- Openamp-rp - lists.openampproject.org

Re: [Openamp-rp] Activity on core 0 affects performance on core 1

by Stefano Stabellini

Hi Gaute, When you say that you are running Linux on core 0 and FreeRTOS on core 1, do you mean that you are running both Linux and FreeRTOS on the Cortex-A cluster in separate virtual machines, using a hypervisor? Or are they running on separate clusters, such as FreeRTOS on Cortex-M/R and Linux on a Cortex-A? Cheers, Stefano On Mon, 7 Sep 2020, Gaute Nilsson via Openamp-rp wrote: > We are running Linux on core 0 and FreeRTOS on core 1. > > We have a high priority FreeRTOS task that needs to run with a 100 µs > interval. When the system is up and running, no cross core > communication takes place. > > If I login with a ssh session into the Linux core, this will cause a > performance hit on the FreeRTOS core and cause the high priority task > to miss its deadline. > > I can see that even an ISR on FreeRTOS core that normally takes 4.5 µs > to execute, temporarily will take 7.6 µs to execute just when I open a > ssh session into Linux core. So everything will temporarily run slower > on FreeRTOS core before it goes back to normal. > > > What can cause this, and how can we prevent activity on Linux core > affecting performance on FreeRTOS core? > > -Gaute Nilsson >

4 years, 7 months

1
0
0 0

Activity on core 0 affects performance on core 1

by Gaute Nilsson

We are running Linux on core 0 and FreeRTOS on core 1. We have a high priority FreeRTOS task that needs to run with a 100 µs interval. When the system is up and running, no cross core communication takes place. If I login with a ssh session into the Linux core, this will cause a performance hit on the FreeRTOS core and cause the high priority task to miss its deadline. I can see that even an ISR on FreeRTOS core that normally takes 4.5 µs to execute, temporarily will take 7.6 µs to execute just when I open a ssh session into Linux core. So everything will temporarily run slower on FreeRTOS core before it goes back to normal. What can cause this, and how can we prevent activity on Linux core affecting performance on FreeRTOS core? -Gaute Nilsson We are running Linux on core 0 and FreeRTOS on core 1. We have a high priority FreeRTOS task that needs to run with a 100 µs interval. When the system is up and running, no cross core communication takes place. If I login with a ssh session into the Linux core, this will cause a performance hit on the FreeRTOS core and cause the high priority task to miss its deadline. I can see that even an ISR on FreeRTOS core that normally takes 4.5 µs to execute, temporarily will take 7.6 µs to execute just when I open a ssh session into Linux core. So everything will temporarily run slower on FreeRTOS core before it goes back to normal. What can cause this, and how can we prevent activity on Linux core affecting performance on FreeRTOS core? -Gaute Nilsson

4 years, 7 months

1
0
0 0

[OpenAMP/libmetal] 8dfbba: Use <atomic> when compiled as C++

by Simon Leiner

Branch: refs/heads/master Home: https://github.com/OpenAMP/libmetal Commit: 8dfbbadc2d92fb06ba0ee0bf1837b281e7a021c8 https://github.com/OpenAMP/libmetal/commit/8dfbbadc2d92fb06ba0ee0bf1837b281… Author: Simon Leiner <simon(a)leiner.me> Date: 2020-09-07 (Mon, 07 Sep 2020) Changed paths: M lib/atomic.h Log Message: ----------- Use <atomic> when compiled as C++ This is necessary for the use of libmetal together with C++ clients that require <atomic> as the replacement headers create a conflict with <atomic>. Signed-off-by: Simon Leiner <simon(a)leiner.me> Commit: db77c464376e603bc46122b38a7956c87597895b https://github.com/OpenAMP/libmetal/commit/db77c464376e603bc46122b38a7956c8… Author: Simon Leiner <simon(a)leiner.me> Date: 2020-09-07 (Mon, 07 Sep 2020) Changed paths: M lib/io.h Log Message: ----------- Perform strict typecasts for atomic_store_explicit This is required by the implementation of atomic_store_explicit in the C++ standard library. Signed-off-by: Simon Leiner <simon(a)leiner.me> Compare: https://github.com/OpenAMP/libmetal/compare/9a3162ecc0a3...db77c464376e

4 years, 7 months

1
0
0 0

[OpenAMP/libmetal] 9a3162: Stop using atomic_flag_*() for atomic_int in mutex

by Simon Leiner

Branch: refs/heads/master Home: https://github.com/OpenAMP/libmetal Commit: 9a3162ecc0a353e4b14fe9e81a0d3d69f9c67eca https://github.com/OpenAMP/libmetal/commit/9a3162ecc0a353e4b14fe9e81a0d3d69… Author: Simon Leiner <simon(a)leiner.me> Date: 2020-09-07 (Mon, 07 Sep 2020) Changed paths: M lib/system/freertos/mutex.h M lib/system/generic/mutex.h Log Message: ----------- Stop using atomic_flag_*() for atomic_int in mutex The atomic_flag_* functions are only defined for operands of the type atomic_flag so the previous implementation relied on undefined behavior. Because of the need for a read without side effects (in __metal_mutex_is_acquired), it is not possible to use the atomic_flags type for the mutex implementation. Because of this, all atomic_flag_* functions are replaced with the appropriate functions for atomic value types. Signed-off-by: Simon Leiner <simon(a)leiner.me>

