clang: selftests/mm gup_longterm error while loading shared libraries liburing.so.2 cannot open shared object file No such file or directory

clang: selftests/mm gup_longterm error while loading shared libraries liburing.so.2 cannot open shared object file No such file or directory

[Naresh Kamboju]

The following test regressions noticed while running selftests/mm gup_longterm
test cases on Dragonboard-845c, Dragonboard-410c, rock-pi-4, qemu-arm64 and
qemu-x86_64 this build have required selftest/mm/configs included and toolchain
is clang nightly.

Regressions found on Dragonboard-845c, Dragonboard-410c, rock-pi-4,
qemu-arm64 and qemu-x86_64
  -  selftests mm gup_longterm fails

Regression Analysis:
 - New regression? Yes
 - Reproducibility? Yes

Test regression: selftests mm gup_longterm error while loading shared
libraries liburing.so.2 cannot open shared object file No such file or
directory
Test regression: selftests mm cow error while loading shared libraries
liburing.so.2 cannot open shared object file No such file or directory
Test regression: selftests mm mlock-random-test exit=139
Test regression: selftests mm pagemap_ioctl exit=1
Test regression: selftests mm guard_regions file hole_punch

Reported-by: Linux Kernel Functional Testing <lkft@linaro.org>


## Test log
Linux version 6.15.0-next-20250606 (tuxmake@tuxmake) (Debian clang
version 21.0.0 (++20250602112323+c5a56f74fef7-1~exp1~20250602112342.1487),
Debian LLD 21.0.0) #1 SMP PREEMPT @1749190532

running ./gup_longterm
----------------------
./gup_longterm: error while loading shared libraries: liburing.so.2:
cannot open shared object file: No such file or directory
[FAIL]
 not ok 14 gup_longterm # exit=127

./cow: error while loading shared libraries: liburing.so.2: cannot
open shared object file: No such file or directory
[FAIL]
not ok 50 cow # exit=127

running ./mlock-random-test
---------------------------
TAP version 13
1..2
[  311.408456] traps: mlock-random-te[21661] general protection fault
ip:7f63210dbf0f sp:7ffdff6fca28 error:0 in
libc.so.6[adf0f,7f6321056000+165000]
[FAIL]
not ok 23 mlock-random-test # exit=139

running ./pagemap_ioctl

...
ok 53 Huge page testing: only two middle pages dirty
ok 54 # SKIP Hugetlb shmem testing: all new pages must not be written (dirty)
ok 55 # SKIP Hugetlb shmem testing: all pages must be written (dirty)
ok 56 # SKIP Hugetlb shmem testing: all pages dirty other than first
and the last one
ok 57 # SKIP Hugetlb shmem testing: PM_SCAN_WP_MATCHING | PM_SCAN_CHECK_WPASYNC
ok 58 # SKIP Hugetlb shmem testing: only middle page dirty
ok 59 # SKIP Hugetlb shmem testing: only two middle pages dirty
ok 60 # SKIP Hugetlb mem testing: all new pages must not be written (dirty)
ok 61 # SKIP Hugetlb mem testing: all pages must be written (dirty)
ok 62 # SKIP Hugetlb mem testing: all pages dirty other than first and
the last one
ok 63 # SKIP Hugetlb mem testing: PM_SCAN_WP_MATCHING |
PM_SCAN_CHECK_WPASYNC[  241.731600] run_vmtests.sh (456): drop_caches:
3
ok 64 # SKIP Hugetlb mem testing: only middle page dirty
ok 65 # SKIP Hugetlb mem testing: only two middle pages dirty
Bail out! uffd-test creation failed 12 Cannot allocate memory
12 skipped test(s) detected. Consider enabling relevant config options
to improve coverage.
Planned tests != run tests (115 != 65)
Totals: pass:53 fail:0 xfail:0 xpass:0 skip:12 error:0
[FAIL]
# not ok 48 pagemap_ioctl # exit=1

running ./guard-regions
...
RUN           guard_regions.file.hole_punch ...
guard-regions.c:1905:hole_punch:Expected madvise(&ptr[3 * page_size],
4 * page_size, MADV_REMOVE) (-1) == 0 (0)
hole_punch: Test terminated by assertion
         FAIL  guard_regions.file.hole_punch
not ok 80 guard_regions.file.hole_punch


## Source
* Kernel version: 6.16.0-rc2
* Git tree: https://kernel.googlesource.com/pub/scm/linux/kernel/git/next/linux-next.git
* Git sha: 050f8ad7b58d9079455af171ac279c4b9b828c11
* Git describe: next-20250616
* Project details:
https://qa-reports.linaro.org/lkft/linux-next-master/build/next-20250616/
* Architectures: arm64, x86_64
* Test environments: Dragonboard-845c, Dragonboard-410c,  rock-pi-4,
qemu-arm64, qemu-x86_64 and x86
* Toolchains: clang nightly
* Kconfigs: selftest/mm/config+defconfig+

