|
From: <oli...@on...> - 2004-09-03 13:48:08
|
I was running valgrind-2.2.0 with the memcheck tool. At a certain point of the application it does suddenly create multiple instances of the process (I had 8 processes of my application running out of nowhere) and it creashes with an error, that something si writing into unmapped space: ==20733== Process terminating with default action of signal 11 (SIGSEGV) ==20733== Access not within mapped region at address 0x2D42DFCF I haven't tried memcheck of 2.1.2 with this application. The addrcheck tool does work fine since 2.2.0 and doesn't crash on my application. |
|
From: Tom H. <th...@cy...> - 2004-09-03 14:24:10
|
In message <5074873$109...@co...>
oli...@on... wrote:
> I was running valgrind-2.2.0 with the memcheck tool. At a certain point
> of the application it does suddenly create multiple instances of the
> process (I had 8 processes of my application running out of nowhere)
> and it creashes with an error, that something si writing into unmapped
> space:
>
> ==20733== Process terminating with default action of signal 11 (SIGSEGV)
> ==20733== Access not within mapped region at address 0x2D42DFCF
What do you mean by "instances" exactly? How are you identifying
these instances?
The point is that valgrind won't randomly decide to start a few
extra copies your program. If your program forks then you will
appear to have multiple valgrind's running, one for each process
that you have forked.
Likewise if you are using ps then what you may be seeing is actually
threads rather than processes. You will always have two threads when
running under valgrind, and one more will appear each time you create
a thread in your program.
As to that error, that looks like a problem in your program. Valgrind
is simply informing you of the details when your programs dies because
of a bad memory access.
Tom
--
Tom Hughes (th...@cy...)
Software Engineer, Cyberscience Corporation
http://www.cyberscience.com/
|