NDK23c does not provide any libgcc libraries. This is reasonable, since it also
does not provide a gcc: it uses clang instead. I’ve never tried to build
Valgrind for Android, but hopefully someone else can tell you how to do it with
a modern NDK.
--
John Dallman
From: $rik@nth
Sent: 23 June 2
> However, I've also been told that the g++ compiler will initialize all stack
> objects to zero when compiling for debug (the -g option). Yet, valgrind
> still detects the un-init condition.
I think whoever told you that was confusing it with Microsoft Visual Studio.
The default debug-build c
The memory requirement for Valgrind is “several times as much memory as the
application you’re testing needs without Valgrind.”
How much memory does your application use normally?
--
John Dallman
From: Padala Dileep [mailto:padala.dil...@gmail.com]
Sent: 25 January 2019 10:00
To: valgrind-users
> As a Valgrind user, I see few ways to get different results with valgrind :
Another possibility is that the OP has a memory management bug, and the
inevitably slightly different memory layout when running under Valgrind can
trigger it. I have not seen this with Valgrind, but I have seen it whe
man
Siemens Industry Sector
Siemens Industry Software Limited
Francis House, 112 Hills Road,
Cambridge CB2 1PH, United Kingdom
Tel. :+44 (1223) 371554
Fax :+44 (1223) 371700
john.dall...@siemens.com
www.siemens.com/plm
-Original Message-
From: FEVOTTE Francois [mailto:francois.fe
use, 112 Hills Road,
Cambridge CB2 1PH, United Kingdom
Tel. :+44 (1223) 371554
Fax :+44 (1223) 371700
john.dall...@siemens.com <mailto:john.dall...@siemens.com>
www.siemens.com/plm
From: [ext] Dallman, John [mailto:john.dall...@siemens.com]
Sent: Thursday, June 01, 2017 12:29 PM
To: &#
This is potentially interesting for what I do. Is the documentation at
http://edf-hpc.github.io/verrou/vr-manual.html up to date?
Something that would be very useful would be control of the seed of the random
number generator, so that we could repeat and debug cases that gave strange
results.
> Whilst 3.12.0 continues to support the 32-bit x86 instruction set, we would
> prefer users to migrate to 64-bit x86 (a.k.a amd64 or x86_64) where possible.
Sound move. I was able to give up shipping 32-bit x86 Mac and Linux software
earlier this year, and our customers tend to be quite conservat
I'm starting to look at fuzz testing the mathematical modelling library I work
on, which reads complicated data files that are produced by end-users, and
could plausibly be used to stage buffer overflow attacks. The basics obviously
come first: use -fstack-protector, take care with string manipu
> I've got a program which seems to cause heap corruption so I thought valgrind
> would be
> able to tell me where this occurs. The trouble is the program uses data files
> from the
> running directory for configuration purposes and running the program under
> valgrind
> seems to prevent it find
Julian Seward wrote:
> One other option is to play around with optimisation settings for your app,
> to see if it changes g++'s inline/no-inline decisions.
g++'s -fno-inline option may help, although it has the same limitation of not
applying to libraries.
--
John Dallman
-
Sie
> I have a lot of calls to functions with names that I recognise,
> but which have a '2 appended to their names. These don't exist
> in the source: any idea what they signify?
They're in the callgrind documentation; they're about recursion, and
I clearly need to accumulate their call counts into t
Philippe Waroquiers wrote:
> Starting callgrind with default options, and then using
> kcachegrind on the resulting callgrind.out file, kcachegrind
> shows the nr of calls to the functions.
I'm trying to avoid using kcachegrind, because I need to automate
this process. We have tens of thousands
Julian Seward wrote:
> I suspect not. In any case counting insns is not terribly expensive,
> so the gain you'd get would be modest.
OK.
> I suspect the names are present in the callgrind.out file, and so
> callgrind_annotate just slices out those that you list. It doesn't
> itself know the na
I'm looking at using callgrind to replace an expensive Windows coverage tool,
and for some other work where the ideas are only part-formed.
The default callgrind run and callgrind_annonate display shows Ir events, but
I only need to count function entries. Is there a way to only collect those
and
> Do any common platforms, other than x86/x86_64, offer more-than-64-bit "long
> double"?
Not that they support as full speed hardware operations, AFAIK. SPARC has
defined
registers and instructions for 128-bit floating point, but implements them as
sequences of operations on 64-bit floats, so t
Julian Seward [mailto:jsew...@acm.org] wrote:
> Even then it's not simple. The x87 control word has a bit-pair that
> controls the default x87 FP precision. Currently V ignores all attempts to
> change it, and just does its thing at 64 bits. If 80-bit arithmetic
> becomes supported, and V ignor
> Because the recompilation time of one patched php is nearly 20 seconds, so if
> the
> sum of degradation time is more than 20 seconds, then our method of
> indirecting
> function will perorm worse than native regression
> testing.
You must expect performance under Valgrind to be worse than an
John Reiser [mailto:jrei...@bitwagon.com] wrote:
> > I'm having an issue with valgrind - mine regression takes 4 hours instead
> > of 45 mins when run in valgrind.
> An elapsed time ratio of 240:45 (5.3 : 1) is not unusual, and may be better
> than average.
That's definitely better than I see: I'
> there is a multicore platform named tile64 http://en.wikipedia.org/wiki/TILE64
> Now, I want to let valgrind execute onto this platform, does anyone know the
> exact way?
