[Valgrind-users] Does anyone use valgrind work on a dpdk code with huge page?

[Datong Li <osd...@gm...>]

When I use valgrind in my program, which use dpdk (a intel library) and
open a hugepage to allocate memory, it arises a error when I run it like
below:

PANIC in rte_eal_init():
Cannot init memory
7: [./output/bdsec_ids() [0x41ebe1]]
6: [/lib64/libc.so.6(__libc_start_main+0xfd) [0x318ae1ecdd]]
5: [./output/bdsec_ids(main_run+0x182) [0x41ee22]]
4: [./output/bdsec_ids(pal_init+0x1e6) [0x428626]]
3: [./output/bdsec_ids(rte_eal_init+0x1410) [0x493bc0]]
2: [./output/bdsec_ids(__rte_panic+0xc4) [0x41eb24]]
1: [./output/bdsec_ids(rte_dump_stack+0x1e) [0x498aee]]


What I found in system 's log :

Detected 12 lcore(s)
EAL: Setting up memory...
EAL: map_all_hugepages(): mmap failed: Invalid argument
EAL: Failed to mmap 1024 MB hugepages


So I guess this error maybe cause by hugepage when use valgrind. Does
anybody know something about this?

With Best wishes,
Datong Li