libffi: add ptest support
PATCH 1/1 libffi: add ptest support
Add ptest support for libffi using the upstream bhaible test suite which validates FFI call and callback interfaces across various calling conventions and data types.
Tests are compiled on-target against the installed libffi. Headers and a linker symlink are installed in the ptest directory to avoid depending on libffi-dev.
Tested on qemux86-64 with ptest-runner: 2 PASS, 0 FAIL (29s).
--- meta/conf/distro/include/ptest-packagelists.inc | 1 + meta/recipes-support/libffi/libffi/run-ptest | 16 ++++++++++++++++ meta/recipes-support/libffi/libffi_3.5.2.bb | 16 +++++++++++++++- 3 files changed, 32 insertions(+), 1 deletion(-) create mode 100755 meta/recipes-support/libffi/libffi/run-ptest
meta/conf/distro/include/ptest-packagelists.inc +1 -0
@@ -36,6 +36,7 @@ PTESTS_FAST = "\ libconfig \ libconvert-asn1-perl \ libexif \ + libffi \ libgpg-error\ libksba \ libmd \
meta/recipes-support/libffi/libffi/run-ptest new file +16 -0
@@ -0,0 +1,16 @@ +#!/bin/sh +# SPDX-License-Identifier: MIT + +PTEST_DIR=$(cd "$(dirname "$0")" && pwd) +cd "$PTEST_DIR/testsuite/libffi.bhaible" || exit 1 + +for test in test-call test-callback; do + gcc -O2 -Wall -I$PTEST_DIR/include -L$PTEST_DIR/lib -o $test $test.c -lffi + if ./$test > $test.out 2>&1 && ! LC_ALL=C uniq -u < $test.out | grep -q .; then + echo "PASS: $test" + else + echo "FAIL: $test" + LC_ALL=C uniq -u < $test.out 2>/dev/null + fi + rm -f $test $test.out +done
meta/recipes-support/libffi/libffi_3.5.2.bb +15 -1
@@ -12,6 +12,7 @@ LIC_FILES_CHKSUM = "file://LICENSE;md5=ce4763670c5b7756000561f9af1ab178" SRC_URI = "${GITHUB_BASE_URI}/download/v${PV}/${BPN}-${PV}.tar.gz \ file://not-win32.patch \ + file://run-ptest \ " SRC_URI[sha256sum] = "f3a3082a23b37c293a4fcd1053147b371f2ff91fa7ea1b2a52e335676bac82dc"
@@ -19,7 +20,7 @@ EXTRA_OECONF = "--disable-builddir" EXTRA_OECONF:class-native += "--with-gcc-arch=generic" EXTRA_OEMAKE:class-target = "LIBTOOLFLAGS='--tag=CC'" -inherit autotools texinfo multilib_header github-releases +inherit autotools texinfo multilib_header github-releases ptest do_install:append() { oe_multilib_header ffi.h ffitarget.h
@@ -30,3 +31,16 @@ do_install:append() { MIPS_INSTRUCTION_SET = "mips" BBCLASSEXTEND = "native nativesdk" + +RDEPENDS:${PN}-ptest += "packagegroup-core-buildessential" + +do_install_ptest() { + install -d ${D}${PTEST_PATH}/testsuite + cp -r ${S}/testsuite/libffi.bhaible ${D}${PTEST_PATH}/testsuite/ + + install -d ${D}${PTEST_PATH}/include + install -m 0644 ${B}/include/ffi.h ${D}${PTEST_PATH}/include/ + install -m 0644 ${B}/include/ffitarget.h ${D}${PTEST_PATH}/include/ + install -d ${D}${PTEST_PATH}/lib + ln -sf ${libdir}/libffi.so.8 ${D}${PTEST_PATH}/lib/libffi.so +}
Signature · 1 line
2.34.1
PATCH v2 1/1 libffi: add ptest support
Add ptest support for libffi using all upstream test suites: bhaible, call, closures, complex, go, and threads.
Tests are compiled on-target against the installed libffi. Headers (including the generated fficonfig.h) and a linker symlink are installed in the ptest directory to avoid depending on libffi-dev.
Tested on qemux86-64 with ptest-runner: 199 PASS, 0 FAIL, 2 SKIP (349s).
---
Changes in v2 :
- Include all upstream test suites instead of only bhaible
- Install fficonfig.h required by ffitest.h
- Add support for C++ tests (closures/unwindtest*.cc)
- Add -lpthread for threads suite
--- .../distro/include/ptest-packagelists.inc | 1 + meta/recipes-support/libffi/libffi/run-ptest | 53 +++++++++++++++++++ meta/recipes-support/libffi/libffi_3.5.2.bb | 21 +++++++- 3 files changed, 74 insertions(+), 1 deletion(-) create mode 100755 meta/recipes-support/libffi/libffi/run-ptest
meta/conf/distro/include/ptest-packagelists.inc +1 -0
@@ -34,6 +34,7 @@ PTESTS_FAST = "\ libconfig \ libconvert-asn1-perl \ libexif \ + libffi \ libgpg-error\ libksba \ libmd \
meta/recipes-support/libffi/libffi/run-ptest new file +53 -0
53 lines not on this page · raw message
meta/recipes-support/libffi/libffi_3.5.2.bb +20 -1
@@ -12,6 +12,7 @@ LIC_FILES_CHKSUM = "file://LICENSE;md5=ce4763670c5b7756000561f9af1ab178" SRC_URI = "${GITHUB_BASE_URI}/download/v${PV}/${BPN}-${PV}.tar.gz \ file://not-win32.patch \ + file://run-ptest \ " SRC_URI[sha256sum] = "f3a3082a23b37c293a4fcd1053147b371f2ff91fa7ea1b2a52e335676bac82dc"
@@ -19,7 +20,7 @@ EXTRA_OECONF = "--disable-builddir" EXTRA_OECONF:class-native += "--with-gcc-arch=generic" EXTRA_OEMAKE:class-target = "LIBTOOLFLAGS='--tag=CC'" -inherit autotools texinfo multilib_header github-releases +inherit autotools texinfo multilib_header github-releases ptest do_install:append() { oe_multilib_header ffi.h ffitarget.h
@@ -30,3 +31,21 @@ do_install:append() { MIPS_INSTRUCTION_SET = "mips" BBCLASSEXTEND = "native nativesdk" + +RDEPENDS:${PN}-ptest += "packagegroup-core-buildessential" + +do_install_ptest() { + install -d ${D}${PTEST_PATH}/testsuite + for suite in libffi.bhaible libffi.call libffi.closures libffi.complex \ + libffi.go libffi.threads; do + cp -r ${S}/testsuite/$suite ${D}${PTEST_PATH}/testsuite/ + done + + install -d ${D}${PTEST_PATH}/include + install -m 0644 ${B}/fficonfig.h ${D}${PTEST_PATH}/include/ + install -m 0644 ${B}/include/ffi.h ${D}${PTEST_PATH}/include/ + install -m 0644 ${B}/include/ffitarget.h ${D}${PTEST_PATH}/include/ + + install -d ${D}${PTEST_PATH}/lib + ln -sf ${libdir}/libffi.so.8 ${D}${PTEST_PATH}/lib/libffi.so +}
Signature · 1 line
2.43.0
meta/conf/distro/include/ptest-packagelists.inc PTESTS_FAST = "\
Quoted 39 lines
-----Original Message-----
From: [email protected] <openembedded- [email protected]> On Behalf Of Pratik Farkase
Sent: Friday, May 8, 2026 12:24 PM
To: [email protected]
Cc: [email protected]; Pratik Farkase <[email protected]>
Subject: [OE-core][PATCH v2] libffi: add ptest supportAdd ptest support for libffi using all upstream test suites: bhaible, call, closures, complex, go, and threads.
Tests are compiled on-target against the installed libffi. Headers (including the generated
fficonfig.h) and a linker symlink are installed in the ptest directory to avoid depending on libffi-dev.Tested on qemux86-64 with ptest-runner: 199 PASS, 0 FAIL, 2 SKIP (349s).
