Skip to contentSkip to conversation

mk: fix unsupported flag error on armhf architercture

PATCH In review Phil Yang 5 participants 16 messages Nov 11, 2019 – Jul 2, 2020

lore
Share & tools
lore

Phil Yang

author
#0 lore raw

PATCH 1/2 mk: fix unsupported flag error on armhf architercture

The older version (e.g. version 7.4.0 ) of GNU C compiler for the armhf architecture doesn't support the flag '-Wno-address-of-packed-member', so remove this flag for aarch32.

Fixes a385972c3675 ("mk: disable warning for packed member pointer")
---
 mk/toolchain/gcc/rte.vars.mk | 2 ++
 1 file changed, 2 insertions(+)
mk/toolchain/gcc/rte.vars.mk +2 -0
index 9fc7041..ec0cbbf 100644
@@ -100,7 +100,9 @@ WERROR_FLAGS += -Wno-format-truncation
 endif
 
 # disable packed member unalign warnings
+ifneq ($(CONFIG_RTE_ARCH_ARM), y)
 WERROR_FLAGS += -Wno-address-of-packed-member
+endif
 
 export CC AS AR LD OBJCOPY OBJDUMP STRIP READELF
 export TOOLCHAIN_CFLAGS TOOLCHAIN_LDFLAGS TOOLCHAIN_ASFLAGS
Signature · 1 line
2.7.4
author
#1 lore raw

PATCH 2/2 doc: add aarch32 build guidance

Add a guidance for aarch64 natively build 32-bit aarch32 DPDK.

---
 .../linux_gsg/cross_build_dpdk_for_arm64.rst       | 36 ++++++++++++++++++----
 1 file changed, 30 insertions(+), 6 deletions(-)
doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst +30 -6
index e799b0b..409d444 100644
@@ -1,15 +1,15 @@
 ..  SPDX-License-Identifier: BSD-3-Clause
     Copyright(c) 2018 ARM Corporation.
 
-Cross compile DPDK for ARM64
-============================
-This chapter describes how to cross compile DPDK for ARM64 from x86 build hosts.
+Cross compile DPDK for aarch64 and aarch32
+==========================================
+This chapter describes how to cross-compile DPDK for aarch64 from x86 and compile 32-bit aarch32 DPDK from aarch64 build hosts.
 
 .. note::
 
-   Whilst it is recommended to natively build DPDK on ARM64 (just
-   like with x86), it is also possible to cross-build DPDK for ARM64. An
-   ARM64 cross compile GNU toolchain is used for this.
+   Whilst it is recommended to natively build DPDK on aarch64 (just
+   like with x86), it is also possible to cross-build DPDK for aarch64. An
+   aarch64 cross compile GNU toolchain is used for this.
 
 Obtain the cross tool chain
 ---------------------------
@@ -133,3 +133,27 @@ command::
 
 	meson arm64-build --cross-file config/arm/arm64_armv8_linux_gcc
 	ninja -C arm64-build
+
+Compiling DPDK for aarch32/armv7 on aarch64 host
+------------------------------------------
+
+Some aarch64 platforms support EL0 aarch32 mode, so the 32-bit aarch32
+applications and the legacy 32-bit armv7 applications can run in this mode.
+With the GNU C compiler for the armhf architecture toolchain
+gcc-arm-linux-gnueabihf, it can generate aarch32 binary on aarch64 natively.
+
+For Example, in 64-bit Debian, we can use the following command to build 32-bit armv7 DPDK on aarch64 natively:
+
+Install cross compile toolchain for the armhf architecure.
+command::
+
+	dpkg --add-architecture armhf
+	apt-get update
+	apt-get install -y gcc gcc-arm-linux-gnueabihf libc6:armhf make binutils
+
+Build DPDK arm-armv7a-linuxapp-gcc executables.
+command::
+
+	cd <dpdk_folder>
+	make config CROSS=arm-linux-gnueabihf- T=arm-armv7a-linuxapp-gcc
+	make -j CROSS=arm-linux-gnueabihf- T=arm-armv7a-linuxapp-gcc RTE_DEVEL_BUILD=n
Signature · 1 line
2.7.4
#2 lore raw

Re: [dpdk-dev] [PATCH 2/2] doc: add aarch32 build guidance

11/11/2019 11:34, Phil Yang:

Quoted 11 lines

Add a guidance for aarch64 natively build 32-bit aarch32 DPDK.

---
 .../linux_gsg/cross_build_dpdk_for_arm64.rst       | 36 ++++++++++++++++++----
 1 file changed, 30 insertions(+), 6 deletions(-)

What is the status of this patch? The patch 1/2 was rejected. There are new works around aarch32. Is it still relevant?

author
#3 lore raw

Re: [dpdk-dev] [PATCH 2/2] doc: add aarch32 build guidance

Quoted 21 lines

-----Original Message-----
From: Thomas Monjalon <[email protected]>
Sent: Monday, May 25, 2020 5:21 AM
To: Phil Yang <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; Honnappa Nagarahalli <[email protected]>; Gavin Hu <[email protected]>; Joyce Kong <[email protected]>; nd <[email protected]>; [email protected]
Subject: Re: [dpdk-dev] [PATCH 2/2] doc: add aarch32 build guidance

11/11/2019 11:34, Phil Yang:

Quoted 10 lines

Add a guidance for aarch64 natively build 32-bit aarch32 DPDK.

---
 .../linux_gsg/cross_build_dpdk_for_arm64.rst       | 36
Quoted 7 lines

++++++++++++++++++----

Quoted 1 line
 1 file changed, 30 insertions(+), 6 deletions(-)

What is the status of this patch?
The patch 1/2 was rejected.
There are new works around aarch32.
Is it still relevant?

Thank you for your kindly reminder.
Yes. It is relevant.

This document is intended to guide the user to cross-compile aarch32 DPDK.
Usually the cross-compile host is x86, but cross-compiling on aarch64 is convenient as well, but not everyone knows how to do it. So it is helpful to provide relevant reference documents.

Thanks,
Phil

Acked-by on v1 2/2
#4 lore raw

Re: [dpdk-dev] [PATCH 2/2] doc: add aarch32 build guidance

In reply to Phil Yang
Quoted 87 lines

-----Original Message-----
From: dev <[email protected]> On Behalf Of Phil Yang
Sent: Monday, November 11, 2019 6:34 PM
To: [email protected]
Cc: [email protected]; [email protected]; [email protected]; Honnappa Nagarahalli <[email protected]>; Gavin Hu (Arm Technology China) <[email protected]>; Joyce Kong (Arm Technology China) <[email protected]>; nd <[email protected]>; [email protected]
Subject: [dpdk-dev] [PATCH 2/2] doc: add aarch32 build guidance

Add a guidance for aarch64 natively build 32-bit aarch32 DPDK.

---
 .../linux_gsg/cross_build_dpdk_for_arm64.rst       | 36
++++++++++++++++++----
 1 file changed, 30 insertions(+), 6 deletions(-)
doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst +0 -3
index e799b0b..409d444 100644
@@ -1,15 +1,15 @@
 ..  SPDX-License-Identifier: BSD-3-Clause
     Copyright(c) 2018 ARM Corporation.
 
-Cross compile DPDK for ARM64
-============================
-This chapter describes how to cross compile DPDK for ARM64 from x86 build

hosts.

doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst +3 -0
+Cross compile DPDK for aarch64 and aarch32
+==========================================
+This chapter describes how to cross-compile DPDK for aarch64 from x86 and

compile 32-bit aarch32 DPDK from aarch64 build hosts.

.. note::

doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst +13 -3
-   Whilst it is recommended to natively build DPDK on ARM64 (just
-   like with x86), it is also possible to cross-build DPDK for ARM64. An
-   ARM64 cross compile GNU toolchain is used for this.
+   Whilst it is recommended to natively build DPDK on aarch64 (just
+   like with x86), it is also possible to cross-build DPDK for aarch64. An
+   aarch64 cross compile GNU toolchain is used for this.
 
