From mboxrd@z Thu Jan 1 00:00:00 1970 Content-Type: multipart/mixed; boundary="===============4985509297221958071==" MIME-Version: 1.0 From: Denis Kenzior Subject: Re: iwd 1.17 test suite is failing Date: Mon, 23 Aug 2021 08:58:54 -0500 Message-ID: In-Reply-To: List-Id: To: iwd@lists.01.org --===============4985509297221958071== Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Hi Tomasz, On 8/22/21 1:55 PM, Tomasz K=C5=82oczko wrote: > Hi, > = > I think that it started around 1.15. > iwd and eel compiled with LTO. > = > [tkloczko(a)barrel iwd-1.17]$ make check > make --no-print-directory check-am > make --no-print-directory check-TESTS > PASS: unit/test-cmac-aes > PASS: unit/test-hmac-md5 > PASS: unit/test-hmac-sha1 > PASS: unit/test-hmac-sha256 > PASS: unit/test-prf-sha1 > PASS: unit/test-kdf-sha256 > PASS: unit/test-crypto > ./build-aux/test-driver: line 107: 3860965 Aborted =C2=A0 =C2=A0 =C2=A0 = =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 (core dumped) = > "$@" > $log_file 2>&1 > FAIL: unit/test-eapol There was indeed a failure that slipped through the cracks. This was due t= o the = test not being updated and is now fixed upstream with commit ed10b00afa3f = ("unit: Fix eapol IP Allocation test failure") > PASS: unit/test-mpdu > PASS: unit/test-ie > PASS: unit/test-util > PASS: unit/test-ssid-security > PASS: unit/test-arc4 > PASS: unit/test-wsc > PASS: unit/test-eap-mschapv2 > PASS: unit/test-eap-sim > /bin/sh ./libtool =C2=A0--tag=3DCC =C2=A0 --mode=3Dlink /usr/bin/gcc =C2= =A0-fvisibility=3Dhidden = > -DUNITDIR=3D\""./unit/"\" -DCERTDIR=3D\""./unit/"\" =C2=A0-O2 -g -grecord= -gcc-switches = > -pipe -Wall -Werror=3Dformat-security -Wp,-D_FORTIFY_SOURCE=3D2 = > -Wp,-D_GLIBCXX_ASSERTIONS -specs=3D/usr/lib/rpm/redhat/redhat-hardened-cc= 1 = > -fstack-protector-strong -specs=3D/usr/lib/rpm/redhat/redhat-annobin-cc1 = -m64 = > -mtune=3Dgeneric -fasynchronous-unwind-tables -fstack-clash-protection = > -fcf-protection -fdata-sections -ffunction-sections -flto=3Dauto = > -flto-partition=3Dnone -Os -Wl,-wrap,l_ecc_supported_ike_groups -Wl,-z,re= lro = > -Wl,--as-needed -Wl,--gc-sections -Wl,-z,now = > -specs=3D/usr/lib/rpm/redhat/redhat-hardened-ld -flto=3Dauto -flto-partit= ion=3Dnone = > -fuse-linker-plugin -o unit/test-sae unit/test-sae.o src/sae.o src/crypto= .o = > src/ie.o src/handshake.o src/erp.o src/util.o src/mpdu.o -lell > libtool: link: /usr/bin/gcc -fvisibility=3Dhidden -DUNITDIR=3D\"./unit/\" = > -DCERTDIR=3D\"./unit/\" -O2 -g -grecord-gcc-switches -pipe -Wall = > -Werror=3Dformat-security -Wp,-D_FORTIFY_SOURCE=3D2 -Wp,-D_GLIBCXX_ASSERT= IONS = > -specs=3D/usr/lib/rpm/redhat/redhat-hardened-cc1 -fstack-protector-strong = > -specs=3D/usr/lib/rpm/redhat/redhat-annobin-cc1 -m64 -mtune=3Dgeneric = > -fasynchronous-unwind-tables -fstack-clash-protection -fcf-protection = > -fdata-sections -ffunction-sections -flto=3Dauto -flto-partition=3Dnone -= Os = > -Wl,-wrap -Wl,l_ecc_supported_ike_groups -Wl,-z -Wl,relro -Wl,--as-needed = > -Wl,--gc-sections -Wl,-z -Wl,now -specs=3D/usr/lib/rpm/redhat/redhat-hard= ened-ld = > -flto=3Dauto -flto-partition=3Dnone -fuse-linker-plugin -o unit/test-sae = > unit/test-sae.o src/sae.o src/crypto.o src/ie.o src/handshake.o src/erp.o = > src/util.o src/mpdu.o =C2=A0-lell > /usr/bin/ld: /tmp/cc8uiWzL.lto.o: in function `sae_choose_next_group': > /home/tkloczko/rpmbuild/BUILD/iwd-1.17/src/sae.c:139: undefined reference= to = > `__wrap_l_ecc_supported_ike_groups' > /usr/bin/ld: /tmp/cc8uiWzL.lto.o: in function `handshake_state_free': > /home/tkloczko/rpmbuild/BUILD/iwd-1.17/src/handshake.c:48: undefined refe= rence = > to `__wrap_l_ecc_supported_ike_groups' > /usr/bin/ld: /tmp/cc8uiWzL.lto.o: in function `sae_rx_authenticate': > /home/tkloczko/rpmbuild/BUILD/iwd-1.17/src/sae.c:161: undefined reference= to = > `__wrap_l_ecc_supported_ike_groups' This would have shown up earliest at iwd 1.16. However, I'm not sure what = this = is about? We use unit_test_sae_LDFLAGS =3D -Wl,-wrap,l_ecc_supported_ike_groups Perhaps LTO somehow interferes with that? ell has a unit test that uses th= e = same mechanism for quite some time... > = > BTW is it possible to organize iwd/eel mirrors on github/gitlab to be abl= e = > submit such an issue without being forced to subscribe to yet another mai= ling = > list or put in README files details how to submit the issue? > = We'll look into it. Regards, -Denis --===============4985509297221958071==--