gdb.threads/attach-many-short-lived-threads.exp FAIL

gdb.threads/attach-many-short-lived-threads.exp FAIL

[Carl Love]

Luis, Guinevere:

I am also seeing a recent increase in the attach-many-short-lived-
threads.exp test failures on both Power 9 and 10 buildbot results.  

I have seen a little "noise" in the past with a threads or signal test
occasionally failing and then being fine on the next run.  I don't
remember this particular test being noisy in the past.  But this test
seems to have recently gotten really "noisy".

I saw that you both briefly discussed the failures on IRC.
Here is what I am seeing for the attach-many-short-lived-threads.exp
test failures.  The issues started on Friday Nov 31.

               	Power 9          	Power 10
11/31/2023	39 new fail, 1 fixed	0 new, 0 fixed
12/1/2023	0 new, 39 fixed		50 new, 0 fixed
12/2/2023	52 new, 0 fixed		87 new, 40 fixed
12/3/2023	checkout fail		12 new, 6 fixed
	
I ran the full regression tests manually today.  It the attach-many-
short-lived-threads.exp test ran fine on Power 9 and failed 103 times
on Power 10.

I took a quick look at the git log and nothing jumped out as an obvious
culprit.  It sounded like Luis was going to look into the test.  I
thought I would start by reaching out to see what the status is?  Do we
have a bugzilla for the issue?

                        Carl 

Re: gdb.threads/attach-many-short-lived-threads.exp FAIL

[Luis Machado]

On 12/5/23 18:01, Carl Love wrote:
> Luis, Guinevere:
> 
> I am also seeing a recent increase in the attach-many-short-lived-
> threads.exp test failures on both Power 9 and 10 buildbot results.  
> 
> I have seen a little "noise" in the past with a threads or signal test
> occasionally failing and then being fine on the next run.  I don't
> remember this particular test being noisy in the past.  But this test
> seems to have recently gotten really "noisy".
> 
> I saw that you both briefly discussed the failures on IRC.
> Here is what I am seeing for the attach-many-short-lived-threads.exp
> test failures.  The issues started on Friday Nov 31.
> 
>                	Power 9          	Power 10
> 11/31/2023	39 new fail, 1 fixed	0 new, 0 fixed
> 12/1/2023	0 new, 39 fixed		50 new, 0 fixed
> 12/2/2023	52 new, 0 fixed		87 new, 40 fixed
> 12/3/2023	checkout fail		12 new, 6 fixed
> 	
> I ran the full regression tests manually today.  It the attach-many-
> short-lived-threads.exp test ran fine on Power 9 and failed 103 times
> on Power 10.
> 
> I took a quick look at the git log and nothing jumped out as an obvious
> culprit.  It sounded like Luis was going to look into the test.  I
> thought I would start by reaching out to see what the status is?  Do we
> have a bugzilla for the issue?
> 
>                         Carl 
> 

I bisected these new FAIL's to a particular patch, and replied to this thread:

https://inbox.sourceware.org/gdb-patches/a6be03ef-9337-4693-a835-99ae5a0bc16e@arm.com/