## Test
* Test log: https://qa-reports.linaro.org/api/testruns/28766026/log_file/
* Test log 2: https://qa-reports.linaro.org/api/testruns/28743077/log_file/
* Build details:
https://regressions.linaro.org/lkft/linux-next-master/next-20250616/kselftest-mm/mm_run_vmtests_sh_gup_longterm/
* Build link: https://storage.tuxsuite.com/public/linaro/lkft/builds/2ya0viPHafKAe0u89drIv5fcwu2/
* Kernel config:
https://storage.tuxsuite.com/public/linaro/lkft/builds/2ya0viPHafKAe0u89drIv5fcwu2/config


## Steps to reproduce
  - tuxrun \
   --runtime podman \
   --device qemu-x86_64 \
   --boot-args rw  \
   --kernel https://storage.tuxsuite.com/public/linaro/lkft/builds/2ya0wmVl0eHb9koWyQYC7myXjpX/bzImage
\
   --rootfs https://storage.tuxboot.com/debian/20250605/trixie/amd64/rootfs.ext4.xz
\
   --modules https://storage.tuxsuite.com/public/linaro/lkft/builds/2ya0wmVl0eHb9koWyQYC7myXjpX/modules.tar.xz
/usr/ \
   --parameters MODULES_PATH=/usr/ \
   --parameters
SQUAD_URL=https://qa-reports.linaro.org//api/submit/lkft/linux-next-master/next-20250616/env/
\
   --parameters SKIPFILE=skipfile-lkft.yaml \
   --parameters
KSELFTEST=https://storage.tuxsuite.com/public/linaro/lkft/builds/2ya0wmVl0eHb9koWyQYC7myXjpX/kselftest.tar.xz
\
   --image docker.io/linaro/tuxrun-dispatcher:v1.2.2 \
   --tests kselftest-mm \
   --timeouts boot=15

--
Linaro LKFT
https://lkft.linaro.org

Re: clang: selftests/mm gup_longterm error while loading shared libraries liburing.so.2 cannot open shared object file No such file or directory

[Christian Heusel]

[-- Attachment #1: Type: text/plain, Size: 6013 bytes --]

On 25/06/16 11:02PM, Naresh Kamboju wrote:
> The following test regressions noticed while running selftests/mm gup_longterm
> test cases on Dragonboard-845c, Dragonboard-410c, rock-pi-4, qemu-arm64 and
> qemu-x86_64 this build have required selftest/mm/configs included and toolchain
> is clang nightly.
> 
> Regressions found on Dragonboard-845c, Dragonboard-410c, rock-pi-4,
> qemu-arm64 and qemu-x86_64
>   -  selftests mm gup_longterm fails
> 
> Regression Analysis:
>  - New regression? Yes
>  - Reproducibility? Yes
> 
> Test regression: selftests mm gup_longterm error while loading shared
> libraries liburing.so.2 cannot open shared object file No such file or
> directory
> Test regression: selftests mm cow error while loading shared libraries
> liburing.so.2 cannot open shared object file No such file or directory

These do not really look like kernel regressions, rather like a bug in
the userspace testing tool 🤔 Could it be that the tests were not
rebuilt for the new liburing or that the dependency is not installed in
the test environment?

