ch3: valgrind "Syscall param writev points to uninitialised byte(s)"

[jczhang07]

$ /scratch/jczhang/petsc/arch-kokkos-dbg/bin/mpichversion 
MPICH Version:      4.1.2
MPICH Release date: Wed Jun  7 15:22:45 CDT 2023
MPICH ABI:          15:1:3
MPICH Device:       ch3:sock
MPICH configure:    --prefix=/scratch/jczhang/petsc/arch-kokkos-dbg MAKE=/usr/bin/gmake --libdir=/scratch/jczhang/petsc/arch-kokkos-dbg/lib CC=gcc CFLAGS=-fPIC -Wno-lto-type-mismatch -Wno-stringop-overflow -g -O0 AR=/usr/bin/ar ARFLAGS=cr CXX=g++ CXXFLAGS=-Wno-lto-type-mismatch -Wno-psabi -g -O0 -std=gnu++20 -fPIC FFLAGS=-fPIC -ffree-line-length-none -ffree-line-length-0 -Wno-lto-type-mismatch -g -O0 -fallow-argument-mismatch FC=gfortran F77=gfortran FCFLAGS=-fPIC -ffree-line-length-none -ffree-line-length-0 -Wno-lto-type-mismatch -g -O0 -fallow-argument-mismatch --enable-shared --with-pm=hydra --disable-java --with-hwloc=embedded --enable-fast=no --enable-error-messages=all --with-device=ch3:sock --enable-g=meminit,dbg PYTHON=/usr/bin/python3 --disable-maintainer-mode --disable-dependency-tracking
MPICH CC:           gcc -fPIC -Wno-lto-type-mismatch -Wno-stringop-overflow -g -O0   -O0
MPICH CXX:          g++ -Wno-lto-type-mismatch -Wno-psabi -g -O0 -std=gnu++20 -fPIC  -O0
MPICH F77:          gfortran -fPIC -ffree-line-length-none -ffree-line-length-0 -Wno-lto-type-mismatch -g -O0 -fallow-argument-mismatch  -O0
MPICH FC:           gfortran -fPIC -ffree-line-length-none -ffree-line-length-0 -Wno-lto-type-mismatch -g -O0 -fallow-argument-mismatch  -O0

Use this mpich to compile the MPI_Bcast example at https://rookiehpc.org/mpi/docs/mpi_bcast/index.html, then run it with valgrind

$ /scratch/jczhang/petsc/arch-kokkos-dbg/bin/mpirun  -n 2 valgrind ./ex100
==1866237== Memcheck, a memory error detector
==1866237== Copyright (C) 2002-2022, and GNU GPL'd, by Julian Seward et al.
==1866237== Using Valgrind-3.22.0 and LibVEX; rerun with -h for copyright info
==1866237== Command: ./ex100
==1866237== 
==1866238== Memcheck, a memory error detector
==1866238== Copyright (C) 2002-2022, and GNU GPL'd, by Julian Seward et al.
==1866238== Using Valgrind-3.22.0 and LibVEX; rerun with -h for copyright info
==1866238== Command: ./ex100
==1866238== 
[MPI process 0] I am the broadcast root, and send value 12345.
