This is the mail archive of the
systemtap@sourceware.org
mailing list for the systemtap project.
Re: Recent ARM systemtap test results using a kernel with uprobe support
- From: David Long <dave dot long at linaro dot org>
- To: William Cohen <wcohen at redhat dot com>
- Cc: systemtap at sourceware dot org, Naresh Kamboju <naresh dot kamboju at linaro dot org>
- Date: Wed, 30 Oct 2013 14:53:18 -0400
- Subject: Re: Recent ARM systemtap test results using a kernel with uprobe support
- Authentication-results: sourceware.org; auth=none
- References: <52713A6D dot 2050208 at redhat dot com>
On 10/30/13 12:57, William Cohen wrote:
Hi All,
I attempted to take a look over the recent arm test results on a kernel with
uprobes supoort:
http://web.elastic.org/~dejazilla/viewsummary.php?summary=%3D%27%3C5271348F.8020508%40redhat.com%3E%27
A lot of the tests passed as seen in the summary.
=== systemtap Summary ===
# of expected passes 3106
# of unexpected failures 113
# of expected failures 283
# of known failures 37
# of untested testcases 174
# of unsupported tests 5
Below is a breakdown of of how things are failing:
When there are multiple uprobes on an executable the userspace application gets a sigsegv. Have talked with David Long about this and have a replicator:
Thanks for the info.
It looks like I have a bad merge with recent-(ish)
arch_uretprobe_hijack_return_address() changes, which is cause uretprobe
to fail. This can be seen in the system log during the at_var test.
What I thought was the obvious fix has not immediately solved the
problem, but I am working on it. I don't know if a fix to this will
solve the sigv problem or not but it would be interesting to see if it
fixes other test failures.
-dl