[libbsd] EPOCH(9): Fix epoch call and drain

Since the epoch call callbacks may lock/unlock a mutex the callbacks must be called from within thread context with thread dispatching enabled. Use thread pinning to ensure that the calling thread stays on the right processor. Use the interrupt server to ensure the thread context for the epoch dr

Re: [PATCH v3] Update Strong APA Scheduler

On 11/06/2021 07:45, Richi Dubey wrote: diff --git a/cpukit/include/rtems/score/schedulernodeimpl.h b/cpukit/include/rtems/score/schedulernodeimpl.h index 5d6f795912..ba116077c7 100644 --- a/cpukit/include/rtems/score/schedulernodeimpl.h +++ b/cpukit/include/rtems/score/schedulernodeimpl.h @@ -4

Re: [PATCH v2] Update Strong APA Scheduler

Thanks again for the prompt response. I have sent v3. Can you please check? On Thu, Jun 10, 2021 at 6:50 PM Sebastian Huber < sebastian.hu...@embedded-brains.de> wrote: > On 10/06/2021 10:49, Richi Dubey wrote: > > +/* Forward Declaration of Per_CPU_Control */ > > +struct Per_CPU_Control; > > +ty

[PATCH v3] Update Strong APA Scheduler

This change allows for the migration of higher priority tasks on the arrival of a lower priority task limited by affinity constraints. --- cpukit/include/rtems/scheduler.h | 10 +- .../include/rtems/score/schedulernodeimpl.h | 3 + .../include/rtems/score/schedulerstrongapa.h |

Re: [PATCH v2] Update Strong APA Scheduler

I understand. Thanks. On Thu, Jun 10, 2021 at 7:20 PM Gedare Bloom wrote: > We'll need the entire patch set resent for commit, but you can work > through the scheduler addition first if you like, and then resend the > tests as their own patch set. > > On Thu, Jun 10, 2021 at 7:20 AM Sebastian Hu

Re: [PATCH] score: Add PER_CPU_DATA_NEED_INITIALIZATION()

On 10/06/2021 15:48, Gedare Bloom wrote: +/* + * SPDX-License-Identifier: BSD-2-Clause + * The SPDX should go by itself on the first line of the file, according to our templates? Hm, yes. I was lucky enough to copy from the wrong template. -- embedded brains GmbH Herr Sebastian HUBER Dorniers

Re: [PATCH v2] Update Strong APA Scheduler

We'll need the entire patch set resent for commit, but you can work through the scheduler addition first if you like, and then resend the tests as their own patch set. On Thu, Jun 10, 2021 at 7:20 AM Sebastian Huber wrote: > > On 10/06/2021 10:49, Richi Dubey wrote: > > +/* Forward Declaration of

Re: [PATCH] EPOCH(9): Use PER_CPU_DATA_NEED_INITIALIZATION()

ok On Thu, Jun 10, 2021 at 2:09 AM Sebastian Huber wrote: > > Use the PER_CPU_DATA_NEED_INITIALIZATION() macro if necessary to request an > initialization of the per-CPU data. > --- > rtemsbsd/rtems/rtems-kernel-epoch.c | 4 > 1 file changed, 4 insertions(+) > > diff --git a/rtemsbsd/rtems/

Re: [PATCH] score: Add PER_CPU_DATA_NEED_INITIALIZATION()

On Thu, Jun 10, 2021 at 2:06 AM Sebastian Huber wrote: > > Make the initialization of the per-CPU data optional. > --- > cpukit/Makefile.am | 1 + > cpukit/include/rtems/score/percpudata.h | 14 > cpukit/sapi/src/exinit.c | 13 > cpukit/score/src/

[PATCH 5/7] grlib: Add apbuart_outbyte_wait()

--- bsps/include/grlib/apbuart.h | 8 +++--- bsps/riscv/griscv/console/printk_support.c | 3 ++- bsps/shared/grlib/uart/apbuart_cons.c | 2 +- bsps/shared/grlib/uart/apbuart_polled.c| 30 +++--- bsps/shared/grlib/uart/apbuart_termios.c | 2 +- bsps/spar

[PATCH 7/7] grlib: Simplify apbuart_inbyte_nonblocking()

--- bsps/shared/grlib/uart/apbuart_polled.c | 10 +- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/bsps/shared/grlib/uart/apbuart_polled.c b/bsps/shared/grlib/uart/apbuart_polled.c index 587dacc731..49e137af85 100644 --- a/bsps/shared/grlib/uart/apbuart_polled.c +++ b/bsps

