h | 436 stm32/ethernet

[PATCH rtems-lwip v2 5/7] RTEMS port of lwIP for STM32 and STM32F4 BSP

2022-09-06 Thread Duc Doan
--- rtemslwip/stm32f4/lwipopts.h | 24 ++- rtemslwip/stm32f4/netstart.c | 29 +++- rtemslwip/stm32f4/stm32f4_lwip.c | 14 ++ rtemslwip/stm32f4/stm32f4_lwip.h | 9 + stm32/ethernetif.c | 288 +-- stm32/ethernetif.h | 14 +- stm32

[PATCH rtems-lwip v2 6/7] lwip.py: Add STM32 lwIP port to build

2022-09-06 Thread Duc Doan
--- lwip.py | 22 ++ 1 file changed, 22 insertions(+) diff --git a/lwip.py b/lwip.py index 9425dc8..bb382cd 100644 --- a/lwip.py +++ b/lwip.py @@ -92,6 +92,21 @@ common_source_files = [ 'rtemslwip/bsd_compat/rtems-kernel-program.c' ] +stm32_drv_incl = [ +'stm32', +

[PATCH rtems-lwip v2 4/7] rtemslwip: Add STM32F4 lwipopts.h and netstart.c

2022-09-06 Thread Duc Doan
--- rtemslwip/stm32f4/lwipopts.h | 141 +++ rtemslwip/stm32f4/netstart.c | 52 + 2 files changed, 193 insertions(+) create mode 100644 rtemslwip/stm32f4/lwipopts.h create mode 100644 rtemslwip/stm32f4/netstart.c diff --git a/rtemslwip/stm32f4/lwipopt

Re: [PATCH rtems-lwip v2 1/7] lwip.py: Change arch and bsp check method

2022-09-06 Thread Kinsey Moore
I'm not opposed to this change since it reduces repetition a bit, but was there a particular reason for it beyond that? Kinsey On 9/6/2022 11:20, Duc Doan wrote: --- lwip.py | 30 +- 1 file changed, 17 insertions(+), 13 deletions(-) diff --git a/lwip.py b/lwip.

Re: [PATCH rtems-lwip v2 3/7] Add STM32 Ethernet source

2022-09-06 Thread Kinsey Moore
This should also add a ORIGIN.stm32 and a COPYING.stm32 to describe the imported sources. The information in the commit message you have here should suffice as the majority of the content of ORIGIN.stm32. Kinsey On 9/6/2022 11:20, Duc Doan wrote: This patch adds ST's Ethernet and lwIP port a

Re: [PATCH rtems-lwip v2 4/7] rtemslwip: Add STM32F4 lwipopts.h and netstart.c

2022-09-06 Thread Kinsey Moore
On 9/6/2022 11:20, Duc Doan wrote: --- rtemslwip/stm32f4/lwipopts.h | 141 +++ rtemslwip/stm32f4/netstart.c | 52 + 2 files changed, 193 insertions(+) create mode 100644 rtemslwip/stm32f4/lwipopts.h create mode 100644 rtemslwip/stm32f4/netstart

Re: [PATCH rtems-lwip v2 5/7] RTEMS port of lwIP for STM32 and STM32F4 BSP

2022-09-06 Thread Kinsey Moore
I think you may have accidentally squashed the lwipopts.h and netstart.c changes into this commit instead of 4/7. Changes to imported code should follow the same rules as rtems-libbsd for easier updating. It looks like you did this in most places, but a few instances are called out below where

Re: [PATCH rtems-lwip v2 1/7] lwip.py: Change arch and bsp check method

2022-09-06 Thread Chris Johns
On 7/9/2022 4:41 am, Kinsey Moore wrote: I'm not opposed to this change since it reduces repetition a bit, but was there a particular reason for it beyond that? Yes there is. The arch/bsp checks ... -    if bld.env.RTEMS_ARCH_BSP.startswith('arm-rtems6-tms570ls3137_hdk'): ... included the m