 Obtain the cross tool chain
 ---------------------------
@@ -133,3 +133,27 @@ command::
 
 	meson arm64-build --cross-file config/arm/arm64_armv8_linux_gcc
 	ninja -C arm64-build
+
+Compiling DPDK for aarch32/armv7 on aarch64 host
+------------------------------------------
+
+Some aarch64 platforms support EL0 aarch32 mode, so the 32-bit aarch32
+applications and the legacy 32-bit armv7 applications can run in this mode.
+With the GNU C compiler for the armhf architecture toolchain
+gcc-arm-linux-gnueabihf, it can generate aarch32 binary on aarch64 natively.
+
+For Example, in 64-bit Debian, we can use the following command to build

32-bit armv7 DPDK on aarch64 natively:

doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst +14 -0
+
+Install cross compile toolchain for the armhf architecure.
+command::
+
+	dpkg --add-architecture armhf
+	apt-get update
+	apt-get install -y gcc gcc-arm-linux-gnueabihf libc6:armhf make
+binutils
+
+Build DPDK arm-armv7a-linuxapp-gcc executables.
+command::
+
+	cd <dpdk_folder>
+	make config CROSS=arm-linux-gnueabihf- T=arm-armv7a-linuxapp-

gcc

doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst +2 -1
+	make -j CROSS=arm-linux-gnueabihf- T=arm-armv7a-linuxapp-gcc
+RTE_DEVEL_BUILD=n
--
Quoted 1 line

2.7.4

author
#5 lore raw

Re: [dpdk-dev] [PATCH 2/2] doc: add aarch32 build guidance

Hi Jerin,

Your insight would be much appreciated.

Thanks,
Phil

Quoted 103 lines

-----Original Message-----
From: Ruifeng Wang <[email protected]>
Sent: Wednesday, May 27, 2020 4:29 PM
To: Phil Yang <[email protected]>; [email protected]
Cc: [email protected]; [email protected]; [email protected]; Honnappa Nagarahalli <[email protected]>; Joyce Kong <[email protected]>; nd <[email protected]>; [email protected]
Subject: RE: [dpdk-dev] [PATCH 2/2] doc: add aarch32 build guidance

Quoted 30 lines

-----Original Message-----
From: dev <[email protected]> On Behalf Of Phil Yang
Sent: Monday, November 11, 2019 6:34 PM
To: [email protected]
Cc: [email protected]; [email protected]; [email protected]; Honnappa Nagarahalli <[email protected]>; Gavin Hu (Arm Technology China) <[email protected]>; Joyce Kong (Arm Technology China) <[email protected]>; nd <[email protected]>; [email protected]
Subject: [dpdk-dev] [PATCH 2/2] doc: add aarch32 build guidance

Add a guidance for aarch64 natively build 32-bit aarch32 DPDK.

---
 .../linux_gsg/cross_build_dpdk_for_arm64.rst       | 36
++++++++++++++++++----
 1 file changed, 30 insertions(+), 6 deletions(-)
doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst +0 -0
index e799b0b..409d444 100644

doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst line 6

Quoted 6 lines
doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst +0 -3
@@ -1,15 +1,15 @@
 ..  SPDX-License-Identifier: BSD-3-Clause
     Copyright(c) 2018 ARM Corporation.
 
-Cross compile DPDK for ARM64
-============================
-This chapter describes how to cross compile DPDK for ARM64 from x86

build

doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst

Quoted 4 lines

hosts.

doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst +3 -0
+Cross compile DPDK for aarch64 and aarch32
+==========================================
+This chapter describes how to cross-compile DPDK for aarch64 from x86

and

Quoted 15 lines

compile 32-bit aarch32 DPDK from aarch64 build hosts.

.. note::

doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst +3 -3
-   Whilst it is recommended to natively build DPDK on ARM64 (just
-   like with x86), it is also possible to cross-build DPDK for ARM64. An
-   ARM64 cross compile GNU toolchain is used for this.
+   Whilst it is recommended to natively build DPDK on aarch64 (just
+   like with x86), it is also possible to cross-build DPDK for aarch64. An
+   aarch64 cross compile GNU toolchain is used for this.
 
 Obtain the cross tool chain
 ---------------------------
@@ -133,3 +133,27 @@ command::
 

doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst command:: · line 143

Quoted 10 lines
doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst +8 -0
 	meson arm64-build --cross-file config/arm/arm64_armv8_linux_gcc
 	ninja -C arm64-build
+
+Compiling DPDK for aarch32/armv7 on aarch64 host
+------------------------------------------
+
+Some aarch64 platforms support EL0 aarch32 mode, so the 32-bit aarch32
+applications and the legacy 32-bit armv7 applications can run in this mode.
+With the GNU C compiler for the armhf architecture toolchain
+gcc-arm-linux-gnueabihf, it can generate aarch32 binary on aarch64

natively.

Quoted 21 lines
doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst +2 -0
+
+For Example, in 64-bit Debian, we can use the following command to build

32-bit armv7 DPDK on aarch64 natively:

doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst +14 -0
+
+Install cross compile toolchain for the armhf architecure.
+command::
+
+	dpkg --add-architecture armhf
+	apt-get update
+	apt-get install -y gcc gcc-arm-linux-gnueabihf libc6:armhf make
+binutils
+
+Build DPDK arm-armv7a-linuxapp-gcc executables.
+command::
+
+	cd <dpdk_folder>
+	make config CROSS=arm-linux-gnueabihf- T=arm-armv7a-linuxapp-

gcc

doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst +2 -1
+	make -j CROSS=arm-linux-gnueabihf- T=arm-armv7a-linuxapp-gcc
+RTE_DEVEL_BUILD=n
--
Quoted 1 line

2.7.4

#6 lore raw

Re: [dpdk-dev] [PATCH 2/2] doc: add aarch32 build guidance

Quoted 87 lines · Phil Yang

Hi Jerin,

Your insight would be much appreciated.

Thanks,
Phil

Quoted 79 lines

-----Original Message-----
From: Ruifeng Wang <[email protected]>
Sent: Wednesday, May 27, 2020 4:29 PM
To: Phil Yang <[email protected]>; [email protected]
Cc: [email protected]; [email protected]; [email protected]; Honnappa Nagarahalli <[email protected]>; Joyce Kong <[email protected]>; nd <[email protected]>; [email protected]
Subject: RE: [dpdk-dev] [PATCH 2/2] doc: add aarch32 build guidance

Quoted 30 lines

-----Original Message-----
From: dev <[email protected]> On Behalf Of Phil Yang
Sent: Monday, November 11, 2019 6:34 PM
To: [email protected]
Cc: [email protected]; [email protected]; [email protected]; Honnappa Nagarahalli <[email protected]>; Gavin Hu (Arm Technology China) <[email protected]>; Joyce Kong (Arm Technology China) <[email protected]>; nd <[email protected]>; [email protected]
Subject: [dpdk-dev] [PATCH 2/2] doc: add aarch32 build guidance

Add a guidance for aarch64 natively build 32-bit aarch32 DPDK.

---
 .../linux_gsg/cross_build_dpdk_for_arm64.rst       | 36
++++++++++++++++++----
 1 file changed, 30 insertions(+), 6 deletions(-)
doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst +0 -0
index e799b0b..409d444 100644

doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst line 6

Quoted 6 lines
doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst +0 -3
@@ -1,15 +1,15 @@
 ..  SPDX-License-Identifier: BSD-3-Clause
     Copyright(c) 2018 ARM Corporation.
 
-Cross compile DPDK for ARM64
-============================
-This chapter describes how to cross compile DPDK for ARM64 from x86

build

doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst

Quoted 4 lines

hosts.

doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst +3 -0
+Cross compile DPDK for aarch64 and aarch32
+==========================================
+This chapter describes how to cross-compile DPDK for aarch64 from x86

and

Quoted 15 lines

compile 32-bit aarch32 DPDK from aarch64 build hosts.

.. note::

doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst +3 -3
-   Whilst it is recommended to natively build DPDK on ARM64 (just
-   like with x86), it is also possible to cross-build DPDK for ARM64. An
-   ARM64 cross compile GNU toolchain is used for this.
+   Whilst it is recommended to natively build DPDK on aarch64 (just
+   like with x86), it is also possible to cross-build DPDK for aarch64. An
+   aarch64 cross compile GNU toolchain is used for this.
 
