Skip to contentSkip to conversation

mk: introduce helper to check valid compiler argument

PATCH Superseded Jerin Jacob Kollanukkaran 8 participants 95 messages Jan 6, 2019 – Apr 15, 2019 Arm/Amlogic Meson SoC support

lore
Share & tools
lore
#0 lore raw

PATCH 1/3 mk: introduce helper to check valid compiler argument

Introduce rte_cc_has_argument() Makefile helper to check a given argument is support by the compiler.

Example Usage:

include $(RTE_SDK)/mk/rte.helper.mk
MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=octeontx2)

This would allow adding -mcpu=octeontx2 in MACHINE_CFLAGS if it is only supported by the compiler. The use case for such scheme is to enable the mcpu optimization if the compiler supports else it needs to compile the source code without any errors.

This patch also moves inclusion of toolchain's rte.vars.mk to before the machine's rte.vars.mk inclusion to make correct CC available for the cross compile case.

---
 mk/rte.helper.mk              | 12 ++++++++++++
 mk/target/generic/rte.vars.mk | 22 +++++++++++-----------
 2 files changed, 23 insertions(+), 11 deletions(-)
 create mode 100644 mk/rte.helper.mk
mk/rte.helper.mk new file +12 -0
new file mode 100644
index 000000000..2c5d5275e
@@ -0,0 +1,12 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International Ltd
+
+# rte_cc_has_argument
+# Usage: MACHINE_CFLAGS += $(call rte_cc_has_argument, -mno-avx512f)
+# Return the argument if the argument is supported by the compiler.
+#
+define rte_cc_has_argument
+	$(shell $(CC) -Werror $(1) -c -x c /dev/null -o tmp$$ 2> /dev/null && rm -f tmp$$ && echo $(1) | xargs echo -n)
+endef
+
+
mk/target/generic/rte.vars.mk +11 -11
index dd149acc9..25a578ad7 100644
@@ -7,6 +7,17 @@
 # executive environment.
 #
 
+#
+# toolchain:
+#
+#   - define CC, LD, AR, AS, ...
+#   - define TOOLCHAIN_CFLAGS variable (overridden by cmdline value)
+#   - define TOOLCHAIN_LDFLAGS variable (overridden by cmdline value)
+#   - define TOOLCHAIN_ASFLAGS variable (overridden by cmdline value)
+#   - may override any previously defined variable
+#
+include $(RTE_SDK)/mk/toolchain/$(RTE_TOOLCHAIN)/rte.vars.mk
+
 #
 # machine:
 #
@@ -45,17 +56,6 @@ endif
 #
 include $(RTE_SDK)/mk/arch/$(RTE_ARCH)/rte.vars.mk
 
-#
-# toolchain:
-#
-#   - define CC, LD, AR, AS, ...
-#   - define TOOLCHAIN_CFLAGS variable (overridden by cmdline value)
-#   - define TOOLCHAIN_LDFLAGS variable (overridden by cmdline value)
-#   - define TOOLCHAIN_ASFLAGS variable (overridden by cmdline value)
-#   - may override any previously defined variable
-#
-include $(RTE_SDK)/mk/toolchain/$(RTE_TOOLCHAIN)/rte.vars.mk
-
 #
 # exec-env:
 #
Signature · 1 line
2.20.1
#1 lore raw

PATCH 2/3 config: add thunderx2 machine config

Optimized configuration for Marvell thunderx2 SoC.

Product details are here:

https://www.marvell.com/server-processors/thunderx2-arm-processors/

---
 config/defconfig_arm64-thunderx2-linuxapp-gcc | 11 ++++++
 mk/machine/thunderx2/rte.vars.mk              | 34 +++++++++++++++++++
 2 files changed, 45 insertions(+)
 create mode 100644 config/defconfig_arm64-thunderx2-linuxapp-gcc
 create mode 100644 mk/machine/thunderx2/rte.vars.mk
config/defconfig_arm64-thunderx2-linuxapp-gcc new file +11 -0
new file mode 100644
index 000000000..27db58e50
@@ -0,0 +1,11 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International Ltd
+#
+
+#include "defconfig_arm64-armv8a-linuxapp-gcc"
+
+CONFIG_RTE_MACHINE="thunderx2"
+
+CONFIG_RTE_CACHE_LINE_SIZE=64
+CONFIG_RTE_MAX_NUMA_NODES=2
+CONFIG_RTE_MAX_LCORE=256
mk/machine/thunderx2/rte.vars.mk new file +34 -0
new file mode 100644
index 000000000..b80dc8680
@@ -0,0 +1,34 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International 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=armv8.1-a+crc+crypto)
+MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=thunderx2t99)
Signature · 1 line
2.20.1
#2 lore raw

Re: [PATCH 2/3] config: add thunderx2 machine config

Hi Jerin,

06/01/2019 14:20, Jerin Jacob Kollanukkaran:

Quoted 10 lines

Optimized configuration for Marvell thunderx2 SoC.

Product details are here:

https://www.marvell.com/server-processors/thunderx2-arm-processors/

---
 config/defconfig_arm64-thunderx2-linuxapp-gcc | 11 ++++++
 mk/machine/thunderx2/rte.vars.mk              | 34 +++++++++++++++++++

You should enable meson build at the same time.

#3 lore raw

Re: [EXT] Re: [PATCH 2/3] config: add thunderx2 machine config

Quoted 1 line · Thomas Monjalon

Hi Jerin,

Hi Thomas,

Quoted 7 lines

06/01/2019 14:20, Jerin Jacob Kollanukkaran:

Quoted 6 lines

Optimized configuration for Marvell thunderx2 SoC.

Product details are here:

https://www.marvell.com/server-processors/thunderx2-arm-processors/

Quoted 7 lines
Quoted 5 lines
---
 config/defconfig_arm64-thunderx2-linuxapp-gcc | 11 ++++++
 mk/machine/thunderx2/rte.vars.mk              | 34

+++++++++++++++++++

You should enable meson build at the same time.

OK. Will send the v2.

Quoted 3 lines
#4 lore raw

Re: [PATCH 2/3] config: add thunderx2 machine config

In reply to Jerin Jacob Kollanukkaran
Quoted 30 lines

-----Original Message-----
From: Jerin Jacob Kollanukkaran <[email protected]>
Sent: Sunday, January 6, 2019 9:20 PM
To: [email protected]
Cc: [email protected]; Gavin Hu (Arm Technology China) <[email protected]>; [email protected]
Subject: [dpdk-dev] [PATCH 2/3] config: add thunderx2 machine config

Optimized configuration for Marvell thunderx2 SoC.

Product details are here:

https://www.marvell.com/server-processors/thunderx2-arm-processors/

---
 config/defconfig_arm64-thunderx2-linuxapp-gcc | 11 ++++++
 mk/machine/thunderx2/rte.vars.mk              | 34 +++++++++++++++++++
 2 files changed, 45 insertions(+)
 create mode 100644 config/defconfig_arm64-thunderx2-linuxapp-gcc
 create mode 100644 mk/machine/thunderx2/rte.vars.mk
config/defconfig_arm64-thunderx2-linuxapp-gcc new file +0 -0
new file mode 100644
index 000000000..27db58e50

config/defconfig_arm64-thunderx2-linuxapp-gcc line 9

Quoted 9 lines
config/defconfig_arm64-thunderx2-linuxapp-gcc new file +9 -0
@@ -0,0 +1,11 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International Ltd
+#
+
+#include "defconfig_arm64-armv8a-linuxapp-gcc"
+
+CONFIG_RTE_MACHINE="thunderx2"
+
+CONFIG_RTE_CACHE_LINE_SIZE=64

ThunderX is 128, ThunderX2 downsized the cache line?

Quoted 46 lines
config/defconfig_arm64-thunderx2-linuxapp-gcc new file +2 -0
+CONFIG_RTE_MAX_NUMA_NODES=2
+CONFIG_RTE_MAX_LCORE=256
mk/machine/thunderx2/rte.vars.mk new file +33 -0
new file mode 100644
index 000000000..b80dc8680
@@ -0,0 +1,34 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International 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=armv8.1-

a+crc+crypto)
+MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=thunderx2t99) --
2.20.1

#5 lore raw

Re: [EXT] RE: [PATCH 2/3] config: add thunderx2 machine config

Quoted 35 lines · Gavin Hu (Arm Technology China)
Quoted 35 lines

-----Original Message-----
From: Jerin Jacob Kollanukkaran <[email protected]>
Sent: Sunday, January 6, 2019 9:20 PM
To: [email protected]
Cc: [email protected]; Gavin Hu (Arm Technology China) <[email protected]

Quoted 1 line

;

[email protected]
Subject: [dpdk-dev] [PATCH 2/3] config: add thunderx2 machine config

Optimized configuration for Marvell thunderx2 SoC.

Product details are here:

https://www.marvell.com/server-processors/thunderx2-arm-processors/

---
 config/defconfig_arm64-thunderx2-linuxapp-gcc | 11 ++++++
 mk/machine/thunderx2/rte.vars.mk              | 34
+++++++++++++++++++
 2 files changed, 45 insertions(+)
 create mode 100644 config/defconfig_arm64-thunderx2-linuxapp-gcc
 create mode 100644 mk/machine/thunderx2/rte.vars.mk
config/defconfig_arm64-thunderx2-linuxapp-gcc new file +3 -0
new file mode 100644
index 000000000..27db58e50
@@ -0,0 +1,11 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International Ltd
+#
Quoted 8 lines

config/defconfig_arm64-thunderx2-linuxapp-gcc line 9

Quoted 6 lines
config/defconfig_arm64-thunderx2-linuxapp-gcc new file +6 -0
+
+#include "defconfig_arm64-armv8a-linuxapp-gcc"
+
+CONFIG_RTE_MACHINE="thunderx2"
+
+CONFIG_RTE_CACHE_LINE_SIZE=64

ThunderX is 128, ThunderX2 downsized the cache line?

Yes. Tunned for server work load.

#6 lore raw

PATCH 3/3 config: add octeontx2 machine config

In reply to Jerin Jacob Kollanukkaran

Optimized configuration for Marvell octeontx2 SoC.

---
 config/defconfig_arm64-octeontx2-linuxapp-gcc | 18 ++++++++++
 mk/machine/octeontx2/rte.vars.mk              | 34 +++++++++++++++++++
 2 files changed, 52 insertions(+)
 create mode 100644 config/defconfig_arm64-octeontx2-linuxapp-gcc
 create mode 100644 mk/machine/octeontx2/rte.vars.mk
config/defconfig_arm64-octeontx2-linuxapp-gcc new file +18 -0
new file mode 100644
index 000000000..9a99eada1
@@ -0,0 +1,18 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International Ltd
+#
+
+#include "defconfig_arm64-armv8a-linuxapp-gcc"
+
+CONFIG_RTE_MACHINE="octeontx2"
+
+CONFIG_RTE_CACHE_LINE_SIZE=128
+CONFIG_RTE_MAX_NUMA_NODES=1
+CONFIG_RTE_MAX_LCORE=24
+
+# Doesn't support NUMA
+CONFIG_RTE_EAL_NUMA_AWARE_HUGEPAGES=n
+CONFIG_RTE_LIBRTE_VHOST_NUMA=n
+
+# Recommend to use VFIO as co-processors needs SMMU/IOMMU
+CONFIG_RTE_EAL_IGB_UIO=n
mk/machine/octeontx2/rte.vars.mk new file +34 -0
new file mode 100644
index 000000000..e209cf492
@@ -0,0 +1,34 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International 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+lse)
+MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=octeontx2)
Signature · 1 line
2.20.1
#7 lore raw

PATCH v2 1/6 mk: introduce helper to check valid compiler argument

In reply to Jerin Jacob Kollanukkaran

From: Jerin Jacob Kollanukkaran <[email protected]>

Introduce rte_cc_has_argument() Makefile helper to check a given argument is support by the compiler.

Example Usage:

include $(RTE_SDK)/mk/rte.helper.mk
MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=octeontx2)

This would allow adding -mcpu=octeontx2 in MACHINE_CFLAGS if it is only supported by the compiler. The use case for such scheme is to enable the mcpu optimization if the compiler supports else it needs to compile the source code without any errors.

This patch also moves inclusion of toolchain's rte.vars.mk to before the machine's rte.vars.mk inclusion to make correct CC available for the cross compile case.

---
 mk/rte.helper.mk              | 12 ++++++++++++
 mk/target/generic/rte.vars.mk | 22 +++++++++++-----------
 2 files changed, 23 insertions(+), 11 deletions(-)
 create mode 100644 mk/rte.helper.mk
mk/rte.helper.mk new file +12 -0
new file mode 100644
index 000000000..2c5d5275e
@@ -0,0 +1,12 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International Ltd
+
+# rte_cc_has_argument
+# Usage: MACHINE_CFLAGS += $(call rte_cc_has_argument, -mno-avx512f)
+# Return the argument if the argument is supported by the compiler.
+#
+define rte_cc_has_argument
+	$(shell $(CC) -Werror $(1) -c -x c /dev/null -o tmp$$ 2> /dev/null && rm -f tmp$$ && echo $(1) | xargs echo -n)
+endef
+
+
mk/target/generic/rte.vars.mk +11 -11
index dd149acc9..25a578ad7 100644
@@ -7,6 +7,17 @@
 # executive environment.
 #
 
+#
+# toolchain:
+#
+#   - define CC, LD, AR, AS, ...
+#   - define TOOLCHAIN_CFLAGS variable (overridden by cmdline value)
+#   - define TOOLCHAIN_LDFLAGS variable (overridden by cmdline value)
+#   - define TOOLCHAIN_ASFLAGS variable (overridden by cmdline value)
+#   - may override any previously defined variable
+#
+include $(RTE_SDK)/mk/toolchain/$(RTE_TOOLCHAIN)/rte.vars.mk
+
 #
 # machine:
 #
@@ -45,17 +56,6 @@ endif
 #
 include $(RTE_SDK)/mk/arch/$(RTE_ARCH)/rte.vars.mk
 
-#
-# toolchain:
-#
-#   - define CC, LD, AR, AS, ...
-#   - define TOOLCHAIN_CFLAGS variable (overridden by cmdline value)
-#   - define TOOLCHAIN_LDFLAGS variable (overridden by cmdline value)
-#   - define TOOLCHAIN_ASFLAGS variable (overridden by cmdline value)
-#   - may override any previously defined variable
-#
-include $(RTE_SDK)/mk/toolchain/$(RTE_TOOLCHAIN)/rte.vars.mk
-
 #
 # exec-env:
 #
Signature · 1 line
2.20.1
#8 lore raw

PATCH v2 2/6 config: add thunderx2 machine config

From: Jerin Jacob Kollanukkaran <[email protected]>

Optimized configuration for Marvell thunderx2 SoC.

Product details are here:

https://www.marvell.com/server-processors/thunderx2-arm-processors/

---
 config/defconfig_arm64-thunderx2-linuxapp-gcc | 11 ++++++
 mk/machine/thunderx2/rte.vars.mk              | 34 +++++++++++++++++++
 2 files changed, 45 insertions(+)
 create mode 100644 config/defconfig_arm64-thunderx2-linuxapp-gcc
 create mode 100644 mk/machine/thunderx2/rte.vars.mk
config/defconfig_arm64-thunderx2-linuxapp-gcc new file +11 -0
new file mode 100644
index 000000000..27db58e50
@@ -0,0 +1,11 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International Ltd
+#
+
+#include "defconfig_arm64-armv8a-linuxapp-gcc"
+
+CONFIG_RTE_MACHINE="thunderx2"
+
+CONFIG_RTE_CACHE_LINE_SIZE=64
+CONFIG_RTE_MAX_NUMA_NODES=2
+CONFIG_RTE_MAX_LCORE=256
mk/machine/thunderx2/rte.vars.mk new file +34 -0
new file mode 100644
index 000000000..b80dc8680
@@ -0,0 +1,34 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International 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=armv8.1-a+crc+crypto)
+MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=thunderx2t99)
Signature · 1 line
2.20.1
#9 lore raw

PATCH v2 3/6 config: add octeontx2 machine config

In reply to Pavan Nikhilesh Bhagavatula

From: Jerin Jacob Kollanukkaran <[email protected]>

Optimized configuration for Marvell octeontx2 SoC.

---
 config/defconfig_arm64-octeontx2-linuxapp-gcc | 18 ++++++++++
 mk/machine/octeontx2/rte.vars.mk              | 34 +++++++++++++++++++
 2 files changed, 52 insertions(+)
 create mode 100644 config/defconfig_arm64-octeontx2-linuxapp-gcc
 create mode 100644 mk/machine/octeontx2/rte.vars.mk
config/defconfig_arm64-octeontx2-linuxapp-gcc new file +18 -0
new file mode 100644
index 000000000..9a99eada1
@@ -0,0 +1,18 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International Ltd
+#
+
+#include "defconfig_arm64-armv8a-linuxapp-gcc"
+
+CONFIG_RTE_MACHINE="octeontx2"
+
+CONFIG_RTE_CACHE_LINE_SIZE=128
+CONFIG_RTE_MAX_NUMA_NODES=1
+CONFIG_RTE_MAX_LCORE=24
+
+# Doesn't support NUMA
+CONFIG_RTE_EAL_NUMA_AWARE_HUGEPAGES=n
+CONFIG_RTE_LIBRTE_VHOST_NUMA=n
+
+# Recommend to use VFIO as co-processors needs SMMU/IOMMU
+CONFIG_RTE_EAL_IGB_UIO=n
mk/machine/octeontx2/rte.vars.mk new file +34 -0
new file mode 100644
index 000000000..e209cf492
@@ -0,0 +1,34 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International 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+lse)
+MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=octeontx2)
Signature · 1 line
2.20.1
#10 lore raw

PATCH v2 4/6 meson: add infra to support machine specific flags

In reply to Pavan Nikhilesh Bhagavatula

From: Pavan Nikhilesh <[email protected]>

Currently, RTE_* flags are set based on the implementer ID but there might be some micro arch specific differences from the same vendor eg. CACHE_LINESIZE.
Add support to set micro arch specific flags.

---
 config/arm/meson.build | 53 +++++++++++++++++++++++++-----------------
 1 file changed, 32 insertions(+), 21 deletions(-)
config/arm/meson.build +32 -21
index dae55d6b2..576363fc0 100644
@@ -7,23 +7,6 @@ march_opt = '-march=@0@'.format(machine)
 
 arm_force_native_march = false
 
-machine_args_generic = [
-	['default', ['-march=armv8-a+crc+crypto']],
-	['native', ['-march=native']],
-	['0xd03', ['-mcpu=cortex-a53']],
-	['0xd04', ['-mcpu=cortex-a35']],
-	['0xd07', ['-mcpu=cortex-a57']],
-	['0xd08', ['-mcpu=cortex-a72']],
-	['0xd09', ['-mcpu=cortex-a73']],
-	['0xd0a', ['-mcpu=cortex-a75']],
-]
-machine_args_cavium = [
-	['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
-	['native', ['-march=native']],
-	['0xa1', ['-mcpu=thunderxt88']],
-	['0xa2', ['-mcpu=thunderxt81']],
-	['0xa3', ['-mcpu=thunderxt83']]]
-
 flags_common_default = [
 	# Accelarate rte_memcpy. Be sure to run unit test (memcpy_perf_autotest)
 	# to determine the best threshold in code. Refer to notes in source file
@@ -50,12 +33,10 @@ flags_generic = [
 	['RTE_USE_C11_MEM_MODEL', true],
 	['RTE_CACHE_LINE_SIZE', 128]]
 flags_cavium = [
-	['RTE_MACHINE', '"thunderx"'],
 	['RTE_CACHE_LINE_SIZE', 128],
 	['RTE_MAX_NUMA_NODES', 2],
 	['RTE_MAX_LCORE', 96],
-	['RTE_MAX_VFIO_GROUPS', 128],
-	['RTE_USE_C11_MEM_MODEL', false]]
+	['RTE_MAX_VFIO_GROUPS', 128]]
 flags_dpaa = [
 	['RTE_MACHINE', '"dpaa"'],
 	['RTE_USE_C11_MEM_MODEL', true],
@@ -69,6 +50,27 @@ flags_dpaa2 = [
 	['RTE_MAX_NUMA_NODES', 1],
 	['RTE_MAX_LCORE', 16],
 	['RTE_LIBRTE_DPAA2_USE_PHYS_IOVA', false]]
+flags_default_extra = []
+flags_thunderx_extra = [
+	['RTE_MACHINE', '"thunderx"'],
+	['RTE_USE_C11_MEM_MODEL', false]]
+
+machine_args_generic = [
+	['default', ['-march=armv8-a+crc+crypto']],
+	['native', ['-march=native']],
+	['0xd03', ['-mcpu=cortex-a53']],
+	['0xd04', ['-mcpu=cortex-a35']],
+	['0xd07', ['-mcpu=cortex-a57']],
+	['0xd08', ['-mcpu=cortex-a72']],
+	['0xd09', ['-mcpu=cortex-a73']],
+	['0xd0a', ['-mcpu=cortex-a75']]]
+
+machine_args_cavium = [
+	['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
+	['native', ['-march=native']],
+	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
+	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
+	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra]]
 
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page G7-5321)
 impl_generic = ['Generic armv8', flags_generic, machine_args_generic]
@@ -157,10 +159,19 @@ else
 	foreach marg: machine[2]
 		if marg[0] == impl_pn
 			foreach f: marg[1]
-				machine_args += f
+				if cc.has_argument(f)
+					machine_args += f
+				endif
 			endforeach
 		endif
 	endforeach
+
+	# Apply any extra machine specific flags.
+	foreach flag: marg.get(2, flags_default_extra)
+		if flag.length() > 0
+			dpdk_conf.set(flag[0], flag[1])
+		endif
+	endforeach
 endif
 message(machine_args)
 
Signature · 1 line
2.20.1
#11 lore raw

PATCH v2 5/6 meson: add thunderx2 machine config

In reply to Pavan Nikhilesh Bhagavatula

From: Pavan Nikhilesh <[email protected]>

Meson configuration for Marvell thunderx2 SoC.

---
 config/arm/meson.build | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)
config/arm/meson.build +8 -1
index 576363fc0..aca285b6a 100644
@@ -54,6 +54,12 @@ flags_default_extra = []
 flags_thunderx_extra = [
 	['RTE_MACHINE', '"thunderx"'],
 	['RTE_USE_C11_MEM_MODEL', false]]
+flags_thunderx2_extra = [
+	['RTE_MACHINE', '"thunderx2"'],
+	['RTE_CACHE_LINE_SIZE', 64],
+	['RTE_MAX_NUMA_NODES', 2],
+	['RTE_MAX_LCORE', 256],
+	['RTE_USE_C11_MEM_MODEL', true]]
 
 machine_args_generic = [
 	['default', ['-march=armv8-a+crc+crypto']],
@@ -70,7 +76,8 @@ machine_args_cavium = [
 	['native', ['-march=native']],
 	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
 	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
-	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra]]
+	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra],
+	['0xaf', ['-mcpu=thunderx2t99'], flags_thunderx2_extra]]
 
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page G7-5321)
 impl_generic = ['Generic armv8', flags_generic, machine_args_generic]
Signature · 1 line
2.20.1
#12 lore raw

Re: [PATCH v2 5/6] meson: add thunderx2 machine config

07/01/2019 16:42, Pavan Nikhilesh Bhagavatula:

Quoted 8 lines

From: Pavan Nikhilesh <[email protected]>

Meson configuration for Marvell thunderx2 SoC.

---
 config/arm/meson.build | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

I think it can be merged with the patch doing the same for make.

#13 lore raw

PATCH v2 6/6 meson: add octeontx2 machine config

In reply to Pavan Nikhilesh Bhagavatula

From: Pavan Nikhilesh <[email protected]>

Meson configuration for Marvell octeontx2 SoC.

---
 config/arm/meson.build | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)
config/arm/meson.build +9 -1
index aca285b6a..8086357a1 100644
@@ -60,6 +60,13 @@ flags_thunderx2_extra = [
 	['RTE_MAX_NUMA_NODES', 2],
 	['RTE_MAX_LCORE', 256],
 	['RTE_USE_C11_MEM_MODEL', true]]
+flags_octeontx2_extra = [
+	['RTE_MACHINE', '"octeontx2"'],
+	['RTE_MAX_NUMA_NODES', 1],
+	['RTE_MAX_LCORE', 24],
+	['RTE_EAL_NUMA_AWARE_HUGEPAGES', false],
+	['RTE_LIBRTE_VHOST_NUMA', false],
+	['RTE_EAL_IGB_UIO', false]]
 
 machine_args_generic = [
 	['default', ['-march=armv8-a+crc+crypto']],
@@ -77,7 +84,8 @@ machine_args_cavium = [
 	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
 	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
 	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra],
-	['0xaf', ['-mcpu=thunderx2t99'], flags_thunderx2_extra]]
+	['0xaf', ['-mcpu=thunderx2t99'], flags_thunderx2_extra],
+	['0xb2', ['-mcpu=octeontx2'], flags_octeontx2_extra]]
 
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page G7-5321)
 impl_generic = ['Generic armv8', flags_generic, machine_args_generic]
Signature · 1 line
2.20.1
#14 lore raw

PATCH v3 1/4 mk: introduce helper to check valid compiler argument

In reply to Jerin Jacob Kollanukkaran

From: Jerin Jacob <[email protected]>

Introduce rte_cc_has_argument() Makefile helper to check a given argument is support by the compiler.

Example Usage:

include $(RTE_SDK)/mk/rte.helper.mk
MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=octeontx2)

This would allow adding -mcpu=octeontx2 in MACHINE_CFLAGS if it is only supported by the compiler. The use case for such scheme is to enable the mcpu optimization if the compiler supports else it needs to compile the source code without any errors.

This patch also moves inclusion of toolchain's rte.vars.mk to before the machine's rte.vars.mk inclusion to make correct CC available for the cross compile case.

---
 mk/rte.helper.mk              | 12 ++++++++++++
 mk/target/generic/rte.vars.mk | 22 +++++++++++-----------
 2 files changed, 23 insertions(+), 11 deletions(-)
 create mode 100644 mk/rte.helper.mk
mk/rte.helper.mk new file +12 -0
new file mode 100644
index 000000000..2c5d5275e
@@ -0,0 +1,12 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International Ltd
+
+# rte_cc_has_argument
+# Usage: MACHINE_CFLAGS += $(call rte_cc_has_argument, -mno-avx512f)
+# Return the argument if the argument is supported by the compiler.
+#
+define rte_cc_has_argument
+	$(shell $(CC) -Werror $(1) -c -x c /dev/null -o tmp$$ 2> /dev/null && rm -f tmp$$ && echo $(1) | xargs echo -n)
+endef
+
+
mk/target/generic/rte.vars.mk +11 -11
index dd149acc9..25a578ad7 100644
@@ -7,6 +7,17 @@
 # executive environment.
 #
 
+#
+# toolchain:
+#
+#   - define CC, LD, AR, AS, ...
+#   - define TOOLCHAIN_CFLAGS variable (overridden by cmdline value)
+#   - define TOOLCHAIN_LDFLAGS variable (overridden by cmdline value)
+#   - define TOOLCHAIN_ASFLAGS variable (overridden by cmdline value)
+#   - may override any previously defined variable
+#
+include $(RTE_SDK)/mk/toolchain/$(RTE_TOOLCHAIN)/rte.vars.mk
+
 #
 # machine:
 #
@@ -45,17 +56,6 @@ endif
 #
 include $(RTE_SDK)/mk/arch/$(RTE_ARCH)/rte.vars.mk
 
-#
-# toolchain:
-#
-#   - define CC, LD, AR, AS, ...
-#   - define TOOLCHAIN_CFLAGS variable (overridden by cmdline value)
-#   - define TOOLCHAIN_LDFLAGS variable (overridden by cmdline value)
-#   - define TOOLCHAIN_ASFLAGS variable (overridden by cmdline value)
-#   - may override any previously defined variable
-#
-include $(RTE_SDK)/mk/toolchain/$(RTE_TOOLCHAIN)/rte.vars.mk
-
 #
 # exec-env:
 #
Signature · 1 line
2.20.1
#15 lore raw

PATCH v3 2/4 meson: add infra to support machine specific flags

From: Pavan Nikhilesh <[email protected]>

Currently, RTE_* flags are set based on the implementer ID but there might be some micro arch specific differences from the same vendor eg. CACHE_LINESIZE.
Add support to set micro arch specific flags.

---
v2 Changes:
- Add meson build support.

 config/arm/meson.build | 53 +++++++++++++++++++++++++-----------------
 1 file changed, 32 insertions(+), 21 deletions(-)
config/arm/meson.build +32 -21
index dae55d6b2..576363fc0 100644
@@ -7,23 +7,6 @@ march_opt = '-march=@0@'.format(machine)
 
 arm_force_native_march = false
 
-machine_args_generic = [
-	['default', ['-march=armv8-a+crc+crypto']],
-	['native', ['-march=native']],
-	['0xd03', ['-mcpu=cortex-a53']],
-	['0xd04', ['-mcpu=cortex-a35']],
-	['0xd07', ['-mcpu=cortex-a57']],
-	['0xd08', ['-mcpu=cortex-a72']],
-	['0xd09', ['-mcpu=cortex-a73']],
-	['0xd0a', ['-mcpu=cortex-a75']],
-]
-machine_args_cavium = [
-	['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
-	['native', ['-march=native']],
-	['0xa1', ['-mcpu=thunderxt88']],
-	['0xa2', ['-mcpu=thunderxt81']],
-	['0xa3', ['-mcpu=thunderxt83']]]
-
 flags_common_default = [
 	# Accelarate rte_memcpy. Be sure to run unit test (memcpy_perf_autotest)
 	# to determine the best threshold in code. Refer to notes in source file
@@ -50,12 +33,10 @@ flags_generic = [
 	['RTE_USE_C11_MEM_MODEL', true],
 	['RTE_CACHE_LINE_SIZE', 128]]
 flags_cavium = [
-	['RTE_MACHINE', '"thunderx"'],
 	['RTE_CACHE_LINE_SIZE', 128],
 	['RTE_MAX_NUMA_NODES', 2],
 	['RTE_MAX_LCORE', 96],
-	['RTE_MAX_VFIO_GROUPS', 128],
-	['RTE_USE_C11_MEM_MODEL', false]]
+	['RTE_MAX_VFIO_GROUPS', 128]]
 flags_dpaa = [
 	['RTE_MACHINE', '"dpaa"'],
 	['RTE_USE_C11_MEM_MODEL', true],
@@ -69,6 +50,27 @@ flags_dpaa2 = [
 	['RTE_MAX_NUMA_NODES', 1],
 	['RTE_MAX_LCORE', 16],
 	['RTE_LIBRTE_DPAA2_USE_PHYS_IOVA', false]]
+flags_default_extra = []
+flags_thunderx_extra = [
+	['RTE_MACHINE', '"thunderx"'],
+	['RTE_USE_C11_MEM_MODEL', false]]
+
+machine_args_generic = [
+	['default', ['-march=armv8-a+crc+crypto']],
+	['native', ['-march=native']],
+	['0xd03', ['-mcpu=cortex-a53']],
+	['0xd04', ['-mcpu=cortex-a35']],
+	['0xd07', ['-mcpu=cortex-a57']],
+	['0xd08', ['-mcpu=cortex-a72']],
+	['0xd09', ['-mcpu=cortex-a73']],
+	['0xd0a', ['-mcpu=cortex-a75']]]
+
+machine_args_cavium = [
+	['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
+	['native', ['-march=native']],
+	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
+	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
+	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra]]
 
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page G7-5321)
 impl_generic = ['Generic armv8', flags_generic, machine_args_generic]
@@ -157,10 +159,19 @@ else
 	foreach marg: machine[2]
 		if marg[0] == impl_pn
 			foreach f: marg[1]
-				machine_args += f
+				if cc.has_argument(f)
+					machine_args += f
+				endif
 			endforeach
 		endif
 	endforeach
+
+	# Apply any extra machine specific flags.
+	foreach flag: marg.get(2, flags_default_extra)
+		if flag.length() > 0
+			dpdk_conf.set(flag[0], flag[1])
+		endif
+	endforeach
 endif
 message(machine_args)
 

--
2.20.1

#16 lore raw

PATCH v3 3/4 config: add thunderx2 machine config

In reply to Pavan Nikhilesh Bhagavatula

From: Jerin Jacob <[email protected]>

Optimized configuration for Marvell thunderx2 SoC. Update meson build to support Marvell thunderx2 SoC.

Product details are here:

https://www.marvell.com/server-processors/thunderx2-arm-processors/

---
v3 Changes:
- Squash meson build support into config support for thunderx2/octeontx2.

 config/arm/meson.build                        |  9 ++++-
 config/defconfig_arm64-thunderx2-linuxapp-gcc | 11 ++++++
 mk/machine/thunderx2/rte.vars.mk              | 34 +++++++++++++++++++
 3 files changed, 53 insertions(+), 1 deletion(-)
 create mode 100644 config/defconfig_arm64-thunderx2-linuxapp-gcc
 create mode 100644 mk/machine/thunderx2/rte.vars.mk
config/arm/meson.build +8 -1
index 576363fc0..aca285b6a 100644
@@ -54,6 +54,12 @@ flags_default_extra = []
 flags_thunderx_extra = [
 	['RTE_MACHINE', '"thunderx"'],
 	['RTE_USE_C11_MEM_MODEL', false]]
+flags_thunderx2_extra = [
+	['RTE_MACHINE', '"thunderx2"'],
+	['RTE_CACHE_LINE_SIZE', 64],
+	['RTE_MAX_NUMA_NODES', 2],
+	['RTE_MAX_LCORE', 256],
+	['RTE_USE_C11_MEM_MODEL', true]]
 
 machine_args_generic = [
 	['default', ['-march=armv8-a+crc+crypto']],
@@ -70,7 +76,8 @@ machine_args_cavium = [
 	['native', ['-march=native']],
 	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
 	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
-	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra]]
+	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra],
+	['0xaf', ['-mcpu=thunderx2t99'], flags_thunderx2_extra]]
 
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page G7-5321)
 impl_generic = ['Generic armv8', flags_generic, machine_args_generic]
config/defconfig_arm64-thunderx2-linuxapp-gcc new file +11 -0
new file mode 100644
index 000000000..27db58e50
@@ -0,0 +1,11 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International Ltd
+#
+
+#include "defconfig_arm64-armv8a-linuxapp-gcc"
+
+CONFIG_RTE_MACHINE="thunderx2"
+
+CONFIG_RTE_CACHE_LINE_SIZE=64
+CONFIG_RTE_MAX_NUMA_NODES=2
+CONFIG_RTE_MAX_LCORE=256
mk/machine/thunderx2/rte.vars.mk new file +34 -0
new file mode 100644
index 000000000..b80dc8680
@@ -0,0 +1,34 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International 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=armv8.1-a+crc+crypto)
+MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=thunderx2t99)

--
2.20.1

#17 lore raw

PATCH v3 4/4 config: add octeontx2 machine config

In reply to Pavan Nikhilesh Bhagavatula

From: Jerin Jacob <[email protected]>

Optimized configuration for Marvell octeontx2 SoC. Update meson build to support Marvell octeontx2 SoC.

---
 config/arm/meson.build                        | 10 +++++-
 config/defconfig_arm64-octeontx2-linuxapp-gcc | 18 ++++++++++
 mk/machine/octeontx2/rte.vars.mk              | 34 +++++++++++++++++++
 3 files changed, 61 insertions(+), 1 deletion(-)
 create mode 100644 config/defconfig_arm64-octeontx2-linuxapp-gcc
 create mode 100644 mk/machine/octeontx2/rte.vars.mk
config/arm/meson.build +9 -1
index aca285b6a..8086357a1 100644
@@ -60,6 +60,13 @@ flags_thunderx2_extra = [
 	['RTE_MAX_NUMA_NODES', 2],
 	['RTE_MAX_LCORE', 256],
 	['RTE_USE_C11_MEM_MODEL', true]]
+flags_octeontx2_extra = [
+	['RTE_MACHINE', '"octeontx2"'],
+	['RTE_MAX_NUMA_NODES', 1],
+	['RTE_MAX_LCORE', 24],
+	['RTE_EAL_NUMA_AWARE_HUGEPAGES', false],
+	['RTE_LIBRTE_VHOST_NUMA', false],
+	['RTE_EAL_IGB_UIO', false]]
 
 machine_args_generic = [
 	['default', ['-march=armv8-a+crc+crypto']],
@@ -77,7 +84,8 @@ machine_args_cavium = [
 	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
 	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
 	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra],
-	['0xaf', ['-mcpu=thunderx2t99'], flags_thunderx2_extra]]
+	['0xaf', ['-mcpu=thunderx2t99'], flags_thunderx2_extra],
+	['0xb2', ['-mcpu=octeontx2'], flags_octeontx2_extra]]
 
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page G7-5321)
 impl_generic = ['Generic armv8', flags_generic, machine_args_generic]
config/defconfig_arm64-octeontx2-linuxapp-gcc new file +18 -0
new file mode 100644
index 000000000..9a99eada1
@@ -0,0 +1,18 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International Ltd
+#
+
+#include "defconfig_arm64-armv8a-linuxapp-gcc"
+
+CONFIG_RTE_MACHINE="octeontx2"
+
+CONFIG_RTE_CACHE_LINE_SIZE=128
+CONFIG_RTE_MAX_NUMA_NODES=1
+CONFIG_RTE_MAX_LCORE=24
+
+# Doesn't support NUMA
+CONFIG_RTE_EAL_NUMA_AWARE_HUGEPAGES=n
+CONFIG_RTE_LIBRTE_VHOST_NUMA=n
+
+# Recommend to use VFIO as co-processors needs SMMU/IOMMU
+CONFIG_RTE_EAL_IGB_UIO=n
mk/machine/octeontx2/rte.vars.mk new file +34 -0
new file mode 100644
index 000000000..e209cf492
@@ -0,0 +1,34 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International 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+lse)
+MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=octeontx2)
Signature · 1 line
2.20.1
#18 lore raw

PATCH v3 5/5 meson: add cross build targets for thunderx2 and octeontx2

In reply to Pavan Nikhilesh Bhagavatula

From: Pavan Nikhilesh <[email protected]>

Add Thunderx2 and Octeontx2 SoC cross build targets.
meson build --cross-file config/arm/arm64_<cpu>_<platform>_<compiler>

---
 config/arm/arm64_octeontx2_linuxapp_gcc | 14 ++++++++++++++
 config/arm/arm64_thunderx2_linuxapp_gcc | 14 ++++++++++++++
 2 files changed, 28 insertions(+)
 create mode 100644 config/arm/arm64_octeontx2_linuxapp_gcc
 create mode 100644 config/arm/arm64_thunderx2_linuxapp_gcc
config/arm/arm64_octeontx2_linuxapp_gcc new file +14 -0
new file mode 100644
index 000000000..83e28e817
@@ -0,0 +1,14 @@
+[binaries]
+c = 'aarch64-linux-gnu-gcc'
+cpp = 'aarch64-linux-gnu-cpp'
+ar = 'aarch64-linux-gnu-gcc-ar'
+strip = 'aarch64-linux-gnu-strip'
+
+[host_machine]
+system = 'linux'
+cpu_family = 'aarch64'
+cpu = 'armv8-a'
+endian = 'little'
+
+[properties]
+implementor_id = '0xb2'
config/arm/arm64_thunderx2_linuxapp_gcc new file +14 -0
new file mode 100644
index 000000000..de1146639
@@ -0,0 +1,14 @@
+[binaries]
+c = 'aarch64-linux-gnu-gcc'
+cpp = 'aarch64-linux-gnu-cpp'
+ar = 'aarch64-linux-gnu-gcc-ar'
+strip = 'aarch64-linux-gnu-strip'
+
+[host_machine]
+system = 'linux'
+cpu_family = 'aarch64'
+cpu = 'armv8-a'
+endian = 'little'
+
+[properties]
+implementor_id = '0xaf'

--
2.20.1

#19 lore raw

PATCH v4 1/5 mk: introduce helper to check valid compiler argument

In reply to Jerin Jacob Kollanukkaran

From: Jerin Jacob <[email protected]>

Introduce rte_cc_has_argument() Makefile helper to check a given argument is support by the compiler.

