https://bugs.kde.org/show_bug.cgi?id=469146
Paul Floyd changed:
What|Removed |Added
Resolution|--- |FIXED
Status|REPORTED
https://bugs.kde.org/show_bug.cgi?id=469146
--- Comment #17 from Nick Nethercote ---
It has been a long time since I looked at this code. I don't have much to add
other than the changes seem plausible
--
You are receiving this mail because:
You are watching all bug changes.
https://bugs.kde.org/show_bug.cgi?id=469146
--- Comment #16 from Paul Floyd ---
> Both failures also occur in the mainline run. Your branch fixes the massif
> tests cleanly.
That's good.
I had a look at implementing a reset function for the "inlined IP cursor". It's
not too difficult but the
https://bugs.kde.org/show_bug.cgi?id=469146
--- Comment #15 from Carl Love ---
Paul,
I did the following tests on a Power 10 system with Fedora release 36.
I ran the mainline Valgrind with the 3.21.0 final commit at the top (commit
d97fed7c3e4aa7c910dfa0b6c5de12fd6cf08155 )
The results are:
https://bugs.kde.org/show_bug.cgi?id=469146
--- Comment #14 from Carl Love ---
per comment 10, 11
Paul, I see I cloned GDB not valgrind when trying to install your test tree.
Sigh, had spent all day working on a GDB patch and failed to completely switch
gears late in the day to look at this
https://bugs.kde.org/show_bug.cgi?id=469146
--- Comment #13 from Paul Floyd ---
I forgot, one thing I was thinking of was to try to speed up the allocs/frees
iipc = VG_(new_IIPC)(ep, ips[i]);
...
VG_(delete_IIPC)(iipc);
that are in the two stack loops. My plan is to split up VG_(new_IIPC) so
https://bugs.kde.org/show_bug.cgi?id=469146
--- Comment #12 from Paul Floyd ---
Nick could you take a look at this please and give me your opinion?
The main benefit I see is that it will make the --ignore-fn more consistent
with the massif.out. Specifically users will be able to copy and paste
https://bugs.kde.org/show_bug.cgi?id=469146
Paul Floyd changed:
What|Removed |Added
CC||n.netherc...@gmail.com
--
You are receiving this
https://bugs.kde.org/show_bug.cgi?id=469146
--- Comment #11 from Paul Floyd ---
git clone https://sourceware.org/git/valgrind.git && git checkout
users/paulf/try-massif_inline
should do the job
--
You are receiving this mail because:
You are watching all bug changes.
https://bugs.kde.org/show_bug.cgi?id=469146
--- Comment #10 from Carl Love ---
I cloned a test tree with:
git clone git://sourceware.org/git/binutils-gdb.git gdb-test-massif
I then looked for the remote
git fetch origin
git branch -v -a > junk
then looked in junk but don't see your tree?
Is
https://bugs.kde.org/show_bug.cgi?id=469146
--- Comment #9 from Paul Floyd ---
Carl, I've just pushed the patchset to a try branch
remotes/origin/users/paulf/try-massif_inline
I'm getting clean results for massif on gcc203, which is
paulf@gcc203:~/scratch/valgrind$ lscpu
Architecture:
https://bugs.kde.org/show_bug.cgi?id=469146
--- Comment #8 from Paul Floyd ---
+ for (i = 0; i < n_ips && !top_has_fnname; ++i) {
That needs to be "for (i = *top;"
+ iipc = VG_(new_IIPC)(ep, ips[i]);
+ do {
+ top_has_fnname = VG_(get_fnname_inl)(ep, ips[i], , iipc);
+
https://bugs.kde.org/show_bug.cgi?id=469146
Paul Floyd changed:
What|Removed |Added
Attachment #158542|0 |1
is obsolete|
https://bugs.kde.org/show_bug.cgi?id=469146
--- Comment #7 from Paul Floyd ---
(In reply to Paul Floyd from comment #6)
> Created attachment 158635 [details]
> Update patch to handle stacked and inlined functions
>
> Just the code fix. Will also need to update expecteds, maybe add a testcase
>
https://bugs.kde.org/show_bug.cgi?id=469146
--- Comment #6 from Paul Floyd ---
Created attachment 158635
--> https://bugs.kde.org/attachment.cgi?id=158635=edit
Update patch to handle stacked and inlined functions
Just the code fix. Will also need to update expecteds, maybe add a testcase and
https://bugs.kde.org/show_bug.cgi?id=469146
--- Comment #5 from Paul Floyd ---
My mistake was to replace loops that iterated over the callstack (but skipped
inlined functions) with a loops that iterate over inline functions only.
We need to do both.
--
You are receiving this mail because:
You
https://bugs.kde.org/show_bug.cgi?id=469146
--- Comment #4 from Paul Floyd ---
Bah. I still think I'm on the right lines for the cause. I was expecting the
extra fails.
I'll do some more debugging of new-cpp.
--
You are receiving this mail because:
You are watching all bug changes.
https://bugs.kde.org/show_bug.cgi?id=469146
--- Comment #3 from Carl Love ---
I tried down loading and applying the patch from Paul to the Valgrind mainline.
Head of the tree was:
commit d97fed7c3e4aa7c910dfa0b6c5de12fd6cf08155 (tag: VALGRIND_3_21_0,
origin/master, origin/HEAD)
Author: Mark
https://bugs.kde.org/show_bug.cgi?id=469146
--- Comment #2 from Carl Love ---
Per comment in the Valgrind emailI see what is happening now. The stack in
question is
---
==2756940==at 0x48A4C8C: malloc (vg_replace_malloc.c:431)
==2756940==by
https://bugs.kde.org/show_bug.cgi?id=469146
Carl Love changed:
What|Removed |Added
CC||c...@us.ibm.com
--
You are receiving this mail
https://bugs.kde.org/show_bug.cgi?id=469146
--- Comment #1 from Paul Floyd ---
Created attachment 158542
--> https://bugs.kde.org/attachment.cgi?id=158542=edit
Patch so that massif filter_IPs handles inline functions
Looks better now
Does affect a few other regtests, but they also look
21 matches
Mail list logo