 Obtain the cross tool chain
 ---------------------------
@@ -133,3 +133,27 @@ command::
 

doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst command:: · line 143

Quoted 10 lines
doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst +8 -0
    meson arm64-build --cross-file config/arm/arm64_armv8_linux_gcc
    ninja -C arm64-build
+
+Compiling DPDK for aarch32/armv7 on aarch64 host
+------------------------------------------
+
+Some aarch64 platforms support EL0 aarch32 mode, so the 32-bit aarch32
+applications and the legacy 32-bit armv7 applications can run in this mode.
+With the GNU C compiler for the armhf architecture toolchain
+gcc-arm-linux-gnueabihf, it can generate aarch32 binary on aarch64

natively.

Quoted 2 lines
Quoted 2 lines

doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst command:: · line 145

Quoted 2 lines
doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst +2 -0
+
+For Example, in 64-bit Debian, we can use the following command to build

I think, "we" usage should be removed from the doc.

Quoted 10 lines
Quoted 10 lines
Quoted 10 lines

32-bit armv7 DPDK on aarch64 natively:

doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst +9 -0
+
+Install cross compile toolchain for the armhf architecure.
+command::
+
+   dpkg --add-architecture armhf
+   apt-get update
+   apt-get install -y gcc gcc-arm-linux-gnueabihf libc6:armhf make
+binutils
+
Quoted 8 lines
Quoted 8 lines

doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst

Quoted 8 lines
doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst +5 -0
+Build DPDK arm-armv7a-linuxapp-gcc executables.
+command::
+
+   cd <dpdk_folder>
+   make config CROSS=arm-linux-gnueabihf- T=arm-armv7a-linuxapp-

gcc

doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst +2 -0
+   make -j CROSS=arm-linux-gnueabihf- T=arm-armv7a-linuxapp-gcc
+RTE_DEVEL_BUILD=n

make based build is deprecated, Please change this to meson.

The rest looks good to me.

Quoted 4 lines
Quoted 4 lines
Quoted 2 lines

--
2.7.4

#7 lore raw
In reply to Phil Yang
Quoted 20 lines · Phil Yang

The older version (e.g. version 7.4.0 ) of GNU C compiler for the armhf architecture doesn't support the flag '-Wno-address-of-packed-member', so remove this flag for aarch32.

Fixes a385972c3675 ("mk: disable warning for packed member pointer")
---
 mk/toolchain/gcc/rte.vars.mk | 2 ++
 1 file changed, 2 insertions(+)
mk/toolchain/gcc/rte.vars.mk +0 -0
index 9fc7041..ec0cbbf 100644

mk/toolchain/gcc/rte.vars.mk WERROR_FLAGS += -Wno-format-truncation · line 106

Quoted 7 lines
mk/toolchain/gcc/rte.vars.mk +2 -0
@@ -100,7 +100,9 @@ WERROR_FLAGS += -Wno-format-truncation
 endif
 
 # disable packed member unalign warnings
+ifneq ($(CONFIG_RTE_ARCH_ARM), y)
 WERROR_FLAGS += -Wno-address-of-packed-member
+endif
 

You don't need to do this, as gcc will not complain about this unknown flag unless you have other issues in your code.[1] I think it's better to keep the code clean in this case, otherwise we'll have the code littered with conditionals for various flags.

/Bruce

[1] https://gcc.gnu.org/onlinedocs/gcc/Warning-Options.html
"When an unrecognized warning option is requested (e.g., -Wunknown-warning), GCC emits a diagnostic stating that the option is not recognized. However, if the -Wno- form is used, the behavior is slightly
different: no diagnostic is produced for -Wno-unknown-warning unless other diagnostics are being produced. This allows the use of new -Wno- options with old compilers, but if something goes wrong, the compiler warns that an unrecognized option is present."

#8 lore raw
Quoted 47 lines

-----Original Message-----
From: Bruce Richardson <[email protected]>
Sent: Monday, November 11, 2019 6:47 PM
To: Phil Yang (Arm Technology China) <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; Honnappa Nagarahalli <[email protected]>; Gavin Hu (Arm Technology China) <[email protected]>; Joyce Kong (Arm Technology China) <[email protected]>; nd <[email protected]>; [email protected]
Subject: Re: [dpdk-dev] [PATCH 1/2] mk: fix unsupported flag error on armhf architercture

Quoted 20 lines · Phil Yang

The older version (e.g. version 7.4.0 ) of GNU C compiler for the armhf architecture doesn't support the flag '-Wno-address-of-packed-member', so remove this flag for aarch32.

Fixes a385972c3675 ("mk: disable warning for packed member pointer")
---
 mk/toolchain/gcc/rte.vars.mk | 2 ++
 1 file changed, 2 insertions(+)
mk/toolchain/gcc/rte.vars.mk +0 -0
index 9fc7041..ec0cbbf 100644

mk/toolchain/gcc/rte.vars.mk WERROR_FLAGS += -Wno-format-truncation · line 106

Quoted 7 lines
mk/toolchain/gcc/rte.vars.mk +2 -0
@@ -100,7 +100,9 @@ WERROR_FLAGS += -Wno-format-truncation
 endif
 
 # disable packed member unalign warnings
+ifneq ($(CONFIG_RTE_ARCH_ARM), y)
 WERROR_FLAGS += -Wno-address-of-packed-member
+endif
 

You don't need to do this, as gcc will not complain about this unknown flag unless you have other issues in your code.[1] I think it's better to keep the code clean in this case, otherwise we'll have the code littered with conditionals for various flags.

/Bruce

Quoted 8 lines

[1] https://gcc.gnu.org/onlinedocs/gcc/Warning-Options.html
"When an unrecognized warning option is requested (e.g., -Wunknown-warning), GCC emits a diagnostic stating that the option is not recognized. However, if the -Wno- form is used, the behavior is slightly
different: no diagnostic is produced for -Wno-unknown-warning unless other diagnostics are being produced. This allows the use of new -Wno- options with old compilers, but if something goes wrong, the compiler warns that an unrecognized option is present."

Thanks, Bruce.

There are thousands of warnings with alignment when compiling for ARMv7 (ARMv7 supports unaligned memory access). Without this new flag, the build system works fine for ARMv7. So I add this conditional for ARMv7 only. I agree with you we should clean up the code, but it needs a lot of effort to achieve that. This patch is going to make the ARMv7 target build successfully during this period.

Thanks,
Phil

#9 lore raw

12/11/2019 06:25, Phil Yang (Arm Technology China):

Quoted 45 lines

From: Bruce Richardson <[email protected]>

Quoted 35 lines · Phil Yang
Quoted 20 lines

The older version (e.g. version 7.4.0 ) of GNU C compiler for the armhf architecture doesn't support the flag '-Wno-address-of-packed-member', so remove this flag for aarch32.

Fixes a385972c3675 ("mk: disable warning for packed member pointer")
---
 mk/toolchain/gcc/rte.vars.mk | 2 ++
 1 file changed, 2 insertions(+)
mk/toolchain/gcc/rte.vars.mk +0 -0
index 9fc7041..ec0cbbf 100644

mk/toolchain/gcc/rte.vars.mk WERROR_FLAGS += -Wno-format-truncation · line 106

Quoted 7 lines
mk/toolchain/gcc/rte.vars.mk +2 -0
@@ -100,7 +100,9 @@ WERROR_FLAGS += -Wno-format-truncation
 endif
 
 # disable packed member unalign warnings
+ifneq ($(CONFIG_RTE_ARCH_ARM), y)
 WERROR_FLAGS += -Wno-address-of-packed-member
+endif
 

You don't need to do this, as gcc will not complain about this unknown flag unless you have other issues in your code.[1] I think it's better to keep the code clean in this case, otherwise we'll have the code littered with conditionals for various flags.

/Bruce

Quoted 8 lines

[1] https://gcc.gnu.org/onlinedocs/gcc/Warning-Options.html
"When an unrecognized warning option is requested (e.g., -Wunknown-warning), GCC emits a diagnostic stating that the option is not recognized. However, if the -Wno- form is used, the behavior is slightly
different: no diagnostic is produced for -Wno-unknown-warning unless other diagnostics are being produced. This allows the use of new -Wno- options with old compilers, but if something goes wrong, the compiler warns that an unrecognized option is present."

Quoted 6 lines

Thanks, Bruce.

