https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=273956
--- Comment #18 from Ed Maste ---
(In reply to Paul Floyd from comment #17)
Thank you for keeping Valgrind on FreeBSD in good shape!
--
You are receiving this mail because:
You are the assignee for the bug.
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=273956
Paul Floyd changed:
What|Removed |Added
Resolution|--- |Overcome By Events
Status
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=273956
--- Comment #16 from Paul Floyd ---
(In reply to Ed Maste from comment #15)
I can't think of anything that would have changed the size of the RW PT_LOAD,
though it might have been a small change that tipped the scales. I can now
reproduce
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=273956
--- Comment #15 from Ed Maste ---
(In reply to Paul Floyd from comment #14)
> How have things changed between 13.2 and 14.0 to cause this?
I don't see anything kernel related that's different between 13.2 and 14.0. Is
it possible that we h
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=273956
--- Comment #14 from Paul Floyd ---
There are 3 ways that Valgrind triggers reading debuginfo
1. from its own image as here
2. from the guest ld-elf.so.1 and exe which Valgrind loads
3. from anything that the guest loads once it is running
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=273956
--- Comment #13 from Ed Maste ---
(In reply to Konstantin Belousov from comment #12)
The issue is just in Valgrind's assumptions of what the mappings will be.
I spotted this comment in its src:
/* A segment in the initial /proc/self/map
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=273956
Konstantin Belousov changed:
What|Removed |Added
CC||k...@freebsd.org
--- Comment
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=273956
Ed Maste changed:
What|Removed |Added
Attachment #245082|text/x-csrc |text/plain
mime type|
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=273956
--- Comment #10 from Ed Maste ---
Interesting - I made a little program that uses KERN_PROC_VMMAP and dumps its
own map. When I run it normally I see the 4th PT_LOAD as swap:
$ ./a.out
1 0x0020 - 0x00201000 (0x1000
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=273956
--- Comment #9 from Paul Floyd ---
And for reference, if I trigger an assert inside a Valgrind built on Alpine I
get
host stacktrace:
==10726==at 0x5804379A: show_sched_status_wrk (m_libcassert.c:407)
==10726==by 0x580438B7: report
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=273956
--- Comment #8 from Paul Floyd ---
(In reply to Ed Maste from comment #6)
I know a thing or two about this.
Initially Valgrind would only trigger reading debuginfo when it considered a
file to have finished mapping. That used to mean seei
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=273956
--- Comment #7 from Paul Floyd ---
(In reply to Ed Maste from comment #5)
Ah so the RW PT_LOAD section is starting out swapped so the sysctl doesn't see
the file.
--
You are receiving this mail because:
You are the assignee for the bug.
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=273956
--- Comment #6 from Ed Maste ---
On 13.2 for /bin/sleep I see:
PID STARTEND PRT RES PRES REF SHD FLAG TP PATH
1120 0x2a57e7f66000 0x2a57e7f67000 r--12 3 1 CN--- vn
/bin/sleep
1120
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=273956
--- Comment #5 from Ed Maste ---
Ok, I can stop it there:
(lldb) b parse_procselfmaps
Breakpoint 1: where = memcheck-amd64-freebsd`parse_procselfmaps + 37 at
aspacemgr-linux.c:3926:12, address = 0x38133885
(lldb) run
Process 30476
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=273956
Graham Perrin changed:
What|Removed |Added
CC||grahamper...@gmail.com
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=273956
--- Comment #4 from Ed Maste ---
Hi Paul, I'm using valgrind-3.21.0,1 from pkg, and a slightly older CURRENT
around the time of the stable/14 branch. I will update one or both if necessary
to reproduce, but want to know what steps I should
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=273956
--- Comment #3 from Paul Floyd ---
Hi Ed
Which versions of FreeBSD and Valgrind are you using?
I was running my dev buuild of Valgrind under gdb when I ran procstat.
The easiest thing is to use a dummy exe that just does a big sleep() an
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=273956
Ed Maste changed:
What|Removed |Added
CC||ema...@freebsd.org
--- Comment #2 from
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=273956
--- Comment #1 from Paul Floyd ---
Will try soon-ish 14.0-BETA2
--
You are receiving this mail because:
You are the assignee for the bug.
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=273956
Bug ID: 273956
Summary: KERN_PROC_VMMAP does not read all mapped files
Product: Base System
Version: 14.0-CURRENT
Hardware: Any
OS: Any
Status: New
20 matches
Mail list logo