[PATCH 0/7] AMBA Plug & Play and APBUART improvements

Sebastian Huber (7): grlib: Customizable allocation in ambapp_scan() grlib: Add ambapp_plb() bsps/leon3: Auto initialization for printk() grlib: Remove NL -> CR in apbuart_outbyte_polled() grlib: Add apbuart_outbyte_wait() grlib: Fix apbuart_outbyte_polled() grlib: Simplify apbuart_in

[PATCH 3/7] bsps/leon3: Auto initialization for printk()

Get rid of the hack to output into the idle stack during the early system initialization. This fixes also a couple of test program failures which fail due to missing output. --- bsps/sparc/leon3/console/printk_support.c | 42 --- 1 file changed, 14 insertions(+), 28 deletions(

[PATCH 2/7] grlib: Add ambapp_plb()

Replace the global variable ambapp_plb with a function to allow an automatic on demand initialization. --- bsps/include/grlib/ambapp.h| 5 ++ bsps/include/grlib/grlib_impl.h| 2 - bsps/riscv/griscv/console/console.c| 4 +- bsps/riscv/griscv/console/printk_sup

[PATCH 1/7] grlib: Customizable allocation in ambapp_scan()

Make the memory allocations in ambapp_scan() customizable via the new struct ambapp_context parameter which generalizes the memory copy handler. --- bsps/include/grlib/ambapp.h | 27 ++ bsps/shared/grlib/amba/ambapp.c | 41 - 2 files changed,

[PATCH 6/7] grlib: Fix apbuart_outbyte_polled()

Callers usually want to output a character. --- bsps/include/grlib/apbuart.h| 2 +- bsps/shared/grlib/uart/apbuart_polled.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/bsps/include/grlib/apbuart.h b/bsps/include/grlib/apbuart.h index 68bcf1bffa..f54689abc2 10

[PATCH 4/7] grlib: Remove NL -> CR in apbuart_outbyte_polled()

This is already done in rtems_putc(). --- bsps/include/grlib/apbuart.h | 1 - bsps/riscv/griscv/console/printk_support.c | 2 +- bsps/shared/grlib/uart/apbuart_cons.c | 2 +- bsps/shared/grlib/uart/apbuart_polled.c| 7 --- bsps/shared/grlib/uart/apbuart_termios.c | 2 +

Re: [PATCH v2] Update Strong APA Scheduler

On 10/06/2021 10:49, Richi Dubey wrote: +/* Forward Declaration of Per_CPU_Control */ +struct Per_CPU_Control; +typedef struct Per_CPU_Control Per_CPU_Control; Please remove the typedef. + +#define STRONG_SCHEDULER_NODE_OF_CHAIN( node ) \ + RTEMS_CONTAINER_OF( node, Scheduler_strong_APA_Node

Re: [PATCH 1/4] Update Strong APA Scheduler

Hello Sebastian, Thanks for the quick review. I have sent a new v2 patch. Can you please check if it looks good? On Wed, Jun 9, 2021 at 7:25 PM Sebastian Huber < sebastian.hu...@embedded-brains.de> wrote: > Hello Richi, > > could you please adjust your new Doxygen comments so that they are in >

[PATCH v2] Update Strong APA Scheduler

This change allows for the migration of higher priority tasks on the arrival of a lower priority task limited by affinity constraints. --- cpukit/include/rtems/scheduler.h | 10 +- .../include/rtems/score/schedulerstrongapa.h | 166 ++- cpukit/score/src/schedulerstrongapa.c

[PATCH] EPOCH(9): Use PER_CPU_DATA_NEED_INITIALIZATION()

Use the PER_CPU_DATA_NEED_INITIALIZATION() macro if necessary to request an initialization of the per-CPU data. --- rtemsbsd/rtems/rtems-kernel-epoch.c | 4 1 file changed, 4 insertions(+) diff --git a/rtemsbsd/rtems/rtems-kernel-epoch.c b/rtemsbsd/rtems/rtems-kernel-epoch.c index f4fff0ea.

[PATCH] score: Add PER_CPU_DATA_NEED_INITIALIZATION()

Make the initialization of the per-CPU data optional. --- cpukit/Makefile.am | 1 + cpukit/include/rtems/score/percpudata.h | 14 cpukit/sapi/src/exinit.c | 13 cpukit/score/src/percpudata.c| 94 cpukit/score/sr