summaryrefslogtreecommitdiff
path: root/tools/perf/scripts/python/libxed.py
diff options
context:
space:
mode:
authorLinus Torvalds <torvalds@linux-foundation.org>2025-04-02 21:33:20 +0300
committerLinus Torvalds <torvalds@linux-foundation.org>2025-04-02 21:33:20 +0300
commit6cb094583aef6533d0fd47ae7d134718da3399da (patch)
tree4c96b5e4526287ee36875785c60a2840f1207064 /tools/perf/scripts/python/libxed.py
parent92b71befc349587d58fdbbe6cdd68fb67f4933a8 (diff)
parente8f45927ee5d99fa52f14205a2c7ac3820c64457 (diff)
downloadlinux-6cb094583aef6533d0fd47ae7d134718da3399da.tar.xz
Merge tag 'x86_tdx_for_6.15-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull x86 TDX updates from Dave Hansen: "Avoid direct HLT instruction execution in TDX guests. TDX guests aren't expected to use the HLT instruction directly. It causes a virtualization exception (#VE). While the #VE _can_ be handled, the current handling is slow and buggy and the easiest thing is just to avoid HLT in the first place. Plus, the kernel already has paravirt infrastructure that makes it relatively painless. Make TDX guests require paravirt and add some TDX-specific paravirt handlers which avoid HLT in the normal halt routines. Also add a warning in case another HLT sneaks in. There was a report that this leads to a "major performance improvement" on specjbb2015, probably because of the extra #VE overhead or missed wakeups from the buggy HLT handling" * tag 'x86_tdx_for_6.15-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: x86/tdx: Emit warning if IRQs are enabled during HLT #VE handling x86/tdx: Fix arch_safe_halt() execution for TDX VMs x86/paravirt: Move halt paravirt calls under CONFIG_PARAVIRT
Diffstat (limited to 'tools/perf/scripts/python/libxed.py')
0 files changed, 0 insertions, 0 deletions