[Buildroot] [PATCH/master 0/7] Qemu and gobject-introspection fixes

[Buildroot] [PATCH/master 0/7] Qemu and gobject-introspection fixes

[Adam Duskett]

First of all, sorry this took so long!

This series is an accumulation of several conversations between Thomas, Yann,
and myself. The basics are this:

- GOI fails to build with the corei7 architecture because calling qemu results
  in qemu trying to use instructions that may not exist on the host processor.
  This causes a core dump.

- OpenEmbedded gets around this issue by adding options to qemu on certain
  CPU and architecture combinations, Corei7 being one of those architectures.

- As discussed with Thomas, a new string "BR2_PACKAGE_HOST_QEMU_USER_MODE_ARGS"
  in qemu/Config.in.host is added, which we then sed to g-ir-scanner.in.

In addition, I was anonyed that there were so many shellcheck issues with the
various wrappers, and Yann was annoyed that the wrappers were calling to bash
when they didn't need to do so, so those are also fixed in this patch series.

And last (but not least), Yann noticed a latent bug in g-ir-scanner-qemuwrapper.
If GIR_EXTRA_LIBS_PATH is empty, then the LD_LIBRARY_PATH is prepended with a
colon, so that is also fixed.

This will lead the way to NodeJS 14 which is one step closer to removing
Python2!

Tested in a CentOS7 docker container:
                     andes-nds32 [ 1/45]: SKIPPED
                     arm-aarch64 [ 2/45]: OK
           bootlin-aarch64-glibc [ 3/45]: OK
       bootlin-arcle-hs38-uclibc [ 4/45]: SKIPPED
            bootlin-armv5-uclibc [ 5/45]: SKIPPED
             bootlin-armv7-glibc [ 6/45]: OK
           bootlin-armv7m-uclibc [ 7/45]: SKIPPED
              bootlin-armv7-musl [ 8/45]: SKIPPED
        bootlin-m68k-5208-uclibc [ 9/45]: SKIPPED
       bootlin-m68k-68040-uclibc [10/45]: SKIPPED
     bootlin-microblazeel-uclibc [11/45]: SKIPPED
        bootlin-mipsel32r6-glibc [12/45]: OK
           bootlin-mipsel-uclibc [13/45]: SKIPPED
             bootlin-nios2-glibc [14/45]: OK
         bootlin-openrisc-uclibc [15/45]: SKIPPED
bootlin-powerpc64le-power8-glibc [16/45]: OK
   bootlin-powerpc-e500mc-uclibc [17/45]: SKIPPED
           bootlin-riscv32-glibc [18/45]: OK
           bootlin-riscv64-glibc [19/45]: OK
            bootlin-riscv64-musl [20/45]: SKIPPED
              bootlin-sh4-uclibc [21/45]: SKIPPED
           bootlin-sparc64-glibc [22/45]: OK
            bootlin-sparc-uclibc [23/45]: SKIPPED
            bootlin-x86-64-glibc [24/45]: OK
             bootlin-x86-64-musl [25/45]: SKIPPED
           bootlin-x86-64-uclibc [26/45]: SKIPPED
           bootlin-xtensa-uclibc [27/45]: SKIPPED
                    br-arm-basic [28/45]: SKIPPED
            br-arm-full-nothread [29/45]: SKIPPED
              br-arm-full-static [30/45]: SKIPPED
           br-i386-pentium4-full [31/45]: SKIPPED
        br-i386-pentium-mmx-musl [32/45]: SKIPPED
              br-mips64-n64-full [33/45]: SKIPPED
         br-mips64r6-el-hf-glibc [34/45]: SKIPPED
       br-powerpc-603e-basic-cpp [35/45]: SKIPPED
       br-powerpc64-power7-glibc [36/45]: OK
               linaro-aarch64-be [37/45]: SKIPPED
                  linaro-aarch64 [38/45]: OK
                      linaro-arm [39/45]: OK
             sourcery-arm-armv4t [40/45]: SKIPPED
                    sourcery-arm [41/45]: SKIPPED
             sourcery-arm-thumb2 [42/45]: SKIPPED
                 sourcery-mips64 [43/45]: SKIPPED
                   sourcery-mips [44/45]: OK
                  sourcery-nios2 [45/45]: OK


Adam Duskett (7):
  g-ir-scanner-qemuwrapper.in: Fix latent bug in LD_LIBRARY_PATH string
  g-ir-scanner-qemuwrapper.in: Fix shellcheck warnings, switch to sh
  g-ir-scanner.in: Fix shellcheck warnings, switch to sh
  g-ir-scanner-lddwrapper.in: Fix shellcheck warnings, switch to sh
  g-ir-compiler.in: Fix shellcheck warnings, switch to sh
  package/qemu: add BR2_PACKAGE_HOST_QEMU_USER_MODE_ARGS string
  package/gobject-introspection: Add QEMU_USERMODE_ARGS support

 .../gobject-introspection/g-ir-compiler.in    |  4 ++--
 .../g-ir-scanner-lddwrapper.in                |  4 ++--
 .../g-ir-scanner-qemuwrapper.in               | 19 ++++++++++++-------
 package/gobject-introspection/g-ir-scanner.in | 12 ++++++------
 .../gobject-introspection.mk                  |  2 ++
 package/qemu/Config.in.host                   |  4 ++++
 6 files changed, 28 insertions(+), 17 deletions(-)

-- 
2.30.2

_______________________________________________
buildroot mailing list
buildroot@lists.buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

[Buildroot] [PATCH/master 1/7] g-ir-scanner-qemuwrapper.in: Fix latent bug in

[Adam Duskett]

If GIR_EXTRA_LIBS_PATH is empty, the LD_LIBRARY_PATH variable is
":.libs:$(dirname $0)/../lib:$(dirname $0)/../../lib".

As discussed with Yann Morin, add a new variable, "GOI_LIBRARY_PATH" and
prepend ${GIR_EXTRA_LIBS_PATH:+${GIR_EXTRA_LIBS_PATH}:} to the path to fix the
above bug.

Signed-off-by: Adam Duskett <aduskett@gmail.com>
---
 package/gobject-introspection/g-ir-scanner-qemuwrapper.in | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/package/gobject-introspection/g-ir-scanner-qemuwrapper.in b/package/gobject-introspection/g-ir-scanner-qemuwrapper.in
index cad7ef875e..ec066062e2 100644
--- a/package/gobject-introspection/g-ir-scanner-qemuwrapper.in
+++ b/package/gobject-introspection/g-ir-scanner-qemuwrapper.in
@@ -4,10 +4,11 @@
 # is older than the target kernel.
 # Use a modules directory which does not exist so we don't load random things
 # which may then get deleted (or their dependencies) and potentially segfault
+GOI_LIBRARY_PATH="${GIR_EXTRA_LIBS_PATH:+${GIR_EXTRA_LIBS_PATH}:}.libs:$(dirname "$0")/../lib:$(dirname "$0")/../../lib"
 GIO_MODULE_DIR=$(dirname $0)/../lib/gio/modules-dummy \
 @QEMU_USER@ -r @TOOLCHAIN_HEADERS_VERSION@ \
 -L $(dirname $0)/../../ \
--E LD_LIBRARY_PATH=${GIR_EXTRA_LIBS_PATH}:.libs:$(dirname $0)/../lib:$(dirname $0)/../../lib \
+-E LD_LIBRARY_PATH="${GOI_LIBRARY_PATH}" \
 "$@"
 
 if [[ $? -ne 0 ]]; then
-- 
2.30.2

_______________________________________________
buildroot mailing list
buildroot@lists.buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

[Buildroot] [PATCH/master 2/7] g-ir-scanner-qemuwrapper.in: Fix shellcheck warnings, switch to sh

[Adam Duskett]

  - Add double quotes to prevent globbing and word splitting.
  - Add indentations of continuation lines.
  - Disable SC2181 and SC2016 as we explicitly do not want the variables
    expanded in the echo.