Signed-off-by Pratik Farkase <[email protected]>---
Changes in v2 :
- Include all upstream test suites instead of only bhaible
- Installfficonfig.hrequired byffitest.h
- Add support for C++ tests (closures/unwindtest*.cc)
- Add -lpthread for threads suite--- .../distro/include/ptest-packagelists.inc | 1 + meta/recipes-support/libffi/libffi/run-ptest | 53 +++++++++++++++++++ meta/recipes-support/libffi/libffi_3.5.2.bb | 21 +++++++- 3 files changed, 74 insertions(+), 1 deletion(-) create mode 100755 meta/recipes-support/libffi/libffi/run-ptestmeta/conf/distro/include/ptest-packagelists.inc +0
-0@@ -34,6 +34,7 @@ PTESTS_FAST = "\
Comment for PTESTS_FASTS is: "ptests which take less than ~30s each", While commit message suggests this takes 10x more (349s)
Quoted 115 lines
meta/conf/distro/include/ptest-packagelists.inc +1
-0libconfig \ libconvert-asn1-perl \ libexif \ + libffi \ libgpg-error\ libksba \ libmd \meta/recipes- renamed from meta/recipes-support/libffi/libffi/run-ptest +0
-0support/libffi/libffi/run-ptest
new file mode 100755
index 0000000000..9f72becda1meta/recipes-support/libffi/libffi/run-ptest new file +53
-0@@ -0,0 +1,53 @@ +#!/bin/sh +# SPDX-License-Identifier: MIT + +PTEST_DIR=$(cd "$(dirname "$0")" && pwd) +CFLAGS="-O2 -Wall -I$PTEST_DIR/include -L$PTEST_DIR/lib" + +# bhaible tests validate by comparing paired output lines +for t in test-call test-callback; do + src="$PTEST_DIR/testsuite/libffi.bhaible/$t.c" + if gcc $CFLAGS -o "$PTEST_DIR/$t" "$src" -lffi 2>&1; then + "$PTEST_DIR/$t" > "$PTEST_DIR/$t.out" 2>&1 + if ! LC_ALL=C uniq -u < "$PTEST_DIR/$t.out" | grep -q .; then + echo "PASS: libffi.bhaible/$t" + else + echo "FAIL: libffi.bhaible/$t" + cat "$PTEST_DIR/$t.out" + fi + else + echo "SKIP: libffi.bhaible/$t" + fi + rm -f "$PTEST_DIR/$t" "$PTEST_DIR/$t.out" +done + +# Individual test programs from upstream dejagnu suites +for suite in libffi.call libffi.closures libffi.complex libffi.go libffi.threads; do + dir="$PTEST_DIR/testsuite/$suite" + [ -d "$dir" ] || continue + + extra="" + [ "$suite" = "libffi.threads" ] && extra="-lpthread" + + for src in "$dir"/*.c "$dir"/*.cc; do + [ -f "$src" ] || continue + grep -q "dg-do run" "$src" || continue + + name=$(basename "${src%.*}") + case "$src" in + *.cc) cc="g++";; + *) cc="gcc";; + esac + + if $cc $CFLAGS -o "$PTEST_DIR/$name" "$src" -lffi $extra 2>&1; then + if "$PTEST_DIR/$name" 2>&1; then + echo "PASS: $suite/$name" + else + echo "FAIL: $suite/$name" + fi + else + echo "SKIP: $suite/$name" + fi + rm -f "$PTEST_DIR/$name" + done +donemeta/recipes- renamed from meta/recipes-support/libffi/libffi_3.5.2.bb +0
-0support/libffi/libffi_3.5.2.bb
indexd3aa90d429..ec2b8032d3100644meta/recipes-support/libffi/libffi_3.5.2.bb +0
-0@@ -12,6 +12,7 @@ LIC_FILES_CHKSUM ="file://LICENSE;md5=ce4763670c5b7756000561f9af1ab178"
SRC_URI = "${GITHUB_BASE_URI}/download/v${PV}/${BPN}-${PV}.tar.gz \ file://not-win32.patch \ + file://run-ptest \ "SRC_URI[sha256sum] =
"f3a3082a23b37c293a4fcd1053147b371f2ff91fa7ea1b2a52e335676bac82dc"meta/recipes-support/libffi/libffi_3.5.2.bb +19
-2@@ -19,7 +20,7 @@ EXTRA_OECONF = "--disable-builddir" EXTRA_OECONF:class-native += "--with-gcc-arch=generic" EXTRA_OEMAKE:class-target = "LIBTOOLFLAGS='--tag=CC'" -inherit autotools texinfo multilib_header github-releases +inherit autotools texinfo multilib_header github-releases ptest do_install:append() { oe_multilib_header ffi.h ffitarget.h@@ -30,3 +31,21 @@ do_install:append() { MIPS_INSTRUCTION_SET = "mips" BBCLASSEXTEND = "native nativesdk" + +RDEPENDS:${PN}-ptest += "packagegroup-core-buildessential" + +do_install_ptest() { + install -d ${D}${PTEST_PATH}/testsuite + for suite in libffi.bhaible libffi.call libffi.closures libffi.complex \ + libffi.go libffi.threads; do + cp -r ${S}/testsuite/$suite ${D}${PTEST_PATH}/testsuite/ + done + + install -d ${D}${PTEST_PATH}/include + install -m 0644 ${B}/fficonfig.h ${D}${PTEST_PATH}/include/ + install -m 0644 ${B}/include/ffi.h ${D}${PTEST_PATH}/include/ + install -m 0644 ${B}/include/ffitarget.h ${D}${PTEST_PATH}/include/ + + install -d ${D}${PTEST_PATH}/lib + ln -sf ${libdir}/libffi.so.8 ${D}${PTEST_PATH}/lib/libffi.so +} --2.43.0
Hi Peter,
Good catch, thanks. v1 only had 2 tests (bhaible suite) which ran in ~29s, but v2 added all upstream test suites (199 tests) pushing it to 349s. I've moved libffi to PTESTS_SLOW in v3.
Best Regards,
Pratik
________________________________________
From: Marko, Peter <[email protected]>
Sent: Friday, May 8, 2026 1:25 PM
To: Pratik Farkase; [email protected]
Cc: [email protected]
Subject: RE: [OE-core][PATCH v2] libffi: add ptest support
meta/conf/distro/include/ptest-packagelists.inc PTESTS_FAST = "\
Quoted 37 lines
-----Original Message-----
From: [email protected] <openembedded- [email protected]> On Behalf Of Pratik Farkase
Sent: Friday, May 8, 2026 12:24 PM
To: [email protected]
Cc: [email protected]; Pratik Farkase <[email protected]>
Subject: [OE-core][PATCH v2] libffi: add ptest supportAdd ptest support for libffi using all upstream test suites: bhaible, call, closures, complex, go, and threads.
Tests are compiled on-target against the installed libffi. Headers (including the generated
fficonfig.h) and a linker symlink are installed in the ptest directory to avoid depending on libffi-dev.Tested on qemux86-64 with ptest-runner: 199 PASS, 0 FAIL, 2 SKIP (349s).
