test/rcu: fix the compiling error for armv8.2
PATCH 1/2 test/rcu: fix the compiling error for armv8.2
With "-march=armv8.2-a" specified, a compiling error generated:app/test/test_rcu_qsbr.c:234:10: error: comparison of integer expressions of different signedness: ‘unsigned int’ and ‘int’ [-Werror=sign-compare]
--- app/test/test_rcu_qsbr.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
app/test/test_rcu_qsbr.c +1 -1
@@ -231,7 +231,7 @@ test_rcu_qsbr_thread_unregister(void)/* Update quiescent state counter */ for (i =0 ; i < num_threads[j]; i++) {/* Skip one update */ -if (i == (RTE_MAX_LCORE -10 )) +if (i == (unsigned int )(RTE_MAX_LCORE -10 ))continue ;rte_rcu_qsbr_quiescent (t[0 ], (j ==2 ) ? (RTE_MAX_LCORE -1 ) : i);
Signature · 1 line
2.7.4
PATCH 2/2 config: add N1SDP configuration
For more information of this platform, refer to:
https://community.arm.com/developer/tools-software/oss-platforms/w/
docs/440/neoverse-n1-sdp
--- config/arm/meson.build | 4 +++- config/defconfig_arm64-neoversen1-linux-gcc | 16 ++++++++++++++++ 2 files changed, 19 insertions(+), 1 deletion(-) create mode 100644 config/defconfig_arm64-neoversen1-linux-gcc
config/arm/meson.build +3 -1
@@ -87,7 +87,9 @@ machine_args_generic = [ ['0xd07', ['-mcpu=cortex-a57']], ['0xd08', ['-mcpu=cortex-a72']], ['0xd09', ['-mcpu=cortex-a73']], - ['0xd0a', ['-mcpu=cortex-a75']]] + ['0xd0a', ['-mcpu=cortex-a75']], + ['0xd0b', ['-mcpu=cortex-a76']], + ['0xd0c', ['-march=armv8.2-a', '-mcpu=neoverse-n1']]] machine_args_cavium = [ ['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
config/defconfig_arm64-neoversen1-linux-gcc new file +16 -0
@@ -0,0 +1,16 @@ +# SPDX-License-Identifier: BSD-3-Clause +# Copyright(c) 2019 Arm Ltd. +# + +#include "defconfig_arm64-armv8a-linux-gcc" + +CONFIG_RTE_MACHINE="armv8.2-a" +# CONFIG_RTE_ARCH_ARM_TUNE="cortex-a57" +CONFIG_RTE_ARCH_ARM_TUNE="neoverse-n1" +CONFIG_RTE_CACHE_LINE_SIZE=64 +CONFIG_RTE_MAX_NUMA_NODES=1 +CONFIG_RTE_MAX_LCORE=4 + +# Doesn't support NUMA +CONFIG_RTE_EAL_NUMA_AWARE_HUGEPAGES=n +CONFIG_RTE_LIBRTE_VHOST_NUMA=n
Signature · 1 line
2.7.4
v2 cover add arm neoverse N1 SDP configuration
In reply to Gavin HuNerverse N1 is a new SoC based on armv8.2-a and with other advanced features. N1 SDP is the infrastructure segment development platform with N1 SoC inside. Here is more information about it:
https://community.arm.com/developer/tools-software/oss-platforms/w/docs/440/neoverse-n1-sdp
V2:
Add the cover letter and the patch for cortex-a76 configuration
Gavin Hu (3):
test/rcu: fix the compiling error for armv8.2
config: add arm neoverse N1 SDP configuration
config: add cortex-a76 configuration
app/test/test_rcu_qsbr.c | 2 +- config/arm/meson.build | 10 +++++++- config/defconfig_arm64-neoversen1-linux-gcc | 1 + config/defconfig_arm64-neoversen1-linuxapp-gcc | 15 ++++++++++++ mk/machine/neoversen1/rte.vars.mk | 34 ++++++++++++++++++++++++++ 5 files changed, 60 insertions(+), 2 deletions(-) create mode 120000 config/defconfig_arm64-neoversen1-linux-gcc create mode 100644 config/defconfig_arm64-neoversen1-linuxapp-gcc create mode 100644 mk/machine/neoversen1/rte.vars.mk
Signature · 1 line
2.7.4
PATCH v2 1/3 test/rcu: fix the compiling error for armv8.2
With "-march=armv8.2-a" specified, a compiling error generated:app/test/test_rcu_qsbr.c:234:10: error: comparison of integer expressions of different signedness: ‘unsigned int’ and ‘int’ [-Werror=sign-compare]
--- app/test/test_rcu_qsbr.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
app/test/test_rcu_qsbr.c +1 -1
@@ -231,7 +231,7 @@ test_rcu_qsbr_thread_unregister(void)/* Update quiescent state counter */ for (i =0 ; i < num_threads[j]; i++) {/* Skip one update */ -if (i == (RTE_MAX_LCORE -10 )) +if (i == (unsigned int )(RTE_MAX_LCORE -10 ))continue ;rte_rcu_qsbr_quiescent (t[0 ], (j ==2 ) ? (RTE_MAX_LCORE -1 ) : i);
Signature · 1 line
2.7.4
PATCH v2 2/3 config: add arm neoverse N1 SDP configuration
In reply to Gavin HuArm N1 SDP is an infrastructure segment development platform based on armv8.2-a Neoverse N1 CPU. For more information, refer to:
https://community.arm.com/developer/tools-software/oss-platforms/w/
docs/440/neoverse-n1-sdp
--- config/arm/meson.build | 9 ++++++- config/defconfig_arm64-neoversen1-linux-gcc | 1 + config/defconfig_arm64-neoversen1-linuxapp-gcc | 15 ++++++++++++ mk/machine/neoversen1/rte.vars.mk | 34 ++++++++++++++++++++++++++ 4 files changed, 58 insertions(+), 1 deletion(-) create mode 120000 config/defconfig_arm64-neoversen1-linux-gcc create mode 100644 config/defconfig_arm64-neoversen1-linuxapp-gcc create mode 100644 mk/machine/neoversen1/rte.vars.mk
config/arm/meson.build +8 -1
@@ -63,6 +63,12 @@ flags_armada = [ ['RTE_MAX_LCORE', 16]] flags_default_extra = [] +flags_neoversen1_extra = [ + ['RTE_MACHINE', '"neoversen1"'], + ['RTE_MAX_NUMA_NODES', 1], + ['RTE_MAX_LCORE', 4], + ['RTE_EAL_NUMA_AWARE_HUGEPAGES', false], + ['RTE_LIBRTE_VHOST_NUMA', false]] flags_thunderx_extra = [ ['RTE_MACHINE', '"thunderx"'], ['RTE_USE_C11_MEM_MODEL', false]]
@@ -87,7 +93,8 @@ machine_args_generic = [ ['0xd07', ['-mcpu=cortex-a57']], ['0xd08', ['-mcpu=cortex-a72']], ['0xd09', ['-mcpu=cortex-a73']], - ['0xd0a', ['-mcpu=cortex-a75']]] + ['0xd0a', ['-mcpu=cortex-a75']], + ['0xd0c', ['-march=armv8.2-a+crc+crypto', '-mcpu=neoverse-n1'], flags_neoversen1_extra]] machine_args_cavium = [ ['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
config/defconfig_arm64-neoversen1-linux-gcc new file +1 -0
@@ -0,0 +1 @@ +defconfig_arm64-neoversen1-linuxapp-gcc
config/defconfig_arm64-neoversen1-linuxapp-gcc new file +15 -0
@@ -0,0 +1,15 @@ +# SPDX-License-Identifier: BSD-3-Clause +# Copyright(c) 2019 Arm Ltd. +# + +#include "defconfig_arm64-armv8a-linux-gcc" + +CONFIG_RTE_MACHINE="neoversen1" +CONFIG_RTE_ARCH_ARM_TUNE="neoverse-n1" +CONFIG_RTE_MAX_LCORE=4 +CONFIG_RTE_MAX_NUMA_NODES=1 +CONFIG_RTE_CACHE_LINE_SIZE=64 + +# Doesn't support NUMA +CONFIG_RTE_EAL_NUMA_AWARE_HUGEPAGES=n +CONFIG_RTE_LIBRTE_VHOST_NUMA=n
mk/machine/neoversen1/rte.vars.mk new file +34 -0
@@ -0,0 +1,34 @@ +# SPDX-License-Identifier: BSD-3-Clause +# Copyright(c) 2019 Arm Ltd +# + +# +# machine: +# +# - can define ARCH variable (overridden by cmdline value) +# - can define CROSS variable (overridden by cmdline value) +# - define MACHINE_CFLAGS variable (overridden by cmdline value) +# - define MACHINE_LDFLAGS variable (overridden by cmdline value) +# - define MACHINE_ASFLAGS variable (overridden by cmdline value) +# - can define CPU_CFLAGS variable (overridden by cmdline value) that +# overrides the one defined in arch. +# - can define CPU_LDFLAGS variable (overridden by cmdline value) that +# overrides the one defined in arch. +# - can define CPU_ASFLAGS variable (overridden by cmdline value) that +# overrides the one defined in arch. +# - may override any previously defined variable +# + +# ARCH = +# CROSS = +# MACHINE_CFLAGS = +# MACHINE_LDFLAGS = +# MACHINE_ASFLAGS = +# CPU_CFLAGS = +# CPU_LDFLAGS = +# CPU_ASFLAGS = + +include $(RTE_SDK)/mk/rte.helper.mk + +MACHINE_CFLAGS += $(call rte_cc_has_argument, -march=armv8.2-a+crc+crypto) +MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=neoverse-n1)
Signature · 1 line
2.7.4
Re: [dpdk-dev] [PATCH v2 2/3] config: add arm neoverse N1 SDP configuration
Quoted 24 lines · Gavin Hu
Arm N1 SDP is an infrastructure segment development platform based on armv8.2-a Neoverse N1 CPU. For more information, refer to:
https://community.arm.com/developer/tools-software/oss-platforms/w/
docs/440/neoverse-n1-sdpSigned-off-by Gavin Hu <[email protected]>Reviewed-by Honnappa Nagarahalli <[email protected]>Reviewed-by Steve Capper <[email protected]>--- config/arm/meson.build | 9 ++++++- config/defconfig_arm64-neoversen1-linux-gcc | 1 + config/defconfig_arm64-neoversen1-linuxapp-gcc | 15 ++++++++++++ mk/machine/neoversen1/rte.vars.mk | 34 ++++++++++++++++++++++++++ 4 files changed, 58 insertions(+), 1 deletion(-) create mode 120000 config/defconfig_arm64-neoversen1-linux-gcc create mode 100644 config/defconfig_arm64-neoversen1-linuxapp-gcc create mode 100644 mk/machine/neoversen1/rte.vars.mkconfig/arm/meson.build +0
-0
config/arm/meson.build flags_armada = [ · line 67
Quoted 5 lines
config/arm/meson.build +2
-0@@ -63,6 +63,12 @@ flags_armada = [ ['RTE_MAX_LCORE', 16]] flags_default_extra = [] +flags_neoversen1_extra = [ + ['RTE_MACHINE', '"neoversen1"'],
What does RTE_MACHINE stand for?
