mirror of
https://sourceware.org/git/glibc.git
synced 2024-11-27 11:43:34 +08:00
849274d48f
The force_first parameter was ineffective because the dlclose'd object was not necessarily the first in the maps array. Also enable force_first handling unconditionally, regardless of namespace. The initial object in a namespace should be destructed first, too. The _dl_sort_maps_dfs function had early returns for relocation dependency processing which broke force_first handling, too, and this is fixed in this change as well. Reviewed-by: Adhemerval Zanella <adhemerval.zanella@linaro.org>
76 lines
3.5 KiB
Modula-2
76 lines
3.5 KiB
Modula-2
# DSO sorting test descriptions.
|
|
# This file is to be processed by ../scripts/dso-ordering-test.py, see usage
|
|
# in elf/Makefile for how it is executed.
|
|
|
|
# We test both dynamic loader sorting algorithms
|
|
tunable_option: glibc.rtld.dynamic_sort=1
|
|
tunable_option: glibc.rtld.dynamic_sort=2
|
|
|
|
# Sequence of single dependencies with no cycles.
|
|
tst-dso-ordering1: a->b->c
|
|
output: c>b>a>{}<a<b<c
|
|
|
|
# Sequence including 2 dependent DSOs not at the end of the graph.
|
|
tst-dso-ordering2: a->b->[cd]->e
|
|
output: e>d>c>b>a>{}<a<b<c<d<e
|
|
|
|
# Complex order with 3 "layers" of full dependencies
|
|
tst-dso-ordering3: a->[bc]->[def]->[gh]->i
|
|
output: i>h>g>f>e>d>c>b>a>{}<a<b<c<d<e<f<g<h<i
|
|
|
|
# Sequence including 2 dependent DSOs at the end of the graph.
|
|
# Additionally the same dependencies appear in two paths.
|
|
tst-dso-ordering4: a->b->[de];a->c->d->e
|
|
output: e>d>c>b>a>{}<a<b<c<d<e
|
|
|
|
# Test that b->c cross link is respected correctly
|
|
tst-dso-ordering5: a!->[bc]->d;b->c
|
|
output: d>c>b>a>{}<a<b<c<d
|
|
|
|
# First DSO fully dependent on 4 DSOs, with another DSO at the end of chain.
|
|
tst-dso-ordering6: a->[bcde]->f
|
|
output: f>e>d>c>b>a>{}<a<b<c<d<e<f
|
|
|
|
# Sequence including 2 dependent and 3 dependent DSOs, and one of the
|
|
# dependent DSOs is dependent on an earlier DSO.
|
|
tst-dso-ordering7: a->[bc];b->[cde];e->f
|
|
output: f>e>d>c>b>a>{}<a<b<c<d<e<f
|
|
|
|
# Sequence where the DSO c is unerlinked and calls a function in DSO a which
|
|
# is technically a cycle. The main executable depends on the first two DSOs.
|
|
# Note: This test has unspecified behavior.
|
|
tst-dso-ordering8: a->b->c=>a;{}->[ba]
|
|
output: c>b>a>{}<a<b<c
|
|
|
|
# Generate the permutation of DT_NEEDED order between the main binary and
|
|
# all 5 DSOs; all link orders should produce exact same init/fini ordering
|
|
tst-dso-ordering9: a->b->c->d->e;{}!->[abcde]
|
|
output: e>d>c>b>a>{}<a<b<c<d<e
|
|
|
|
# Test if init/fini ordering behavior is proper, despite main program with
|
|
# an soname that may cause confusion
|
|
tst-dso-ordering10: {}->a->b->c;soname({})=c
|
|
output: b>a>{}<a<b
|
|
|
|
# Complex example from Bugzilla #15311, under-linked and with circular
|
|
# relocation(dynamic) dependencies. While this is technically unspecified, the
|
|
# presumed reasonable practical behavior is for the destructor order to respect
|
|
# the static DT_NEEDED links (here this means the a->b->c->d order).
|
|
# The older dynamic_sort=1 algorithm originally did not achieve this,
|
|
# but this was a bug in the way _dl_sort_maps was called from _dl_close_worker,
|
|
# effectively disabling proper force_first handling.
|
|
# The new dynamic_sort=2 algorithm shows the effect of the simpler force_first
|
|
# handling: the a object is simply moved to the front.
|
|
# The below expected outputs are what the two algorithms currently produce
|
|
# respectively, for regression testing purposes.
|
|
tst-bz15311: {+a;+e;+f;+g;+d;%d;-d;-g;-f;-e;-a};a->b->c->d;d=>[ba];c=>a;b=>e=>a;c=>f=>b;d=>g=>c
|
|
output(glibc.rtld.dynamic_sort=1): {+a[d>c>b>a>];+e[e>];+f[f>];+g[g>];+d[];%d(b(e(a()))a()g(c(a()f(b(e(a()))))));-d[];-g[];-f[];-e[];-a[<a<b<c<d<g<f<e];}
|
|
output(glibc.rtld.dynamic_sort=2): {+a[d>c>b>a>];+e[e>];+f[f>];+g[g>];+d[];%d(b(e(a()))a()g(c(a()f(b(e(a()))))));-d[];-g[];-f[];-e[];-a[<a<g<f<b<c<d<e];}
|
|
|
|
# Test that even in the presence of dependency loops involving dlopen'ed
|
|
# object, that object is initialized last (and not unloaded prematurely).
|
|
# Final destructor order is indeterminate due to the cycle.
|
|
tst-bz28937: {+a;+b;-b;+c;%c};a->a1;a->a2;a2->a;b->b1;c->a1;c=>a1
|
|
output(glibc.rtld.dynamic_sort=1): {+a[a2>a1>a>];+b[b1>b>];-b[<b<b1];+c[c>];%c(a1());}<a<a2<c<a1
|
|
output(glibc.rtld.dynamic_sort=2): {+a[a2>a1>a>];+b[b1>b>];-b[<b<b1];+c[c>];%c(a1());}<a2<a<c<a1
|