linux-kernel - [RFC PATCH 00/10] x86: Allocate AVX512 xstate ondemand

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-Id: <20211203003636.11417-1-jiaxun.yang@flygoat.com>
Date:   Fri,  3 Dec 2021 00:36:26 +0000
From:   Jiaxun Yang <jiaxun.yang@...goat.com>
To:     x86@...nel.org
Cc:     tglx@...utronix.de, mingo@...hat.com, bp@...en8.de,
        dave.hansen@...ux.intel.com, hpa@...or.com,
        chang.seok.bae@...el.com, linux-kernel@...r.kernel.org,
        Jiaxun Yang <jiaxun.yang@...goat.com>
Subject: [RFC PATCH 00/10] x86: Allocate AVX512 xstate ondemand

This series makes allocation of AVX512 xstate buffer ondemand.
It can save some memory (~2k for a thread not using AVX512).

Also we are going to have heterogeneous processors that
only some cores support AVX512, it can be helpful when
dealing with such processors.

Patch 1~6 are some preparations.
Patch 7 moves reallocation of xstate buffer to save_fpregs_to_fpstate.
Patch 8~10 are some cleanup and enablement work.

Tested on a Skylake-X system with AVX512 intensive y-cruncher and numpy,
the performance impact seems neglectable.

Any sugguestions are welcomed.

Cheers!
---
Jiaxun Yang
Year 1 ECS Undergraduate
University of Edinburgh

Jiaxun Yang (10):
  x86/fpu: Remove duplicated declaration of __fpu_state_size_dynamic
  x86/fpu: Split fpu_xfd_enabled from fpu_state_size_dynamic
  x86/fpu: Calculate xsave state addr on fly
  x86/fpu: Remove xstate_comp_offsets cache
  x86/fpu: Cache xstate_is_aligned
  x86/fpu/xcr: Return all enabled xfeature if CPU doesn't support
  x86/fpu: Rellocate fpstate on save_fpregs_to_fpstate
  x86/fpu: Don't keep state_size in perm struct
  x86/fpu: Split out XFEATURE_MASK_KERNEL_DYNAMIC
  x86/fpu: Mark AVX512 xfeatures as kernel dynamic

 arch/x86/include/asm/fpu/types.h  |   8 -
 arch/x86/include/asm/fpu/xcr.h    |  16 +-
 arch/x86/include/asm/fpu/xstate.h |  19 +-
 arch/x86/kernel/fpu/core.c        |  21 ++-
 arch/x86/kernel/fpu/xstate.c      | 281 ++++++++++++++----------------
 arch/x86/kernel/fpu/xstate.h      |   3 +-
 6 files changed, 181 insertions(+), 167 deletions(-)

-- 
2.30.2