Neoverse N1 is a CPU core, not a platform, not a SoC architecture etc. Many different SoC's can be designed using the N1 core.
Some will be small with few cores and no NUMA, others might be large with many cores and multi-socket (NUMA) support.
The N1SDP (N1 System Development Platform) is using such a small SoC (specially designed for the N1DSP, it is not going to show up in other platforms). The configuration below matches N1SDP but will most likely not match other future N1-based SoC's. What do we do when other quite different "machines" using the Neoverse N1 core shows up? Is it OK to qualitatively change these configurations later?
Quoted 25 lines
config/arm/meson.build +6
-1+ ['RTE_MAX_NUMA_NODES', 1], + ['RTE_MAX_LCORE', 4], + ['RTE_EAL_NUMA_AWARE_HUGEPAGES', false], + ['RTE_LIBRTE_VHOST_NUMA', false]] flags_thunderx_extra = [ ['RTE_MACHINE', '"thunderx"'], ['RTE_USE_C11_MEM_MODEL', false]]@@ -87,7 +93,8 @@ machine_args_generic = [ ['0xd07', ['-mcpu=cortex-a57']], ['0xd08', ['-mcpu=cortex-a72']], ['0xd09', ['-mcpu=cortex-a73']], - ['0xd0a', ['-mcpu=cortex-a75']]] + ['0xd0a', ['-mcpu=cortex-a75']], + ['0xd0c', ['-march=armv8.2-a+crc+crypto', '-mcpu=neoverse-n1'],
flags_neoversen1_extra]]
machine_args_cavium= [config/arm/meson.build +0
-0['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],config/defconfig_arm64-neoversen1-linux-gcc new file +0
-0
config/defconfig_arm64-neoversen1-linux-gcc line 1
Quoted 2 lines
config/defconfig_arm64-neoversen1-linux-gcc new file +1
-0@@ -0,0 +1 @@ +defconfig_arm64-neoversen1-linuxapp-gcc
!
Quoted 8 lines
config/defconfig_arm64-neoversen1-linuxapp-gcc new file +0
-0Quoted 1 line
config/defconfig_arm64-neoversen1-linuxapp-gcc line 7
Quoted 7 lines
config/defconfig_arm64-neoversen1-linuxapp-gcc new file +7
-0@@ -0,0 +1,15 @@ +# SPDX-License-Identifier: BSD-3-Clause +# Copyright(c) 2019 Arm Ltd. +# + +#include "defconfig_arm64-armv8a-linux-gcc" + +CONFIG_RTE_MACHINE="neoversen1"
This should probably be "n1sdp" as this is the name of the platform that matches the below configuration.
Quoted 51 lines
config/defconfig_arm64-neoversen1-linuxapp-gcc new file +8
-0+CONFIG_RTE_ARCH_ARM_TUNE="neoverse-n1" +CONFIG_RTE_MAX_LCORE=4 +CONFIG_RTE_MAX_NUMA_NODES=1 +CONFIG_RTE_CACHE_LINE_SIZE=64 + +# Doesn't support NUMA +CONFIG_RTE_EAL_NUMA_AWARE_HUGEPAGES=n +CONFIG_RTE_LIBRTE_VHOST_NUMA=nmk/machine/neoversen1/rte.vars.mk new file +34
-0@@ -0,0 +1,34 @@ +# SPDX-License-Identifier: BSD-3-Clause +# Copyright(c) 2019 Arm Ltd +# + +# +# machine: +# +# - can define ARCH variable (overridden by cmdline value) +# - can define CROSS variable (overridden by cmdline value) +# - define MACHINE_CFLAGS variable (overridden by cmdline value) +# - define MACHINE_LDFLAGS variable (overridden by cmdline value) +# - define MACHINE_ASFLAGS variable (overridden by cmdline value) +# - can define CPU_CFLAGS variable (overridden by cmdline value) that +# overrides the one defined in arch. +# - can define CPU_LDFLAGS variable (overridden by cmdline value) that +# overrides the one defined in arch. +# - can define CPU_ASFLAGS variable (overridden by cmdline value) that +# overrides the one defined in arch. +# - may override any previously defined variable +# + +# ARCH = +# CROSS = +# MACHINE_CFLAGS = +# MACHINE_LDFLAGS = +# MACHINE_ASFLAGS = +# CPU_CFLAGS = +# CPU_LDFLAGS = +# CPU_ASFLAGS = + +include $(RTE_SDK)/mk/rte.helper.mk + +MACHINE_CFLAGS += $(call rte_cc_has_argument, -march=armv8.2-a+crc+crypto) +MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=neoverse-n1)
Signature · 2 lines
Ola Liljedahl, Networking System Architect, Arm Phone +46706866373, Skype ola.liljedahl
Re: [dpdk-dev] [PATCH v2 2/3] config: add arm neoverse N1 SDP configuration
<snip>
Quoted 28 lines · Gavin Hu
Quoted 26 lines
Arm N1 SDP is an infrastructure segment development platform based on armv8.2-a Neoverse N1 CPU. For more information, refer to:
https://community.arm.com/developer/tools-software/oss-platforms/w/
docs/440/neoverse-n1-sdpSigned-off-by Gavin Hu <[email protected]>Reviewed-by Honnappa Nagarahalli <[email protected]>Reviewed-by Steve Capper <[email protected]>--- config/arm/meson.build | 9 ++++++- config/defconfig_arm64-neoversen1-linux-gcc | 1 + config/defconfig_arm64-neoversen1-linuxapp-gcc | 15 ++++++++++++ mk/machine/neoversen1/rte.vars.mk | 34 ++++++++++++++++++++++++++ 4 files changed, 58 insertions(+), 1 deletion(-) create mode 120000 config/defconfig_arm64-neoversen1-linux-gcc create mode 100644 config/defconfig_arm64-neoversen1-linuxapp-gcc create mode 100644 mk/machine/neoversen1/rte.vars.mkconfig/arm/meson.build index renamed from config/arm/meson.build +0
-0979018e..995d321 100644
config/arm/meson.build +0
-0@@ -63,6 +63,12 @@ flags_armada = [ ['RTE_MAX_LCORE', 16]]
Quoted 4 lines
config/arm/meson.build flags_armada = [ · line 67
Quoted 3 lines
config/arm/meson.build +2
-0flags_default_extra = [] +flags_neoversen1_extra = [ +['RTE_MACHINE', '"neoversen1"'],What does
RTE_MACHINEstand for?
Not sure, what it stands for. I cannot find a consistent assignment for it. Some time it takes 'armv8' (in the case of BlueField), 'dpaa', 'dpaa2', 'thunderx' etc (architecture, SoC, micro-architecture). I also cannot find an instance of its usage.
Quoted 11 lines
Neoverse N1 is a CPU core, not a platform, not a SoC architecture etc. Many different SoC's can be designed using the N1 core.
Some will be small with few cores and no NUMA, others might be large with many cores and multi-socket (NUMA) support.
The N1SDP (N1 System Development Platform) is using such a small SoC (specially designed for the N1DSP, it is not going to show up in other platforms). The configuration below matches N1SDP but will most likely not match other future N1-based SoC's. What do we do when other quite different "machines" using the Neoverse N1 core shows up? Is it OK to qualitatively change these configurations later?
All other implementations of N1 based SoCs should have their own implementor ID and variant. Hence they should not clash with these configurations. IIRC, BlueField had the same implementation ID as Arm (0x41), so ' flags_arm' was modified to reflect BlueField configuration. Hopefully, this will be corrected in the future.
Quoted 36 lines
config/arm/meson.build flags_armada = [ · line 70
Quoted 4 lines
config/arm/meson.build +3
-0+['RTE_MAX_NUMA_NODES', 1], +['RTE_MAX_LCORE', 4], +['RTE_EAL_NUMA_AWARE_HUGEPAGES', false],['
RTE_LIBRTE_VHOST_NUMA',Quoted 20 lines
config/arm/meson.build +1
-0+false]] flags_thunderx_extra = [ ['RTE_MACHINE', '"thunderx"'], ['RTE_USE_C11_MEM_MODEL', false]]@@ -87,7 +93,8 @@ machine_args_generic = [ ['0xd07', ['-mcpu=cortex-a57']], ['0xd08', ['-mcpu=cortex-a72']], ['0xd09',['-mcpu=cortex-a73']], -['0xd0a', ['-mcpu=cortex-a75']]] +['0xd0a', ['-mcpu=cortex-a75']],
+['0xd0c', ['-march=armv8.2-a+crc+crypto', '-mcpu=neoverse-n1'],flags_neoversen1_extra]]
machine_args_cavium= [
['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],config/defconfig_arm64-neoversen1-linux-gcc new file +0
-0config/defconfig_arm64-neoversen1-linux-gcc line 1
Quoted 2 lines
config/defconfig_arm64-neoversen1-linux-gcc new file +1
-0@@ -0,0 +1 @@ +defconfig_arm64-neoversen1-linuxapp-gcc!