There are thousands of warnings with alignment when compiling for ARMv7 (ARMv7 supports unaligned memory access). Without this new flag, the build system works fine for ARMv7. So I add this conditional for ARMv7 only. I agree with you we should clean up the code, but it needs a lot of effort to achieve that. This patch is going to make the ARMv7 target build successfully during this period.

I would go the opposite direction.
I think we should remove all warnings disablement like "-Wno-something", and clean the code.

If you want to allow building ARMv7, you can just allow warnings as not errors. So we will still see something is wrong but it won't prevent from building.

#10 lore raw
Quoted 71 lines

-----Original Message-----
From: Thomas Monjalon <[email protected]>
Sent: Tuesday, November 12, 2019 2:03 PM
To: Phil Yang (Arm Technology China) <[email protected]>
Cc: Bruce Richardson <[email protected]>; [email protected]; [email protected]; [email protected]; Honnappa Nagarahalli <[email protected]>; Gavin Hu (Arm Technology China) <[email protected]>; Joyce Kong (Arm Technology China) <[email protected]>; nd <[email protected]>; [email protected]
Subject: Re: [dpdk-dev] [PATCH 1/2] mk: fix unsupported flag error on armhf architercture

12/11/2019 06:25, Phil Yang (Arm Technology China):

Quoted 4 lines

From: Bruce Richardson <[email protected]>

Quoted 3 lines · Phil Yang
Quoted 2 lines

The older version (e.g. version 7.4.0 ) of GNU C compiler for the armhf architecture doesn't support the flag '-Wno-address-of-packed-

member',

Quoted 9 lines
Quoted 9 lines
Quoted 9 lines

so remove this flag for aarch32.

Fixes a385972c3675 ("mk: disable warning for packed member pointer")
Quoted 5 lines
Quoted 5 lines

Review

Quoted 5 lines
---
 mk/toolchain/gcc/rte.vars.mk | 2 ++
 1 file changed, 2 insertions(+)
(quoted diff) +0 -0

b/mk/toolchain/gcc/rte.vars.mk

Quoted 6 lines
Quoted 6 lines
Quoted 6 lines

index 9fc7041..ec0cbbf 100644

mk/toolchain/gcc/rte.vars.mk +0 -0
@@ -100,7 +100,9 @@ WERROR_FLAGS += -Wno-format-truncation
 endif
 
Quoted 6 lines
Quoted 6 lines

mk/toolchain/gcc/rte.vars.mk WERROR_FLAGS += -Wno-format-truncation · line 106

Quoted 5 lines
mk/toolchain/gcc/rte.vars.mk +2 -0
 # disable packed member unalign warnings
+ifneq ($(CONFIG_RTE_ARCH_ARM), y)
 WERROR_FLAGS += -Wno-address-of-packed-member
+endif
 

You don't need to do this, as gcc will not complain about this unknown

flag

Quoted 1 line
Quoted 1 line

unless you have other issues in your code.[1] I think it's better to keep

the

Quoted 8 lines
Quoted 8 lines

code clean in this case, otherwise we'll have the code littered with conditionals for various flags.

/Bruce

[1] https://gcc.gnu.org/onlinedocs/gcc/Warning-Options.html
"When an unrecognized warning option is requested (e.g., -Wunknown-warning), GCC emits a diagnostic stating that the option is

not

Quoted 2 lines
Quoted 2 lines

recognized. However, if the -Wno- form is used, the behavior is slightly
different: no diagnostic is produced for -Wno-unknown-warning unless

other

Quoted 2 lines
Quoted 2 lines

diagnostics are being produced. This allows the use of new -Wno- options with old compilers, but if something goes wrong, the compiler warns that

an

Quoted 5 lines
Quoted 1 line

unrecognized option is present."

Thanks, Bruce.

There are thousands of warnings with alignment when compiling for

ARMv7 (ARMv7 supports unaligned memory access).

Quoted 1 line

Without this new flag, the build system works fine for ARMv7. So I add this

conditional for ARMv7 only.

Quoted 8 lines
Quoted 1 line

I agree with you we should clean up the code, but it needs a lot of effort to

achieve that.

Quoted 1 line

This patch is going to make the ARMv7 target build successfully during this

period.

I would go the opposite direction.
I think we should remove all warnings disablement like "-Wno-something", and clean the code.

Thanks, Thomas.

I think it is a compiler issue rather than a code defect in this case, as It cannot recognize the '-Waddress-of-packed-member' flag. "arm-linux-gnueabihf-gcc: error: unrecognized command line option '-Waddress-of-packed-member'"

Quoted 4 lines

If you want to allow building ARMv7, you can just allow warnings as not errors.
So we will still see something is wrong but it won't prevent from building.

Yes. We did this already, all the alignment issues are emitted as warnings on ARMv7. Please see these two commits:
79fa5c2a9992 ("mk: ignore alignment errors for ARMv7")
6433c19105e8 ("mk: generalize strict alignment warning handling")

Quoted 2 lines
#11 lore raw

12/11/2019 07:40, Phil Yang (Arm Technology China):

Quoted 86 lines
Quoted 71 lines

-----Original Message-----
From: Thomas Monjalon <[email protected]>
Sent: Tuesday, November 12, 2019 2:03 PM
To: Phil Yang (Arm Technology China) <[email protected]>
Cc: Bruce Richardson <[email protected]>; [email protected]; [email protected]; [email protected]; Honnappa Nagarahalli <[email protected]>; Gavin Hu (Arm Technology China) <[email protected]>; Joyce Kong (Arm Technology China) <[email protected]>; nd <[email protected]>; [email protected]
Subject: Re: [dpdk-dev] [PATCH 1/2] mk: fix unsupported flag error on armhf architercture

12/11/2019 06:25, Phil Yang (Arm Technology China):

Quoted 4 lines

From: Bruce Richardson <[email protected]>

Quoted 3 lines · Phil Yang
Quoted 2 lines

The older version (e.g. version 7.4.0 ) of GNU C compiler for the armhf architecture doesn't support the flag '-Wno-address-of-packed-

member',

Quoted 9 lines
Quoted 9 lines
Quoted 9 lines

so remove this flag for aarch32.

Fixes a385972c3675 ("mk: disable warning for packed member pointer")
Quoted 5 lines
Quoted 5 lines

Review

Quoted 5 lines
---
 mk/toolchain/gcc/rte.vars.mk | 2 ++
 1 file changed, 2 insertions(+)
(quoted diff) +0 -0

b/mk/toolchain/gcc/rte.vars.mk

Quoted 6 lines
Quoted 6 lines
Quoted 6 lines

index 9fc7041..ec0cbbf 100644

mk/toolchain/gcc/rte.vars.mk +0 -0
@@ -100,7 +100,9 @@ WERROR_FLAGS += -Wno-format-truncation
 endif
 
Quoted 6 lines
Quoted 6 lines

mk/toolchain/gcc/rte.vars.mk WERROR_FLAGS += -Wno-format-truncation · line 106

Quoted 5 lines
mk/toolchain/gcc/rte.vars.mk +2 -0
 # disable packed member unalign warnings
+ifneq ($(CONFIG_RTE_ARCH_ARM), y)
 WERROR_FLAGS += -Wno-address-of-packed-member
+endif
 

You don't need to do this, as gcc will not complain about this unknown

flag

Quoted 1 line
Quoted 1 line

unless you have other issues in your code.[1] I think it's better to keep

the

Quoted 8 lines
Quoted 8 lines

code clean in this case, otherwise we'll have the code littered with conditionals for various flags.

/Bruce

[1] https://gcc.gnu.org/onlinedocs/gcc/Warning-Options.html
"When an unrecognized warning option is requested (e.g., -Wunknown-warning), GCC emits a diagnostic stating that the option is

not

Quoted 2 lines
Quoted 2 lines

recognized. However, if the -Wno- form is used, the behavior is slightly
different: no diagnostic is produced for -Wno-unknown-warning unless

other

Quoted 2 lines
Quoted 2 lines

diagnostics are being produced. This allows the use of new -Wno- options with old compilers, but if something goes wrong, the compiler warns that

an

Quoted 5 lines
Quoted 1 line

unrecognized option is present."

Thanks, Bruce.