> Test regression: selftests mm mlock-random-test exit=139
> Test regression: selftests mm pagemap_ioctl exit=1
> Test regression: selftests mm guard_regions file hole_punch
> 
> Reported-by: Linux Kernel Functional Testing <lkft@linaro.org>
> 
> 
> ## Test log
> Linux version 6.15.0-next-20250606 (tuxmake@tuxmake) (Debian clang
> version 21.0.0 (++20250602112323+c5a56f74fef7-1~exp1~20250602112342.1487),
> Debian LLD 21.0.0) #1 SMP PREEMPT @1749190532
> 
> running ./gup_longterm
> ----------------------
> ./gup_longterm: error while loading shared libraries: liburing.so.2:
> cannot open shared object file: No such file or directory
> [FAIL]
>  not ok 14 gup_longterm # exit=127
> 
> ./cow: error while loading shared libraries: liburing.so.2: cannot
> open shared object file: No such file or directory
> [FAIL]
> not ok 50 cow # exit=127
> 
> running ./mlock-random-test
> ---------------------------
> TAP version 13
> 1..2
> [  311.408456] traps: mlock-random-te[21661] general protection fault
> ip:7f63210dbf0f sp:7ffdff6fca28 error:0 in
> libc.so.6[adf0f,7f6321056000+165000]
> [FAIL]
> not ok 23 mlock-random-test # exit=139
> 
> running ./pagemap_ioctl
> 
> ...
> ok 53 Huge page testing: only two middle pages dirty
> ok 54 # SKIP Hugetlb shmem testing: all new pages must not be written (dirty)
> ok 55 # SKIP Hugetlb shmem testing: all pages must be written (dirty)
> ok 56 # SKIP Hugetlb shmem testing: all pages dirty other than first
> and the last one
> ok 57 # SKIP Hugetlb shmem testing: PM_SCAN_WP_MATCHING | PM_SCAN_CHECK_WPASYNC
> ok 58 # SKIP Hugetlb shmem testing: only middle page dirty
> ok 59 # SKIP Hugetlb shmem testing: only two middle pages dirty
> ok 60 # SKIP Hugetlb mem testing: all new pages must not be written (dirty)
> ok 61 # SKIP Hugetlb mem testing: all pages must be written (dirty)
> ok 62 # SKIP Hugetlb mem testing: all pages dirty other than first and
> the last one
> ok 63 # SKIP Hugetlb mem testing: PM_SCAN_WP_MATCHING |
> PM_SCAN_CHECK_WPASYNC[  241.731600] run_vmtests.sh (456): drop_caches:
> 3
> ok 64 # SKIP Hugetlb mem testing: only middle page dirty
> ok 65 # SKIP Hugetlb mem testing: only two middle pages dirty
> Bail out! uffd-test creation failed 12 Cannot allocate memory
> 12 skipped test(s) detected. Consider enabling relevant config options
> to improve coverage.
> Planned tests != run tests (115 != 65)
> Totals: pass:53 fail:0 xfail:0 xpass:0 skip:12 error:0
> [FAIL]
> # not ok 48 pagemap_ioctl # exit=1
> 
> running ./guard-regions
> ...
> RUN           guard_regions.file.hole_punch ...
> guard-regions.c:1905:hole_punch:Expected madvise(&ptr[3 * page_size],
> 4 * page_size, MADV_REMOVE) (-1) == 0 (0)
> hole_punch: Test terminated by assertion
>          FAIL  guard_regions.file.hole_punch
> not ok 80 guard_regions.file.hole_punch
> 
> 
> ## Source
> * Kernel version: 6.16.0-rc2
> * Git tree: https://kernel.googlesource.com/pub/scm/linux/kernel/git/next/linux-next.git
> * Git sha: 050f8ad7b58d9079455af171ac279c4b9b828c11
> * Git describe: next-20250616
> * Project details:
> https://qa-reports.linaro.org/lkft/linux-next-master/build/next-20250616/
> * Architectures: arm64, x86_64
> * Test environments: Dragonboard-845c, Dragonboard-410c,  rock-pi-4,
> qemu-arm64, qemu-x86_64 and x86
> * Toolchains: clang nightly
> * Kconfigs: selftest/mm/config+defconfig+
> 
> ## Test
> * Test log: https://qa-reports.linaro.org/api/testruns/28766026/log_file/
> * Test log 2: https://qa-reports.linaro.org/api/testruns/28743077/log_file/
> * Build details:
> https://regressions.linaro.org/lkft/linux-next-master/next-20250616/kselftest-mm/mm_run_vmtests_sh_gup_longterm/
> * Build link: https://storage.tuxsuite.com/public/linaro/lkft/builds/2ya0viPHafKAe0u89drIv5fcwu2/
> * Kernel config:
> https://storage.tuxsuite.com/public/linaro/lkft/builds/2ya0viPHafKAe0u89drIv5fcwu2/config
> 
> 
> ## Steps to reproduce
>   - tuxrun \
>    --runtime podman \
>    --device qemu-x86_64 \
>    --boot-args rw  \
>    --kernel https://storage.tuxsuite.com/public/linaro/lkft/builds/2ya0wmVl0eHb9koWyQYC7myXjpX/bzImage
> \
>    --rootfs https://storage.tuxboot.com/debian/20250605/trixie/amd64/rootfs.ext4.xz
> \
>    --modules https://storage.tuxsuite.com/public/linaro/lkft/builds/2ya0wmVl0eHb9koWyQYC7myXjpX/modules.tar.xz
> /usr/ \
>    --parameters MODULES_PATH=/usr/ \
>    --parameters
> SQUAD_URL=https://qa-reports.linaro.org//api/submit/lkft/linux-next-master/next-20250616/env/
> \
>    --parameters SKIPFILE=skipfile-lkft.yaml \
>    --parameters
> KSELFTEST=https://storage.tuxsuite.com/public/linaro/lkft/builds/2ya0wmVl0eHb9koWyQYC7myXjpX/kselftest.tar.xz
> \
>    --image docker.io/linaro/tuxrun-dispatcher:v1.2.2 \
>    --tests kselftest-mm \
>    --timeouts boot=15
> 
> --
> Linaro LKFT
> https://lkft.linaro.org
> 