Example Usage:

include $(RTE_SDK)/mk/rte.helper.mk
MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=octeontx2)

This would allow adding -mcpu=octeontx2 in MACHINE_CFLAGS if it is only supported by the compiler. The use case for such scheme is to enable the mcpu optimization if the compiler supports else it needs to compile the source code without any errors.

This patch also moves inclusion of toolchain's rte.vars.mk to before the machine's rte.vars.mk inclusion to make correct CC available for the cross compile case.

---

v2 Changes:
- Add meson build support.

v3 Changes:
- Squash meson build support into config support for thunderx2/octeontx2.

v4 Changes:
- Fix incorrect signoff marrvell -> marvell.

 mk/rte.helper.mk              | 12 ++++++++++++
 mk/target/generic/rte.vars.mk | 22 +++++++++++-----------
 2 files changed, 23 insertions(+), 11 deletions(-)
 create mode 100644 mk/rte.helper.mk
mk/rte.helper.mk new file +12 -0
new file mode 100644
index 000000000..2c5d5275e
@@ -0,0 +1,12 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International Ltd
+
+# rte_cc_has_argument
+# Usage: MACHINE_CFLAGS += $(call rte_cc_has_argument, -mno-avx512f)
+# Return the argument if the argument is supported by the compiler.
+#
+define rte_cc_has_argument
+	$(shell $(CC) -Werror $(1) -c -x c /dev/null -o tmp$$ 2> /dev/null && rm -f tmp$$ && echo $(1) | xargs echo -n)
+endef
+
+
mk/target/generic/rte.vars.mk +11 -11
index dd149acc9..25a578ad7 100644
@@ -7,6 +7,17 @@
 # executive environment.
 #
 
+#
+# toolchain:
+#
+#   - define CC, LD, AR, AS, ...
+#   - define TOOLCHAIN_CFLAGS variable (overridden by cmdline value)
+#   - define TOOLCHAIN_LDFLAGS variable (overridden by cmdline value)
+#   - define TOOLCHAIN_ASFLAGS variable (overridden by cmdline value)
+#   - may override any previously defined variable
+#
+include $(RTE_SDK)/mk/toolchain/$(RTE_TOOLCHAIN)/rte.vars.mk
+
 #
 # machine:
 #
@@ -45,17 +56,6 @@ endif
 #
 include $(RTE_SDK)/mk/arch/$(RTE_ARCH)/rte.vars.mk
 
-#
-# toolchain:
-#
-#   - define CC, LD, AR, AS, ...
-#   - define TOOLCHAIN_CFLAGS variable (overridden by cmdline value)
-#   - define TOOLCHAIN_LDFLAGS variable (overridden by cmdline value)
-#   - define TOOLCHAIN_ASFLAGS variable (overridden by cmdline value)
-#   - may override any previously defined variable
-#
-include $(RTE_SDK)/mk/toolchain/$(RTE_TOOLCHAIN)/rte.vars.mk
-
 #
 # exec-env:
 #

--
2.20.1

#20 lore raw

PATCH v4 2/5 meson: add infra to support machine specific flags

From: Pavan Nikhilesh <[email protected]>

Currently, RTE_* flags are set based on the implementer ID but there might be some micro arch specific differences from the same vendor eg. CACHE_LINESIZE.
Add support to set micro arch specific flags.

---
 config/arm/meson.build | 53 +++++++++++++++++++++++++-----------------
 1 file changed, 32 insertions(+), 21 deletions(-)
config/arm/meson.build +32 -21
index dae55d6b2..576363fc0 100644
@@ -7,23 +7,6 @@ march_opt = '-march=@0@'.format(machine)
 
 arm_force_native_march = false
 
-machine_args_generic = [
-	['default', ['-march=armv8-a+crc+crypto']],
-	['native', ['-march=native']],
-	['0xd03', ['-mcpu=cortex-a53']],
-	['0xd04', ['-mcpu=cortex-a35']],
-	['0xd07', ['-mcpu=cortex-a57']],
-	['0xd08', ['-mcpu=cortex-a72']],
-	['0xd09', ['-mcpu=cortex-a73']],
-	['0xd0a', ['-mcpu=cortex-a75']],
-]
-machine_args_cavium = [
-	['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
-	['native', ['-march=native']],
-	['0xa1', ['-mcpu=thunderxt88']],
-	['0xa2', ['-mcpu=thunderxt81']],
-	['0xa3', ['-mcpu=thunderxt83']]]
-
 flags_common_default = [
 	# Accelarate rte_memcpy. Be sure to run unit test (memcpy_perf_autotest)
 	# to determine the best threshold in code. Refer to notes in source file
@@ -50,12 +33,10 @@ flags_generic = [
 	['RTE_USE_C11_MEM_MODEL', true],
 	['RTE_CACHE_LINE_SIZE', 128]]
 flags_cavium = [
-	['RTE_MACHINE', '"thunderx"'],
 	['RTE_CACHE_LINE_SIZE', 128],
 	['RTE_MAX_NUMA_NODES', 2],
 	['RTE_MAX_LCORE', 96],
-	['RTE_MAX_VFIO_GROUPS', 128],
-	['RTE_USE_C11_MEM_MODEL', false]]
+	['RTE_MAX_VFIO_GROUPS', 128]]
 flags_dpaa = [
 	['RTE_MACHINE', '"dpaa"'],
 	['RTE_USE_C11_MEM_MODEL', true],
@@ -69,6 +50,27 @@ flags_dpaa2 = [
 	['RTE_MAX_NUMA_NODES', 1],
 	['RTE_MAX_LCORE', 16],
 	['RTE_LIBRTE_DPAA2_USE_PHYS_IOVA', false]]
+flags_default_extra = []
+flags_thunderx_extra = [
+	['RTE_MACHINE', '"thunderx"'],
+	['RTE_USE_C11_MEM_MODEL', false]]
+
+machine_args_generic = [
+	['default', ['-march=armv8-a+crc+crypto']],
+	['native', ['-march=native']],
+	['0xd03', ['-mcpu=cortex-a53']],
+	['0xd04', ['-mcpu=cortex-a35']],
+	['0xd07', ['-mcpu=cortex-a57']],
+	['0xd08', ['-mcpu=cortex-a72']],
+	['0xd09', ['-mcpu=cortex-a73']],
+	['0xd0a', ['-mcpu=cortex-a75']]]
+
+machine_args_cavium = [
+	['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
+	['native', ['-march=native']],
+	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
+	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
+	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra]]
 
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page G7-5321)
 impl_generic = ['Generic armv8', flags_generic, machine_args_generic]
@@ -157,10 +159,19 @@ else
 	foreach marg: machine[2]
 		if marg[0] == impl_pn
 			foreach f: marg[1]
-				machine_args += f
+				if cc.has_argument(f)
+					machine_args += f
+				endif
 			endforeach
 		endif
 	endforeach
+
+	# Apply any extra machine specific flags.
+	foreach flag: marg.get(2, flags_default_extra)
+		if flag.length() > 0
+			dpdk_conf.set(flag[0], flag[1])
+		endif
+	endforeach
 endif
 message(machine_args)
 
Signature · 1 line
2.20.1
#21 lore raw

Re: [PATCH v4 2/5] meson: add infra to support machine specific flags

09/01/2019 11:39, Pavan Nikhilesh Bhagavatula:

Quoted 18 lines

From: Pavan Nikhilesh <[email protected]>

Currently, RTE_* flags are set based on the implementer ID but there might be some micro arch specific differences from the same vendor eg. CACHE_LINESIZE.
Add support to set micro arch specific flags.

---
flags_cavium = [

config/arm/meson.build +1 -3
-	['RTE_MACHINE', '"thunderx"'],
 	['RTE_CACHE_LINE_SIZE', 128],
 	['RTE_MAX_NUMA_NODES', 2],
 	['RTE_MAX_LCORE', 96],
-	['RTE_MAX_VFIO_GROUPS', 128],
-	['RTE_USE_C11_MEM_MODEL', false]]
+	['RTE_MAX_VFIO_GROUPS', 128]]

[...]

Quoted 3 lines
config/arm/meson.build +3 -0
+flags_thunderx_extra = [
+	['RTE_MACHINE', '"thunderx"'],
+	['RTE_USE_C11_MEM_MODEL', false]]

[..]

config/arm/meson.build flags_dpaa2 = [ · line 73

Quoted 6 lines
config/arm/meson.build +6 -0
+machine_args_cavium = [
+	['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
+	['native', ['-march=native']],
+	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
+	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
+	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra]]

This patch is not only adding infra, it also change some Cavium config. What about splitting, and explaining the change of RTE_USE_C11_MEM_MODEL?

#22 lore raw

Re: [PATCH v4 2/5] meson: add infra to support machine specific flags

In reply to Pavan Nikhilesh Bhagavatula

Hi Pavan,

Quoted 100 lines

-----Original Message-----
From: dev <[email protected]> On Behalf Of Pavan Nikhilesh Bhagavatula
Sent: Wednesday, January 9, 2019 6:40 PM
To: [email protected]; Gavin Hu (Arm Technology China) <[email protected]>; [email protected]; [email protected]
Cc: [email protected]; Pavan Nikhilesh Bhagavatula <[email protected]>
Subject: [dpdk-dev] [PATCH v4 2/5] meson: add infra to support machine specific flags

From: Pavan Nikhilesh <[email protected]>

Currently, RTE_* flags are set based on the implementer ID but there might be some micro arch specific differences from the same vendor eg. CACHE_LINESIZE. Add support to set micro arch specific flags.

---
 config/arm/meson.build | 53 +++++++++++++++++++++++++-----------------
 1 file changed, 32 insertions(+), 21 deletions(-)
config/arm/meson.build index renamed from config/arm/meson.build +0 -0

dae55d6b2..576363fc0 100644

config/arm/meson.build +0 -17
@@ -7,23 +7,6 @@ march_opt = '-march=@0@'.format(machine)
 
 arm_force_native_march = false
 
-machine_args_generic = [
-	['default', ['-march=armv8-a+crc+crypto']],
-	['native', ['-march=native']],
-	['0xd03', ['-mcpu=cortex-a53']],
-	['0xd04', ['-mcpu=cortex-a35']],
-	['0xd07', ['-mcpu=cortex-a57']],
-	['0xd08', ['-mcpu=cortex-a72']],
-	['0xd09', ['-mcpu=cortex-a73']],
-	['0xd0a', ['-mcpu=cortex-a75']],
-]
-machine_args_cavium = [
-	['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
-	['native', ['-march=native']],
-	['0xa1', ['-mcpu=thunderxt88']],
-	['0xa2', ['-mcpu=thunderxt81']],
-	['0xa3', ['-mcpu=thunderxt83']]]
-
 flags_common_default = [
 	# Accelarate rte_memcpy. Be sure to run unit test
 (memcpy_perf_autotest)
config/arm/meson.build +22 -3
 	# to determine the best threshold in code. Refer to notes in source file
@@ -50,12 +33,10 @@ flags_generic = [
 	['RTE_USE_C11_MEM_MODEL', true],
 	['RTE_CACHE_LINE_SIZE', 128]]
 flags_cavium = [
-	['RTE_MACHINE', '"thunderx"'],
 	['RTE_CACHE_LINE_SIZE', 128],
 	['RTE_MAX_NUMA_NODES', 2],
 	['RTE_MAX_LCORE', 96],
-	['RTE_MAX_VFIO_GROUPS', 128],
-	['RTE_USE_C11_MEM_MODEL', false]]
+	['RTE_MAX_VFIO_GROUPS', 128]]
 flags_dpaa = [
 	['RTE_MACHINE', '"dpaa"'],
 	['RTE_USE_C11_MEM_MODEL', true],
@@ -69,6 +50,27 @@ flags_dpaa2 = [
 	['RTE_MAX_NUMA_NODES', 1],
 	['RTE_MAX_LCORE', 16],
 	['RTE_LIBRTE_DPAA2_USE_PHYS_IOVA', false]]
+flags_default_extra = []
+flags_thunderx_extra = [
+	['RTE_MACHINE', '"thunderx"'],
+	['RTE_USE_C11_MEM_MODEL', false]]
+
+machine_args_generic = [
+	['default', ['-march=armv8-a+crc+crypto']],
+	['native', ['-march=native']],
+	['0xd03', ['-mcpu=cortex-a53']],
+	['0xd04', ['-mcpu=cortex-a35']],
+	['0xd07', ['-mcpu=cortex-a57']],
+	['0xd08', ['-mcpu=cortex-a72']],
+	['0xd09', ['-mcpu=cortex-a73']],
+	['0xd0a', ['-mcpu=cortex-a75']]]
+
+machine_args_cavium = [
+	['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
+	['native', ['-march=native']],
+	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
+	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
+	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra]]
 
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page G7-5321)

impl_generic = ['Generic armv8', flags_generic, machine_args_generic] @@ - 157,10 +159,19 @@ else

config/arm/meson.build +3 -1
 	foreach marg: machine[2]
 		if marg[0] == impl_pn
 			foreach f: marg[1]
-				machine_args += f
+				if cc.has_argument(f)
+					machine_args += f
+				endif

config/arm/meson.build else · line 174

Quoted 10 lines
config/arm/meson.build +7 -0
 			endforeach
 		endif
 	endforeach
+
+	# Apply any extra machine specific flags.
+	foreach flag: marg.get(2, flags_default_extra)
+		if flag.length() > 0
+			dpdk_conf.set(flag[0], flag[1])
+		endif
+	endforeach

I think this loop should put inside the 'if marg[0] == impl_pn' condition. The right logic should be:
If marg[0] == impl_pn

	# update machine_args
	# Apply any extra machine specific flags
endif

I tested this patch on thunderx2, but it set the octeontx2 extra flags into the rte_build_config.h. Because octeontx2 is the last item of 'machine_args_cavium' table.

Quoted 5 lines

endif
message(machine_args)

--
2.20.1

Thanks,
Phil Yang.

#23 lore raw

Re: [EXT] RE: [PATCH v4 2/5] meson: add infra to support machine specific flags

Hi Phil,

Quoted 130 lines · Phil Yang (Arm Technology China)

External Email

-------------------------------------------------------------------
---
Hi Pavan,

Quoted 109 lines

-----Original Message-----
From: dev <[email protected]> On Behalf Of Pavan Nikhilesh Bhagavatula
Sent: Wednesday, January 9, 2019 6:40 PM
To: [email protected]; Gavin Hu (Arm Technology China) <[email protected]>; [email protected]; [email protected]
Cc: [email protected]; Pavan Nikhilesh Bhagavatula < [email protected]>
Subject: [dpdk-dev] [PATCH v4 2/5] meson: add infra to support machine specific
flags

From: Pavan Nikhilesh <[email protected]>

Currently, RTE_* flags are set based on the implementer ID but there might be
some micro arch specific differences from the same vendor eg. CACHE_LINESIZE.
Add support to set micro arch specific flags.

---
 config/arm/meson.build | 53 +++++++++++++++++++++++++-------------
----
 1 file changed, 32 insertions(+), 21 deletions(-)
config/arm/meson.build index renamed from config/arm/meson.build +0 -0

dae55d6b2..576363fc0 100644

config/arm/meson.build +0 -17
@@ -7,23 +7,6 @@ march_opt = '-march=@0@'.format(machine)
 
 arm_force_native_march = false
 
-machine_args_generic = [
-	['default', ['-march=armv8-a+crc+crypto']],
-	['native', ['-march=native']],
-	['0xd03', ['-mcpu=cortex-a53']],
-	['0xd04', ['-mcpu=cortex-a35']],
-	['0xd07', ['-mcpu=cortex-a57']],
-	['0xd08', ['-mcpu=cortex-a72']],
-	['0xd09', ['-mcpu=cortex-a73']],
-	['0xd0a', ['-mcpu=cortex-a75']],
-]
-machine_args_cavium = [
-	['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
-	['native', ['-march=native']],
-	['0xa1', ['-mcpu=thunderxt88']],
-	['0xa2', ['-mcpu=thunderxt81']],
-	['0xa3', ['-mcpu=thunderxt83']]]
-
 flags_common_default = [
 	# Accelarate rte_memcpy. Be sure to run unit test
 (memcpy_perf_autotest)
 	# to determine the best threshold in code. Refer to notes in
source file
config/arm/meson.build +22 -3
@@ -50,12 +33,10 @@ flags_generic = [
 	['RTE_USE_C11_MEM_MODEL', true],
 	['RTE_CACHE_LINE_SIZE', 128]]
 flags_cavium = [
-	['RTE_MACHINE', '"thunderx"'],
 	['RTE_CACHE_LINE_SIZE', 128],
 	['RTE_MAX_NUMA_NODES', 2],
 	['RTE_MAX_LCORE', 96],
-	['RTE_MAX_VFIO_GROUPS', 128],
-	['RTE_USE_C11_MEM_MODEL', false]]
+	['RTE_MAX_VFIO_GROUPS', 128]]
 flags_dpaa = [
 	['RTE_MACHINE', '"dpaa"'],
 	['RTE_USE_C11_MEM_MODEL', true],
@@ -69,6 +50,27 @@ flags_dpaa2 = [
 	['RTE_MAX_NUMA_NODES', 1],
 	['RTE_MAX_LCORE', 16],
 	['RTE_LIBRTE_DPAA2_USE_PHYS_IOVA', false]]
+flags_default_extra = []
+flags_thunderx_extra = [
+	['RTE_MACHINE', '"thunderx"'],
+	['RTE_USE_C11_MEM_MODEL', false]]
+
+machine_args_generic = [
+	['default', ['-march=armv8-a+crc+crypto']],
+	['native', ['-march=native']],
+	['0xd03', ['-mcpu=cortex-a53']],
+	['0xd04', ['-mcpu=cortex-a35']],
+	['0xd07', ['-mcpu=cortex-a57']],
+	['0xd08', ['-mcpu=cortex-a72']],
+	['0xd09', ['-mcpu=cortex-a73']],
+	['0xd0a', ['-mcpu=cortex-a75']]]
+
+machine_args_cavium = [
+	['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
+	['native', ['-march=native']],
+	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
+	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
+	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra]]
 
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page G7-
 5321)

impl_generic = ['Generic armv8', flags_generic,
machine_args_generic] @@ -
157,10 +159,19 @@ else

config/arm/meson.build +3 -1
 	foreach marg: machine[2]
 		if marg[0] == impl_pn
 			foreach f: marg[1]
-				machine_args += f
+				if cc.has_argument(f)
+					machine_args += f
+				endif

config/arm/meson.build else · line 174

Quoted 10 lines
config/arm/meson.build +7 -0
 			endforeach
 		endif
 	endforeach
+
+	# Apply any extra machine specific flags.
+	foreach flag: marg.get(2, flags_default_extra)
+		if flag.length() > 0
+			dpdk_conf.set(flag[0], flag[1])
+		endif
+	endforeach

I think this loop should put inside the 'if marg[0] == impl_pn' condition.
The right logic should be:
If marg[0] == impl_pn

Quoted 3 lines
	# update machine_args
	# Apply any extra machine specific flags
endif

Yes, you are correct I will modify this in the next version.

Thanks,
Pavan.

Quoted 13 lines

I tested this patch on thunderx2, but it set the octeontx2 extra flags into the rte_build_config.h. Because octeontx2 is the last item of 'machine_args_cavium' table.

Quoted 5 lines

endif
message(machine_args)

--
2.20.1

Thanks,
Phil Yang.

#24 lore raw

PATCH v4 3/5 config: add thunderx2 machine config

In reply to Pavan Nikhilesh Bhagavatula

From: Jerin Jacob <[email protected]>

Optimized configuration for Marvell thunderx2 SoC. Update meson build to support Marvell thunderx2 SoC.

Product details are here:

https://www.marvell.com/server-processors/thunderx2-arm-processors/

---
 config/arm/meson.build                        |  9 ++++-
 config/defconfig_arm64-thunderx2-linuxapp-gcc | 11 ++++++
 mk/machine/thunderx2/rte.vars.mk              | 34 +++++++++++++++++++
 3 files changed, 53 insertions(+), 1 deletion(-)
 create mode 100644 config/defconfig_arm64-thunderx2-linuxapp-gcc
 create mode 100644 mk/machine/thunderx2/rte.vars.mk
config/arm/meson.build +8 -1
index 576363fc0..aca285b6a 100644
@@ -54,6 +54,12 @@ flags_default_extra = []
 flags_thunderx_extra = [
 	['RTE_MACHINE', '"thunderx"'],
 	['RTE_USE_C11_MEM_MODEL', false]]
+flags_thunderx2_extra = [
+	['RTE_MACHINE', '"thunderx2"'],
+	['RTE_CACHE_LINE_SIZE', 64],
+	['RTE_MAX_NUMA_NODES', 2],
+	['RTE_MAX_LCORE', 256],
+	['RTE_USE_C11_MEM_MODEL', true]]
 
 machine_args_generic = [
 	['default', ['-march=armv8-a+crc+crypto']],
@@ -70,7 +76,8 @@ machine_args_cavium = [
 	['native', ['-march=native']],
 	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
 	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
-	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra]]
+	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra],
+	['0xaf', ['-mcpu=thunderx2t99'], flags_thunderx2_extra]]
 
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page G7-5321)
 impl_generic = ['Generic armv8', flags_generic, machine_args_generic]
config/defconfig_arm64-thunderx2-linuxapp-gcc new file +11 -0
new file mode 100644
index 000000000..27db58e50
@@ -0,0 +1,11 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International Ltd
+#
+
+#include "defconfig_arm64-armv8a-linuxapp-gcc"
+
+CONFIG_RTE_MACHINE="thunderx2"
+
+CONFIG_RTE_CACHE_LINE_SIZE=64
+CONFIG_RTE_MAX_NUMA_NODES=2
+CONFIG_RTE_MAX_LCORE=256
mk/machine/thunderx2/rte.vars.mk new file +34 -0
new file mode 100644
index 000000000..b80dc8680
@@ -0,0 +1,34 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International 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=armv8.1-a+crc+crypto)
+MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=thunderx2t99)
Signature · 1 line
2.20.1
#25 lore raw

PATCH v4 4/5 config: add octeontx2 machine config

In reply to Pavan Nikhilesh Bhagavatula

From: Jerin Jacob <[email protected]>

Optimized configuration for Marvell octeontx2 SoC. Update meson build to support Marvell octeontx2 SoC.

---
 config/arm/meson.build                        | 10 +++++-
 config/defconfig_arm64-octeontx2-linuxapp-gcc | 18 ++++++++++
 mk/machine/octeontx2/rte.vars.mk              | 34 +++++++++++++++++++
 3 files changed, 61 insertions(+), 1 deletion(-)
 create mode 100644 config/defconfig_arm64-octeontx2-linuxapp-gcc
 create mode 100644 mk/machine/octeontx2/rte.vars.mk
config/arm/meson.build +9 -1
index aca285b6a..8086357a1 100644
@@ -60,6 +60,13 @@ flags_thunderx2_extra = [
 	['RTE_MAX_NUMA_NODES', 2],
 	['RTE_MAX_LCORE', 256],
 	['RTE_USE_C11_MEM_MODEL', true]]
+flags_octeontx2_extra = [
+	['RTE_MACHINE', '"octeontx2"'],
+	['RTE_MAX_NUMA_NODES', 1],
+	['RTE_MAX_LCORE', 24],
+	['RTE_EAL_NUMA_AWARE_HUGEPAGES', false],
+	['RTE_LIBRTE_VHOST_NUMA', false],
+	['RTE_EAL_IGB_UIO', false]]
 
 machine_args_generic = [
 	['default', ['-march=armv8-a+crc+crypto']],
@@ -77,7 +84,8 @@ machine_args_cavium = [
 	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
 	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
 	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra],
-	['0xaf', ['-mcpu=thunderx2t99'], flags_thunderx2_extra]]
+	['0xaf', ['-mcpu=thunderx2t99'], flags_thunderx2_extra],
+	['0xb2', ['-mcpu=octeontx2'], flags_octeontx2_extra]]
 
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page G7-5321)
 impl_generic = ['Generic armv8', flags_generic, machine_args_generic]
config/defconfig_arm64-octeontx2-linuxapp-gcc new file +18 -0
new file mode 100644
index 000000000..9a99eada1
@@ -0,0 +1,18 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International Ltd
+#
+
+#include "defconfig_arm64-armv8a-linuxapp-gcc"
+
+CONFIG_RTE_MACHINE="octeontx2"
+
+CONFIG_RTE_CACHE_LINE_SIZE=128
+CONFIG_RTE_MAX_NUMA_NODES=1
+CONFIG_RTE_MAX_LCORE=24
+
+# Doesn't support NUMA
+CONFIG_RTE_EAL_NUMA_AWARE_HUGEPAGES=n
+CONFIG_RTE_LIBRTE_VHOST_NUMA=n
+
+# Recommend to use VFIO as co-processors needs SMMU/IOMMU
+CONFIG_RTE_EAL_IGB_UIO=n
mk/machine/octeontx2/rte.vars.mk new file +34 -0
new file mode 100644
index 000000000..e209cf492
@@ -0,0 +1,34 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International 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+lse)
+MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=octeontx2)
Signature · 1 line
2.20.1
#26 lore raw

Re: [PATCH v4 4/5] config: add octeontx2 machine config

Hi Jerin/Pavan,

Quoted 106 lines

-----Original Message-----
From: dev <[email protected]> On Behalf Of Pavan Nikhilesh Bhagavatula
Sent: Wednesday, January 9, 2019 6:40 PM
To: [email protected]; Gavin Hu (Arm Technology China) <[email protected]>; [email protected]; [email protected]
Cc: [email protected]; Pavan Nikhilesh Bhagavatula <[email protected]>
Subject: [dpdk-dev] [PATCH v4 4/5] config: add octeontx2 machine config

From: Jerin Jacob <[email protected]>

Optimized configuration for Marvell octeontx2 SoC. Update meson build to support Marvell octeontx2 SoC.

---
 config/arm/meson.build                        | 10 +++++-
 config/defconfig_arm64-octeontx2-linuxapp-gcc | 18 ++++++++++
 mk/machine/octeontx2/rte.vars.mk              | 34 +++++++++++++++++++
 3 files changed, 61 insertions(+), 1 deletion(-)  create mode 100644
config/defconfig_arm64-octeontx2-linuxapp-gcc
 create mode 100644 mk/machine/octeontx2/rte.vars.mk
config/arm/meson.build index renamed from config/arm/meson.build +0 -0

aca285b6a..8086357a1 100644

config/arm/meson.build +7 -0
@@ -60,6 +60,13 @@ flags_thunderx2_extra = [
 	['RTE_MAX_NUMA_NODES', 2],
 	['RTE_MAX_LCORE', 256],
 	['RTE_USE_C11_MEM_MODEL', true]]
+flags_octeontx2_extra = [
+	['RTE_MACHINE', '"octeontx2"'],
+	['RTE_MAX_NUMA_NODES', 1],
+	['RTE_MAX_LCORE', 24],
+	['RTE_EAL_NUMA_AWARE_HUGEPAGES', false],
+	['RTE_LIBRTE_VHOST_NUMA', false],
+	['RTE_EAL_IGB_UIO', false]]
 
 machine_args_generic = [
 	['default', ['-march=armv8-a+crc+crypto']], @@ -77,7 +84,8 @@

machine_args_cavium = [

config/arm/meson.build +2 -1
 	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
 	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
 	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra],
-	['0xaf', ['-mcpu=thunderx2t99'], flags_thunderx2_extra]]
+	['0xaf', ['-mcpu=thunderx2t99'], flags_thunderx2_extra],
+	['0xb2', ['-mcpu=octeontx2'], flags_octeontx2_extra]]
 
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page G7-5321)

impl_generic = ['Generic armv8', flags_generic, machine_args_generic] diff --git a/config/defconfig_arm64-octeontx2-linuxapp-gcc b/config/defconfig_arm64- octeontx2-linuxapp-gcc
new file mode 100644
index 000000000..9a99eada1

config/defconfig_arm64-octeontx2-linuxapp-gcc new file +17 -0
@@ -0,0 +1,18 @@
+# SPDX-License-Identifier: BSD-3-Clause # Copyright(c) 2018 Marvell
+International Ltd #
+
+#include "defconfig_arm64-armv8a-linuxapp-gcc"
+
+CONFIG_RTE_MACHINE="octeontx2"
+
+CONFIG_RTE_CACHE_LINE_SIZE=128
+CONFIG_RTE_MAX_NUMA_NODES=1
+CONFIG_RTE_MAX_LCORE=24
+
+# Doesn't support NUMA
+CONFIG_RTE_EAL_NUMA_AWARE_HUGEPAGES=n
+CONFIG_RTE_LIBRTE_VHOST_NUMA=n
+
+# Recommend to use VFIO as co-processors needs SMMU/IOMMU
+CONFIG_RTE_EAL_IGB_UIO=n
mk/machine/octeontx2/rte.vars.mk new file +24 -0
new file mode 100644
index 000000000..e209cf492
@@ -0,0 +1,34 @@
+# SPDX-License-Identifier: BSD-3-Clause # Copyright(c) 2018 Marvell
+International 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/octeontx2/rte.vars.mk line 34

Quoted 10 lines
mk/machine/octeontx2/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+lse) MACHINE_CFLAGS += $(call
+rte_cc_has_argument, -mcpu=octeontx2)

Why do you choose to expose armv8.2-a flag other than using armv8a combine with the extension flags here?
The value 'armv8.2-a' implies 'armv8.1-a' and enables compiler support for the ARMv8.2-A architecture extensions. And the 'lse' extension is the default feature for 'armv8.1-a'. So it seems there is no need to specify 'armv8.2-a' with 'lse'.

According to the meson build code, the default -march config for Cavium is 'armv8-a'. So I think it seems better to keep 'armv8-a' here and add flags for the specific extensions. The same for thunderx2 configuration.

Quoted 2 lines

--
2.20.1

Thanks,
Phil Yang

#27 lore raw

PATCH v4 5/5 meson: add cross build targets for thunderx2 and octeontx2

In reply to Pavan Nikhilesh Bhagavatula

From: Pavan Nikhilesh <[email protected]>

Add Thunderx2 and Octeontx2 SoC cross build targets.
meson build --cross-file config/arm/arm64_<cpu>_<platform>_<compiler>

---
 config/arm/arm64_octeontx2_linuxapp_gcc | 14 ++++++++++++++
 config/arm/arm64_thunderx2_linuxapp_gcc | 14 ++++++++++++++
 2 files changed, 28 insertions(+)
 create mode 100644 config/arm/arm64_octeontx2_linuxapp_gcc
 create mode 100644 config/arm/arm64_thunderx2_linuxapp_gcc
config/arm/arm64_octeontx2_linuxapp_gcc new file +14 -0
new file mode 100644
index 000000000..83e28e817
@@ -0,0 +1,14 @@
+[binaries]
+c = 'aarch64-linux-gnu-gcc'
+cpp = 'aarch64-linux-gnu-cpp'
+ar = 'aarch64-linux-gnu-gcc-ar'
+strip = 'aarch64-linux-gnu-strip'
+
+[host_machine]
+system = 'linux'
+cpu_family = 'aarch64'
+cpu = 'armv8-a'
+endian = 'little'
+
+[properties]
+implementor_id = '0xb2'
config/arm/arm64_thunderx2_linuxapp_gcc new file +14 -0
new file mode 100644
index 000000000..de1146639
@@ -0,0 +1,14 @@
+[binaries]
+c = 'aarch64-linux-gnu-gcc'
+cpp = 'aarch64-linux-gnu-cpp'
+ar = 'aarch64-linux-gnu-gcc-ar'
+strip = 'aarch64-linux-gnu-strip'
+
+[host_machine]
+system = 'linux'
+cpu_family = 'aarch64'
+cpu = 'armv8-a'
+endian = 'little'
+
+[properties]
+implementor_id = '0xaf'
Signature · 1 line
2.20.1
#28 lore raw

Re: [PATCH v4 5/5] meson: add cross build targets for thunderx2 and octeontx2

09/01/2019 11:39, Pavan Nikhilesh Bhagavatula:

Quoted 10 lines

From: Pavan Nikhilesh <[email protected]>

Add Thunderx2 and Octeontx2 SoC cross build targets.
meson build --cross-file config/arm/arm64_<cpu>_<platform>_<compiler>

---
 config/arm/arm64_octeontx2_linuxapp_gcc | 14 ++++++++++++++
 config/arm/arm64_thunderx2_linuxapp_gcc | 14 ++++++++++++++
 2 files changed, 28 insertions(+)

Don't you think it could be merged in patches 3 & 4 for more clarity?

#29 lore raw
In reply to Pavan Nikhilesh Bhagavatula

09/01/2019 11:39, Pavan Nikhilesh Bhagavatula:

Quoted 26 lines

From: Jerin Jacob <[email protected]>

Introduce rte_cc_has_argument() Makefile helper to check a given argument is support by the compiler.

Example Usage:

include $(RTE_SDK)/mk/rte.helper.mk
MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=octeontx2)

This would allow adding -mcpu=octeontx2 in MACHINE_CFLAGS if it is only supported by the compiler. The use case for such scheme is to enable the mcpu optimization if the compiler supports else it needs to compile the source code without any errors.

This patch also moves inclusion of toolchain's rte.vars.mk to before the machine's rte.vars.mk inclusion to make correct CC available for the cross compile case.

---

mk/rte.helper.mk line 10

Quoted 10 lines
mk/rte.helper.mk new file +10 -0
@@ -0,0 +1,12 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International Ltd
+
+# rte_cc_has_argument
+# Usage: MACHINE_CFLAGS += $(call rte_cc_has_argument, -mno-avx512f)
+# Return the argument if the argument is supported by the compiler.
+#
+define rte_cc_has_argument
+	$(shell $(CC) -Werror $(1) -c -x c /dev/null -o tmp$$ 2> /dev/null && rm -f tmp$$ && echo $(1) | xargs echo -n)
+endef

What is tmp$$ ?

If the command is interrupted in the middle, temp file is not cleaned. We could fix it with "trap".
Is it possible to just avoid creating a temporary file?

#30 lore raw
Quoted 36 lines · Thomas Monjalon

-------------------------------------------------------------------
---
09/01/2019 11:39, Pavan Nikhilesh Bhagavatula:

Quoted 33 lines

From: Jerin Jacob <[email protected]>

Introduce rte_cc_has_argument() Makefile helper to check a given argument is support by the compiler.

Example Usage:

include $(RTE_SDK)/mk/rte.helper.mk
MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=octeontx2)

This would allow adding -mcpu=octeontx2 in MACHINE_CFLAGS if it is only supported by the compiler. The use case for such scheme is to enable the mcpu optimization if the compiler supports else it needs to compile the source code without any errors.

This patch also moves inclusion of toolchain's rte.vars.mk to before the machine's rte.vars.mk inclusion to make correct CC available for the cross compile case.

---

mk/rte.helper.mk new file +5 -0
@@ -0,0 +1,12 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International Ltd
+
+# rte_cc_has_argument
+# Usage: MACHINE_CFLAGS += $(call rte_cc_has_argument, -mno-

avx512f)
+# Return the argument if the argument is supported by the

Quoted 8 lines
Quoted 6 lines

compiler.

mk/rte.helper.mk new file +3 -0
+#
+define rte_cc_has_argument
+	$(shell $(CC) -Werror $(1) -c -x c /dev/null -o tmp$$ 2>

/dev/null && rm -f tmp$$ && echo $(1) | xargs echo -n) +endef

What is tmp$$ ?

It is created per process with pid value.

Quoted 3 lines

If the command is interrupted in the middle, temp file is not cleaned.

Yes. I think we can move to RTE_OUTPUT. Even it is not cleaned then it is file, I think, that would be easiest solution.

Quoted 1 line

We could fix it with "trap".

Can we do it in Makefile?

Quoted 1 line

Is it possible to just avoid creating a temporary file?

I tried it but gcc creates one.

Quoted 3 lines
#31 lore raw

14/01/2019 12:56, Jerin Jacob Kollanukkaran:

Quoted 46 lines · Thomas Monjalon
Quoted 36 lines

-------------------------------------------------------------------
---
09/01/2019 11:39, Pavan Nikhilesh Bhagavatula:

Quoted 33 lines

From: Jerin Jacob <[email protected]>

Introduce rte_cc_has_argument() Makefile helper to check a given argument is support by the compiler.

Example Usage:

include $(RTE_SDK)/mk/rte.helper.mk
MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=octeontx2)

This would allow adding -mcpu=octeontx2 in MACHINE_CFLAGS if it is only supported by the compiler. The use case for such scheme is to enable the mcpu optimization if the compiler supports else it needs to compile the source code without any errors.

This patch also moves inclusion of toolchain's rte.vars.mk to before the machine's rte.vars.mk inclusion to make correct CC available for the cross compile case.

---

mk/rte.helper.mk new file +5 -0
@@ -0,0 +1,12 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International Ltd
+
+# rte_cc_has_argument
+# Usage: MACHINE_CFLAGS += $(call rte_cc_has_argument, -mno-

avx512f)
+# Return the argument if the argument is supported by the

Quoted 8 lines
Quoted 6 lines

compiler.

mk/rte.helper.mk new file +3 -0
+#
+define rte_cc_has_argument
+	$(shell $(CC) -Werror $(1) -c -x c /dev/null -o tmp$$ 2>

/dev/null && rm -f tmp$$ && echo $(1) | xargs echo -n) +endef

What is tmp$$ ?

