On 02/21/2024 03:16 PM, Maxim Kuvyrkov wrote:
On Feb 21, 2024, at 05:46, Tiezhu Yang <yangtie...@loongson.cn> wrote:
On 02/21/2024 03:52 AM, ci_not...@linaro.org wrote:
If you can't get what you need from our CI within minutes, let us know and we
will be happy to help.
We can see "Operation not permitted" in the log info,
please try one of the following processes to test:
(1) set ptrace_scope as 0
$ echo 0 | sudo tee /proc/sys/kernel/yama/ptrace_scope
$ make check-gdb TESTS="gdb.threads/attach-many-short-lived-threads.exp"
Hi Tiezhu,
We already use the above approach for testing. Also, our CI reports only
regressions, not all failures, and the environment, generally, does not change
whether the test passes or fails.
The problem appears to be the fact that gdb.threads/attach-many-short-lived-threads.exp
tests are flaky, and detected as such in [1] -- search for "delete all
breakpoints".
However, because your patch renames the tests, the flaky entries do not match,
and failures are seen as regressions.
OK, I see. Are there any regressions tested with the following change
on top of the patch?
diff --git a/gdb/testsuite/lib/gdb.exp b/gdb/testsuite/lib/gdb.exp
index 7357d56f89a..7e14de44609 100644
--- a/gdb/testsuite/lib/gdb.exp
+++ b/gdb/testsuite/lib/gdb.exp
@@ -373,7 +373,7 @@ proc delete_breakpoints {} {
#
set timeout 100
- set msg "delete all breakpoints, watchpoints, tracepoints, and
catchpoints in delete_breakpoints"
+ set msg "delete all breakpoints in delete_breakpoints"
set deleted 0
gdb_test_multiple "delete breakpoints" "$msg" {
-re "Delete all breakpoints, watchpoints, tracepoints, and
catchpoints.*y or n.*$" {
If it is OK for you to avoid regressions, I will squash the above change
in the patch and then send a new version.
Thanks,
Tiezhu
_______________________________________________
linaro-toolchain mailing list -- linaro-toolchain@lists.linaro.org
To unsubscribe send an email to linaro-toolchain-le...@lists.linaro.org