Signed-off-by: Adam Duskett <aduskett@gmail.com>
---
 .../g-ir-scanner-qemuwrapper.in                  | 16 +++++++++-------
 1 file changed, 9 insertions(+), 7 deletions(-)

diff --git a/package/gobject-introspection/g-ir-scanner-qemuwrapper.in b/package/gobject-introspection/g-ir-scanner-qemuwrapper.in
index ec066062e2..5ece75d0a4 100644
--- a/package/gobject-introspection/g-ir-scanner-qemuwrapper.in
+++ b/package/gobject-introspection/g-ir-scanner-qemuwrapper.in
@@ -1,18 +1,20 @@
-#!/usr/bin/env bash
+#!/usr/bin/env sh
 
-# Pass -r to qemu-user as to trick glibc into not errorings out if the host kernel
+# Pass -r to qemu-user as to trick glibc into not erroring out if the host kernel
 # is older than the target kernel.
 # Use a modules directory which does not exist so we don't load random things
 # which may then get deleted (or their dependencies) and potentially segfault
 GOI_LIBRARY_PATH="${GIR_EXTRA_LIBS_PATH:+${GIR_EXTRA_LIBS_PATH}:}.libs:$(dirname "$0")/../lib:$(dirname "$0")/../../lib"
-GIO_MODULE_DIR=$(dirname $0)/../lib/gio/modules-dummy \
+GIO_MODULE_DIR="$(dirname "$0")/../lib/gio/modules-dummy" \
 @QEMU_USER@ -r @TOOLCHAIN_HEADERS_VERSION@ \
--L $(dirname $0)/../../ \
--E LD_LIBRARY_PATH="${GOI_LIBRARY_PATH}" \
-"$@"
+    -L "$(dirname "$0")/../../" \
+    -E LD_LIBRARY_PATH="${GOI_LIBRARY_PATH}" \
+    "$@"
 
-if [[ $? -ne 0 ]]; then
+# shellcheck disable=SC2181
+if [ $? -ne 0 ]; then
     echo "If the above error message is about missing .so libraries, then setting up GIR_EXTRA_LIBS_PATH in the .mk file should help."
+    # shellcheck disable=SC2016
     echo 'Typically like this: PKG_MAKE_ENV += GIR_EXTRA_LIBS_PATH="$(@D)/.libs"'
     exit 1
 fi
-- 
2.30.2

_______________________________________________
buildroot mailing list
buildroot@lists.buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

[Buildroot] [PATCH/master 3/7] g-ir-scanner.in: Fix shellcheck warnings, switch to sh

[Adam Duskett]

Signed-off-by: Adam Duskett <aduskett@gmail.com>
---
 package/gobject-introspection/g-ir-scanner.in | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/package/gobject-introspection/g-ir-scanner.in b/package/gobject-introspection/g-ir-scanner.in
index 22df7be309..013eb15700 100644
--- a/package/gobject-introspection/g-ir-scanner.in
+++ b/package/gobject-introspection/g-ir-scanner.in
@@ -1,4 +1,4 @@
-#!/usr/bin/env bash
+#!/usr/bin/env sh
 
 # These environment variables are empty by default, and as such ccompiler.py
 # defaults to either using the system tools (CC, CXX, and CPP) or leaving
@@ -15,8 +15,8 @@ export LDFLAGS="${LDFLAGS:-@TARGET_LDFLAGS@}"
 
 export GI_SCANNER_DISABLE_CACHE=1
 
-${HOST_DIR}/bin/g-ir-scanner \
---lib-dirs-envvar=GIR_EXTRA_LIBS_PATH \
---use-binary-wrapper=$(dirname $0)/g-ir-scanner-qemuwrapper \
---use-ldd-wrapper=$(dirname $0)/g-ir-scanner-lddwrapper \
---add-include-path=$(dirname $0)/../share/gir-1.0 "$@"
+"${HOST_DIR}"/bin/g-ir-scanner \
+    --lib-dirs-envvar=GIR_EXTRA_LIBS_PATH \
+    --use-binary-wrapper="$(dirname "$0")"/g-ir-scanner-qemuwrapper \
+    --use-ldd-wrapper="$(dirname "$0")"/g-ir-scanner-lddwrapper \
+    --add-include-path="$(dirname "$0")"/../share/gir-1.0 "$@"
-- 
2.30.2

_______________________________________________
buildroot mailing list
buildroot@lists.buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

[Buildroot] [PATCH/master 4/7] g-ir-scanner-lddwrapper.in: Fix shellcheck warnings, switch to sh

[Adam Duskett]

Signed-off-by: Adam Duskett <aduskett@gmail.com>
---
 package/gobject-introspection/g-ir-scanner-lddwrapper.in | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/package/gobject-introspection/g-ir-scanner-lddwrapper.in b/package/gobject-introspection/g-ir-scanner-lddwrapper.in
index c19314566a..2f09bca77a 100644
--- a/package/gobject-introspection/g-ir-scanner-lddwrapper.in
+++ b/package/gobject-introspection/g-ir-scanner-lddwrapper.in
@@ -1,3 +1,3 @@
-#!/usr/bin/env bash
+#!/usr/bin/env sh
 
-${HOST_DIR}/sbin/prelink-rtld --root=$(dirname $0)/../../ "$@"
+"${HOST_DIR}"/sbin/prelink-rtld --root="$(dirname "$0")/../../" "$@"
-- 
2.30.2

_______________________________________________
buildroot mailing list
buildroot@lists.buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

[Buildroot] [PATCH/master 5/7] g-ir-compiler.in: Fix shellcheck warnings, switch to sh

[Adam Duskett]

Signed-off-by: Adam Duskett <aduskett@gmail.com>
---
 package/gobject-introspection/g-ir-compiler.in | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/package/gobject-introspection/g-ir-compiler.in b/package/gobject-introspection/g-ir-compiler.in
index 3313cc501f..712753023a 100644
--- a/package/gobject-introspection/g-ir-compiler.in
+++ b/package/gobject-introspection/g-ir-compiler.in
@@ -1,3 +1,3 @@
-#!/usr/bin/env bash
+#!/usr/bin/env sh
 
-$(dirname $0)/g-ir-scanner-qemuwrapper $(dirname $0)/g-ir-compiler.real "$@"
+"$(dirname "$0")"/g-ir-scanner-qemuwrapper "$(dirname "$0")"/g-ir-compiler.real "$@"
-- 
2.30.2

_______________________________________________
buildroot mailing list
buildroot@lists.buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

[Buildroot] [PATCH/master 6/7] package/qemu: add BR2_PACKAGE_HOST_QEMU_USER_MODE_ARGS string

[Adam Duskett]

For specific architectures, running qemu in user mode without any additional
options may fail if the host processor does not have the necessary instructions
to properly run qemu in user mode, which results in the following error:
"qemu: uncaught target signal 4 (Illegal instruction) - core dumped"

CoreI7 is one such architecture that has had consistent auto-build failures.

Add a new string in qemu/Config.in.host: BR2_PACKAGE_HOST_QEMU_USER_MODE_ARGS.
The default for the Corei7 architecture is directly from the OpenEmbedded
project found in meta/conf/machine/include/x86/tune-corei7.inc:
"-cpu Nehalem,check=false." Other architectures may be added to this string at
a later date if other failures occure.