Quoted 1 line

It is created per process with pid value.

I see. The file is in current directory with name tmp + PID.

Quoted 9 lines
Quoted 2 lines

If the command is interrupted in the middle, temp file is not cleaned.

Yes. I think we can move to RTE_OUTPUT. Even it is not cleaned then it is file, I think, that would be easiest solution.

Quoted 1 line

We could fix it with "trap".

Can we do it in Makefile?

Yes, it is just one more command separated with ;

Quoted 3 lines
Quoted 1 line

Is it possible to just avoid creating a temporary file?

I tried it but gcc creates one.

It does not create a file in my test.

A better command is:

	$(CC) -E $(1) -xc /dev/null >/dev/null

Then you can just check the return value.

If you want rte_cc_has_argument returns a string or empty as true/false, you can do this:

	$(CC) -E $(1) -xc /dev/null >/dev/null && echo $(1)
#32 lore raw

PATCH v5 1/4 mk: introduce helper to check valid compiler argument

In reply to Pavan Nikhilesh Bhagavatula

Introduce rte_cc_has_argument() Makefile helper to check a given argument is support by the compiler.

Example Usage:

include $(RTE_SDK)/mk/rte.helper.mk
MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=octeontx2)

This would allow adding -mcpu=octeontx2 in MACHINE_CFLAGS if it is only supported by the compiler. The use case for such scheme is to enable the mcpu optimization if the compiler supports else it needs to compile the source code without any errors.

This patch also moves inclusion of toolchain's rte.vars.mk to before the machine's rte.vars.mk inclusion to make correct CC available for the cross compile case.

---

Change history of this series:

v2 Changes:
- Add meson build support.

v3 Changes:
- Squash meson build support into config support for thunderx2/octeontx2.

v4 Changes:
- Fix incorrect signoff marrvell -> marvell.

v5 Changes:
- Fix incorrect meson flag parsing(Phil Yang)
- Squash meson cross build patch(5/5) into configuration update patches for thunderx2(3/5) and octeontx2(4/5)(Thomas)
- Changed octeontx2's march as armv8-a and added the extension required instead of armv8-2a(Phil Yang)
- Improved rte_cc_has_argument() implementaion by removing the temp file(Thomas)

---
 mk/rte.helper.mk              | 10 ++++++++++
 mk/target/generic/rte.vars.mk | 22 +++++++++++-----------
 2 files changed, 21 insertions(+), 11 deletions(-)
 create mode 100644 mk/rte.helper.mk
mk/rte.helper.mk new file +10 -0
new file mode 100644
index 000000000..6e7fd03d7
@@ -0,0 +1,10 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International Ltd
+
+# rte_cc_has_argument
+# Usage: MACHINE_CFLAGS += $(call rte_cc_has_argument, -mno-avx512f)
+# Return the argument if the argument is supported by the compiler.
+#
+define rte_cc_has_argument
+	$(shell $(CC) -E $(1) -xc /dev/null 1>/dev/null 2>/dev/null && echo $(1))
+endef
mk/target/generic/rte.vars.mk +11 -11
index dd149acc9..25a578ad7 100644
@@ -7,6 +7,17 @@
 # executive environment.
 #
 
+#
+# toolchain:
+#
+#   - define CC, LD, AR, AS, ...
+#   - define TOOLCHAIN_CFLAGS variable (overridden by cmdline value)
+#   - define TOOLCHAIN_LDFLAGS variable (overridden by cmdline value)
+#   - define TOOLCHAIN_ASFLAGS variable (overridden by cmdline value)
+#   - may override any previously defined variable
+#
+include $(RTE_SDK)/mk/toolchain/$(RTE_TOOLCHAIN)/rte.vars.mk
+
 #
 # machine:
 #
@@ -45,17 +56,6 @@ endif
 #
 include $(RTE_SDK)/mk/arch/$(RTE_ARCH)/rte.vars.mk
 
-#
-# toolchain:
-#
-#   - define CC, LD, AR, AS, ...
-#   - define TOOLCHAIN_CFLAGS variable (overridden by cmdline value)
-#   - define TOOLCHAIN_LDFLAGS variable (overridden by cmdline value)
-#   - define TOOLCHAIN_ASFLAGS variable (overridden by cmdline value)
-#   - may override any previously defined variable
-#
-include $(RTE_SDK)/mk/toolchain/$(RTE_TOOLCHAIN)/rte.vars.mk
-
 #
 # exec-env:
 #
Signature · 1 line
2.20.1
#33 lore raw

PATCH v5 2/4 meson: add infra to support machine specific flags

From: Pavan Nikhilesh <[email protected]>

Currently, RTE_* flags are set based on the implementer ID but there might be some micro arch specific differences from the same vendor eg. CACHE_LINESIZE. Add support to set micro arch specific flags.

---
 config/arm/meson.build | 54 +++++++++++++++++++++++++-----------------
 1 file changed, 32 insertions(+), 22 deletions(-)
config/arm/meson.build +32 -22
index dae55d6b2..7eb6f5c99 100644
@@ -7,23 +7,6 @@ march_opt = '-march=@0@'.format(machine)
 
 arm_force_native_march = false
 
-machine_args_generic = [
-	['default', ['-march=armv8-a+crc+crypto']],
-	['native', ['-march=native']],
-	['0xd03', ['-mcpu=cortex-a53']],
-	['0xd04', ['-mcpu=cortex-a35']],
-	['0xd07', ['-mcpu=cortex-a57']],
-	['0xd08', ['-mcpu=cortex-a72']],
-	['0xd09', ['-mcpu=cortex-a73']],
-	['0xd0a', ['-mcpu=cortex-a75']],
-]
-machine_args_cavium = [
-	['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
-	['native', ['-march=native']],
-	['0xa1', ['-mcpu=thunderxt88']],
-	['0xa2', ['-mcpu=thunderxt81']],
-	['0xa3', ['-mcpu=thunderxt83']]]
-
 flags_common_default = [
 	# Accelarate rte_memcpy. Be sure to run unit test (memcpy_perf_autotest)
 	# to determine the best threshold in code. Refer to notes in source file
@@ -50,12 +33,10 @@ flags_generic = [
 	['RTE_USE_C11_MEM_MODEL', true],
 	['RTE_CACHE_LINE_SIZE', 128]]
 flags_cavium = [
-	['RTE_MACHINE', '"thunderx"'],
 	['RTE_CACHE_LINE_SIZE', 128],
 	['RTE_MAX_NUMA_NODES', 2],
 	['RTE_MAX_LCORE', 96],
-	['RTE_MAX_VFIO_GROUPS', 128],
-	['RTE_USE_C11_MEM_MODEL', false]]
+	['RTE_MAX_VFIO_GROUPS', 128]]
 flags_dpaa = [
 	['RTE_MACHINE', '"dpaa"'],
 	['RTE_USE_C11_MEM_MODEL', true],
@@ -69,6 +50,27 @@ flags_dpaa2 = [
 	['RTE_MAX_NUMA_NODES', 1],
 	['RTE_MAX_LCORE', 16],
 	['RTE_LIBRTE_DPAA2_USE_PHYS_IOVA', false]]
+flags_default_extra = []
+flags_thunderx_extra = [
+	['RTE_MACHINE', '"thunderx"'],
+	['RTE_USE_C11_MEM_MODEL', false]]
+
+machine_args_generic = [
+	['default', ['-march=armv8-a+crc+crypto']],
+	['native', ['-march=native']],
+	['0xd03', ['-mcpu=cortex-a53']],
+	['0xd04', ['-mcpu=cortex-a35']],
+	['0xd07', ['-mcpu=cortex-a57']],
+	['0xd08', ['-mcpu=cortex-a72']],
+	['0xd09', ['-mcpu=cortex-a73']],
+	['0xd0a', ['-mcpu=cortex-a75']]]
+
+machine_args_cavium = [
+	['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
+	['native', ['-march=native']],
+	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
+	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
+	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra]]
 
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page G7-5321)
 impl_generic = ['Generic armv8', flags_generic, machine_args_generic]
@@ -156,8 +158,16 @@ else
 	endif
 	foreach marg: machine[2]
 		if marg[0] == impl_pn
-			foreach f: marg[1]
-				machine_args += f
+			foreach flag: marg[1]
+				if cc.has_argument(flag)
+					machine_args += flag
+				endif
+			endforeach
+			# Apply any extra machine specific flags.
+			foreach flag: marg.get(2, flags_default_extra)
+				if flag.length() > 0
+					dpdk_conf.set(flag[0], flag[1])
+				endif
 			endforeach
 		endif
 	endforeach
Signature · 1 line
2.20.1
#34 lore raw

PATCH v5 3/4 config: add thunderx2 machine config

In reply to Jerin Jacob Kollanukkaran

Optimized configuration for Marvell thunderx2 SoC. Updated meson build to support Marvell thunderx2 SoC. Added meson cross compile target.

Product details are here:

https://www.marvell.com/server-processors/thunderx2-arm-processors/

---
 config/arm/arm64_thunderx2_linuxapp_gcc       | 15 ++++++++
 config/arm/meson.build                        |  9 ++++-
 config/defconfig_arm64-thunderx2-linuxapp-gcc | 11 ++++++
 mk/machine/thunderx2/rte.vars.mk              | 34 +++++++++++++++++++
 4 files changed, 68 insertions(+), 1 deletion(-)
 create mode 100644 config/arm/arm64_thunderx2_linuxapp_gcc
 create mode 100644 config/defconfig_arm64-thunderx2-linuxapp-gcc
 create mode 100644 mk/machine/thunderx2/rte.vars.mk
config/arm/arm64_thunderx2_linuxapp_gcc new file +15 -0
new file mode 100644
index 000000000..48b07a40c
@@ -0,0 +1,15 @@
+[binaries]
+c = 'aarch64-linux-gnu-gcc'
+cpp = 'aarch64-linux-gnu-cpp'
+ar = 'aarch64-linux-gnu-gcc-ar'
+strip = 'aarch64-linux-gnu-strip'
+
+[host_machine]
+system = 'linux'
+cpu_family = 'aarch64'
+cpu = 'armv8-a'
+endian = 'little'
+
+[properties]
+implementor_id = '0x43'
+implementor_pn = '0xaf'
config/arm/meson.build +8 -1
index 7eb6f5c99..f34913291 100644
@@ -54,6 +54,12 @@ flags_default_extra = []
 flags_thunderx_extra = [
 	['RTE_MACHINE', '"thunderx"'],
 	['RTE_USE_C11_MEM_MODEL', false]]
+flags_thunderx2_extra = [
+	['RTE_MACHINE', '"thunderx2"'],
+	['RTE_CACHE_LINE_SIZE', 64],
+	['RTE_MAX_NUMA_NODES', 2],
+	['RTE_MAX_LCORE', 256],
+	['RTE_USE_C11_MEM_MODEL', true]]
 
 machine_args_generic = [
 	['default', ['-march=armv8-a+crc+crypto']],
@@ -70,7 +76,8 @@ machine_args_cavium = [
 	['native', ['-march=native']],
 	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
 	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
-	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra]]
+	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra],
+	['0xaf', ['-mcpu=thunderx2t99'], flags_thunderx2_extra]]
 
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page G7-5321)
 impl_generic = ['Generic armv8', flags_generic, machine_args_generic]
config/defconfig_arm64-thunderx2-linuxapp-gcc new file +11 -0
new file mode 100644
index 000000000..27db58e50
@@ -0,0 +1,11 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International Ltd
+#
+
+#include "defconfig_arm64-armv8a-linuxapp-gcc"
+
+CONFIG_RTE_MACHINE="thunderx2"
+
+CONFIG_RTE_CACHE_LINE_SIZE=64
+CONFIG_RTE_MAX_NUMA_NODES=2
+CONFIG_RTE_MAX_LCORE=256
mk/machine/thunderx2/rte.vars.mk new file +34 -0
new file mode 100644
index 000000000..b80dc8680
@@ -0,0 +1,34 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International 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=armv8.1-a+crc+crypto)
+MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=thunderx2t99)
Signature · 1 line
2.20.1
#35 lore raw

PATCH v5 4/4 config: add octeontx2 machine config

In reply to Jerin Jacob Kollanukkaran

Optimized configuration for Marvell octeontx2 SoC. Updated meson build to support Marvell octeontx2 SoC. Added meson cross build target for octeontx2.

---
 config/arm/arm64_octeontx2_linuxapp_gcc       | 15 ++++++++
 config/arm/meson.build                        |  9 ++++-
 config/defconfig_arm64-octeontx2-linuxapp-gcc | 18 ++++++++++
 mk/machine/octeontx2/rte.vars.mk              | 34 +++++++++++++++++++
 4 files changed, 75 insertions(+), 1 deletion(-)
 create mode 100644 config/arm/arm64_octeontx2_linuxapp_gcc
 create mode 100644 config/defconfig_arm64-octeontx2-linuxapp-gcc
 create mode 100644 mk/machine/octeontx2/rte.vars.mk
config/arm/arm64_octeontx2_linuxapp_gcc new file +15 -0
new file mode 100644
index 000000000..7a6429b1b
@@ -0,0 +1,15 @@
+[binaries]
+c = 'aarch64-linux-gnu-gcc'
+cpp = 'aarch64-linux-gnu-cpp'
+ar = 'aarch64-linux-gnu-gcc-ar'
+strip = 'aarch64-linux-gnu-strip'
+
+[host_machine]
+system = 'linux'
+cpu_family = 'aarch64'
+cpu = 'armv8-a'
+endian = 'little'
+
+[properties]
+implementor_id = '0x43'
+implementor_pn = '0xb2'
config/arm/meson.build +8 -1
index f34913291..5edaeec13 100644
@@ -60,6 +60,12 @@ flags_thunderx2_extra = [
 	['RTE_MAX_NUMA_NODES', 2],
 	['RTE_MAX_LCORE', 256],
 	['RTE_USE_C11_MEM_MODEL', true]]
+flags_octeontx2_extra = [
+	['RTE_MACHINE', '"octeontx2"'],
+	['RTE_MAX_NUMA_NODES', 1],
+	['RTE_MAX_LCORE', 24],
+	['RTE_EAL_IGB_UIO', false],
+	['RTE_USE_C11_MEM_MODEL', true]]
 
 machine_args_generic = [
 	['default', ['-march=armv8-a+crc+crypto']],
@@ -77,7 +83,8 @@ machine_args_cavium = [
 	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
 	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
 	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra],
-	['0xaf', ['-mcpu=thunderx2t99'], flags_thunderx2_extra]]
+	['0xaf', ['-mcpu=thunderx2t99'], flags_thunderx2_extra],
+	['0xb2', ['-mcpu=octeontx2'], flags_octeontx2_extra]]
 
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page G7-5321)
 impl_generic = ['Generic armv8', flags_generic, machine_args_generic]
config/defconfig_arm64-octeontx2-linuxapp-gcc new file +18 -0
new file mode 100644
index 000000000..9a99eada1
@@ -0,0 +1,18 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International Ltd
+#
+
+#include "defconfig_arm64-armv8a-linuxapp-gcc"
+
+CONFIG_RTE_MACHINE="octeontx2"
+
+CONFIG_RTE_CACHE_LINE_SIZE=128
+CONFIG_RTE_MAX_NUMA_NODES=1
+CONFIG_RTE_MAX_LCORE=24
+
+# Doesn't support NUMA
+CONFIG_RTE_EAL_NUMA_AWARE_HUGEPAGES=n
+CONFIG_RTE_LIBRTE_VHOST_NUMA=n
+
+# Recommend to use VFIO as co-processors needs SMMU/IOMMU
+CONFIG_RTE_EAL_IGB_UIO=n
mk/machine/octeontx2/rte.vars.mk new file +34 -0
new file mode 100644
index 000000000..e0a7498a3
@@ -0,0 +1,34 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International 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-a+crc+crypto+lse)
+MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=octeontx2)
Signature · 1 line
2.20.1
#36 lore raw

PATCH v6 1/4 mk: introduce helper to check valid compiler argument

In reply to Jerin Jacob Kollanukkaran

From: Jerin Jacob <[email protected]>

Introduce rte_cc_has_argument() Makefile helper to check a given argument is support by the compiler.

Example Usage:

include $(RTE_SDK)/mk/rte.helper.mk
MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=octeontx2)

This would allow adding -mcpu=octeontx2 in MACHINE_CFLAGS if it is only supported by the compiler. The use case for such scheme is to enable the mcpu optimization if the compiler supports else it needs to compile the source code without any errors.

This patch also moves inclusion of toolchain's rte.vars.mk to before the machine's rte.vars.mk inclusion to make correct CC available for the cross compile case.

---

Change history of this series:

v2 Changes:
- Add meson build support.

v3 Changes:
- Squash meson build support into config support for thunderx2/octeontx2.

v4 Changes:
- Fix incorrect signoff marrvell -> marvell.

v5 Changes:
- Fix incorrect meson flag parsing(Phil Yang)
- Squash meson cross build patch(5/5) into configuration update patches for thunderx2(3/5) and octeontx2(4/5)(Thomas)
- Changed octeontx2's march as armv8-a and added the extension required instead of armv8-2a(Phil Yang)
- Improved rte_cc_has_argument() implementaion by removing the temp file(Thomas)

v6 Changes:
- Rework to change the config files to sync with "mk: use linux and freebsd in config names"
- Fix the following error with latest gcc by fixing the mcpu type
cc1: error: switch -mcpu=armv8.2-a conflicts with -march=armv8-a switch

---
 mk/rte.helper.mk              | 10 ++++++++++
 mk/target/generic/rte.vars.mk | 22 +++++++++++-----------
 2 files changed, 21 insertions(+), 11 deletions(-)
 create mode 100644 mk/rte.helper.mk
mk/rte.helper.mk new file +10 -0
new file mode 100644
index 000000000..6e7fd03d7
@@ -0,0 +1,10 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International Ltd
+
+# rte_cc_has_argument
+# Usage: MACHINE_CFLAGS += $(call rte_cc_has_argument, -mno-avx512f)
+# Return the argument if the argument is supported by the compiler.
+#
+define rte_cc_has_argument
+	$(shell $(CC) -E $(1) -xc /dev/null 1>/dev/null 2>/dev/null && echo $(1))
+endef
mk/target/generic/rte.vars.mk +11 -11
index dd149acc9..25a578ad7 100644
@@ -7,6 +7,17 @@
 # executive environment.
 #
 
+#
+# toolchain:
+#
+#   - define CC, LD, AR, AS, ...
+#   - define TOOLCHAIN_CFLAGS variable (overridden by cmdline value)
+#   - define TOOLCHAIN_LDFLAGS variable (overridden by cmdline value)
+#   - define TOOLCHAIN_ASFLAGS variable (overridden by cmdline value)
+#   - may override any previously defined variable
+#
+include $(RTE_SDK)/mk/toolchain/$(RTE_TOOLCHAIN)/rte.vars.mk
+
 #
 # machine:
 #
@@ -45,17 +56,6 @@ endif
 #
 include $(RTE_SDK)/mk/arch/$(RTE_ARCH)/rte.vars.mk
 
-#
-# toolchain:
-#
-#   - define CC, LD, AR, AS, ...
-#   - define TOOLCHAIN_CFLAGS variable (overridden by cmdline value)
-#   - define TOOLCHAIN_LDFLAGS variable (overridden by cmdline value)
-#   - define TOOLCHAIN_ASFLAGS variable (overridden by cmdline value)
-#   - may override any previously defined variable
-#
-include $(RTE_SDK)/mk/toolchain/$(RTE_TOOLCHAIN)/rte.vars.mk
-
 #
 # exec-env:
 #
Signature · 1 line
2.21.0
#37 lore raw

PATCH v6 2/4 meson: add infra to support machine specific flags

From: Pavan Nikhilesh <[email protected]>

Currently, RTE_* flags are set based on the implementer ID but there might be some micro arch specific differences from the same vendor eg. CACHE_LINESIZE. Add support to set micro arch specific flags.

---
 config/arm/meson.build | 37 ++++++++++++++++++++++++++++++++-----
 1 file changed, 32 insertions(+), 5 deletions(-)
config/arm/meson.build +32 -5
index 8e892fa77..3db6d6445 100644
@@ -52,12 +52,10 @@ flags_generic = [
 	['RTE_USE_C11_MEM_MODEL', true],
 	['RTE_CACHE_LINE_SIZE', 128]]
 flags_cavium = [
-	['RTE_MACHINE', '"thunderx"'],
 	['RTE_CACHE_LINE_SIZE', 128],
 	['RTE_MAX_NUMA_NODES', 2],
 	['RTE_MAX_LCORE', 96],
-	['RTE_MAX_VFIO_GROUPS', 128],
-	['RTE_USE_C11_MEM_MODEL', false]]
+	['RTE_MAX_VFIO_GROUPS', 128]]
 flags_dpaa = [
 	['RTE_MACHINE', '"dpaa"'],
 	['RTE_USE_C11_MEM_MODEL', true],
@@ -71,6 +69,27 @@ flags_dpaa2 = [
 	['RTE_MAX_NUMA_NODES', 1],
 	['RTE_MAX_LCORE', 16],
 	['RTE_LIBRTE_DPAA2_USE_PHYS_IOVA', false]]
+flags_default_extra = []
+flags_thunderx_extra = [
+	['RTE_MACHINE', '"thunderx"'],
+	['RTE_USE_C11_MEM_MODEL', false]]
+
+machine_args_generic = [
+	['default', ['-march=armv8-a+crc+crypto']],
+	['native', ['-march=native']],
+	['0xd03', ['-mcpu=cortex-a53']],
+	['0xd04', ['-mcpu=cortex-a35']],
+	['0xd07', ['-mcpu=cortex-a57']],
+	['0xd08', ['-mcpu=cortex-a72']],
+	['0xd09', ['-mcpu=cortex-a73']],
+	['0xd0a', ['-mcpu=cortex-a75']]]
+
+machine_args_cavium = [
+	['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
+	['native', ['-march=native']],
+	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
+	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
+	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra]]
 
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page G7-5321)
 impl_generic = ['Generic armv8', flags_generic, machine_args_generic]
@@ -157,8 +176,16 @@ else
 	endif
 	foreach marg: machine[2]
 		if marg[0] == impl_pn
-			foreach f: marg[1]
-				machine_args += f
+			foreach flag: marg[1]
+				if cc.has_argument(flag)
+					machine_args += flag
+				endif
+			endforeach
+			# Apply any extra machine specific flags.
+			foreach flag: marg.get(2, flags_default_extra)
+				if flag.length() > 0
+					dpdk_conf.set(flag[0], flag[1])
+				endif
 			endforeach
 		endif
 	endforeach
Signature · 1 line
2.21.0
#38 lore raw

Re: [PATCH v6 2/4] meson: add infra to support machine specific flags

Quoted 6 lines · Jerin Jacob Kollanukkaran

From: Pavan Nikhilesh <[email protected]>

Currently, RTE_* flags are set based on the implementer ID but there might be some micro arch specific differences from the same vendor eg. CACHE_LINESIZE. Add support to set micro arch specific flags.

Quoted 6 lines
---
 config/arm/meson.build | 37 ++++++++++++++++++++++++++++++++-----
 1 file changed, 32 insertions(+), 5 deletions(-)

This looks ok to me, but I think review and ack from another Arm vendor would be good to have.

/Bruce

#39 lore raw

Re: [PATCH v6 2/4] meson: add infra to support machine specific flags

19/03/2019 10:40, Bruce Richardson:

Quoted 7 lines · Jerin Jacob Kollanukkaran
Quoted 6 lines

From: Pavan Nikhilesh <[email protected]>

Currently, RTE_* flags are set based on the implementer ID but there might be some micro arch specific differences from the same vendor eg. CACHE_LINESIZE. Add support to set micro arch specific flags.

Quoted 8 lines
Quoted 6 lines
---
 config/arm/meson.build | 37 ++++++++++++++++++++++++++++++++-----
 1 file changed, 32 insertions(+), 5 deletions(-)

This looks ok to me, but I think review and ack from another Arm vendor would be good to have.

Adding Cc's for more reviews

#40 lore raw

PATCH v6 3/4 config: add thunderx2 machine config

In reply to Jerin Jacob Kollanukkaran

From: Jerin Jacob <[email protected]>

Optimized configuration for Marvell thunderx2 SoC. Updated meson build to support Marvell thunderx2 SoC. Added meson cross compile target.

Product details are here:

https://www.marvell.com/server-processors/thunderx2-arm-processors/

---
 config/arm/arm64_thunderx2_linux_gcc          | 15 ++++++++
 config/arm/meson.build                        |  9 ++++-
 config/defconfig_arm64-thunderx2-linux-gcc    |  1 +
 config/defconfig_arm64-thunderx2-linuxapp-gcc | 11 ++++++
 mk/machine/thunderx2/rte.vars.mk              | 34 +++++++++++++++++++
 5 files changed, 69 insertions(+), 1 deletion(-)
 create mode 100644 config/arm/arm64_thunderx2_linux_gcc
 create mode 120000 config/defconfig_arm64-thunderx2-linux-gcc
 create mode 100644 config/defconfig_arm64-thunderx2-linuxapp-gcc
 create mode 100644 mk/machine/thunderx2/rte.vars.mk
config/arm/arm64_thunderx2_linux_gcc new file +15 -0
new file mode 100644
index 000000000..48b07a40c
@@ -0,0 +1,15 @@
+[binaries]
+c = 'aarch64-linux-gnu-gcc'
+cpp = 'aarch64-linux-gnu-cpp'
+ar = 'aarch64-linux-gnu-gcc-ar'
+strip = 'aarch64-linux-gnu-strip'
+
+[host_machine]
+system = 'linux'
+cpu_family = 'aarch64'
+cpu = 'armv8-a'
+endian = 'little'
+
+[properties]
+implementor_id = '0x43'
+implementor_pn = '0xaf'
config/arm/meson.build +8 -1
index 3db6d6445..5c391ed3c 100644
@@ -73,6 +73,12 @@ flags_default_extra = []
 flags_thunderx_extra = [
 	['RTE_MACHINE', '"thunderx"'],
 	['RTE_USE_C11_MEM_MODEL', false]]
+flags_thunderx2_extra = [
+	['RTE_MACHINE', '"thunderx2"'],
+	['RTE_CACHE_LINE_SIZE', 64],
+	['RTE_MAX_NUMA_NODES', 2],
+	['RTE_MAX_LCORE', 256],
+	['RTE_USE_C11_MEM_MODEL', true]]
 
 machine_args_generic = [
 	['default', ['-march=armv8-a+crc+crypto']],
@@ -89,7 +95,8 @@ machine_args_cavium = [
 	['native', ['-march=native']],
 	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
 	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
-	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra]]
+	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra],
+	['0xaf', ['-mcpu=thunderx2t99'], flags_thunderx2_extra]]
 
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page G7-5321)
 impl_generic = ['Generic armv8', flags_generic, machine_args_generic]
config/defconfig_arm64-thunderx2-linux-gcc new file +1 -0
new file mode 120000
index 000000000..b40a760b1
@@ -0,0 +1 @@
+defconfig_arm64-thunderx2-linuxapp-gcc
\ No newline at end of file
config/defconfig_arm64-thunderx2-linuxapp-gcc new file +11 -0
new file mode 100644
index 000000000..cc5c64ba0
@@ -0,0 +1,11 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International Ltd
+#
+
+#include "defconfig_arm64-armv8a-linux-gcc"
+
+CONFIG_RTE_MACHINE="thunderx2"
+
+CONFIG_RTE_CACHE_LINE_SIZE=64
+CONFIG_RTE_MAX_NUMA_NODES=2
+CONFIG_RTE_MAX_LCORE=256
mk/machine/thunderx2/rte.vars.mk new file +34 -0
new file mode 100644
index 000000000..b80dc8680
@@ -0,0 +1,34 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International 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=armv8.1-a+crc+crypto)
+MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=thunderx2t99)
Signature · 1 line
2.21.0
Reviewed-by on v6 3/4
#41 lore raw

Re: [PATCH v6 3/4] config: add thunderx2 machine config

Quoted 151 lines

-----Original Message-----
From: Jerin Jacob Kollanukkaran <[email protected]>
Sent: Tuesday, March 19, 2019 12:50 AM
To: [email protected]
Cc: [email protected]; Gavin Hu (Arm Technology China) <[email protected]>; [email protected]; Pavan Nikhilesh Bhagavatula <[email protected]>
Subject: [dpdk-dev] [PATCH v6 3/4] config: add thunderx2 machine config

From: Jerin Jacob <[email protected]>

Optimized configuration for Marvell thunderx2 SoC. Updated meson build to support Marvell thunderx2 SoC. Added meson cross compile target.

Product details are here:

https://www.marvell.com/server-processors/thunderx2-arm-processors/

---
 config/arm/arm64_thunderx2_linux_gcc          | 15 ++++++++
 config/arm/meson.build                        |  9 ++++-
 config/defconfig_arm64-thunderx2-linux-gcc    |  1 +
 config/defconfig_arm64-thunderx2-linuxapp-gcc | 11 ++++++
 mk/machine/thunderx2/rte.vars.mk              | 34 +++++++++++++++++++
 5 files changed, 69 insertions(+), 1 deletion(-)
 create mode 100644 config/arm/arm64_thunderx2_linux_gcc
 create mode 120000 config/defconfig_arm64-thunderx2-linux-gcc
 create mode 100644 config/defconfig_arm64-thunderx2-linuxapp-gcc
 create mode 100644 mk/machine/thunderx2/rte.vars.mk
config/arm/arm64_thunderx2_linux_gcc new file +15 -0
new file mode 100644
index 000000000..48b07a40c
@@ -0,0 +1,15 @@
+[binaries]
+c = 'aarch64-linux-gnu-gcc'
+cpp = 'aarch64-linux-gnu-cpp'
+ar = 'aarch64-linux-gnu-gcc-ar'
+strip = 'aarch64-linux-gnu-strip'
+
+[host_machine]
+system = 'linux'
+cpu_family = 'aarch64'
+cpu = 'armv8-a'
+endian = 'little'
+
+[properties]
+implementor_id = '0x43'
+implementor_pn = '0xaf'
config/arm/meson.build +8 -1
index 3db6d6445..5c391ed3c 100644
@@ -73,6 +73,12 @@ flags_default_extra = []
 flags_thunderx_extra = [
 ['RTE_MACHINE', '"thunderx"'],
 ['RTE_USE_C11_MEM_MODEL', false]]
+flags_thunderx2_extra = [
+['RTE_MACHINE', '"thunderx2"'],
+['RTE_CACHE_LINE_SIZE', 64],
+['RTE_MAX_NUMA_NODES', 2],
+['RTE_MAX_LCORE', 256],
+['RTE_USE_C11_MEM_MODEL', true]]
 
 machine_args_generic = [
 ['default', ['-march=armv8-a+crc+crypto']],
@@ -89,7 +95,8 @@ machine_args_cavium = [
 ['native', ['-march=native']],
 ['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
 ['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
-['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra]]
+['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra],
+['0xaf', ['-mcpu=thunderx2t99'], flags_thunderx2_extra]]
 
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page G7-5321)
 impl_generic = ['Generic armv8', flags_generic, machine_args_generic]
config/defconfig_arm64-thunderx2-linux-gcc new file +1 -0
new file mode 120000
index 000000000..b40a760b1
@@ -0,0 +1 @@
+defconfig_arm64-thunderx2-linuxapp-gcc
\ No newline at end of file
config/defconfig_arm64-thunderx2-linuxapp-gcc new file +11 -0
new file mode 100644
index 000000000..cc5c64ba0
@@ -0,0 +1,11 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International Ltd
+#
+
+#include "defconfig_arm64-armv8a-linux-gcc"
+
+CONFIG_RTE_MACHINE="thunderx2"
+
+CONFIG_RTE_CACHE_LINE_SIZE=64
+CONFIG_RTE_MAX_NUMA_NODES=2
+CONFIG_RTE_MAX_LCORE=256
mk/machine/thunderx2/rte.vars.mk new file +33 -0
new file mode 100644
index 000000000..b80dc8680
@@ -0,0 +1,34 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International 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=armv8.1-
Quoted 4 lines

a+crc+crypto)
+MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=thunderx2t99) --
2.21.0

IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.

#42 lore raw

Re: [dpdk-dev] [PATCH v6 3/4] config: add thunderx2 machine config

In reply to Jerin Jacob Kollanukkaran

18/03/2019 17:50, Jerin Jacob Kollanukkaran:

Quoted 6 lines

From: Jerin Jacob <[email protected]>

Optimized configuration for Marvell thunderx2 SoC. Updated meson build to support Marvell thunderx2 SoC. Added meson cross compile target.

Quoted 6 lines

Product details are here:

https://www.marvell.com/server-processors/thunderx2-arm-processors/

I see a very strange error when compiling with meson:

In file included from /usr/include/inttypes.h:27,

                 from /usr/include/pcap/pcap-inttypes.h:114,
                 from /usr/include/pcap/pcap.h:74,
                 from /usr/include/pcap.h:43,
                 from ../drivers/net/pcap/rte_eth_pcap.c:19:

/usr/include/stdint.h:109: error: "__INT64_C" redefined [-Werror]

Could we try to debug it and merge these patches in -rc2?

#43 lore raw

Re: [dpdk-dev] [PATCH v6 3/4] config: add thunderx2 machine config

Quoted 25 lines

-----Original Message-----
From: Thomas Monjalon <[email protected]>
Sent: Friday, April 5, 2019 11:43 PM
To: Jerin Jacob Kollanukkaran <[email protected]>; Pavan Nikhilesh Bhagavatula <[email protected]>
Cc: [email protected]; [email protected]
Subject: Re: [dpdk-dev] [PATCH v6 3/4] config: add thunderx2 machine config

18/03/2019 17:50, Jerin Jacob Kollanukkaran:

Quoted 6 lines

From: Jerin Jacob <[email protected]>

Optimized configuration for Marvell thunderx2 SoC. Updated meson build to support Marvell thunderx2 SoC. Added meson cross compile target.

Quoted 6 lines

Product details are here:

https://www.marvell.com/server-processors/thunderx2-arm-processors/

I see a very strange error when compiling with meson:

Quoted 8 lines

In file included from /usr/include/inttypes.h:27,

                 from /usr/include/pcap/pcap-inttypes.h:114,
                 from /usr/include/pcap/pcap.h:74,
                 from /usr/include/pcap.h:43,
                 from ../drivers/net/pcap/rte_eth_pcap.c:19:

/usr/include/stdint.h:109: error: "__INT64_C" redefined [-Werror]

Could we try to debug it and merge these patches in -rc2?

Could you please give more details to reproduce this issue
Compiler version - gcc or clang version?
Build - Native or cross ?
Build OS ?
Meson and Ninja - Versions?
The command use to reproduce this issue as we are not able to get this Issue in our local setup?

Quoted 1 line
#44 lore raw

Re: [dpdk-dev] [PATCH v6 3/4] config: add thunderx2 machine config

05/04/2019 20:47, Jerin Jacob Kollanukkaran:

Quoted 24 lines

From: Thomas Monjalon <[email protected]>

Quoted 23 lines

18/03/2019 17:50, Jerin Jacob Kollanukkaran:

Quoted 6 lines

From: Jerin Jacob <[email protected]>

Optimized configuration for Marvell thunderx2 SoC. Updated meson build to support Marvell thunderx2 SoC. Added meson cross compile target.

Quoted 6 lines

Product details are here:

https://www.marvell.com/server-processors/thunderx2-arm-processors/

I see a very strange error when compiling with meson:

In file included from /usr/include/inttypes.h:27,

                 from /usr/include/pcap/pcap-inttypes.h:114,
                 from /usr/include/pcap/pcap.h:74,
                 from /usr/include/pcap.h:43,
                 from ../drivers/net/pcap/rte_eth_pcap.c:19:

/usr/include/stdint.h:109: error: "__INT64_C" redefined [-Werror]

Quoted 3 lines
Quoted 1 line

Could we try to debug it and merge these patches in -rc2?

Could you please give more details to reproduce this issue

First of all, we need to understand why it is getting pcap includes from the native system while compiling for thunderx2.

Quoted 1 line

Compiler version - gcc or clang version?

aarch64-linux-gnu-gcc (GCC) 8.3.0

Quoted 1 line

Build - Native or cross ?

cross-compilation

Quoted 1 line

Build OS ?

My host is an archlinux

Quoted 1 line

Meson and Ninja - Versions?

meson 0.49.2
ninja 1.9.0

Quoted 1 line

The command use to reproduce this issue as we are not able to get this Issue in our local setup?

I just run devtools/test-meson-builds.sh
and it compiles for config/arm/arm64_thunderx_linux_gcc because I have an Arm toolchain installed.

#45 lore raw

Re: [dpdk-dev] [PATCH v6 3/4] config: add thunderx2 machine config

Quoted 18 lines · Thomas Monjalon

05/04/2019 20:47, Jerin Jacob Kollanukkaran:

Quoted 17 lines

From: Thomas Monjalon <[email protected]>

Quoted 16 lines

18/03/2019 17:50, Jerin Jacob Kollanukkaran:

Quoted 6 lines

From: Jerin Jacob <[email protected]>

Optimized configuration for Marvell thunderx2 SoC. Updated meson build to support Marvell thunderx2 SoC. Added meson cross compile target.

Quoted 6 lines

Product details are here:

https://www.marvell.com/server-processors/thunderx2-arm-processors/

I see a very strange error when compiling with meson:

cross compiling config files needs an update since "build: improve pcap dependency handling" changeset.

I will spin another version to fix this.

Quoted 9 lines
Quoted 9 lines
Quoted 8 lines

In file included from /usr/include/inttypes.h:27,

                 from /usr/include/pcap/pcap-inttypes.h:114,
                 from /usr/include/pcap/pcap.h:74,
                 from /usr/include/pcap.h:43,
                 from ../drivers/net/pcap/rte_eth_pcap.c:19:

/usr/include/stdint.h:109: error: "__INT64_C" redefined [-Werror]

Could we try to debug it and merge these patches in -rc2?

#46 lore raw

PATCH v6 4/4 config: add octeontx2 machine config

In reply to Jerin Jacob Kollanukkaran

From: Jerin Jacob <[email protected]>

Optimized configuration for Marvell octeontx2 SoC. Updated meson build to support Marvell octeontx2 SoC. Added meson cross build target for octeontx2.

