On 8/19/25 8:22 AM, André Almeida wrote:
Hi Waiman,
Em 10/08/2025 19:27, Waiman Long escreveu:
The "Memory out of range" subtest of futex_numa_mpol assumes that memory
access outside of the mmap'ed area is invalid. That may not be the case
depending on the actual memory layout of the test application. When
that subtest was run on an x86-64 system with latest upstream kernel,
the test passed as an error was returned from futex_wake(). On another
powerpc system, the same subtest failed because futex_wake() returned 0.
Bail out! futex2_wake(64, 0x86) should fail, but didn't
Looking further into the passed subtest on x86-64, it was found that an
-EINVAL was returned instead of -EFAULT. The -EINVAL error was returned
because the node value test with FLAGS_NUMA set failed with a node value
of 0x7f7f. IOW, the futex memory was accessible and futex_wake() failed
because the supposed node number wasn't valid. If that memory location
happens to have a very small value (e.g. 0), the test will pass and no
error will be returned.
Since this subtest is non-deterministic, it is dropped unless we
explicitly set a guard page beyond the mmap region.
I had proposed a refactor of the futex selftests[1] and I spotted the
same issue with the memory out of range test. My solution for this was
to create a "buffer zone" with PROT_NONE to ensure that I would have a
invalid memory access:
/*
* test_harness_run() calls mmap(..., MAP_SHARED, ...), which can create
* a valid access memory region just bellow the mmap() issue here. Then,
* the test for "Memory out of range" will fail because it will succeed
* accessing the memory address after the range. To avoid this we create
* a "Buffer zone" with PROT_NONE between the two mmap's.
*/
buffer_zone = mmap(NULL, mem_size, PROT_NONE, MAP_PRIVATE |
MAP_ANONYMOUS, 0, 0);
[1]
https://lore.kernel.org/lkml/20250704-tonyk-robust_test_cleanup-v1-13-c0ff4f24c...@igalia.com/
That should work too.
I am fine with either my patch or your getting merged as it should
address the test failure that I had encountered.
Cheers,
Longman