Signed-off-by Pratik Farkase <[email protected]>---
Changes in v2 :
- Include all upstream test suites instead of only bhaible
- Installfficonfig.hrequired byffitest.h
- Add support for C++ tests (closures/unwindtest*.cc)
- Add -lpthread for threads suite--- .../distro/include/ptest-packagelists.inc | 1 + meta/recipes-support/libffi/libffi/run-ptest | 53 +++++++++++++++++++ meta/recipes-support/libffi/libffi_3.5.2.bb | 21 +++++++- 3 files changed, 74 insertions(+), 1 deletion(-) create mode 100755 meta/recipes-support/libffi/libffi/run-ptestmeta/conf/distro/include/ptest-packagelists.inc +0
-0@@ -34,6 +34,7 @@ PTESTS_FAST = "\
Comment for PTESTS_FASTS is: "ptests which take less than ~30s each", While commit message suggests this takes 10x more (349s)
Quoted 115 lines
meta/conf/distro/include/ptest-packagelists.inc +1
-0libconfig \ libconvert-asn1-perl \ libexif \ + libffi \ libgpg-error\ libksba \ libmd \meta/recipes- renamed from meta/recipes-support/libffi/libffi/run-ptest +0
-0support/libffi/libffi/run-ptest
new file mode 100755
index 0000000000..9f72becda1meta/recipes-support/libffi/libffi/run-ptest new file +53
-0@@ -0,0 +1,53 @@ +#!/bin/sh +# SPDX-License-Identifier: MIT + +PTEST_DIR=$(cd "$(dirname "$0")" && pwd) +CFLAGS="-O2 -Wall -I$PTEST_DIR/include -L$PTEST_DIR/lib" + +# bhaible tests validate by comparing paired output lines +for t in test-call test-callback; do + src="$PTEST_DIR/testsuite/libffi.bhaible/$t.c" + if gcc $CFLAGS -o "$PTEST_DIR/$t" "$src" -lffi 2>&1; then + "$PTEST_DIR/$t" > "$PTEST_DIR/$t.out" 2>&1 + if ! LC_ALL=C uniq -u < "$PTEST_DIR/$t.out" | grep -q .; then + echo "PASS: libffi.bhaible/$t" + else + echo "FAIL: libffi.bhaible/$t" + cat "$PTEST_DIR/$t.out" + fi + else + echo "SKIP: libffi.bhaible/$t" + fi + rm -f "$PTEST_DIR/$t" "$PTEST_DIR/$t.out" +done + +# Individual test programs from upstream dejagnu suites +for suite in libffi.call libffi.closures libffi.complex libffi.go libffi.threads; do + dir="$PTEST_DIR/testsuite/$suite" + [ -d "$dir" ] || continue + + extra="" + [ "$suite" = "libffi.threads" ] && extra="-lpthread" + + for src in "$dir"/*.c "$dir"/*.cc; do + [ -f "$src" ] || continue + grep -q "dg-do run" "$src" || continue + + name=$(basename "${src%.*}") + case "$src" in + *.cc) cc="g++";; + *) cc="gcc";; + esac + + if $cc $CFLAGS -o "$PTEST_DIR/$name" "$src" -lffi $extra 2>&1; then + if "$PTEST_DIR/$name" 2>&1; then + echo "PASS: $suite/$name" + else + echo "FAIL: $suite/$name" + fi + else + echo "SKIP: $suite/$name" + fi + rm -f "$PTEST_DIR/$name" + done +donemeta/recipes- renamed from meta/recipes-support/libffi/libffi_3.5.2.bb +0
-0support/libffi/libffi_3.5.2.bb
indexd3aa90d429..ec2b8032d3100644meta/recipes-support/libffi/libffi_3.5.2.bb +0
-0@@ -12,6 +12,7 @@ LIC_FILES_CHKSUM ="file://LICENSE;md5=ce4763670c5b7756000561f9af1ab178"
SRC_URI = "${GITHUB_BASE_URI}/download/v${PV}/${BPN}-${PV}.tar.gz \ file://not-win32.patch \ + file://run-ptest \ "SRC_URI[sha256sum] =
"f3a3082a23b37c293a4fcd1053147b371f2ff91fa7ea1b2a52e335676bac82dc"meta/recipes-support/libffi/libffi_3.5.2.bb +19
-2@@ -19,7 +20,7 @@ EXTRA_OECONF = "--disable-builddir" EXTRA_OECONF:class-native += "--with-gcc-arch=generic" EXTRA_OEMAKE:class-target = "LIBTOOLFLAGS='--tag=CC'" -inherit autotools texinfo multilib_header github-releases +inherit autotools texinfo multilib_header github-releases ptest do_install:append() { oe_multilib_header ffi.h ffitarget.h@@ -30,3 +31,21 @@ do_install:append() { MIPS_INSTRUCTION_SET = "mips" BBCLASSEXTEND = "native nativesdk" + +RDEPENDS:${PN}-ptest += "packagegroup-core-buildessential" + +do_install_ptest() { + install -d ${D}${PTEST_PATH}/testsuite + for suite in libffi.bhaible libffi.call libffi.closures libffi.complex \ + libffi.go libffi.threads; do + cp -r ${S}/testsuite/$suite ${D}${PTEST_PATH}/testsuite/ + done + + install -d ${D}${PTEST_PATH}/include + install -m 0644 ${B}/fficonfig.h ${D}${PTEST_PATH}/include/ + install -m 0644 ${B}/include/ffi.h ${D}${PTEST_PATH}/include/ + install -m 0644 ${B}/include/ffitarget.h ${D}${PTEST_PATH}/include/ + + install -d ${D}${PTEST_PATH}/lib + ln -sf ${libdir}/libffi.so.8 ${D}${PTEST_PATH}/lib/libffi.so +} --2.43.0
PATCH v3 1/1 libffi: add ptest support
In reply to Pratik FarkaseAdd ptest support for libffi using all upstream test suites: bhaible, call, closures, complex, go, and threads.
Tests are compiled on-target against the installed libffi. Headers (including the generated fficonfig.h) and a linker symlink are installed in the ptest directory to avoid depending on libffi-dev.
Tested on qemux86-64 with ptest-runner: 199 PASS, 0 FAIL, 2 SKIP (349s).
---
Changes in v3:
- Move libffi from PTESTS_FAST to PTESTS_SLOW (349s) runtime
Changes in v2 :
- Include all upstream test suites instead of only bhaible
- Install fficonfig.h required by ffitest.h
- Add support for C++ tests (closures/unwindtest*.cc)
- Add -lpthread for threads suite
--- .../distro/include/ptest-packagelists.inc | 1 + meta/recipes-support/libffi/libffi/run-ptest | 53 +++++++++++++++++++ meta/recipes-support/libffi/libffi_3.5.2.bb | 21 +++++++- 3 files changed, 74 insertions(+), 1 deletion(-) create mode 100755 meta/recipes-support/libffi/libffi/run-ptest
meta/conf/distro/include/ptest-packagelists.inc +1 -0
@@ -126,6 +126,7 @@ PTESTS_SLOW = "\ less \ libevent \ libgcrypt \ + libffi \ libmodule-build-perl \ libpng \ ${@bb.utils.contains('DISTRO_FEATURES', 'seccomp', 'libseccomp', '',d)} \
meta/recipes-support/libffi/libffi/run-ptest new file +53 -0
53 lines not on this page · raw message
meta/recipes-support/libffi/libffi_3.5.2.bb +20 -1
@@ -12,6 +12,7 @@ LIC_FILES_CHKSUM = "file://LICENSE;md5=ce4763670c5b7756000561f9af1ab178" SRC_URI = "${GITHUB_BASE_URI}/download/v${PV}/${BPN}-${PV}.tar.gz \ file://not-win32.patch \ + file://run-ptest \ " SRC_URI[sha256sum] = "f3a3082a23b37c293a4fcd1053147b371f2ff91fa7ea1b2a52e335676bac82dc"
@@ -19,7 +20,7 @@ EXTRA_OECONF = "--disable-builddir" EXTRA_OECONF:class-native += "--with-gcc-arch=generic" EXTRA_OEMAKE:class-target = "LIBTOOLFLAGS='--tag=CC'" -inherit autotools texinfo multilib_header github-releases +inherit autotools texinfo multilib_header github-releases ptest do_install:append() { oe_multilib_header ffi.h ffitarget.h
@@ -30,3 +31,21 @@ do_install:append() { MIPS_INSTRUCTION_SET = "mips" BBCLASSEXTEND = "native nativesdk" + +RDEPENDS:${PN}-ptest += "packagegroup-core-buildessential" + +do_install_ptest() { + install -d ${D}${PTEST_PATH}/testsuite + for suite in libffi.bhaible libffi.call libffi.closures libffi.complex \ + libffi.go libffi.threads; do + cp -r ${S}/testsuite/$suite ${D}${PTEST_PATH}/testsuite/ + done + + install -d ${D}${PTEST_PATH}/include + install -m 0644 ${B}/fficonfig.h ${D}${PTEST_PATH}/include/ + install -m 0644 ${B}/include/ffi.h ${D}${PTEST_PATH}/include/ + install -m 0644 ${B}/include/ffitarget.h ${D}${PTEST_PATH}/include/ + + install -d ${D}${PTEST_PATH}/lib + ln -sf ${libdir}/libffi.so.8 ${D}${PTEST_PATH}/lib/libffi.so +}
Signature · 1 line
2.43.0
Quoted 11 lines · Pratik Farkase
Add ptest support for libffi using all upstream test suites: bhaible, call, closures, complex, go, and threads.
