@@ -57,6 +57,19 @@ do_install_ptest() {
-i ${D}/${PTEST_PATH}/${TESTDIR}/Makefile
}
+# Tests that go too long without output under QEMU for the qemuriscv64 ptest
+# runs. strace is not machine-specific, so this also skips them on riscv64
+# hardware.
+STRACE_PTEST_SKIP ?= ""
+STRACE_PTEST_SKIP:riscv64 = "qual_fault-syscall"
+
+do_install_ptest:append() {
+ for t in ${STRACE_PTEST_SKIP}; do
+ sed -i -e '1a . "${srcdir=.}/init.sh"\nskip_ "Too slow under QEMU on riscv64"' \
+ ${D}${PTEST_PATH}/${TESTDIR}/$t.test
+ done
+}
+
RDEPENDS:${PN}-ptest += "make coreutils grep gawk sed locale-base-en-us"
BBCLASSEXTEND = "native"
On qemuriscv64, the strace ptest goes 283 s without output while qual_fault-syscall.test finishes. That is 63% of the 450 s ptest-runner limit, leaving little margin on a loaded autobuilder, so skip it there before strace is re-enabled for riscv64. The longest gaps between outputs in core-image-ptest-strace on qemuriscv64 (make -j4, every output line timestamped), with the test that ended each: - qual_fault-syscall.test (283 s) - filtering_syscall-syntax.test (139 s) - trace_creds.gen.test (90 s) - strace--tips-full.test (65 s) The rest are under 50 s. strace is not machine-specific, so the skip is keyed on riscv64, as the gnutls skips in cc848c403a are. It therefore also applies to riscv64 hardware, where the test passes (VisionFive 2). Signed-off-by: OOHehir <oo.hehir@gmail.com> --- meta/recipes-devtools/strace/strace_7.2.bb | 13 +++++++++++++ 1 file changed, 13 insertions(+)