Signed-off-by: Adam Duskett <aduskett@gmail.com>
---
 package/qemu/Config.in.host | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/package/qemu/Config.in.host b/package/qemu/Config.in.host
index 2dcf5631e5..fedf90d8b7 100644
--- a/package/qemu/Config.in.host
+++ b/package/qemu/Config.in.host
@@ -73,6 +73,10 @@ config BR2_PACKAGE_HOST_QEMU_LINUX_USER_MODE
 	  the running host kernel, you may run into invalid system
 	  calls, which may yield surprising effects.
 
+config BR2_PACKAGE_HOST_QEMU_USER_MODE_ARGS
+	string
+	default "-cpu Nehalem,check=false" if BR2_x86_corei7
+
 config BR2_PACKAGE_HOST_QEMU_VDE2
 	bool "VDE2 support"
 	help
-- 
2.30.2

_______________________________________________
buildroot mailing list
buildroot@lists.buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

[Buildroot] [PATCH/master 7/7] package/gobject-introspection: Add QEMU_USERMODE_ARGS support

[Adam Duskett]

Now that the qemu package has the BR2_PACKAGE_HOST_QEMU_USER_MODE_ARGS string,
the gobject-introspection g-ir-scanner-qemuwrapper script can pass that string
to qemu.

Add the QEMU_USERMODE_ARGS to g-ir-scanner-qemuwrapper.in and unconditionally
sed @QEMU_USERMODE_ARGS@ with BR2_PACKAGE_HOST_QEMU_USER_MODE_ARGS.

Fixes:
http://autobuild.buildroot.org/results/2e9dbc0d36600c09fa9e59ab1b1903c9f40661e8

Signed-off-by: Adam Duskett <aduskett@gmail.com>
---
 package/gobject-introspection/g-ir-scanner-qemuwrapper.in | 2 ++
 package/gobject-introspection/gobject-introspection.mk    | 2 ++
 2 files changed, 4 insertions(+)

diff --git a/package/gobject-introspection/g-ir-scanner-qemuwrapper.in b/package/gobject-introspection/g-ir-scanner-qemuwrapper.in
index 5ece75d0a4..e434502e2b 100644
--- a/package/gobject-introspection/g-ir-scanner-qemuwrapper.in
+++ b/package/gobject-introspection/g-ir-scanner-qemuwrapper.in
@@ -5,8 +5,10 @@
 # Use a modules directory which does not exist so we don't load random things
 # which may then get deleted (or their dependencies) and potentially segfault
 GOI_LIBRARY_PATH="${GIR_EXTRA_LIBS_PATH:+${GIR_EXTRA_LIBS_PATH}:}.libs:$(dirname "$0")/../lib:$(dirname "$0")/../../lib"
+QEMU_USERMODE_ARGS="@QEMU_USERMODE_ARGS@"
 GIO_MODULE_DIR="$(dirname "$0")/../lib/gio/modules-dummy" \
 @QEMU_USER@ -r @TOOLCHAIN_HEADERS_VERSION@ \
+    ${QEMU_USERMODE_ARGS:+${QEMU_USERMODE_ARGS}} \
     -L "$(dirname "$0")/../../" \
     -E LD_LIBRARY_PATH="${GOI_LIBRARY_PATH}" \
     "$@"
diff --git a/package/gobject-introspection/gobject-introspection.mk b/package/gobject-introspection/gobject-introspection.mk
index 81938b14df..23d6a8231e 100644
--- a/package/gobject-introspection/gobject-introspection.mk
+++ b/package/gobject-introspection/gobject-introspection.mk
@@ -84,6 +84,8 @@ define GOBJECT_INTROSPECTION_INSTALL_PRE_WRAPPERS
 		$(STAGING_DIR)/usr/bin/g-ir-scanner-qemuwrapper
 	$(SED) "s%@QEMU_USER@%$(QEMU_USER)%g" \
 		$(STAGING_DIR)/usr/bin/g-ir-scanner-qemuwrapper
+	$(SED) "s%@QEMU_USERMODE_ARGS@%$(call qstrip,$(BR2_PACKAGE_HOST_QEMU_USER_MODE_ARGS))%g" \
+		$(STAGING_DIR)/usr/bin/g-ir-scanner-qemuwrapper
 	$(SED) "s%@TOOLCHAIN_HEADERS_VERSION@%$(BR2_TOOLCHAIN_HEADERS_AT_LEAST)%g" \
 		$(STAGING_DIR)/usr/bin/g-ir-scanner-qemuwrapper
 
-- 
2.30.2

_______________________________________________
buildroot mailing list
buildroot@lists.buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

Re: [Buildroot] [PATCH/master 1/7] g-ir-scanner-qemuwrapper.in: Fix latent bug in

[Thomas Petazzoni]

Hello Adam,

Thanks for this patch series!

On Fri,  3 Sep 2021 09:20:21 -0700
Adam Duskett <aduskett@gmail.com> wrote:

> If GIR_EXTRA_LIBS_PATH is empty, the LD_LIBRARY_PATH variable is
> ":.libs:$(dirname $0)/../lib:$(dirname $0)/../../lib".
> 
> As discussed with Yann Morin, add a new variable, "GOI_LIBRARY_PATH" and
> prepend ${GIR_EXTRA_LIBS_PATH:+${GIR_EXTRA_LIBS_PATH}:} to the path to fix the
> above bug.
> 
> Signed-off-by: Adam Duskett <aduskett@gmail.com>

The commit title should have a prefix of
package/gobject-introspection/g-ir-scanner-qemuwrapper.in, and also a
description that isn't cut of in the middle.

These details can be fixed when applying, no need to resend for that.

Thanks!

Thomas
-- 
Thomas Petazzoni, co-owner and CEO, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
_______________________________________________
buildroot mailing list
buildroot@lists.buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

Re: [Buildroot] [PATCH/master 0/7] Qemu and gobject-introspection fixes

[Thomas Petazzoni]

On Fri,  3 Sep 2021 09:20:20 -0700
Adam Duskett <aduskett@gmail.com> wrote:

> First of all, sorry this took so long!
> 
> This series is an accumulation of several conversations between Thomas, Yann,
> and myself. The basics are this:
> 
> - GOI fails to build with the corei7 architecture because calling qemu results
>   in qemu trying to use instructions that may not exist on the host processor.
>   This causes a core dump.
> 
> - OpenEmbedded gets around this issue by adding options to qemu on certain
>   CPU and architecture combinations, Corei7 being one of those architectures.
> 
> - As discussed with Thomas, a new string "BR2_PACKAGE_HOST_QEMU_USER_MODE_ARGS"
>   in qemu/Config.in.host is added, which we then sed to g-ir-scanner.in.
> 
> In addition, I was anonyed that there were so many shellcheck issues with the
> various wrappers, and Yann was annoyed that the wrappers were calling to bash
> when they didn't need to do so, so those are also fixed in this patch series.
> 
> And last (but not least), Yann noticed a latent bug in g-ir-scanner-qemuwrapper.
> If GIR_EXTRA_LIBS_PATH is empty, then the LD_LIBRARY_PATH is prepended with a
> colon, so that is also fixed.
> 
> This will lead the way to NodeJS 14 which is one step closer to removing
> Python2!

Thanks a lot for all this work! Could you remind me what's the
relationship between this and NodeJS 14 ?

Thanks!