Tests are compiled on-target against the installed libffi. Headers (including the generated
fficonfig.h) and a linker symlink are installed in the ptest directory to avoid depending on libffi-dev.Tested on qemux86-64 with ptest-runner: 199 PASS, 0 FAIL, 2 SKIP (349s).
Signed-off-by Pratik Farkase <[email protected]>---
Hi Pratik,
Thanks for the patch.
Do you have any clue about why this is triggering some selftest failures? I'm a bit surprised by that, but I confirm it comes from this patch:
2026-05-10 16:16:13,784 - oe-selftest - INFO - locales.LocalesTest.test_locales_off (subunit.RemotedTestCase) 2026-05-10 16:16:13,787 - oe-selftest - INFO - ... FAIL
...
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-c, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-el-gr, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-en-gb, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-es-es, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-fa-ir, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-hr-hr, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-ja-jp.euc-jp, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-lt-lt, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-pl-pl, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-pl-pl.iso-8859-2, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-ru-ru, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: Fatal QA errors were found, failing task.
...
2026-05-10 16:19:03,004 - oe-selftest - INFO - locales.LocalesTest.test_locales_on (subunit.RemotedTestCase) 2026-05-10 16:19:03,005 - oe-selftest - INFO - ... FAIL
...
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-c, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-el-gr, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-en-gb, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-es-es, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-fa-ir, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-hr-hr, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-ja-jp.euc-jp, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-lt-lt, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-pl-pl, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-pl-pl.iso-8859-2, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-ru-ru, but it isn't a build dependency? [build-deps]
NOTE: Setscene tasks completed
https://autobuilder.yoctoproject.org/valkyrie/#/builders/23/builds/3929
https://autobuilder.yoctoproject.org/valkyrie/#/builders/35/builds/3827
https://autobuilder.yoctoproject.org/valkyrie/#/builders/48/builds/3696
I also note some is install error:
ERROR: lib64-core-image-sato-sdk-1.0-r0 do_rootfs: Could not invoke dnf. Command '...' returned 1:
...
Error: Transaction test error:
file /usr/bin/g++ conflicts between attempted installs of lib64-g++-symlinks-15.2.0-r0.x86_64 and g++-symlinks-15.2.0-r0.core2_32 file /usr/bin/cpp conflicts between attempted installs of cpp-symlinks-15.2.0-r0.core2_32 and lib64-cpp-symlinks-15.2.0-r0.x86_64 file /usr/bin/gcc conflicts between attempted installs of lib64-gcc-symlinks-15.2.0-r0.x86_64 and gcc-symlinks-15.2.0-r0.core2_32
Finally, I note a ptest failures (on riscv64). This one look like intermittent, if it happens too often, we might want to exclude the failing test:
Failed ptests:
{'libffi': ['libffi.complex/complex_int']}
https://autobuilder.yoctoproject.org/valkyrie/#/builders/56/builds/1576
I will try to run more tests, as again, I'm a bit surprised by some of these failures.
Thanks,
Mathieu
Signature · 3 lines
Mathieu Dubois-Briand, Bootlin Embedded Linux and Kernel engineering https://bootlin.com
Hi Mathieu,
Thanks for looking into this.
The glib-2.0 locale issue looks pre-existing — my patch only modifies ptest-packagelists.inc which likely invalidated sstate and exposed it.
The multilib symlinks conflict I can address in v4 if confirmed.
The complex_int failure on riscv64 could be arch-specific — happy to exclude it if it reproduces consistently.
Let me know what you find after further testing.
Best Regards,
Pratik
________________________________________
From: Mathieu Dubois-Briand <[email protected]>
Sent: Monday, May 11, 2026 8:09 AM
To: Pratik Farkase; [email protected]
Cc: [email protected]
Subject: Re: [OE-core][PATCH v3] libffi: add ptest support
Quoted 11 lines · Pratik Farkase
Add ptest support for libffi using all upstream test suites: bhaible, call, closures, complex, go, and threads.
Tests are compiled on-target against the installed libffi. Headers (including the generated
fficonfig.h) and a linker symlink are installed in the ptest directory to avoid depending on libffi-dev.Tested on qemux86-64 with ptest-runner: 199 PASS, 0 FAIL, 2 SKIP (349s).
Signed-off-by Pratik Farkase <[email protected]>---
Hi Pratik,
Thanks for the patch.
Do you have any clue about why this is triggering some selftest failures? I'm a bit surprised by that, but I confirm it comes from this patch:
2026-05-10 16:16:13,784 - oe-selftest - INFO - locales.LocalesTest.test_locales_off (subunit.RemotedTestCase) 2026-05-10 16:16:13,787 - oe-selftest - INFO - ... FAIL
...
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-c, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-el-gr, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-en-gb, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-es-es, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-fa-ir, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-hr-hr, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-ja-jp.euc-jp, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-lt-lt, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-pl-pl, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-pl-pl.iso-8859-2, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-ru-ru, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: Fatal QA errors were found, failing task.
...
2026-05-10 16:19:03,004 - oe-selftest - INFO - locales.LocalesTest.test_locales_on (subunit.RemotedTestCase) 2026-05-10 16:19:03,005 - oe-selftest - INFO - ... FAIL
...
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-c, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-el-gr, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-en-gb, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-es-es, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-fa-ir, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-hr-hr, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-ja-jp.euc-jp, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-lt-lt, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-pl-pl, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-pl-pl.iso-8859-2, but it isn't a build dependency? [build-deps]
ERROR: glib-2.0-1_2.88.0-r0 do_package_qa: QA Issue: glib-2.0-ptest rdepends on locale-base-ru-ru, but it isn't a build dependency? [build-deps]
NOTE: Setscene tasks completed
https://autobuilder.yoctoproject.org/valkyrie/#/builders/23/builds/3929
https://autobuilder.yoctoproject.org/valkyrie/#/builders/35/builds/3827
https://autobuilder.yoctoproject.org/valkyrie/#/builders/48/builds/3696
I also note some is install error:
ERROR: lib64-core-image-sato-sdk-1.0-r0 do_rootfs: Could not invoke dnf. Command '...' returned 1:
...
Error: Transaction test error:
file /usr/bin/g++ conflicts between attempted installs of lib64-g++-symlinks-15.2.0-r0.x86_64 and g++-symlinks-15.2.0-r0.core2_32 file /usr/bin/cpp conflicts between attempted installs of cpp-symlinks-15.2.0-r0.core2_32 and lib64-cpp-symlinks-15.2.0-r0.x86_64 file /usr/bin/gcc conflicts between attempted installs of lib64-gcc-symlinks-15.2.0-r0.x86_64 and gcc-symlinks-15.2.0-r0.core2_32
Finally, I note a ptest failures (on riscv64). This one look like intermittent, if it happens too often, we might want to exclude the failing test:
Failed ptests:
{'libffi': ['libffi.complex/complex_int']}
https://autobuilder.yoctoproject.org/valkyrie/#/builders/56/builds/1576
I will try to run more tests, as again, I'm a bit surprised by some of these failures.
Thanks,
Mathieu
--
Mathieu Dubois-Briand, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
Quoted 9 lines · Pratik Farkase
Hi Mathieu,
Thanks for looking into this.
The glib-2.0 locale issue looks pre-existing — my patch only modifies ptest-packagelists.inc which likely invalidated sstate and exposed it.
The multilib symlinks conflict I can address in v4 if confirmed.
Quoted 8 lines
The
complex_intfailure on riscv64 could be arch-specific — happy to exclude it if it reproduces consistently.Let me know what you find after further testing.
Best Regards,
Pratik
So I tired again and applied this series on top of an otherwise fully working branch. I basically got all the same failures:
https://autobuilder.yoctoproject.org/valkyrie/#/builders/56/builds/1577
This does include the ptest one, so I believe we should disable the 'libffi.complex/complex_int' test.
Thanks,
Mathieu
Signature · 3 lines
Mathieu Dubois-Briand, Bootlin Embedded Linux and Kernel engineering https://bootlin.com
Hi Mathieu,
I tested v3 on qemux86-64 against current master and was unable to reproduce the glib-2.0 locale or multilib symlinks failures.