There are thousands of warnings with alignment when compiling for

ARMv7 (ARMv7 supports unaligned memory access).

Quoted 1 line

Without this new flag, the build system works fine for ARMv7. So I add this

conditional for ARMv7 only.

Quoted 8 lines
Quoted 1 line

I agree with you we should clean up the code, but it needs a lot of effort to

achieve that.

Quoted 1 line

This patch is going to make the ARMv7 target build successfully during this

period.

I would go the opposite direction.
I think we should remove all warnings disablement like "-Wno-something", and clean the code.

Thanks, Thomas.

I think it is a compiler issue rather than a code defect in this case, as It cannot recognize the '-Waddress-of-packed-member' flag. "arm-linux-gnueabihf-gcc: error: unrecognized command line option '-Waddress-of-packed-member'"

Quoted 1 line
Quoted 8 lines
Quoted 3 lines

If you want to allow building ARMv7, you can just allow warnings as not errors.
So we will still see something is wrong but it won't prevent from building.

Yes. We did this already, all the alignment issues are emitted as warnings on ARMv7. Please see these two commits:
79fa5c2a9992 ("mk: ignore alignment errors for ARMv7")
6433c19105e8 ("mk: generalize strict alignment warning handling")

I don't know what to do with this patch. I think it is not how it should work, but you say your compiler has a real bug.

Can you confirm which compiler is it? Does it prevent to compile even with -Wno-error ?

Any other opinion?

#12 lore raw
Quoted 120 lines

-----Original Message-----
From: Thomas Monjalon <[email protected]>
Sent: Wednesday, November 27, 2019 1:02 AM
To: Phil Yang (Arm Technology China) <[email protected]>
Cc: [email protected]; Bruce Richardson <[email protected]>; [email protected]; [email protected]; Honnappa Nagarahalli <[email protected]>; Gavin Hu (Arm Technology China) <[email protected]>; Joyce Kong (Arm Technology China) <[email protected]>; nd <[email protected]>; [email protected]
Subject: Re: [dpdk-dev] [PATCH 1/2] mk: fix unsupported flag error on armhf architercture

12/11/2019 07:40, Phil Yang (Arm Technology China):

Quoted 6 lines
Quoted 6 lines

-----Original Message-----
From: Thomas Monjalon <[email protected]>
Sent: Tuesday, November 12, 2019 2:03 PM
To: Phil Yang (Arm Technology China) <[email protected]>
Cc: Bruce Richardson <[email protected]>; [email protected]; [email protected]; [email protected]; Honnappa

Nagarahalli

Quoted 4 lines
Quoted 4 lines

<[email protected]>; Gavin Hu (Arm Technology China) <[email protected]>; Joyce Kong (Arm Technology China) <[email protected]>; nd <[email protected]>; [email protected]
Subject: Re: [dpdk-dev] [PATCH 1/2] mk: fix unsupported flag error on

armhf

Quoted 6 lines
Quoted 6 lines

architercture

12/11/2019 06:25, Phil Yang (Arm Technology China):

Quoted 3 lines

From: Bruce Richardson <[email protected]>

Quoted 2 lines · Phil Yang
Quoted 1 line

The older version (e.g. version 7.4.0 ) of GNU C compiler for the

armhf

Quoted 5 lines
Quoted 5 lines
Quoted 1 line
Quoted 1 line
Quoted 1 line

architecture doesn't support the flag '-Wno-address-of-packed-

member',

Quoted 3 lines
Quoted 3 lines
Quoted 3 lines

so remove this flag for aarch32.

Fixes a385972c3675 ("mk: disable warning for packed member

pointer")

Quoted 18 lines
Quoted 18 lines
Quoted 11 lines
Quoted 11 lines

Review

Quoted 11 lines
---
 mk/toolchain/gcc/rte.vars.mk | 2 ++
 1 file changed, 2 insertions(+)
(quoted diff) +0 -0

b/mk/toolchain/gcc/rte.vars.mk

Quoted 6 lines
Quoted 6 lines
Quoted 6 lines

index 9fc7041..ec0cbbf 100644

mk/toolchain/gcc/rte.vars.mk +0 -0
@@ -100,7 +100,9 @@ WERROR_FLAGS += -Wno-format-truncation
 endif
 
Quoted 8 lines
Quoted 8 lines
Quoted 6 lines
Quoted 6 lines

mk/toolchain/gcc/rte.vars.mk WERROR_FLAGS += -Wno-format-truncation · line 106

Quoted 5 lines
mk/toolchain/gcc/rte.vars.mk +2 -0
 # disable packed member unalign warnings
+ifneq ($(CONFIG_RTE_ARCH_ARM), y)
 WERROR_FLAGS += -Wno-address-of-packed-member
+endif
 

You don't need to do this, as gcc will not complain about this unknown

flag

Quoted 1 line
Quoted 1 line

unless you have other issues in your code.[1] I think it's better to

keep

Quoted 9 lines
Quoted 9 lines

the

Quoted 8 lines
Quoted 8 lines

code clean in this case, otherwise we'll have the code littered with conditionals for various flags.

/Bruce

[1] https://gcc.gnu.org/onlinedocs/gcc/Warning-Options.html
"When an unrecognized warning option is requested (e.g., -Wunknown-warning), GCC emits a diagnostic stating that the option

is

Quoted 2 lines
Quoted 2 lines

not

Quoted 1 line
Quoted 1 line

recognized. However, if the -Wno- form is used, the behavior is

slightly

Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

different: no diagnostic is produced for -Wno-unknown-warning

unless

Quoted 2 lines
Quoted 2 lines

other

Quoted 1 line
Quoted 1 line

diagnostics are being produced. This allows the use of new -Wno-

options

Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

with old compilers, but if something goes wrong, the compiler warns

that

Quoted 8 lines
Quoted 8 lines

an

Quoted 5 lines
Quoted 1 line

unrecognized option is present."

Thanks, Bruce.

There are thousands of warnings with alignment when compiling for

ARMv7 (ARMv7 supports unaligned memory access).

Quoted 1 line

Without this new flag, the build system works fine for ARMv7. So I add

this

Quoted 2 lines
Quoted 2 lines

conditional for ARMv7 only.

Quoted 1 line

I agree with you we should clean up the code, but it needs a lot of

effort to

Quoted 2 lines
Quoted 2 lines

achieve that.

Quoted 1 line

This patch is going to make the ARMv7 target build successfully during

this

Quoted 4 lines
Quoted 4 lines

period.

I would go the opposite direction.
I think we should remove all warnings disablement like "-Wno-

something",

Quoted 5 lines
Quoted 1 line

and clean the code.

Thanks, Thomas.

I think it is a compiler issue rather than a code defect in this case, as It

cannot recognize the '-Waddress-of-packed-member' flag.

Quoted 1 line

"arm-linux-gnueabihf-gcc: error: unrecognized command line option '-

Waddress-of-packed-member'"

Quoted 7 lines
Quoted 5 lines

If you want to allow building ARMv7, you can just allow warnings as not errors.
So we will still see something is wrong but it won't prevent from building.

Yes. We did this already, all the alignment issues are emitted as warnings

Quoted 8 lines

on ARMv7.

Quoted 3 lines

Please see these two commits:
79fa5c2a9992 ("mk: ignore alignment errors for ARMv7") 6433c19105e8 ("mk: generalize strict alignment warning handling")

I don't know what to do with this patch.
I think it is not how it should work,
but you say your compiler has a real bug.

Agree with you, however I haven't found a better solution for this issue yet.

Quoted 2 lines

Can you confirm which compiler is it?

Please check the packet information dumped below, it is the latest version in Ubuntu 18.04 LTS.

$ sudo apt show gcc-arm-linux-gnueabihf
Package: gcc-arm-linux-gnueabihf
Version: 4:7.4.0-1ubuntu2.3
Priority: optional
Section: devel
Source: gcc-defaults (1.176ubuntu2.3)
Origin: Ubuntu
Quoted 1 line

Does it prevent to compile even with -Wno-error ?

Yes, it does. I put one piece of the compiling log bellow. Please take a look.

