stress-ng fork updated to 8a8add42 (V0.14.01+)

[Po-Hsu Lin]

Hello folks,

I have our stress-ng fork updated to 8a8add42 (V0.14.01+)
Apart from Colin's tests, I have It tested with s390x / Power8 / Power9 /
ARM64 / RISCV64 systems in our pool ranging from Xenial to Jammy.

We're using commit 8a8add42 for the sysfs fix [0].

Here is the note from Colin for this update:
== V0.14.00 ==
> This release contains additional floppy device, block and tty ioctls
> tests, shared memory mutex stressors for cache coherency testing, a jpeg
> compressor stressor, support for modern x86 cpus (gcc optimizations).
>
> It also contains a new --klog-check option that checks for kernel
> errors/oopses; this may be useful for your regression tests.
>
> Other changes worth noting:
>    - urandom end of page reads (tests an upstream kernel fix)
>    - /dev/ndctl and /dev/nmem ioctl test fixes
>    - fix misuse of TCOOFF, TCOON, TCOIOFF, TCION
>    - add racy madvise/read checks
>    - core-helper: bump minimum stack size for signal handlers to 64K
>    - stress-physpage: Add mtrr add/del on physical page for x86
>   - stress-dev: add block ioctls BLKRASET, BLKFRAGET, BLKFRASET,
> BLKROSET, BLKBSZSET
>    - Add CDROMCLOSETRAY and CDROM_SELECT_DISC ioctls
>    - stress-sysfs: increase rate of sysfs files being exercised
> (increases test coverage)
>   - stress-rtc: exercise RTC_PARAM_GET and RTC_PARAM_SET
>   - add FS_IOC_FSGETXATTR/FS_IOC_FSSETXATTR ioctls
>   - stress-pci: attempt to read pci ROM via mmap
>   - stress-upd: add --upd-gro to enable Generic Receive Offload
>   - stress-madvise: add MADV_POPULATE_READ MADV_POPULATE_WRITE

== V0.14.01 ==
> Changes worth noting:
>
>   * New --msyncmany stressor to perform concurrent msync's on mappings
>   * New --mutex stressor to perform mutex stressing when changing CPU
> affinity and trying to force priority inversion issues
>   * af-alg: ignore deprecated skciper ecb(arc4) cipher
>   * All the network stressors allow one to specify the network device, e.g.
>     stress-ng --udp 1 --udp-if eth0
>   * The rawsock stressor has a workaround for spinlock contention
> issues,  see bug: https://bugzilla.kernel.org/show_bug.cgi?id=215888

Cheers
Sam

[0] https://github.com/ColinIanKing/stress-ng/commit/8a8add425a4