Could you share the steps or configuration you used to reproduce the these two issues? That would help me investigate further.
Best Regards,
Pratik
________________________________________
From: Mathieu Dubois-Briand <[email protected]>
Sent: Monday, May 11, 2026 2:25 PM
To: Pratik Farkase; [email protected]
Cc: [email protected]
Subject: Re: [OE-core][PATCH v3] libffi: add ptest support
Quoted 9 lines · Pratik Farkase
Hi Mathieu,
Thanks for looking into this.
The glib-2.0 locale issue looks pre-existing — my patch only modifies ptest-packagelists.inc which likely invalidated sstate and exposed it.
The multilib symlinks conflict I can address in v4 if confirmed.
Quoted 8 lines
The
complex_intfailure on riscv64 could be arch-specific — happy to exclude it if it reproduces consistently.Let me know what you find after further testing.
Best Regards,
Pratik
So I tired again and applied this series on top of an otherwise fully working branch. I basically got all the same failures:
https://autobuilder.yoctoproject.org/valkyrie/#/builders/56/builds/1577
This does include the ptest one, so I believe we should disable the 'libffi.complex/complex_int' test.
Thanks,
Mathieu
--
Mathieu Dubois-Briand, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
Quoted 11 lines · Pratik Farkase
Hi Mathieu,
I tested v3 on qemux86-64 against current master and was unable to reproduce the glib-2.0 locale or multilib symlinks failures.
Could you share the steps or configuration you used to reproduce the these two issues? That would help me investigate further.
Best Regards,
Pratik
Hi Pratik,
Regarding the multilib one, I just managed to reproduce it locally:
Using tag oecore/autobuilder.yoctoproject.org/valkyrie/a-full-3808 from poky-ci-archive git:
https://git.yoctoproject.org/poky-ci-archive/tag/?h=oecore/autobuilder.yoctoproject.org/valkyrie/a-full-3808
Default local.conf template, with these lines added:PACKAGE_CLASSES = "package_rpm package_deb package_ipk"OE_FRAGMENTS += 'core/yocto-autobuilder/multilib-x86-lib64'SANITY_TESTED_DISTROS = ''OEQA_TESTDISPLAY = ':1'OE_FRAGMENTS += 'core/yocto-autobuilder/autobuilder core/yocto-autobuilder/autobuilder-resource-constraints' EXTRA_IMAGE_FEATURES ?= 'allow-empty-password empty-root-password allow-root-login' OE_FRAGMENTS += 'machine/qemux86 distro/poky'
Then:
bitbake lib64-core-image-sato-sdk
...
ERROR: lib64-core-image-sato-sdk-1.0-r0 do_rootfs: Could not invoke dnf. Command '...' returned 1:
...
Running transaction check
Transaction check succeeded.
Running transaction test
Error: Transaction test error:
file /usr/bin/g++ conflicts between attempted installs of lib64-g++-symlinks-15.2.0-r0.x86_64 and g++-symlinks-15.2.0-r0.core2_32 file /usr/bin/cpp conflicts between attempted installs of cpp-symlinks-15.2.0-r0.core2_32 and lib64-cpp-symlinks-15.2.0-r0.x86_64 file /usr/bin/gcc conflicts between attempted installs of lib64-gcc-symlinks-15.2.0-r0.x86_64 and gcc-symlinks-15.2.0-r0.core2_32
I will try to reproduce the selftest failures.
Thanks,
Mathieu
Signature · 3 lines
Mathieu Dubois-Briand, Bootlin Embedded Linux and Kernel engineering https://bootlin.com
Quoted 43 lines · Mathieu Dubois-Briand
Quoted 11 lines · Pratik Farkase
Hi Mathieu,
I tested v3 on qemux86-64 against current master and was unable to reproduce the glib-2.0 locale or multilib symlinks failures.
Could you share the steps or configuration you used to reproduce the these two issues? That would help me investigate further.
Best Regards,
PratikHi Pratik,
Regarding the multilib one, I just managed to reproduce it locally:
Using tag oecore/autobuilder.yoctoproject.org/valkyrie/a-full-3808 from poky-ci-archive git:
https://git.yoctoproject.org/poky-ci-archive/tag/?h=oecore/autobuilder.yoctoproject.org/valkyrie/a-full-3808Default local.conf template, with these lines added:
PACKAGE_CLASSES= "package_rpmpackage_debpackage_ipk"OE_FRAGMENTS+= 'core/yocto-autobuilder/multilib-x86-lib64'SANITY_TESTED_DISTROS= ''OEQA_TESTDISPLAY= ':1'OE_FRAGMENTS+= 'core/yocto-autobuilder/autobuilder core/yocto-autobuilder/autobuilder-resource-constraints'EXTRA_IMAGE_FEATURES?= 'allow-empty-password empty-root-password allow-root-login'OE_FRAGMENTS+= 'machine/qemux86 distro/poky'Then:
bitbake lib64-core-image-sato-sdk
...
ERROR: lib64-core-image-sato-sdk-1.0-r0do_rootfs: Could not invoke dnf. Command '...' returned 1:
...
Running transaction check
Transaction check succeeded.
Running transaction test
Error: Transaction test error:
file/usr/bin/g++conflicts between attempted installs of lib64-g++-symlinks-15.2.0-r0.x86_64 and g++-symlinks-15.2.0-r0.core2_32 file/usr/bin/cppconflicts between attempted installs of cpp-symlinks-15.2.0-r0.core2_32 and lib64-cpp-symlinks-15.2.0-r0.x86_64 file/usr/bin/gccconflicts between attempted installs of lib64-gcc-symlinks-15.2.0-r0.x86_64 and gcc-symlinks-15.2.0-r0.core2_32
Quoted 2 lines
I will try to reproduce the selftest failures.
I don't have the selftest issue locally. We can assume it is some sstate pollution for now, we will go back to it later if it happens again.
Signature · 3 lines
Mathieu Dubois-Briand, Bootlin Embedded Linux and Kernel engineering https://bootlin.com
Hi Mathieu,
I was able to reproduce the multilib symlinks conflict locally with the same configuration. The root cause is that packagegroup-core-buildessential pulls in gcc-symlinks, g++-symlinks, and cpp-symlinks which conflict in multilib images (both arches try to own /usr/bin/gcc).
The fix is to replace:
RDEPENDS:${PN}-ptest += "packagegroup-core-buildessential"
with:
RDEPENDS:${PN}-ptest += "gcc g++ make"
And update run-ptest to find the compiler dynamically rather than hardcoding gcc/g++. I've verified this resolves the multilib conflict -lib64-core-image-sato-sdk builds cleanly with the fix applied.
Once you've had a chance to reproduce the glib-2.0 locale selftest issue, I'll send v4 addressing all three problems:
- glib-2.0 locale selftest issue
- Multilib symlinks conflict (use gcc/g++/make instead of packagegroup)
- complex_int exclusion (fails on riscv64)
Best Regards,
Pratik
________________________________________
From: Mathieu Dubois-Briand <[email protected]>
Sent: Wednesday, May 13, 2026 7:33 AM
To: Pratik Farkase; [email protected]
Cc: [email protected]
Subject: Re: [OE-core][PATCH v3] libffi: add ptest support
Quoted 43 lines · Mathieu Dubois-Briand
Quoted 11 lines · Pratik Farkase
Hi Mathieu,
I tested v3 on qemux86-64 against current master and was unable to reproduce the glib-2.0 locale or multilib symlinks failures.
Could you share the steps or configuration you used to reproduce the these two issues? That would help me investigate further.
Best Regards,
PratikHi Pratik,
Regarding the multilib one, I just managed to reproduce it locally:
Using tag oecore/autobuilder.yoctoproject.org/valkyrie/a-full-3808 from poky-ci-archive git:
https://git.yoctoproject.org/poky-ci-archive/tag/?h=oecore/autobuilder.yoctoproject.org/valkyrie/a-full-3808Default local.conf template, with these lines added:
PACKAGE_CLASSES= "package_rpmpackage_debpackage_ipk"OE_FRAGMENTS+= 'core/yocto-autobuilder/multilib-x86-lib64'SANITY_TESTED_DISTROS= ''OEQA_TESTDISPLAY= ':1'OE_FRAGMENTS+= 'core/yocto-autobuilder/autobuilder core/yocto-autobuilder/autobuilder-resource-constraints'EXTRA_IMAGE_FEATURES?= 'allow-empty-password empty-root-password allow-root-login'OE_FRAGMENTS+= 'machine/qemux86 distro/poky'Then:
bitbake lib64-core-image-sato-sdk
...