Quoted 7 lines
config/defconfig_arm64-neoversen1-linuxapp-gcc new file +0
-0Quoted 1 line
Quoted 9 lines
config/defconfig_arm64-neoversen1-linuxapp-gcc line 6
Quoted 7 lines
config/defconfig_arm64-neoversen1-linuxapp-gcc new file +6
-0@@ -0,0 +1,15 @@ +# SPDX-License-Identifier: BSD-3-Clause # Copyright(c) 2019 Arm Ltd. +# + +#include "defconfig_arm64-armv8a-linux-gcc" + +CONFIG_RTE_MACHINE="neoversen1"This should probably be "n1sdp" as this is the name of the platform that matches the below configuration.
A clear definition of RTE_MACHINE is required. Jerin?
Quoted 54 lines
Quoted 41 lines
config/defconfig_arm64-neoversen1-linuxapp-gcc new file +8
-0+CONFIG_RTE_ARCH_ARM_TUNE="neoverse-n1" +CONFIG_RTE_MAX_LCORE=4 +CONFIG_RTE_MAX_NUMA_NODES=1 +CONFIG_RTE_CACHE_LINE_SIZE=64 + +# Doesn't support NUMA +CONFIG_RTE_EAL_NUMA_AWARE_HUGEPAGES=n +CONFIG_RTE_LIBRTE_VHOST_NUMA=nmk/machine/neoversen1/rte.vars.mk new file +23
-0@@ -0,0 +1,34 @@ +# SPDX-License-Identifier: BSD-3-Clause # Copyright(c) 2019 Arm Ltd # + +# +# machine: +# +# - can define ARCH variable (overridden by cmdline value) +# - can define CROSS variable (overridden by cmdline value) +# - define MACHINE_CFLAGS variable (overridden by cmdline value) +# - define MACHINE_LDFLAGS variable (overridden by cmdline value) +# - define MACHINE_ASFLAGS variable (overridden by cmdline value) +# - can define CPU_CFLAGS variable (overridden by cmdline value) that +# overrides the one defined in arch. +# - can define CPU_LDFLAGS variable (overridden by cmdline value) that +# overrides the one defined in arch. +# - can define CPU_ASFLAGS variable (overridden by cmdline value) that +# overrides the one defined in arch. +# - may override any previously defined variable +# + +# ARCH = +# CROSS = +# MACHINE_CFLAGS = +# MACHINE_LDFLAGS =mk/machine/neoversen1/rte.vars.mk line 33
Quoted 10 lines
mk/machine/neoversen1/rte.vars.mk new file +10
-0+# MACHINE_ASFLAGS = +# CPU_CFLAGS = +# CPU_LDFLAGS = +# CPU_ASFLAGS = + +include $(RTE_SDK)/mk/rte.helper.mk + +MACHINE_CFLAGS += $(call rte_cc_has_argument, +-march=armv8.2-a+crc+crypto) MACHINE_CFLAGS += $(call +rte_cc_has_argument, -mcpu=neoverse-n1)--
Ola Liljedahl, Networking System Architect, Arm Phone +46706866373, Skype ola.liljedahl
Re: [dpdk-dev] [PATCH v2 2/3] config: add arm neoverse N1 SDP configuration
Quoted 36 lines · Honnappa Nagarahalli
<snip>
Quoted 24 lines · Gavin Hu
Quoted 22 lines
Arm N1 SDP is an infrastructure segment development platform based on armv8.2-a Neoverse N1 CPU. For more information, refer to:
https://community.arm.com/developer/tools-software/oss-platforms/w/
docs/440/neoverse-n1-sdpSigned-off-by Gavin Hu <[email protected]>Reviewed-by Honnappa Nagarahalli <[email protected]>Reviewed-by Steve Capper <[email protected]>--- config/arm/meson.build | 9 ++++++- config/defconfig_arm64-neoversen1-linux-gcc | 1 + config/defconfig_arm64-neoversen1-linuxapp-gcc | 15 ++++++++++++ mk/machine/neoversen1/rte.vars.mk | 34config/defconfig_arm64-neoversen1-linuxapp-gcc new file +0
-0Quoted 1 line
Quoted 9 lines
config/defconfig_arm64-neoversen1-linuxapp-gcc line 6
Quoted 7 lines
config/defconfig_arm64-neoversen1-linuxapp-gcc new file +6
-0@@ -0,0 +1,15 @@ +# SPDX-License-Identifier: BSD-3-Clause # Copyright(c) 2019 Arm Ltd. +# + +#include "defconfig_arm64-armv8a-linux-gcc" + +CONFIG_RTE_MACHINE="neoversen1"This should probably be "n1sdp" as this is the name of the platform that matches the below configuration.
Quoted 1 line
A clear definition of
RTE_MACHINEis required. Jerin?
I think, In the existing scheme of things, RTE_MACHINE defines, where to take the MACHINE_CFLAGSmk/machine/xxxx/rte.vars.mk
Considering the fact that there will be a lot of reusable IPs(for CPU) from ARM for armv8, I think,
it would make sense to introduce RTE_MICRO_ARCH to avoid a lot of code duplications and
confusion.
RTE_ARCH example: "x86" or "arm64"RTE_MICRO_ARCH example: "a72" or "thunderx3" - defines
mcpu and armv8 verion arch etcRTE_MACHINE example: "bluefield" or "thunderx3"
- defines, number of cores, NUMA or not? etc
RTE_MACHINE should be probe based on "implementation ID" for arm64 and reuse already defined RTE_MICRO_ARCH
to avoid code duplication.
Quoted 56 lines
Quoted 55 lines
Quoted 42 lines
config/defconfig_arm64-neoversen1-linuxapp-gcc new file +8
-0+CONFIG_RTE_ARCH_ARM_TUNE="neoverse-n1" +CONFIG_RTE_MAX_LCORE=4 +CONFIG_RTE_MAX_NUMA_NODES=1 +CONFIG_RTE_CACHE_LINE_SIZE=64 + +# Doesn't support NUMA +CONFIG_RTE_EAL_NUMA_AWARE_HUGEPAGES=n +CONFIG_RTE_LIBRTE_VHOST_NUMA=nmk/machine/neoversen1/rte.vars.mk new file +23
-0@@ -0,0 +1,34 @@ +# SPDX-License-Identifier: BSD-3-Clause # Copyright(c) 2019 Arm Ltd # + +# +# machine: +# +# - can define ARCH variable (overridden by cmdline value) +# - can define CROSS variable (overridden by cmdline value) +# - define MACHINE_CFLAGS variable (overridden by cmdline value) +# - define MACHINE_LDFLAGS variable (overridden by cmdline value) +# - define MACHINE_ASFLAGS variable (overridden by cmdline value) +# - can define CPU_CFLAGS variable (overridden by cmdline value) that +# overrides the one defined in arch. +# - can define CPU_LDFLAGS variable (overridden by cmdline value) that +# overrides the one defined in arch. +# - can define CPU_ASFLAGS variable (overridden by cmdline value) that +# overrides the one defined in arch. +# - may override any previously defined variable +# + +# ARCH = +# CROSS = +# MACHINE_CFLAGS = +# MACHINE_LDFLAGS =mk/machine/neoversen1/rte.vars.mk line 33
Quoted 10 lines
mk/machine/neoversen1/rte.vars.mk new file +10
-0+# MACHINE_ASFLAGS = +# CPU_CFLAGS = +# CPU_LDFLAGS = +# CPU_ASFLAGS = + +include $(RTE_SDK)/mk/rte.helper.mk + +MACHINE_CFLAGS += $(call rte_cc_has_argument, +-march=armv8.2-a+crc+crypto) MACHINE_CFLAGS += $(call +rte_cc_has_argument, -mcpu=neoverse-n1)--
Ola Liljedahl, Networking System Architect, Arm Phone +46706866373, Skype ola.liljedahl
Re: [dpdk-dev] [PATCH v2 2/3] config: add arm neoverse N1 SDP configuration
Quoted 36 lines
Quoted 36 lines
<snip>
Quoted 34 lines · Gavin Hu
Quoted 24 lines
Arm N1 SDP is an infrastructure segment development platform based on armv8.2-a Neoverse N1 CPU. For more information, refer to:
https://community.arm.com/developer/tools-software/oss-platforms/w
/
docs/440/neoverse-n1-sdpSigned-off-by Gavin Hu <[email protected]>Reviewed-by Honnappa Nagarahalli <[email protected]>Reviewed-by Steve Capper <[email protected]>--- config/arm/meson.build | 9 ++++++- config/defconfig_arm64-neoversen1-linux-gcc | 1 + config/defconfig_arm64-neoversen1-linuxapp-gcc | 15 ++++++++++++ mk/machine/neoversen1/rte.vars.mk | 34config/defconfig_arm64-neoversen1-linuxapp-gcc new file +0
-0Quoted 1 line
config/defconfig_arm64-neoversen1-linuxapp-gcc line 6
Quoted 6 lines
config/defconfig_arm64-neoversen1-linuxapp-gcc new file +6
-0@@ -0,0 +1,15 @@ +# SPDX-License-Identifier: BSD-3-Clause # Copyright(c) 2019 Arm Ltd. +# + +#include "defconfig_arm64-armv8a-linux-gcc" + +CONFIG_RTE_MACHINE="neoversen1"This should probably be "n1sdp" as this is the name of the platform that matches the below configuration.
Quoted 4 lines
Quoted 1 line
A clear definition of
RTE_MACHINEis required. Jerin?I think, In the existing scheme of things,
RTE_MACHINEdefines, where to take theMACHINE_CFLAGSmk/machine/xxxx/rte.vars.mk
Ok, thank you
Quoted 10 lines
Considering the fact that there will be a lot of reusable IPs(for CPU) from ARM for armv8, I think, it would make sense to introduce
RTE_MICRO_ARCHto avoid a lot of code duplications and confusion.
RTE_ARCHexample: "x86" or "arm64"RTE_MICRO_ARCHexample: "a72" or "thunderx3" - defines
mcpu and armv8 verion arch etcRTE_MACHINEexample: "bluefield" or "thunderx3"
- defines, number of cores, NUMA or not? etc
Looking at mk/machine/ directory, looks like RTE_MACHINE seems to be defining micro-architecture for Intel. For ex: hsw, nhm, wsm. I see the same for Arm as well. Are you suggesting that we use RTE_MICRO_ARCH to pick mk/micro-arch/xxxx/rte.vars.mk? and RTE_MACHINE would pick mk/machine/xxxx/rte.vars.mk, but contain NUMA, #of cores etc?