==1866237== Syscall param writev(vector[0]) points to uninitialised byte(s)
==1866237==    at 0x5191867: writev (writev.c:26)
==1866237==    by 0x4DA2EA8: MPL_large_writev (mpl_sock.c:31)
==1866237==    by 0x4D75A8B: MPIDI_CH3I_Socki_handle_write (sock.c:3688)
==1866237==    by 0x4D74834: MPIDI_CH3I_Sock_wait (sock.c:3367)
==1866237==    by 0x4D7D09B: MPIDI_CH3i_Progress_wait (ch3_progress.c:173)
==1866237==    by 0x4D7F248: MPIDI_CH3I_Progress (ch3_progress.c:808)
==1866237==    by 0x4C30234: MPIR_Wait_state (request_impl.c:885)
==1866237==    by 0x4C3036F: MPIR_Wait_impl (request_impl.c:908)
==1866237==    by 0x4BAC801: MPID_Wait (mpidpost.h:273)
==1866237==    by 0x4BACA58: MPIC_Wait (helper_fns.c:63)
==1866237==    by 0x4BACD35: MPIC_Send (helper_fns.c:130)
==1866237==    by 0x4ABC5D5: MPIR_Bcast_intra_binomial (bcast_intra_binomial.c:146)
==1866237==  Address 0x5b5cd6c is 60 bytes inside a block of size 176 alloc'd
==1866237==    at 0x4848889: malloc (in /home/jczhang/homebrew/Cellar/valgrind/3.22.0/libexec/valgrind/vgpreload_memcheck-amd64-linux.so)
==1866237==    by 0x4D6A6D7: MPL_malloc (mpl_trmem.h:373)
==1866237==    by 0x4D6A92D: MPIDI_CH3I_Connection_alloc (ch3u_connect_sock.c:150)
==1866237==    by 0x4D6D243: MPIDI_CH3I_Sock_connect (ch3u_connect_sock.c:1070)
==1866237==    by 0x4D6D0C7: MPIDI_CH3I_VC_post_sockconnect (ch3u_connect_sock.c:1016)
==1866237==    by 0x4D7EB8D: MPIDI_CH3I_VC_post_connect (ch3_progress.c:652)
==1866237==    by 0x4D7A1A4: MPIDI_CH3_iStartMsg (ch3_istartmsg.c:154)
==1866237==    by 0x4D29ABC: MPIDI_CH3_EagerContigShortSend (ch3u_eager.c:262)
==1866237==    by 0x4D4C7FB: MPID_Send (mpid_send.c:119)
==1866237==    by 0x4D4CCF9: MPID_Send_coll (mpid_send.c:206)
==1866237==    by 0x4BACC93: MPIC_Send (helper_fns.c:126)
==1866237==    by 0x4ABC5D5: MPIR_Bcast_intra_binomial (bcast_intra_binomial.c:146)
==1866237== 
[MPI process 1] I am a broadcast receiver, and obtained value 12345.
==1866237== Syscall param write(buf) points to uninitialised byte(s)
==1866237==    at 0x518B697: write (write.c:26)
==1866237==    by 0x4D72FE1: MPIDI_CH3I_Sock_write (sock.c:2614)
==1866237==    by 0x4D79B72: MPIDI_CH3_iStartMsg (ch3_istartmsg.c:68)
==1866237==    by 0x4CA39BD: MPIDI_CH3U_VC_SendClose (ch3u_handle_connection.c:237)
==1866237==    by 0x4D6931B: MPIDI_PG_Close_VCs (mpidi_pg.c:973)
==1866237==    by 0x4D3A9EF: MPID_Finalize (mpid_finalize.c:99)
==1866237==    by 0x4C228B8: MPII_Finalize (mpir_init.c:399)
==1866237==    by 0x4C22B5F: MPIR_Finalize_impl (mpir_init.c:454)
==1866237==    by 0x49C8727: internal_Finalize (finalize.c:37)
==1866237==    by 0x49C879E: PMPI_Finalize (finalize.c:83)
==1866237==    by 0x10927B: main (ex100.c:40)
==1866237==  Address 0x1ffeffcdd8 is on thread 1's stack
==1866237==  in frame #3, created by MPIDI_CH3U_VC_SendClose (ch3u_handle_connection.c:199)
==1866237== 
==1866238== Syscall param write(buf) points to uninitialised byte(s)
==1866238==    at 0x518B697: write (write.c:26)
==1866238==    by 0x4D72FE1: MPIDI_CH3I_Sock_write (sock.c:2614)
==1866238==    by 0x4D79B72: MPIDI_CH3_iStartMsg (ch3_istartmsg.c:68)
==1866238==    by 0x4CA39BD: MPIDI_CH3U_VC_SendClose (ch3u_handle_connection.c:237)
