Segfault in getrs! during testing of MultiScaleArrays.jl

[maleadt]

PkgEval encounters the following segfault during testing of MultiScaleArrays.jl: https://s3.amazonaws.com/julialang-reports/nanosoldier/pkgeval/by_date/2023-04/22/MultiScaleArrays.primary.log

[44] signal (11.128): Segmentation fault
in expression starting at /home/pkgeval/.julia/packages/MultiScaleArrays/02k83/test/dynamic_diffeq.jl:66
dlaswp_plus at /opt/julia/bin/../lib/julia/libopenblas64_.so (unknown line)
dgetrs_N_single at /opt/julia/bin/../lib/julia/libopenblas64_.so (unknown line)
dgetrs_64_ at /opt/julia/bin/../lib/julia/libopenblas64_.so (unknown line)
getrs! at /cache/build/default-amdci5-3/julialang/julia-master/usr/share/julia/stdlib/v1.10/LinearAlgebra/src/lapack.jl:1015
ldiv! at /cache/build/default-amdci5-3/julialang/julia-master/usr/share/julia/stdlib/v1.10/LinearAlgebra/src/lu.jl:425 [inlined]
ldiv! at /home/pkgeval/.julia/packages/MultiScaleArrays/02k83/src/math.jl:99
ldiv! at /cache/build/default-amdci5-3/julialang/julia-master/usr/share/julia/stdlib/v1.10/LinearAlgebra/src/factorization.jl:188
#solve#14 at /home/pkgeval/.julia/packages/LinearSolve/LD2dF/src/factorization.jl:451 [inlined]
solve at /home/pkgeval/.julia/packages/LinearSolve/LD2dF/src/factorization.jl:442 [inlined]
#solve#6 at /home/pkgeval/.julia/packages/LinearSolve/LD2dF/src/common.jl:224 [inlined]
solve at /home/pkgeval/.julia/packages/LinearSolve/LD2dF/src/common.jl:223 [inlined]
#dolinsolve#3 at /home/pkgeval/.julia/packages/OrdinaryDiffEq/gjQVg/src/misc_utils.jl:106 [inlined]
dolinsolve at /home/pkgeval/.julia/packages/OrdinaryDiffEq/gjQVg/src/misc_utils.jl:83 [inlined]
perform_step! at /home/pkgeval/.julia/packages/OrdinaryDiffEq/gjQVg/src/perform_step/rosenbrock_perform_step.jl:58
perform_step! at /home/pkgeval/.julia/packages/OrdinaryDiffEq/gjQVg/src/perform_step/rosenbrock_perform_step.jl:29 [inlined]
solve! at /home/pkgeval/.julia/packages/OrdinaryDiffEq/gjQVg/src/solve.jl:520

The segfault can be reproduced by calling getrs! with the following inputs:

factors = [-1.0375496667427269 -0.005364238020436056 -0.00536423806812862 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0; 0.01034020502789305 -1.0428584375621481 -0.010673008814814361 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0; 0.015510307310650843 0.015351568942073743 -1.048031094361553 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0; -0.0 -0.0 -0.0 -1.0643210765001119 -0.01979110098964682 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0; -0.0 -0.0 -0.0 0.023243809028434988 -1.0688088317993285 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0; -0.0 -0.0 -0.0 -0.0 -0.0 -1.0375496667427269 -0.005364238020436056 -0.00536423806812862 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0; -0.0 -0.0 -0.0 -0.0 -0.0 0.01034020502789305 -1.0428584375621481 -0.010673008814814361 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0; -0.0 -0.0 -0.0 -0.0 -0.0 0.015510307310650843 0.015351568942073743 -1.048031094361553 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0; -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -1.0522803778482694 -0.012064702619429851 -0.012064702481971475 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0; -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 0.007643528838868881 -1.048166593095576 -0.007950918043954668 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0; -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 0.019108822299936577 0.018963871173813467 -1.0599421902245667 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0; -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -1.064310907564103 -0.01837454501830832 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0; -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 0.025896397144739576 -1.0730223461184658 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0; -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -1.0522803778482694 -0.012064702619429851 -0.012064702481971475 0.0 0.0 0.0 0.0 0.0 0.0 0.0; -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 0.007643528838868881 -1.048166593095576 -0.007950918043954668 0.0 0.0 0.0 0.0 0.0 0.0 0.0; -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 0.019108822299936577 0.018963871173813467 -1.0599421902245667 0.0 0.0 0.0 0.0 0.0 0.0 0.0; -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -1.064310907564103 -0.01837454501830832 0.0 0.0 0.0 0.0 0.0; -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 0.025896397144739576 -1.0730223461184658 0.0 0.0 0.0 0.0 0.0; -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -1.0375496667427269 -0.005364238020436056 -0.00536423806812862 0.0 0.0; -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 0.01034020502789305 -1.0428584375621481 -0.010673008814814361 0.0 0.0; -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 0.015510307310650843 0.015351568942073743 -1.048194941792193 0.0 0.0; -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 0.0 -1.0643210765001119 -0.01979110098964682; -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 0.0 -0.024738875847106944 -1.0692688523711944]
ipiv = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20]
B = [-0.4997096221673574, -0.9994192443333407, -1.499128866497099, -1.7005190903124199, -2.1256488629121635, -0.4997096221673574, -0.9994192443333407, -1.499128866497099, -1.4043079260213087, -0.9362052840017713, -2.3405132100204273, -1.628667034221737, -2.4430005513335136, -1.4043079260213087, -0.9362052840017713, -2.3405132100204273, -1.628667034221737, -2.4430005513335136, -0.4997096221673574, -0.9994192443333407, -1.499128866497099, -1.7005190903124199, -2.1256488629121635]

using LinearAlgebra
LinearAlgebra.LAPACK.getrs!('N', factors, ipiv, B)

Seeing how these inputs (which I extracted from the MultiScaleArrays failure) make Julia crash all the way back to 1.6, I'm guessing there's something wrong with them. However, our LAPACK wrappers are AFAIK supposed to be safe wrappers that validate their inputs, so probably shouldn't simply segfault.

[gbaraldi]

Would be interesting to try this with MKL as well.

[maleadt]

Without calling getrs! directly:

using LinearAlgebra
x = rand(23, 23)
y = collect(1:20)
A = LU(x, y, 0)
LinearAlgebra.ldiv!(A, rand(23))

Calling LinearAlgebra.ldiv!(A, x) also segfaults, but differently (not in libopenblas).

[gbaraldi]

MKL doesn't segfault but does print Intel MKL ERROR: Parameter 6 was incorrect on entry to DLASWP. on both cases.

[aravindh-krishnamoorthy]

Hello, here is the problem isolated. I'll take a look later if possible. Just copy the code below to getrs.jl and execute julia> include("getrs.jl"); see below after the code.

using LinearAlgebra
using LinearAlgebra.BLAS: @blasfunc, libblastrampoline, chkuplo
using LinearAlgebra: BlasFloat, BlasInt, LAPACKException, DimensionMismatch,
    SingularException, PosDefException, chkstride1, checksquare,triu, tril, dot
using LinearAlgebra.LAPACK: chktrans

