Re: [linux-next:master] [mm/vma] 6898c9039b: ltp.memfd_create03.fail

[Lorenzo Stoakes <lorenzo.stoakes@oracle.com>]

On Sun, Sep 08, 2024 at 09:15:45PM GMT, kernel test robot wrote:
>
>
> Hello,
>
> kernel test robot noticed "ltp.memfd_create03.fail" on:

Hi,

Thanks for the report.

This looks _probably_ to be a problem in the test rather than the kernel.

An munmap() is failing (as expected by the test) to unmap invalid sizes
from a hugetlb allocation.

It expects EINVAL:

			if (errno == EINVAL) {
				tst_res(TINFO,
					"munmap(%p, %dkB) failed as expected",
					mem, i/1024);

But is instead receiving ENOMEM.

This is because we rearrange how we perform unmaps.

In both cases we are failing when we go ahead and attempt a split via
__split_vma(), which invokes vm_ops->may_split() which is equal to
hugetlb_vm_op_split() which returns EINVAL.

The difference is that original implementation, in do_vmi_align_munmap() we
do:

		error = __split_vma(vmi, vma, start, 1);
		if (error)
			goto start_split_failed;

Which forwards the EINVAL.

Now in vms_gather_munmap_vmas(), we do:

	int error = -ENOMEM;

	...

		if (__split_vma(vms->vmi, vms->vma, vms->start, 1))
			goto start_split_failed;

So we filter the error code to ENOMEM.

Liam - could you confirm if this needs an actual change in the kernel? If
not the test should be changed.

Test fix is obviously as simple as:

			if (errno == EINVAL || error == ENOMEM) {
				...

Cheers, Lorenzo

>
> commit: 6898c9039bc8e3027ae0fcd0f05fc2b82ccc8be0 ("mm/vma: extract the gathering of vmas from do_vmi_align_munmap()")
> https://git.kernel.org/cgit/linux/kernel/git/next/linux-next.git master
>
> [test failed on linux-next/master 9aaeb87ce1e966169a57f53a02ba05b30880ffb8]
>
> in testcase: ltp
> version: ltp-x86_64-14c1f76-1_20240831
> with following parameters:
>
> 	disk: 1HDD
> 	fs: btrfs
> 	test: syscalls-03/memfd_create03
>
>
>
> compiler: gcc-12
> test machine: 4 threads 1 sockets Intel(R) Core(TM) i3-3220 CPU @ 3.30GHz (Ivy Bridge) with 8G memory
>
> (please refer to attached dmesg/kmsg for entire log/backtrace)
>
>
>
>
> If you fix the issue in a separate patch/commit (i.e. not just a new version of
> the same patch/commit), kindly add following tags
> | Reported-by: kernel test robot <oliver.sang@intel.com>
> | Closes: https://lore.kernel.org/oe-lkp/202409081536.d283a0fb-oliver.sang@intel.com
>
>
>
> Running tests.......
> <<<test_start>>>
> tag=memfd_create03 stime=1725647495
> cmdline="memfd_create03"
> contacts=""
> analysis=exit
> <<<test_output>>>
> tst_hugepage.c:84: TINFO: 1 hugepage(s) reserved
> tst_test.c:1809: TINFO: LTP version: 20240524-209-g9a6f3896f
> tst_test.c:1813: TINFO: Tested kernel: 6.11.0-rc6-00249-g6898c9039bc8 #1 SMP PREEMPT_DYNAMIC Thu Sep  5 20:53:00 CST 2024 x86_64
> tst_test.c:1652: TINFO: Timeout per run is 0h 02m 30s
> memfd_create03.c:171: TINFO: --TESTING WRITE CALL IN HUGEPAGES--
> memfd_create03.c:176: TINFO: memfd_create() succeeded
> memfd_create03.c:70: TPASS: write(4, "LTP", 3) failed as expected
>
> memfd_create03.c:171: TINFO: --TESTING PAGE SIZE OF CREATED FILE--
> memfd_create03.c:176: TINFO: memfd_create() succeeded
> memfd_create03.c:43: TINFO: mmap((nil), 2097152, 2, 2, 4, 0) succeeded
> memfd_create03.c:96: TFAIL: munmap(0x7f9b26c00000, 512kB) failed unexpectedly: ENOMEM (12)
> memfd_create03.c:171: TINFO: --TESTING HUGEPAGE ALLOCATION LIMIT--
> memfd_create03.c:176: TINFO: memfd_create() succeeded
> memfd_create03.c:39: TBROK: mmap((nil),0,PROT_WRITE(2),2,4,0) failed: EINVAL (22)
>
> Summary:
> passed   1
> failed   1
> broken   1
> skipped  0
> warnings 0
> incrementing stop
> <<<execution_status>>>
> initiation_status="ok"
> duration=2 termination_type=exited termination_id=3 corefile=no
> cutime=0 cstime=187
> <<<test_end>>>
> INFO: ltp-pan reported some tests FAIL
> LTP Version: 20240524-209-g9a6f3896f
>
>        ###############################################################
>
>             Done executing testcases.
>             LTP Version:  20240524-209-g9a6f3896f
>        ###############################################################
>
> -------------------------------------------
> INFO: runltp script is deprecated, try kirk
> https://github.com/linux-test-project/kirk
> -------------------------------------------
>
>
>
> The kernel config and materials to reproduce are available at:
> https://download.01.org/0day-ci/archive/20240908/202409081536.d283a0fb-oliver.sang@intel.com
>
>
>
> --
> 0-DAY CI Kernel Test Service
> https://github.com/intel/lkp-tests/wiki
>