Thomas
-- 
Thomas Petazzoni, co-owner and CEO, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
_______________________________________________
buildroot mailing list
buildroot@lists.buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

Re: [Buildroot] [PATCH/master 6/7] package/qemu: add BR2_PACKAGE_HOST_QEMU_USER_MODE_ARGS string

[Thomas Petazzoni]

Hello Adam,

On Fri,  3 Sep 2021 09:20:26 -0700
Adam Duskett <aduskett@gmail.com> wrote:

> For specific architectures, running qemu in user mode without any additional
> options may fail if the host processor does not have the necessary instructions
> to properly run qemu in user mode, which results in the following error:
> "qemu: uncaught target signal 4 (Illegal instruction) - core dumped"
> 
> CoreI7 is one such architecture that has had consistent auto-build failures.
> 
> Add a new string in qemu/Config.in.host: BR2_PACKAGE_HOST_QEMU_USER_MODE_ARGS.
> The default for the Corei7 architecture is directly from the OpenEmbedded
> project found in meta/conf/machine/include/x86/tune-corei7.inc:
> "-cpu Nehalem,check=false." Other architectures may be added to this string at
> a later date if other failures occure.
> 
> Signed-off-by: Adam Duskett <aduskett@gmail.com>

So on the principle, I'm obviously OK as this is something I have
suggested. However, what bothers me here is that we are handling only
the Core i7 case, because it caused some issue in the autobuilder.

What is qemu doing when no -cpu is provided? Does it emulate the host
CPU in this case? If so, then it is going to be wrong in a lot of other
cases than Core i7, no?

Or put differently, shouldn't we essentially have a value for
BR2_PACKAGE_HOST_QEMU_USER_MODE_ARGS for pretty much all CPU
architectures that we want to support for this use-case ?

Thomas
-- 
Thomas Petazzoni, co-owner and CEO, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
_______________________________________________
buildroot mailing list
buildroot@lists.buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

Re: [Buildroot] [PATCH/master 0/7] Qemu and gobject-introspection fixes

[Adam Duskett]

Hey Thomas;

NodeJS14 ALSO requires a qemu wrapper to compile, so this work is
necessary to allows NodeJS14 to
not crash when compiling against a corei7 target.

Adam

On Fri, Sep 3, 2021 at 3:27 PM Thomas Petazzoni
<thomas.petazzoni@bootlin.com> wrote:
>
> On Fri,  3 Sep 2021 09:20:20 -0700
> Adam Duskett <aduskett@gmail.com> wrote:
>
> > First of all, sorry this took so long!
> >
> > This series is an accumulation of several conversations between Thomas, Yann,
> > and myself. The basics are this:
> >
> > - GOI fails to build with the corei7 architecture because calling qemu results
> >   in qemu trying to use instructions that may not exist on the host processor.
> >   This causes a core dump.
> >
> > - OpenEmbedded gets around this issue by adding options to qemu on certain
> >   CPU and architecture combinations, Corei7 being one of those architectures.
> >
> > - As discussed with Thomas, a new string "BR2_PACKAGE_HOST_QEMU_USER_MODE_ARGS"
> >   in qemu/Config.in.host is added, which we then sed to g-ir-scanner.in.
> >
> > In addition, I was anonyed that there were so many shellcheck issues with the
> > various wrappers, and Yann was annoyed that the wrappers were calling to bash
> > when they didn't need to do so, so those are also fixed in this patch series.
> >
> > And last (but not least), Yann noticed a latent bug in g-ir-scanner-qemuwrapper.
> > If GIR_EXTRA_LIBS_PATH is empty, then the LD_LIBRARY_PATH is prepended with a
> > colon, so that is also fixed.
> >
> > This will lead the way to NodeJS 14 which is one step closer to removing
> > Python2!
>
> Thanks a lot for all this work! Could you remind me what's the
> relationship between this and NodeJS 14 ?
>
> Thanks!
>
> Thomas
> --
> Thomas Petazzoni, co-owner and CEO, Bootlin
> Embedded Linux and Kernel engineering
> https://bootlin.com
_______________________________________________
buildroot mailing list
buildroot@lists.buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

Re: [Buildroot] [PATCH/master 0/7] Qemu and gobject-introspection fixes

[Thomas Petazzoni]

Hello Adam,

On Fri, 3 Sep 2021 16:09:35 -0700
Adam Duskett <aduskett@gmail.com> wrote:

> NodeJS14 ALSO requires a qemu wrapper to compile, so this work is
> necessary to allows NodeJS14 to
> not crash when compiling against a corei7 target.

Ah, OK. So there's no relationship between NodeJS and GOI, but it's
NodeJS that needs a qemu wrapper. OK makes sense!

Thomas
-- 
Thomas Petazzoni, co-owner and CEO, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
_______________________________________________
buildroot mailing list
buildroot@lists.buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

Re: [Buildroot] [PATCH/master 6/7] package/qemu: add BR2_PACKAGE_HOST_QEMU_USER_MODE_ARGS string

[Adam Duskett]

Hey Thomas;

As briefly discussed on IRC, this is mostly due to x86 discrepancies,
however, as I do not have access to * every * CPU
that buildroot + GOI supports, it's better to commit these changes
now, and if others find issues with their CPU/GOI
builds, they can submit a patch far easier thanks to this patch series!

Adam

On Fri, Sep 3, 2021 at 3:37 PM Thomas Petazzoni
<thomas.petazzoni@bootlin.com> wrote:
>
> Hello Adam,
>
> On Fri,  3 Sep 2021 09:20:26 -0700
> Adam Duskett <aduskett@gmail.com> wrote:
>
> > For specific architectures, running qemu in user mode without any additional
> > options may fail if the host processor does not have the necessary instructions
> > to properly run qemu in user mode, which results in the following error:
> > "qemu: uncaught target signal 4 (Illegal instruction) - core dumped"
> >
> > CoreI7 is one such architecture that has had consistent auto-build failures.
> >
> > Add a new string in qemu/Config.in.host: BR2_PACKAGE_HOST_QEMU_USER_MODE_ARGS.
> > The default for the Corei7 architecture is directly from the OpenEmbedded
> > project found in meta/conf/machine/include/x86/tune-corei7.inc:
> > "-cpu Nehalem,check=false." Other architectures may be added to this string at
> > a later date if other failures occure.
> >
> > Signed-off-by: Adam Duskett <aduskett@gmail.com>
>
> So on the principle, I'm obviously OK as this is something I have
> suggested. However, what bothers me here is that we are handling only
> the Core i7 case, because it caused some issue in the autobuilder.
>
> What is qemu doing when no -cpu is provided? Does it emulate the host
> CPU in this case? If so, then it is going to be wrong in a lot of other
> cases than Core i7, no?
>
> Or put differently, shouldn't we essentially have a value for
> BR2_PACKAGE_HOST_QEMU_USER_MODE_ARGS for pretty much all CPU
> architectures that we want to support for this use-case ?
>
> Thomas
> --
> Thomas Petazzoni, co-owner and CEO, Bootlin
> Embedded Linux and Kernel engineering
> https://bootlin.com
_______________________________________________
buildroot mailing list
buildroot@lists.buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

Re: [Buildroot] [PATCH/master 1/7] g-ir-scanner-qemuwrapper.in: Fix latent bug in

[Arnout Vandecappelle]

 Applied to master, thanks.

 I've tried to fix the summary line.