using Base: iszero, require_one_based_indexing

#     SUBROUTINE DGETRS( TRANS, N, NRHS, A, LDA, IPIV, B, LDB, INFO )
#*     .. Scalar Arguments ..
#      CHARACTER          TRANS
#      INTEGER            INFO, LDA, LDB, N, NRHS
#     .. Array Arguments ..
#      INTEGER            IPIV( * )
#      DOUBLE PRECISION   A( LDA, * ), B( LDB, * )
function getrs!(trans::AbstractChar, A::AbstractMatrix{Float64}, ipiv::AbstractVector{BlasInt}, B::AbstractVecOrMat{Float64})
    require_one_based_indexing(A, ipiv, B)
    chktrans(trans)
    chkstride1(A, B, ipiv)
    n = checksquare(A)
    if n != size(B, 1)
        throw(DimensionMismatch("B has leading dimension $(size(B,1)), but needs $n"))
    end
    nrhs = size(B, 2)
    info = Ref{BlasInt}()
    ccall((@blasfunc(dgetrs_), libblastrampoline), Cvoid,
          (Ref{UInt8}, Ref{BlasInt}, Ref{BlasInt}, Ptr{Float64}, Ref{BlasInt},
           Ptr{BlasInt}, Ptr{Float64}, Ref{BlasInt}, Ptr{BlasInt}, Clong),
          trans, n, size(B,2), A, max(1,stride(A,2)), ipiv, B, max(1,stride(B,2)), info, 1)
    chklapackerror(info[])
    B
end

factors = [-1.0375496667427269 -0.005364238020436056 -0.00536423806812862 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0; 0.01034020502789305 -1.0428584375621481 -0.010673008814814361 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0; 0.015510307310650843 0.015351568942073743 -1.048031094361553 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0; -0.0 -0.0 -0.0 -1.0643210765001119 -0.01979110098964682 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0; -0.0 -0.0 -0.0 0.023243809028434988 -1.0688088317993285 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0; -0.0 -0.0 -0.0 -0.0 -0.0 -1.0375496667427269 -0.005364238020436056 -0.00536423806812862 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0; -0.0 -0.0 -0.0 -0.0 -0.0 0.01034020502789305 -1.0428584375621481 -0.010673008814814361 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0; -0.0 -0.0 -0.0 -0.0 -0.0 0.015510307310650843 0.015351568942073743 -1.048031094361553 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0; -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -1.0522803778482694 -0.012064702619429851 -0.012064702481971475 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0; -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 0.007643528838868881 -1.048166593095576 -0.007950918043954668 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0; -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 0.019108822299936577 0.018963871173813467 -1.0599421902245667 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0; -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -1.064310907564103 -0.01837454501830832 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0; -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 0.025896397144739576 -1.0730223461184658 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0; -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -1.0522803778482694 -0.012064702619429851 -0.012064702481971475 0.0 0.0 0.0 0.0 0.0 0.0 0.0; -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 0.007643528838868881 -1.048166593095576 -0.007950918043954668 0.0 0.0 0.0 0.0 0.0 0.0 0.0; -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 0.019108822299936577 0.018963871173813467 -1.0599421902245667 0.0 0.0 0.0 0.0 0.0 0.0 0.0; -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -1.064310907564103 -0.01837454501830832 0.0 0.0 0.0 0.0 0.0; -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 0.025896397144739576 -1.0730223461184658 0.0 0.0 0.0 0.0 0.0; -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -1.0375496667427269 -0.005364238020436056 -0.00536423806812862 0.0 0.0; -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 0.01034020502789305 -1.0428584375621481 -0.010673008814814361 0.0 0.0; -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 0.015510307310650843 0.015351568942073743 -1.048194941792193 0.0 0.0; -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 0.0 -1.0643210765001119 -0.01979110098964682; -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 -0.0 0.0 -0.024738875847106944 -1.0692688523711944]
ipiv = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20]
B = [-0.4997096221673574, -0.9994192443333407, -1.499128866497099, -1.7005190903124199, -2.1256488629121635, -0.4997096221673574, -0.9994192443333407, -1.499128866497099, -1.4043079260213087, -0.9362052840017713, -2.3405132100204273, -1.628667034221737, -2.4430005513335136, -1.4043079260213087, -0.9362052840017713, -2.3405132100204273, -1.628667034221737, -2.4430005513335136, -0.4997096221673574, -0.9994192443333407, -1.499128866497099, -1.7005190903124199, -2.1256488629121635]

getrs!('N', factors, ipiv, B)