Quoted 62 lines
RTE_MACHINEshould be probe based on "implementation ID" for arm64 and reuse already definedRTE_MICRO_ARCHto avoid code duplication.Quoted 57 lines
Quoted 56 lines
Quoted 43 lines
config/defconfig_arm64-neoversen1-linuxapp-gcc new file +8
-0+CONFIG_RTE_ARCH_ARM_TUNE="neoverse-n1" +CONFIG_RTE_MAX_LCORE=4 +CONFIG_RTE_MAX_NUMA_NODES=1 +CONFIG_RTE_CACHE_LINE_SIZE=64 + +# Doesn't support NUMA +CONFIG_RTE_EAL_NUMA_AWARE_HUGEPAGES=n +CONFIG_RTE_LIBRTE_VHOST_NUMA=nmk/machine/neoversen1/rte.vars.mk new file +24
-0@@ -0,0 +1,34 @@ +# SPDX-License-Identifier: BSD-3-Clause # Copyright(c) 2019 Arm +Ltd # + +# +# machine: +# +# - can define ARCH variable (overridden by cmdline value) +# - can define CROSS variable (overridden by cmdline value) +# - define MACHINE_CFLAGS variable (overridden by cmdline value) +# - define MACHINE_LDFLAGS variable (overridden by cmdline value) +# - define MACHINE_ASFLAGS variable (overridden by cmdline value) +# - can define CPU_CFLAGS variable (overridden by cmdline value) that +# overrides the one defined in arch. +# - can define CPU_LDFLAGS variable (overridden by cmdline value) that +# overrides the one defined in arch. +# - can define CPU_ASFLAGS variable (overridden by cmdline value) that +# overrides the one defined in arch. +# - may override any previously defined variable +# + +# ARCH = +# CROSS = +# MACHINE_CFLAGS = +# MACHINE_LDFLAGS =mk/machine/neoversen1/rte.vars.mk line 34
Quoted 10 lines
mk/machine/neoversen1/rte.vars.mk new file +10
-0+# MACHINE_ASFLAGS = +# CPU_CFLAGS = +# CPU_LDFLAGS = +# CPU_ASFLAGS = + +include $(RTE_SDK)/mk/rte.helper.mk + +MACHINE_CFLAGS += $(call rte_cc_has_argument, +-march=armv8.2-a+crc+crypto) MACHINE_CFLAGS += $(call +rte_cc_has_argument, -mcpu=neoverse-n1)--
Ola Liljedahl, Networking System Architect, Arm Phone +46706866373, Skype ola.liljedahl
Re: [dpdk-dev] [PATCH v2 2/3] config: add arm neoverse N1 SDP configuration
Quoted 48 lines · Honnappa Nagarahalli
Quoted 37 lines
Quoted 37 lines
<snip>
Quoted 34 lines · Gavin Hu
Quoted 24 lines
Arm N1 SDP is an infrastructure segment development platform based on armv8.2-a Neoverse N1 CPU. For more information, refer to:
https://community.arm.com/developer/tools-software/oss-platforms/w
/
docs/440/neoverse-n1-sdpSigned-off-by Gavin Hu <[email protected]>Reviewed-by Honnappa Nagarahalli <[email protected]>Reviewed-by Steve Capper <[email protected]>--- config/arm/meson.build | 9 ++++++- config/defconfig_arm64-neoversen1-linux-gcc | 1 + config/defconfig_arm64-neoversen1-linuxapp-gcc | 15 ++++++++++++ mk/machine/neoversen1/rte.vars.mk | 34config/defconfig_arm64-neoversen1-linuxapp-gcc new file +0
-0Quoted 1 line
config/defconfig_arm64-neoversen1-linuxapp-gcc line 6
Quoted 6 lines
config/defconfig_arm64-neoversen1-linuxapp-gcc new file +6
-0@@ -0,0 +1,15 @@ +# SPDX-License-Identifier: BSD-3-Clause # Copyright(c) 2019 Arm Ltd. +# + +#include "defconfig_arm64-armv8a-linux-gcc" + +CONFIG_RTE_MACHINE="neoversen1"This should probably be "n1sdp" as this is the name of the platform that matches the below configuration.
Quoted 4 lines
Quoted 1 line
A clear definition of
RTE_MACHINEis required. Jerin?I think, In the existing scheme of things,
RTE_MACHINEdefines, where to take theMACHINE_CFLAGSmk/machine/xxxx/rte.vars.mkOk, thank you
Quoted 5 lines
Considering the fact that there will be a lot of reusable IPs(for CPU) from ARM for armv8, I think, it would make sense to introduce
RTE_MICRO_ARCHto avoid a lot of code duplications and confusion.
Quoted 7 lines
Quoted 5 lines
RTE_ARCHexample: "x86" or "arm64"RTE_MICRO_ARCHexample: "a72" or "thunderx3" - defines
mcpu and armv8 verion arch etcRTE_MACHINEexample: "bluefield" or "thunderx3"
- defines, number of cores, NUMA or not? etcLooking at mk/machine/ directory, looks like
RTE_MACHINEseems to be defining micro-architecture for Intel. For ex: hsw, nhm, wsm. I see the same for Arm as well. Are you suggesting that we useRTE_MICRO_ARCHto pickmk/micro-arch/xxxx/rte.vars.mk? andRTE_MACHINEwould pickmk/machine/xxxx/rte.vars.mk, but contain NUMA, #of cores etc?
Yes for Make build. I think, it is deprecated soon, so we need a similar solution for meson.
Quoted 63 lines
Quoted 63 lines
RTE_MACHINEshould be probe based on "implementation ID" for arm64 and reuse already definedRTE_MICRO_ARCHto avoid code duplication.Quoted 57 lines
Quoted 56 lines
Quoted 43 lines
config/defconfig_arm64-neoversen1-linuxapp-gcc new file +8
-0+CONFIG_RTE_ARCH_ARM_TUNE="neoverse-n1" +CONFIG_RTE_MAX_LCORE=4 +CONFIG_RTE_MAX_NUMA_NODES=1 +CONFIG_RTE_CACHE_LINE_SIZE=64 + +# Doesn't support NUMA +CONFIG_RTE_EAL_NUMA_AWARE_HUGEPAGES=n +CONFIG_RTE_LIBRTE_VHOST_NUMA=nmk/machine/neoversen1/rte.vars.mk new file +24
-0@@ -0,0 +1,34 @@ +# SPDX-License-Identifier: BSD-3-Clause # Copyright(c) 2019 Arm +Ltd # + +# +# machine: +# +# - can define ARCH variable (overridden by cmdline value) +# - can define CROSS variable (overridden by cmdline value) +# - define MACHINE_CFLAGS variable (overridden by cmdline value) +# - define MACHINE_LDFLAGS variable (overridden by cmdline value) +# - define MACHINE_ASFLAGS variable (overridden by cmdline value) +# - can define CPU_CFLAGS variable (overridden by cmdline value) that +# overrides the one defined in arch. +# - can define CPU_LDFLAGS variable (overridden by cmdline value) that +# overrides the one defined in arch. +# - can define CPU_ASFLAGS variable (overridden by cmdline value) that +# overrides the one defined in arch. +# - may override any previously defined variable +# + +# ARCH = +# CROSS = +# MACHINE_CFLAGS = +# MACHINE_LDFLAGS =mk/machine/neoversen1/rte.vars.mk line 34
Quoted 10 lines
mk/machine/neoversen1/rte.vars.mk new file +10
-0+# MACHINE_ASFLAGS = +# CPU_CFLAGS = +# CPU_LDFLAGS = +# CPU_ASFLAGS = + +include $(RTE_SDK)/mk/rte.helper.mk + +MACHINE_CFLAGS += $(call rte_cc_has_argument, +-march=armv8.2-a+crc+crypto) MACHINE_CFLAGS += $(call +rte_cc_has_argument, -mcpu=neoverse-n1)--
Ola Liljedahl, Networking System Architect, Arm Phone +46706866373, Skype ola.liljedahl
Re: [dpdk-dev] [PATCH v2 2/3] config: add arm neoverse N1 SDP configuration
23/10/2019 07:03, Jerin Jacob:
Quoted 32 lines
Quoted 30 lines · Honnappa Nagarahalli
Quoted 12 lines
Quoted 12 lines
Quoted 12 lines · Gavin Hu
Quoted 11 lines
Arm N1 SDP is an infrastructure segment development platform based on armv8.2-a Neoverse N1 CPU. For more information, refer to:
https://community.arm.com/developer/tools-software/oss-platforms/w
/
docs/440/neoverse-n1-sdpSigned-off-by Gavin Hu <[email protected]>Reviewed-by Honnappa Nagarahalli <[email protected]>Reviewed-by Steve Capper <[email protected]>---
+CONFIG_RTE_MACHINE="neoversen1"Quoted 6 lines
Quoted 3 lines
Quoted 2 lines
This should probably be "n1sdp" as this is the name of the platform that matches the below configuration.
A clear definition of
RTE_MACHINEis required. Jerin?I think, In the existing scheme of things,
RTE_MACHINEdefines, where to take theMACHINE_CFLAGSmk/machine/xxxx/rte.vars.mkOk, thank you
Quoted 10 lines
Considering the fact that there will be a lot of reusable IPs(for CPU) from ARM for armv8, I think, it would make sense to introduce
RTE_MICRO_ARCHto avoid a lot of code duplications and confusion.
RTE_ARCHexample: "x86" or "arm64"RTE_MICRO_ARCHexample: "a72" or "thunderx3" - defines
mcpu and armv8 verion arch etcRTE_MACHINEexample: "bluefield" or "thunderx3"
- defines, number of cores, NUMA or not? etc
Quoted 5 lines
Quoted 2 lines
Looking at mk/machine/ directory, looks like
RTE_MACHINEseems to be defining micro-architecture for Intel. For ex: hsw, nhm, wsm. I see the same for Arm as well. Are you suggesting that we useRTE_MICRO_ARCHto pickmk/micro-arch/xxxx/rte.vars.mk? andRTE_MACHINEwould pickmk/machine/xxxx/rte.vars.mk, but contain NUMA, #of cores etc?Yes for Make build. I think, it is deprecated soon, so we need a similar solution for meson.