---
 config/arm/arm64_octeontx2_linux_gcc          | 15 ++++++++
 config/arm/meson.build                        |  9 ++++-
 config/defconfig_arm64-octeontx2-linux-gcc    |  1 +
 config/defconfig_arm64-octeontx2-linuxapp-gcc | 18 ++++++++++
 mk/machine/octeontx2/rte.vars.mk              | 34 +++++++++++++++++++
 5 files changed, 76 insertions(+), 1 deletion(-)
 create mode 100644 config/arm/arm64_octeontx2_linux_gcc
 create mode 120000 config/defconfig_arm64-octeontx2-linux-gcc
 create mode 100644 config/defconfig_arm64-octeontx2-linuxapp-gcc
 create mode 100644 mk/machine/octeontx2/rte.vars.mk
config/arm/arm64_octeontx2_linux_gcc new file +15 -0
new file mode 100644
index 000000000..7a6429b1b
@@ -0,0 +1,15 @@
+[binaries]
+c = 'aarch64-linux-gnu-gcc'
+cpp = 'aarch64-linux-gnu-cpp'
+ar = 'aarch64-linux-gnu-gcc-ar'
+strip = 'aarch64-linux-gnu-strip'
+
+[host_machine]
+system = 'linux'
+cpu_family = 'aarch64'
+cpu = 'armv8-a'
+endian = 'little'
+
+[properties]
+implementor_id = '0x43'
+implementor_pn = '0xb2'
config/arm/meson.build +8 -1
index 5c391ed3c..87ae40c24 100644
@@ -79,6 +79,12 @@ flags_thunderx2_extra = [
 	['RTE_MAX_NUMA_NODES', 2],
 	['RTE_MAX_LCORE', 256],
 	['RTE_USE_C11_MEM_MODEL', true]]
+flags_octeontx2_extra = [
+	['RTE_MACHINE', '"octeontx2"'],
+	['RTE_MAX_NUMA_NODES', 1],
+	['RTE_MAX_LCORE', 24],
+	['RTE_EAL_IGB_UIO', false],
+	['RTE_USE_C11_MEM_MODEL', true]]
 
 machine_args_generic = [
 	['default', ['-march=armv8-a+crc+crypto']],
@@ -96,7 +102,8 @@ machine_args_cavium = [
 	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
 	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
 	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra],
-	['0xaf', ['-mcpu=thunderx2t99'], flags_thunderx2_extra]]
+	['0xaf', ['-mcpu=thunderx2t99'], flags_thunderx2_extra],
+	['0xb2', ['-mcpu=octeontx2'], flags_octeontx2_extra]]
 
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page G7-5321)
 impl_generic = ['Generic armv8', flags_generic, machine_args_generic]
config/defconfig_arm64-octeontx2-linux-gcc new file +1 -0
new file mode 120000
index 000000000..e25150531
@@ -0,0 +1 @@
+defconfig_arm64-octeontx2-linuxapp-gcc
\ No newline at end of file
config/defconfig_arm64-octeontx2-linuxapp-gcc new file +18 -0
new file mode 100644
index 000000000..9eae84538
@@ -0,0 +1,18 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International Ltd
+#
+
+#include "defconfig_arm64-armv8a-linux-gcc"
+
+CONFIG_RTE_MACHINE="octeontx2"
+
+CONFIG_RTE_CACHE_LINE_SIZE=128
+CONFIG_RTE_MAX_NUMA_NODES=1
+CONFIG_RTE_MAX_LCORE=24
+
+# Doesn't support NUMA
+CONFIG_RTE_EAL_NUMA_AWARE_HUGEPAGES=n
+CONFIG_RTE_LIBRTE_VHOST_NUMA=n
+
+# Recommend to use VFIO as co-processors needs SMMU/IOMMU
+CONFIG_RTE_EAL_IGB_UIO=n
mk/machine/octeontx2/rte.vars.mk new file +34 -0
new file mode 100644
index 000000000..cbec7f14d
@@ -0,0 +1,34 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International 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=-mcpu=armv8.2-a+crc+crypto+lse)
+MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=octeontx2)
Signature · 1 line
2.21.0
Reviewed-by on v6 4/4
#47 lore raw

Re: [PATCH v6 4/4] config: add octeontx2 machine config

Quoted 153 lines

-----Original Message-----
From: Jerin Jacob Kollanukkaran <[email protected]>
Sent: Tuesday, March 19, 2019 12:50 AM
To: [email protected]
Cc: [email protected]; Gavin Hu (Arm Technology China) <[email protected]>; [email protected]; Pavan Nikhilesh Bhagavatula <[email protected]>
Subject: [dpdk-dev] [PATCH v6 4/4] config: add octeontx2 machine config

From: Jerin Jacob <[email protected]>

Optimized configuration for Marvell octeontx2 SoC. Updated meson build to support Marvell octeontx2 SoC. Added meson cross build target for octeontx2.

---
 config/arm/arm64_octeontx2_linux_gcc          | 15 ++++++++
 config/arm/meson.build                        |  9 ++++-
 config/defconfig_arm64-octeontx2-linux-gcc    |  1 +
 config/defconfig_arm64-octeontx2-linuxapp-gcc | 18 ++++++++++
 mk/machine/octeontx2/rte.vars.mk              | 34 +++++++++++++++++++
 5 files changed, 76 insertions(+), 1 deletion(-)
 create mode 100644 config/arm/arm64_octeontx2_linux_gcc
 create mode 120000 config/defconfig_arm64-octeontx2-linux-gcc
 create mode 100644 config/defconfig_arm64-octeontx2-linuxapp-gcc
 create mode 100644 mk/machine/octeontx2/rte.vars.mk
config/arm/arm64_octeontx2_linux_gcc new file +15 -0
new file mode 100644
index 000000000..7a6429b1b
@@ -0,0 +1,15 @@
+[binaries]
+c = 'aarch64-linux-gnu-gcc'
+cpp = 'aarch64-linux-gnu-cpp'
+ar = 'aarch64-linux-gnu-gcc-ar'
+strip = 'aarch64-linux-gnu-strip'
+
+[host_machine]
+system = 'linux'
+cpu_family = 'aarch64'
+cpu = 'armv8-a'
+endian = 'little'
+
+[properties]
+implementor_id = '0x43'
+implementor_pn = '0xb2'
config/arm/meson.build +8 -1
index 5c391ed3c..87ae40c24 100644
@@ -79,6 +79,12 @@ flags_thunderx2_extra = [
 ['RTE_MAX_NUMA_NODES', 2],
 ['RTE_MAX_LCORE', 256],
 ['RTE_USE_C11_MEM_MODEL', true]]
+flags_octeontx2_extra = [
+['RTE_MACHINE', '"octeontx2"'],
+['RTE_MAX_NUMA_NODES', 1],
+['RTE_MAX_LCORE', 24],
+['RTE_EAL_IGB_UIO', false],
+['RTE_USE_C11_MEM_MODEL', true]]
 
 machine_args_generic = [
 ['default', ['-march=armv8-a+crc+crypto']],
@@ -96,7 +102,8 @@ machine_args_cavium = [
 ['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
 ['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
 ['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra],
-['0xaf', ['-mcpu=thunderx2t99'], flags_thunderx2_extra]]
+['0xaf', ['-mcpu=thunderx2t99'], flags_thunderx2_extra],
+['0xb2', ['-mcpu=octeontx2'], flags_octeontx2_extra]]
 
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page G7-5321)
 impl_generic = ['Generic armv8', flags_generic, machine_args_generic]
config/defconfig_arm64-octeontx2-linux-gcc new file +1 -0
new file mode 120000
index 000000000..e25150531
@@ -0,0 +1 @@
+defconfig_arm64-octeontx2-linuxapp-gcc
\ No newline at end of file
config/defconfig_arm64-octeontx2-linuxapp-gcc new file +18 -0
new file mode 100644
index 000000000..9eae84538
@@ -0,0 +1,18 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International Ltd
+#
+
+#include "defconfig_arm64-armv8a-linux-gcc"
+
+CONFIG_RTE_MACHINE="octeontx2"
+
+CONFIG_RTE_CACHE_LINE_SIZE=128
+CONFIG_RTE_MAX_NUMA_NODES=1
+CONFIG_RTE_MAX_LCORE=24
+
+# Doesn't support NUMA
+CONFIG_RTE_EAL_NUMA_AWARE_HUGEPAGES=n
+CONFIG_RTE_LIBRTE_VHOST_NUMA=n
+
+# Recommend to use VFIO as co-processors needs SMMU/IOMMU
+CONFIG_RTE_EAL_IGB_UIO=n
mk/machine/octeontx2/rte.vars.mk new file +33 -0
new file mode 100644
index 000000000..cbec7f14d
@@ -0,0 +1,34 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International 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=-
Quoted 4 lines

mcpu=armv8.2-a+crc+crypto+lse)
+MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=octeontx2) --
2.21.0

IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.

#48 lore raw

PATCH v7 1/4 mk: introduce helper to check valid compiler argument

In reply to Jerin Jacob Kollanukkaran

From: Jerin Jacob <[email protected]>

Introduce rte_cc_has_argument() Makefile helper to check a given argument is support by the compiler.

Example Usage:

include $(RTE_SDK)/mk/rte.helper.mk
MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=octeontx2)

This would allow adding -mcpu=octeontx2 in MACHINE_CFLAGS if it is only supported by the compiler. The use case for such scheme is to enable the mcpu optimization if the compiler supports else it needs to compile the source code without any errors.

This patch also moves inclusion of toolchain's rte.vars.mk to before the machine's rte.vars.mk inclusion to make correct CC available for the cross compile case.

---

Change history of this series:

v7 Changes:
- Updated cross compile config files align with
"build: improve pcap dependency handling" changeset to fix build issue with meson

- Some compiler needs the following depended patch to compile with meson
http://patches.dpdk.org/patch/52367/

v6 Changes:
- Rework to change the config files to sync with "mk: use linux and freebsd in config names"
- Fix the following error with latest gcc by fixing the mcpu type
cc1: error: switch -mcpu=armv8.2-a conflicts with -march=armv8-a switch

v5 Changes:
- Fix incorrect meson flag parsing(Phil Yang)
- Squash meson cross build patch(5/5) into configuration update patches for thunderx2(3/5) and octeontx2(4/5)(Thomas)
- Changed octeontx2's march as armv8-a and added the extension required instead of armv8-2a(Phil Yang)
- Improved rte_cc_has_argument() implementaion by removing the temp file(Thomas)

v4 Changes:
- Fix incorrect signoff marrvell -> marvell.

v3 Changes:
- Squash meson build support into config support for thunderx2/octeontx2.

v2 Changes:
- Add meson build support.

---
 mk/rte.helper.mk              | 10 ++++++++++
 mk/target/generic/rte.vars.mk | 22 +++++++++++-----------
 2 files changed, 21 insertions(+), 11 deletions(-)
 create mode 100644 mk/rte.helper.mk
mk/rte.helper.mk new file +10 -0
new file mode 100644
index 000000000..6e7fd03d7
@@ -0,0 +1,10 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International Ltd
+
+# rte_cc_has_argument
+# Usage: MACHINE_CFLAGS += $(call rte_cc_has_argument, -mno-avx512f)
+# Return the argument if the argument is supported by the compiler.
+#
+define rte_cc_has_argument
+	$(shell $(CC) -E $(1) -xc /dev/null 1>/dev/null 2>/dev/null && echo $(1))
+endef
mk/target/generic/rte.vars.mk +11 -11
index dd149acc9..25a578ad7 100644
@@ -7,6 +7,17 @@
 # executive environment.
 #
 
+#
+# toolchain:
+#
+#   - define CC, LD, AR, AS, ...
+#   - define TOOLCHAIN_CFLAGS variable (overridden by cmdline value)
+#   - define TOOLCHAIN_LDFLAGS variable (overridden by cmdline value)
+#   - define TOOLCHAIN_ASFLAGS variable (overridden by cmdline value)
+#   - may override any previously defined variable
+#
+include $(RTE_SDK)/mk/toolchain/$(RTE_TOOLCHAIN)/rte.vars.mk
+
 #
 # machine:
 #
@@ -45,17 +56,6 @@ endif
 #
 include $(RTE_SDK)/mk/arch/$(RTE_ARCH)/rte.vars.mk
 
-#
-# toolchain:
-#
-#   - define CC, LD, AR, AS, ...
-#   - define TOOLCHAIN_CFLAGS variable (overridden by cmdline value)
-#   - define TOOLCHAIN_LDFLAGS variable (overridden by cmdline value)
-#   - define TOOLCHAIN_ASFLAGS variable (overridden by cmdline value)
-#   - may override any previously defined variable
-#
-include $(RTE_SDK)/mk/toolchain/$(RTE_TOOLCHAIN)/rte.vars.mk
-
 #
 # exec-env:
 #
Signature · 1 line
2.21.0
#49 lore raw

PATCH v7 2/4 meson: add infra to support machine specific flags

From: Pavan Nikhilesh <[email protected]>

Currently, RTE_* flags are set based on the implementer ID but there might be some micro arch specific differences from the same vendor eg. CACHE_LINESIZE. Add support to set micro arch specific flags.

---
 config/arm/meson.build | 37 ++++++++++++++++++++++++++++++++-----
 1 file changed, 32 insertions(+), 5 deletions(-)
config/arm/meson.build +32 -5
index 170a4981a..8de3f3e3a 100644
@@ -52,12 +52,10 @@ flags_generic = [
 	['RTE_USE_C11_MEM_MODEL', true],
 	['RTE_CACHE_LINE_SIZE', 128]]
 flags_cavium = [
-	['RTE_MACHINE', '"thunderx"'],
 	['RTE_CACHE_LINE_SIZE', 128],
 	['RTE_MAX_NUMA_NODES', 2],
 	['RTE_MAX_LCORE', 96],
-	['RTE_MAX_VFIO_GROUPS', 128],
-	['RTE_USE_C11_MEM_MODEL', false]]
+	['RTE_MAX_VFIO_GROUPS', 128]]
 flags_dpaa = [
 	['RTE_MACHINE', '"dpaa"'],
 	['RTE_USE_C11_MEM_MODEL', true],
@@ -71,6 +69,27 @@ flags_dpaa2 = [
 	['RTE_MAX_NUMA_NODES', 1],
 	['RTE_MAX_LCORE', 16],
 	['RTE_LIBRTE_DPAA2_USE_PHYS_IOVA', false]]
+flags_default_extra = []
+flags_thunderx_extra = [
+	['RTE_MACHINE', '"thunderx"'],
+	['RTE_USE_C11_MEM_MODEL', false]]
+
+machine_args_generic = [
+	['default', ['-march=armv8-a+crc+crypto']],
+	['native', ['-march=native']],
+	['0xd03', ['-mcpu=cortex-a53']],
+	['0xd04', ['-mcpu=cortex-a35']],
+	['0xd07', ['-mcpu=cortex-a57']],
+	['0xd08', ['-mcpu=cortex-a72']],
+	['0xd09', ['-mcpu=cortex-a73']],
+	['0xd0a', ['-mcpu=cortex-a75']]]
+
+machine_args_cavium = [
+	['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
+	['native', ['-march=native']],
+	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
+	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
+	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra]]
 
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page G7-5321)
 impl_generic = ['Generic armv8', flags_generic, machine_args_generic]
@@ -157,8 +176,16 @@ else
 	endif
 	foreach marg: machine[2]
 		if marg[0] == impl_pn
-			foreach f: marg[1]
-				machine_args += f
+			foreach flag: marg[1]
+				if cc.has_argument(flag)
+					machine_args += flag
+				endif
+			endforeach
+			# Apply any extra machine specific flags.
+			foreach flag: marg.get(2, flags_default_extra)
+				if flag.length() > 0
+					dpdk_conf.set(flag[0], flag[1])
+				endif
 			endforeach
 		endif
 	endforeach
Signature · 1 line
2.21.0
#50 lore raw

Re: [dpdk-dev] [PATCH v7 2/4] meson: add infra to support machine specific flags

Quoted 48 lines · On Apr 6, 2019, at 7:27 AM, [email protected] wrote:

From: Pavan Nikhilesh <[email protected]>

Currently, RTE_* flags are set based on the implementer ID but there might be some micro arch specific differences from the same vendor eg. CACHE_LINESIZE. Add support to set micro arch specific flags.

---
config/arm/meson.build | 37 ++++++++++++++++++++++++++++++++-----

1 file changed, 32 insertions(+), 5 deletions(-)
config/arm/meson.build +0 -0
index 170a4981a..8de3f3e3a 100644
@@ -52,12 +52,10 @@ flags_generic = [
 	['RTE_USE_C11_MEM_MODEL', true],
 	['RTE_CACHE_LINE_SIZE', 128]]

flags_cavium = [

config/arm/meson.build +0 -1
-	['RTE_MACHINE', '"thunderx"'],
	['RTE_CACHE_LINE_SIZE', 128],
	['RTE_MAX_NUMA_NODES', 2],
	['RTE_MAX_LCORE', 96],
config/arm/meson.build +1 -2
-	['RTE_MAX_VFIO_GROUPS', 128],
-	['RTE_USE_C11_MEM_MODEL', false]]
+	['RTE_MAX_VFIO_GROUPS', 128]]

flags_dpaa = [

	['RTE_MACHINE', '"dpaa"'],
	['RTE_USE_C11_MEM_MODEL', true],
config/arm/meson.build +11 -0
@@ -71,6 +69,27 @@ flags_dpaa2 = [
 	['RTE_MAX_NUMA_NODES', 1],
 	['RTE_MAX_LCORE', 16],
 	['RTE_LIBRTE_DPAA2_USE_PHYS_IOVA', false]]
+flags_default_extra = []
+flags_thunderx_extra = [
+	['RTE_MACHINE', '"thunderx"'],
+	['RTE_USE_C11_MEM_MODEL', false]]
+
+machine_args_generic = [
+	['default', ['-march=armv8-a+crc+crypto']],
+	['native', ['-march=native']],
+	['0xd03', ['-mcpu=cortex-a53']],
+	['0xd04', ['-mcpu=cortex-a35']],
+	['0xd07', ['-mcpu=cortex-a57']],

config/arm/meson.build flags_dpaa2 = [ · line 92

Quoted 10 lines
config/arm/meson.build +10 -0
+	['0xd08', ['-mcpu=cortex-a72']],
+	['0xd09', ['-mcpu=cortex-a73']],
+	['0xd0a', ['-mcpu=cortex-a75']]]
+
+machine_args_cavium = [
+	['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
+	['native', ['-march=native']],
+	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
+	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
+	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra]]

Looks like there's a mistake in rebasing it?
You should've removed machine_args_generic and machine_args_cavium in the beginning of this file.

Other than that, it looks good to me.

BTW, thanks for the patch. I raised this issue before and I was supposed to make the change but you have taken it.

Yongseok

Quoted 24 lines
config/arm/meson.build +0 -0
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page G7-5321)

impl_generic = ['Generic armv8', flags_generic, machine_args_generic]

config/arm/meson.build +10 -2
@@ -157,8 +176,16 @@ else
 	endif
 	foreach marg: machine[2]
 		if marg[0] == impl_pn
-			foreach f: marg[1]
-				machine_args += f
+			foreach flag: marg[1]
+				if cc.has_argument(flag)
+					machine_args += flag
+				endif
+			endforeach
+			# Apply any extra machine specific flags.
+			foreach flag: marg.get(2, flags_default_extra)
+				if flag.length() > 0
+					dpdk_conf.set(flag[0], flag[1])
+				endif
 			endforeach
 		endif
 	endforeach
Signature · 1 line
2.21.0
#51 lore raw

Re: [dpdk-dev] [PATCH v7 2/4] meson: add infra to support machine specific flags

Quoted 83 lines · Yongseok Koh
Quoted 49 lines · On Apr 6, 2019, at 7:27 AM, [email protected] wrote:

From: Pavan Nikhilesh <[email protected]>

Currently, RTE_* flags are set based on the implementer ID but there might be some micro arch specific differences from the same vendor eg. CACHE_LINESIZE. Add support to set micro arch specific flags.

---
config/arm/meson.build | 37 ++++++++++++++++++++++++++++++++-----

1 file changed, 32 insertions(+), 5 deletions(-)
config/arm/meson.build +0 -0
index 170a4981a..8de3f3e3a 100644
@@ -52,12 +52,10 @@ flags_generic = [
 	['RTE_USE_C11_MEM_MODEL', true],
 	['RTE_CACHE_LINE_SIZE', 128]]

flags_cavium = [

config/arm/meson.build +0 -1
-	['RTE_MACHINE', '"thunderx"'],
	['RTE_CACHE_LINE_SIZE', 128],
	['RTE_MAX_NUMA_NODES', 2],
	['RTE_MAX_LCORE', 96],
config/arm/meson.build +1 -2
-	['RTE_MAX_VFIO_GROUPS', 128],
-	['RTE_USE_C11_MEM_MODEL', false]]
+	['RTE_MAX_VFIO_GROUPS', 128]]

flags_dpaa = [

	['RTE_MACHINE', '"dpaa"'],
	['RTE_USE_C11_MEM_MODEL', true],
config/arm/meson.build +11 -0
@@ -71,6 +69,27 @@ flags_dpaa2 = [
 	['RTE_MAX_NUMA_NODES', 1],
 	['RTE_MAX_LCORE', 16],
 	['RTE_LIBRTE_DPAA2_USE_PHYS_IOVA', false]]
+flags_default_extra = []
+flags_thunderx_extra = [
+	['RTE_MACHINE', '"thunderx"'],
+	['RTE_USE_C11_MEM_MODEL', false]]
+
+machine_args_generic = [
+	['default', ['-march=armv8-a+crc+crypto']],
+	['native', ['-march=native']],
+	['0xd03', ['-mcpu=cortex-a53']],
+	['0xd04', ['-mcpu=cortex-a35']],
+	['0xd07', ['-mcpu=cortex-a57']],

config/arm/meson.build flags_dpaa2 = [ · line 92

Quoted 10 lines
config/arm/meson.build +10 -0
+	['0xd08', ['-mcpu=cortex-a72']],
+	['0xd09', ['-mcpu=cortex-a73']],
+	['0xd0a', ['-mcpu=cortex-a75']]]
+
+machine_args_cavium = [
+	['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
+	['native', ['-march=native']],
+	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
+	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
+	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra]]

Looks like there's a mistake in rebasing it?
You should've removed machine_args_generic and machine_args_cavium in the beginning of this file.

Other than that, it looks good to me.

BTW, thanks for the patch. I raised this issue before and I was supposed to make the change but you have taken it.

Yongseok

Quoted 10 lines
config/arm/meson.build +0 -0
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page G7-5321)

impl_generic = ['Generic armv8', flags_generic, machine_args_generic]

config/arm/meson.build +2 -2
@@ -157,8 +176,16 @@ else
 	endif
 	foreach marg: machine[2]
 		if marg[0] == impl_pn
-			foreach f: marg[1]
-				machine_args += f
+			foreach flag: marg[1]
+				if cc.has_argument(flag)
Quoted 8 lines

config/arm/meson.build else · line 188

Quoted 8 lines
config/arm/meson.build +8 -0
+					machine_args += flag
+				endif
+			endforeach
+			# Apply any extra machine specific flags.
+			foreach flag: marg.get(2, flags_default_extra)
+				if flag.length() > 0
+					dpdk_conf.set(flag[0], flag[1])
+				endif

And setting the extra flags doesn't work well with gcc < 7 because of the following,

# Primary part number based mcpu flags are supported
# for gcc versions > 7
if cc.version().version_compare(
'<7.0') or cmd_output.length() == 0
if not meson.is_cross_build() and arm_force_native_march == true
impl_pn = 'native'
else
impl_pn = 'default'
endif
endif

Thanks,
Yongseok

Quoted 5 lines
Quoted 5 lines
config/arm/meson.build +0 -0
 			endforeach
 		endif
 	endforeach
Signature · 1 line
2.21.0
#52 lore raw

Re: [dpdk-dev] [PATCH v7 2/4] meson: add infra to support machine specific flags

Quoted 12 lines

-----Original Message-----
From: Yongseok Koh <[email protected]>
Sent: Wednesday, April 10, 2019 7:45 AM
To: [email protected]; Pavan Nikhilesh Bhagavatula <[email protected]>
Cc: Thomas Monjalon <[email protected]>; dev <[email protected]>; Jerin Jacob Kollanukkaran <[email protected]>
Subject: [EXT] Re: [dpdk-dev] [PATCH v7 2/4] meson: add infra to support machine specific flags

Quoted 66 lines

External Email

----------------------------------------------------------------------

Quoted 62 lines · Yongseok Koh
Quoted 49 lines · On Apr 6, 2019, at 7:27 AM, [email protected] wrote:

From: Pavan Nikhilesh <[email protected]>

Currently, RTE_* flags are set based on the implementer ID but there might be some micro arch specific differences from the same vendor eg. CACHE_LINESIZE. Add support to set micro arch specific flags.

---
config/arm/meson.build | 37 ++++++++++++++++++++++++++++++++-----

1 file changed, 32 insertions(+), 5 deletions(-)
config/arm/meson.build index renamed from config/arm/meson.build +0 -0

170a4981a..8de3f3e3a 100644

config/arm/meson.build +0 -0
@@ -52,12 +52,10 @@ flags_generic = [
 	['RTE_USE_C11_MEM_MODEL', true],
 	['RTE_CACHE_LINE_SIZE', 128]]

flags_cavium = [

config/arm/meson.build +0 -1
-	['RTE_MACHINE', '"thunderx"'],
	['RTE_CACHE_LINE_SIZE', 128],
	['RTE_MAX_NUMA_NODES', 2],
	['RTE_MAX_LCORE', 96],
config/arm/meson.build +1 -2
-	['RTE_MAX_VFIO_GROUPS', 128],
-	['RTE_USE_C11_MEM_MODEL', false]]
+	['RTE_MAX_VFIO_GROUPS', 128]]

flags_dpaa = [

	['RTE_MACHINE', '"dpaa"'],
	['RTE_USE_C11_MEM_MODEL', true],
config/arm/meson.build +11 -0
@@ -71,6 +69,27 @@ flags_dpaa2 = [
 	['RTE_MAX_NUMA_NODES', 1],
 	['RTE_MAX_LCORE', 16],
 	['RTE_LIBRTE_DPAA2_USE_PHYS_IOVA', false]]
+flags_default_extra = []
+flags_thunderx_extra = [
+	['RTE_MACHINE', '"thunderx"'],
+	['RTE_USE_C11_MEM_MODEL', false]]
+
+machine_args_generic = [
+	['default', ['-march=armv8-a+crc+crypto']],
+	['native', ['-march=native']],
+	['0xd03', ['-mcpu=cortex-a53']],
+	['0xd04', ['-mcpu=cortex-a35']],
+	['0xd07', ['-mcpu=cortex-a57']],

config/arm/meson.build flags_dpaa2 = [ · line 92

Quoted 10 lines
config/arm/meson.build +10 -0
+	['0xd08', ['-mcpu=cortex-a72']],
+	['0xd09', ['-mcpu=cortex-a73']],
+	['0xd0a', ['-mcpu=cortex-a75']]]
+
+machine_args_cavium = [
+	['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
+	['native', ['-march=native']],
+	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
+	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
+	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra]]

Looks like there's a mistake in rebasing it?

Seems so will send out v8.

Quoted 23 lines
Quoted 23 lines

You should've removed machine_args_generic and machine_args_cavium in the beginning of this file.

Other than that, it looks good to me.

BTW, thanks for the patch. I raised this issue before and I was supposed to make the change but you have taken it.

Yongseok

Quoted 13 lines
config/arm/meson.build +0 -0
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page
 G7-5321) impl_generic = ['Generic armv8', flags_generic,

machine_args_generic] @@ -157,8 +176,16 @@ else

	endif
	foreach marg: machine[2]
		if marg[0] == impl_pn
config/arm/meson.build +5 -2
-			foreach f: marg[1]
-				machine_args += f
+			foreach flag: marg[1]
+				if cc.has_argument(flag)
+					machine_args += flag
+				endif
+			endforeach
Quoted 8 lines
Quoted 5 lines

config/arm/meson.build else · line 188

Quoted 5 lines
config/arm/meson.build +5 -0
+			# Apply any extra machine specific flags.
+			foreach flag: marg.get(2, flags_default_extra)
+				if flag.length() > 0
+					dpdk_conf.set(flag[0], flag[1])
+				endif

And setting the extra flags doesn't work well with gcc < 7 because of the following,

Extra flags aren't set in this case as the third variable in the list is missing when 'native' or 'default' is selected generic=

        ['default', ['-march=armv8-a+crc+crypto']],
        ['native', ['-march=native']]
cavium=
        ['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
        ['native', ['-march=native']]

And marg.get falls back to flags_default_extra = [].

Quoted 20 lines

# Primary part number based mcpu flags are supported
# for gcc versions > 7
if cc.version().version_compare(
'<7.0') or cmd_output.length() == 0
if not meson.is_cross_build() and arm_force_native_march == true
impl_pn = 'native'
else
impl_pn = 'default'
endif
endif

Thanks,
Yongseok

Quoted 5 lines
Quoted 5 lines

endforeach
endif
endforeach
--
2.21.0

#53 lore raw

PATCH v7 3/4 config: add thunderx2 machine config

In reply to [email protected]

From: Jerin Jacob <[email protected]>

Optimized configuration for Marvell thunderx2 SoC. Updated meson build to support Marvell thunderx2 SoC. Added meson cross compile target.

Product details are here:

https://www.marvell.com/server-processors/thunderx2-arm-processors/

---
 config/arm/arm64_thunderx2_linux_gcc          | 16 +++++++++
 config/arm/meson.build                        |  9 ++++-
 config/defconfig_arm64-thunderx2-linux-gcc    |  1 +
 config/defconfig_arm64-thunderx2-linuxapp-gcc | 11 ++++++
 mk/machine/thunderx2/rte.vars.mk              | 34 +++++++++++++++++++
 5 files changed, 70 insertions(+), 1 deletion(-)
 create mode 100644 config/arm/arm64_thunderx2_linux_gcc
 create mode 120000 config/defconfig_arm64-thunderx2-linux-gcc
 create mode 100644 config/defconfig_arm64-thunderx2-linuxapp-gcc
 create mode 100644 mk/machine/thunderx2/rte.vars.mk
config/arm/arm64_thunderx2_linux_gcc new file +16 -0
new file mode 100644
index 000000000..0dc275644
@@ -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 = '0x43'
+implementor_pn = '0xaf'
config/arm/meson.build +8 -1
index 8de3f3e3a..9282bbf33 100644
@@ -73,6 +73,12 @@ flags_default_extra = []
 flags_thunderx_extra = [
 	['RTE_MACHINE', '"thunderx"'],
 	['RTE_USE_C11_MEM_MODEL', false]]
+flags_thunderx2_extra = [
+	['RTE_MACHINE', '"thunderx2"'],
+	['RTE_CACHE_LINE_SIZE', 64],
+	['RTE_MAX_NUMA_NODES', 2],
+	['RTE_MAX_LCORE', 256],
+	['RTE_USE_C11_MEM_MODEL', true]]
 
 machine_args_generic = [
 	['default', ['-march=armv8-a+crc+crypto']],
@@ -89,7 +95,8 @@ machine_args_cavium = [
 	['native', ['-march=native']],
 	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
 	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
-	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra]]
+	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra],
+	['0xaf', ['-mcpu=thunderx2t99'], flags_thunderx2_extra]]
 
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page G7-5321)
 impl_generic = ['Generic armv8', flags_generic, machine_args_generic]
config/defconfig_arm64-thunderx2-linux-gcc new file +1 -0
new file mode 120000
index 000000000..b40a760b1
@@ -0,0 +1 @@
+defconfig_arm64-thunderx2-linuxapp-gcc
\ No newline at end of file
config/defconfig_arm64-thunderx2-linuxapp-gcc new file +11 -0
new file mode 100644
index 000000000..cc5c64ba0
@@ -0,0 +1,11 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International Ltd
+#
+
+#include "defconfig_arm64-armv8a-linux-gcc"
+
+CONFIG_RTE_MACHINE="thunderx2"
+
+CONFIG_RTE_CACHE_LINE_SIZE=64
+CONFIG_RTE_MAX_NUMA_NODES=2
+CONFIG_RTE_MAX_LCORE=256
mk/machine/thunderx2/rte.vars.mk new file +34 -0
new file mode 100644
index 000000000..b80dc8680
@@ -0,0 +1,34 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International 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=armv8.1-a+crc+crypto)
+MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=thunderx2t99)
Signature · 1 line
2.21.0
#54 lore raw

Re: [dpdk-dev] [PATCH v7 3/4] config: add thunderx2 machine config

06/04/2019 16:27, [email protected]:

config/arm/arm64_thunderx2_linux_gcc line 6

Quoted 9 lines
config/arm/arm64_thunderx2_linux_gcc new file +6 -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 = ''

Strangely, I still see the same issue:

In file included from /usr/include/inttypes.h:27,

                 from /usr/include/pcap/pcap-inttypes.h:114,
                 from /usr/include/pcap/pcap.h:74,
                 from /usr/include/pcap.h:43,
                 from ../drivers/net/pcap/rte_eth_pcap.c:19:

/usr/include/stdint.h:109: error: "__INT64_C" redefined [-Werror]

How may I help to debug it?

#55 lore raw

Re: [dpdk-dev] [PATCH v7 3/4] config: add thunderx2 machine config

08/04/2019 12:32, Thomas Monjalon:

Quoted 13 lines

06/04/2019 16:27, [email protected]:

config/arm/arm64_thunderx2_linux_gcc line 6

Quoted 9 lines
config/arm/arm64_thunderx2_linux_gcc new file +6 -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 = ''

Strangely, I still see the same issue:

Quoted 8 lines

In file included from /usr/include/inttypes.h:27,

                 from /usr/include/pcap/pcap-inttypes.h:114,
                 from /usr/include/pcap/pcap.h:74,
                 from /usr/include/pcap.h:43,
                 from ../drivers/net/pcap/rte_eth_pcap.c:19:

/usr/include/stdint.h:109: error: "__INT64_C" redefined [-Werror]

How may I help to debug it?

Thanks Bruce for telling me to remove the build directory. Now it works!

#56 lore raw

Re: [dpdk-dev] [PATCH v7 3/4] config: add thunderx2 machine config

Quoted 21 lines · Thomas Monjalon

08/04/2019 12:32, Thomas Monjalon:

Quoted 20 lines

06/04/2019 16:27, [email protected]:

config/arm/arm64_thunderx2_linux_gcc line 6

Quoted 9 lines
config/arm/arm64_thunderx2_linux_gcc new file +6 -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 = ''

Strangely, I still see the same issue:

In file included from /usr/include/inttypes.h:27,

                 from /usr/include/pcap/pcap-inttypes.h:114,
                 from /usr/include/pcap/pcap.h:74,
                 from /usr/include/pcap.h:43,
                 from ../drivers/net/pcap/rte_eth_pcap.c:19:

/usr/include/stdint.h:109: error: "__INT64_C" redefined [-Werror]

Quoted 5 lines
Quoted 1 line

How may I help to debug it?

Thanks Bruce for telling me to remove the build directory. Now it works!

For reference, official documentation on this point - unlike most things in meson, changing the cross-build file does require you to start from scratch with a new build folder.

https://mesonbuild.com/Cross-compilation.html#changing-cross-file-settings

/Bruce

#57 lore raw

PATCH v7 4/4 config: add octeontx2 machine config

In reply to [email protected]

From: Jerin Jacob <[email protected]>

Optimized configuration for Marvell octeontx2 SoC. Updated meson build to support Marvell octeontx2 SoC. Added meson cross build target for octeontx2.

---
 config/arm/arm64_octeontx2_linux_gcc          | 16 +++++++++
 config/arm/meson.build                        |  9 ++++-
 config/defconfig_arm64-octeontx2-linux-gcc    |  1 +
 config/defconfig_arm64-octeontx2-linuxapp-gcc | 18 ++++++++++
 mk/machine/octeontx2/rte.vars.mk              | 34 +++++++++++++++++++
 5 files changed, 77 insertions(+), 1 deletion(-)
 create mode 100644 config/arm/arm64_octeontx2_linux_gcc
 create mode 120000 config/defconfig_arm64-octeontx2-linux-gcc
 create mode 100644 config/defconfig_arm64-octeontx2-linuxapp-gcc
 create mode 100644 mk/machine/octeontx2/rte.vars.mk
config/arm/arm64_octeontx2_linux_gcc new file +16 -0
new file mode 100644
index 000000000..e2c0b8f72
@@ -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 = '0x43'
+implementor_pn = '0xb2'
config/arm/meson.build +8 -1
index 9282bbf33..0d76b2554 100644
@@ -79,6 +79,12 @@ flags_thunderx2_extra = [
 	['RTE_MAX_NUMA_NODES', 2],
 	['RTE_MAX_LCORE', 256],
 	['RTE_USE_C11_MEM_MODEL', true]]
+flags_octeontx2_extra = [
+	['RTE_MACHINE', '"octeontx2"'],
+	['RTE_MAX_NUMA_NODES', 1],
+	['RTE_MAX_LCORE', 24],
+	['RTE_EAL_IGB_UIO', false],
+	['RTE_USE_C11_MEM_MODEL', true]]
 
 machine_args_generic = [
 	['default', ['-march=armv8-a+crc+crypto']],
@@ -96,7 +102,8 @@ machine_args_cavium = [
 	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
 	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
 	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra],
-	['0xaf', ['-mcpu=thunderx2t99'], flags_thunderx2_extra]]
+	['0xaf', ['-mcpu=thunderx2t99'], flags_thunderx2_extra],
+	['0xb2', ['-mcpu=octeontx2'], flags_octeontx2_extra]]
 
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page G7-5321)
 impl_generic = ['Generic armv8', flags_generic, machine_args_generic]
config/defconfig_arm64-octeontx2-linux-gcc new file +1 -0
new file mode 120000
index 000000000..e25150531
@@ -0,0 +1 @@
+defconfig_arm64-octeontx2-linuxapp-gcc
\ No newline at end of file
config/defconfig_arm64-octeontx2-linuxapp-gcc new file +18 -0
new file mode 100644
index 000000000..9eae84538
@@ -0,0 +1,18 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International Ltd
+#
+
+#include "defconfig_arm64-armv8a-linux-gcc"
+
+CONFIG_RTE_MACHINE="octeontx2"
+
+CONFIG_RTE_CACHE_LINE_SIZE=128
+CONFIG_RTE_MAX_NUMA_NODES=1
+CONFIG_RTE_MAX_LCORE=24
+
+# Doesn't support NUMA
+CONFIG_RTE_EAL_NUMA_AWARE_HUGEPAGES=n
+CONFIG_RTE_LIBRTE_VHOST_NUMA=n
+
+# Recommend to use VFIO as co-processors needs SMMU/IOMMU
+CONFIG_RTE_EAL_IGB_UIO=n
mk/machine/octeontx2/rte.vars.mk new file +34 -0
new file mode 100644
index 000000000..cbec7f14d
@@ -0,0 +1,34 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International 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=-mcpu=armv8.2-a+crc+crypto+lse)
+MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=octeontx2)
Signature · 1 line
2.21.0
#58 lore raw

Re: [dpdk-dev] [PATCH v7 4/4] config: add octeontx2 machine config

06/04/2019 16:27, [email protected]:

Quoted 5 lines

From: Jerin Jacob <[email protected]>

Optimized configuration for Marvell octeontx2 SoC. Updated meson build to support Marvell octeontx2 SoC. Added meson cross build target for octeontx2.