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: clang: selftests/mm gup_longterm error while loading shared libraries liburing.so.2 cannot open shared object file No such file or directory

[David Hildenbrand]

On 16.06.25 21:14, Christian Heusel wrote:
> On 25/06/16 11:02PM, Naresh Kamboju wrote:
>> The following test regressions noticed while running selftests/mm gup_longterm
>> test cases on Dragonboard-845c, Dragonboard-410c, rock-pi-4, qemu-arm64 and
>> qemu-x86_64 this build have required selftest/mm/configs included and toolchain
>> is clang nightly.
>>
>> Regressions found on Dragonboard-845c, Dragonboard-410c, rock-pi-4,
>> qemu-arm64 and qemu-x86_64
>>    -  selftests mm gup_longterm fails
>>
>> Regression Analysis:
>>   - New regression? Yes
>>   - Reproducibility? Yes
>>
>> Test regression: selftests mm gup_longterm error while loading shared
>> libraries liburing.so.2 cannot open shared object file No such file or
>> directory
 >> Test regression: selftests mm cow error while loading shared 
libraries>> liburing.so.2 cannot open shared object file No such file or 
directory
> 
> These do not really look like kernel regressions, rather like a bug in
> the userspace testing tool 🤔 Could it be that the tests were not
> rebuilt for the new liburing or that the dependency is not installed in
> the test environment?

It looks like the tests were build with liburing around, and then ran 
without liburing around.

Note that the file for example has:

#ifdef LOCAL_CONFIG_HAVE_LIBURING
#include <liburing.h>
#endif /* LOCAL_CONFIG_HAVE_LIBURING */

You should be running into similar issues with cow.c, which uses the 
exact same approach for detecting+linking liburing.

So seems like something is off in your testing environment?

-- 
Cheers,

David / dhildenb

Re: clang: selftests/mm gup_longterm error while loading shared libraries liburing.so.2 cannot open shared object file No such file or directory

[Naresh Kamboju]

On Tue, 17 Jun 2025 at 01:58, David Hildenbrand <david@redhat.com> wrote:
>
> On 16.06.25 21:14, Christian Heusel wrote:
> > On 25/06/16 11:02PM, Naresh Kamboju wrote:
> >> The following test regressions noticed while running selftests/mm gup_longterm
> >> test cases on Dragonboard-845c, Dragonboard-410c, rock-pi-4, qemu-arm64 and
> >> qemu-x86_64 this build have required selftest/mm/configs included and toolchain
> >> is clang nightly.
> >>
> >> Regressions found on Dragonboard-845c, Dragonboard-410c, rock-pi-4,
> >> qemu-arm64 and qemu-x86_64
> >>    -  selftests mm gup_longterm fails
> >>
> >> Regression Analysis:
> >>   - New regression? Yes
> >>   - Reproducibility? Yes
> >>
> >> Test regression: selftests mm gup_longterm error while loading shared
> >> libraries liburing.so.2 cannot open shared object file No such file or
> >> directory
>  >> Test regression: selftests mm cow error while loading shared
> libraries>> liburing.so.2 cannot open shared object file No such file or
> directory
> >
> > These do not really look like kernel regressions, rather like a bug in
> > the userspace testing tool 🤔 Could it be that the tests were not
> > rebuilt for the new liburing or that the dependency is not installed in
> > the test environment?
>
> It looks like the tests were build with liburing around, and then ran
> without liburing around.
>
> Note that the file for example has:
>
> #ifdef LOCAL_CONFIG_HAVE_LIBURING
> #include <liburing.h>
> #endif /* LOCAL_CONFIG_HAVE_LIBURING */
>
> You should be running into similar issues with cow.c, which uses the
> exact same approach for detecting+linking liburing.
>
> So seems like something is off in your testing environment?

The kselftest built with cross toolchain((gcc-13 and clang) vs (arm64 / x86_64))
and tar / zipped and exported to NFS and mounted by DUT the
device under test by using overlayfs.

The tar / zipped file has all installed binaries and script files to run tests.

However, The missing debian package onto the userspace has been installed
 + liburing-dev today. This will fix the reported problem.

- Naresh

>
> --
> Cheers,
>
> David / dhildenb
>