ERROR: lib64-core-image-sato-sdk-1.0-r0do_rootfs: Could not invoke dnf. Command '...' returned 1:
...
Running transaction check
Transaction check succeeded.
Running transaction test
Error: Transaction test error:
file/usr/bin/g++conflicts between attempted installs of lib64-g++-symlinks-15.2.0-r0.x86_64 and g++-symlinks-15.2.0-r0.core2_32 file/usr/bin/cppconflicts between attempted installs of cpp-symlinks-15.2.0-r0.core2_32 and lib64-cpp-symlinks-15.2.0-r0.x86_64 file/usr/bin/gccconflicts between attempted installs of lib64-gcc-symlinks-15.2.0-r0.x86_64 and gcc-symlinks-15.2.0-r0.core2_32
Quoted 2 lines
I will try to reproduce the selftest failures.
I don't have the selftest issue locally. We can assume it is some sstate pollution for now, we will go back to it later if it happens again.
--
Mathieu Dubois-Briand, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
Quoted 11 lines · Pratik Farkase
Hi Mathieu,
I was able to reproduce the multilib symlinks conflict locally with the same configuration. The root cause is that packagegroup-core-buildessential pulls in gcc-symlinks, g++-symlinks, and cpp-symlinks which conflict in multilib images (both arches try to own
/usr/bin/gcc).The fix is to replace:
RDEPENDS:${PN}-ptest += "packagegroup-core-buildessential"
with:
RDEPENDS:${PN}-ptest += "gcc g++ make"And update run-ptest to find the compiler dynamically rather than hardcoding gcc/g++. I've verified this resolves the multilib conflict -lib64-core-image-sato-sdk builds cleanly with the fix applied.
Thanks.
Quoted 5 lines
Once you've had a chance to reproduce the glib-2.0 locale selftest issue, I'll send v4 addressing all three problems:
- glib-2.0 locale selftest issue
- Multilib symlinks conflict (use gcc/g++/make instead of packagegroup)
-complex_intexclusion (fails on riscv64)
I failed to do that yesterday, and I won't have time to look at this in the coming days. As I said, it might be some other issue that just shows here.
Send your v4 when you are ready, if we are a bit lucky we won't see this issue anymore.
Quoted 65 lines
Best Regards,
Pratik________________________________________
From: Mathieu Dubois-Briand <[email protected]>
Sent: Wednesday, May 13, 2026 7:33 AM
To: Pratik Farkase; [email protected]
Cc: [email protected]
Subject: Re: [OE-core][PATCH v3] libffi: add ptest supportQuoted 43 lines · Mathieu Dubois-Briand
Quoted 11 lines · Pratik Farkase
Hi Mathieu,
I tested v3 on qemux86-64 against current master and was unable to reproduce the glib-2.0 locale or multilib symlinks failures.
Could you share the steps or configuration you used to reproduce the these two issues? That would help me investigate further.
Best Regards,
PratikHi Pratik,
Regarding the multilib one, I just managed to reproduce it locally:
Using tag oecore/autobuilder.yoctoproject.org/valkyrie/a-full-3808 from poky-ci-archive git:
https://git.yoctoproject.org/poky-ci-archive/tag/?h=oecore/autobuilder.yoctoproject.org/valkyrie/a-full-3808Default local.conf template, with these lines added:
PACKAGE_CLASSES= "package_rpmpackage_debpackage_ipk"OE_FRAGMENTS+= 'core/yocto-autobuilder/multilib-x86-lib64'SANITY_TESTED_DISTROS= ''OEQA_TESTDISPLAY= ':1'OE_FRAGMENTS+= 'core/yocto-autobuilder/autobuilder core/yocto-autobuilder/autobuilder-resource-constraints'EXTRA_IMAGE_FEATURES?= 'allow-empty-password empty-root-password allow-root-login'OE_FRAGMENTS+= 'machine/qemux86 distro/poky'Then:
bitbake lib64-core-image-sato-sdk
...
ERROR: lib64-core-image-sato-sdk-1.0-r0do_rootfs: Could not invoke dnf. Command '...' returned 1:
...
Running transaction check
Transaction check succeeded.
Running transaction test
Error: Transaction test error:
file/usr/bin/g++conflicts between attempted installs of lib64-g++-symlinks-15.2.0-r0.x86_64 and g++-symlinks-15.2.0-r0.core2_32 file/usr/bin/cppconflicts between attempted installs of cpp-symlinks-15.2.0-r0.core2_32 and lib64-cpp-symlinks-15.2.0-r0.x86_64 file/usr/bin/gccconflicts between attempted installs of lib64-gcc-symlinks-15.2.0-r0.x86_64 and gcc-symlinks-15.2.0-r0.core2_32Quoted 2 lines
I will try to reproduce the selftest failures.
I don't have the selftest issue locally. We can assume it is some sstate pollution for now, we will go back to it later if it happens again.
--
Mathieu Dubois-Briand, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
Signature · 3 lines
Mathieu Dubois-Briand, Bootlin Embedded Linux and Kernel engineering https://bootlin.com
Hi Mathieu,
I have sent a v4 which addresses the issues you reported:
- Replaced packagegroup-core-buildessential with "gcc g++ make" to fix the multilib symlinks conflict in lib64-core-image-sato-sdk
- Excluded complex_int test (fails on riscv64)
- Compiler is detected dynamically in run-ptest (works without symlinks)
Best Regards,
Pratik
________________________________________
From: Mathieu Dubois-Briand <[email protected]>
Sent: Wednesday, May 13, 2026 1:31 PM
To: Pratik Farkase; [email protected]
Cc: [email protected]
Subject: Re: [OE-core][PATCH v3] libffi: add ptest support
Quoted 11 lines · Pratik Farkase
Hi Mathieu,
I was able to reproduce the multilib symlinks conflict locally with the same configuration. The root cause is that packagegroup-core-buildessential pulls in gcc-symlinks, g++-symlinks, and cpp-symlinks which conflict in multilib images (both arches try to own
/usr/bin/gcc).The fix is to replace:
RDEPENDS:${PN}-ptest += "packagegroup-core-buildessential"
with:
RDEPENDS:${PN}-ptest += "gcc g++ make"And update run-ptest to find the compiler dynamically rather than hardcoding gcc/g++. I've verified this resolves the multilib conflict -lib64-core-image-sato-sdk builds cleanly with the fix applied.
Thanks.
Quoted 5 lines
Once you've had a chance to reproduce the glib-2.0 locale selftest issue, I'll send v4 addressing all three problems:
- glib-2.0 locale selftest issue
- Multilib symlinks conflict (use gcc/g++/make instead of packagegroup)
-complex_intexclusion (fails on riscv64)
I failed to do that yesterday, and I won't have time to look at this in the coming days. As I said, it might be some other issue that just shows here.
Send your v4 when you are ready, if we are a bit lucky we won't see this issue anymore.
Quoted 58 lines
Best Regards,
Pratik________________________________________
From: Mathieu Dubois-Briand <[email protected]>
Sent: Wednesday, May 13, 2026 7:33 AM
To: Pratik Farkase; [email protected]
Cc: [email protected]
Subject: Re: [OE-core][PATCH v3] libffi: add ptest supportQuoted 43 lines · Mathieu Dubois-Briand
Quoted 11 lines · Pratik Farkase
Hi Mathieu,
I tested v3 on qemux86-64 against current master and was unable to reproduce the glib-2.0 locale or multilib symlinks failures.
Could you share the steps or configuration you used to reproduce the these two issues? That would help me investigate further.