I see this error with meson when compiling softnic for octeontx2:

-c ../drivers/net/softnic/rte_eth_softnic_action.c
{standard input}: Assembler messages:
{standard input}:16: Error: selected processor does not support crc32cx w3,w3,x0' {standard input}:38: Error: selected processor does not support crc32cx w1,w1,x3' {standard input}:44: Error: selected processor does not support crc32cx w0,w0,x4' {standard input}:69: Error: selected processor does not support crc32cx w2,w2,x3'
...

#59 lore raw

Re: [dpdk-dev] [PATCH v7 4/4] config: add octeontx2 machine config

Quoted 9 lines · Thomas Monjalon

06/04/2019 16:27, [email protected]:

Quoted 5 lines

From: Jerin Jacob <[email protected]>

Optimized configuration for Marvell octeontx2 SoC. Updated meson build to support Marvell octeontx2 SoC. Added meson cross build target for octeontx2.

I see this error with meson when compiling softnic for octeontx2:

Quoted 11 lines

-c ../drivers/net/softnic/rte_eth_softnic_action.c {standard input}: Assembler messages:
{standard input}:16: Error: selected processor does not support `crc32cx w3,w3,x0'
{standard input}:38: Error: selected processor does not support `crc32cx w1,w1,x3'
{standard input}:44: Error: selected processor does not support `crc32cx w0,w0,x4'
{standard input}:69: Error: selected processor does not support `crc32cx w2,w2,x3'
...

I have seen it too. Fix is here

Snippet from cover letter:

v7 Changes:

- Some compiler needs the following depended patch to compile with meson
http://patches.dpdk.org/patch/52367/

Quoted 1 line
#60 lore raw

PATCH v8 1/4 mk: introduce helper to check valid compiler argument

In reply to [email protected]

From: Jerin Jacob <[email protected]>

Introduce rte_cc_has_argument() Makefile helper to check a given argument is support by the compiler.

Example Usage:

include $(RTE_SDK)/mk/rte.helper.mk
MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=octeontx2)

This would allow adding -mcpu=octeontx2 in MACHINE_CFLAGS if it is only supported by the compiler. The use case for such scheme is to enable the mcpu optimization if the compiler supports else it needs to compile the source code without any errors.

This patch also moves inclusion of toolchain's rte.vars.mk to before the machine's rte.vars.mk inclusion to make correct CC available for the cross compile case.

---
Change history of this series:

v8 Changes:
- Remove redudant lists (rebase aritfacts). (Yongseok Koh)

v7 Changes:
- Updated cross compile config files align with
"build: improve pcap dependency handling" changeset to fix build issue with meson
- Some compiler needs the following depended patch to compile with meson
http://patches.dpdk.org/patch/52367/

v6 Changes:
- Rework to change the config files to sync with "mk: use linux and freebsd in config names"
- Fix the following error with latest gcc by fixing the mcpu type
cc1: error: switch -mcpu=armv8.2-a conflicts with -march=armv8-a switch

v5 Changes:
- Fix incorrect meson flag parsing(Phil Yang)
- Squash meson cross build patch(5/5) into configuration update patches for thunderx2(3/5) and octeontx2(4/5)(Thomas)
- Changed octeontx2's march as armv8-a and added the extension required instead of armv8-2a(Phil Yang)
- Improved rte_cc_has_argument() implementaion by removing the temp file(Thomas)

v4 Changes:
- Fix incorrect signoff marrvell -> marvell.

v3 Changes:
- Squash meson build support into config support for thunderx2/octeontx2.

v2 Changes:
- Add meson build support.

---
 mk/rte.helper.mk              | 10 ++++++++++
 mk/target/generic/rte.vars.mk | 22 +++++++++++-----------
 2 files changed, 21 insertions(+), 11 deletions(-)
 create mode 100644 mk/rte.helper.mk
mk/rte.helper.mk new file +10 -0
new file mode 100644
index 000000000..6e7fd03d7
@@ -0,0 +1,10 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International Ltd
+
+# rte_cc_has_argument
+# Usage: MACHINE_CFLAGS += $(call rte_cc_has_argument, -mno-avx512f)
+# Return the argument if the argument is supported by the compiler.
+#
+define rte_cc_has_argument
+	$(shell $(CC) -E $(1) -xc /dev/null 1>/dev/null 2>/dev/null && echo $(1))
+endef
mk/target/generic/rte.vars.mk +11 -11
index dd149acc9..25a578ad7 100644
@@ -7,6 +7,17 @@
 # executive environment.
 #
 
+#
+# toolchain:
+#
+#   - define CC, LD, AR, AS, ...
+#   - define TOOLCHAIN_CFLAGS variable (overridden by cmdline value)
+#   - define TOOLCHAIN_LDFLAGS variable (overridden by cmdline value)
+#   - define TOOLCHAIN_ASFLAGS variable (overridden by cmdline value)
+#   - may override any previously defined variable
+#
+include $(RTE_SDK)/mk/toolchain/$(RTE_TOOLCHAIN)/rte.vars.mk
+
 #
 # machine:
 #
@@ -45,17 +56,6 @@ endif
 #
 include $(RTE_SDK)/mk/arch/$(RTE_ARCH)/rte.vars.mk
 
-#
-# toolchain:
-#
-#   - define CC, LD, AR, AS, ...
-#   - define TOOLCHAIN_CFLAGS variable (overridden by cmdline value)
-#   - define TOOLCHAIN_LDFLAGS variable (overridden by cmdline value)
-#   - define TOOLCHAIN_ASFLAGS variable (overridden by cmdline value)
-#   - may override any previously defined variable
-#
-include $(RTE_SDK)/mk/toolchain/$(RTE_TOOLCHAIN)/rte.vars.mk
-
 #
 # exec-env:
 #
Signature · 1 line
2.21.0
#61 lore raw

PATCH v8 2/4 meson: add infra to support machine specific flags

From: Pavan Nikhilesh <[email protected]>

Currently, RTE_* flags are set based on the implementer ID but there might be some micro arch specific differences from the same vendor eg. CACHE_LINESIZE. Add support to set micro arch specific flags.

---
 config/arm/meson.build | 56 ++++++++++++++++++++++++------------------
 1 file changed, 32 insertions(+), 24 deletions(-)
config/arm/meson.build +32 -24
index 170a4981a..24bce2b39 100644
@@ -7,25 +7,6 @@ march_opt = '-march=@0@'.format(machine)
 
 arm_force_native_march = false
 
-machine_args_generic = [
-	['default', ['-march=armv8-a+crc+crypto']],
-	['native', ['-march=native']],
-	['0xd03', ['-mcpu=cortex-a53']],
-	['0xd04', ['-mcpu=cortex-a35']],
-	['0xd05', ['-mcpu=cortex-a55']],
-	['0xd07', ['-mcpu=cortex-a57']],
-	['0xd08', ['-mcpu=cortex-a72']],
-	['0xd09', ['-mcpu=cortex-a73']],
-	['0xd0a', ['-mcpu=cortex-a75']],
-	['0xd0b', ['-mcpu=cortex-a76']],
-]
-machine_args_cavium = [
-	['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
-	['native', ['-march=native']],
-	['0xa1', ['-mcpu=thunderxt88']],
-	['0xa2', ['-mcpu=thunderxt81']],
-	['0xa3', ['-mcpu=thunderxt83']]]
-
 flags_common_default = [
 	# Accelarate rte_memcpy. Be sure to run unit test (memcpy_perf_autotest)
 	# to determine the best threshold in code. Refer to notes in source file
@@ -52,12 +33,10 @@ flags_generic = [
 	['RTE_USE_C11_MEM_MODEL', true],
 	['RTE_CACHE_LINE_SIZE', 128]]
 flags_cavium = [
-	['RTE_MACHINE', '"thunderx"'],
 	['RTE_CACHE_LINE_SIZE', 128],
 	['RTE_MAX_NUMA_NODES', 2],
 	['RTE_MAX_LCORE', 96],
-	['RTE_MAX_VFIO_GROUPS', 128],
-	['RTE_USE_C11_MEM_MODEL', false]]
+	['RTE_MAX_VFIO_GROUPS', 128]]
 flags_dpaa = [
 	['RTE_MACHINE', '"dpaa"'],
 	['RTE_USE_C11_MEM_MODEL', true],
@@ -71,6 +50,27 @@ flags_dpaa2 = [
 	['RTE_MAX_NUMA_NODES', 1],
 	['RTE_MAX_LCORE', 16],
 	['RTE_LIBRTE_DPAA2_USE_PHYS_IOVA', false]]
+flags_default_extra = []
+flags_thunderx_extra = [
+	['RTE_MACHINE', '"thunderx"'],
+	['RTE_USE_C11_MEM_MODEL', false]]
+
+machine_args_generic = [
+	['default', ['-march=armv8-a+crc+crypto']],
+	['native', ['-march=native']],
+	['0xd03', ['-mcpu=cortex-a53']],
+	['0xd04', ['-mcpu=cortex-a35']],
+	['0xd07', ['-mcpu=cortex-a57']],
+	['0xd08', ['-mcpu=cortex-a72']],
+	['0xd09', ['-mcpu=cortex-a73']],
+	['0xd0a', ['-mcpu=cortex-a75']]]
+
+machine_args_cavium = [
+	['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
+	['native', ['-march=native']],
+	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
+	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
+	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra]]
 
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page G7-5321)
 impl_generic = ['Generic armv8', flags_generic, machine_args_generic]
@@ -157,8 +157,16 @@ else
 	endif
 	foreach marg: machine[2]
 		if marg[0] == impl_pn
-			foreach f: marg[1]
-				machine_args += f
+			foreach flag: marg[1]
+				if cc.has_argument(flag)
+					machine_args += flag
+				endif
+			endforeach
+			# Apply any extra machine specific flags.
+			foreach flag: marg.get(2, flags_default_extra)
+				if flag.length() > 0
+					dpdk_conf.set(flag[0], flag[1])
+				endif
 			endforeach
 		endif
 	endforeach
Signature · 1 line
2.21.0
#62 lore raw

Re: [dpdk-dev] [PATCH v8 2/4] meson: add infra to support machine specific flags

Quoted 93 lines · On Apr 10, 2019, at 9:13 AM, [email protected] wrote:

From: Pavan Nikhilesh <[email protected]>

Currently, RTE_* flags are set based on the implementer ID but there might be some micro arch specific differences from the same vendor eg. CACHE_LINESIZE. Add support to set micro arch specific flags.

---
config/arm/meson.build | 56 ++++++++++++++++++++++++------------------

1 file changed, 32 insertions(+), 24 deletions(-)
config/arm/meson.build +0 -0
index 170a4981a..24bce2b39 100644
@@ -7,25 +7,6 @@ march_opt = '-march=@0@'.format(machine)
 

arm_force_native_march = false

config/arm/meson.build +0 -19
-machine_args_generic = [
-	['default', ['-march=armv8-a+crc+crypto']],
-	['native', ['-march=native']],
-	['0xd03', ['-mcpu=cortex-a53']],
-	['0xd04', ['-mcpu=cortex-a35']],
-	['0xd05', ['-mcpu=cortex-a55']],
-	['0xd07', ['-mcpu=cortex-a57']],
-	['0xd08', ['-mcpu=cortex-a72']],
-	['0xd09', ['-mcpu=cortex-a73']],
-	['0xd0a', ['-mcpu=cortex-a75']],
-	['0xd0b', ['-mcpu=cortex-a76']],
-]
-machine_args_cavium = [
-	['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
-	['native', ['-march=native']],
-	['0xa1', ['-mcpu=thunderxt88']],
-	['0xa2', ['-mcpu=thunderxt81']],
-	['0xa3', ['-mcpu=thunderxt83']]]
-

flags_common_default = [
# Accelarate rte_memcpy. Be sure to run unit test (memcpy_perf_autotest) # to determine the best threshold in code. Refer to notes in source file

config/arm/meson.build +0 -0
@@ -52,12 +33,10 @@ flags_generic = [
 	['RTE_USE_C11_MEM_MODEL', true],
 	['RTE_CACHE_LINE_SIZE', 128]]

flags_cavium = [

config/arm/meson.build +0 -1
-	['RTE_MACHINE', '"thunderx"'],
	['RTE_CACHE_LINE_SIZE', 128],
	['RTE_MAX_NUMA_NODES', 2],
	['RTE_MAX_LCORE', 96],
config/arm/meson.build +1 -2
-	['RTE_MAX_VFIO_GROUPS', 128],
-	['RTE_USE_C11_MEM_MODEL', false]]
+	['RTE_MAX_VFIO_GROUPS', 128]]

flags_dpaa = [

	['RTE_MACHINE', '"dpaa"'],
	['RTE_USE_C11_MEM_MODEL', true],
config/arm/meson.build +21 -0
@@ -71,6 +50,27 @@ flags_dpaa2 = [
 	['RTE_MAX_NUMA_NODES', 1],
 	['RTE_MAX_LCORE', 16],
 	['RTE_LIBRTE_DPAA2_USE_PHYS_IOVA', false]]
+flags_default_extra = []
+flags_thunderx_extra = [
+	['RTE_MACHINE', '"thunderx"'],
+	['RTE_USE_C11_MEM_MODEL', false]]
+
+machine_args_generic = [
+	['default', ['-march=armv8-a+crc+crypto']],
+	['native', ['-march=native']],
+	['0xd03', ['-mcpu=cortex-a53']],
+	['0xd04', ['-mcpu=cortex-a35']],
+	['0xd07', ['-mcpu=cortex-a57']],
+	['0xd08', ['-mcpu=cortex-a72']],
+	['0xd09', ['-mcpu=cortex-a73']],
+	['0xd0a', ['-mcpu=cortex-a75']]]
+
+machine_args_cavium = [
+	['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
+	['native', ['-march=native']],
+	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
+	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
+	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra]]
 
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page G7-5321)

impl_generic = ['Generic armv8', flags_generic, machine_args_generic]

config/arm/meson.build +0 -2
@@ -157,8 +157,16 @@ else
 	endif
 	foreach marg: machine[2]
 		if marg[0] == impl_pn
-			foreach f: marg[1]
-				machine_args += f

config/arm/meson.build else · line 169

Quoted 10 lines
config/arm/meson.build +10 -0
+			foreach flag: marg[1]
+				if cc.has_argument(flag)
+					machine_args += flag
+				endif
+			endforeach
+			# Apply any extra machine specific flags.
+			foreach flag: marg.get(2, flags_default_extra)
+				if flag.length() > 0
+					dpdk_conf.set(flag[0], flag[1])
+				endif

Let me continue the discussion from v7 here.
Seems I wan't clear enough.

Let me take an example. If the host is thunderx2 (0xaf) and compiler is older than v7, flags_thunderx2_extra isn't set. This means, for example, RTE_CACHE_LINE_SIZE will still be 128. Is that what you want? RTE_CACHE_LINE_SIZE has nothing to do with compiler support and you might want to set it regardless of gcc version. You could skip setting -mcpu with setting the extra flags.

Thoughts?

Thanks,
Yongseok

#63 lore raw

Re: [dpdk-dev] [PATCH v8 2/4] meson: add infra to support machine specific flags

Hi Yongseok,

Quoted 120 lines

-----Original Message-----
From: Yongseok Koh <[email protected]>
Sent: Wednesday, April 10, 2019 11:08 PM
To: Pavan Nikhilesh Bhagavatula <[email protected]>
Cc: Thomas Monjalon <[email protected]>; dev <[email protected]>; Jerin Jacob Kollanukkaran <[email protected]>; [email protected]
Subject: [EXT] Re: [dpdk-dev] [PATCH v8 2/4] meson: add infra to support machine specific flags

External Email

----------------------------------------------------------------------

Quoted 41 lines · On Apr 10, 2019, at 9:13 AM, [email protected] wrote:

From: Pavan Nikhilesh <[email protected]>

Currently, RTE_* flags are set based on the implementer ID but there might be some micro arch specific differences from the same vendor eg. CACHE_LINESIZE. Add support to set micro arch specific flags.

---
config/arm/meson.build | 56 ++++++++++++++++++++++++------------------

1 file changed, 32 insertions(+), 24 deletions(-)
config/arm/meson.build index renamed from config/arm/meson.build +0 -0

170a4981a..24bce2b39 100644

config/arm/meson.build +0 -0
@@ -7,25 +7,6 @@ march_opt = '-march=@0@'.format(machine)
 

arm_force_native_march = false

config/arm/meson.build +0 -19
-machine_args_generic = [
-	['default', ['-march=armv8-a+crc+crypto']],
-	['native', ['-march=native']],
-	['0xd03', ['-mcpu=cortex-a53']],
-	['0xd04', ['-mcpu=cortex-a35']],
-	['0xd05', ['-mcpu=cortex-a55']],
-	['0xd07', ['-mcpu=cortex-a57']],
-	['0xd08', ['-mcpu=cortex-a72']],
-	['0xd09', ['-mcpu=cortex-a73']],
-	['0xd0a', ['-mcpu=cortex-a75']],
-	['0xd0b', ['-mcpu=cortex-a76']],
-]
-machine_args_cavium = [
-	['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
-	['native', ['-march=native']],
-	['0xa1', ['-mcpu=thunderxt88']],
-	['0xa2', ['-mcpu=thunderxt81']],
-	['0xa3', ['-mcpu=thunderxt83']]]
-
Quoted 2 lines

flags_common_default = [
# Accelarate rte_memcpy. Be sure to run unit test

(memcpy_perf_autotest)

Quoted 49 lines
	# to determine the best threshold in code. Refer to notes in source
file @@ -52,12 +33,10 @@ flags_generic = [
	['RTE_USE_C11_MEM_MODEL', true],
	['RTE_CACHE_LINE_SIZE', 128]]
flags_cavium = [
config/arm/meson.build +0 -1
-	['RTE_MACHINE', '"thunderx"'],
	['RTE_CACHE_LINE_SIZE', 128],
	['RTE_MAX_NUMA_NODES', 2],
	['RTE_MAX_LCORE', 96],
config/arm/meson.build +1 -2
-	['RTE_MAX_VFIO_GROUPS', 128],
-	['RTE_USE_C11_MEM_MODEL', false]]
+	['RTE_MAX_VFIO_GROUPS', 128]]

flags_dpaa = [

	['RTE_MACHINE', '"dpaa"'],
	['RTE_USE_C11_MEM_MODEL', true],
config/arm/meson.build +21 -0
@@ -71,6 +50,27 @@ flags_dpaa2 = [
 	['RTE_MAX_NUMA_NODES', 1],
 	['RTE_MAX_LCORE', 16],
 	['RTE_LIBRTE_DPAA2_USE_PHYS_IOVA', false]]
+flags_default_extra = []
+flags_thunderx_extra = [
+	['RTE_MACHINE', '"thunderx"'],
+	['RTE_USE_C11_MEM_MODEL', false]]
+
+machine_args_generic = [
+	['default', ['-march=armv8-a+crc+crypto']],
+	['native', ['-march=native']],
+	['0xd03', ['-mcpu=cortex-a53']],
+	['0xd04', ['-mcpu=cortex-a35']],
+	['0xd07', ['-mcpu=cortex-a57']],
+	['0xd08', ['-mcpu=cortex-a72']],
+	['0xd09', ['-mcpu=cortex-a73']],
+	['0xd0a', ['-mcpu=cortex-a75']]]
+
+machine_args_cavium = [
+	['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
+	['native', ['-march=native']],
+	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
+	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
+	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra]]
 
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page
 G7-5321) impl_generic = ['Generic armv8', flags_generic,

machine_args_generic] @@ -157,8 +157,16 @@ else

	endif
	foreach marg: machine[2]
		if marg[0] == impl_pn
config/arm/meson.build +0 -2
-			foreach f: marg[1]
-				machine_args += f

config/arm/meson.build else · line 169

Quoted 10 lines
config/arm/meson.build +10 -0
+			foreach flag: marg[1]
+				if cc.has_argument(flag)
+					machine_args += flag
+				endif
+			endforeach
+			# Apply any extra machine specific flags.
+			foreach flag: marg.get(2, flags_default_extra)
+				if flag.length() > 0
+					dpdk_conf.set(flag[0], flag[1])
+				endif

Let me continue the discussion from v7 here.
Seems I wan't clear enough.

Quoted 7 lines

Let me take an example. If the host is thunderx2 (0xaf) and compiler is older than v7, flags_thunderx2_extra isn't set. This means, for example, RTE_CACHE_LINE_SIZE will still be 128. Is that what you want? RTE_CACHE_LINE_SIZE has nothing to do with compiler support and you might want to set it regardless of gcc version. You could skip setting -mcpu with setting the extra flags.

Thanks for the detailed explanation.
I think since we have the check to skip mcpu flag when cc doesn't support it (cc.has_argument(flag)) It will be safe to remove
`
# Primary part number based mcpu flags are supported
# for gcc versions > 7
if cc.version().version_compare(
'<7.0') or cmd_output.length() == 0
if not meson.is_cross_build() and arm_force_native_march == true
impl_pn = 'native'
else
impl_pn = 'default'
endif
endif
`

The command output check can also be removed as it is handled when calling the command script itself.

Thoughts?

PS. I think the safest way to set CACHELINE_SIZE is to read the cache type register[1] but sadly only few latest kernels have the support through sysfs (/sys/devices/system/cpu/cpu0/cache/index0/coherency_line_size)

Quoted 7 lines

Thoughts?

Thanks,
Yongseok

Regards,
Pavan.

#64 lore raw

Re: [dpdk-dev] [PATCH v8 2/4] meson: add infra to support machine specific flags

Quoted 133 lines · Pavan Nikhilesh Bhagavatula

Hi Yongseok,

Quoted 120 lines

-----Original Message-----
From: Yongseok Koh <[email protected]>
Sent: Wednesday, April 10, 2019 11:08 PM
To: Pavan Nikhilesh Bhagavatula <[email protected]>
Cc: Thomas Monjalon <[email protected]>; dev <[email protected]>; Jerin Jacob Kollanukkaran <[email protected]>; [email protected]
Subject: [EXT] Re: [dpdk-dev] [PATCH v8 2/4] meson: add infra to support machine specific flags

External Email

----------------------------------------------------------------------

Quoted 41 lines · On Apr 10, 2019, at 9:13 AM, [email protected] wrote:

From: Pavan Nikhilesh <[email protected]>

Currently, RTE_* flags are set based on the implementer ID but there might be some micro arch specific differences from the same vendor eg. CACHE_LINESIZE. Add support to set micro arch specific flags.

---
config/arm/meson.build | 56 ++++++++++++++++++++++++------------------

1 file changed, 32 insertions(+), 24 deletions(-)
config/arm/meson.build index renamed from config/arm/meson.build +0 -0

170a4981a..24bce2b39 100644

config/arm/meson.build +0 -0
@@ -7,25 +7,6 @@ march_opt = '-march=@0@'.format(machine)
 

arm_force_native_march = false

config/arm/meson.build +0 -19
-machine_args_generic = [
-	['default', ['-march=armv8-a+crc+crypto']],
-	['native', ['-march=native']],
-	['0xd03', ['-mcpu=cortex-a53']],
-	['0xd04', ['-mcpu=cortex-a35']],
-	['0xd05', ['-mcpu=cortex-a55']],
-	['0xd07', ['-mcpu=cortex-a57']],
-	['0xd08', ['-mcpu=cortex-a72']],
-	['0xd09', ['-mcpu=cortex-a73']],
-	['0xd0a', ['-mcpu=cortex-a75']],
-	['0xd0b', ['-mcpu=cortex-a76']],
-]
-machine_args_cavium = [
-	['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
-	['native', ['-march=native']],
-	['0xa1', ['-mcpu=thunderxt88']],
-	['0xa2', ['-mcpu=thunderxt81']],
-	['0xa3', ['-mcpu=thunderxt83']]]
-
Quoted 2 lines

flags_common_default = [
# Accelarate rte_memcpy. Be sure to run unit test

(memcpy_perf_autotest)

Quoted 49 lines
	# to determine the best threshold in code. Refer to notes in source
file @@ -52,12 +33,10 @@ flags_generic = [
	['RTE_USE_C11_MEM_MODEL', true],
	['RTE_CACHE_LINE_SIZE', 128]]
flags_cavium = [
config/arm/meson.build +0 -1
-	['RTE_MACHINE', '"thunderx"'],
	['RTE_CACHE_LINE_SIZE', 128],
	['RTE_MAX_NUMA_NODES', 2],
	['RTE_MAX_LCORE', 96],
config/arm/meson.build +1 -2
-	['RTE_MAX_VFIO_GROUPS', 128],
-	['RTE_USE_C11_MEM_MODEL', false]]
+	['RTE_MAX_VFIO_GROUPS', 128]]

flags_dpaa = [

	['RTE_MACHINE', '"dpaa"'],
	['RTE_USE_C11_MEM_MODEL', true],
config/arm/meson.build +21 -0
@@ -71,6 +50,27 @@ flags_dpaa2 = [
 	['RTE_MAX_NUMA_NODES', 1],
 	['RTE_MAX_LCORE', 16],
 	['RTE_LIBRTE_DPAA2_USE_PHYS_IOVA', false]]
+flags_default_extra = []
+flags_thunderx_extra = [
+	['RTE_MACHINE', '"thunderx"'],
+	['RTE_USE_C11_MEM_MODEL', false]]
+
+machine_args_generic = [
+	['default', ['-march=armv8-a+crc+crypto']],
+	['native', ['-march=native']],
+	['0xd03', ['-mcpu=cortex-a53']],
+	['0xd04', ['-mcpu=cortex-a35']],
+	['0xd07', ['-mcpu=cortex-a57']],
+	['0xd08', ['-mcpu=cortex-a72']],
+	['0xd09', ['-mcpu=cortex-a73']],
+	['0xd0a', ['-mcpu=cortex-a75']]]
+
+machine_args_cavium = [
+	['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
+	['native', ['-march=native']],
+	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
+	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
+	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra]]
 
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page
 G7-5321) impl_generic = ['Generic armv8', flags_generic,

machine_args_generic] @@ -157,8 +157,16 @@ else

	endif
	foreach marg: machine[2]
		if marg[0] == impl_pn
config/arm/meson.build +0 -2
-			foreach f: marg[1]
-				machine_args += f

config/arm/meson.build else · line 169

Quoted 10 lines
config/arm/meson.build +10 -0
+			foreach flag: marg[1]
+				if cc.has_argument(flag)
+					machine_args += flag
+				endif
+			endforeach
+			# Apply any extra machine specific flags.
+			foreach flag: marg.get(2, flags_default_extra)
+				if flag.length() > 0
+					dpdk_conf.set(flag[0], flag[1])
+				endif

Let me continue the discussion from v7 here.
Seems I wan't clear enough.

Quoted 7 lines

Let me take an example. If the host is thunderx2 (0xaf) and compiler is older than v7, flags_thunderx2_extra isn't set. This means, for example, RTE_CACHE_LINE_SIZE will still be 128. Is that what you want? RTE_CACHE_LINE_SIZE has nothing to do with compiler support and you might want to set it regardless of gcc version. You could skip setting -mcpu with setting the extra flags.

Quoted 15 lines

Thanks for the detailed explanation.
I think since we have the check to skip mcpu flag when cc doesn't support it (cc.has_argument(flag)) It will be safe to remove
`
# Primary part number based mcpu flags are supported
# for gcc versions > 7
if cc.version().version_compare(
'<7.0') or cmd_output.length() == 0
if not meson.is_cross_build() and arm_force_native_march == true
impl_pn = 'native'
else
impl_pn = 'default'
endif
endif
`

+1

Quoted 2 lines

The command output check can also be removed as it is handled when calling the command script itself.

+1

Quoted 5 lines

Thoughts?

PS. I think the safest way to set CACHELINE_SIZE is to read the cache type register[1] but sadly only few latest kernels have the support through sysfs (/sys/devices/system/cpu/cpu0/cache/index0/coherency_line_size)

+1

In summary, +3. LoL

I'll also submit a patch to change the default cacheline size of cortex-a72 with the new flags_*_extra[]

thanks,
Yongseok

#65 lore raw

Re: [dpdk-dev] [PATCH v8 2/4] meson: add infra to support machine specific flags

Quoted 152 lines · Yongseok Koh
Quoted 134 lines · Pavan Nikhilesh Bhagavatula

Hi Yongseok,

Quoted 120 lines

-----Original Message-----
From: Yongseok Koh <[email protected]>
Sent: Wednesday, April 10, 2019 11:08 PM
To: Pavan Nikhilesh Bhagavatula <[email protected]>
Cc: Thomas Monjalon <[email protected]>; dev <[email protected]>; Jerin Jacob Kollanukkaran <[email protected]>; [email protected]
Subject: [EXT] Re: [dpdk-dev] [PATCH v8 2/4] meson: add infra to support machine specific flags

External Email

----------------------------------------------------------------------

Quoted 41 lines · On Apr 10, 2019, at 9:13 AM, [email protected] wrote:

From: Pavan Nikhilesh <[email protected]>

Currently, RTE_* flags are set based on the implementer ID but there might be some micro arch specific differences from the same vendor eg. CACHE_LINESIZE. Add support to set micro arch specific flags.

---
config/arm/meson.build | 56 ++++++++++++++++++++++++------------------

1 file changed, 32 insertions(+), 24 deletions(-)
config/arm/meson.build index renamed from config/arm/meson.build +0 -0

170a4981a..24bce2b39 100644

config/arm/meson.build +0 -0
@@ -7,25 +7,6 @@ march_opt = '-march=@0@'.format(machine)
 

arm_force_native_march = false

config/arm/meson.build +0 -19
-machine_args_generic = [
-	['default', ['-march=armv8-a+crc+crypto']],
-	['native', ['-march=native']],
-	['0xd03', ['-mcpu=cortex-a53']],
-	['0xd04', ['-mcpu=cortex-a35']],
-	['0xd05', ['-mcpu=cortex-a55']],
-	['0xd07', ['-mcpu=cortex-a57']],
-	['0xd08', ['-mcpu=cortex-a72']],
-	['0xd09', ['-mcpu=cortex-a73']],
-	['0xd0a', ['-mcpu=cortex-a75']],
-	['0xd0b', ['-mcpu=cortex-a76']],
-]
-machine_args_cavium = [
-	['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
-	['native', ['-march=native']],
-	['0xa1', ['-mcpu=thunderxt88']],
-	['0xa2', ['-mcpu=thunderxt81']],
-	['0xa3', ['-mcpu=thunderxt83']]]
-
Quoted 2 lines

flags_common_default = [
# Accelarate rte_memcpy. Be sure to run unit test

(memcpy_perf_autotest)

Quoted 49 lines
	# to determine the best threshold in code. Refer to notes in source
file @@ -52,12 +33,10 @@ flags_generic = [
	['RTE_USE_C11_MEM_MODEL', true],
	['RTE_CACHE_LINE_SIZE', 128]]
flags_cavium = [
config/arm/meson.build +0 -1
-	['RTE_MACHINE', '"thunderx"'],
	['RTE_CACHE_LINE_SIZE', 128],
	['RTE_MAX_NUMA_NODES', 2],
	['RTE_MAX_LCORE', 96],
config/arm/meson.build +1 -2
-	['RTE_MAX_VFIO_GROUPS', 128],
-	['RTE_USE_C11_MEM_MODEL', false]]
+	['RTE_MAX_VFIO_GROUPS', 128]]

flags_dpaa = [

	['RTE_MACHINE', '"dpaa"'],
	['RTE_USE_C11_MEM_MODEL', true],
config/arm/meson.build +21 -0
@@ -71,6 +50,27 @@ flags_dpaa2 = [
 	['RTE_MAX_NUMA_NODES', 1],
 	['RTE_MAX_LCORE', 16],
 	['RTE_LIBRTE_DPAA2_USE_PHYS_IOVA', false]]
+flags_default_extra = []
+flags_thunderx_extra = [
+	['RTE_MACHINE', '"thunderx"'],
+	['RTE_USE_C11_MEM_MODEL', false]]
+
+machine_args_generic = [
+	['default', ['-march=armv8-a+crc+crypto']],
+	['native', ['-march=native']],
+	['0xd03', ['-mcpu=cortex-a53']],
+	['0xd04', ['-mcpu=cortex-a35']],
+	['0xd07', ['-mcpu=cortex-a57']],
+	['0xd08', ['-mcpu=cortex-a72']],
+	['0xd09', ['-mcpu=cortex-a73']],
+	['0xd0a', ['-mcpu=cortex-a75']]]
+
+machine_args_cavium = [
+	['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
+	['native', ['-march=native']],
+	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
+	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
+	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra]]
 
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page
 G7-5321) impl_generic = ['Generic armv8', flags_generic,

machine_args_generic] @@ -157,8 +157,16 @@ else

	endif
	foreach marg: machine[2]
		if marg[0] == impl_pn
config/arm/meson.build +0 -2
-			foreach f: marg[1]
-				machine_args += f

config/arm/meson.build else · line 169

Quoted 10 lines
config/arm/meson.build +10 -0
+			foreach flag: marg[1]
+				if cc.has_argument(flag)
+					machine_args += flag
+				endif
+			endforeach
+			# Apply any extra machine specific flags.
+			foreach flag: marg.get(2, flags_default_extra)
+				if flag.length() > 0
+					dpdk_conf.set(flag[0], flag[1])
+				endif

Let me continue the discussion from v7 here.
Seems I wan't clear enough.

Quoted 7 lines

Let me take an example. If the host is thunderx2 (0xaf) and compiler is older than v7, flags_thunderx2_extra isn't set. This means, for example, RTE_CACHE_LINE_SIZE will still be 128. Is that what you want? RTE_CACHE_LINE_SIZE has nothing to do with compiler support and you might want to set it regardless of gcc version. You could skip setting -mcpu with setting the extra flags.

Quoted 15 lines

Thanks for the detailed explanation.
I think since we have the check to skip mcpu flag when cc doesn't support it (cc.has_argument(flag)) It will be safe to remove
`
# Primary part number based mcpu flags are supported
# for gcc versions > 7
if cc.version().version_compare(
'<7.0') or cmd_output.length() == 0
if not meson.is_cross_build() and arm_force_native_march == true
impl_pn = 'native'
else
impl_pn = 'default'
endif
endif
`

Quoted 1 line

+1

I've tested it but still have an issue with old gcc.
Even if -mcpu isn't set due to cc.has_argument(), -march isn't set either. So, it spews error due to lack of CRC feature.
-march should have '+crc'. The error I got was:

Quoted 4 lines

ninja: Entering directory `build'
[942/1452] Compiling C object 'drivers/drivers...c@sta/net_softnic_rte_eth_softnic_action.c.o'.
FAILED: drivers/drivers@@tmp_rte_pmd_softnic@sta/net_softnic_rte_eth_softnic_action.c.o
cc -Idrivers/drivers@@tmp_rte_pmd_softnic@sta -Idrivers -I../drivers -Idrivers/net/softnic -I../drivers/net/softnic -Ilib/librte_ethdev -I../lib/librte_ethdev -I. -I../ -Iconfig -I../config-Ilib/librte_eal/common/include -I../lib/librte_eal/common/include -I../lib/librte_eal/linux/eal/include -Ilib/librte_eal/common -I../lib/librte_eal/common -Ilib/librte_eal/

Quoted 8 lines

common/include/arch/arm -I../lib/librte_eal/common/include/arch/arm -Ilib/librte_eal -I../lib/librte_eal -Ilib/librte_kvargs -I../lib/librte_kvargs -Ilib/librte_net -I../lib/librte_net -Ilib/librte_mbuf -I../lib/librte_mbuf -Ilib/librte_mempool -I../lib/librte_mempool -Ilib/librte_ring -I../lib/librte_ring -Ilib/librte_cmdline -I../lib/librte_cmdline -Ilib/lib rte_meter -I../lib/librte_meter -Idrivers/bus/pci -I../drivers/bus/pci -I../drivers/bus/pci/linux -Ilib/librte_pci -I../lib/librte_pci -Idrivers/bus/vdev -I../drivers/bus/vdev -Ilib/librte_pipeline -I../lib/librte_pipeline -Ilib/librte_port -I../lib/librte_port -Ilib/librte_sched -I../lib/librte_sched -Ilib/librte_ip_frag -I../lib/librte_ip_frag -Ilib/librte_h ash -I../lib/librte_hash -Ilib/librte_cryptodev -I../lib/librte_cryptodev -Ilib/librte_kni -I../lib/librte_kni -Ilib/librte_table -I../lib/librte_table -Ilib/librte_lpm -I../lib/librte_lpm -Ilib/librte_acl -I../lib/librte_acl -pipe -D_FILE_OFFSET_BITS=64 -Wall -Winvalid-pch -O3 -include rte_config.h -Wsign-compare -Wcast-qual -fPIC -D_GNU_SOURCE -DALLOW_EXPERI MENTAL_API -MD -MQ 'drivers/drivers@@tmp_rte_pmd_softnic@sta/net_softnic_rte_eth_softnic_action.c.o' -MF 'drivers/drivers@@tmp_rte_pmd_softnic@sta/net_softnic_rte_eth_softnic_action.c.o.d' -o 'drivers/drivers@@tmp_rte_pmd_softnic@sta/net_softnic_rte_eth_softnic_action.c.o' -c ../drivers/net/softnic/rte_eth_softnic_action.c {standard input}: Assembler messages:
{standard input}:14: Error: selected processor does not support `crc32cx w3,w3,x0'
{standard input}:37: Error: selected processor does not support `crc32cx w1,w1,x3'
{standard input}:40: Error: selected processor does not support `crc32cx w0,w0,x2'

My machine has 0x41(Arm) and 0xd08(cortex-a72). gcc is '4.8.5 20150623 (Red Hat 4.8.5-28)'

Thanks,
Yongseok

Quoted 21 lines
Quoted 3 lines

The command output check can also be removed as it is handled when calling the command script itself.

