This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [PATCH 13/17] btrace: non-stop
- From: Pedro Alves <palves at redhat dot com>
- To: "Metzger, Markus T" <markus dot t dot metzger at intel dot com>
- Cc: "gdb-patches at sourceware dot org" <gdb-patches at sourceware dot org>
- Date: Thu, 10 Sep 2015 12:05:01 +0100
- Subject: Re: [PATCH 13/17] btrace: non-stop
- Authentication-results: sourceware.org; auth=none
- References: <1441794909-32718-1-git-send-email-markus dot t dot metzger at intel dot com> <1441794909-32718-14-git-send-email-markus dot t dot metzger at intel dot com> <55F01DEC dot 4030209 at redhat dot com> <A78C989F6D9628469189715575E55B23331AB7A1 at IRSMSX104 dot ger dot corp dot intel dot com> <55F03852 dot 7030200 at redhat dot com> <A78C989F6D9628469189715575E55B23331AB873 at IRSMSX104 dot ger dot corp dot intel dot com> <55F04765 dot 9020206 at redhat dot com> <A78C989F6D9628469189715575E55B23331ACADF at IRSMSX104 dot ger dot corp dot intel dot com>
On 09/10/2015 08:47 AM, Metzger, Markus T wrote:
> That seems to be the case. I'm still getting the error I'm checking for before
> the prompt, though, and my test will run into a timeout.
>
> thread apply all reverse-stepi 4 &
>
> Thread 2 (Thread 0x7ffff74fb700 (LWP 70895)):
>
> Thread 1 (Thread 0x7ffff7fcc740 (LWP 70891)):
> Cannot execute this command while the selected thread is running.
> (gdb) PASS: gdb.btrace/non-stop.exp: reverse-step: all: thread apply all reverse-stepi 4: prompt
> 0x0000000000400671 28 for (; i < 10; ++i) global += i; /* loop */
> PASS: gdb.btrace/non-stop.exp: reverse-step: all: thread apply all reverse-stepi 4: thread 0
> FAIL: gdb.btrace/non-stop.exp: reverse-step: all: thread apply all reverse-stepi 4: thread 1 (timeout)
>
> A failing run might take a bit longer, but that should be it.
Odd. I'm running the test now for over 20 minutes, and it doesn't ever fail.
Before I ran it against gdbserver for 10 minutes, never failed.
This is with an i7-2620M; I hacked linux-btrace.c:intel_supports_bts to
enable btrace.
If you put an "info threads" just before that reverse-stepi, does it
show thread 1 as "(running)"? I wonder if the problem happened
sooner / above in the log, and you're seeing a consequence.
A "set debug infrun 1" log will likely help.
>
> I changed the test, removed the "Cannot execute ..." check, and pushed the
> series into users/mmetzger/btrace-non-stop.
This looks good to me. The comments above gdb_cont_to are stale though.
Thanks,
Pedro Alves