[scudo] Skip AllocAfterFork test on machines with low max_map_count

Reducing the number of iterations in that test with D111342 helped,
but the failure still occured flakily when the test is ran as part
of a large test suite.

Reducing further the number of iterations might not be good enough,
so we will skip the test if the `max_map_count` variable can be
read, and if lower than a given threshold.

Differential Revision: https://reviews.llvm.org/D111465
This commit is contained in:
Kostya Kortchinsky 2021-10-08 14:12:34 -07:00
parent b4b7e605a6
commit 56a9effc42
1 changed files with 17 additions and 0 deletions

View File

@ -11,6 +11,7 @@
#include <atomic>
#include <condition_variable>
#include <fstream>
#include <memory>
#include <mutex>
#include <thread>
@ -131,6 +132,22 @@ TEST(ScudoWrappersCppTest, ThreadedNew) {
#if !SCUDO_FUCHSIA
TEST(ScudoWrappersCppTest, AllocAfterFork) {
// This test can fail flakily when ran as a part of large number of
// other tests if the maxmimum number of mappings allowed is low.
// We tried to reduce the number of iterations of the loops with
// moderate success, so we will now skip this test under those
// circumstances.
if (SCUDO_LINUX) {
long MaxMapCount = 0;
// If the file can't be accessed, we proceed with the test.
std::ifstream Stream("/proc/sys/vm/max_map_count");
if (Stream.good()) {
Stream >> MaxMapCount;
if (MaxMapCount < 200000)
return;
}
}
std::atomic_bool Stop;
// Create threads that simply allocate and free different sizes.