Yes I would prefer we clean the mess in Meson, instead of talking about the makefile system.
And honestly, N1 is not needed in the legacy makefile system.
So focusing on config/arm/meson.build,
I think RTE_MACHINE is defined only for API compatibility with makefile. However, I doubt this value is used by any application.
I think we can try to remove RTE_MACHINE from meson builds in DPDK 19.11, or use RTE_MACHINE as micro-arch (my preference).
Re: [dpdk-dev] [PATCH v2 2/3] config: add arm neoverse N1 SDP configuration
Quoted 44 lines
23/10/2019 07:03, Jerin Jacob:
Quoted 5 lines
Quoted 3 lines · Honnappa Nagarahalli
Quoted 3 lines
Quoted 3 lines
Quoted 3 lines · Gavin Hu
Quoted 2 lines
Arm N1 SDP is an infrastructure segment development platform based on armv8.2-a Neoverse N1 CPU. For more information, refer
to:
Quoted 15 lines
Quoted 15 lines
Quoted 15 lines
Quoted 14 lines
Quoted 5 lines
Quoted 5 lines
https://community.arm.com/developer/tools-software/oss-platf
orms/w
/
docs/440/neoverse-n1-sdpQuoted 8 lines
Quoted 6 lines
Signed-off-by Gavin Hu <[email protected]>Reviewed-by Honnappa NagarahalliReviewed-by Steve Capper <[email protected]>---
+CONFIG_RTE_MACHINE="neoversen1"This should probably be "n1sdp" as this is the name of the platform that matches the below configuration.
A clear definition of
RTE_MACHINEis required. Jerin?Quoted 8 lines
Quoted 8 lines
Quoted 2 lines
I think, In the existing scheme of things,
RTE_MACHINEdefines, where to take theMACHINE_CFLAGSmk/machine/xxxx/rte.vars.mkOk, thank you
Quoted 4 lines
Considering the fact that there will be a lot of reusable IPs(for CPU) from ARM for armv8, I think, it would make sense to introduce
RTE_MICRO_ARCHto avoid a lot of code duplications andconfusion.
Quoted 7 lines
Quoted 7 lines
Quoted 6 lines
RTE_ARCHexample: "x86" or "arm64"RTE_MICRO_ARCHexample: "a72" or "thunderx3" - defines
mcpu and armv8 verion arch etcRTE_MACHINEexample: "bluefield" or "thunderx3"
- defines, number of cores, NUMA or not? etcLooking at mk/machine/ directory, looks like
RTE_MACHINEseems to bedefining micro-architecture for Intel. For ex: hsw, nhm, wsm. I see the same for Arm as well.
Quoted 1 line
Quoted 1 line
Are you suggesting that we use
RTE_MICRO_ARCHto pick mk/micro-
arch/xxxx/rte.vars.mk? andRTE_MACHINEwould pickmk/machine/xxxx/rte.vars.mk, but contain NUMA, #of cores etc?Quoted 1 line
Quoted 6 lines
Quoted 2 lines
Yes for Make build. I think, it is deprecated soon, so we need a similar solution for meson.
Yes I would prefer we clean the mess in Meson, instead of talking about the makefile system.
And honestly, N1 is not needed in the legacy makefile system.
Unfortunately, most of the guys I talk to are still on makefile. When is makefile system getting removed?
Quoted 6 lines
So focusing on config/arm/meson.build,
I thinkRTE_MACHINEis defined only for API compatibility with makefile. However, I doubt this value is used by any application.
I think we can try to removeRTE_MACHINEfrom meson builds in DPDK 19.11, or useRTE_MACHINEas micro-arch (my preference).
'MACHINE' means different things to different people, which is the root cause of this discussion. 'MICRO-ARCH' has a very clear meaning. Do you see any problem going with MICRO-ARCH instead?
Quoted 1 line
Re: [dpdk-dev] [PATCH v2 2/3] config: add arm neoverse N1 SDP configuration
28/10/2019 04:24, Honnappa Nagarahalli:
Quoted 44 lines
Quoted 44 lines
23/10/2019 07:03, Jerin Jacob:
Quoted 5 lines
Quoted 3 lines · Honnappa Nagarahalli
Quoted 3 lines
Quoted 3 lines
Quoted 3 lines · Gavin Hu
Quoted 2 lines
Arm N1 SDP is an infrastructure segment development platform based on armv8.2-a Neoverse N1 CPU. For more information, refer
to:
Quoted 15 lines
Quoted 15 lines
Quoted 15 lines
Quoted 14 lines
Quoted 5 lines
Quoted 5 lines
https://community.arm.com/developer/tools-software/oss-platf
orms/w
/
docs/440/neoverse-n1-sdpQuoted 8 lines
Quoted 6 lines
Signed-off-by Gavin Hu <[email protected]>Reviewed-by Honnappa NagarahalliReviewed-by Steve Capper <[email protected]>---
+CONFIG_RTE_MACHINE="neoversen1"This should probably be "n1sdp" as this is the name of the platform that matches the below configuration.
A clear definition of
RTE_MACHINEis required. Jerin?Quoted 8 lines
Quoted 8 lines
Quoted 2 lines
I think, In the existing scheme of things,
RTE_MACHINEdefines, where to take theMACHINE_CFLAGSmk/machine/xxxx/rte.vars.mkOk, thank you
Quoted 4 lines
Considering the fact that there will be a lot of reusable IPs(for CPU) from ARM for armv8, I think, it would make sense to introduce
RTE_MICRO_ARCHto avoid a lot of code duplications andconfusion.
Quoted 7 lines
Quoted 7 lines
Quoted 6 lines
RTE_ARCHexample: "x86" or "arm64"RTE_MICRO_ARCHexample: "a72" or "thunderx3" - defines
mcpu and armv8 verion arch etcRTE_MACHINEexample: "bluefield" or "thunderx3"
- defines, number of cores, NUMA or not? etcLooking at mk/machine/ directory, looks like
RTE_MACHINEseems to bedefining micro-architecture for Intel. For ex: hsw, nhm, wsm. I see the same for Arm as well.
Quoted 1 line
Quoted 1 line
Are you suggesting that we use
RTE_MICRO_ARCHto pick mk/micro-
arch/xxxx/rte.vars.mk? andRTE_MACHINEwould pickmk/machine/xxxx/rte.vars.mk, but contain NUMA, #of cores etc?Quoted 1 line
Quoted 7 lines
Quoted 6 lines
Quoted 2 lines
Yes for Make build. I think, it is deprecated soon, so we need a similar solution for meson.
Yes I would prefer we clean the mess in Meson, instead of talking about the makefile system.
And honestly, N1 is not needed in the legacy makefile system.Unfortunately, most of the guys I talk to are still on makefile.
You need to help them to switch.
Adding new targets in meson-only can be a good motivation :)
Quoted 1 line
When is makefile system getting removed?
It must be clearly decided and announced.
The previous techboard discussions were about making makefile hardly usable during 2020, i.e. very soon.
Quoted 7 lines
Quoted 5 lines
So focusing on config/arm/meson.build,
I thinkRTE_MACHINEis defined only for API compatibility with makefile. However, I doubt this value is used by any application.
I think we can try to removeRTE_MACHINEfrom meson builds in DPDK 19.11, or useRTE_MACHINEas micro-arch (my preference).'MACHINE' means different things to different people, which is the root cause of this discussion. 'MICRO-ARCH' has a very clear meaning. Do you see any problem going with MICRO-ARCH instead?
Some applications may use RTE_MACHINE for this purpose. It is part of the API since the befinning of DPDK. I don't see a real motivation to break this API now.
Re: [dpdk-dev] [PATCH v2 2/3] config: add arm neoverse N1 SDP configuration
<snip>
Quoted 40 lines
28/10/2019 04:24, Honnappa Nagarahalli:
Quoted 38 lines
Quoted 38 lines
23/10/2019 07:03, Jerin Jacob:
Quoted 6 lines
Quoted 4 lines · Honnappa Nagarahalli
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines · Gavin Hu
Quoted 3 lines
Arm N1 SDP is an infrastructure segment development platform based on armv8.2-a Neoverse N1 CPU. For more information, refer
to:
Quoted 19 lines
Quoted 19 lines
Quoted 19 lines
Quoted 12 lines
Quoted 12 lines
Quoted 12 lines
https://community.arm.com/developer/tools-software/oss-p
latf
orms/w
/
docs/440/neoverse-n1-sdpSigned-off-by Gavin Hu <[email protected]>Reviewed-by Honnappa NagarahalliReviewed-by Steve Capper <[email protected]>---
+CONFIG_RTE_MACHINE="neoversen1"Quoted 3 lines
Quoted 2 lines
This should probably be "n1sdp" as this is the name of the platform that matches the below configuration.
A clear definition of
RTE_MACHINEis required. Jerin?I think, In the existing scheme of things,
RTE_MACHINEdefines, where to take theMACHINE_CFLAGSmk/machine/xxxx/rte.vars.mkQuoted 8 lines
Quoted 8 lines
Ok, thank you
Quoted 6 lines
Considering the fact that there will be a lot of reusable IPs(for
CPU) from ARM for armv8, I think, it would make sense to introduceRTE_MICRO_ARCHto avoid a lot of code duplications andconfusion.
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
RTE_ARCHexample: "x86" or "arm64"
I see that there are already RTE_ARCH_X86, RTE_ARCH_ARM, RTE_ARCH_ARM64, RTE_ARCH_PPC_64 etc. I believe they should be sufficient.
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
RTE_MICRO_ARCHexample: "a72" or "thunderx3" - defines
mcpu and armv8 verion arch etc
Are you suggesting this just for Arm platforms?
My understanding is your intention was to clean up the config/arm/meson.build file.
Quoted 32 lines
Quoted 10 lines
Quoted 10 lines
Quoted 4 lines
Quoted 4 lines
Quoted 2 lines
RTE_MACHINEexample: "bluefield" or "thunderx3"
- defines, number of cores, NUMA or not? etcLooking at mk/machine/ directory, looks like
RTE_MACHINEseems to bedefining micro-architecture for Intel. For ex: hsw, nhm, wsm. I see the same for Arm as well.