Example run:

julia> include("getrs.jl")

signal (11): Segmentation fault
in expression starting at /home/ark/getrs.jl:38
dlaswp_plus at /home/ark/.julia/juliaup/julia-1.8.5+0.x64.linux.gnu/bin/../lib/julia/libopenblas64_.so (unknown line)
dgetrs_N_parallel at /home/ark/.julia/juliaup/julia-1.8.5+0.x64.linux.gnu/bin/../lib/julia/libopenblas64_.so (unknown line)
dgetrs_64_ at /home/ark/.julia/juliaup/julia-1.8.5+0.x64.linux.gnu/bin/../lib/julia/libopenblas64_.so (unknown line)
getrs! at /home/ark/getrs.jl:26
unknown function (ip: 0x7f87536cade0)
_jl_invoke at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/src/gf.c:2377 [inlined]
ijl_apply_generic at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/src/gf.c:2559
jl_apply at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/src/julia.h:1843 [inlined]
do_call at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/src/interpreter.c:126
eval_value at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/src/interpreter.c:215
eval_stmt_value at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/src/interpreter.c:166 [inlined]
eval_body at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/src/interpreter.c:612
jl_interpret_toplevel_thunk at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/src/interpreter.c:750
jl_toplevel_eval_flex at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/src/toplevel.c:906
jl_toplevel_eval_flex at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/src/toplevel.c:850
ijl_toplevel_eval_in at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/src/toplevel.c:965
eval at ./boot.jl:368 [inlined]
include_string at ./loading.jl:1428
_jl_invoke at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/src/gf.c:2377 [inlined]
ijl_apply_generic at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/src/gf.c:2559
_include at ./loading.jl:1488
include at ./client.jl:476
unknown function (ip: 0x7f87536ca582)
_jl_invoke at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/src/gf.c:2377 [inlined]
ijl_apply_generic at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/src/gf.c:2559
jl_apply at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/src/julia.h:1843 [inlined]
do_call at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/src/interpreter.c:126
eval_value at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/src/interpreter.c:215
eval_stmt_value at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/src/interpreter.c:166 [inlined]
eval_body at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/src/interpreter.c:612
jl_interpret_toplevel_thunk at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/src/interpreter.c:750
jl_toplevel_eval_flex at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/src/toplevel.c:906
jl_toplevel_eval_flex at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/src/toplevel.c:850
ijl_toplevel_eval_in at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/src/toplevel.c:965
eval at ./boot.jl:368 [inlined]
eval_user_input at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/usr/share/julia/stdlib/v1.8/REPL/src/REPL.jl:151
repl_backend_loop at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/usr/share/julia/stdlib/v1.8/REPL/src/REPL.jl:247
start_repl_backend at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/usr/share/julia/stdlib/v1.8/REPL/src/REPL.jl:232
#run_repl#47 at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/usr/share/julia/stdlib/v1.8/REPL/src/REPL.jl:369
run_repl at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/usr/share/julia/stdlib/v1.8/REPL/src/REPL.jl:355
jfptr_run_repl_65104.clone_1 at /home/ark/.julia/juliaup/julia-1.8.5+0.x64.linux.gnu/lib/julia/sys.so (unknown line)
_jl_invoke at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/src/gf.c:2377 [inlined]
ijl_apply_generic at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/src/gf.c:2559
JuliaLang/julia#967 at ./client.jl:419
jfptr_YY.967_33139.clone_1 at /home/ark/.julia/juliaup/julia-1.8.5+0.x64.linux.gnu/lib/julia/sys.so (unknown line)
_jl_invoke at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/src/gf.c:2377 [inlined]
ijl_apply_generic at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/src/gf.c:2559
jl_apply at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/src/julia.h:1843 [inlined]
jl_f__call_latest at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/src/builtins.c:774
#invokelatest#2 at ./essentials.jl:729 [inlined]
invokelatest at ./essentials.jl:726 [inlined]
run_main_repl at ./client.jl:404
exec_options at ./client.jl:318
_start at ./client.jl:522
jfptr__start_38041.clone_1 at /home/ark/.julia/juliaup/julia-1.8.5+0.x64.linux.gnu/lib/julia/sys.so (unknown line)
_jl_invoke at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/src/gf.c:2377 [inlined]
ijl_apply_generic at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/src/gf.c:2559
jl_apply at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/src/julia.h:1843 [inlined]
true_main at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/src/jlapi.c:575
jl_repl_entrypoint at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/src/jlapi.c:719
main at /cache/build/default-amdci4-2/julialang/julia-release-1-dot-8/cli/loader_exe.c:59
__libc_start_main at /lib/x86_64-linux-gnu/libc.so.6 (unknown line)
unknown function (ip: 0x401098)
Allocations: 2908 (Pool: 2895; Big: 13); GC: 0
Segmentation fault

