[PATCH v9 0/6] elf: THP-aware load segment alignment

[WANG Rui]

On Tue, Apr 14, 2026 at 12:23 PM H.J. Lu <hjl.tools@gmail.com> wrote:
>
> On Mon, Apr 13, 2026 at 1:35 PM H.J. Lu <hjl.tools@gmail.com> wrote:
> >
> > On Mon, Apr 13, 2026 at 1:44 AM WANG Rui <wangrui@loongson.cn> wrote:
> > >
> >
> > Hi,
> >
> > Please take a look at users/hjl/hugepage/master branch:
> >
> > https://gitlab.com/x86-glibc/glibc/-/commits/users/hjl/hugepage/marker
> >
> > It includes
> >
> > 30cd1f2158 elf: Support THP segment load with THP enabled with madvise
> > 6db7f33e4e Check if linker supports -z separate-code
> > 562148f88e x86-64: Disable THP segment load for tst-valgrind-smoke
> >
> > Please test its performance under madvise THP kernel.
>
> I updated users/hjl/hugepage/master branch to fix a few bugs and
> remove merging 2 PT_LOAD segments.

Sorry for the late reply.

>From benchmarking a clang build of the Linux kernel on x86_64 with
your patch in THP madvise mode, I observed that iTLB misses were
reduced, similar to what we see in THP always mode.

Thanks,
Rui

>
> >
> > > Changes since [v8]:
> > >   * Add a new test case.
> > >   * Minor formatting fixes.
> > >   * Mark _dl_map_segment_align with attribute_hidden.
> > >
> > > Changes since [v7]:
> > >   * Rename tunable glibc.elf.hugetlb to glibc.elf.thp.
> > >   * Rebase on current master.
> > >
> > > Changes since [v6]:
> > >   * Move MAX_THP_PAGESIZE to hugepages.
> > >   * Skip THP mode probing when DL_MAP_DEFAULT_THP_PAGESIZE is non-zero.
> > >
> > > Changes since [v5]:
> > >   * No functional changes.
> > >   * Add benchmark results to the commit message.
> > >
> > > Changes since [v4]:
> > >   * Merge malloc-hugepages into hugepages.
> > >   * Limit the glibc.elf.hugetlb tunable maximum value to 1.
> > >
> > > Changes since [v3]:
> > >   * Rebased on current master.
> > >   * Resolved conflicts with recently merged changes.
> > >   * No functional changes intended.
> > >
> > > Changes since [v2]:
> > >   * Refactor THP detection into a new generic hugepages abstraction,
> > >     moving helpers out of malloc-hugepages.
> > >   * Add a new tunable, `glibc.elf.hugetlb`, to control THP-aware ELF
> > >     segment alignment.
> > >   * Move the Linux implementation of `_dl_map_segment_align` to
> > >     sysdeps/unix/sysv/linux/, making it generic for Linux (including
> > >     32-bit) and avoiding wordsize-64-specific overrides.
> > >   * Use `static inline` instead of `static __always_inline`.
> > >
> > > Changes since [v1]:
> > >   * Fix CI build failure (-Wunused-function).
> > >
> > > This patch series introduces a small extension point in the ELF loader to allow
> > > architecture-specific adjustment of load segment alignment, and uses it to
> > > improve Transparent Huge Page (THP) usage on Linux.
> > >
> > > Patch 1 removes a redundant declaration of `_dl_map_segments` from
> > > dl-load.h, avoiding `-Wunused-function` build failures and keeping the
> > > prototype colocated with its definition.
> > >
> > > Patch 2 adds a new helper, `_dl_map_segment_align`, which is called when
> > > determining the maximum alignment for ELF load segments. The generic
> > > implementation is a no-op and preserves existing behavior.
> > >
> > > Patch 3 introduces a new tunable, `glibc.elf.hugetlb`, which controls
> > > THP-aware alignment of ELF loadable segments. By default, the value is 0
> > > and existing behavior is preserved.
> > >
> > > Patch 4 provides a Linux implementation of `_dl_map_segment_align` that
> > > opportunistically aligns large, suitably aligned, non-writable `PT_LOAD`
> > > segments to the system’s default THP page size when THP is configured to
> > > be used unconditionally and the tunable is enabled.
> > >
> > > Patch 5 enables this behavior by default on LoongArch64 Linux and defines
> > > the default THP page size (32MB), matching the architecture’s PMD huge page
> > > geometry.
> > >
> > > Patch 6 adds a new test for THP alignment of large load segments.
> > >
> > > [v8]: https://sourceware.org/pipermail/libc-alpha/2026-April/176440.html
> > > [v7]: https://sourceware.org/pipermail/libc-alpha/2026-March/175776.html
> > > [v6]: https://sourceware.org/pipermail/libc-alpha/2026-March/175737.html
> > > [v5]: https://sourceware.org/pipermail/libc-alpha/2026-March/175694.html
> > > [v4]: https://sourceware.org/pipermail/libc-alpha/2026-March/175644.html
> > > [v3]: https://sourceware.org/pipermail/libc-alpha/2026-February/175464.html
> > > [v2]: https://sourceware.org/pipermail/libc-alpha/2026-February/175394.html
> > > [v1]: https://sourceware.org/pipermail/libc-alpha/2026-February/175359.html
> > >
> > > WANG Rui (6):
> > >   elf: Remove redundant _dl_map_segments declaration from dl-load.h
> > >   elf: Introduce _dl_map_segment_align hook for segment alignment tuning
> > >   tunables: Add glibc.elf.thp tunable for THP-aware segment alignment
> > >   elf: Align large load segments to PMD huge page size for THP
> > >   loongarch: Enable THP-aligned load segments by default on 64-bit
> > >   elf: Add test for THP alignment of large load segments
> > >
> > >  elf/dl-load.c                                 |   4 +
> > >  elf/dl-load.h                                 |   5 +-
> > >  elf/dl-tunables.list                          |   8 +
> > >  manual/tunables.texi                          |  24 +++
> > >  sysdeps/generic/dl-map-segment-align.h        |  26 +++
> > >  sysdeps/unix/sysv/linux/Makefile              |   8 +-
> > >  .../unix/sysv/linux/dl-map-segment-align.c    |  55 +++++++
> > >  .../unix/sysv/linux/dl-map-segment-align.h    |  27 ++++
> > >  .../unix/sysv/linux/loongarch/cpu-features.c  |   6 +
> > >  .../loongarch/lp64/dl-map-segment-align.h     |  22 +++
> > >  sysdeps/unix/sysv/linux/tst-thp-align.c       | 149 ++++++++++++++++++
> > >  sysdeps/unix/sysv/linux/tst-thp-size-mod.S    |  22 +++
> > >  12 files changed, 352 insertions(+), 4 deletions(-)
> > >  create mode 100644 sysdeps/generic/dl-map-segment-align.h
> > >  create mode 100644 sysdeps/unix/sysv/linux/dl-map-segment-align.c
> > >  create mode 100644 sysdeps/unix/sysv/linux/dl-map-segment-align.h
> > >  create mode 100644 sysdeps/unix/sysv/linux/loongarch/lp64/dl-map-segment-align.h
> > >  create mode 100644 sysdeps/unix/sysv/linux/tst-thp-align.c
> > >  create mode 100644 sysdeps/unix/sysv/linux/tst-thp-size-mod.S
> > >
> > > --
> > > 2.53.0
> > >
> >
> >
> > --
> > H.J.
>
>
>
> --
> H.J.