This is due to a dynamically loaded library that has been closed. You want to set the 'TOPONODLCLOSE' environment variable to avoid this in this particular situation.
- Eric On Thu, Jul 17, 2008 at 01:47:03PM -0700, Scott Davenport wrote: > Hello, > > I'm trying to trace down a memory leak, and ::findleaks -d > isn't decoding all of the symbols. Ex: > > # UMEM_DEBUG=default /usr/lib/fm/fmd/fmtopo -C > /dev/null > ... > > ::findleaks -v > ... > umem_alloc_16 leak: 2 buffers, 16 bytes each, 32 bytes total > ADDR BUFADDR TIMESTAMP THREAD > ^[[4m CACHE LASTLOG > CONTENTS^[[m > ^[[1m 2d6c30^[[m 2cfaa0 b85d1620cf > 1 > 82008 0 0 > libumem.so.1`umem_cache_alloc+0x148 > libumem.so.1`umem_alloc+0xa8 > 0xfe8633d0 > 0xfe863a8c > 0xfe861bb0 > libtopo.so.1`rtld_init+0x160 > libtopo.so.1`topo_mod_start+0x40 > libtopo.so.1`topo_mod_create+0xcc > libtopo.so.1`topo_modhash_load+0x20 > libtopo.so.1`topo_mod_load+0xc0 > libtopo.so.1`mem_enum+0x28 > libtopo.so.1`topo_mod_enumerate+0xac > libtopo.so.1`topo_builtin_create+0xa0 > libtopo.so.1`topo_open+0x384 > main+0x238 > > Is this just an instance of "6398977 findleaks not resolving symbol > names"? Or are there othe > > Meanwhile starting to sift through nm output. > > Thanks > -- scott > http://blogs.sun.com/sdaven > _______________________________________________ > mdb-discuss mailing list > mdb-discuss at opensolaris.org -- Eric Schrock, Fishworks http://blogs.sun.com/eschrock