On 03/09/2021 18:20, Adam Duskett wrote:
> If GIR_EXTRA_LIBS_PATH is empty, the LD_LIBRARY_PATH variable is
> ":.libs:$(dirname $0)/../lib:$(dirname $0)/../../lib".
> 
> As discussed with Yann Morin, add a new variable, "GOI_LIBRARY_PATH" and
> prepend ${GIR_EXTRA_LIBS_PATH:+${GIR_EXTRA_LIBS_PATH}:} to the path to fix the
> above bug.
> 
> Signed-off-by: Adam Duskett <aduskett@gmail.com>
> ---
>  package/gobject-introspection/g-ir-scanner-qemuwrapper.in | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/package/gobject-introspection/g-ir-scanner-qemuwrapper.in b/package/gobject-introspection/g-ir-scanner-qemuwrapper.in
> index cad7ef875e..ec066062e2 100644
> --- a/package/gobject-introspection/g-ir-scanner-qemuwrapper.in
> +++ b/package/gobject-introspection/g-ir-scanner-qemuwrapper.in
> @@ -4,10 +4,11 @@
>  # is older than the target kernel.
>  # Use a modules directory which does not exist so we don't load random things
>  # which may then get deleted (or their dependencies) and potentially segfault
> +GOI_LIBRARY_PATH="${GIR_EXTRA_LIBS_PATH:+${GIR_EXTRA_LIBS_PATH}:}.libs:$(dirname "$0")/../lib:$(dirname "$0")/../../lib"

 Since the comment above is not related to this assignment, I moved the
assignment above the comment.

 I haven't tested the result so I hope I made no mistakes :-)

 Regards,
 Arnout

>  GIO_MODULE_DIR=$(dirname $0)/../lib/gio/modules-dummy \
>  @QEMU_USER@ -r @TOOLCHAIN_HEADERS_VERSION@ \
>  -L $(dirname $0)/../../ \
> --E LD_LIBRARY_PATH=${GIR_EXTRA_LIBS_PATH}:.libs:$(dirname $0)/../lib:$(dirname $0)/../../lib \
> +-E LD_LIBRARY_PATH="${GOI_LIBRARY_PATH}" \
>  "$@"
>  
>  if [[ $? -ne 0 ]]; then
> 
_______________________________________________
buildroot mailing list
buildroot@lists.buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

Re: [Buildroot] [PATCH/master 2/7] g-ir-scanner-qemuwrapper.in: Fix shellcheck warnings, switch to sh

[Arnout Vandecappelle]

On 03/09/2021 18:20, Adam Duskett wrote:
>   - Add double quotes to prevent globbing and word splitting.
>   - Add indentations of continuation lines.
>   - Disable SC2181 and SC2016 as we explicitly do not want the variables
>     expanded in the echo.
> 
> Signed-off-by: Adam Duskett <aduskett@gmail.com>
> ---
>  .../g-ir-scanner-qemuwrapper.in                  | 16 +++++++++-------
>  1 file changed, 9 insertions(+), 7 deletions(-)
> 
> diff --git a/package/gobject-introspection/g-ir-scanner-qemuwrapper.in b/package/gobject-introspection/g-ir-scanner-qemuwrapper.in
> index ec066062e2..5ece75d0a4 100644
> --- a/package/gobject-introspection/g-ir-scanner-qemuwrapper.in
> +++ b/package/gobject-introspection/g-ir-scanner-qemuwrapper.in
> @@ -1,18 +1,20 @@
> -#!/usr/bin/env bash
> +#!/usr/bin/env sh
>  
> -# Pass -r to qemu-user as to trick glibc into not errorings out if the host kernel
> +# Pass -r to qemu-user as to trick glibc into not erroring out if the host kernel
>  # is older than the target kernel.
>  # Use a modules directory which does not exist so we don't load random things
>  # which may then get deleted (or their dependencies) and potentially segfault
>  GOI_LIBRARY_PATH="${GIR_EXTRA_LIBS_PATH:+${GIR_EXTRA_LIBS_PATH}:}.libs:$(dirname "$0")/../lib:$(dirname "$0")/../../lib"
> -GIO_MODULE_DIR=$(dirname $0)/../lib/gio/modules-dummy \
> +GIO_MODULE_DIR="$(dirname "$0")/../lib/gio/modules-dummy" \
>  @QEMU_USER@ -r @TOOLCHAIN_HEADERS_VERSION@ \
> --L $(dirname $0)/../../ \
> --E LD_LIBRARY_PATH="${GOI_LIBRARY_PATH}" \
> -"$@"
> +    -L "$(dirname "$0")/../../" \
> +    -E LD_LIBRARY_PATH="${GOI_LIBRARY_PATH}" \
> +    "$@"
>  
> -if [[ $? -ne 0 ]]; then
> +# shellcheck disable=SC2181

 Instead of adding an exception, we could just do the right thing:

if ! @QEMU_USER@ -r ... \
    ... \
then
    ...

 But since I wasn't going to test this change, I left it as is.

 Regards,
 Arnout

> +if [ $? -ne 0 ]; then
>      echo "If the above error message is about missing .so libraries, then setting up GIR_EXTRA_LIBS_PATH in the .mk file should help."
> +    # shellcheck disable=SC2016
>      echo 'Typically like this: PKG_MAKE_ENV += GIR_EXTRA_LIBS_PATH="$(@D)/.libs"'
>      exit 1
>  fi
> 
_______________________________________________
buildroot mailing list
buildroot@lists.buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

Re: [Buildroot] [PATCH/master 3/7] g-ir-scanner.in: Fix shellcheck warnings, switch to sh

[Arnout Vandecappelle]

On 03/09/2021 18:20, Adam Duskett wrote:
> Signed-off-by: Adam Duskett <aduskett@gmail.com>

 Applied to master, thanks.

 ... And now I notice that I forgot to add the package/gobject-introspection/
prefix to the previous patch. I added it here and will hopefully remember to do
it for the following ones.

 Regards,
 Arnout

> ---
>  package/gobject-introspection/g-ir-scanner.in | 12 ++++++------
>  1 file changed, 6 insertions(+), 6 deletions(-)
> 
> diff --git a/package/gobject-introspection/g-ir-scanner.in b/package/gobject-introspection/g-ir-scanner.in
> index 22df7be309..013eb15700 100644
> --- a/package/gobject-introspection/g-ir-scanner.in
> +++ b/package/gobject-introspection/g-ir-scanner.in
> @@ -1,4 +1,4 @@
> -#!/usr/bin/env bash
> +#!/usr/bin/env sh
>  
>  # These environment variables are empty by default, and as such ccompiler.py
>  # defaults to either using the system tools (CC, CXX, and CPP) or leaving
> @@ -15,8 +15,8 @@ export LDFLAGS="${LDFLAGS:-@TARGET_LDFLAGS@}"
>  
>  export GI_SCANNER_DISABLE_CACHE=1
>  
> -${HOST_DIR}/bin/g-ir-scanner \
> ---lib-dirs-envvar=GIR_EXTRA_LIBS_PATH \
> ---use-binary-wrapper=$(dirname $0)/g-ir-scanner-qemuwrapper \
> ---use-ldd-wrapper=$(dirname $0)/g-ir-scanner-lddwrapper \
> ---add-include-path=$(dirname $0)/../share/gir-1.0 "$@"
> +"${HOST_DIR}"/bin/g-ir-scanner \
> +    --lib-dirs-envvar=GIR_EXTRA_LIBS_PATH \
> +    --use-binary-wrapper="$(dirname "$0")"/g-ir-scanner-qemuwrapper \
> +    --use-ldd-wrapper="$(dirname "$0")"/g-ir-scanner-lddwrapper \
> +    --add-include-path="$(dirname "$0")"/../share/gir-1.0 "$@"
> 
_______________________________________________
buildroot mailing list
buildroot@lists.buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

Re: [Buildroot] [PATCH/master 4/7] g-ir-scanner-lddwrapper.in: Fix shellcheck warnings, switch to sh

[Arnout Vandecappelle]

On 03/09/2021 18:20, Adam Duskett wrote:
> Signed-off-by: Adam Duskett <aduskett@gmail.com>

 Applied to master, thanks.

