[petsc-users] valgrind reports with --leak-check=full

[Dominik Szczerba]

I am getting the following reports running with --leak-check=full
(which could be mentioned somewhere in FAQ). Is this something to
worry about?

Thanks
Dominik


==10052== 300 (60 direct, 240 indirect) bytes in 1 blocks are
definitely lost in loss record 14 of 15
==10052==    at 0x4C28F9F: malloc (vg_replace_malloc.c:236)
==10052==    by 0x67AB855: nss_parse_service_list (nsswitch.c:626)
==10052==    by 0x67ABE39: __nss_database_lookup (nsswitch.c:167)
==10052==    by 0x74B6823: ???
==10052==    by 0x67648AC: getpwuid_r@@GLIBC_2.2.5 (getXXbyYY_r.c:256)
==10052==    by 0x67641A2: getpwuid (getXXbyYY.c:117)
==10052==    by 0x99BE9D: PetscGetUserName (fuser.c:66)
==10052==    by 0x5D2A20: PetscErrorPrintfInitialize (errtrace.c:68)
==10052==    by 0x5F1057: PetscInitialize (pinit.c:697)

==10051== 8 bytes in 1 blocks are definitely lost in loss record 2 of 15
==10051==    at 0x4C28F9F: malloc (vg_replace_malloc.c:236)
==10051==    by 0x1296EB9: GKmalloc__ (util.c:151)
==10051==    by 0x1296E00: fmalloc__ (util.c:112)
==10051==    by 0x1295CBC: CheckInputs__ (weird.c:188)
==10051==    by 0x126F8B8: ParMETIS_V3_PartKway (kmetis.c:56)
==10051==    by 0xD26874: MatPartitioningApply_Parmetis (pmetis.c:96)
==10051==    by 0x6DDCFC: MatPartitioningApply (partition.c:226)