$ make -j1 CROSS=arm-linux-gnueabihf- T=arm-armv7a-linuxapp-gcc RTE_DEVEL_BUILD=y
...
/home/phil/aarch32_dpdk/dpdk/lib/librte_eal/linux/eal/eal_vfio.c:1178:7: warning: cast increases required alignment of target type [-Wcast-align]
   p = (struct vfio_mp_param *)mp_rep->param;
       ^
/home/phil/aarch32_dpdk/dpdk/lib/librte_eal/linux/eal/eal_vfio.c: At top level:
cc1: error: unrecognized command line option '-Wno-address-of-packed-member' [-Werror]
...
Quoted 3 lines

Any other opinion?

#13 lore raw
Quoted 142 lines · Phil Yang (Arm Technology China)
Quoted 120 lines

-----Original Message-----
From: Thomas Monjalon <[email protected]>
Sent: Wednesday, November 27, 2019 1:02 AM
To: Phil Yang (Arm Technology China) <[email protected]>
Cc: [email protected]; Bruce Richardson <[email protected]>; [email protected]; [email protected]; Honnappa Nagarahalli <[email protected]>; Gavin Hu (Arm Technology China) <[email protected]>; Joyce Kong (Arm Technology China) <[email protected]>; nd <[email protected]>; [email protected]
Subject: Re: [dpdk-dev] [PATCH 1/2] mk: fix unsupported flag error on armhf architercture

12/11/2019 07:40, Phil Yang (Arm Technology China):

Quoted 6 lines
Quoted 6 lines

-----Original Message-----
From: Thomas Monjalon <[email protected]>
Sent: Tuesday, November 12, 2019 2:03 PM
To: Phil Yang (Arm Technology China) <[email protected]>
Cc: Bruce Richardson <[email protected]>; [email protected]; [email protected]; [email protected]; Honnappa

Nagarahalli

Quoted 4 lines
Quoted 4 lines

<[email protected]>; Gavin Hu (Arm Technology China) <[email protected]>; Joyce Kong (Arm Technology China) <[email protected]>; nd <[email protected]>; [email protected]
Subject: Re: [dpdk-dev] [PATCH 1/2] mk: fix unsupported flag error on

armhf

Quoted 6 lines
Quoted 6 lines

architercture

12/11/2019 06:25, Phil Yang (Arm Technology China):

Quoted 3 lines

From: Bruce Richardson <[email protected]>

Quoted 2 lines · Phil Yang
Quoted 1 line

The older version (e.g. version 7.4.0 ) of GNU C compiler for the

armhf

Quoted 5 lines
Quoted 5 lines
Quoted 1 line
Quoted 1 line
Quoted 1 line

architecture doesn't support the flag '-Wno-address-of-packed-

member',

Quoted 3 lines
Quoted 3 lines
Quoted 3 lines

so remove this flag for aarch32.

Fixes a385972c3675 ("mk: disable warning for packed member

pointer")

Quoted 18 lines
Quoted 18 lines
Quoted 11 lines
Quoted 11 lines

Review

Quoted 11 lines
---
 mk/toolchain/gcc/rte.vars.mk | 2 ++
 1 file changed, 2 insertions(+)
(quoted diff) +0 -0

b/mk/toolchain/gcc/rte.vars.mk

Quoted 6 lines
Quoted 6 lines
Quoted 6 lines

index 9fc7041..ec0cbbf 100644

mk/toolchain/gcc/rte.vars.mk +0 -0
@@ -100,7 +100,9 @@ WERROR_FLAGS += -Wno-format-truncation
 endif
 
Quoted 8 lines
Quoted 8 lines
Quoted 6 lines
Quoted 6 lines

mk/toolchain/gcc/rte.vars.mk WERROR_FLAGS += -Wno-format-truncation · line 106

Quoted 5 lines
mk/toolchain/gcc/rte.vars.mk +2 -0
 # disable packed member unalign warnings
+ifneq ($(CONFIG_RTE_ARCH_ARM), y)
 WERROR_FLAGS += -Wno-address-of-packed-member
+endif
 

You don't need to do this, as gcc will not complain about this unknown

flag

Quoted 1 line
Quoted 1 line

unless you have other issues in your code.[1] I think it's better to

keep

Quoted 9 lines
Quoted 9 lines

the

Quoted 8 lines
Quoted 8 lines

code clean in this case, otherwise we'll have the code littered with conditionals for various flags.

/Bruce

[1] https://gcc.gnu.org/onlinedocs/gcc/Warning-Options.html
"When an unrecognized warning option is requested (e.g., -Wunknown-warning), GCC emits a diagnostic stating that the option

is

Quoted 2 lines
Quoted 2 lines

not

Quoted 1 line
Quoted 1 line

recognized. However, if the -Wno- form is used, the behavior is

slightly

Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

different: no diagnostic is produced for -Wno-unknown-warning

unless

Quoted 2 lines
Quoted 2 lines

other

Quoted 1 line
Quoted 1 line

diagnostics are being produced. This allows the use of new -Wno-

options

Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

with old compilers, but if something goes wrong, the compiler warns

that

Quoted 8 lines
Quoted 8 lines

an

Quoted 5 lines
Quoted 1 line

unrecognized option is present."

Thanks, Bruce.

There are thousands of warnings with alignment when compiling for

ARMv7 (ARMv7 supports unaligned memory access).

Quoted 1 line

Without this new flag, the build system works fine for ARMv7. So I add

this

Quoted 2 lines
Quoted 2 lines

conditional for ARMv7 only.

Quoted 1 line

I agree with you we should clean up the code, but it needs a lot of

effort to

Quoted 2 lines
Quoted 2 lines

achieve that.

Quoted 1 line

This patch is going to make the ARMv7 target build successfully during

this

Quoted 4 lines
Quoted 4 lines

period.

I would go the opposite direction.
I think we should remove all warnings disablement like "-Wno-

something",

Quoted 5 lines
Quoted 1 line

and clean the code.

Thanks, Thomas.

I think it is a compiler issue rather than a code defect in this case, as It

cannot recognize the '-Waddress-of-packed-member' flag.

Quoted 1 line

"arm-linux-gnueabihf-gcc: error: unrecognized command line option '-

Waddress-of-packed-member'"

Quoted 7 lines
Quoted 5 lines

If you want to allow building ARMv7, you can just allow warnings as not errors.
So we will still see something is wrong but it won't prevent from building.

Yes. We did this already, all the alignment issues are emitted as warnings

Quoted 8 lines

on ARMv7.

Quoted 3 lines

Please see these two commits:
79fa5c2a9992 ("mk: ignore alignment errors for ARMv7") 6433c19105e8 ("mk: generalize strict alignment warning handling")

I don't know what to do with this patch.
I think it is not how it should work,
but you say your compiler has a real bug.

Agree with you, however I haven't found a better solution for this issue yet.

Quoted 2 lines

Can you confirm which compiler is it?

Please check the packet information dumped below, it is the latest version in Ubuntu 18.04 LTS.

$ sudo apt show gcc-arm-linux-gnueabihf
Package: gcc-arm-linux-gnueabihf
Version: 4:7.4.0-1ubuntu2.3
Priority: optional
Section: devel
Source: gcc-defaults (1.176ubuntu2.3)
Origin: Ubuntu
Quoted 4 lines
Quoted 1 line

Does it prevent to compile even with -Wno-error ?

Yes, it does. I put one piece of the compiling log bellow. Please take a look.

$ make -j1 CROSS=arm-linux-gnueabihf- T=arm-armv7a-linuxapp-gcc RTE_DEVEL_BUILD=y

RTE_DEVEL_BUILD turns on -Werror, while what Thomas is asking about is using the opposite flag "-Wno-error". Can you do a non-developer build or put "-Wno-error" in EXTRA_CFLAGS to test this.