Quoted 1 line
Quoted 1 line
Are you suggesting that we use
RTE_MICRO_ARCHto pick mk/micro-
arch/xxxx/rte.vars.mk? andRTE_MACHINEwould pickmk/machine/xxxx/rte.vars.mk, but contain NUMA, #of cores etc?Quoted 1 line
Quoted 7 lines
Quoted 6 lines
Quoted 2 lines
Yes for Make build. I think, it is deprecated soon, so we need a similar solution for meson.
Yes I would prefer we clean the mess in Meson, instead of talking about the makefile system.
And honestly, N1 is not needed in the legacy makefile system.Unfortunately, most of the guys I talk to are still on makefile.
You need to help them to switch.
Adding new targets in meson-only can be a good motivation :)Quoted 1 line
When is makefile system getting removed?
It must be clearly decided and announced.
The previous techboard discussions were about making makefile hardly usable during 2020, i.e. very soon.Quoted 5 lines
Quoted 5 lines
So focusing on config/arm/meson.build, I think
RTE_MACHINEis defined only for API compatibility with makefile. However, I doubt this value is used by any application. I think we can try to removeRTE_MACHINEfrom meson builds in DPDK 19.11, or useRTE_MACHINEas micro-arch (my preference).
Quoted 8 lines
Quoted 1 line
'MACHINE' means different things to different people, which is the root
cause of this discussion.
Quoted 1 line
'MICRO-ARCH' has a very clear meaning. Do you see any problem going
with MICRO-ARCH instead?
Some applications may use
RTE_MACHINEfor this purpose. It is part of the API since the befinning of DPDK. I don't see a real motivation to break this API now.
The suggestions are not clear to me. The original suggestion was to introduce RTE_MICRO_ARCH and contain all the micro-architecture related compiler flags in that.
Now, the suggestion is to use RTE_MACHINE to contain micro-architecture related compiler flags. Will it contain NUMA, number of cores as well (as suggested earlier)? If yes, I do not see it changing anything.
I am not a meson expert. However, I looked at various meson.build files. I have few questions/concerns.
1) Are these suggestions are for all the platforms? IMO, these need to be the same across all the architectures.
2) I am looking at config/meson.build. Here RTE_MACHINE is defined to indicate Architecture for Arm (armv7-a, aarchxx) and micro-architecture for x86 (corei7?). Is the understanding correct? IMO, this should this have a common meaning across all the platforms?
3) If the changes are for all the platforms, is the risk high for 19.11 release?
4) The N1 config patch as such conforms to the current conventions. What is being asked here is an enhancement, is the understanding correct?
Quoted 1 line
Re: [dpdk-dev] [PATCH v2 2/3] config: add arm neoverse N1 SDP configuration
Hi Honnappa,
Quoted 116 lines
-----Original Message-----
From: Honnappa Nagarahalli <[email protected]>
Sent: Tuesday, October 29, 2019 1:47 PM
To: [email protected]
Cc: Jerin Jacob <[email protected]>; Ola Liljedahl <[email protected]>; Gavin Hu (Arm Technology China) <[email protected]>; [email protected]; [email protected]; nd <[email protected]>; [email protected]; [email protected]; [email protected]; Honnappa Nagarahalli <[email protected]>; nd <[email protected]>
Subject: RE: [dpdk-dev] [PATCH v2 2/3] config: add arm neoverse N1 SDP configuration<snip>
Quoted 11 lines
28/10/2019 04:24, Honnappa Nagarahalli:
Quoted 9 lines
Quoted 9 lines
23/10/2019 07:03, Jerin Jacob:
Quoted 6 lines
Quoted 4 lines · Honnappa Nagarahalli
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines · Gavin Hu
Quoted 3 lines
Arm N1 SDP is an infrastructure segment development platform based on armv8.2-a Neoverse N1 CPU. For more information, refer
to:
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
p
Quoted 21 lines
Quoted 21 lines
Quoted 21 lines
Quoted 21 lines
Quoted 21 lines
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines
Quoted 11 lines
latf
orms/w
/
docs/440/neoverse-n1-sdpSigned-off-by Gavin Hu <[email protected]>Reviewed-by Honnappa NagarahalliReviewed-by Steve Capper <[email protected]>---
+CONFIG_RTE_MACHINE="neoversen1"Quoted 7 lines
Quoted 3 lines
Quoted 2 lines
This should probably be "n1sdp" as this is the name of the platform that matches the below configuration.
A clear definition of
RTE_MACHINEis required. Jerin?I think, In the existing scheme of things,
RTE_MACHINEdefines, where to take theMACHINE_CFLAGSmk/machine/xxxx/rte.vars.mkOk, thank you
Quoted 1 line
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
Considering the fact that there will be a lot of reusable IPs(for
CPU) from ARM for armv8, I think, it would make sense to introduceRTE_MICRO_ARCHto avoid a lot of code duplications andconfusion.
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
RTE_ARCHexample: "x86" or "arm64"I see that there are already
RTE_ARCH_X86,RTE_ARCH_ARM,RTE_ARCH_ARM64,RTE_ARCH_PPC_64etc. I believe they should be sufficient.Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
RTE_MICRO_ARCHexample: "a72" or "thunderx3" -defines
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
mcpu and armv8 verion arch etc
Are you suggesting this just for Arm platforms? My understanding is your intention was to clean up the config/arm/meson.build file.
Quoted 7 lines
Quoted 7 lines
Quoted 7 lines
Quoted 4 lines
Quoted 4 lines
Quoted 2 lines
RTE_MACHINEexample: "bluefield" or "thunderx3"
- defines, number of cores, NUMA or not? etcLooking at mk/machine/ directory, looks like
RTE_MACHINEseems to bedefining micro-architecture for Intel. For ex: hsw, nhm, wsm. I see the same for Arm as well.
Quoted 1 line
Quoted 1 line
Are you suggesting that we use
RTE_MICRO_ARCHto pickmk/micro-
Quoted 11 lines
Quoted 10 lines
Quoted 9 lines
arch/xxxx/rte.vars.mk? andRTE_MACHINEwould pickmk/machine/xxxx/rte.vars.mk, but contain NUMA, #of cores etc?Quoted 3 lines
Yes for Make build. I think, it is deprecated soon, so we need a similar solution for meson.
Yes I would prefer we clean the mess in Meson, instead of talking about the makefile system.
And honestly, N1 is not needed in the legacy makefile system.Unfortunately, most of the guys I talk to are still on makefile.
Quoted 13 lines
You need to help them to switch.
Adding new targets in meson-only can be a good motivation :)Quoted 1 line
When is makefile system getting removed?
It must be clearly decided and announced.
The previous techboard discussions were about making makefile hardly usable during 2020, i.e. very soon.Quoted 4 lines
Quoted 4 lines
So focusing on config/arm/meson.build, I think
RTE_MACHINEis defined only for API compatibility with makefile. However, I doubt this value is used by any application. I think we can try to removeRTE_MACHINEfrom meson builds inDPDK
Quoted 9 lines
Quoted 2 lines
Quoted 1 line
19.11, or use
RTE_MACHINEas micro-arch (my preference).'MACHINE' means different things to different people, which is the root
cause of this discussion.
Quoted 1 line
'MICRO-ARCH' has a very clear meaning. Do you see any problem going
with MICRO-ARCH instead?
Some applications may use
RTE_MACHINEfor this purpose. It is part of the API since the befinning of DPDK. I don't see a real motivation to break this API now.The suggestions are not clear to me. The original suggestion was to introduce
RTE_MICRO_ARCHand contain all the micro-architecture related compiler flags in that.
Now, the suggestion is to useRTE_MACHINEto contain micro-architecture related compiler flags. Will it contain NUMA, number of cores as well (as suggested earlier)? If yes, I do not see it changing anything.
Quoted 12 lines
I am not a meson expert. However, I looked at various meson.build files. I have few questions/concerns.
1) Are these suggestions are for all the platforms? IMO, these need to be the same across all the architectures.
2) I am looking at config/meson.build. HereRTE_MACHINEis defined to indicate Architecture for Arm (armv7-a, aarchxx) and micro-architecture for x86 (corei7?). Is the understanding correct? IMO, this should this have a common meaning across all the platforms?
3) If the changes are for all the platforms, is the risk high for 19.11 release?
4) The N1 config patch as such conforms to the current conventions. What is being asked here is an enhancement, is the understanding correct?
We consider it is an enhancement to introduce the RTE_MICRO_ARCH and I understand it is a high risk and we did not agree here yet. I will submit v3 to change the configuration name from neoverse-n1 to n1sdp, to address Honnappa's comment. /Gavin
PATCH v2 3/3 config: add cortex-a76 configuration
In reply to Gavin HuTo make the list complete and consistent, add cortex-a76 configuration.