+1

Quoted 5 lines

Thoughts?

PS. I think the safest way to set CACHELINE_SIZE is to read the cache type register[1] but sadly only few latest kernels have the support through sysfs (/sys/devices/system/cpu/cpu0/cache/index0/coherency_line_size)

+1

In summary, +3. LoL

I'll also submit a patch to change the default cacheline size of cortex-a72 with the new flags_*_extra[]

thanks,
Yongseok

#66 lore raw

Re: [dpdk-dev] [PATCH v8 2/4] meson: add infra to support machine specific flags

Quoted 65 lines

-----Original Message-----
From: Yongseok Koh <[email protected]>
Sent: Friday, April 12, 2019 7:35 AM
To: Pavan Nikhilesh Bhagavatula <[email protected]>
Cc: Thomas Monjalon <[email protected]>; dev <[email protected]>; Jerin Jacob Kollanukkaran <[email protected]>; [email protected]
Subject: [EXT] Re: [dpdk-dev] [PATCH v8 2/4] meson: add infra to support machine specific flags

External Email

I've tested it but still have an issue with old gcc.
Even if -mcpu isn't set due to cc.has_argument(), -march isn't set either. So, it spews error due to lack of CRC feature.
-march should have '+crc'. The error I got was:

Quoted 2 lines

ninja: Entering directory `build'
[942/1452] Compiling C object

'drivers/drivers...c@sta/net_softnic_rte_eth_softnic_action.c.o'.

Quoted 2 lines

FAILED:

drivers/drivers@@tmp_rte_pmd_softnic@sta/net_softnic_rte_eth_softnic _a

Quoted 24 lines

ction.c.o cc -Idrivers/drivers@@tmp_rte_pmd_softnic@sta -Idrivers

(quoted diff) +0 -23
-I../drivers -Idrivers/net/softnic -I../drivers/net/softnic
-Ilib/librte_ethdev -I../lib/librte_ethdev -I. -I../ -Iconfig
-I../config-Ilib/librte_eal/common/include
-I../lib/librte_eal/common/include
-I../lib/librte_eal/linux/eal/include -Ilib/librte_eal/common
-I../lib/librte_eal/common -Ilib/librte_eal/ common/include/arch/arm
-I../lib/librte_eal/common/include/arch/arm -Ilib/librte_eal
-I../lib/librte_eal -Ilib/librte_kvargs -I../lib/librte_kvargs
-Ilib/librte_net -I../lib/librte_net -Ilib/librte_mbuf
-I../lib/librte_mbuf -Ilib/librte_mempool -I../lib/librte_mempool
-Ilib/librte_ring -I../lib/librte_ring -Ilib/librte_cmdline
-I../lib/librte_cmdline -Ilib/lib rte_meter -I../lib/librte_meter
-Idrivers/bus/pci -I../drivers/bus/pci -I../drivers/bus/pci/linux
-Ilib/librte_pci -I../lib/librte_pci -Idrivers/bus/vdev
-I../drivers/bus/vdev -Ilib/librte_pipeline -I../lib/librte_pipeline
-Ilib/librte_port -I../lib/librte_port -Ilib/librte_sched
-I../lib/librte_sched -Ilib/librte_ip_frag -I../lib/librte_ip_frag
-Ilib/librte_h ash -I../lib/librte_hash -Ilib/librte_cryptodev
-I../lib/librte_cryptodev -Ilib/librte_kni -I../lib/librte_kni
-Ilib/librte_table -I../lib/librte_table -Ilib/librte_lpm
-I../lib/librte_lpm -Ilib/librte_acl -I../lib/librte_acl -pipe
-D_FILE_OFFSET_BITS=64 -Wall -Winvalid-pch -O3 -include rte_config.h
-Wsign-compare -Wcast-qual -fPIC -D_GNU_SOURCE -DALLOW_EXPERI
Quoted 2 lines

MENTAL_API -MD -MQ

'drivers/drivers@@tmp_rte_pmd_softnic@sta/net_softnic_rte_eth_softnic _

Quoted 2 lines

action.c.o' -MF

'drivers/drivers@@tmp_rte_pmd_softnic@sta/net_softnic_rte_eth_softnic _

Quoted 2 lines

action.c.o.d' -o

'drivers/drivers@@tmp_rte_pmd_softnic@sta/net_softnic_rte_eth_softnic _

Quoted 3 lines

action.c.o' -c ../drivers/net/softnic/rte_eth_softnic_action.c {standard input}: Assembler messages:
{standard input}:14: Error: selected processor does not support `crc32cx

w3,w3,x0'

Quoted 8 lines
Quoted 1 line

{standard input}:37: Error: selected processor does not support `crc32cx

w1,w1,x3'

Quoted 1 line

{standard input}:40: Error: selected processor does not support `crc32cx

w0,w0,x2'

My machine has 0x41(Arm) and 0xd08(cortex-a72). gcc is '4.8.5 20150623 (Red Hat 4.8.5-28)'

Are you testing with very latest master where the following patch available in build?
http://patches.dpdk.org/patch/52367/
It should fix that issue.

Quoted 30 lines

Thanks,
Yongseok

Quoted 3 lines
Quoted 3 lines

The command output check can also be removed as it is handled when

calling the command script itself.

Quoted 21 lines

+1

Quoted 7 lines

Thoughts?

PS. I think the safest way to set CACHELINE_SIZE is to read the cache type register[1] but sadly only few latest kernels have the support through sysfs (/sys/devices/system/cpu/cpu0/cache/index0/coherency_line_size)

+1

In summary, +3. LoL

I'll also submit a patch to change the default cacheline size of cortex-a72 with the new flags_*_extra[]

thanks,
Yongseok

#67 lore raw

Re: [dpdk-dev] [PATCH v8 2/4] meson: add infra to support machine specific flags

Quoted 74 lines · Jerin Jacob Kollanukkaran
Quoted 72 lines

-----Original Message-----
From: Yongseok Koh <[email protected]>
Sent: Friday, April 12, 2019 7:35 AM
To: Pavan Nikhilesh Bhagavatula <[email protected]>
Cc: Thomas Monjalon <[email protected]>; dev <[email protected]>; Jerin Jacob Kollanukkaran <[email protected]>; [email protected]
Subject: [EXT] Re: [dpdk-dev] [PATCH v8 2/4] meson: add infra to support machine specific flags

External Email

I've tested it but still have an issue with old gcc.
Even if -mcpu isn't set due to cc.has_argument(), -march isn't set either. So, it spews error due to lack of CRC feature.
-march should have '+crc'. The error I got was:

Quoted 2 lines

ninja: Entering directory `build'
[942/1452] Compiling C object

'drivers/drivers...c@sta/net_softnic_rte_eth_softnic_action.c.o'.

Quoted 2 lines

FAILED:

drivers/drivers@@tmp_rte_pmd_softnic@sta/net_softnic_rte_eth_softnic _a

Quoted 24 lines

ction.c.o cc -Idrivers/drivers@@tmp_rte_pmd_softnic@sta -Idrivers

(quoted diff) +0 -23
-I../drivers -Idrivers/net/softnic -I../drivers/net/softnic
-Ilib/librte_ethdev -I../lib/librte_ethdev -I. -I../ -Iconfig
-I../config-Ilib/librte_eal/common/include
-I../lib/librte_eal/common/include
-I../lib/librte_eal/linux/eal/include -Ilib/librte_eal/common
-I../lib/librte_eal/common -Ilib/librte_eal/ common/include/arch/arm
-I../lib/librte_eal/common/include/arch/arm -Ilib/librte_eal
-I../lib/librte_eal -Ilib/librte_kvargs -I../lib/librte_kvargs
-Ilib/librte_net -I../lib/librte_net -Ilib/librte_mbuf
-I../lib/librte_mbuf -Ilib/librte_mempool -I../lib/librte_mempool
-Ilib/librte_ring -I../lib/librte_ring -Ilib/librte_cmdline
-I../lib/librte_cmdline -Ilib/lib rte_meter -I../lib/librte_meter
-Idrivers/bus/pci -I../drivers/bus/pci -I../drivers/bus/pci/linux
-Ilib/librte_pci -I../lib/librte_pci -Idrivers/bus/vdev
-I../drivers/bus/vdev -Ilib/librte_pipeline -I../lib/librte_pipeline
-Ilib/librte_port -I../lib/librte_port -Ilib/librte_sched
-I../lib/librte_sched -Ilib/librte_ip_frag -I../lib/librte_ip_frag
-Ilib/librte_h ash -I../lib/librte_hash -Ilib/librte_cryptodev
-I../lib/librte_cryptodev -Ilib/librte_kni -I../lib/librte_kni
-Ilib/librte_table -I../lib/librte_table -Ilib/librte_lpm
-I../lib/librte_lpm -Ilib/librte_acl -I../lib/librte_acl -pipe
-D_FILE_OFFSET_BITS=64 -Wall -Winvalid-pch -O3 -include rte_config.h
-Wsign-compare -Wcast-qual -fPIC -D_GNU_SOURCE -DALLOW_EXPERI
Quoted 2 lines

MENTAL_API -MD -MQ

'drivers/drivers@@tmp_rte_pmd_softnic@sta/net_softnic_rte_eth_softnic _

Quoted 2 lines

action.c.o' -MF

'drivers/drivers@@tmp_rte_pmd_softnic@sta/net_softnic_rte_eth_softnic _

Quoted 2 lines

action.c.o.d' -o

'drivers/drivers@@tmp_rte_pmd_softnic@sta/net_softnic_rte_eth_softnic _

Quoted 3 lines

action.c.o' -c ../drivers/net/softnic/rte_eth_softnic_action.c {standard input}: Assembler messages:
{standard input}:14: Error: selected processor does not support `crc32cx

w3,w3,x0'

Quoted 1 line

{standard input}:37: Error: selected processor does not support `crc32cx

w1,w1,x3'

Quoted 1 line

{standard input}:40: Error: selected processor does not support `crc32cx

w0,w0,x2'

Quoted 6 lines
Quoted 2 lines

My machine has 0x41(Arm) and 0xd08(cortex-a72). gcc is '4.8.5 20150623 (Red Hat 4.8.5-28)'

Are you testing with very latest master where the following patch available in build?
https://eur03.safelinks.protection.outlook.com/?url=http%3A%2F%2Fpatches.dpdk.org%2Fpatch%2F52367%2F&amp;data=02%7C01%7Cyskoh%40mellanox.com%7C9adebf1529ab4bbc189708d6bf0d1d00%7Ca652971c7d2e4d9ba6a4d149256f461b%7C0%7C0%7C636906460384104429&amp;sdata=vG65lezE%2BZacEpn38cUoozwEYm%2BBUGvuYBQ2ToEKnSI%3D&amp;reserved=0
It should fix that issue.

Thanks, that fixes the issue.
But I've encountered another one. Are you aware of this?

ninja: Entering directory `build'
[1151/1452] Compiling C object 'drivers/drivers@@tmp_r...d_octeontx_event@sta/event_octeontx_timvf_worker.c.o'.
FAILED: drivers/drivers@@tmp_rte_pmd_octeontx_event@sta/event_octeontx_timvf_worker.c.o
cc -Idrivers/drivers@@tmp_rte_pmd_octeontx_event@sta -Idrivers -I../drivers -Idrivers/event/octeontx -I../drivers/event/octeontx -Ilib/librte_eventdev -I../lib/librte_eventdev -I. -I../ -Iconfig -I../config -Ilib/librte_eal/common/include -I../lib/librte_eal/common/include -I../lib/librte_eal/linux/eal/include -Ilib/librte_eal/common -I../lib/librte_eal/common -Ilib/librte_eal/common/include/arch/arm -I../lib/librte_eal/co mmon/include/arch/arm -Ilib/librte_eal -I../lib/librte_eal -Ilib/librte_kvargs -I../lib/librte_kvargs -Ilib/librte_ring -I../lib/librte_ring -Ilib/librte_ethdev -I../lib/librte_ethdev -Ilib/librte_net -I../lib/librte_net -Ilib/librte_mbuf -I../lib/librte_mbuf -Ilib/librte_mempool -I../lib/librte_mempool -Ilib/librte_cmdline -I../lib/librte_cmdline -Ilib/librte_meter -I../lib/librte_meter -Ilib/librte_hash -I../lib/librte_h ash -Ilib/librte_timer -I../lib/librte_timer -Ilib/librte_cryptodev -I../lib/librte_cryptodev -Idrivers/common/octeontx -I../drivers/common/octeontx -Idrivers/mempool/octeontx -I../drivers/mempool/octeontx -Idrivers/bus/pci -I../drivers/bus/pci -I../drivers/bus/pci/linux -Ilib/librte_pci -I../lib/librte_pci -Idrivers/bus/vdev -I../drivers/bus/vdev -Idrivers/net/octeontx -I../drivers/net/octeontx -Idrivers/net/octeontx/base
-I../drivers/net/octeontx/base -pipe -D_FILE_OFFSET_BITS=64 -Wall -Winvalid-pch -O3 -include rte_config.h -Wsign-compare -Wcast-qual -fPIC -D_GNU_SOURCE -DALLOW_EXPERIMENTAL_API -MD -MQ 'drivers/drivers@@tmp_rte_pmd_octeontx_event@sta/event_octeontx_timvf_worker.c.o' -MF 'drivers/drivers@@tmp_rte_pmd_octeontx_event@sta/event_octeontx_timvf_worker.c.o.d' -o 'drivers/drivers@@tmp_rte_pmd_octeontx_event@sta/event_octeontx_t
imvf_worker.c.o' -c ../drivers/event/octeontx/timvf_worker.c
../drivers/event/octeontx/timvf_worker.c: In function ‘timvf_timer_arm_burst_sp’:
../drivers/event/octeontx/timvf_worker.c:88:1: error: could not split insn
}
^
(insn 95 98 99 (parallel [

            (set (reg:DI 3 x3 [orig:98 D.8656 ] [98])
                (mem/v:DI (reg/f:DI 21 x21 [orig:88 D.8662 ] [88]) [-1  S8 A64]))
            (set (mem/v:DI (reg/f:DI 21 x21 [orig:88 D.8662 ] [88]) [-1  S8 A64])
                (unspec_volatile:DI [
                        (plus:DI (mem/v:DI (reg/f:DI 21 x21 [orig:88 D.8662 ] [88]) [-1  S8 A64])
                            (const_int -281474976710656 [0xffff000000000000]))
                        (const_int 0 [0])
                    ] UNSPECV_ATOMIC_OP))
            (clobber (reg:CC 66 cc))
            (clobber (reg:DI 0 x0))
            (clobber (reg:SI 1 x1))
        ]) ../drivers/event/octeontx/timvf_worker.h:95 1832 {atomic_fetch_adddi}
     (expr_list:REG_UNUSED (reg:CC 66 cc)
        (expr_list:REG_UNUSED (reg:SI 1 x1)
            (expr_list:REG_UNUSED (reg:DI 0 x0)
                (nil)))))

../drivers/event/octeontx/timvf_worker.c:88:1: internal compiler error: in final_scan_insn, at final.c:2897 Please submit a full bug report,
with preprocessed source if appropriate.
See <http://bugzilla.redhat.com/bugzilla> for instructions.
{standard input}: Assembler messages:
{standard input}: Error: open CFI at the end of file; missing .cfi_endproc directive Preprocessed source stored into /tmp/ccnQRbOm.out file, please attach this to your bugreport. [1168/1452] Compiling C object 'drivers/drivers@@tmp_r...ntx_crypto@sta/crypto_octeontx_otx_cryptodev_ops.c.o'.
ninja: build stopped: subcommand failed.

Thanks
Yongseok

author Acked-by Ferruh Yigit
#68 lore raw

Re: [dpdk-dev] [PATCH v8 2/4] meson: add infra to support machine specific flags

Quoted 11 lines

-----Original Message-----
From: Yongseok Koh <[email protected]>
Sent: Friday, April 12, 2019 12:14 PM
To: Jerin Jacob Kollanukkaran <[email protected]>
Cc: Pavan Nikhilesh Bhagavatula <[email protected]>; Thomas Monjalon <[email protected]>; dev <[email protected]>; [email protected]
Subject: [EXT] Re: [dpdk-dev] [PATCH v8 2/4] meson: add infra to support machine specific flags

Quoted 2 lines

Thanks, that fixes the issue.
But I've encountered another one. Are you aware of this?

Yes. It is a compiler bug. This patch set is NOT introducing this.

Fixed same on legacy build with arm64 . If you are using < 4.8.6 compiler For meson, this patched needs to be ported to meson

$ git show f3af3e44a444cdfe3fa7b3e2c042be351401eb23
commit f3af3e44a444cdfe3fa7b3e2c042be351401eb23
Author: Jerin Jacob <[email protected]>
Date:   Mon Sep 3 15:01:10 2018 +0530

mk: disable OcteonTx for buggy compilers only on arm64

Disable octeontx for gcc 4.8.5 as the compiler is emitting "internal compiler error" for aarch64. The GCC "internal compiler error" was observed only for arm64 architecture so disable the PMD only for arm64.

Fixes 4f760550a093 ("mk: disable OcteonTx for buggy compilers")
Quoted 73 lines

ninja: Entering directory `build'
[1151/1452] Compiling C object 'drivers/drivers@@tmp_r...d_octeontx_event@sta/event_octeontx_timvf_ worker.c.o'.
FAILED:
drivers/drivers@@tmp_rte_pmd_octeontx_event@sta/event_octeontx_ti mvf_worker.c.o
cc -Idrivers/drivers@@tmp_rte_pmd_octeontx_event@sta -Idrivers - I../drivers -Idrivers/event/octeontx -I../drivers/event/octeontx - Ilib/librte_eventdev -I../lib/librte_eventdev -I. -I../ -Iconfig -I../config - Ilib/librte_eal/common/include -I../lib/librte_eal/common/include - I../lib/librte_eal/linux/eal/include -Ilib/librte_eal/common - I../lib/librte_eal/common -Ilib/librte_eal/common/include/arch/arm - I../lib/librte_eal/co mmon/include/arch/arm -Ilib/librte_eal -I../lib/librte_eal -Ilib/librte_kvargs -I../lib/librte_kvargs -Ilib/librte_ring -I../lib/librte_ring - Ilib/librte_ethdev -I../lib/librte_ethdev -Ilib/librte_net -I../lib/librte_net - Ilib/librte_mbuf -I../lib/librte_mbuf -Ilib/librte_mempool - I../lib/librte_mempool -Ilib/librte_cmdline -I../lib/librte_cmdline - Ilib/librte_meter -I../lib/librte_meter -Ilib/librte_hash -I../lib/librte_h ash - Ilib/librte_timer -I../lib/librte_timer -Ilib/librte_cryptodev - I../lib/librte_cryptodev -Idrivers/common/octeontx - I../drivers/common/octeontx -Idrivers/mempool/octeontx - I../drivers/mempool/octeontx -Idrivers/bus/pci -I../drivers/bus/pci - I../drivers/bus/pci/linux -Ilib/librte_pci -I../lib/librte_pci -Idrivers/bus/vdev - I../drivers/bus/vdev -Idrivers/net/octeontx -I../drivers/net/octeontx - Idrivers/net/octeontx/base -I../drivers/net/octeontx/base -pipe - D_FILE_OFFSET_BITS=64 -Wall -Winvalid-pch -O3 -include rte_config.h - Wsign-compare -Wcast-qual -fPIC -D_GNU_SOURCE - DALLOW_EXPERIMENTAL_API -MD -MQ 'drivers/drivers@@tmp_rte_pmd_octeontx_event@sta/event_octeontx_ti mvf_worker.c.o' -MF 'drivers/drivers@@tmp_rte_pmd_octeontx_event@sta/event_octeontx_ti mvf_worker.c.o.d' -o 'drivers/drivers@@tmp_rte_pmd_octeontx_event@sta/event_octeontx_t imvf_worker.c.o' -c ../drivers/event/octeontx/timvf_worker.c ../drivers/event/octeontx/timvf_worker.c: In function ‘timvf_timer_arm_burst_sp’:
../drivers/event/octeontx/timvf_worker.c:88:1: error: could not split insn } ^ (insn 95 98 99 (parallel [

            (set (reg:DI 3 x3 [orig:98 D.8656 ] [98])
                (mem/v:DI (reg/f:DI 21 x21 [orig:88 D.8662 ] [88]) [-1  S8 A64]))
            (set (mem/v:DI (reg/f:DI 21 x21 [orig:88 D.8662 ] [88]) [-1  S8 A64])
                (unspec_volatile:DI [
                        (plus:DI (mem/v:DI (reg/f:DI 21 x21 [orig:88 D.8662 ] [88]) [-1  S8
A64])
                            (const_int -281474976710656 [0xffff000000000000]))
                        (const_int 0 [0])
                    ] UNSPECV_ATOMIC_OP))
            (clobber (reg:CC 66 cc))
            (clobber (reg:DI 0 x0))
            (clobber (reg:SI 1 x1))
        ]) ../drivers/event/octeontx/timvf_worker.h:95 1832
{atomic_fetch_adddi}
     (expr_list:REG_UNUSED (reg:CC 66 cc)
        (expr_list:REG_UNUSED (reg:SI 1 x1)
            (expr_list:REG_UNUSED (reg:DI 0 x0)
                (nil)))))

../drivers/event/octeontx/timvf_worker.c:88:1: internal compiler error: in final_scan_insn, at final.c:2897 Please submit a full bug report, with preprocessed source if appropriate.
See <http://bugzilla.redhat.com/bugzilla> for instructions. {standard input}: Assembler messages:
{standard input}: Error: open CFI at the end of file; missing .cfi_endproc directive Preprocessed source stored into /tmp/ccnQRbOm.out file, please attach this to your bugreport.
[1168/1452] Compiling C object 'drivers/drivers@@tmp_r...ntx_crypto@sta/crypto_octeontx_otx_cryptode v_ops.c.o'.
ninja: build stopped: subcommand failed.

Thanks
Yongseok

#69 lore raw

Re: [dpdk-dev] [PATCH v8 2/4] meson: add infra to support machine specific flags

Quoted 14 lines · Jerin Jacob Kollanukkaran
Quoted 12 lines

-----Original Message-----
From: Yongseok Koh <[email protected]>
Sent: Friday, April 12, 2019 12:14 PM
To: Jerin Jacob Kollanukkaran <[email protected]>
Cc: Pavan Nikhilesh Bhagavatula <[email protected]>; Thomas Monjalon <[email protected]>; dev <[email protected]>; [email protected]
Subject: [EXT] Re: [dpdk-dev] [PATCH v8 2/4] meson: add infra to support machine specific flags

Quoted 7 lines
Quoted 2 lines

Thanks, that fixes the issue.
But I've encountered another one. Are you aware of this?

Yes. It is a compiler bug. This patch set is NOT introducing this.

Fixed same on legacy build with arm64 . If you are using < 4.8.6 compiler For meson, this patched needs to be ported to meson

Okay, I've ported it. Will submit it soon.

Yongseok

Quoted 93 lines
$ git show f3af3e44a444cdfe3fa7b3e2c042be351401eb23
commit f3af3e44a444cdfe3fa7b3e2c042be351401eb23
Author: Jerin Jacob <[email protected]>
Date:   Mon Sep 3 15:01:10 2018 +0530

mk: disable OcteonTx for buggy compilers only on arm64

Disable octeontx for gcc 4.8.5 as the compiler is emitting "internal compiler error" for aarch64. The GCC "internal compiler error" was observed only for arm64 architecture so disable the PMD only for arm64.

Fixes 4f760550a093 ("mk: disable OcteonTx for buggy compilers")
Quoted 73 lines

ninja: Entering directory `build'
[1151/1452] Compiling C object 'drivers/drivers@@tmp_r...d_octeontx_event@sta/event_octeontx_timvf_ worker.c.o'.
FAILED:
drivers/drivers@@tmp_rte_pmd_octeontx_event@sta/event_octeontx_ti mvf_worker.c.o
cc -Idrivers/drivers@@tmp_rte_pmd_octeontx_event@sta -Idrivers - I../drivers -Idrivers/event/octeontx -I../drivers/event/octeontx - Ilib/librte_eventdev -I../lib/librte_eventdev -I. -I../ -Iconfig -I../config - Ilib/librte_eal/common/include -I../lib/librte_eal/common/include - I../lib/librte_eal/linux/eal/include -Ilib/librte_eal/common - I../lib/librte_eal/common -Ilib/librte_eal/common/include/arch/arm - I../lib/librte_eal/co mmon/include/arch/arm -Ilib/librte_eal -I../lib/librte_eal -Ilib/librte_kvargs -I../lib/librte_kvargs -Ilib/librte_ring -I../lib/librte_ring - Ilib/librte_ethdev -I../lib/librte_ethdev -Ilib/librte_net -I../lib/librte_net - Ilib/librte_mbuf -I../lib/librte_mbuf -Ilib/librte_mempool - I../lib/librte_mempool -Ilib/librte_cmdline -I../lib/librte_cmdline - Ilib/librte_meter -I../lib/librte_meter -Ilib/librte_hash -I../lib/librte_h ash - Ilib/librte_timer -I../lib/librte_timer -Ilib/librte_cryptodev - I../lib/librte_cryptodev -Idrivers/common/octeontx - I../drivers/common/octeontx -Idrivers/mempool/octeontx - I../drivers/mempool/octeontx -Idrivers/bus/pci -I../drivers/bus/pci - I../drivers/bus/pci/linux -Ilib/librte_pci -I../lib/librte_pci -Idrivers/bus/vdev - I../drivers/bus/vdev -Idrivers/net/octeontx -I../drivers/net/octeontx - Idrivers/net/octeontx/base -I../drivers/net/octeontx/base -pipe - D_FILE_OFFSET_BITS=64 -Wall -Winvalid-pch -O3 -include rte_config.h - Wsign-compare -Wcast-qual -fPIC -D_GNU_SOURCE - DALLOW_EXPERIMENTAL_API -MD -MQ 'drivers/drivers@@tmp_rte_pmd_octeontx_event@sta/event_octeontx_ti mvf_worker.c.o' -MF 'drivers/drivers@@tmp_rte_pmd_octeontx_event@sta/event_octeontx_ti mvf_worker.c.o.d' -o 'drivers/drivers@@tmp_rte_pmd_octeontx_event@sta/event_octeontx_t imvf_worker.c.o' -c ../drivers/event/octeontx/timvf_worker.c ../drivers/event/octeontx/timvf_worker.c: In function ‘timvf_timer_arm_burst_sp’:
../drivers/event/octeontx/timvf_worker.c:88:1: error: could not split insn } ^ (insn 95 98 99 (parallel [

           (set (reg:DI 3 x3 [orig:98 D.8656 ] [98])
               (mem/v:DI (reg/f:DI 21 x21 [orig:88 D.8662 ] [88]) [-1  S8 A64]))
           (set (mem/v:DI (reg/f:DI 21 x21 [orig:88 D.8662 ] [88]) [-1  S8 A64])
               (unspec_volatile:DI [
                       (plus:DI (mem/v:DI (reg/f:DI 21 x21 [orig:88 D.8662 ] [88]) [-1  S8
A64])
                           (const_int -281474976710656 [0xffff000000000000]))
                       (const_int 0 [0])
                   ] UNSPECV_ATOMIC_OP))
           (clobber (reg:CC 66 cc))
           (clobber (reg:DI 0 x0))
           (clobber (reg:SI 1 x1))
       ]) ../drivers/event/octeontx/timvf_worker.h:95 1832
{atomic_fetch_adddi}
    (expr_list:REG_UNUSED (reg:CC 66 cc)
       (expr_list:REG_UNUSED (reg:SI 1 x1)
           (expr_list:REG_UNUSED (reg:DI 0 x0)
               (nil)))))

../drivers/event/octeontx/timvf_worker.c:88:1: internal compiler error: in
final_scan_insn, at final.c:2897 Please submit a full bug report, with
preprocessed source if appropriate.
See <https://eur03.safelinks.protection.outlook.com/?url=http%3A%2F%2Fbugzilla.redhat.com%2Fbugzilla&amp;data=02%7C01%7Cyskoh%40mellanox.com%7C557fa46858104e15443508d6bf1494e9%7Ca652971c7d2e4d9ba6a4d149256f461b%7C0%7C0%7C636906492451079366&amp;sdata=EpdRkARHJDQCd6LRSe7gWjCghjIec%2Fx%2BJjbJiDNEYm4%3D&amp;reserved=0> for instructions. {standard input}: Assembler messages:
{standard input}: Error: open CFI at the end of file; missing .cfi_endproc
directive Preprocessed source stored into /tmp/ccnQRbOm.out file, please
attach this to your bugreport.
[1168/1452] Compiling C object
'drivers/drivers@@tmp_r...ntx_crypto@sta/crypto_octeontx_otx_cryptode
v_ops.c.o'.
ninja: build stopped: subcommand failed.

Thanks
Yongseok

#70 lore raw

Re: [dpdk-dev] [PATCH v8 2/4] meson: add infra to support machine specific flags

In reply to Yongseok Koh
Quoted 116 lines · Yongseok Koh
Quoted 75 lines · Jerin Jacob Kollanukkaran
Quoted 72 lines

-----Original Message-----
From: Yongseok Koh <[email protected]>
Sent: Friday, April 12, 2019 7:35 AM
To: Pavan Nikhilesh Bhagavatula <[email protected]>
Cc: Thomas Monjalon <[email protected]>; dev <[email protected]>; Jerin Jacob Kollanukkaran <[email protected]>; [email protected]
Subject: [EXT] Re: [dpdk-dev] [PATCH v8 2/4] meson: add infra to support machine specific flags

External Email

I've tested it but still have an issue with old gcc.
Even if -mcpu isn't set due to cc.has_argument(), -march isn't set either. So, it spews error due to lack of CRC feature.
-march should have '+crc'. The error I got was:

Quoted 2 lines

ninja: Entering directory `build'
[942/1452] Compiling C object

'drivers/drivers...c@sta/net_softnic_rte_eth_softnic_action.c.o'.

Quoted 2 lines

FAILED:

drivers/drivers@@tmp_rte_pmd_softnic@sta/net_softnic_rte_eth_softnic _a

Quoted 24 lines

ction.c.o cc -Idrivers/drivers@@tmp_rte_pmd_softnic@sta -Idrivers

(quoted diff) +0 -23
-I../drivers -Idrivers/net/softnic -I../drivers/net/softnic
-Ilib/librte_ethdev -I../lib/librte_ethdev -I. -I../ -Iconfig
-I../config-Ilib/librte_eal/common/include
-I../lib/librte_eal/common/include
-I../lib/librte_eal/linux/eal/include -Ilib/librte_eal/common
-I../lib/librte_eal/common -Ilib/librte_eal/ common/include/arch/arm
-I../lib/librte_eal/common/include/arch/arm -Ilib/librte_eal
-I../lib/librte_eal -Ilib/librte_kvargs -I../lib/librte_kvargs
-Ilib/librte_net -I../lib/librte_net -Ilib/librte_mbuf
-I../lib/librte_mbuf -Ilib/librte_mempool -I../lib/librte_mempool
-Ilib/librte_ring -I../lib/librte_ring -Ilib/librte_cmdline
-I../lib/librte_cmdline -Ilib/lib rte_meter -I../lib/librte_meter
-Idrivers/bus/pci -I../drivers/bus/pci -I../drivers/bus/pci/linux
-Ilib/librte_pci -I../lib/librte_pci -Idrivers/bus/vdev
-I../drivers/bus/vdev -Ilib/librte_pipeline -I../lib/librte_pipeline
-Ilib/librte_port -I../lib/librte_port -Ilib/librte_sched
-I../lib/librte_sched -Ilib/librte_ip_frag -I../lib/librte_ip_frag
-Ilib/librte_h ash -I../lib/librte_hash -Ilib/librte_cryptodev
-I../lib/librte_cryptodev -Ilib/librte_kni -I../lib/librte_kni
-Ilib/librte_table -I../lib/librte_table -Ilib/librte_lpm
-I../lib/librte_lpm -Ilib/librte_acl -I../lib/librte_acl -pipe
-D_FILE_OFFSET_BITS=64 -Wall -Winvalid-pch -O3 -include rte_config.h
-Wsign-compare -Wcast-qual -fPIC -D_GNU_SOURCE -DALLOW_EXPERI
Quoted 2 lines

MENTAL_API -MD -MQ

'drivers/drivers@@tmp_rte_pmd_softnic@sta/net_softnic_rte_eth_softnic _

Quoted 2 lines

action.c.o' -MF

'drivers/drivers@@tmp_rte_pmd_softnic@sta/net_softnic_rte_eth_softnic _

Quoted 2 lines

action.c.o.d' -o

'drivers/drivers@@tmp_rte_pmd_softnic@sta/net_softnic_rte_eth_softnic _

Quoted 3 lines

action.c.o' -c ../drivers/net/softnic/rte_eth_softnic_action.c {standard input}: Assembler messages:
{standard input}:14: Error: selected processor does not support `crc32cx

w3,w3,x0'

Quoted 1 line

{standard input}:37: Error: selected processor does not support `crc32cx

w1,w1,x3'

Quoted 1 line

{standard input}:40: Error: selected processor does not support `crc32cx

w0,w0,x2'

Quoted 6 lines
Quoted 2 lines

My machine has 0x41(Arm) and 0xd08(cortex-a72). gcc is '4.8.5 20150623 (Red Hat 4.8.5-28)'

Are you testing with very latest master where the following patch available in build?
https://eur03.safelinks.protection.outlook.com/?url=http%3A%2F%2Fpatches.dpdk.org%2Fpatch%2F52367%2F&amp;data=02%7C01%7Cyskoh%40mellanox.com%7C5909260f30a64e07a95108d6bf123bde%7Ca652971c7d2e4d9ba6a4d149256f461b%7C0%7C0%7C636906482386596820&amp;sdata=4%2BfRfELXK37SNY4wdFNGPF8lpU7S3DEfPoDfAH5K7GE%3D&amp;reserved=0
It should fix that issue.

Thanks, that fixes the issue.
But I've encountered another one. Are you aware of this?

ninja: Entering directory `build'
[1151/1452] Compiling C object 'drivers/drivers@@tmp_r...d_octeontx_event@sta/event_octeontx_timvf_worker.c.o'.
FAILED: drivers/drivers@@tmp_rte_pmd_octeontx_event@sta/event_octeontx_timvf_worker.c.o
cc -Idrivers/drivers@@tmp_rte_pmd_octeontx_event@sta -Idrivers -I../drivers -Idrivers/event/octeontx -I../drivers/event/octeontx -Ilib/librte_eventdev -I../lib/librte_eventdev -I. -I../ -Iconfig -I../config -Ilib/librte_eal/common/include -I../lib/librte_eal/common/include -I../lib/librte_eal/linux/eal/include -Ilib/librte_eal/common -I../lib/librte_eal/common -Ilib/librte_eal/common/include/arch/arm -I../lib/librte_eal/co mmon/include/arch/arm -Ilib/librte_eal -I../lib/librte_eal -Ilib/librte_kvargs -I../lib/librte_kvargs -Ilib/librte_ring -I../lib/librte_ring -Ilib/librte_ethdev -I../lib/librte_ethdev -Ilib/librte_net -I../lib/librte_net -Ilib/librte_mbuf -I../lib/librte_mbuf -Ilib/librte_mempool -I../lib/librte_mempool -Ilib/librte_cmdline -I../lib/librte_cmdline -Ilib/librte_meter -I../lib/librte_meter -Ilib/librte_hash -I../lib/librte_h ash -Ilib/librte_timer -I../lib/librte_timer -Ilib/librte_cryptodev -I../lib/librte_cryptodev -Idrivers/common/octeontx -I../drivers/common/octeontx -Idrivers/mempool/octeontx -I../drivers/mempool/octeontx -Idrivers/bus/pci -I../drivers/bus/pci -I../drivers/bus/pci/linux -Ilib/librte_pci -I../lib/librte_pci -Idrivers/bus/vdev -I../drivers/bus/vdev -Idrivers/net/octeontx -I../drivers/net/octeontx -Idrivers/net/octeontx/base -I../drivers/net/octeontx/base -pipe -D_FILE_OFFSET_BITS=64 -Wall -Winvalid-pch -O3 -include rte_config.h -Wsign-compare -Wcast-qual -fPIC -D_GNU_SOURCE -DALLOW_EXPERIMENTAL_API -MD -MQ 'drivers/drivers@@tmp_rte_pmd_octeontx_event@sta/event_octeontx_timvf_worker.c.o' -MF 'drivers/drivers@@tmp_rte_pmd_octeontx_event@sta/event_octeontx_timvf_worker.c.o.d' -o 'drivers/drivers@@tmp_rte_pmd_octeontx_event@sta/event_octeontx_t imvf_worker.c.o' -c ../drivers/event/octeontx/timvf_worker.c
../drivers/event/octeontx/timvf_worker.c: In function ‘timvf_timer_arm_burst_sp’:
../drivers/event/octeontx/timvf_worker.c:88:1: error: could not split insn
}
^
(insn 95 98 99 (parallel [

           (set (reg:DI 3 x3 [orig:98 D.8656 ] [98])
               (mem/v:DI (reg/f:DI 21 x21 [orig:88 D.8662 ] [88]) [-1  S8 A64]))
           (set (mem/v:DI (reg/f:DI 21 x21 [orig:88 D.8662 ] [88]) [-1  S8 A64])
               (unspec_volatile:DI [
                       (plus:DI (mem/v:DI (reg/f:DI 21 x21 [orig:88 D.8662 ] [88]) [-1  S8 A64])
                           (const_int -281474976710656 [0xffff000000000000]))
                       (const_int 0 [0])
                   ] UNSPECV_ATOMIC_OP))
           (clobber (reg:CC 66 cc))
           (clobber (reg:DI 0 x0))
           (clobber (reg:SI 1 x1))
       ]) ../drivers/event/octeontx/timvf_worker.h:95 1832 {atomic_fetch_adddi}
    (expr_list:REG_UNUSED (reg:CC 66 cc)
       (expr_list:REG_UNUSED (reg:SI 1 x1)
           (expr_list:REG_UNUSED (reg:DI 0 x0)
               (nil)))))
Quoted 9 lines

../drivers/event/octeontx/timvf_worker.c:88:1: internal compiler error: in final_scan_insn, at final.c:2897
Please submit a full bug report,
with preprocessed source if appropriate.
See <https://eur03.safelinks.protection.outlook.com/?url=http%3A%2F%2Fbugzilla.redhat.com%2Fbugzilla&amp;data=02%7C01%7Cyskoh%40mellanox.com%7C5909260f30a64e07a95108d6bf123bde%7Ca652971c7d2e4d9ba6a4d149256f461b%7C0%7C0%7C636906482386596820&amp;sdata=2AH1gInkxui7UEDb7LLNppMxDEaf%2F5N5TEHhDRTSDJY%3D&amp;reserved=0> for instructions. {standard input}: Assembler messages:
{standard input}: Error: open CFI at the end of file; missing .cfi_endproc directive
Preprocessed source stored into /tmp/ccnQRbOm.out file, please attach this to your bugreport.
[1168/1452] Compiling C object 'drivers/drivers@@tmp_r...ntx_crypto@sta/crypto_octeontx_otx_cryptodev_ops.c.o'.
ninja: build stopped: subcommand failed.