 Regards,
 Arnout

> ---
>  package/gobject-introspection/g-ir-scanner-lddwrapper.in | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/package/gobject-introspection/g-ir-scanner-lddwrapper.in b/package/gobject-introspection/g-ir-scanner-lddwrapper.in
> index c19314566a..2f09bca77a 100644
> --- a/package/gobject-introspection/g-ir-scanner-lddwrapper.in
> +++ b/package/gobject-introspection/g-ir-scanner-lddwrapper.in
> @@ -1,3 +1,3 @@
> -#!/usr/bin/env bash
> +#!/usr/bin/env sh
>  
> -${HOST_DIR}/sbin/prelink-rtld --root=$(dirname $0)/../../ "$@"
> +"${HOST_DIR}"/sbin/prelink-rtld --root="$(dirname "$0")/../../" "$@"
> 
_______________________________________________
buildroot mailing list
buildroot@lists.buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

Re: [Buildroot] [PATCH/master 5/7] g-ir-compiler.in: Fix shellcheck warnings, switch to sh

[Arnout Vandecappelle]

On 03/09/2021 18:20, Adam Duskett wrote:
> Signed-off-by: Adam Duskett <aduskett@gmail.com>

 Applied to master, thanks.

 Regards,
 Arnout

> ---
>  package/gobject-introspection/g-ir-compiler.in | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/package/gobject-introspection/g-ir-compiler.in b/package/gobject-introspection/g-ir-compiler.in
> index 3313cc501f..712753023a 100644
> --- a/package/gobject-introspection/g-ir-compiler.in
> +++ b/package/gobject-introspection/g-ir-compiler.in
> @@ -1,3 +1,3 @@
> -#!/usr/bin/env bash
> +#!/usr/bin/env sh
>  
> -$(dirname $0)/g-ir-scanner-qemuwrapper $(dirname $0)/g-ir-compiler.real "$@"
> +"$(dirname "$0")"/g-ir-scanner-qemuwrapper "$(dirname "$0")"/g-ir-compiler.real "$@"
> 
_______________________________________________
buildroot mailing list
buildroot@lists.buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

Re: [Buildroot] [PATCH/master 6/7] package/qemu: add BR2_PACKAGE_HOST_QEMU_USER_MODE_ARGS string

[Arnout Vandecappelle]

On 10/09/2021 19:42, Adam Duskett wrote:
> Hey Thomas;
> 
> As briefly discussed on IRC, this is mostly due to x86 discrepancies,
> however, as I do not have access to * every * CPU
> that buildroot + GOI supports, it's better to commit these changes
> now, and if others find issues with their CPU/GOI
> builds, they can submit a patch far easier thanks to this patch series!

 I still think Thomas is right: we should have a mapping for each architecture
option to the Qemu emulation option that corresponds to it.

 In an ideal world, we would have BR2_ARCH_QEMU_CPU that is set in
arch/Config.in.* for all (sub)architectures.

 However, I doubt that anyone is going to take the time to do that (and test it,
without even knowing how to test it because *most* code will run just fine when
the wrong -cpu option is used).

 So, the pragmatic approach is to use this patch.

 Therefore, applied to master, thanks.

 Regards,
 Arnout

> 
> Adam
> 
> On Fri, Sep 3, 2021 at 3:37 PM Thomas Petazzoni
> <thomas.petazzoni@bootlin.com> wrote:
>>
>> Hello Adam,
>>
>> On Fri,  3 Sep 2021 09:20:26 -0700
>> Adam Duskett <aduskett@gmail.com> wrote:
>>
>>> For specific architectures, running qemu in user mode without any additional
>>> options may fail if the host processor does not have the necessary instructions
>>> to properly run qemu in user mode, which results in the following error:
>>> "qemu: uncaught target signal 4 (Illegal instruction) - core dumped"
>>>
>>> CoreI7 is one such architecture that has had consistent auto-build failures.
>>>
>>> Add a new string in qemu/Config.in.host: BR2_PACKAGE_HOST_QEMU_USER_MODE_ARGS.
>>> The default for the Corei7 architecture is directly from the OpenEmbedded
>>> project found in meta/conf/machine/include/x86/tune-corei7.inc:
>>> "-cpu Nehalem,check=false." Other architectures may be added to this string at
>>> a later date if other failures occure.
>>>
>>> Signed-off-by: Adam Duskett <aduskett@gmail.com>
>>
>> So on the principle, I'm obviously OK as this is something I have
>> suggested. However, what bothers me here is that we are handling only
>> the Core i7 case, because it caused some issue in the autobuilder.
>>
>> What is qemu doing when no -cpu is provided? Does it emulate the host
>> CPU in this case? If so, then it is going to be wrong in a lot of other
>> cases than Core i7, no?
>>
>> Or put differently, shouldn't we essentially have a value for
>> BR2_PACKAGE_HOST_QEMU_USER_MODE_ARGS for pretty much all CPU
>> architectures that we want to support for this use-case ?
>>
>> Thomas
>> --
>> Thomas Petazzoni, co-owner and CEO, Bootlin
>> Embedded Linux and Kernel engineering
>> https://bootlin.com
> _______________________________________________
> buildroot mailing list
> buildroot@lists.buildroot.org
> https://lists.buildroot.org/mailman/listinfo/buildroot
> 
_______________________________________________
buildroot mailing list
buildroot@lists.buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

Re: [Buildroot] [PATCH/master 6/7] package/qemu: add BR2_PACKAGE_HOST_QEMU_USER_MODE_ARGS string

[Arnout Vandecappelle]

On 11/09/2021 21:39, Arnout Vandecappelle wrote:
> 
> 
> On 10/09/2021 19:42, Adam Duskett wrote:
>> Hey Thomas;
>>
>> As briefly discussed on IRC, this is mostly due to x86 discrepancies,
>> however, as I do not have access to * every * CPU
>> that buildroot + GOI supports, it's better to commit these changes
>> now, and if others find issues with their CPU/GOI
>> builds, they can submit a patch far easier thanks to this patch series!
> 
>  I still think Thomas is right: we should have a mapping for each architecture
> option to the Qemu emulation option that corresponds to it.
> 
>  In an ideal world, we would have BR2_ARCH_QEMU_CPU that is set in
> arch/Config.in.* for all (sub)architectures.

 And even better: a qemu wrapper would make sure that the correct cpu option is
*always* passed to qemu.

 Dream on...