/Bruce

#14 lore raw
Quoted 170 lines

-----Original Message-----
From: Bruce Richardson <[email protected]>
Sent: Wednesday, November 27, 2019 5:30 PM
To: Phil Yang (Arm Technology China) <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; Honnappa Nagarahalli <[email protected]>; Gavin Hu (Arm Technology China) <[email protected]>; Joyce Kong (Arm Technology China) <[email protected]>; nd <[email protected]>; [email protected]
Subject: Re: [dpdk-dev] [PATCH 1/2] mk: fix unsupported flag error on armhf architercture

Quoted 6 lines · Phil Yang (Arm Technology China)
Quoted 6 lines

-----Original Message-----
From: Thomas Monjalon <[email protected]>
Sent: Wednesday, November 27, 2019 1:02 AM
To: Phil Yang (Arm Technology China) <[email protected]>
Cc: [email protected]; Bruce Richardson <[email protected]>; [email protected]; [email protected]; Honnappa

Nagarahalli

Quoted 4 lines
Quoted 4 lines

<[email protected]>; Gavin Hu (Arm Technology China) <[email protected]>; Joyce Kong (Arm Technology China) <[email protected]>; nd <[email protected]>; [email protected]
Subject: Re: [dpdk-dev] [PATCH 1/2] mk: fix unsupported flag error on

armhf

Quoted 9 lines
Quoted 9 lines

architercture

12/11/2019 07:40, Phil Yang (Arm Technology China):

Quoted 6 lines
Quoted 6 lines

-----Original Message-----
From: Thomas Monjalon <[email protected]>
Sent: Tuesday, November 12, 2019 2:03 PM
To: Phil Yang (Arm Technology China) <[email protected]>
Cc: Bruce Richardson <[email protected]>; [email protected]; [email protected]; [email protected]; Honnappa

Quoted 5 lines
Quoted 5 lines

Nagarahalli

Quoted 4 lines
Quoted 4 lines

<[email protected]>; Gavin Hu (Arm Technology China) <[email protected]>; Joyce Kong (Arm Technology China) <[email protected]>; nd <[email protected]>; [email protected]
Subject: Re: [dpdk-dev] [PATCH 1/2] mk: fix unsupported flag error

on

Quoted 24 lines
Quoted 24 lines

armhf

Quoted 6 lines
Quoted 6 lines

architercture

12/11/2019 06:25, Phil Yang (Arm Technology China):

Quoted 3 lines

From: Bruce Richardson <[email protected]>

Quoted 2 lines · Phil Yang
Quoted 1 line

The older version (e.g. version 7.4.0 ) of GNU C compiler for the

armhf

Quoted 5 lines
Quoted 5 lines
Quoted 1 line
Quoted 1 line
Quoted 1 line

architecture doesn't support the flag '-Wno-address-of-packed-

member',

Quoted 3 lines
Quoted 3 lines
Quoted 3 lines

so remove this flag for aarch32.

Fixes a385972c3675 ("mk: disable warning for packed member

pointer")

Quoted 10 lines
Quoted 10 lines
Quoted 10 lines
Quoted 10 lines
Quoted 10 lines
---
 mk/toolchain/gcc/rte.vars.mk | 2 ++
 1 file changed, 2 insertions(+)
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 1 line
Quoted 1 line

Review

Quoted 1 line
(quoted diff) +0 -0

b/mk/toolchain/gcc/rte.vars.mk

Quoted 4 lines
Quoted 4 lines

mk/toolchain/gcc/rte.vars.mk WERROR_FLAGS += -Wno-format-

Quoted 4 lines

index 9fc7041..ec0cbbf 100644

mk/toolchain/gcc/rte.vars.mk +0 -0
@@ -100,7 +100,9 @@ WERROR_FLAGS += -Wno-format-

truncation

Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines

mk/toolchain/gcc/rte.vars.mk WERROR_FLAGS += -Wno-format- · line 106

Quoted 7 lines
mk/toolchain/gcc/rte.vars.mk +2 -0
 endif
 
 # disable packed member unalign warnings
+ifneq ($(CONFIG_RTE_ARCH_ARM), y)
 WERROR_FLAGS += -Wno-address-of-packed-member
+endif
 

You don't need to do this, as gcc will not complain about this

unknown

Quoted 4 lines
Quoted 4 lines
Quoted 2 lines
Quoted 2 lines

flag

Quoted 1 line
Quoted 1 line

unless you have other issues in your code.[1] I think it's better to

keep

Quoted 1 line
Quoted 1 line

the

Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines

code clean in this case, otherwise we'll have the code littered with conditionals for various flags.

/Bruce

[1] https://gcc.gnu.org/onlinedocs/gcc/Warning-Options.html
"When an unrecognized warning option is requested (e.g., -Wunknown-warning), GCC emits a diagnostic stating that the

option

Quoted 10 lines
Quoted 10 lines

is

Quoted 2 lines
Quoted 2 lines

not

Quoted 1 line
Quoted 1 line

recognized. However, if the -Wno- form is used, the behavior is

slightly

Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

different: no diagnostic is produced for -Wno-unknown-warning

unless

Quoted 2 lines
Quoted 2 lines

other

Quoted 1 line
Quoted 1 line

diagnostics are being produced. This allows the use of new -Wno-

options

Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

with old compilers, but if something goes wrong, the compiler

warns

Quoted 9 lines
Quoted 9 lines

that

Quoted 8 lines
Quoted 8 lines

an

Quoted 5 lines
Quoted 1 line

unrecognized option is present."

Thanks, Bruce.

There are thousands of warnings with alignment when compiling for

ARMv7 (ARMv7 supports unaligned memory access).

Quoted 1 line

Without this new flag, the build system works fine for ARMv7. So I

add

Quoted 6 lines
Quoted 6 lines

this

Quoted 2 lines
Quoted 2 lines

conditional for ARMv7 only.

Quoted 1 line

I agree with you we should clean up the code, but it needs a lot of

effort to

Quoted 2 lines
Quoted 2 lines

achieve that.

Quoted 1 line

This patch is going to make the ARMv7 target build successfully

during

Quoted 10 lines
Quoted 10 lines

this

Quoted 4 lines
Quoted 4 lines

period.

I would go the opposite direction.
I think we should remove all warnings disablement like "-Wno-

something",

Quoted 4 lines
Quoted 1 line

and clean the code.

Thanks, Thomas.

Quoted 7 lines
Quoted 7 lines
Quoted 1 line

I think it is a compiler issue rather than a code defect in this case, as It

cannot recognize the '-Waddress-of-packed-member' flag.

Quoted 1 line

"arm-linux-gnueabihf-gcc: error: unrecognized command line option '-

Waddress-of-packed-member'"

Quoted 3 lines
Quoted 3 lines

If you want to allow building ARMv7, you can just allow warnings as

not

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

errors.
So we will still see something is wrong but it won't prevent from

building.

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

Yes. We did this already, all the alignment issues are emitted as

warnings

Quoted 8 lines
Quoted 8 lines

on ARMv7.

Quoted 3 lines

Please see these two commits:
79fa5c2a9992 ("mk: ignore alignment errors for ARMv7") 6433c19105e8 ("mk: generalize strict alignment warning handling")

I don't know what to do with this patch.
I think it is not how it should work,
but you say your compiler has a real bug.

Quoted 6 lines

Agree with you, however I haven't found a better solution for this issue yet.

Quoted 2 lines

Can you confirm which compiler is it?

Please check the packet information dumped below, it is the latest version

in Ubuntu 18.04 LTS.

Quoted 9 lines
$ sudo apt show gcc-arm-linux-gnueabihf
Package: gcc-arm-linux-gnueabihf
Version: 4:7.4.0-1ubuntu2.3
Priority: optional
Section: devel
Source: gcc-defaults (1.176ubuntu2.3)
Origin: Ubuntu
Quoted 1 line

Does it prevent to compile even with -Wno-error ?

Quoted 8 lines
Quoted 3 lines

Yes, it does. I put one piece of the compiling log bellow. Please take a look.

$ make -j1 CROSS=arm-linux-gnueabihf- T=arm-armv7a-linuxapp-gcc

RTE_DEVEL_BUILD=y

RTE_DEVEL_BUILD turns on -Werror, while what Thomas is asking about is using the opposite flag "-Wno-error". Can you do a non-developer build or put "-Wno-error" in EXTRA_CFLAGS to test this.

Thanks, Bruce.
There is no compiling issue for the non-developer build and "-Wno-error" flag build.