--- config/arm/meson.build | 1 + 1 file changed, 1 insertion(+)
config/arm/meson.build +1 -0
@@ -94,6 +94,7 @@ machine_args_generic = [ ['0xd08', ['-mcpu=cortex-a72']], ['0xd09', ['-mcpu=cortex-a73']], ['0xd0a', ['-mcpu=cortex-a75']], + ['0xd0b', ['-mcpu=cortex-a76']], ['0xd0c', ['-march=armv8.2-a+crc+crypto', '-mcpu=neoverse-n1'], flags_neoversen1_extra]] machine_args_cavium = [
Signature · 1 line
2.7.4
v3 cover add arm N1SDP and A76 configurations
In reply to Gavin HuNerverse N1 is a new SoC based on armv8.2-a and with other advanced features. N1 SDP is the infrastructure segment development platform with N1 SoC inside. Here is more information about it:
https://community.arm.com/developer/tools-software/oss-platforms/w/docs/440/neoverse-n1-sdp
V3
-change the configuration name from "neoversen1" to "n1sdp" to be platform specific other than microarchitecture specific -add the missing config/arm/arm64_n1sdp_linux-gcc file for meson build
V2:
-add the cover letter and the patch for cortex-a76 configuration
Gavin Hu (3):
test/rcu: fix the compiling error for armv8.2
config: add arm neoverse N1 SDP configuration
config: add cortex-a76 configuration
app/test/test_rcu_qsbr.c | 2 +- config/arm/arm64_n1sdp_linux_gcc | 16 +++++++++++++++ config/arm/meson.build | 10 ++++++++- config/defconfig_arm64-n1sdp-linux-gcc | 1 + config/defconfig_arm64-n1sdp-linuxapp-gcc | 14 +++++++++++++ mk/machine/n1sdp/rte.vars.mk | 34 +++++++++++++++++++++++++++++++ 6 files changed, 75 insertions(+), 2 deletions(-) create mode 100644 config/arm/arm64_n1sdp_linux_gcc create mode 120000 config/defconfig_arm64-n1sdp-linux-gcc create mode 100644 config/defconfig_arm64-n1sdp-linuxapp-gcc create mode 100644 mk/machine/n1sdp/rte.vars.mk
Signature · 1 line
2.7.4
PATCH v3 1/3 test/rcu: fix the compiling error for armv8.2
With "-march=armv8.2-a" specified, a compiling error generated:app/test/test_rcu_qsbr.c:234:10: error: comparison of integer expressions of different signedness: ‘unsigned int’ and ‘int’ [-Werror=sign-compare]
--- app/test/test_rcu_qsbr.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
app/test/test_rcu_qsbr.c +1 -1
@@ -230,7 +230,7 @@ test_rcu_qsbr_thread_unregister(void)/* Update quiescent state counter */ for (i =0 ; i < num_threads[j]; i++) {/* Skip one update */ -if (i == (RTE_MAX_LCORE -10 )) +if (i == (unsigned int )(RTE_MAX_LCORE -10 ))continue ;rte_rcu_qsbr_quiescent (t[0 ], (j ==2 ) ? (RTE_MAX_LCORE -1 ) : i);
Signature · 1 line
2.7.4
11/11/2019 06:41, Gavin Hu:
Quoted 14 lines
With "-march=armv8.2-a" specified, a compiling error generated:
app/test/test_rcu_qsbr.c:234:10: error: comparison of integer expressions of different signedness: ‘unsigned int’ and ‘int’ [-Werror=sign-compare]Fixes b87089b0bb19 ("test/rcu: add API and functional tests")Signed-off-by Gavin Hu <[email protected]>Reviewed-by Honnappa Nagarahalli <[email protected]>Reviewed-by Steve Capper <[email protected]>---
app/test/test_rcu_qsbr.c test_rcu_qsbr_thread_unregister(void) · line 233
Quoted 3 lines
app/test/test_rcu_qsbr.c +1
-1/* Skip one update */ -if (i == (RTE_MAX_LCORE -10 )) +if (i == (unsigned int )(RTE_MAX_LCORE -10 ))
For N1 SDP, RTE_MAX_LCORE = 4.
So this test weird.
Why not using rand() to find a number between 0 and max ?
You are hiding a real failure with a forced type casting.
Please reviewers, explain why you think it is the right thing to do?
Quoted 36 lines
-----Original Message-----
From: Thomas Monjalon <[email protected]>
Sent: Thursday, November 21, 2019 6:19 AM
To: Gavin Hu (Arm Technology China) <[email protected]>; Honnappa Nagarahalli <[email protected]>; Steve Capper <[email protected]>
Cc: [email protected]; [email protected]; nd <[email protected]>; [email protected]
Subject: Re: [dpdk-stable] [PATCH v3 1/3] test/rcu: fix the compiling error for armv8.211/11/2019 06:41, Gavin Hu:
Quoted 14 lines
With "-march=armv8.2-a" specified, a compiling error generated:
app/test/test_rcu_qsbr.c:234:10: error: comparison of integer expressions of different signedness: ‘unsigned int’ and ‘int’ [-Werror=sign-compare]Fixes b87089b0bb19 ("test/rcu: add API and functional tests")Signed-off-by Gavin Hu <[email protected]>Reviewed-by Honnappa Nagarahalli <[email protected]>Reviewed-by Steve Capper <[email protected]>---
app/test/test_rcu_qsbr.c test_rcu_qsbr_thread_unregister(void) · line 233
Quoted 3 lines
app/test/test_rcu_qsbr.c +1
-1/* Skip one update */ -if (i == (RTE_MAX_LCORE -10 )) +if (i == (unsigned int )(RTE_MAX_LCORE -10 ))For N1 SDP,
RTE_MAX_LCORE= 4.
So this test weird.
Why not usingrand()to find a number between 0 and max ?
Quoted 4 lines
You are hiding a real failure with a forced type casting.
Please reviewers, explain why you think it is the right thing to do?
Thanks for pointing out this issue, will fix in next release.
PATCH v4 1/1 test/rcu: fix build for small number of cores
In reply to Gavin HuIf the RTE_MAX_LCORE is less than 10, a compilation error is generated:app/test/test_rcu_qsbr.c:234:10: error: comparison of integer expressions of different signedness: ‘unsigned int’ and ‘int’ [-Werror=sign-compare]
The cause is (RTE_MAX_LCORE - 10) results in a negative value.
To fix, use rte_rand() to find a number between 0 and RTE_MAX_LCORE.
---
v4:
- separate from the N1SDP configuration patch series
- fix the build for smaller number of cores(Thomas Monjalon)
--- app/test/test_rcu_qsbr.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-)
app/test/test_rcu_qsbr.c +5 -2
@@ -9,6 +9,7 @@#include <rte_hash_crc.h> #include <rte_malloc.h> #include <rte_cycles.h> +#include <rte_random.h> #include <unistd.h> #include "test.h"
@@ -168,6 +169,7 @@ test_rcu_qsbr_thread_unregister(void) {unsigned int num_threads[3 ] = {1 ,RTE_MAX_LCORE ,1 };unsigned int i, j; +unsigned int skip_thread_id;uint64_t token;int ret;
@@ -227,10 +229,11 @@ test_rcu_qsbr_thread_unregister(void) token =rte_rcu_qsbr_start (t[0 ]);TEST_RCU_QSBR_RETURN_IF_ERROR ( (token != (TEST_RCU_QSBR_CNT_INIT +1 )),"QSBR Start" ); + skip_thread_id =rte_rand () %RTE_MAX_LCORE ;/* Update quiescent state counter */ for (i =0 ; i < num_threads[j]; i++) {/* Skip one update */ -if (i == (RTE_MAX_LCORE -10 )) +if ((j ==1 ) && (i == skip_thread_id))continue ;rte_rcu_qsbr_quiescent (t[0 ], (j ==2 ) ? (RTE_MAX_LCORE -1 ) : i);
@@ -242,7 +245,7 @@ test_rcu_qsbr_thread_unregister(void)TEST_RCU_QSBR_RETURN_IF_ERROR ((ret ==0 ),"Non-blocking QSBR check" );/* Update the previously skipped thread */ -rte_rcu_qsbr_quiescent (t[0 ],RTE_MAX_LCORE -10 ); +rte_rcu_qsbr_quiescent (t[0 ], skip_thread_id); }/* Validate the updates */
Signature · 1 line
2.7.4
Thomas Monjalon
AppliedRe: [dpdk-dev] [dpdk-stable] [PATCH v4] test/rcu: fix build for small number of cores
25/11/2019 04:36, Gavin Hu:
Quoted 9 lines
If the
RTE_MAX_LCOREis less than 10, a compilation error is generated:app/test/test_rcu_qsbr.c:234:10: error: comparison of integer expressions of different signedness: ‘unsigned int’ and ‘int’ [-Werror=sign-compare]The cause is (
RTE_MAX_LCORE- 10) results in a negative value.To fix, use
rte_rand()to find a number between 0 andRTE_MAX_LCORE.