 Regards,
 Arnout

> 
>  However, I doubt that anyone is going to take the time to do that (and test it,
> without even knowing how to test it because *most* code will run just fine when
> the wrong -cpu option is used).
> 
>  So, the pragmatic approach is to use this patch.
> 
>  Therefore, applied to master, thanks.
> 
>  Regards,
>  Arnout
> 
>>
>> Adam
>>
>> On Fri, Sep 3, 2021 at 3:37 PM Thomas Petazzoni
>> <thomas.petazzoni@bootlin.com> wrote:
>>>
>>> Hello Adam,
>>>
>>> On Fri,  3 Sep 2021 09:20:26 -0700
>>> Adam Duskett <aduskett@gmail.com> wrote:
>>>
>>>> For specific architectures, running qemu in user mode without any additional
>>>> options may fail if the host processor does not have the necessary instructions
>>>> to properly run qemu in user mode, which results in the following error:
>>>> "qemu: uncaught target signal 4 (Illegal instruction) - core dumped"
>>>>
>>>> CoreI7 is one such architecture that has had consistent auto-build failures.
>>>>
>>>> Add a new string in qemu/Config.in.host: BR2_PACKAGE_HOST_QEMU_USER_MODE_ARGS.
>>>> The default for the Corei7 architecture is directly from the OpenEmbedded
>>>> project found in meta/conf/machine/include/x86/tune-corei7.inc:
>>>> "-cpu Nehalem,check=false." Other architectures may be added to this string at
>>>> a later date if other failures occure.
>>>>
>>>> Signed-off-by: Adam Duskett <aduskett@gmail.com>
>>>
>>> So on the principle, I'm obviously OK as this is something I have
>>> suggested. However, what bothers me here is that we are handling only
>>> the Core i7 case, because it caused some issue in the autobuilder.
>>>
>>> What is qemu doing when no -cpu is provided? Does it emulate the host
>>> CPU in this case? If so, then it is going to be wrong in a lot of other
>>> cases than Core i7, no?
>>>
>>> Or put differently, shouldn't we essentially have a value for
>>> BR2_PACKAGE_HOST_QEMU_USER_MODE_ARGS for pretty much all CPU
>>> architectures that we want to support for this use-case ?
>>>
>>> Thomas
>>> --
>>> Thomas Petazzoni, co-owner and CEO, Bootlin
>>> Embedded Linux and Kernel engineering
>>> https://bootlin.com
>> _______________________________________________
>> buildroot mailing list
>> buildroot@lists.buildroot.org
>> https://lists.buildroot.org/mailman/listinfo/buildroot
>>
_______________________________________________
buildroot mailing list
buildroot@lists.buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

Re: [Buildroot] [PATCH/master 7/7] package/gobject-introspection: Add QEMU_USERMODE_ARGS support

[Arnout Vandecappelle]

On 03/09/2021 18:20, Adam Duskett wrote:
> Now that the qemu package has the BR2_PACKAGE_HOST_QEMU_USER_MODE_ARGS string,
> the gobject-introspection g-ir-scanner-qemuwrapper script can pass that string
> to qemu.
> 
> Add the QEMU_USERMODE_ARGS to g-ir-scanner-qemuwrapper.in and unconditionally
> sed @QEMU_USERMODE_ARGS@ with BR2_PACKAGE_HOST_QEMU_USER_MODE_ARGS.
> 
> Fixes:
> http://autobuild.buildroot.org/results/2e9dbc0d36600c09fa9e59ab1b1903c9f40661e8
> 
> Signed-off-by: Adam Duskett <aduskett@gmail.com>
> ---
>  package/gobject-introspection/g-ir-scanner-qemuwrapper.in | 2 ++
>  package/gobject-introspection/gobject-introspection.mk    | 2 ++
>  2 files changed, 4 insertions(+)
> 
> diff --git a/package/gobject-introspection/g-ir-scanner-qemuwrapper.in b/package/gobject-introspection/g-ir-scanner-qemuwrapper.in
> index 5ece75d0a4..e434502e2b 100644
> --- a/package/gobject-introspection/g-ir-scanner-qemuwrapper.in
> +++ b/package/gobject-introspection/g-ir-scanner-qemuwrapper.in
> @@ -5,8 +5,10 @@
>  # Use a modules directory which does not exist so we don't load random things
>  # which may then get deleted (or their dependencies) and potentially segfault
>  GOI_LIBRARY_PATH="${GIR_EXTRA_LIBS_PATH:+${GIR_EXTRA_LIBS_PATH}:}.libs:$(dirname "$0")/../lib:$(dirname "$0")/../../lib"
> +QEMU_USERMODE_ARGS="@QEMU_USERMODE_ARGS@"
>  GIO_MODULE_DIR="$(dirname "$0")/../lib/gio/modules-dummy" \
>  @QEMU_USER@ -r @TOOLCHAIN_HEADERS_VERSION@ \
> +    ${QEMU_USERMODE_ARGS:+${QEMU_USERMODE_ARGS}} \

 I'm surprised this doesn't trigger SC2086... But actually, there's no need for
the additional variable, right? If you put @QEMU_USERMODE_ARGS@ here directly,
it should do the right thing:

- if empty, it just expands to

qemu-user-x86_64 -r 5.10.1 \
     \
    -L ...

- if set to "-cpu Nehalem,check=false" it expands to

qemu-user-x86_64 -r 5.10.1 \
    -cpu Nehalem,check=false \
    -L ...


 So I changed that and I'm doing a test build before pushing to be sure it works
as expected.

 Regards,
 Arnout

>      -L "$(dirname "$0")/../../" \
>      -E LD_LIBRARY_PATH="${GOI_LIBRARY_PATH}" \
>      "$@"
> diff --git a/package/gobject-introspection/gobject-introspection.mk b/package/gobject-introspection/gobject-introspection.mk
> index 81938b14df..23d6a8231e 100644
> --- a/package/gobject-introspection/gobject-introspection.mk
> +++ b/package/gobject-introspection/gobject-introspection.mk
> @@ -84,6 +84,8 @@ define GOBJECT_INTROSPECTION_INSTALL_PRE_WRAPPERS
>  		$(STAGING_DIR)/usr/bin/g-ir-scanner-qemuwrapper
>  	$(SED) "s%@QEMU_USER@%$(QEMU_USER)%g" \
>  		$(STAGING_DIR)/usr/bin/g-ir-scanner-qemuwrapper
> +	$(SED) "s%@QEMU_USERMODE_ARGS@%$(call qstrip,$(BR2_PACKAGE_HOST_QEMU_USER_MODE_ARGS))%g" \
> +		$(STAGING_DIR)/usr/bin/g-ir-scanner-qemuwrapper
>  	$(SED) "s%@TOOLCHAIN_HEADERS_VERSION@%$(BR2_TOOLCHAIN_HEADERS_AT_LEAST)%g" \
>  		$(STAGING_DIR)/usr/bin/g-ir-scanner-qemuwrapper
>  
> 
_______________________________________________
buildroot mailing list
buildroot@lists.buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

Re: [Buildroot] [PATCH/master 2/7] g-ir-scanner-qemuwrapper.in: Fix shellcheck warnings, switch to sh

[Yann E. MORIN]