One more issue.
With gcc7.2, crypto isn't enabled if -mcpu is set.
How about thunderx/octeon?
Looks it should be like -mcpu=cortex-a72+crypto
I'll take care of this in a separate patch.
Because I want to add a new option to control it.
The reason is a binary having crypto support can't be run on a cpu w/o crypto extension, it is panicked. And -mcpu=cortex-a72 includes 'crc' support by default.

Thanks,
Yongseok

#71 lore raw

Re: [dpdk-dev] [PATCH v8 2/4] meson: add infra to support machine specific flags

Quoted 14 lines

-----Original Message-----
From: Yongseok Koh <[email protected]>
Sent: Friday, April 12, 2019 12:39 PM
To: Jerin Jacob Kollanukkaran <[email protected]>
Cc: Pavan Nikhilesh Bhagavatula <[email protected]>; Thomas Monjalon <[email protected]>; dev <[email protected]>; [email protected]
Subject: [EXT] Re: [dpdk-dev] [PATCH v8 2/4] meson: add infra to support machine specific flags

Quoted 3 lines

One more issue.
With gcc7.2, crypto isn't enabled if -mcpu is set. How about thunderx/octeon?

It is enabled when when mcpu=thunderxt88 is provided.
In default case, it picks up -march=armv8-a+crc+crypto if mpcu is not matching. So I don’t see any issue.

Quoted 1 line

Looks it should be like -mcpu=cortex-a72+crypto I'll take care of this in a

mcpu should be enough. Looks like like it a compiler bug, Is cortex-a72 available with out crypto? mcpu suppose to tune for specific cpu.

Quoted 5 lines

separate patch.
Because I want to add a new option to control it.
The reason is a binary having crypto support can't be run on a cpu w/o crypto extension, it is panicked.
And -mcpu=cortex-a72 includes 'crc' support by default.

I suggest to add config/arm/arm64_a72_linux_gcc, If we need to catch all this issue in cross compilation otherwise it will come only on native compilation on that specific board. Now the devtools/test-meson-builds.sh goes over all the config/arm/* on the native compilation On arm64 board and x86. If you need set compiler test for a72 then please add the cross compile config.

Quoted 5 lines

Thanks,
Yongseok

#72 lore raw

Re: [dpdk-dev] [PATCH v8 2/4] meson: add infra to support machine specific flags

Quoted 21 lines · Jerin Jacob Kollanukkaran
Quoted 15 lines

-----Original Message-----
From: Yongseok Koh <[email protected]>
Sent: Friday, April 12, 2019 12:39 PM
To: Jerin Jacob Kollanukkaran <[email protected]>
Cc: Pavan Nikhilesh Bhagavatula <[email protected]>; Thomas Monjalon <[email protected]>; dev <[email protected]>; [email protected]
Subject: [EXT] Re: [dpdk-dev] [PATCH v8 2/4] meson: add infra to support machine specific flags

Quoted 3 lines

One more issue.
With gcc7.2, crypto isn't enabled if -mcpu is set. How about thunderx/octeon?

Quoted 8 lines

It is enabled when when mcpu=thunderxt88 is provided.
In default case, it picks up -march=armv8-a+crc+crypto if mpcu is not matching. So I don’t see any issue.

Quoted 1 line

Looks it should be like -mcpu=cortex-a72+crypto I'll take care of this in a

mcpu should be enough. Looks like like it a compiler bug, Is cortex-a72 available with out crypto? mcpu suppose to tune for specific cpu.

Yes, Mellanox BlueField is designed based on a72 and there are two variants in production. One with crypto, the other w/o crypto. That might be the reason why -mcpu=cortex-a72 doesn't enable crypto by default.

Quoted 10 lines
Quoted 5 lines

separate patch.
Because I want to add a new option to control it.
The reason is a binary having crypto support can't be run on a cpu w/o crypto extension, it is panicked.
And -mcpu=cortex-a72 includes 'crc' support by default.

I suggest to add config/arm/arm64_a72_linux_gcc, If we need to catch all this issue in cross compilation otherwise it will come only on native compilation on that specific board. Now the devtools/test-meson-builds.sh goes over all the config/arm/* on the native compilation On arm64 board and x86. If you need set compiler test for a72 then please add the cross compile config.

Yes, I'm going to add config/arm/arm64_bluefield_linux_gcc soon. Patch is ready now.

Thanks,
Yongseok

#73 lore raw

Re: [dpdk-dev] [PATCH v8 2/4] meson: add infra to support machine specific flags

In reply to [email protected]

10/04/2019 18:13, [email protected]:

Quoted 5 lines

From: Pavan Nikhilesh <[email protected]>

Currently, RTE_* flags are set based on the implementer ID but there might be some micro arch specific differences from the same vendor eg. CACHE_LINESIZE. Add support to set micro arch specific flags.

I don't like how flags are set in config/arm/meson.build. It is a real mess to find which flag applies to which machine. Adding the flags_*_extra in the machine_args_* is adding more mess.

[...]

Quoted 32 lines

flags_common_default = [

config/arm/meson.build +12 -3
 	# Accelarate rte_memcpy. Be sure to run unit test (memcpy_perf_autotest)
 	# to determine the best threshold in code. Refer to notes in source file
@@ -52,12 +33,10 @@ flags_generic = [
 	['RTE_USE_C11_MEM_MODEL', true],
 	['RTE_CACHE_LINE_SIZE', 128]]
 flags_cavium = [
-	['RTE_MACHINE', '"thunderx"'],
 	['RTE_CACHE_LINE_SIZE', 128],
 	['RTE_MAX_NUMA_NODES', 2],
 	['RTE_MAX_LCORE', 96],
-	['RTE_MAX_VFIO_GROUPS', 128],
-	['RTE_USE_C11_MEM_MODEL', false]]
+	['RTE_MAX_VFIO_GROUPS', 128]]
 flags_dpaa = [
 	['RTE_MACHINE', '"dpaa"'],
 	['RTE_USE_C11_MEM_MODEL', true],
@@ -71,6 +50,27 @@ flags_dpaa2 = [
 	['RTE_MAX_NUMA_NODES', 1],
 	['RTE_MAX_LCORE', 16],
 	['RTE_LIBRTE_DPAA2_USE_PHYS_IOVA', false]]
+flags_default_extra = []
+flags_thunderx_extra = [
+	['RTE_MACHINE', '"thunderx"'],
+	['RTE_USE_C11_MEM_MODEL', false]]
+
+machine_args_generic = [
+	['default', ['-march=armv8-a+crc+crypto']],
+	['native', ['-march=native']],
+	['0xd03', ['-mcpu=cortex-a53']],
+	['0xd04', ['-mcpu=cortex-a35']],
+	['0xd07', ['-mcpu=cortex-a57']],

config/arm/meson.build flags_dpaa2 = [ · line 73

Quoted 10 lines
config/arm/meson.build +10 -0
+	['0xd08', ['-mcpu=cortex-a72']],
+	['0xd09', ['-mcpu=cortex-a73']],
+	['0xd0a', ['-mcpu=cortex-a75']]]
+
+machine_args_cavium = [
+	['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
+	['native', ['-march=native']],
+	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
+	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
+	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra]]

I think we should have a simpler model.
We need only to know the machine name and get all the related machine config. In native compilation, machine name is guessed from implementor id and pn (from config/arm/armv8_machine.py). We can directly output the machine name from this script and leave the naming logic in this script.
In the cross-compilation config files (config/arm/*),
we can just specify the machine name.
Then every machine config (machine_args and dpdk_conf) would be specified in some arrays based on the machine name.
Of course, we can keep some common default values.

Thoughts?

#74 lore raw

Re: [dpdk-dev] [PATCH v8 2/4] meson: add infra to support machine specific flags

Quoted 53 lines · Thomas Monjalon

10/04/2019 18:13, [email protected]:

Quoted 5 lines

From: Pavan Nikhilesh <[email protected]>

Currently, RTE_* flags are set based on the implementer ID but there might be some micro arch specific differences from the same vendor eg. CACHE_LINESIZE. Add support to set micro arch specific flags.

I don't like how flags are set in config/arm/meson.build. It is a real mess to find which flag applies to which machine. Adding the flags_*_extra in the machine_args_* is adding more mess.

[...]

Quoted 38 lines

flags_common_default = [
# Accelarate rte_memcpy. Be sure to run unit test (memcpy_perf_autotest) # to determine the best threshold in code. Refer to notes in source file

config/arm/meson.build +0 -0
@@ -52,12 +33,10 @@ flags_generic = [
 	['RTE_USE_C11_MEM_MODEL', true],
 	['RTE_CACHE_LINE_SIZE', 128]]

flags_cavium = [

config/arm/meson.build +0 -1
-	['RTE_MACHINE', '"thunderx"'],
	['RTE_CACHE_LINE_SIZE', 128],
	['RTE_MAX_NUMA_NODES', 2],
	['RTE_MAX_LCORE', 96],
config/arm/meson.build +1 -2
-	['RTE_MAX_VFIO_GROUPS', 128],
-	['RTE_USE_C11_MEM_MODEL', false]]
+	['RTE_MAX_VFIO_GROUPS', 128]]

flags_dpaa = [

	['RTE_MACHINE', '"dpaa"'],
	['RTE_USE_C11_MEM_MODEL', true],
config/arm/meson.build +17 -0
@@ -71,6 +50,27 @@ flags_dpaa2 = [
 	['RTE_MAX_NUMA_NODES', 1],
 	['RTE_MAX_LCORE', 16],
 	['RTE_LIBRTE_DPAA2_USE_PHYS_IOVA', false]]
+flags_default_extra = []
+flags_thunderx_extra = [
+	['RTE_MACHINE', '"thunderx"'],
+	['RTE_USE_C11_MEM_MODEL', false]]
+
+machine_args_generic = [
+	['default', ['-march=armv8-a+crc+crypto']],
+	['native', ['-march=native']],
+	['0xd03', ['-mcpu=cortex-a53']],
+	['0xd04', ['-mcpu=cortex-a35']],
+	['0xd07', ['-mcpu=cortex-a57']],
+	['0xd08', ['-mcpu=cortex-a72']],
+	['0xd09', ['-mcpu=cortex-a73']],
+	['0xd0a', ['-mcpu=cortex-a75']]]
+
+machine_args_cavium = [
+	['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
Quoted 8 lines

config/arm/meson.build flags_dpaa2 = [ · line 73

Quoted 4 lines
config/arm/meson.build +4 -0
+	['native', ['-march=native']],
+	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
+	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
+	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra]]

I think we should have a simpler model.
We need only to know the machine name and get all the related machine config. In native compilation, machine name is guessed from implementor id and pn

It can be guessed unless machine name is specified as an option. We can add machine_name option?
One observation. We do have machine option but it isn't used much for arm. And in config/arm/meson.build, march_opt isn't used either.

Quoted 9 lines

(from config/arm/armv8_machine.py). We can directly output the machine name from this script and leave the naming logic in this script.
In the cross-compilation config files (config/arm/*),
we can just specify the machine name.
Then every machine config (machine_args and dpdk_conf) would be specified in some arrays based on the machine name.
Of course, we can keep some common default values.

Thoughts?

Sounds like we can cleanly reorganize it with the suggestion.

Thanks,
Yongseok

#75 lore raw

Re: [dpdk-dev] [PATCH v8 2/4] meson: add infra to support machine specific flags

In reply to Thomas Monjalon
Quoted 69 lines

-----Original Message-----
From: Thomas Monjalon <[email protected]>
Sent: Friday, April 12, 2019 5:07 AM
To: Pavan Nikhilesh Bhagavatula <[email protected]>; Jerin Jacob Kollanukkaran <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]
Subject: Re: [dpdk-dev] [PATCH v8 2/4] meson: add infra to support machine specific flags

10/04/2019 18:13, [email protected]:

Quoted 5 lines

From: Pavan Nikhilesh <[email protected]>

Currently, RTE_* flags are set based on the implementer ID but there might be some micro arch specific differences from the same vendor eg. CACHE_LINESIZE. Add support to set micro arch specific flags.

I don't like how flags are set in config/arm/meson.build. It is a real mess to find which flag applies to which machine. Adding the flags_*_extra in the machine_args_* is adding more mess.

[...]

Quoted 2 lines

flags_common_default = [
# Accelarate rte_memcpy. Be sure to run unit test

(memcpy_perf_autotest)

Quoted 30 lines
 	# to determine the best threshold in code. Refer to notes in source
file @@ -52,12 +33,10 @@ flags_generic = [
config/arm/meson.build +12 -3
 	['RTE_USE_C11_MEM_MODEL', true],
 	['RTE_CACHE_LINE_SIZE', 128]]
 flags_cavium = [
-	['RTE_MACHINE', '"thunderx"'],
 	['RTE_CACHE_LINE_SIZE', 128],
 	['RTE_MAX_NUMA_NODES', 2],
 	['RTE_MAX_LCORE', 96],
-	['RTE_MAX_VFIO_GROUPS', 128],
-	['RTE_USE_C11_MEM_MODEL', false]]
+	['RTE_MAX_VFIO_GROUPS', 128]]
 flags_dpaa = [
 	['RTE_MACHINE', '"dpaa"'],
 	['RTE_USE_C11_MEM_MODEL', true],
@@ -71,6 +50,27 @@ flags_dpaa2 = [
 	['RTE_MAX_NUMA_NODES', 1],
 	['RTE_MAX_LCORE', 16],
 	['RTE_LIBRTE_DPAA2_USE_PHYS_IOVA', false]]
+flags_default_extra = []
+flags_thunderx_extra = [
+	['RTE_MACHINE', '"thunderx"'],
+	['RTE_USE_C11_MEM_MODEL', false]]
+
+machine_args_generic = [
+	['default', ['-march=armv8-a+crc+crypto']],
+	['native', ['-march=native']],
+	['0xd03', ['-mcpu=cortex-a53']],
+	['0xd04', ['-mcpu=cortex-a35']],
+	['0xd07', ['-mcpu=cortex-a57']],

config/arm/meson.build flags_dpaa2 = [ · line 73

Quoted 10 lines
config/arm/meson.build +10 -0
+	['0xd08', ['-mcpu=cortex-a72']],
+	['0xd09', ['-mcpu=cortex-a73']],
+	['0xd0a', ['-mcpu=cortex-a75']]]
+
+machine_args_cavium = [
+	['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
+	['native', ['-march=native']],
+	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
+	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
+	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra]]

I think we should have a simpler model.
We need only to know the machine name and get all the related machine config.

Quoted 8 lines

In native compilation, machine name is guessed from implementor id and pn (from config/arm/armv8_machine.py). We can directly output the machine name from this script and leave the naming logic in this script. In the cross-compilation config files (config/arm/*), we can just specify the machine name.
Then every machine config (machine_args and dpdk_conf) would be specified in some arrays based on the machine name.
Of course, we can keep some common default values.

Thomas,

This patch was around last three months. It reached upto v8.
I think, in that last minute for RC2, We cannot take major rework on this as it needs to tested for Other arm64 platform too. It was pulled out from RC1 because other pcap issue from meson. Now its not fair to say to rework the meson stuff now.
I suggest to take other rework in next release.

Quoted 3 lines

Thoughts?

#76 lore raw

Re: [dpdk-dev] [PATCH v8 2/4] meson: add infra to support machine specific flags

12/04/2019 09:12, Jerin Jacob Kollanukkaran:

Quoted 69 lines

From: Thomas Monjalon <[email protected]>

Quoted 59 lines

10/04/2019 18:13, [email protected]:

Quoted 5 lines

From: Pavan Nikhilesh <[email protected]>

Currently, RTE_* flags are set based on the implementer ID but there might be some micro arch specific differences from the same vendor eg. CACHE_LINESIZE. Add support to set micro arch specific flags.

I don't like how flags are set in config/arm/meson.build. It is a real mess to find which flag applies to which machine. Adding the flags_*_extra in the machine_args_* is adding more mess.

[...]

Quoted 2 lines

flags_common_default = [
# Accelarate rte_memcpy. Be sure to run unit test

(memcpy_perf_autotest)

Quoted 30 lines
 	# to determine the best threshold in code. Refer to notes in source
file @@ -52,12 +33,10 @@ flags_generic = [
config/arm/meson.build +12 -3
 	['RTE_USE_C11_MEM_MODEL', true],
 	['RTE_CACHE_LINE_SIZE', 128]]
 flags_cavium = [
-	['RTE_MACHINE', '"thunderx"'],
 	['RTE_CACHE_LINE_SIZE', 128],
 	['RTE_MAX_NUMA_NODES', 2],
 	['RTE_MAX_LCORE', 96],
-	['RTE_MAX_VFIO_GROUPS', 128],
-	['RTE_USE_C11_MEM_MODEL', false]]
+	['RTE_MAX_VFIO_GROUPS', 128]]
 flags_dpaa = [
 	['RTE_MACHINE', '"dpaa"'],
 	['RTE_USE_C11_MEM_MODEL', true],
@@ -71,6 +50,27 @@ flags_dpaa2 = [
 	['RTE_MAX_NUMA_NODES', 1],
 	['RTE_MAX_LCORE', 16],
 	['RTE_LIBRTE_DPAA2_USE_PHYS_IOVA', false]]
+flags_default_extra = []
+flags_thunderx_extra = [
+	['RTE_MACHINE', '"thunderx"'],
+	['RTE_USE_C11_MEM_MODEL', false]]
+
+machine_args_generic = [
+	['default', ['-march=armv8-a+crc+crypto']],
+	['native', ['-march=native']],
+	['0xd03', ['-mcpu=cortex-a53']],
+	['0xd04', ['-mcpu=cortex-a35']],
+	['0xd07', ['-mcpu=cortex-a57']],

config/arm/meson.build flags_dpaa2 = [ · line 73

Quoted 10 lines
config/arm/meson.build +10 -0
+	['0xd08', ['-mcpu=cortex-a72']],
+	['0xd09', ['-mcpu=cortex-a73']],
+	['0xd0a', ['-mcpu=cortex-a75']]]
+
+machine_args_cavium = [
+	['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
+	['native', ['-march=native']],
+	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
+	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
+	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra]]

I think we should have a simpler model.
We need only to know the machine name and get all the related machine config.

Quoted 8 lines

In native compilation, machine name is guessed from implementor id and pn (from config/arm/armv8_machine.py). We can directly output the machine name from this script and leave the naming logic in this script. In the cross-compilation config files (config/arm/*), we can just specify the machine name.
Then every machine config (machine_args and dpdk_conf) would be specified in some arrays based on the machine name.
Of course, we can keep some common default values.

Quoted 7 lines

Thomas,

This patch was around last three months. It reached upto v8.
I think, in that last minute for RC2, We cannot take major rework on this as it needs to tested for Other arm64 platform too. It was pulled out from RC1 because other pcap issue from meson. Now its not fair to say to rework the meson stuff now.
I suggest to take other rework in next release.

I was not confortable with this patch without being able to say why. Yesterday I spent more time to understand and see what may be improved. I agree it is late, so it won't block this patch for 19.05. Do you agree this file can be improved?
Please would you like to look at reworking during next cycle? Thanks

#77 lore raw

Re: [dpdk-dev] [PATCH v8 2/4] meson: add infra to support machine specific flags

Quoted 24 lines
Quoted 19 lines
Quoted 10 lines

config/arm/meson.build flags_dpaa2 = [ · line 73

Quoted 6 lines
config/arm/meson.build +6 -0
+machine_args_cavium = [
+	['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
+	['native', ['-march=native']],
+	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
+	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
+	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra]]

I think we should have a simpler model.
We need only to know the machine name and get all the related machine config.

Quoted 8 lines

In native compilation, machine name is guessed from implementor id and pn (from config/arm/armv8_machine.py). We can directly output the machine name from this script and leave the naming logic in this script. In the cross-compilation config files (config/arm/*), we can just specify the machine name.
Then every machine config (machine_args and dpdk_conf) would be specified in some arrays based on the machine name.
Of course, we can keep some common default values.

Quoted 5 lines

Thomas,

This patch was around last three months. It reached upto v8.
I think, in that last minute for RC2, We cannot take major rework on this as it needs to tested for Other arm64 platform too. It was pulled out from

Quoted 8 lines

RC1 because other pcap issue from meson.

Quoted 2 lines

Now its not fair to say to rework the meson stuff now. I suggest to take other rework in next release.

I was not confortable with this patch without being able to say why. Yesterday I spent more time to understand and see what may be improved. I agree it is late, so it won't block this patch for 19.05. Do you agree this file can be improved?

Moving to the all to static config file is an option but we lose the flexibility
of runtime detecting the options and few of them are probing at runtime based
on gcc versions and mcpu combination etc.
I am not expert in meson area and not sure meson/python has better data strcture for this other than list/array combo. If Bruce has any feedback on this, then we
will try to prototype it.

Quoted 3 lines

Please would you like to look at reworking during next cycle? Thanks

#78 lore raw

Re: [dpdk-dev] [PATCH v8 2/4] meson: add infra to support machine specific flags

13/04/2019 08:24, Jerin Jacob Kollanukkaran:

Quoted 8 lines
Quoted 4 lines

I was not confortable with this patch without being able to say why. Yesterday I spent more time to understand and see what may be improved. I agree it is late, so it won't block this patch for 19.05. Do you agree this file can be improved?

Moving to the all to static config file is an option but we lose the flexibility of runtime detecting the options and few of them are probing at runtime based on gcc versions and mcpu combination etc.

I think there is a misunderstanding.
I'm suggesting to symplify arrays by indexing only by machine name. It should not change the behaviour.

Quoted 6 lines

I am not expert in meson area and not sure meson/python has better data strcture for this other than list/array combo. If Bruce has any feedback on this, then we
will try to prototype it.

Quoted 2 lines

Please would you like to look at reworking during next cycle? Thanks

#79 lore raw

Re: [dpdk-dev] [PATCH v8 2/4] meson: add infra to support machine specific flags

Hi Thomas,

There is no guarantee of primary part number (machine names) uniqueness between implementors.
If we limit lookups to only machine names through primary part number we would have a lot of repetitive defines. Also, moving the arrays into the python script is not feasible as meson needs to reparse the standard out from the python script

Currently, config is split into three parts :
1. Implementor specific defines.
2. Micro-arch specific compiler flags.
3. Micro-arch specific defines.

I think from a configurability point of view the above three are really important for fine grained control.

Thoughts?

Regards,
Pavan.

Quoted 32 lines

-----Original Message-----
From: Thomas Monjalon <[email protected]>
Sent: Sunday, April 14, 2019 2:13 AM
To: Jerin Jacob Kollanukkaran <[email protected]>
Cc: Pavan Nikhilesh Bhagavatula <[email protected]>; [email protected]; [email protected]; [email protected]; [email protected]
Subject: Re: [dpdk-dev] [PATCH v8 2/4] meson: add infra to support machine specific flags

13/04/2019 08:24, Jerin Jacob Kollanukkaran:

Quoted 8 lines
Quoted 4 lines

I was not confortable with this patch without being able to say why. Yesterday I spent more time to understand and see what may be improved. I agree it is late, so it won't block this patch for 19.05. Do you agree this file can be improved?

Moving to the all to static config file is an option but we lose the flexibility of runtime detecting the options and few of them are probing at runtime based on gcc versions and mcpu combination etc.

I think there is a misunderstanding.
I'm suggesting to symplify arrays by indexing only by machine name. It should not change the behaviour.

Quoted 6 lines

I am not expert in meson area and not sure meson/python has better data strcture for this other than list/array combo. If Bruce has any feedback on this, then we will try to prototype it.

Quoted 2 lines

Please would you like to look at reworking during next cycle? Thanks

#80 lore raw

Re: [dpdk-dev] [PATCH v8 2/4] meson: add infra to support machine specific flags

14/04/2019 16:40, Pavan Nikhilesh Bhagavatula:

Quoted 3 lines

Hi Thomas,

There is no guarantee of primary part number (machine names) uniqueness between implementors.

I think we don't speak the same language :)
By machine name, I mean what we set in RTE_MACHINE, like octeontx2.

Quoted 2 lines

If we limit lookups to only machine names through primary part number we would have a lot of repetitive defines. Also, moving the arrays into the python script is not feasible as meson needs to reparse the standard out from the python script

I will probably need to write a PoC.

Quoted 4 lines

Currently, config is split into three parts :
1. Implementor specific defines.
2. Micro-arch specific compiler flags.
3. Micro-arch specific defines.

This is currently unreadable in my opinion.

Quoted 1 line

I think from a configurability point of view the above three are really important for fine grained control.

I agree fine grain is required.

Quoted 39 lines

Thoughts?

Regards,
Pavan.

Quoted 32 lines

-----Original Message-----
From: Thomas Monjalon <[email protected]>
Sent: Sunday, April 14, 2019 2:13 AM
To: Jerin Jacob Kollanukkaran <[email protected]>
Cc: Pavan Nikhilesh Bhagavatula <[email protected]>; [email protected]; [email protected]; [email protected]; [email protected]
Subject: Re: [dpdk-dev] [PATCH v8 2/4] meson: add infra to support machine specific flags

13/04/2019 08:24, Jerin Jacob Kollanukkaran:

Quoted 8 lines
Quoted 4 lines

I was not confortable with this patch without being able to say why. Yesterday I spent more time to understand and see what may be improved. I agree it is late, so it won't block this patch for 19.05. Do you agree this file can be improved?

Moving to the all to static config file is an option but we lose the flexibility of runtime detecting the options and few of them are probing at runtime based on gcc versions and mcpu combination etc.

I think there is a misunderstanding.
I'm suggesting to symplify arrays by indexing only by machine name. It should not change the behaviour.

Quoted 6 lines

I am not expert in meson area and not sure meson/python has better data strcture for this other than list/array combo. If Bruce has any feedback on this, then we will try to prototype it.

Quoted 2 lines

Please would you like to look at reworking during next cycle? Thanks

#81 lore raw

Re: [dpdk-dev] [PATCH v8 2/4] meson: add infra to support machine specific flags

Quoted 9 lines

-----Original Message-----
From: Thomas Monjalon <[email protected]>
Sent: Sunday, April 14, 2019 11:14 PM
To: Pavan Nikhilesh Bhagavatula <[email protected]>
Cc: Jerin Jacob Kollanukkaran <[email protected]>; [email protected]; [email protected]; [email protected]; [email protected]
Subject: Re: [dpdk-dev] [PATCH v8 2/4] meson: add infra to support machine specific flags

Quoted 8 lines

14/04/2019 16:40, Pavan Nikhilesh Bhagavatula:

Quoted 3 lines

Hi Thomas,

There is no guarantee of primary part number (machine names) uniqueness

between implementors.

I think we don't speak the same language :) By machine name, I mean what we set in RTE_MACHINE, like octeontx2.

As you know, The system probes "implementor_id" and "implementor_pn" Values. There is nothing like machine name in meson and in order to keep Synergy with native build, we need to just follow, "implementor_id" and "implementor_pn". Now, it is possible to have "implemetor_id" to "implementor_pn" to machine name lookup but Unlike, "make" based Build system, meson supports supporting a lot of machines(like RTE_MACHINE), with that structure. So converting to another intermediate called "machine string" will have more overhead IMO.

['0xa1', ['-mcpu=thunderxt88']],
['0xa2', ['-mcpu=thunderxt81']],
['0xa3', ['-mcpu=thunderxt83']]]
['0xd03', ['-mcpu=cortex-a53']],
['0xd04', ['-mcpu=cortex-a35']],
['0xd05', ['-mcpu=cortex-a55']],
['0xd07', ['-mcpu=cortex-a57']],
['0xd08', ['-mcpu=cortex-a72']],
['0xd09', ['-mcpu=cortex-a73']],
['0xd0a', ['-mcpu=cortex-a75']],
['0xd0b', ['-mcpu=cortex-a76']]
impl_0x41 = ['Arm', flags_generic, machine_args_generic] impl_0x42 = ['Broadcom', flags_generic, machine_args_generic] impl_0x43 = ['Cavium', flags_cavium, machine_args_cavium] impl_0x44 = ['DEC', flags_generic, machine_args_generic] impl_0x49 = ['Infineon', flags_generic, machine_args_generic] impl_0x4d = ['Motorola', flags_generic, machine_args_generic] impl_0x4e = ['NVIDIA', flags_generic, machine_args_generic] impl_0x50 = ['AppliedMicro', flags_generic, machine_args_generic] impl_0x51 = ['Qualcomm', flags_generic, machine_args_generic] impl_0x53 = ['Samsung', flags_generic, machine_args_generic] impl_0x56 = ['Marvell', flags_generic, machine_args_generic] impl_0x69 = ['Intel', flags_generic, machine_args_generic] impl_dpaa = ['NXP DPAA', flags_dpaa, machine_args_generic] impl_dpaa2 = ['NXP DPAA2', flags_dpaa2, machine_args_generic]

Quoted 7 lines
Quoted 2 lines

If we limit lookups to only machine names through primary part number we

would have a lot of repetitive defines.

Quoted 2 lines

Also, moving the arrays into the python script is not feasible as meson needs to reparse the standard out from the python script

I will probably need to write a PoC.

Yes. Please

Quoted 7 lines
Quoted 5 lines

Currently, config is split into three parts :
1. Implementor specific defines.
2. Micro-arch specific compiler flags.
3. Micro-arch specific defines.

This is currently unreadable in my opinion.

Bit subjective.
If we want to keep all the fine grained control along with "native" /"cross" build, "distribution" build, cache line differences etc, makes it bit difficult.

But if you think, it can be improved. Please share patch, We are happy to review and test in the platforms we have.

Quoted 50 lines
Quoted 1 line

I think from a configurability point of view the above three are really

important for fine grained control.

I agree fine grain is required.

Quoted 18 lines

Thoughts?

Regards,
Pavan.

Quoted 13 lines

-----Original Message-----
From: Thomas Monjalon <[email protected]>
Sent: Sunday, April 14, 2019 2:13 AM
To: Jerin Jacob Kollanukkaran <[email protected]>
Cc: Pavan Nikhilesh Bhagavatula <[email protected]>; [email protected]; [email protected]; [email protected]; [email protected]
Subject: Re: [dpdk-dev] [PATCH v8 2/4] meson: add infra to support machine specific flags

13/04/2019 08:24, Jerin Jacob Kollanukkaran:

Quoted 2 lines
Quoted 2 lines

I was not confortable with this patch without being able to say why. Yesterday I spent more time to understand and see what may be

improved.

Quoted 21 lines
Quoted 19 lines
Quoted 6 lines
Quoted 2 lines

I agree it is late, so it won't block this patch for 19.05. Do you agree this file can be improved?

Moving to the all to static config file is an option but we lose the flexibility of runtime detecting the options and few of them are probing at runtime based on gcc versions and mcpu combination etc.

I think there is a misunderstanding.
I'm suggesting to symplify arrays by indexing only by machine name. It should not change the behaviour.

Quoted 6 lines

I am not expert in meson area and not sure meson/python has better data strcture for this other than list/array combo. If Bruce has any feedback on this, then we will try to prototype it.

Quoted 2 lines

Please would you like to look at reworking during next cycle? Thanks

#82 lore raw

Re: [dpdk-dev] [PATCH v8 2/4] meson: add infra to support machine specific flags

14/04/2019 20:19, Jerin Jacob Kollanukkaran:

Quoted 10 lines

From: Thomas Monjalon <[email protected]>

Quoted 8 lines

14/04/2019 16:40, Pavan Nikhilesh Bhagavatula:

Quoted 3 lines

Hi Thomas,

There is no guarantee of primary part number (machine names) uniqueness

between implementors.

I think we don't speak the same language :) By machine name, I mean what we set in RTE_MACHINE, like octeontx2.

Quoted 8 lines

As you know, The system probes "implementor_id" and "implementor_pn" Values. There is nothing like machine name in meson and in order to keep Synergy with native build, we need to just follow, "implementor_id" and "implementor_pn". Now, it is possible to have "implemetor_id" to "implementor_pn" to machine name lookup but Unlike, "make" based Build system, meson supports supporting a lot of machines(like RTE_MACHINE), with that structure. So converting to another intermediate called "machine string" will have more overhead IMO.

We already have this string with RTE_MACHINE. You already set RTE_MACHINE based on id and pn. I don't see any overhead.
Anyway, no need to discuss it more without any real code.

#83 lore raw

PATCH v8 3/4 config: add thunderx2 machine config

In reply to [email protected]

From: Jerin Jacob <[email protected]>

Optimized configuration for Marvell thunderx2 SoC. Updated meson build to support Marvell thunderx2 SoC. Added meson cross compile target.

Product details are here:

https://www.marvell.com/server-processors/thunderx2-arm-processors/

---
 config/arm/arm64_thunderx2_linux_gcc          | 16 +++++++++
 config/arm/meson.build                        |  9 ++++-
 config/defconfig_arm64-thunderx2-linux-gcc    |  1 +
 config/defconfig_arm64-thunderx2-linuxapp-gcc | 11 ++++++
 mk/machine/thunderx2/rte.vars.mk              | 34 +++++++++++++++++++
 5 files changed, 70 insertions(+), 1 deletion(-)
 create mode 100644 config/arm/arm64_thunderx2_linux_gcc
 create mode 120000 config/defconfig_arm64-thunderx2-linux-gcc
 create mode 100644 config/defconfig_arm64-thunderx2-linuxapp-gcc
 create mode 100644 mk/machine/thunderx2/rte.vars.mk
config/arm/arm64_thunderx2_linux_gcc new file +16 -0
new file mode 100644
index 000000000..0dc275644
@@ -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 = '0x43'
+implementor_pn = '0xaf'
config/arm/meson.build +8 -1
index 24bce2b39..3220b584f 100644
@@ -54,6 +54,12 @@ flags_default_extra = []
 flags_thunderx_extra = [
 	['RTE_MACHINE', '"thunderx"'],
 	['RTE_USE_C11_MEM_MODEL', false]]
+flags_thunderx2_extra = [
+	['RTE_MACHINE', '"thunderx2"'],
+	['RTE_CACHE_LINE_SIZE', 64],
+	['RTE_MAX_NUMA_NODES', 2],
+	['RTE_MAX_LCORE', 256],
+	['RTE_USE_C11_MEM_MODEL', true]]
 
 machine_args_generic = [
 	['default', ['-march=armv8-a+crc+crypto']],
@@ -70,7 +76,8 @@ machine_args_cavium = [
 	['native', ['-march=native']],
 	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
 	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
-	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra]]
+	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra],
+	['0xaf', ['-mcpu=thunderx2t99'], flags_thunderx2_extra]]
 
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page G7-5321)
 impl_generic = ['Generic armv8', flags_generic, machine_args_generic]
config/defconfig_arm64-thunderx2-linux-gcc new file +1 -0
new file mode 120000
index 000000000..b40a760b1
@@ -0,0 +1 @@
+defconfig_arm64-thunderx2-linuxapp-gcc
\ No newline at end of file
config/defconfig_arm64-thunderx2-linuxapp-gcc new file +11 -0
new file mode 100644
index 000000000..cc5c64ba0
@@ -0,0 +1,11 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International Ltd
+#
+
+#include "defconfig_arm64-armv8a-linux-gcc"
+
+CONFIG_RTE_MACHINE="thunderx2"
+
+CONFIG_RTE_CACHE_LINE_SIZE=64
+CONFIG_RTE_MAX_NUMA_NODES=2
+CONFIG_RTE_MAX_LCORE=256
mk/machine/thunderx2/rte.vars.mk new file +34 -0
new file mode 100644
index 000000000..b80dc8680
@@ -0,0 +1,34 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International 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=armv8.1-a+crc+crypto)
+MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=thunderx2t99)
Signature · 1 line
2.21.0
#84 lore raw

PATCH v8 4/4 config: add octeontx2 machine config

In reply to [email protected]

From: Jerin Jacob <[email protected]>

Optimized configuration for Marvell octeontx2 SoC. Updated meson build to support Marvell octeontx2 SoC. Added meson cross build target for octeontx2.

---
 config/arm/arm64_octeontx2_linux_gcc          | 16 +++++++++
 config/arm/meson.build                        |  9 ++++-
 config/defconfig_arm64-octeontx2-linux-gcc    |  1 +
 config/defconfig_arm64-octeontx2-linuxapp-gcc | 18 ++++++++++
 mk/machine/octeontx2/rte.vars.mk              | 34 +++++++++++++++++++
 5 files changed, 77 insertions(+), 1 deletion(-)
 create mode 100644 config/arm/arm64_octeontx2_linux_gcc
 create mode 120000 config/defconfig_arm64-octeontx2-linux-gcc
 create mode 100644 config/defconfig_arm64-octeontx2-linuxapp-gcc
 create mode 100644 mk/machine/octeontx2/rte.vars.mk
config/arm/arm64_octeontx2_linux_gcc new file +16 -0
new file mode 100644
index 000000000..e2c0b8f72
@@ -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 = '0x43'
+implementor_pn = '0xb2'
config/arm/meson.build +8 -1
index 3220b584f..0708bc64a 100644
@@ -60,6 +60,12 @@ flags_thunderx2_extra = [
 	['RTE_MAX_NUMA_NODES', 2],
 	['RTE_MAX_LCORE', 256],
 	['RTE_USE_C11_MEM_MODEL', true]]
+flags_octeontx2_extra = [
+	['RTE_MACHINE', '"octeontx2"'],
+	['RTE_MAX_NUMA_NODES', 1],
+	['RTE_MAX_LCORE', 24],
+	['RTE_EAL_IGB_UIO', false],
+	['RTE_USE_C11_MEM_MODEL', true]]
 
 machine_args_generic = [
 	['default', ['-march=armv8-a+crc+crypto']],
@@ -77,7 +83,8 @@ machine_args_cavium = [
 	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
 	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
 	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra],
-	['0xaf', ['-mcpu=thunderx2t99'], flags_thunderx2_extra]]
+	['0xaf', ['-mcpu=thunderx2t99'], flags_thunderx2_extra],
+	['0xb2', ['-mcpu=octeontx2'], flags_octeontx2_extra]]
 
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page G7-5321)
 impl_generic = ['Generic armv8', flags_generic, machine_args_generic]
config/defconfig_arm64-octeontx2-linux-gcc new file +1 -0
new file mode 120000
index 000000000..e25150531
@@ -0,0 +1 @@
+defconfig_arm64-octeontx2-linuxapp-gcc
\ No newline at end of file
config/defconfig_arm64-octeontx2-linuxapp-gcc new file +18 -0
new file mode 100644
index 000000000..9eae84538
@@ -0,0 +1,18 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International Ltd
+#
+
+#include "defconfig_arm64-armv8a-linux-gcc"
+
+CONFIG_RTE_MACHINE="octeontx2"
+
+CONFIG_RTE_CACHE_LINE_SIZE=128
+CONFIG_RTE_MAX_NUMA_NODES=1
+CONFIG_RTE_MAX_LCORE=24
+
+# Doesn't support NUMA
+CONFIG_RTE_EAL_NUMA_AWARE_HUGEPAGES=n
+CONFIG_RTE_LIBRTE_VHOST_NUMA=n
+
+# Recommend to use VFIO as co-processors needs SMMU/IOMMU
+CONFIG_RTE_EAL_IGB_UIO=n
mk/machine/octeontx2/rte.vars.mk new file +34 -0
new file mode 100644
index 000000000..cbec7f14d
@@ -0,0 +1,34 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International 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=-mcpu=armv8.2-a+crc+crypto+lse)
+MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=octeontx2)
Signature · 1 line
2.21.0
#85 lore raw

PATCH v9 1/4 mk: introduce helper to check valid compiler argument

In reply to [email protected]

From: Jerin Jacob <[email protected]>

Introduce rte_cc_has_argument() Makefile helper to check a given argument is support by the compiler.