Best Regards,
PratikHi Pratik,
Regarding the multilib one, I just managed to reproduce it locally:
Using tag oecore/autobuilder.yoctoproject.org/valkyrie/a-full-3808 from poky-ci-archive git:
https://git.yoctoproject.org/poky-ci-archive/tag/?h=oecore/autobuilder.yoctoproject.org/valkyrie/a-full-3808Default local.conf template, with these lines added:
PACKAGE_CLASSES= "package_rpmpackage_debpackage_ipk"OE_FRAGMENTS+= 'core/yocto-autobuilder/multilib-x86-lib64'SANITY_TESTED_DISTROS= ''OEQA_TESTDISPLAY= ':1'OE_FRAGMENTS+= 'core/yocto-autobuilder/autobuilder core/yocto-autobuilder/autobuilder-resource-constraints'EXTRA_IMAGE_FEATURES?= 'allow-empty-password empty-root-password allow-root-login'OE_FRAGMENTS+= 'machine/qemux86 distro/poky'Then:
bitbake lib64-core-image-sato-sdk
...
ERROR: lib64-core-image-sato-sdk-1.0-r0do_rootfs: Could not invoke dnf. Command '...' returned 1:
...
Running transaction check
Transaction check succeeded.
Running transaction test
Error: Transaction test error:
file/usr/bin/g++conflicts between attempted installs of lib64-g++-symlinks-15.2.0-r0.x86_64 and g++-symlinks-15.2.0-r0.core2_32 file/usr/bin/cppconflicts between attempted installs of cpp-symlinks-15.2.0-r0.core2_32 and lib64-cpp-symlinks-15.2.0-r0.x86_64 file/usr/bin/gccconflicts between attempted installs of lib64-gcc-symlinks-15.2.0-r0.x86_64 and gcc-symlinks-15.2.0-r0.core2_32Quoted 2 lines
I will try to reproduce the selftest failures.
Quoted 7 lines
I don't have the selftest issue locally. We can assume it is some sstate pollution for now, we will go back to it later if it happens again.
--
Mathieu Dubois-Briand, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
--
Mathieu Dubois-Briand, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
PATCH v4 1/1 libffi: add ptest support
In reply to Pratik FarkaseAdd ptest support for libffi using all upstream test suites: bhaible, call, closures, complex, go, and threads.
Tests are compiled on-target against the installed libffi. Headers (including the generated fficonfig.h) and a linker symlink are installed in the ptest directory to avoid depending on libffi-dev.
The complex_int test is excluded as it uses non-standard _Complex int which is unsupported on some architectures (e.g. riscv64).
Tested on qemux86-64 with ptest-runner: 198 PASS, 0 FAIL, 2 SKIP (327s).
---
Changes in v4:
- Replace packagegroup-core-buildessential with gcc/g++/make to fix multilib symlinks conflict in lib64-core-image-sato-sdk
- Detect compiler dynamically in run-ptest (works without symlinks)
- Exclude complex_int test (fails on riscv64)
Changes in v3:
- Move libffi from PTESTS_FAST to PTESTS_SLOW (349s) runtime
Changes in v2 :
- Include all upstream test suites instead of only bhaible
- Install fficonfig.h required by ffitest.h
- Add support for C++ tests (closures/unwindtest*.cc)
- Add -lpthread for threads suite
--- .../distro/include/ptest-packagelists.inc | 1 + meta/recipes-support/libffi/libffi/run-ptest | 62 +++++++++++++++++++ meta/recipes-support/libffi/libffi_3.5.2.bb | 23 ++++++- 3 files changed, 84 insertions(+), 2 deletions(-) create mode 100755 meta/recipes-support/libffi/libffi/run-ptest
meta/conf/distro/include/ptest-packagelists.inc +1 -0
@@ -126,6 +126,7 @@ PTESTS_SLOW = "\ less \ libevent \ libgcrypt \ + libffi \ libmodule-build-perl \ libpng \ ${@bb.utils.contains('DISTRO_FEATURES', 'seccomp', 'libseccomp', '',d)} \
meta/recipes-support/libffi/libffi/run-ptest new file +62 -0
62 lines not on this page · raw message
meta/recipes-support/libffi/libffi_3.5.2.bb +21 -2
@@ -12,6 +12,7 @@ LIC_FILES_CHKSUM = "file://LICENSE;md5=ce4763670c5b7756000561f9af1ab178" SRC_URI = "${GITHUB_BASE_URI}/download/v${PV}/${BPN}-${PV}.tar.gz \ file://not-win32.patch \ + file://run-ptest \ " SRC_URI[sha256sum] = "f3a3082a23b37c293a4fcd1053147b371f2ff91fa7ea1b2a52e335676bac82dc"
@@ -19,10 +20,10 @@ EXTRA_OECONF = "--disable-builddir" EXTRA_OECONF:class-native += "--with-gcc-arch=generic" EXTRA_OEMAKE:class-target = "LIBTOOLFLAGS='--tag=CC'" -inherit autotools texinfo multilib_header github-releases +inherit autotools texinfo multilib_header github-releases ptest do_install:append() { - oe_multilib_header ffi.h ffitarget.h + oe_multilib_header ffi.h ffitarget.h } # Doesn't compile in MIPS16e mode due to use of hand-written
@@ -30,3 +31,21 @@ do_install:append() { MIPS_INSTRUCTION_SET = "mips" BBCLASSEXTEND = "native nativesdk" + +RDEPENDS:${PN}-ptest += "gcc g++ make" + +do_install_ptest() { + install -d ${D}${PTEST_PATH}/testsuite + for suite in libffi.bhaible libffi.call libffi.closures libffi.complex \ + libffi.go libffi.threads; do + cp -r ${S}/testsuite/$suite ${D}${PTEST_PATH}/testsuite/ + done + + install -d ${D}${PTEST_PATH}/include + install -m 0644 ${B}/fficonfig.h ${D}${PTEST_PATH}/include/ + install -m 0644 ${B}/include/ffi.h ${D}${PTEST_PATH}/include/ + install -m 0644 ${B}/include/ffitarget.h ${D}${PTEST_PATH}/include/ + + install -d ${D}${PTEST_PATH}/lib + ln -sf ${libdir}/libffi.so.8 ${D}${PTEST_PATH}/lib/libffi.so +}
Signature · 1 line
2.43.0
Quoted 7 lines · Pratik Farkase via lists.openembedded.org
Add ptest support for libffi using all upstream test suites: bhaible, call, closures, complex, go, and threads.
Tests are compiled on-target against the installed libffi. Headers (including the generated
fficonfig.h) and a linker symlink are installed in the ptest directory to avoid depending on libffi-dev.
The point of ptests is to do integration testing of what we build, which this does not do. As you’re manually doing the compilation in run-ptest, why can’t this be done at build time?
meta/recipes-support/libffi/libffi/run-ptest line 10
Quoted 6 lines
meta/recipes-support/libffi/libffi/run-ptest new file +6
-0+CC="gcc" +CXX="g++" +if ! command -v gcc >/dev/null 2>&1; then + CC=$(find /usr/bin -name "*-gcc" | head -n 1) + CXX=$(find /usr/bin -name "*-g++" | head -n 1) +fi
Please don’t be so overly defensive. If you were to continue with compilation on the target, then just RDEPEND on gcc-symlinks and then gcc is on PATH.