Arnout, All,

On 2021-09-11 20:57 +0200, Arnout Vandecappelle spake thusly:
> On 03/09/2021 18:20, Adam Duskett wrote:
> >   - Add double quotes to prevent globbing and word splitting.
> >   - Add indentations of continuation lines.
> >   - Disable SC2181 and SC2016 as we explicitly do not want the variables
> >     expanded in the echo.
> > 
> > Signed-off-by: Adam Duskett <aduskett@gmail.com>
> > ---
> >  .../g-ir-scanner-qemuwrapper.in                  | 16 +++++++++-------
> >  1 file changed, 9 insertions(+), 7 deletions(-)
> > 
> > diff --git a/package/gobject-introspection/g-ir-scanner-qemuwrapper.in b/package/gobject-introspection/g-ir-scanner-qemuwrapper.in
> > index ec066062e2..5ece75d0a4 100644
> > --- a/package/gobject-introspection/g-ir-scanner-qemuwrapper.in
> > +++ b/package/gobject-introspection/g-ir-scanner-qemuwrapper.in
> > @@ -1,18 +1,20 @@
> > -#!/usr/bin/env bash
> > +#!/usr/bin/env sh
> >  
> > -# Pass -r to qemu-user as to trick glibc into not errorings out if the host kernel
> > +# Pass -r to qemu-user as to trick glibc into not erroring out if the host kernel
> >  # is older than the target kernel.
> >  # Use a modules directory which does not exist so we don't load random things
> >  # which may then get deleted (or their dependencies) and potentially segfault
> >  GOI_LIBRARY_PATH="${GIR_EXTRA_LIBS_PATH:+${GIR_EXTRA_LIBS_PATH}:}.libs:$(dirname "$0")/../lib:$(dirname "$0")/../../lib"
> > -GIO_MODULE_DIR=$(dirname $0)/../lib/gio/modules-dummy \
> > +GIO_MODULE_DIR="$(dirname "$0")/../lib/gio/modules-dummy" \
> >  @QEMU_USER@ -r @TOOLCHAIN_HEADERS_VERSION@ \
> > --L $(dirname $0)/../../ \
> > --E LD_LIBRARY_PATH="${GOI_LIBRARY_PATH}" \
> > -"$@"
> > +    -L "$(dirname "$0")/../../" \
> > +    -E LD_LIBRARY_PATH="${GOI_LIBRARY_PATH}" \
> > +    "$@"
> >  
> > -if [[ $? -ne 0 ]]; then
> > +# shellcheck disable=SC2181
> 
>  Instead of adding an exception, we could just do the right thing:
> 
> if ! @QEMU_USER@ -r ... \
>     ... \
> then
>     ...
> 
>  But since I wasn't going to test this change, I left it as is.

I also considered it when Adam and I discussed this patch on IRC, but I
side with him to keep the code as-is, but forgot to explain why so that
he could add that to the commit log.

So for the posterity, here is my reasonning:

    Continuation lines in a condition are ugly and make an if-then-else
    condition harder to grok, so we prefer adding an SC2181 exception as
    it leaves the code more readable.

Regards,
Yann E. MORIN.

-- 
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 561 099 427 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'
_______________________________________________
buildroot mailing list
buildroot@lists.buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

[Buildroot] [PATCH/master 0/7] Qemu and gobject-introspection fixes

[Adam Duskett]

From: Adam Duskett <aduskett@gmail.com>

First of all, sorry this took so long!

This series is an accumulation of several conversations between Thomas, Yann,
and myself. The basics are this:

- GOI fails to build with the corei7 architecture because calling qemu results
  in qemu trying to use instructions that may not exist on the host processor.
  This causes a core dump.

- OpenEmbedded gets around this issue by adding options to qemu on certain
  CPU and architecture combinations, Corei7 being one of those architectures.

- As discussed with Thomas, a new string "BR2_PACKAGE_HOST_QEMU_USER_MODE_ARGS"
  in qemu/Config.in.host is added, which we then sed to g-ir-scanner.in.

In addition, I was anonyed that there were so many shellcheck issues with the
various wrappers, and Yann was annoyed that the wrappers were calling to bash
when they didn't need to do so, so those are also fixed in this patch series.

And last (but not least), Yann noticed a latent bug in g-ir-scanner-qemuwrapper.
If GIR_EXTRA_LIBS_PATH is empty, then the LD_LIBRARY_PATH is prepended with a
colon, so that is also fixed.

This will lead the way to NodeJS 14 which is one step closer to removing
Python2!

Tested in a CentOS7 docker container:
                     andes-nds32 [ 1/45]: SKIPPED
                     arm-aarch64 [ 2/45]: OK
           bootlin-aarch64-glibc [ 3/45]: OK
       bootlin-arcle-hs38-uclibc [ 4/45]: SKIPPED
            bootlin-armv5-uclibc [ 5/45]: SKIPPED
             bootlin-armv7-glibc [ 6/45]: OK
           bootlin-armv7m-uclibc [ 7/45]: SKIPPED
              bootlin-armv7-musl [ 8/45]: SKIPPED
        bootlin-m68k-5208-uclibc [ 9/45]: SKIPPED
       bootlin-m68k-68040-uclibc [10/45]: SKIPPED
     bootlin-microblazeel-uclibc [11/45]: SKIPPED
        bootlin-mipsel32r6-glibc [12/45]: OK
           bootlin-mipsel-uclibc [13/45]: SKIPPED
             bootlin-nios2-glibc [14/45]: OK
         bootlin-openrisc-uclibc [15/45]: SKIPPED
bootlin-powerpc64le-power8-glibc [16/45]: OK
   bootlin-powerpc-e500mc-uclibc [17/45]: SKIPPED
           bootlin-riscv32-glibc [18/45]: OK
           bootlin-riscv64-glibc [19/45]: OK
            bootlin-riscv64-musl [20/45]: SKIPPED
              bootlin-sh4-uclibc [21/45]: SKIPPED
           bootlin-sparc64-glibc [22/45]: OK
            bootlin-sparc-uclibc [23/45]: SKIPPED
            bootlin-x86-64-glibc [24/45]: OK
             bootlin-x86-64-musl [25/45]: SKIPPED
           bootlin-x86-64-uclibc [26/45]: SKIPPED
           bootlin-xtensa-uclibc [27/45]: SKIPPED
                    br-arm-basic [28/45]: SKIPPED
            br-arm-full-nothread [29/45]: SKIPPED
              br-arm-full-static [30/45]: SKIPPED
           br-i386-pentium4-full [31/45]: SKIPPED
        br-i386-pentium-mmx-musl [32/45]: SKIPPED
              br-mips64-n64-full [33/45]: SKIPPED
         br-mips64r6-el-hf-glibc [34/45]: SKIPPED
       br-powerpc-603e-basic-cpp [35/45]: SKIPPED
       br-powerpc64-power7-glibc [36/45]: OK
               linaro-aarch64-be [37/45]: SKIPPED
                  linaro-aarch64 [38/45]: OK
                      linaro-arm [39/45]: OK
             sourcery-arm-armv4t [40/45]: SKIPPED
                    sourcery-arm [41/45]: SKIPPED
             sourcery-arm-thumb2 [42/45]: SKIPPED
                 sourcery-mips64 [43/45]: SKIPPED
                   sourcery-mips [44/45]: OK
                  sourcery-nios2 [45/45]: OK


Adam Duskett (7):
  g-ir-scanner-qemuwrapper.in: Fix latent bug in LD_LIBRARY_PATH string
  g-ir-scanner-qemuwrapper.in: Fix shellcheck warnings, switch to sh
  g-ir-scanner.in: Fix shellcheck warnings, switch to sh
  g-ir-scanner-lddwrapper.in: Fix shellcheck warnings, switch to sh
  g-ir-compiler.in: Fix shellcheck warnings, switch to sh
  package/qemu: add BR2_PACKAGE_HOST_QEMU_USER_MODE_ARGS string
  package/gobject-introspection: Add QEMU_USERMODE_ARGS support

 .../gobject-introspection/g-ir-compiler.in    |  4 ++--
 .../g-ir-scanner-lddwrapper.in                |  4 ++--
 .../g-ir-scanner-qemuwrapper.in               | 19 ++++++++++++-------
 package/gobject-introspection/g-ir-scanner.in | 12 ++++++------
 .../gobject-introspection.mk                  |  2 ++
 package/qemu/Config.in.host                   |  4 ++++
 6 files changed, 28 insertions(+), 17 deletions(-)

-- 
2.30.2

_______________________________________________
buildroot mailing list
buildroot@lists.buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot