On Thu, 2012-03-29 at 16:22 +0800, 田东云 wrote:
>
> (gdb) target remote | /usr/local/valgrind/lib/valgrind/../../bin/vgdb
> --pid=10870
> Remote debugging using | /usr/local/valgrind/lib/valgrind/../../bin/vgdb
> --pid=10870
> vgdb error: no FIFO found matching pid 10870
> Remote communication error: Resource temporarily unavailable.
The above is strange. Such behaviour has been seen if the temp
directory used by the Valgrind process and the gdb process
are not the same.
Can you do the following trial :
launch the valgrind on the java jvm (with the -v arg).
while the above runs, type the following command:
vgdb -d -d -d help
vgdb will report which FIFO files it is trying, and one of
these should match the one reported by the valgrind -v java.
Apart of this, if the only thing you want to do is to ask for
a leak report when java jvm is running, you do not (necessarily) need
gdb.
You can use vgdb from shell command line, typing e.g.
vgdb leak_check full reachable any
to trigger a memory leak search showing the full set of
malloc-ed blocks allocated.
Note however that I do not know if the java jvm executable is using
the malloc library. If java executable has its own heap management
build e.g. on top of mmap, then there is not much chance
memcheck will report anything intelligent.
Philippe
------------------------------------------------------------------------------
This SF email is sponsosred by:
Try Windows Azure free for 90 days Click Here
http://p.sf.net/sfu/sfd2d-msazure
_______________________________________________
Valgrind-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/valgrind-users
