On Fri, Apr 3, 2020 at 8:48 AM Sebastian Huber < sebastian.hu...@embedded-brains.de> wrote:
> On 02/04/2020 22:28, Joel Sherrill wrote: > > > > On Thu, Apr 2, 2020 at 9:47 AM Sebastian Huber < > sebastian.hu...@embedded-brains.de> wrote: > >> On 02/04/2020 16:45, Gedare Bloom wrote: >> >> > Can we hold this until 6? >> Joel complained that the CONFIGURE_USE_DEVFS_AS_BASE_FILESYSTEM is >> broken. For me it is not important. >> > > If this fixes that, please push it. > > The same technique needs to be applied to all drivers of this kind: > > bsps/arm/atsam/console/console.c: link(&usart[0], > CONSOLE_DEVICE_NAME); > bsps/arm/atsam/console/console.c: link(&uart[0], CONSOLE_DEVICE_NAME); > bsps/arm/imx/console/console-config.c: link(path, > CONSOLE_DEVICE_NAME); > bsps/arm/xilinx-zynqmp/console/console-config.c: link(&uart[0], > CONSOLE_DEVICE_NAME); > bsps/arm/raspberrypi/console/console-config.c: link(FBCONS, > CONSOLE_DEVICE_NAME); > bsps/arm/raspberrypi/console/console-config.c: link(MINIUART, > CONSOLE_DEVICE_NAME); > bsps/arm/raspberrypi/console/console-config.c: link(PL011, > CONSOLE_DEVICE_NAME); > bsps/arm/raspberrypi/console/console-config.c: link(PL011, > CONSOLE_DEVICE_NAME); > bsps/arm/xilinx-zynq/console/console-init.c: (void) > link(CONSOLE_DEVICE_NAME, &uart[0]); > bsps/riscv/riscv/console/console-config.c: link(htif_path, > CONSOLE_DEVICE_NAME); > bsps/riscv/riscv/console/console-config.c: link(path, > CONSOLE_DEVICE_NAME); > bsps/riscv/riscv/console/console-config.c: link(fe310_path, > CONSOLE_DEVICE_NAME); > bsps/m68k/mrm332/console/console.c: if ( link( "/dev/sci", > "/dev/console") < 0 ) > bsps/powerpc/tqm8xx/console/console.c: rv = link(tty_name, > CONSOLE_DEVICE_NAME); > OK. But in your patch, you swapped the first and second arguments to link, why? - if ( link( ctx->device_file, CONSOLE_DEVICE_NAME ) != 0 ) { - bsp_fatal( BSP_FATAL_CONSOLE_INSTALL_1 ); - } + (void) link( CONSOLE_DEVICE_NAME, ctx->device_file ); The code you cited on mrm332 is disabled. I also can't cipher the exact change. What is the change? Worse, I don't have access to nearly any of those to test. Ignoring time constraints, I don't understand the change enough to make it, and don't have time. You understand it and it should be mechanical. There are only 7 BSPs there and I think you or Christian created or have worked on most of them. Please just you two split the list and do it. You will have it done before I understand the change. --joel
_______________________________________________ devel mailing list devel@rtems.org http://lists.rtems.org/mailman/listinfo/devel