[maleadt]

That is just the definition from LinearAlgebra.LAPACK, I'm not sure how that helps? That getrs! wrapper should just perform additional parameter validation as to not crash when invoking it with the invalid inputs I posted above. MultiScaleArrays.jl is probably still doing something wrong to produce those inputs, but Julia shouldn't crash.

[aravindh-krishnamoorthy]

That is just the definition from LinearAlgebra.LAPACK, I'm not sure how that helps?

That helps because now you can easily step through using the Debugger and validate the inputs. It's the first step to debugging. Also, it's easy to make changes like prints. Otherwise, you'll have to make changes to the Base and recompile, which is arguably much harder.

Do you already know what's wrong with the inputs?

[maleadt]

Otherwise, you'll have to make changes to the Base and recompile, which is arguably much harder.

You use Revise.jl for that (or for quick hacking you can just @eval into another module, or simply redefine the external method).

Do you already know what's wrong with the inputs?

The ipiv array should have the same length as the dimensions of the square input.

[aravindh-krishnamoorthy]

Ah, okay. I thought that it's something deeper with the libraries or the interfacing. I'll leave this at this then.

You use Revise.jl for that (or for quick hacking you can just @eval into another module, or simply redefine the external method).

A redefinition exactly what you see in the post above. Guess each one has our own comfortable ways :) Got used this this style before Revise.jl but maybe I should take a second look at Revise.

[aravindh-krishnamoorthy]

Firstly, this problem seems unlikely to be a library interfacing or BLAS/LAPACK issue.

The problem (ipiv of wrong dimension) is likely introduced much before the code hits MultiScaleArrays.jl in LinearSolve.jl directly or via RecursiveFactorization.jl, both of which directly use the LU() constructor, which has no checks on the dimension of ipiv. From here on, ipiv is passed blindly all the way to XGETRS.

The question is: who has to do the checks? Now, if we want to allow such LU instantiation externally, then the constructor needs the checks? Or do we need a new way of external constructions while not penalizing the good internal ones? Decisions. Decisions. Decisions. (Note: this applies uniformly to all potential callers of XGETRS like LU, QR, SVD.)

In LAPLAS and BLAS, we've always had the GIGO (garbage-in garbage-out) rule for several reasons, with no exceptions for even the "stack-dumping calls."

If anyone has ideas, I (the ranking FNG) can work up the dirty work for all callers of XGETRS and submit the patch.

[aravindh-krishnamoorthy]

@maleadt If you cherry pick aravindh-krishnamoorthy/julia@c10008a into your repository (changes stdlib/LinearAlgebra), you will likely see an exception in LinearSolve.jl during a call to LU, which can be debugged independently further. This should also avoid the segfault.

Could you please let me know if that's right? If so, I can check other entry points and write similar checks.

[maleadt]

Yes, that fix makes the MultiScaleArrays test suite error and not crash. Can you put that in a PR, with also a test for this behavior? Thanks!

[aravindh-krishnamoorthy]

@maleadt Thanks, will do! And thank you very much for the Revise suggestion. Man, Revise is now amazing and helped me save a ton of time!