Quoted 42 lines
meta/recipes-support/libffi/libffi/run-ptest new file +42
-0+CFLAGS="-O2 -Wall -I$PTEST_DIR/include -L$PTEST_DIR/lib" + +# bhaible tests validate by comparing paired output lines +for t in test-call test-callback; do + src="$PTEST_DIR/testsuite/libffi.bhaible/$t.c" + if $CC $CFLAGS -o "$PTEST_DIR/$t" "$src" -lffi 2>&1; then + "$PTEST_DIR/$t" > "$PTEST_DIR/$t.out" 2>&1 + if ! LC_ALL=C uniq -u < "$PTEST_DIR/$t.out" | grep -q .; then + echo "PASS: libffi.bhaible/$t" + else + echo "FAIL: libffi.bhaible/$t" + cat "$PTEST_DIR/$t.out" + fi + else + echo "SKIP: libffi.bhaible/$t" + fi + rm -f "$PTEST_DIR/$t" "$PTEST_DIR/$t.out" +done + +# Individual test programs from upstream dejagnu suites +for suite in libffi.call libffi.closures libffi.complex libffi.go libffi.threads; do + dir="$PTEST_DIR/testsuite/$suite" + [ -d "$dir" ] || continue + + extra="" + [ "$suite" = "libffi.threads" ] && extra="-lpthread" + + for src in "$dir"/*.c "$dir"/*.cc; do + [ -f "$src" ] || continue + grep -q "dg-do run" "$src" || continue + + name=$(basename "${src%.*}") + + # complex_int uses non-standard _Complex int, fails on some arches (riscv64) + [ "$name" = "complex_int" ] && continue + case "$src" in + *.cc) cc="$CXX";; + *) cc="$CC";; + esac + + if $cc $CFLAGS -o "$PTEST_DIR/$name" "$src" -lffi $extra 2>&1; then + if "$PTEST_DIR/$name" 2>&1; then
meta/recipes-support/libffi/libffi/run-ptest line 62
Quoted 10 lines
meta/recipes-support/libffi/libffi/run-ptest new file +10
-0+ echo "PASS: $suite/$name" + else + echo "FAIL: $suite/$name" + fi + else + echo "SKIP: $suite/$name" + fi + rm -f "$PTEST_DIR/$name" + done +done
This is a _lot_ of logic that we have to maintain forever to run some integration tests…
Ross
Hi Ross,
I initially compiled on target thinking it would better to test against the installed library and headers as the user would see them. The tests just validate libffi's functionality, which works the same whether compiled at build time or on target but yes it then does makes the code more complex and hard to maintain too i agree. I'll rework this to compile the tests at build time in do_compile_ptest and just install the binaries so that run-ptest has minimalistic work to do.
Will send v5 once I've simplified and tested it.
Best Regards,
Pratik
________________________________________
From: Ross Burton <[email protected]>
Sent: Wednesday, May 20, 2026 6:02 PM
To: Pratik Farkase
Cc: [email protected]; [email protected]
Subject: Re: [OE-core][PATCH v4] libffi: add ptest support
Quoted 7 lines · Pratik Farkase via lists.openembedded.org
Add ptest support for libffi using all upstream test suites: bhaible, call, closures, complex, go, and threads.
Tests are compiled on-target against the installed libffi. Headers (including the generated
fficonfig.h) and a linker symlink are installed in the ptest directory to avoid depending on libffi-dev.
The point of ptests is to do integration testing of what we build, which this does not do. As you’re manually doing the compilation in run-ptest, why can’t this be done at build time?
meta/recipes-support/libffi/libffi/run-ptest line 10
Quoted 6 lines
meta/recipes-support/libffi/libffi/run-ptest new file +6
-0+CC="gcc" +CXX="g++" +if ! command -v gcc >/dev/null 2>&1; then + CC=$(find /usr/bin -name "*-gcc" | head -n 1) + CXX=$(find /usr/bin -name "*-g++" | head -n 1) +fi
Please don’t be so overly defensive. If you were to continue with compilation on the target, then just RDEPEND on gcc-symlinks and then gcc is on PATH.
Quoted 42 lines
meta/recipes-support/libffi/libffi/run-ptest new file +42
-0+CFLAGS="-O2 -Wall -I$PTEST_DIR/include -L$PTEST_DIR/lib" + +# bhaible tests validate by comparing paired output lines +for t in test-call test-callback; do + src="$PTEST_DIR/testsuite/libffi.bhaible/$t.c" + if $CC $CFLAGS -o "$PTEST_DIR/$t" "$src" -lffi 2>&1; then + "$PTEST_DIR/$t" > "$PTEST_DIR/$t.out" 2>&1 + if ! LC_ALL=C uniq -u < "$PTEST_DIR/$t.out" | grep -q .; then + echo "PASS: libffi.bhaible/$t" + else + echo "FAIL: libffi.bhaible/$t" + cat "$PTEST_DIR/$t.out" + fi + else + echo "SKIP: libffi.bhaible/$t" + fi + rm -f "$PTEST_DIR/$t" "$PTEST_DIR/$t.out" +done + +# Individual test programs from upstream dejagnu suites +for suite in libffi.call libffi.closures libffi.complex libffi.go libffi.threads; do + dir="$PTEST_DIR/testsuite/$suite" + [ -d "$dir" ] || continue + + extra="" + [ "$suite" = "libffi.threads" ] && extra="-lpthread" + + for src in "$dir"/*.c "$dir"/*.cc; do + [ -f "$src" ] || continue + grep -q "dg-do run" "$src" || continue + + name=$(basename "${src%.*}") + + # complex_int uses non-standard _Complex int, fails on some arches (riscv64) + [ "$name" = "complex_int" ] && continue + case "$src" in + *.cc) cc="$CXX";; + *) cc="$CC";; + esac + + if $cc $CFLAGS -o "$PTEST_DIR/$name" "$src" -lffi $extra 2>&1; then + if "$PTEST_DIR/$name" 2>&1; then
meta/recipes-support/libffi/libffi/run-ptest line 62
Quoted 10 lines
meta/recipes-support/libffi/libffi/run-ptest new file +10
-0+ echo "PASS: $suite/$name" + else + echo "FAIL: $suite/$name" + fi + else + echo "SKIP: $suite/$name" + fi + rm -f "$PTEST_DIR/$name" + done +done
This is a _lot_ of logic that we have to maintain forever to run some integration tests…
Ross
PATCH v5 1/1 libffi: add ptest support
In reply to Pratik FarkaseAdd ptest support for libffi using all upstream test suites: bhaible, call, closures, complex, go, and threads.
Tests are cross-compiled at build time in do_compile_ptest and the binaries are installed to the ptest directory. No toolchain is needed on target.
The complex_int test is excluded as it uses non-standard _Complex int which is unsupported on some architectures (e.g. riscv64).
Tested on qemux86-64 with ptest-runner: 198 PASS, 0 FAIL (5s).
---
Changes in v5:
- Cross-compile tests at build time instead of on target
- Minimal run-ptest that just executes pre-compiled binaries
- Moved back to PTESTS_FAST (5s runtime vs 327s previously)
Changes in v4:
- Replace packagegroup-core-buildessential with gcc/g++/make to fix multilib symlinks conflict in lib64-core-image-sato-sdk
- Detect compiler dynamically in run-ptest (works without symlinks)
- Exclude complex_int test (fails on riscv64)
Changes in v3:
- Move libffi from PTESTS_FAST to PTESTS_SLOW (349s) runtime
Changes in v2 :
- Include all upstream test suites instead of only bhaible
- Install fficonfig.h required by ffitest.h
- Add support for C++ tests (closures/unwindtest*.cc)
- Add -lpthread for threads suite
--- .../distro/include/ptest-packagelists.inc | 1 + meta/recipes-support/libffi/libffi/run-ptest | 31 ++++++++++++++ meta/recipes-support/libffi/libffi_3.5.2.bb | 42 ++++++++++++++++++- 3 files changed, 73 insertions(+), 1 deletion(-) create mode 100755 meta/recipes-support/libffi/libffi/run-ptest
meta/conf/distro/include/ptest-packagelists.inc +1 -0
@@ -34,6 +34,7 @@ PTESTS_FAST = "\ libconfig \ libconvert-asn1-perl \ libexif \ + libffi \ libgpg-error\ libksba \ libmd \
meta/recipes-support/libffi/libffi/run-ptest new file +31 -0
@@ -0,0 +1,31 @@ +#!/bin/sh +# SPDX-License-Identifier: MIT + +cd "$(dirname "$0")/tests" || exit 1 + +for t in *; do + [ -x "$t" ] || continue + + suite="${t%%__*}" + name="${t#*__}" + + case "$suite" in + libffi.bhaible) + ./"$t" > "$t.out" 2>&1 + if ! LC_ALL=C uniq -u < "$t.out" | grep -q .; then + echo "PASS: $suite/$name" + else + echo "FAIL: $suite/$name" + cat "$t.out" + fi + rm -f "$t.out" + ;; + *) + if ./"$t" 2>&1; then + echo "PASS: $suite/$name" + else + echo "FAIL: $suite/$name" + fi + ;; + esac +done
meta/recipes-support/libffi/libffi_3.5.2.bb +41 -1
57 lines not on this page · raw message
Signature · 1 line
2.43.0