4 years, 7 months

1
0
0 0

[OpenAMP/libmetal] 9e03c4: Introduce metal_asm

by Simon Leiner

Branch: refs/heads/master Home: https://github.com/OpenAMP/libmetal Commit: 9e03c4fceda40ff037c7f4355c431fd775486199 https://github.com/OpenAMP/libmetal/commit/9e03c4fceda40ff037c7f4355c431fd7… Author: Simon Leiner <simon(a)leiner.me> Date: 2020-09-07 (Mon, 07 Sep 2020) Changed paths: M lib/compiler/gcc/compiler.h M lib/compiler/iar/compiler.h Log Message: ----------- Introduce metal_asm This can be used as a future replacement for asm, which is not allowed in the strict ISO C modes of GCC and IAR. Signed-off-by: Simon Leiner <simon(a)leiner.me> Commit: 5cae751afe8c516b4c533a4516a4e80412e93ac1 https://github.com/OpenAMP/libmetal/commit/5cae751afe8c516b4c533a4516a4e804… Author: Simon Leiner <simon(a)leiner.me> Date: 2020-09-07 (Mon, 07 Sep 2020) Changed paths: M examples/system/freertos/zynqmp_r5/zynqmp_amp_demo/common.h M examples/system/generic/zynqmp_r5/zynqmp_amp_demo/common.h M lib/processor/aarch64/cpu.h M lib/processor/x86/cpu.h M lib/processor/x86_64/cpu.h M lib/system/freertos/zynq7/sys.c M lib/system/freertos/zynqmp_a53/sys.c M lib/system/freertos/zynqmp_r5/sys.c M lib/system/generic/microblaze_generic/sys.c M lib/system/generic/zynq7/sys.c M lib/system/generic/zynqmp_a53/sys.c M lib/system/generic/zynqmp_r5/sys.c M lib/system/zephyr/cortexm/sys.c M test/system/freertos/main.c Log Message: ----------- Use metal_asm instead of asm and __asm__ This enables GCC compilation in strict mode (-std=c11 -pedantic) Signed-off-by: Simon Leiner <simon(a)leiner.me> Compare: https://github.com/OpenAMP/libmetal/compare/821fd4fc3db2...5cae751afe8c

4 years, 7 months

1
0
0 0

Meeting in 25 min

by Bill Mills

Hello, Just a reminder we have our meeting in 25 min. Talk to you soon, Bill

4 years, 8 months

1
0
0 0

Re: [Openamp-rp] Openamp-rp post acknowledgement

by William A. Mills

All, I will be sending out a new invite as soon as I get my computer put back together. Maybe tonight but certainly tomorrow AM. Bill On Tue, Aug 11, 2020, 2:22 PM Milea, Danut Gabriel (Danut) via Openamp-rp < openamp-rp(a)lists.openampproject.org> wrote: > Please ignore the meeting update emails I sent to the list earlier.. > That was an error. > Thanks and sorry, > Dan > __________________________________________________________________ > > From: Openamp-rp <openamp-rp-bounces(a)lists.openampproject.org> on > behalf of openamp-rp-bounces(a)lists.openampproject.org > <openamp-rp-bounces(a)lists.openampproject.org> > Sent: Tuesday, August 11, 2020 20:30 > To: Milea, Danut Gabriel (Danut) <Danut.Milea(a)windriver.com> > Subject: Openamp-rp post acknowledgement > > Your message entitled > [Draft] OpenAMP remoteproc sub-group call > was successfully received by the Openamp-rp mailing list. > List info page: > [1]https://lists.openampproject.org/mailman/listinfo/openamp-rp > Your preferences: > [2]https://lists.openampproject.org/mailman/options/openamp-rp/danut.mi > lea%40windriver.com > > References > > 1. https://lists.openampproject.org/mailman/listinfo/openamp-rp > 2. > https://lists.openampproject.org/mailman/options/openamp-rp/danut.milea@win… > -- > Openamp-rp mailing list > Openamp-rp(a)lists.openampproject.org > https://lists.openampproject.org/mailman/listinfo/openamp-rp > All, I will be sending out a new invite as soon as I get my computer put back together. Maybe tonight but certainly tomorrow AM. Bill On Tue, Aug 11, 2020, 2:22 PM Milea, Danut Gabriel (Danut) via Openamp-rp <[1]openamp-rp(a)lists.openampproject.org> wrote: Please ignore the meeting update emails I sent to the list earlier.. That was an error. Thanks and sorry, Dan __________________________________________________________________ From: Openamp-rp <[2]openamp-rp-bounces(a)lists.openampproject.org> on behalf of [3]openamp-rp-bounces(a)lists.openampproject.org <[4]openamp-rp-bounces(a)lists.openampproject.org> Sent: Tuesday, August 11, 2020 20:30 To: Milea, Danut Gabriel (Danut) <[5]Danut.Milea(a)windriver.com> Subject: Openamp-rp post acknowledgement Your message entitled [Draft] OpenAMP remoteproc sub-group call was successfully received by the Openamp-rp mailing list. List info page: [1][6]https://lists.openampproject.org/mailman/listinfo/openamp-rp Your preferences: [2][7]https://lists.openampproject.org/mailman/options/openamp-rp/da nut.mi lea%[8]40windriver.com References 1. [9]https://lists.openampproject.org/mailman/listinfo/openamp-rp 2. [10]https://lists.openampproject.org/mailman/options/openamp-rp/danu t.milea(a)windriver.com -- Openamp-rp mailing list [11]Openamp-rp(a)lists.openampproject.org [12]https://lists.openampproject.org/mailman/listinfo/openamp-rp References 1. mailto:openamp-rp@lists.openampproject.org 2. mailto:openamp-rp-bounces@lists.openampproject.org 3. mailto:openamp-rp-bounces@lists.openampproject.org 4. mailto:openamp-rp-bounces@lists.openampproject.org 5. mailto:Danut.Milea@windriver.com 6. https://lists.openampproject.org/mailman/listinfo/openamp-rp 7. https://lists.openampproject.org/mailman/options/openamp-rp/danut.mi 8. http://40windriver.com/ 9. https://lists.openampproject.org/mailman/listinfo/openamp-rp 10. https://lists.openampproject.org/mailman/options/openamp-rp/danut.milea@win… 11. mailto:Openamp-rp@lists.openampproject.org 12. https://lists.openampproject.org/mailman/listinfo/openamp-rp

