lib/mbedtls
1
0
Fork 0
mirror of https://github.com/Mbed-TLS/mbedtls.git synced 2025-07-29 11:41:15 +03:00
Code Activity

Merge pull request #9390 from eleuzi01/backport-9327

Browse Source 
[Backport 3.6] Remove hacks about asm vs constant-flow testing
This commit is contained in:
Manuel Pégourié-Gonnard
2024-09-03 07:37:07 +00:00
committed by GitHub
parent 6cda2cae48 bc7bffcb76
commit 33a2918a48
3 changed files with 33 additions and 17 deletions
Download Patch File Download Diff File Expand all files Collapse all files

30
tests/scripts/components-sanitizers.sh
View File

@ -42,6 +42,7 @@ component_test_memsan_constant_flow () {
scripts/config.py set MBEDTLS_TEST_CONSTANT_FLOW_MEMSAN
scripts/config.py unset MBEDTLS_USE_PSA_CRYPTO
scripts/config.py unset MBEDTLS_AESNI_C # memsan doesn't grok asm
scripts/config.py unset MBEDTLS_HAVE_ASM
CC=clang cmake -D CMAKE_BUILD_TYPE:String=MemSan .
make
@ -60,6 +61,7 @@ component_test_memsan_constant_flow_psa () {
scripts/config.py full
scripts/config.py set MBEDTLS_TEST_CONSTANT_FLOW_MEMSAN
scripts/config.py unset MBEDTLS_AESNI_C # memsan doesn't grok asm
scripts/config.py unset MBEDTLS_HAVE_ASM
CC=clang cmake -D CMAKE_BUILD_TYPE:String=MemSan .
make
@ -93,7 +95,6 @@ component_release_test_valgrind_constant_flow () {
# Test asm path in constant time module - by default, it will test the plain C
# path under Valgrind or Memsan. Running only the constant_time tests is fast (<1s)
msg "test: valgrind asm constant_time"
scripts/config.py --force set MBEDTLS_TEST_CONSTANT_FLOW_ASM
skip_all_except_given_suite test_suite_constant_time
cmake -D CMAKE_BUILD_TYPE:String=Release .
make clean
@ -101,6 +102,32 @@ component_release_test_valgrind_constant_flow () {
make memcheck
}
component_release_test_valgrind_constant_flow_no_asm () {
# This tests both (1) everything that valgrind's memcheck usually checks
# (heap buffer overflows, use of uninitialized memory, use-after-free,
# etc.) and (2) branches or memory access depending on secret values,
# which will be reported as uninitialized memory. To distinguish between
# secret and actually uninitialized:
# - unset MBEDTLS_TEST_CONSTANT_FLOW_VALGRIND - does the failure persist?
# - or alternatively, build with debug info and manually run the offending
# test suite with valgrind --track-origins=yes, then check if the origin
# was TEST_CF_SECRET() or something else.
msg "build: cmake release GCC, full config minus MBEDTLS_USE_PSA_CRYPTO, minus MBEDTLS_HAVE_ASM with constant flow testing"
scripts/config.py full
scripts/config.py set MBEDTLS_TEST_CONSTANT_FLOW_VALGRIND
scripts/config.py unset MBEDTLS_USE_PSA_CRYPTO
scripts/config.py unset MBEDTLS_AESNI_C
scripts/config.py unset MBEDTLS_HAVE_ASM
skip_suites_without_constant_flow
cmake -D CMAKE_BUILD_TYPE:String=Release .
make
# this only shows a summary of the results (how many of each type)
# details are left in Testing/<date>/DynamicAnalysis.xml
msg "test: some suites (full minus MBEDTLS_USE_PSA_CRYPTO, minus MBEDTLS_HAVE_ASM, valgrind + constant flow)"
make memcheck
}
component_release_test_valgrind_constant_flow_psa () {
# This tests both (1) everything that valgrind's memcheck usually checks
# (heap buffer overflows, use of uninitialized memory, use-after-free,
@ -145,6 +172,7 @@ component_test_tsan () {
component_test_memsan () {
msg "build: MSan (clang)" # ~ 1 min 20s
scripts/config.py unset MBEDTLS_AESNI_C # memsan doesn't grok asm
scripts/config.py unset MBEDTLS_HAVE_ASM
CC=clang cmake -D CMAKE_BUILD_TYPE:String=MemSan .
make