#15 lore raw

27/11/2019 11:00, Phil Yang (Arm Technology China):

Quoted 173 lines
Quoted 173 lines

-----Original Message-----
From: Bruce Richardson <[email protected]>
Sent: Wednesday, November 27, 2019 5:30 PM
To: Phil Yang (Arm Technology China) <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; Honnappa Nagarahalli <[email protected]>; Gavin Hu (Arm Technology China) <[email protected]>; Joyce Kong (Arm Technology China) <[email protected]>; nd <[email protected]>; [email protected]
Subject: Re: [dpdk-dev] [PATCH 1/2] mk: fix unsupported flag error on armhf architercture

Quoted 6 lines · Phil Yang (Arm Technology China)
Quoted 6 lines

-----Original Message-----
From: Thomas Monjalon <[email protected]>
Sent: Wednesday, November 27, 2019 1:02 AM
To: Phil Yang (Arm Technology China) <[email protected]>
Cc: [email protected]; Bruce Richardson <[email protected]>; [email protected]; [email protected]; Honnappa

Nagarahalli

Quoted 4 lines
Quoted 4 lines

<[email protected]>; Gavin Hu (Arm Technology China) <[email protected]>; Joyce Kong (Arm Technology China) <[email protected]>; nd <[email protected]>; [email protected]
Subject: Re: [dpdk-dev] [PATCH 1/2] mk: fix unsupported flag error on

armhf

Quoted 9 lines
Quoted 9 lines

architercture

12/11/2019 07:40, Phil Yang (Arm Technology China):

Quoted 6 lines
Quoted 6 lines

-----Original Message-----
From: Thomas Monjalon <[email protected]>
Sent: Tuesday, November 12, 2019 2:03 PM
To: Phil Yang (Arm Technology China) <[email protected]>
Cc: Bruce Richardson <[email protected]>; [email protected]; [email protected]; [email protected]; Honnappa

Quoted 5 lines
Quoted 5 lines

Nagarahalli

Quoted 4 lines
Quoted 4 lines

<[email protected]>; Gavin Hu (Arm Technology China) <[email protected]>; Joyce Kong (Arm Technology China) <[email protected]>; nd <[email protected]>; [email protected]
Subject: Re: [dpdk-dev] [PATCH 1/2] mk: fix unsupported flag error

on

Quoted 24 lines
Quoted 24 lines

armhf

Quoted 6 lines
Quoted 6 lines

architercture

12/11/2019 06:25, Phil Yang (Arm Technology China):

Quoted 3 lines

From: Bruce Richardson <[email protected]>

Quoted 2 lines · Phil Yang
Quoted 1 line

The older version (e.g. version 7.4.0 ) of GNU C compiler for the

armhf

Quoted 5 lines
Quoted 5 lines
Quoted 1 line
Quoted 1 line
Quoted 1 line

architecture doesn't support the flag '-Wno-address-of-packed-

member',

Quoted 3 lines
Quoted 3 lines
Quoted 3 lines

so remove this flag for aarch32.

Fixes a385972c3675 ("mk: disable warning for packed member

pointer")

Quoted 10 lines
Quoted 10 lines
Quoted 10 lines
Quoted 10 lines
Quoted 10 lines
---
 mk/toolchain/gcc/rte.vars.mk | 2 ++
 1 file changed, 2 insertions(+)
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
Quoted 1 line
Quoted 1 line

Review

Quoted 1 line
(quoted diff) +0 -0

b/mk/toolchain/gcc/rte.vars.mk

Quoted 4 lines
Quoted 4 lines

mk/toolchain/gcc/rte.vars.mk WERROR_FLAGS += -Wno-format-

Quoted 4 lines

index 9fc7041..ec0cbbf 100644

mk/toolchain/gcc/rte.vars.mk +0 -0
@@ -100,7 +100,9 @@ WERROR_FLAGS += -Wno-format-

truncation

Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines

mk/toolchain/gcc/rte.vars.mk WERROR_FLAGS += -Wno-format- · line 106

Quoted 7 lines
mk/toolchain/gcc/rte.vars.mk +2 -0
 endif
 
 # disable packed member unalign warnings
+ifneq ($(CONFIG_RTE_ARCH_ARM), y)
 WERROR_FLAGS += -Wno-address-of-packed-member
+endif
 

You don't need to do this, as gcc will not complain about this

unknown

Quoted 4 lines
Quoted 4 lines
Quoted 2 lines
Quoted 2 lines

flag

Quoted 1 line
Quoted 1 line

unless you have other issues in your code.[1] I think it's better to

keep

Quoted 1 line
Quoted 1 line

the

Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines

code clean in this case, otherwise we'll have the code littered with conditionals for various flags.

/Bruce

[1] https://gcc.gnu.org/onlinedocs/gcc/Warning-Options.html
"When an unrecognized warning option is requested (e.g., -Wunknown-warning), GCC emits a diagnostic stating that the

option

Quoted 10 lines
Quoted 10 lines

is

Quoted 2 lines
Quoted 2 lines

not

Quoted 1 line
Quoted 1 line

recognized. However, if the -Wno- form is used, the behavior is

slightly

Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

different: no diagnostic is produced for -Wno-unknown-warning

unless

Quoted 2 lines
Quoted 2 lines

other

Quoted 1 line
Quoted 1 line

diagnostics are being produced. This allows the use of new -Wno-

options

Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

with old compilers, but if something goes wrong, the compiler

warns

Quoted 9 lines
Quoted 9 lines

that

Quoted 8 lines
Quoted 8 lines

an

Quoted 5 lines
Quoted 1 line

unrecognized option is present."

Thanks, Bruce.

There are thousands of warnings with alignment when compiling for

ARMv7 (ARMv7 supports unaligned memory access).

Quoted 1 line

Without this new flag, the build system works fine for ARMv7. So I

add

Quoted 6 lines
Quoted 6 lines

this

Quoted 2 lines
Quoted 2 lines

conditional for ARMv7 only.

Quoted 1 line

I agree with you we should clean up the code, but it needs a lot of

effort to

Quoted 2 lines
Quoted 2 lines

achieve that.

Quoted 1 line

This patch is going to make the ARMv7 target build successfully

during

Quoted 10 lines
Quoted 10 lines

this

Quoted 4 lines
Quoted 4 lines

period.

I would go the opposite direction.
I think we should remove all warnings disablement like "-Wno-

something",

Quoted 4 lines
Quoted 1 line

and clean the code.

Thanks, Thomas.

Quoted 7 lines
Quoted 7 lines
Quoted 1 line

I think it is a compiler issue rather than a code defect in this case, as It

cannot recognize the '-Waddress-of-packed-member' flag.

Quoted 1 line

"arm-linux-gnueabihf-gcc: error: unrecognized command line option '-

Waddress-of-packed-member'"

Quoted 3 lines
Quoted 3 lines

If you want to allow building ARMv7, you can just allow warnings as

not

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

errors.
So we will still see something is wrong but it won't prevent from

building.

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

Yes. We did this already, all the alignment issues are emitted as

warnings

Quoted 8 lines
Quoted 8 lines

on ARMv7.

Quoted 3 lines

Please see these two commits:
79fa5c2a9992 ("mk: ignore alignment errors for ARMv7") 6433c19105e8 ("mk: generalize strict alignment warning handling")

I don't know what to do with this patch.
I think it is not how it should work,
but you say your compiler has a real bug.

Quoted 6 lines

Agree with you, however I haven't found a better solution for this issue yet.

Quoted 2 lines

Can you confirm which compiler is it?

Please check the packet information dumped below, it is the latest version

in Ubuntu 18.04 LTS.

Quoted 12 lines
$ sudo apt show gcc-arm-linux-gnueabihf
Package: gcc-arm-linux-gnueabihf
Version: 4:7.4.0-1ubuntu2.3
Priority: optional
Section: devel
Source: gcc-defaults (1.176ubuntu2.3)
Origin: Ubuntu
Quoted 1 line

Does it prevent to compile even with -Wno-error ?

Yes, it does. I put one piece of the compiling log bellow. Please take a look.

$ make -j1 CROSS=arm-linux-gnueabihf- T=arm-armv7a-linuxapp-gcc
Quoted 8 lines
Quoted 5 lines

RTE_DEVEL_BUILD=y

RTE_DEVEL_BUILD turns on -Werror, while what Thomas is asking about is using the opposite flag "-Wno-error". Can you do a non-developer build or put "-Wno-error" in EXTRA_CFLAGS to test this.

Thanks, Bruce.
There is no compiling issue for the non-developer build and "-Wno-error" flag build.

Cool, so we can classify this patch as "rejected".