==1866238==    by 0x4D6931B: MPIDI_PG_Close_VCs (mpidi_pg.c:973)
==1866238==    by 0x4D3A9EF: MPID_Finalize (mpid_finalize.c:99)
==1866238==    by 0x4C228B8: MPII_Finalize (mpir_init.c:399)
==1866238==    by 0x4C22B5F: MPIR_Finalize_impl (mpir_init.c:454)
==1866238==    by 0x49C8727: internal_Finalize (finalize.c:37)
==1866238==    by 0x49C879E: PMPI_Finalize (finalize.c:83)
==1866238==    by 0x10927B: main (ex100.c:40)
==1866238==  Address 0x1ffeffcdd8 is on thread 1's stack
==1866238==  in frame #3, created by MPIDI_CH3U_VC_SendClose (ch3u_handle_connection.c:199)
==1866238== 
==1866237== Syscall param write(buf) points to uninitialised byte(s)
==1866237==    at 0x518B697: write (write.c:26)
==1866237==    by 0x4D72FE1: MPIDI_CH3I_Sock_write (sock.c:2614)
==1866237==    by 0x4D79B72: MPIDI_CH3_iStartMsg (ch3_istartmsg.c:68)
==1866237==    by 0x4CA3B1B: MPIDI_CH3_PktHandler_Close (ch3u_handle_connection.c:274)
==1866237==    by 0x4D7DA5E: MPIDI_CH3I_Progress_handle_sock_event (ch3_progress.c:397)
==1866237==    by 0x4D7D121: MPIDI_CH3i_Progress_wait (ch3_progress.c:185)
==1866237==    by 0x4D7F248: MPIDI_CH3I_Progress (ch3_progress.c:808)
==1866237==    by 0x4CA3CFC: MPIDI_CH3U_VC_WaitForClose (ch3u_handle_connection.c:356)
==1866237==    by 0x4D3AA72: MPID_Finalize (mpid_finalize.c:104)
==1866237==    by 0x4C228B8: MPII_Finalize (mpir_init.c:399)
==1866237==    by 0x4C22B5F: MPIR_Finalize_impl (mpir_init.c:454)
==1866237==    by 0x49C8727: internal_Finalize (finalize.c:37)
==1866237==  Address 0x1ffeffccf8 is on thread 1's stack
==1866237==  in frame #3, created by MPIDI_CH3_PktHandler_Close (ch3u_handle_connection.c:259)
==1866237== 
==1866238== 
==1866238== HEAP SUMMARY:
==1866238==     in use at exit: 149 bytes in 3 blocks
==1866238==   total heap usage: 9,524 allocs, 9,513 frees, 6,207,412 bytes allocated
==1866238== 
==1866237== 
==1866237== HEAP SUMMARY:
==1866237==     in use at exit: 149 bytes in 3 blocks
==1866237==   total heap usage: 9,546 allocs, 9,533 frees, 6,221,330 bytes allocated
==1866237== 
==1866238== LEAK SUMMARY:
==1866238==    definitely lost: 128 bytes in 1 blocks
==1866238==    indirectly lost: 21 bytes in 2 blocks
==1866238==      possibly lost: 0 bytes in 0 blocks
==1866238==    still reachable: 0 bytes in 0 blocks
==1866238==         suppressed: 0 bytes in 0 blocks
==1866238== Rerun with --leak-check=full to see details of leaked memory
==1866238== 
==1866238== Use --track-origins=yes to see where uninitialised values come from
==1866238== For lists of detected and suppressed errors, rerun with: -s
==1866238== ERROR SUMMARY: 1 errors from 1 contexts (suppressed: 0 from 0)
==1866237== LEAK SUMMARY:
==1866237==    definitely lost: 128 bytes in 1 blocks
==1866237==    indirectly lost: 21 bytes in 2 blocks
==1866237==      possibly lost: 0 bytes in 0 blocks
==1866237==    still reachable: 0 bytes in 0 blocks
==1866237==         suppressed: 0 bytes in 0 blocks
==1866237== Rerun with --leak-check=full to see details of leaked memory
==1866237== 
==1866237== Use --track-origins=yes to see where uninitialised values come from
==1866237== For lists of detected and suppressed errors, rerun with: -s
==1866237== ERROR SUMMARY: 4 errors from 3 contexts (suppressed: 0 from 0)