TILE64 uses a MIPS-based instruction set. The platforms page on the Valgrind
website,
http://valgrind.org/info/platforms.ht
Benjamin Schindler [mailto:bschind...@inf.ethz.ch] wrote:
> ... the backtrace size used by valgrind is too small... i.e. it
> just shows 10 functions. Is there a way to see more than just that
Use the --num-callers= command line switch.
best,
--
John Dallman
Parasolid Porting Engineer
-O
> It appears that everything used in that line is set. Might
> valgrind not recognize the values set to idx and length in
> the call to sscanf()?
With puzzling things like this, try simply printing the values in
question with printf(). This usually clarifies the situation.
--
John Dallman
F
Use "--num-callers=25", or any other number. This is in valgrind --help,
but isn't specific to memcheck.
best,
--
John Dallman
Parasolid Porting Engineer
Siemens Product Lifecycle Management Software
Industry Sector
46 Regent Street, Cambridge, CB2 1DP
United Kingdom
Tel: +44-1223-371554
john.d
I don't have any trouble with FP math on 32-bit x86, but I take
a different strategy from you.
I set the x87 FP unit to use 64-bit doubles rather than 80-bit
doubles, since that makes the results far more similar to 64-bit
Linux, and other platforms that use 64-bit doubles. 32-bit Linux
uses 80
ndustry Sector
46 Regent Street, Cambridge, CB2 1DP
United Kingdom
Tel: +44-1223-371554
john.dall...@siemens.com
www.siemens.com/plm
> -Original Message-
> From: Ildar Isaev [mailto:iis...@ispras.ru]
> Sent: Wednesday, May 26, 2010 4:08 PM
> To: Dallman, John
> Cc: valgrin
> So if one wants to find a bug with Avalanche, one should be
> better take shorter files. There's just more chance to detect
> anything.
This is fine with some kinds of data. One can make a smaller bitmap,
or a shorter sound clip. But with what I do - accurate 3D shape
representation - one can'
I'm not clear where the restriction of files to 712 bytes comes from;
is that an arbitrary limit to ensure that analysis takes a sane length
of time?
thanks,
--
John Dallman
Parasolid Porting Engineer
Siemens Product Lifecycle Management Software
Industry Sector
46 Regent Street, Cambridge, C
Julian Seward wrote:
> It sounds interesting. I would like to read more about it and
> perhaps try it out, to get some idea of its effectiveness on
> large programs (ability to find bugs, false error rate, speed
> and memory use).
Same here. I have a fairly basic question: in what terms does
Av
We have adopted Valgrind as a periodic QA measure: every six months,
we run all our tests through it. This takes a couple of months for
everything to run, so we don't feel we can do it more often. What
it finds for us is uninitialized variable errors; our internal memory
management takes big chu
> Now back to my problem. The original code has double instead of int (I
> just tried both types to be sure and forgot to change back) and on
MSVC
> std::swap() of uninitialised doubles triggers a floating point
> exception (if enabled) as MSVC uses the FPU.
>
> I was really surprised that the cod
Julian Seward [mailto:jsew...@acm.org] wrote:
> Harmless .. ignore it.
OK, and thanks.
--
John Dallman
Parasolid Porting Engineer
--
Open Source Business Conference (OSBC), March 24-25, 2009, San Francisco, CA
-OSBC
> They indicate that valgrind's memory manager was asked to
> handle a block (actually a single contiguous address
> interval, which could be a block or a non-block) of more
> than 100,000,000 bytes ... on x86_64 with 64-bit addresses
> and much more latitude for placement of PT_LOAD, this
> m
I'm running Valgrind-3.3.1 on x86-64 on SLES10sp1, memchecking some C++
code
with which I'm not very familiar. I get these messages:
==4213== Warning: set address range perms: large range 102130360
(undefined)
==4213== Warning: set address range perms: large range 102130392
(noaccess)
I can't fin
> 1. How can I find out how much memory is used in each object in firefox?
> 2. How can I find out what are the functions spent the most time during my
> program execution?
Valgrind is not an all-purpose tool. It does not do either of those things.
==2616== Thread 4:
==2616== Invalid read of si
> You are right, I tried to recompile everything and the compilation of
> standard c++ libraries etc. fail. Sorry for wasting your time I guess
> :(
If you want to persist with this idea, don't let me stop you from
starting
work on making everything capable of building with packed structures. It
> IMHO this view is too negative. The proper approach for ensuring that
> a program does not contain data races is to choose a strategy at
> design time that guarantees that a program is data-race free (e.g.
> protecting all shared data through locking and documenting which
> synchronization object
> Oh yes, I forgot about those :) You have a good point, but still, the
> effort for recompiling the libraries and other required code for this
> purpose might be much less than trying to find an uninitialized
> variable in some cumbersome code.
It might. But it seems unlikely. To the best of my k
Tom Hughes [mailto:t...@compton.nu]
> eyurt...@abo.fi wrote:
> > However it is possible to pack the structs without holes...
> You're going to recompile glibc and all the other libraries that your
> program uses with that option are you?
And, of course, if you don't, structs passed between packed
> However it is possible to pack the structs without holes...
You will not like what this does to performance. It also means
that you will be Valgrinding a different build of your software
from the one that gets run in anger.
--
John Dallman
Parasolid Porting Engineer
39 matches
Mail list logo