Quoted 7 lines
Fixes b87089b0bb19 ("test/rcu: add API and functional tests")Signed-off-by Gavin Hu <[email protected]>Reviewed-by Honnappa Nagarahalli <[email protected]>Reviewed-by Steve Capper <[email protected]>Reviewed-by Dharmik Thakkar <[email protected]>
Applied, thanks
PATCH v3 2/3 config: add arm neoverse N1 SDP configuration
In reply to Gavin HuArm N1 SDP is an infrastructure segment development platform based on armv8.2-a Neoverse N1 CPU. For more information, refer to:
https://community.arm.com/developer/tools-software/oss-platforms/w/
docs/440/neoverse-n1-sdp
---
V3
-change the configuration name from "neoversen1" to "n1sdp" to be platform specific other than microarchitecture specific
-add the missing config/arm/arm64_n1sdp_linux-gcc file for meson build
--- config/arm/arm64_n1sdp_linux_gcc | 16 +++++++++++++++ config/arm/meson.build | 9 +++++++- config/defconfig_arm64-n1sdp-linux-gcc | 1 + config/defconfig_arm64-n1sdp-linuxapp-gcc | 14 +++++++++++++ mk/machine/n1sdp/rte.vars.mk | 34 +++++++++++++++++++++++++++++++ 5 files changed, 73 insertions(+), 1 deletion(-) create mode 100644 config/arm/arm64_n1sdp_linux_gcc create mode 120000 config/defconfig_arm64-n1sdp-linux-gcc create mode 100644 config/defconfig_arm64-n1sdp-linuxapp-gcc create mode 100644 mk/machine/n1sdp/rte.vars.mk
config/arm/arm64_n1sdp_linux_gcc new file +16 -0
@@ -0,0 +1,16 @@ +[binaries] +c = 'aarch64-linux-gnu-gcc' +cpp = 'aarch64-linux-gnu-cpp' +ar = 'aarch64-linux-gnu-gcc-ar' +strip = 'aarch64-linux-gnu-strip' +pcap-config = '' + +[host_machine] +system = 'linux' +cpu_family = 'aarch64' +cpu = 'armv8-a' +endian = 'little' + +[properties] +implementor_id = '0x41' +implementor_pn = '0xd0c'
config/arm/meson.build +8 -1
@@ -57,6 +57,12 @@ flags_armada = [ ['RTE_MAX_LCORE', 16]] flags_default_extra = [] +flags_n1sdp_extra = [ + ['RTE_MACHINE', '"n1sdp"'], + ['RTE_MAX_NUMA_NODES', 1], + ['RTE_MAX_LCORE', 4], + ['RTE_EAL_NUMA_AWARE_HUGEPAGES', false], + ['RTE_LIBRTE_VHOST_NUMA', false]] flags_thunderx_extra = [ ['RTE_MACHINE', '"thunderx"'], ['RTE_USE_C11_MEM_MODEL', false]]
@@ -83,7 +89,8 @@ machine_args_generic = [ ['0xd07', ['-mcpu=cortex-a57']], ['0xd08', ['-mcpu=cortex-a72']], ['0xd09', ['-mcpu=cortex-a73']], - ['0xd0a', ['-mcpu=cortex-a75']]] + ['0xd0a', ['-mcpu=cortex-a75']], + ['0xd0c', ['-march=armv8.2-a+crc+crypto', '-mcpu=neoverse-n1'], flags_n1sdp_extra]] machine_args_cavium = [ ['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
config/defconfig_arm64-n1sdp-linux-gcc new file +1 -0
@@ -0,0 +1 @@ +defconfig_arm64-n1sdp-linuxapp-gcc
config/defconfig_arm64-n1sdp-linuxapp-gcc new file +14 -0
@@ -0,0 +1,14 @@ +# SPDX-License-Identifier: BSD-3-Clause +# Copyright(c) 2019 Arm Ltd. +# + +#include "defconfig_arm64-armv8a-linux-gcc" + +CONFIG_RTE_MACHINE="n1sdp" +CONFIG_RTE_MAX_LCORE=4 +CONFIG_RTE_MAX_NUMA_NODES=1 +CONFIG_RTE_CACHE_LINE_SIZE=64 + +# Doesn't support NUMA +CONFIG_RTE_EAL_NUMA_AWARE_HUGEPAGES=n +CONFIG_RTE_LIBRTE_VHOST_NUMA=n
mk/machine/n1sdp/rte.vars.mk new file +34 -0
@@ -0,0 +1,34 @@ +# SPDX-License-Identifier: BSD-3-Clause +# Copyright(c) 2019 Arm Ltd +# + +# +# machine: +# +# - can define ARCH variable (overridden by cmdline value) +# - can define CROSS variable (overridden by cmdline value) +# - define MACHINE_CFLAGS variable (overridden by cmdline value) +# - define MACHINE_LDFLAGS variable (overridden by cmdline value) +# - define MACHINE_ASFLAGS variable (overridden by cmdline value) +# - can define CPU_CFLAGS variable (overridden by cmdline value) that +# overrides the one defined in arch. +# - can define CPU_LDFLAGS variable (overridden by cmdline value) that +# overrides the one defined in arch. +# - can define CPU_ASFLAGS variable (overridden by cmdline value) that +# overrides the one defined in arch. +# - may override any previously defined variable +# + +# ARCH = +# CROSS = +# MACHINE_CFLAGS = +# MACHINE_LDFLAGS = +# MACHINE_ASFLAGS = +# CPU_CFLAGS = +# CPU_LDFLAGS = +# CPU_ASFLAGS = + +include $(RTE_SDK)/mk/rte.helper.mk + +MACHINE_CFLAGS += $(call rte_cc_has_argument, -march=armv8.2-a+crc+crypto) +MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=neoverse-n1)
Signature · 1 line
2.7.4
Re: [dpdk-dev] [PATCH v3 2/3] config: add arm neoverse N1 SDP configuration
Quoted 24 lines · Gavin Hu
Arm N1 SDP is an infrastructure segment development platform based on armv8.2-a Neoverse N1 CPU. For more information, refer to:
https://community.arm.com/developer/tools-software/oss-platforms/w/
docs/440/neoverse-n1-sdpSigned-off-by Gavin Hu <[email protected]>Reviewed-by Honnappa Nagarahalli <[email protected]>Reviewed-by Steve Capper <[email protected]>---
V3
-change the configuration name from "neoversen1" to "n1sdp" to be platform specific other than microarchitecture specificconfig/arm/arm64_n1sdp_linux_gcc new file +3
-2-add the missing config/arm/arm64_n1sdp_linux-gcc file for meson build --- +[properties] +implementor_id = '0x41' +implementor_pn = '0xd0c'config/arm/meson.build +0
-0
config/arm/meson.build flags_armada = [ · line 65
Quoted 9 lines
config/arm/meson.build +6
-0@@ -57,6 +57,12 @@ flags_armada = [ ['RTE_MAX_LCORE', 16]] flags_default_extra = [] +flags_n1sdp_extra = [ + ['RTE_MACHINE', '"n1sdp"'], + ['RTE_MAX_NUMA_NODES', 1], + ['RTE_MAX_LCORE', 4], + ['RTE_EAL_NUMA_AWARE_HUGEPAGES', false], + ['RTE_LIBRTE_VHOST_NUMA', false]]
Shouldn't we add ['RTE_CACHE_LINE_SIZE', 64], here?
Quoted 1 line
config/arm/meson.build +0
-0flags_thunderx_extra = [
Re: [dpdk-dev] [PATCH v3 2/3] config: add arm neoverse N1 SDP configuration
Hi Jerin,
Quoted 37 lines
-----Original Message-----
From: Jerin Jacob <[email protected]>
Sent: Tuesday, November 12, 2019 12:24 PM
To: Gavin Hu (Arm Technology China) <[email protected]>
Cc: dpdk-dev <[email protected]>; nd <[email protected]>; [email protected]; Honnappa Nagarahalli <[email protected]>
Subject: Re: [dpdk-dev] [PATCH v3 2/3] config: add arm neoverse N1 SDP configurationQuoted 6 lines · Gavin Hu
Arm N1 SDP is an infrastructure segment development platform based on armv8.2-a Neoverse N1 CPU. For more information, refer to:
https://community.arm.com/developer/tools-software/oss-platforms/w/
docs/440/neoverse-n1-sdpQuoted 6 lines
Signed-off-by Gavin Hu <[email protected]>Reviewed-by Honnappa Nagarahalli <[email protected]>Reviewed-by Steve Capper <[email protected]>---
V3
-change the configuration name from "neoversen1" to "n1sdp" to beplatform
Quoted 14 lines
specific other than microarchitecture specific
config/arm/arm64_n1sdp_linux_gcc new file +3
-2-add the missing config/arm/arm64_n1sdp_linux-gcc file for meson build --- +[properties] +implementor_id = '0x41' +implementor_pn = '0xd0c'config/arm/meson.build +0
-0@@ -57,6 +57,12 @@ flags_armada = [ ['RTE_MAX_LCORE', 16]]
Quoted 9 lines
config/arm/meson.build flags_armada = [ · line 65
Quoted 7 lines
config/arm/meson.build +6
-0flags_default_extra = [] +flags_n1sdp_extra = [ + ['RTE_MACHINE', '"n1sdp"'], + ['RTE_MAX_NUMA_NODES', 1], + ['RTE_MAX_LCORE', 4], + ['RTE_EAL_NUMA_AWARE_HUGEPAGES', false], + ['RTE_LIBRTE_VHOST_NUMA', false]]Shouldn't we add ['
RTE_CACHE_LINE_SIZE', 64], here?
Arm defines it in the upper implemter_id levelflags_arm = [
['RTE_MACHINE', '"armv8a"'], ['RTE_MAX_LCORE', 16], ['RTE_USE_C11_MEM_MODEL', true], ['RTE_CACHE_LINE_SIZE', 64]]
http://code.dpdk.org/dpdk/latest/source/config/arm/meson.build#L40
Quoted 2 lines
Quoted 2 lines
config/arm/meson.build +0
-0flags_thunderx_extra = [
Re: [dpdk-dev] [PATCH v3 2/3] config: add arm neoverse N1 SDP configuration
In reply to Gavin HuQuoted 10 lines · Gavin Hu
Arm N1 SDP is an infrastructure segment development platform based on armv8.2-a Neoverse N1 CPU. For more information, refer to:
https://community.arm.com/developer/tools-software/oss-platforms/w/
docs/440/neoverse-n1-sdpSigned-off-by Gavin Hu <[email protected]>Reviewed-by Honnappa Nagarahalli <[email protected]>Reviewed-by Steve Capper <[email protected]>---
PATCH v3 3/3 config: add cortex-a76 configuration
In reply to Gavin HuTo make the list complete and consistent, add cortex-a76 configuration.
--- config/arm/meson.build | 1 + 1 file changed, 1 insertion(+)
config/arm/meson.build +1 -0
@@ -90,6 +90,7 @@ machine_args_generic = [ ['0xd08', ['-mcpu=cortex-a72']], ['0xd09', ['-mcpu=cortex-a73']], ['0xd0a', ['-mcpu=cortex-a75']], + ['0xd0b', ['-mcpu=cortex-a76']], ['0xd0c', ['-march=armv8.2-a+crc+crypto', '-mcpu=neoverse-n1'], flags_n1sdp_extra]] machine_args_cavium = [
Signature · 1 line
2.7.4
Re: [dpdk-dev] [PATCH v3 3/3] config: add cortex-a76 configuration
Quoted 5 lines · Gavin Hu
To make the list complete and consistent, add cortex-a76 configuration.
Signed-off-by Gavin Hu <[email protected]>Reviewed-by Honnappa Nagarahalli <[email protected]>
Quoted 19 lines
--- config/arm/meson.build | 1 + 1 file changed, 1 insertion(+)config/arm/meson.build +1
-1@@ -90,6 +90,7 @@ machine_args_generic = [ ['0xd08', ['-mcpu=cortex-a72']], ['0xd09', ['-mcpu=cortex-a73']], ['0xd0a', ['-mcpu=cortex-a75']], + ['0xd0b', ['-mcpu=cortex-a76']], ['0xd0c', ['-march=armv8.2-a+crc+crypto', '-mcpu=neoverse-n1'], flags_n1sdp_extra]] machine_args_cavium = [ --2.7.4
Thomas Monjalon
AppliedRe: [dpdk-dev] [PATCH v3 0/3] add arm N1SDP and A76 configurations
In reply to Gavin Hu11/11/2019 06:41, Gavin Hu:
Quoted 4 lines
Gavin Hu (3):
test/rcu: fix the compiling error for armv8.2
config: add arm neoverse N1 SDP configuration
config: add cortex-a76 configuration
Applied, without first patch.
Please re-submit a fix for RCU test.