Example Usage:

include $(RTE_SDK)/mk/rte.helper.mk
MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=octeontx2)

This would allow adding -mcpu=octeontx2 in MACHINE_CFLAGS if it is only supported by the compiler. The use case for such scheme is to enable the mcpu optimization if the compiler supports else it needs to compile the source code without any errors.

This patch also moves inclusion of toolchain's rte.vars.mk to before the machine's rte.vars.mk inclusion to make correct CC available for the cross compile case.

---
Change history of this series:

v9 Changes:
- Remove compiler version check as it is now done using cc.has_argument().

v8 Changes:
- Remove redudant lists (rebase aritfacts). (Yongseok Koh)

v7 Changes:
- Updated cross compile config files align with
"build: improve pcap dependency handling" changeset to fix build issue with meson
- Some compiler needs the following depended patch to compile with meson
http://patches.dpdk.org/patch/52367/

v6 Changes:
- Rework to change the config files to sync with "mk: use linux and freebsd in config names"
- Fix the following error with latest gcc by fixing the mcpu type
cc1: error: switch -mcpu=armv8.2-a conflicts with -march=armv8-a switch

v5 Changes:
- Fix incorrect meson flag parsing(Phil Yang)
- Squash meson cross build patch(5/5) into configuration update patches for thunderx2(3/5) and octeontx2(4/5)(Thomas)
- Changed octeontx2's march as armv8-a and added the extension required instead of armv8-2a(Phil Yang)
- Improved rte_cc_has_argument() implementaion by removing the temp file(Thomas)

v4 Changes:
- Fix incorrect signoff marrvell -> marvell.

v3 Changes:
- Squash meson build support into config support for thunderx2/octeontx2.

v2 Changes:
- Add meson build support.

 mk/rte.helper.mk              | 10 ++++++++++
 mk/target/generic/rte.vars.mk | 22 +++++++++++-----------
 2 files changed, 21 insertions(+), 11 deletions(-)
 create mode 100644 mk/rte.helper.mk
mk/rte.helper.mk new file +10 -0
new file mode 100644
index 000000000..6e7fd03d7
@@ -0,0 +1,10 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International Ltd
+
+# rte_cc_has_argument
+# Usage: MACHINE_CFLAGS += $(call rte_cc_has_argument, -mno-avx512f)
+# Return the argument if the argument is supported by the compiler.
+#
+define rte_cc_has_argument
+	$(shell $(CC) -E $(1) -xc /dev/null 1>/dev/null 2>/dev/null && echo $(1))
+endef
mk/target/generic/rte.vars.mk +11 -11
index dd149acc9..25a578ad7 100644
@@ -7,6 +7,17 @@
 # executive environment.
 #
 
+#
+# toolchain:
+#
+#   - define CC, LD, AR, AS, ...
+#   - define TOOLCHAIN_CFLAGS variable (overridden by cmdline value)
+#   - define TOOLCHAIN_LDFLAGS variable (overridden by cmdline value)
+#   - define TOOLCHAIN_ASFLAGS variable (overridden by cmdline value)
+#   - may override any previously defined variable
+#
+include $(RTE_SDK)/mk/toolchain/$(RTE_TOOLCHAIN)/rte.vars.mk
+
 #
 # machine:
 #
@@ -45,17 +56,6 @@ endif
 #
 include $(RTE_SDK)/mk/arch/$(RTE_ARCH)/rte.vars.mk
 
-#
-# toolchain:
-#
-#   - define CC, LD, AR, AS, ...
-#   - define TOOLCHAIN_CFLAGS variable (overridden by cmdline value)
-#   - define TOOLCHAIN_LDFLAGS variable (overridden by cmdline value)
-#   - define TOOLCHAIN_ASFLAGS variable (overridden by cmdline value)
-#   - may override any previously defined variable
-#
-include $(RTE_SDK)/mk/toolchain/$(RTE_TOOLCHAIN)/rte.vars.mk
-
 #
 # exec-env:
 #
Signature · 1 line
2.21.0
#86 lore raw

PATCH v9 2/4 meson: add infra to support machine specific flags

From: Pavan Nikhilesh <[email protected]>

Currently, RTE_* flags are set based on the implementer ID but there might be some micro arch specific differences from the same vendor eg. CACHE_LINESIZE. Add support to set micro arch specific flags.

---
 config/arm/meson.build | 67 +++++++++++++++++++++---------------------
 1 file changed, 33 insertions(+), 34 deletions(-)
config/arm/meson.build +33 -34
index 170a4981a..d30a17426 100644
@@ -7,25 +7,6 @@ march_opt = '-march=@0@'.format(machine)
 
 arm_force_native_march = false
 
-machine_args_generic = [
-	['default', ['-march=armv8-a+crc+crypto']],
-	['native', ['-march=native']],
-	['0xd03', ['-mcpu=cortex-a53']],
-	['0xd04', ['-mcpu=cortex-a35']],
-	['0xd05', ['-mcpu=cortex-a55']],
-	['0xd07', ['-mcpu=cortex-a57']],
-	['0xd08', ['-mcpu=cortex-a72']],
-	['0xd09', ['-mcpu=cortex-a73']],
-	['0xd0a', ['-mcpu=cortex-a75']],
-	['0xd0b', ['-mcpu=cortex-a76']],
-]
-machine_args_cavium = [
-	['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
-	['native', ['-march=native']],
-	['0xa1', ['-mcpu=thunderxt88']],
-	['0xa2', ['-mcpu=thunderxt81']],
-	['0xa3', ['-mcpu=thunderxt83']]]
-
 flags_common_default = [
 	# Accelarate rte_memcpy. Be sure to run unit test (memcpy_perf_autotest)
 	# to determine the best threshold in code. Refer to notes in source file
@@ -52,12 +33,10 @@ flags_generic = [
 	['RTE_USE_C11_MEM_MODEL', true],
 	['RTE_CACHE_LINE_SIZE', 128]]
 flags_cavium = [
-	['RTE_MACHINE', '"thunderx"'],
 	['RTE_CACHE_LINE_SIZE', 128],
 	['RTE_MAX_NUMA_NODES', 2],
 	['RTE_MAX_LCORE', 96],
-	['RTE_MAX_VFIO_GROUPS', 128],
-	['RTE_USE_C11_MEM_MODEL', false]]
+	['RTE_MAX_VFIO_GROUPS', 128]]
 flags_dpaa = [
 	['RTE_MACHINE', '"dpaa"'],
 	['RTE_USE_C11_MEM_MODEL', true],
@@ -71,6 +50,27 @@ flags_dpaa2 = [
 	['RTE_MAX_NUMA_NODES', 1],
 	['RTE_MAX_LCORE', 16],
 	['RTE_LIBRTE_DPAA2_USE_PHYS_IOVA', false]]
+flags_default_extra = []
+flags_thunderx_extra = [
+	['RTE_MACHINE', '"thunderx"'],
+	['RTE_USE_C11_MEM_MODEL', false]]
+
+machine_args_generic = [
+	['default', ['-march=armv8-a+crc+crypto']],
+	['native', ['-march=native']],
+	['0xd03', ['-mcpu=cortex-a53']],
+	['0xd04', ['-mcpu=cortex-a35']],
+	['0xd07', ['-mcpu=cortex-a57']],
+	['0xd08', ['-mcpu=cortex-a72']],
+	['0xd09', ['-mcpu=cortex-a73']],
+	['0xd0a', ['-mcpu=cortex-a75']]]
+
+machine_args_cavium = [
+	['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
+	['native', ['-march=native']],
+	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
+	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
+	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra]]
 
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page G7-5321)
 impl_generic = ['Generic armv8', flags_generic, machine_args_generic]
@@ -145,20 +145,19 @@ else
 			dpdk_conf.set(flag[0], flag[1])
 		endif
 	endforeach
-	# Primary part number based mcpu flags are supported
-	# for gcc versions > 7
-	if cc.version().version_compare(
-			'<7.0') or cmd_output.length() == 0
-		if not meson.is_cross_build() and arm_force_native_march == true
-			impl_pn = 'native'
-		else
-			impl_pn = 'default'
-		endif
-	endif
+	
 	foreach marg: machine[2]
 		if marg[0] == impl_pn
-			foreach f: marg[1]
-				machine_args += f
+			foreach flag: marg[1]
+				if cc.has_argument(flag)
+					machine_args += flag
+				endif
+			endforeach
+			# Apply any extra machine specific flags.
+			foreach flag: marg.get(2, flags_default_extra)
+				if flag.length() > 0
+					dpdk_conf.set(flag[0], flag[1])
+				endif
 			endforeach
 		endif
 	endforeach
Signature · 1 line
2.21.0
#87 lore raw

PATCH v9 3/4 config: add thunderx2 machine config

In reply to [email protected]

From: Jerin Jacob <[email protected]>

Optimized configuration for Marvell thunderx2 SoC. Updated meson build to support Marvell thunderx2 SoC. Added meson cross compile target.

Product details are here:

https://www.marvell.com/server-processors/thunderx2-arm-processors/

---
 config/arm/arm64_thunderx2_linux_gcc          | 16 +++++++++
 config/arm/meson.build                        |  9 ++++-
 config/defconfig_arm64-thunderx2-linux-gcc    |  1 +
 config/defconfig_arm64-thunderx2-linuxapp-gcc | 11 ++++++
 mk/machine/thunderx2/rte.vars.mk              | 34 +++++++++++++++++++
 5 files changed, 70 insertions(+), 1 deletion(-)
 create mode 100644 config/arm/arm64_thunderx2_linux_gcc
 create mode 120000 config/defconfig_arm64-thunderx2-linux-gcc
 create mode 100644 config/defconfig_arm64-thunderx2-linuxapp-gcc
 create mode 100644 mk/machine/thunderx2/rte.vars.mk
config/arm/arm64_thunderx2_linux_gcc new file +16 -0
new file mode 100644
index 000000000..0dc275644
@@ -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 = '0x43'
+implementor_pn = '0xaf'
config/arm/meson.build +8 -1
index d30a17426..a5fce7a59 100644
@@ -54,6 +54,12 @@ flags_default_extra = []
 flags_thunderx_extra = [
 	['RTE_MACHINE', '"thunderx"'],
 	['RTE_USE_C11_MEM_MODEL', false]]
+flags_thunderx2_extra = [
+	['RTE_MACHINE', '"thunderx2"'],
+	['RTE_CACHE_LINE_SIZE', 64],
+	['RTE_MAX_NUMA_NODES', 2],
+	['RTE_MAX_LCORE', 256],
+	['RTE_USE_C11_MEM_MODEL', true]]
 
 machine_args_generic = [
 	['default', ['-march=armv8-a+crc+crypto']],
@@ -70,7 +76,8 @@ machine_args_cavium = [
 	['native', ['-march=native']],
 	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
 	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
-	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra]]
+	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra],
+	['0xaf', ['-mcpu=thunderx2t99'], flags_thunderx2_extra]]
 
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page G7-5321)
 impl_generic = ['Generic armv8', flags_generic, machine_args_generic]
config/defconfig_arm64-thunderx2-linux-gcc new file +1 -0
new file mode 120000
index 000000000..b40a760b1
@@ -0,0 +1 @@
+defconfig_arm64-thunderx2-linuxapp-gcc
\ No newline at end of file
config/defconfig_arm64-thunderx2-linuxapp-gcc new file +11 -0
new file mode 100644
index 000000000..cc5c64ba0
@@ -0,0 +1,11 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International Ltd
+#
+
+#include "defconfig_arm64-armv8a-linux-gcc"
+
+CONFIG_RTE_MACHINE="thunderx2"
+
+CONFIG_RTE_CACHE_LINE_SIZE=64
+CONFIG_RTE_MAX_NUMA_NODES=2
+CONFIG_RTE_MAX_LCORE=256
mk/machine/thunderx2/rte.vars.mk new file +34 -0
new file mode 100644
index 000000000..b80dc8680
@@ -0,0 +1,34 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International 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=armv8.1-a+crc+crypto)
+MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=thunderx2t99)
Signature · 1 line
2.21.0
#88 lore raw

PATCH v9 4/4 config: add octeontx2 machine config

In reply to [email protected]

From: Jerin Jacob <[email protected]>

Optimized configuration for Marvell octeontx2 SoC. Updated meson build to support Marvell octeontx2 SoC. Added meson cross build target for octeontx2.

---
 config/arm/arm64_octeontx2_linux_gcc          | 16 +++++++++
 config/arm/meson.build                        |  9 ++++-
 config/defconfig_arm64-octeontx2-linux-gcc    |  1 +
 config/defconfig_arm64-octeontx2-linuxapp-gcc | 18 ++++++++++
 mk/machine/octeontx2/rte.vars.mk              | 34 +++++++++++++++++++
 5 files changed, 77 insertions(+), 1 deletion(-)
 create mode 100644 config/arm/arm64_octeontx2_linux_gcc
 create mode 120000 config/defconfig_arm64-octeontx2-linux-gcc
 create mode 100644 config/defconfig_arm64-octeontx2-linuxapp-gcc
 create mode 100644 mk/machine/octeontx2/rte.vars.mk
config/arm/arm64_octeontx2_linux_gcc new file +16 -0
new file mode 100644
index 000000000..e2c0b8f72
@@ -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 = '0x43'
+implementor_pn = '0xb2'
config/arm/meson.build +8 -1
index a5fce7a59..bdb36ca69 100644
@@ -60,6 +60,12 @@ flags_thunderx2_extra = [
 	['RTE_MAX_NUMA_NODES', 2],
 	['RTE_MAX_LCORE', 256],
 	['RTE_USE_C11_MEM_MODEL', true]]
+flags_octeontx2_extra = [
+	['RTE_MACHINE', '"octeontx2"'],
+	['RTE_MAX_NUMA_NODES', 1],
+	['RTE_MAX_LCORE', 24],
+	['RTE_EAL_IGB_UIO', false],
+	['RTE_USE_C11_MEM_MODEL', true]]
 
 machine_args_generic = [
 	['default', ['-march=armv8-a+crc+crypto']],
@@ -77,7 +83,8 @@ machine_args_cavium = [
 	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
 	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
 	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra],
-	['0xaf', ['-mcpu=thunderx2t99'], flags_thunderx2_extra]]
+	['0xaf', ['-mcpu=thunderx2t99'], flags_thunderx2_extra],
+	['0xb2', ['-mcpu=octeontx2'], flags_octeontx2_extra]]
 
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page G7-5321)
 impl_generic = ['Generic armv8', flags_generic, machine_args_generic]
config/defconfig_arm64-octeontx2-linux-gcc new file +1 -0
new file mode 120000
index 000000000..e25150531
@@ -0,0 +1 @@
+defconfig_arm64-octeontx2-linuxapp-gcc
\ No newline at end of file
config/defconfig_arm64-octeontx2-linuxapp-gcc new file +18 -0
new file mode 100644
index 000000000..9eae84538
@@ -0,0 +1,18 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International Ltd
+#
+
+#include "defconfig_arm64-armv8a-linux-gcc"
+
+CONFIG_RTE_MACHINE="octeontx2"
+
+CONFIG_RTE_CACHE_LINE_SIZE=128
+CONFIG_RTE_MAX_NUMA_NODES=1
+CONFIG_RTE_MAX_LCORE=24
+
+# Doesn't support NUMA
+CONFIG_RTE_EAL_NUMA_AWARE_HUGEPAGES=n
+CONFIG_RTE_LIBRTE_VHOST_NUMA=n
+
+# Recommend to use VFIO as co-processors needs SMMU/IOMMU
+CONFIG_RTE_EAL_IGB_UIO=n
mk/machine/octeontx2/rte.vars.mk new file +34 -0
new file mode 100644
index 000000000..cbec7f14d
@@ -0,0 +1,34 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International 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=-mcpu=armv8.2-a+crc+crypto+lse)
+MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=octeontx2)
Signature · 1 line
2.21.0
#89 lore raw

PATCH v10 1/4 mk: introduce helper to check valid compiler argument

In reply to [email protected]

From: Jerin Jacob <[email protected]>

Introduce rte_cc_has_argument() Makefile helper to check a given argument is support by the compiler.

Example Usage:

include $(RTE_SDK)/mk/rte.helper.mk
MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=octeontx2)

This would allow adding -mcpu=octeontx2 in MACHINE_CFLAGS if it is only supported by the compiler. The use case for such scheme is to enable the mcpu optimization if the compiler supports else it needs to compile the source code without any errors.

This patch also moves inclusion of toolchain's rte.vars.mk to before the machine's rte.vars.mk inclusion to make correct CC available for the cross compile case.

---

Change history of this series:

v10 Changes:
- Fix the following checkpatch warning
http://mails.dpdk.org/archives/test-report/2019-April/080453.html

v9 Changes:
- Remove compiler version check as it is now done using cc.has_argument().

v8 Changes:
- Remove redudant lists (rebase aritfacts). (Yongseok Koh)

v7 Changes:
- Updated cross compile config files align with
"build: improve pcap dependency handling" changeset to fix build issue with meson
- Some compiler needs the following depended patch to compile with meson
http://patches.dpdk.org/patch/52367/

v6 Changes:
- Rework to change the config files to sync with "mk: use linux and freebsd in config names"
- Fix the following error with latest gcc by fixing the mcpu type
cc1: error: switch -mcpu=armv8.2-a conflicts with -march=armv8-a switch

v5 Changes:
- Fix incorrect meson flag parsing(Phil Yang)
- Squash meson cross build patch(5/5) into configuration update patches for thunderx2(3/5) and octeontx2(4/5)(Thomas)
- Changed octeontx2's march as armv8-a and added the extension required instead of armv8-2a(Phil Yang)
- Improved rte_cc_has_argument() implementaion by removing the temp file(Thomas)

v4 Changes:
- Fix incorrect signoff marrvell -> marvell.

v3 Changes:
- Squash meson build support into config support for thunderx2/octeontx2.

v2 Changes:
- Add meson build support.

---

 mk/rte.helper.mk              | 10 ++++++++++
 mk/target/generic/rte.vars.mk | 22 +++++++++++-----------
 2 files changed, 21 insertions(+), 11 deletions(-)
 create mode 100644 mk/rte.helper.mk
mk/rte.helper.mk new file +10 -0
new file mode 100644
index 000000000..6e7fd03d7
@@ -0,0 +1,10 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International Ltd
+
+# rte_cc_has_argument
+# Usage: MACHINE_CFLAGS += $(call rte_cc_has_argument, -mno-avx512f)
+# Return the argument if the argument is supported by the compiler.
+#
+define rte_cc_has_argument
+	$(shell $(CC) -E $(1) -xc /dev/null 1>/dev/null 2>/dev/null && echo $(1))
+endef
mk/target/generic/rte.vars.mk +11 -11
index dd149acc9..25a578ad7 100644
@@ -7,6 +7,17 @@
 # executive environment.
 #
 
+#
+# toolchain:
+#
+#   - define CC, LD, AR, AS, ...
+#   - define TOOLCHAIN_CFLAGS variable (overridden by cmdline value)
+#   - define TOOLCHAIN_LDFLAGS variable (overridden by cmdline value)
+#   - define TOOLCHAIN_ASFLAGS variable (overridden by cmdline value)
+#   - may override any previously defined variable
+#
+include $(RTE_SDK)/mk/toolchain/$(RTE_TOOLCHAIN)/rte.vars.mk
+
 #
 # machine:
 #
@@ -45,17 +56,6 @@ endif
 #
 include $(RTE_SDK)/mk/arch/$(RTE_ARCH)/rte.vars.mk
 
-#
-# toolchain:
-#
-#   - define CC, LD, AR, AS, ...
-#   - define TOOLCHAIN_CFLAGS variable (overridden by cmdline value)
-#   - define TOOLCHAIN_LDFLAGS variable (overridden by cmdline value)
-#   - define TOOLCHAIN_ASFLAGS variable (overridden by cmdline value)
-#   - may override any previously defined variable
-#
-include $(RTE_SDK)/mk/toolchain/$(RTE_TOOLCHAIN)/rte.vars.mk
-
 #
 # exec-env:
 #
Signature · 1 line
2.21.0
#90 lore raw

PATCH v10 2/4 meson: add infra to support machine specific flags

From: Pavan Nikhilesh <[email protected]>

Currently, RTE_* flags are set based on the implementer ID but there might be some micro arch specific differences from the same vendor eg. CACHE_LINESIZE. Add support to set micro arch specific flags.

---
 config/arm/meson.build | 67 +++++++++++++++++++++---------------------
 1 file changed, 33 insertions(+), 34 deletions(-)
config/arm/meson.build +33 -34
index 170a4981a..104acd6d4 100644
@@ -7,25 +7,6 @@ march_opt = '-march=@0@'.format(machine)
 
 arm_force_native_march = false
 
-machine_args_generic = [
-	['default', ['-march=armv8-a+crc+crypto']],
-	['native', ['-march=native']],
-	['0xd03', ['-mcpu=cortex-a53']],
-	['0xd04', ['-mcpu=cortex-a35']],
-	['0xd05', ['-mcpu=cortex-a55']],
-	['0xd07', ['-mcpu=cortex-a57']],
-	['0xd08', ['-mcpu=cortex-a72']],
-	['0xd09', ['-mcpu=cortex-a73']],
-	['0xd0a', ['-mcpu=cortex-a75']],
-	['0xd0b', ['-mcpu=cortex-a76']],
-]
-machine_args_cavium = [
-	['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
-	['native', ['-march=native']],
-	['0xa1', ['-mcpu=thunderxt88']],
-	['0xa2', ['-mcpu=thunderxt81']],
-	['0xa3', ['-mcpu=thunderxt83']]]
-
 flags_common_default = [
 	# Accelarate rte_memcpy. Be sure to run unit test (memcpy_perf_autotest)
 	# to determine the best threshold in code. Refer to notes in source file
@@ -52,12 +33,10 @@ flags_generic = [
 	['RTE_USE_C11_MEM_MODEL', true],
 	['RTE_CACHE_LINE_SIZE', 128]]
 flags_cavium = [
-	['RTE_MACHINE', '"thunderx"'],
 	['RTE_CACHE_LINE_SIZE', 128],
 	['RTE_MAX_NUMA_NODES', 2],
 	['RTE_MAX_LCORE', 96],
-	['RTE_MAX_VFIO_GROUPS', 128],
-	['RTE_USE_C11_MEM_MODEL', false]]
+	['RTE_MAX_VFIO_GROUPS', 128]]
 flags_dpaa = [
 	['RTE_MACHINE', '"dpaa"'],
 	['RTE_USE_C11_MEM_MODEL', true],
@@ -71,6 +50,27 @@ flags_dpaa2 = [
 	['RTE_MAX_NUMA_NODES', 1],
 	['RTE_MAX_LCORE', 16],
 	['RTE_LIBRTE_DPAA2_USE_PHYS_IOVA', false]]
+flags_default_extra = []
+flags_thunderx_extra = [
+	['RTE_MACHINE', '"thunderx"'],
+	['RTE_USE_C11_MEM_MODEL', false]]
+
+machine_args_generic = [
+	['default', ['-march=armv8-a+crc+crypto']],
+	['native', ['-march=native']],
+	['0xd03', ['-mcpu=cortex-a53']],
+	['0xd04', ['-mcpu=cortex-a35']],
+	['0xd07', ['-mcpu=cortex-a57']],
+	['0xd08', ['-mcpu=cortex-a72']],
+	['0xd09', ['-mcpu=cortex-a73']],
+	['0xd0a', ['-mcpu=cortex-a75']]]
+
+machine_args_cavium = [
+	['default', ['-march=armv8-a+crc+crypto','-mcpu=thunderx']],
+	['native', ['-march=native']],
+	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
+	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
+	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra]]
 
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page G7-5321)
 impl_generic = ['Generic armv8', flags_generic, machine_args_generic]
@@ -145,20 +145,19 @@ else
 			dpdk_conf.set(flag[0], flag[1])
 		endif
 	endforeach
-	# Primary part number based mcpu flags are supported
-	# for gcc versions > 7
-	if cc.version().version_compare(
-			'<7.0') or cmd_output.length() == 0
-		if not meson.is_cross_build() and arm_force_native_march == true
-			impl_pn = 'native'
-		else
-			impl_pn = 'default'
-		endif
-	endif
+
 	foreach marg: machine[2]
 		if marg[0] == impl_pn
-			foreach f: marg[1]
-				machine_args += f
+			foreach flag: marg[1]
+				if cc.has_argument(flag)
+					machine_args += flag
+				endif
+			endforeach
+			# Apply any extra machine specific flags.
+			foreach flag: marg.get(2, flags_default_extra)
+				if flag.length() > 0
+					dpdk_conf.set(flag[0], flag[1])
+				endif
 			endforeach
 		endif
 	endforeach
Signature · 1 line
2.21.0
#91 lore raw

PATCH v10 3/4 config: add thunderx2 machine config

In reply to [email protected]

From: Jerin Jacob <[email protected]>

Optimized configuration for Marvell thunderx2 SoC. Updated meson build to support Marvell thunderx2 SoC. Added meson cross compile target.

Product details are here:

https://www.marvell.com/server-processors/thunderx2-arm-processors/

---
 config/arm/arm64_thunderx2_linux_gcc          | 16 +++++++++
 config/arm/meson.build                        |  9 ++++-
 config/defconfig_arm64-thunderx2-linux-gcc    |  1 +
 config/defconfig_arm64-thunderx2-linuxapp-gcc | 11 ++++++
 mk/machine/thunderx2/rte.vars.mk              | 34 +++++++++++++++++++
 5 files changed, 70 insertions(+), 1 deletion(-)
 create mode 100644 config/arm/arm64_thunderx2_linux_gcc
 create mode 120000 config/defconfig_arm64-thunderx2-linux-gcc
 create mode 100644 config/defconfig_arm64-thunderx2-linuxapp-gcc
 create mode 100644 mk/machine/thunderx2/rte.vars.mk
config/arm/arm64_thunderx2_linux_gcc new file +16 -0
new file mode 100644
index 000000000..0dc275644
@@ -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 = '0x43'
+implementor_pn = '0xaf'
config/arm/meson.build +8 -1
index 104acd6d4..ccf806e73 100644
@@ -54,6 +54,12 @@ flags_default_extra = []
 flags_thunderx_extra = [
 	['RTE_MACHINE', '"thunderx"'],
 	['RTE_USE_C11_MEM_MODEL', false]]
+flags_thunderx2_extra = [
+	['RTE_MACHINE', '"thunderx2"'],
+	['RTE_CACHE_LINE_SIZE', 64],
+	['RTE_MAX_NUMA_NODES', 2],
+	['RTE_MAX_LCORE', 256],
+	['RTE_USE_C11_MEM_MODEL', true]]
 
 machine_args_generic = [
 	['default', ['-march=armv8-a+crc+crypto']],
@@ -70,7 +76,8 @@ machine_args_cavium = [
 	['native', ['-march=native']],
 	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
 	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
-	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra]]
+	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra],
+	['0xaf', ['-mcpu=thunderx2t99'], flags_thunderx2_extra]]
 
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page G7-5321)
 impl_generic = ['Generic armv8', flags_generic, machine_args_generic]
config/defconfig_arm64-thunderx2-linux-gcc new file +1 -0
new file mode 120000
index 000000000..b40a760b1
@@ -0,0 +1 @@
+defconfig_arm64-thunderx2-linuxapp-gcc
\ No newline at end of file
config/defconfig_arm64-thunderx2-linuxapp-gcc new file +11 -0
new file mode 100644
index 000000000..cc5c64ba0
@@ -0,0 +1,11 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International Ltd
+#
+
+#include "defconfig_arm64-armv8a-linux-gcc"
+
+CONFIG_RTE_MACHINE="thunderx2"
+
+CONFIG_RTE_CACHE_LINE_SIZE=64
+CONFIG_RTE_MAX_NUMA_NODES=2
+CONFIG_RTE_MAX_LCORE=256
mk/machine/thunderx2/rte.vars.mk new file +34 -0
new file mode 100644
index 000000000..b80dc8680
@@ -0,0 +1,34 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International 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=armv8.1-a+crc+crypto)
+MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=thunderx2t99)
Signature · 1 line
2.21.0
#92 lore raw

PATCH v10 4/4 config: add octeontx2 machine config

In reply to [email protected]

From: Jerin Jacob <[email protected]>

Optimized configuration for Marvell octeontx2 SoC. Updated meson build to support Marvell octeontx2 SoC. Added meson cross build target for octeontx2.

---
 config/arm/arm64_octeontx2_linux_gcc          | 16 +++++++++
 config/arm/meson.build                        |  9 ++++-
 config/defconfig_arm64-octeontx2-linux-gcc    |  1 +
 config/defconfig_arm64-octeontx2-linuxapp-gcc | 18 ++++++++++
 mk/machine/octeontx2/rte.vars.mk              | 34 +++++++++++++++++++
 5 files changed, 77 insertions(+), 1 deletion(-)
 create mode 100644 config/arm/arm64_octeontx2_linux_gcc
 create mode 120000 config/defconfig_arm64-octeontx2-linux-gcc
 create mode 100644 config/defconfig_arm64-octeontx2-linuxapp-gcc
 create mode 100644 mk/machine/octeontx2/rte.vars.mk
config/arm/arm64_octeontx2_linux_gcc new file +16 -0
new file mode 100644
index 000000000..e2c0b8f72
@@ -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 = '0x43'
+implementor_pn = '0xb2'
config/arm/meson.build +8 -1
index ccf806e73..22a062bad 100644
@@ -60,6 +60,12 @@ flags_thunderx2_extra = [
 	['RTE_MAX_NUMA_NODES', 2],
 	['RTE_MAX_LCORE', 256],
 	['RTE_USE_C11_MEM_MODEL', true]]
+flags_octeontx2_extra = [
+	['RTE_MACHINE', '"octeontx2"'],
+	['RTE_MAX_NUMA_NODES', 1],
+	['RTE_MAX_LCORE', 24],
+	['RTE_EAL_IGB_UIO', false],
+	['RTE_USE_C11_MEM_MODEL', true]]
 
 machine_args_generic = [
 	['default', ['-march=armv8-a+crc+crypto']],
@@ -77,7 +83,8 @@ machine_args_cavium = [
 	['0xa1', ['-mcpu=thunderxt88'], flags_thunderx_extra],
 	['0xa2', ['-mcpu=thunderxt81'], flags_thunderx_extra],
 	['0xa3', ['-mcpu=thunderxt83'], flags_thunderx_extra],
-	['0xaf', ['-mcpu=thunderx2t99'], flags_thunderx2_extra]]
+	['0xaf', ['-mcpu=thunderx2t99'], flags_thunderx2_extra],
+	['0xb2', ['-mcpu=octeontx2'], flags_octeontx2_extra]]
 
 ## Arm implementer ID (ARM DDI 0487C.a, Section G7.2.106, Page G7-5321)
 impl_generic = ['Generic armv8', flags_generic, machine_args_generic]
config/defconfig_arm64-octeontx2-linux-gcc new file +1 -0
new file mode 120000
index 000000000..e25150531
@@ -0,0 +1 @@
+defconfig_arm64-octeontx2-linuxapp-gcc
\ No newline at end of file
config/defconfig_arm64-octeontx2-linuxapp-gcc new file +18 -0
new file mode 100644
index 000000000..9eae84538
@@ -0,0 +1,18 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International Ltd
+#
+
+#include "defconfig_arm64-armv8a-linux-gcc"
+
+CONFIG_RTE_MACHINE="octeontx2"
+
+CONFIG_RTE_CACHE_LINE_SIZE=128
+CONFIG_RTE_MAX_NUMA_NODES=1
+CONFIG_RTE_MAX_LCORE=24
+
+# Doesn't support NUMA
+CONFIG_RTE_EAL_NUMA_AWARE_HUGEPAGES=n
+CONFIG_RTE_LIBRTE_VHOST_NUMA=n
+
+# Recommend to use VFIO as co-processors needs SMMU/IOMMU
+CONFIG_RTE_EAL_IGB_UIO=n
mk/machine/octeontx2/rte.vars.mk new file +34 -0
new file mode 100644
index 000000000..cbec7f14d
@@ -0,0 +1,34 @@
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2018 Marvell International 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=-mcpu=armv8.2-a+crc+crypto+lse)
+MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=octeontx2)
Signature · 1 line
2.21.0
#93 lore raw
In reply to [email protected]

13/04/2019 22:19, [email protected]:

Quoted 10 lines

Change history of this series:

v10 Changes:
- Fix the following checkpatch warning
http://mails.dpdk.org/archives/test-report/2019-April/080453.html

v9 Changes:
- Remove compiler version check as it is now done using cc.has_argument().

Quoted 8 lines

v8 Changes:
- Remove redudant lists (rebase aritfacts). (Yongseok Koh)

v7 Changes:
- Updated cross compile config files align with
"build: improve pcap dependency handling" changeset to fix build issue with meson
- Some compiler needs the following depended patch to compile with meson
http://patches.dpdk.org/patch/52367/

Applied, thanks

Reviewed-by on v5 1/4
#94 lore raw
In reply to Jerin Jacob Kollanukkaran
Quoted 113 lines

-----Original Message-----
From: Jerin Jacob Kollanukkaran <[email protected]>
Sent: Monday, February 25, 2019 2:11 AM
To: [email protected]
Cc: [email protected]; Phil Yang (Arm Technology China) <[email protected]>; [email protected]; [email protected]; Pavan Nikhilesh Bhagavatula <[email protected]>
Subject: [dpdk-dev] [PATCH v5 1/4] mk: introduce helper to check valid compiler argument

Introduce rte_cc_has_argument() Makefile helper to check a given argument is support by the compiler.

Example Usage:

include $(RTE_SDK)/mk/rte.helper.mk
MACHINE_CFLAGS += $(call rte_cc_has_argument, -mcpu=octeontx2)

This would allow adding -mcpu=octeontx2 in MACHINE_CFLAGS if it is only supported by the compiler. The use case for such scheme is to enable the mcpu optimization if the compiler supports else it needs to compile the source code without any errors.

This patch also moves inclusion of toolchain's rte.vars.mk to before the machine's rte.vars.mk inclusion to make correct CC available for the cross compile case.

---

Change history of this series:

v2 Changes:
- Add meson build support.

v3 Changes:
- Squash meson build support into config support for thunderx2/octeontx2.

v4 Changes:
- Fix incorrect signoff marrvell -> marvell.

v5 Changes:
- Fix incorrect meson flag parsing(Phil Yang)
- Squash meson cross build patch(5/5) into configuration update patches for thunderx2(3/5) and octeontx2(4/5)(Thomas)
- Changed octeontx2's march as armv8-a and added the extension required instead of armv8-2a(Phil Yang)
- Improved rte_cc_has_argument() implementaion by removing the temp file(Thomas)

---
 mk/rte.helper.mk              | 10 ++++++++++
 mk/target/generic/rte.vars.mk | 22 +++++++++++-----------
 2 files changed, 21 insertions(+), 11 deletions(-)  create mode 100644

mk/rte.helper.mk

mk/rte.helper.mk new file mode 100644 index renamed from mk/rte.helper.mk +0 -0

000000000..6e7fd03d7

mk/rte.helper.mk new file +10 -0
@@ -0,0 +1,10 @@
+# SPDX-License-Identifier: BSD-3-Clause # Copyright(c) 2018 Marvell
+International Ltd
+
+# rte_cc_has_argument
+# Usage: MACHINE_CFLAGS += $(call rte_cc_has_argument, -mno-avx512f) #
+Return the argument if the argument is supported by the compiler.
+#
+define rte_cc_has_argument
+	$(shell $(CC) -E $(1) -xc /dev/null 1>/dev/null 2>/dev/null && echo
+$(1)) endef
mk/target/generic/rte.vars.mk +11 -12
index dd149acc9..25a578ad7 100644
@@ -7,6 +7,17 @@
 # executive environment.
 #
 
+#
+# toolchain:
+#
+#   - define CC, LD, AR, AS, ...
+#   - define TOOLCHAIN_CFLAGS variable (overridden by cmdline value)
+#   - define TOOLCHAIN_LDFLAGS variable (overridden by cmdline value)
+#   - define TOOLCHAIN_ASFLAGS variable (overridden by cmdline value)
+#   - may override any previously defined variable
+#
+include $(RTE_SDK)/mk/toolchain/$(RTE_TOOLCHAIN)/rte.vars.mk
+
 #
 # machine:
 #
@@ -45,17 +56,6 @@ endif
 #
 include $(RTE_SDK)/mk/arch/$(RTE_ARCH)/rte.vars.mk
 
-#
-# toolchain:
-#
-#   - define CC, LD, AR, AS, ...
-#   - define TOOLCHAIN_CFLAGS variable (overridden by cmdline value)
-#   - define TOOLCHAIN_LDFLAGS variable (overridden by cmdline value)
-#   - define TOOLCHAIN_ASFLAGS variable (overridden by cmdline value)
-#   - may override any previously defined variable
-#
-include $(RTE_SDK)/mk/toolchain/$(RTE_TOOLCHAIN)/rte.vars.mk
-
 #
 # exec-env:
 #
--
Quoted 1 line

2.20.1

For this patch set.

Thanks,
Phil