4 years, 8 months

2
1
0 0

Re: [Openamp-rp] Openamp-rp post acknowledgement

by Milea, Danut Gabriel (Danut)

Please ignore the meeting update emails I sent to the list earlier.. That was an error. Thanks and sorry, Dan ________________________________ From: Openamp-rp <openamp-rp-bounces(a)lists.openampproject.org> on behalf of openamp-rp-bounces(a)lists.openampproject.org <openamp-rp-bounces(a)lists.openampproject.org> Sent: Tuesday, August 11, 2020 20:30 To: Milea, Danut Gabriel (Danut) <Danut.Milea(a)windriver.com> Subject: Openamp-rp post acknowledgement Your message entitled [Draft] OpenAMP remoteproc sub-group call was successfully received by the Openamp-rp mailing list. List info page: https://lists.openampproject.org/mailman/listinfo/openamp-rp Your preferences: https://lists.openampproject.org/mailman/options/openamp-rp/danut.milea%40w… Please ignore the meeting update emails I sent to the list earlier.. That was an error. Thanks and sorry, Dan __________________________________________________________________ From: Openamp-rp <openamp-rp-bounces(a)lists.openampproject.org> on behalf of openamp-rp-bounces(a)lists.openampproject.org <openamp-rp-bounces(a)lists.openampproject.org> Sent: Tuesday, August 11, 2020 20:30 To: Milea, Danut Gabriel (Danut) <Danut.Milea(a)windriver.com> Subject: Openamp-rp post acknowledgement Your message entitled [Draft] OpenAMP remoteproc sub-group call was successfully received by the Openamp-rp mailing list. List info page: [1]https://lists.openampproject.org/mailman/listinfo/openamp-rp Your preferences: [2]https://lists.openampproject.org/mailman/options/openamp-rp/danut.mi lea%40windriver.com References 1. https://lists.openampproject.org/mailman/listinfo/openamp-rp 2. https://lists.openampproject.org/mailman/options/openamp-rp/danut.milea@win…

4 years, 8 months

1
0
0 0

Canceled: [Draft] OpenAMP remoteproc sub-group call

by Milea, Danut Gabriel (Danut)

https://zoom.us/j/717134356 This is a new meeting invite for 2020 and a new zoom number. This call occurs every two weeks. (There may be other subgroup calls in this timeslot in other weeks.) Agenda items and notes are collected here: https://github.com/OpenAMP/open-amp/wiki/OpenAMP-remoteproc-Subgroup-Meetin… https://zoom.us/j/717134356 This is a new meeting invite for 2020 and a new zoom number. This call occurs every two weeks. (There may be other subgroup calls in this timeslot in other weeks.) Agenda items and notes are collected here: https://github.com/OpenAMP/open-amp/wiki/OpenAMP-remoteproc-Subgroup-Me eting-Notes-2020

4 years, 8 months

1
0
0 0

[Draft] OpenAMP remoteproc sub-group call

by Milea, Danut Gabriel (Danut)

https://zoom.us/j/717134356 This is a new meeting invite for 2020 and a new zoom number. This call occurs every two weeks. (There may be other subgroup calls in this timeslot in other weeks.) Agenda items and notes are collected here: https://github.com/OpenAMP/open-amp/wiki/OpenAMP-remoteproc-Subgroup-Meetin…

4 years, 8 months

1
0
0 0

2025

2024

2023

2022

2021

2020

2019

Openamp-rp