Re: grub2 on old Linux Distributions (at least CentOS 7)

2021-11-20 Thread Joel Sherrill
On Sat, Nov 20, 2021, 6:50 PM Karel Gardas wrote: > > Hi, > > indeed, centos 7 is bundled with gcc 4.8.5. Now, the question is, is > centos 7 still supported by RTEMS? If so, then we will need to invent > some machinery to check not only for commands but also for versions in > cfg files? Also if

Re: grub2 on old Linux Distributions (at least CentOS 7)

2021-11-20 Thread Karel Gardas
Hi, indeed, centos 7 is bundled with gcc 4.8.5. Now, the question is, is centos 7 still supported by RTEMS? If so, then we will need to invent some machinery to check not only for commands but also for versions in cfg files? Also if we're doing this we need to check firmly we're really on old ce

Re: [PATCH 27/27] score: Restrict affinity for EDF SMP scheduler

2021-11-20 Thread Gedare Bloom
ok other than my comments. nothing looked major. On Mon, Nov 15, 2021 at 10:14 AM Sebastian Huber wrote: > > The SMP EDF scheduler supports a one-to-one and one-to-all thread to > processor affinity. It accepted affinity sets which are a proper > subset of the online processor containing at least

mips gdb/sim build failure on Darwin

2021-11-20 Thread Joel Sherrill
Hi This looks like we need to bump the hash, submit a patch, or change to building with -Wno-error (https://lists.rtems.org/pipermail/build/2021-November/030047.html). Any ideas? FYI Chris.. this one may be on you to see if a hash bump fixes it. /usr/bin/cc -O2 -pipe -fbracket-depth=1024 -I/opt/

dtc build failure on CentOS 7

2021-11-20 Thread Joel Sherrill
Hi Looking in https://lists.rtems.org/pipermail/build/2021-November/030083.html, I'm not sure what explains this error: CC libfdt/fdt.o libfdt/fdt.c: In function 'fdt_ro_probe_': libfdt/fdt.c:27:11: error: 'FDT_ERR_ALIGNMENT' undeclared (first use in this function) return -FDT_ERR_ALIGNMENT;

grub2 on old Linux Distributions (at least CentOS 7)

2021-11-20 Thread Joel Sherrill
Hi Looking over the build sweep failures, I spotted this one (https://lists.rtems.org/pipermail/build/2021-November/030118.html) which would lean to their being a floor on older distributions. There is a software collection for a gcc8 for CentOS 7 but do we want to require it? This is for x86_64 t

Re: [PATCH 15/27] score: Add SMP scheduler make/clean sticky

2021-11-20 Thread Gedare Bloom
On Mon, Nov 15, 2021 at 10:13 AM Sebastian Huber wrote: > > This patch fixes the following broken behaviour: > > While a thread is scheduled on a helping scheduler, while it does not > own a MrsP semaphore, if it obtains a MrsP semaphore, then no > scheduler node using an idle thread and the

Re: [PATCH 14/27] score: Add SMP scheduler idle exchange callback

2021-11-20 Thread Gedare Bloom
On Mon, Nov 15, 2021 at 10:14 AM Sebastian Huber wrote: > > Update #4531. > --- > cpukit/include/rtems/score/schedulersmpimpl.h | 4 > 1 file changed, 4 insertions(+) > > diff --git a/cpukit/include/rtems/score/schedulersmpimpl.h > b/cpukit/include/rtems/score/schedulersmpimpl.h > index 499

Re: [PATCH 13/27] score: Optimize SMP EDF move to ready operation

2021-11-20 Thread Gedare Bloom
On Mon, Nov 15, 2021 at 10:13 AM Sebastian Huber wrote: > > If a node is moved from the scheduled chain to the ready queue, then we > know that it is the highest priority ready node. So, it can be > prepended to the ready queue without doing any comparisons. > I'm not certain on the logic here. I

Re: [PATCH 12/27] score: Rework affine ready queue handling

2021-11-20 Thread Gedare Bloom
On Mon, Nov 15, 2021 at 10:13 AM Sebastian Huber wrote: > > Rework the handling of the affine ready queue for the EDF SMP scheduler. > Do the queue handling in the node insert, move, and extract operations. > Remove the queue handling from _Scheduler_EDF_SMP_Allocate_processor(). > > Update #4531.

Re: [PATCH 08/27] score: Add missing idle thread exchanges

2021-11-20 Thread Gedare Bloom
On Mon, Nov 15, 2021 at 10:13 AM Sebastian Huber wrote: > > Update #4531. > --- > cpukit/include/rtems/score/schedulersmpimpl.h | 18 ++ > 1 file changed, 18 insertions(+) > > diff --git a/cpukit/include/rtems/score/schedulersmpimpl.h > b/cpukit/include/rtems/score/schedulersmpim

Re: [PATCH 02/27] score: Add red-black tree append/prepend

2021-11-20 Thread Gedare Bloom
On Mon, Nov 15, 2021 at 10:13 AM Sebastian Huber wrote: > > These functions are a faster alternative to _RBTree_Insert_inline() if > it is known that the new node is the maximum/minimum node. > > Update #4531. > --- > cpukit/include/rtems/score/rbtreeimpl.h | 26 +++ > cpukit/score/src/rb

Re: [PATCH 1/2] cpukit: Enable debug for SMP AArch64

2021-11-20 Thread Gedare Bloom
This looks ok, along with the copyright clean-up. In the future please separate non-functional (style, copyright, etc) changes from functional patches to simplify review/approval/revision process. One question I do have from this: the minimum APPROX size is 180 (+CPU_PER_CPU_CONTROL_SIZE + CPU_INT

Re: [PATCH 6/6] score: Properly continue the thread during restart

2021-11-20 Thread Gedare Bloom
On Mon, Nov 15, 2021 at 10:09 AM Sebastian Huber wrote: > > The _Thread_queue_Extract() does not deal with potential priority > updates and the SMP locking protocol handling. Use > _Thread_queue_Continue(). For the POSIX signals processing this is > currently probably unnecessary, however, the u

Re: [PATCH 2/6] score: Add _Thread_MP_Extract_proxy()

2021-11-20 Thread Gedare Bloom
On Mon, Nov 15, 2021 at 10:09 AM Sebastian Huber wrote: > > Remove _Thread_queue_Extract_with_proxy() and move the proxy extraction > to _Thread_MP_Extract_proxy(). Move similar code blocks of the previous > caller of _Thread_queue_Extract_with_proxy() to helper functions. > > Update #4546. > ---