From: Andrew Cooper <andrew.cooper3@citrix.com>
To: Xen-devel <xen-devel@lists.xen.org>
Cc: Andrew Cooper <andrew.cooper3@citrix.com>
Subject: [PATCH v5 00/21] x86: Improvements to cpuid handling for guests
Date: Thu, 7 Apr 2016 12:57:05 +0100 [thread overview]
Message-ID: <1460030246-30153-1-git-send-email-andrew.cooper3@citrix.com> (raw)
This series is available in git form at:
http://xenbits.xen.org/git-http/people/andrewcoop/xen.git levelling-v4
Notable changes from v4 are further tweaking of the SSE* dependency tree,
fixing of deep-C and P state collection on some Linux dom0 kernels, fix the
construction of HVM domains on Skylake hardware, XSM hooks for the new
sysctls.
Most patches do now how Acks/Reviews. The remaining patches are #1-3,6-7
(x86), #8,14 (XSM), #15(ARM), #26 (Toolstack).
The current cpuid code, both in the hypervisor and toolstack, has grown
organically for a very long time, and is flawed in many ways. This series
focuses specifically on the fixing the bits pertaining to the visible
features, and I will be fixing other areas in future work (e.g. per-core,
per-package values, auditing of incoming migration values, etc.)
These changes alter the workflow of cpuid handling as follows:
Xen boots and evaluates its current capabilities. It uses this information to
calculate the maximum featuresets it can provide to guests, and provides this
information for toolstack consumption. A toolstack may then calculate a safe
set of features (taking into account migratability), and sets a guests cpuid
policy. Xen then takes care of context switching the levelling state.
In particular, this means that PV guests may have different levels while
running on the same host, an option which was not previously available.
Andrew Cooper (21):
xen/x86: Annotate VM applicability in featureset
xen/x86: Calculate maximum host and guest featuresets
xen/x86: Generate deep dependencies of features
xen/x86: Clear dependent features when clearing a cpu cap
xen/x86: Improve disabling of features which have dependencies
xen/x86: Improvements to in-hypervisor cpuid sanity checks
x86/cpu: Move set_cpumask() calls into c_early_init()
x86/cpu: Sysctl and common infrastructure for levelling context
switching
x86/cpu: Rework AMD masking MSR setup
x86/cpu: Rework Intel masking/faulting setup
x86/cpu: Context switch cpuid masks and faulting state in
context_switch()
x86/pv: Provide custom cpumasks for PV domains
x86/domctl: Update PV domain cpumasks when setting cpuid policy
xen+tools: Export maximum host and guest cpu featuresets via SYSCTL
tools/libxc: Modify bitmap operations to take void pointers
tools/libxc: Use public/featureset.h for cpuid policy generation
tools/libxc: Expose the automatically generated cpu featuremask
information
tools: Utility for dealing with featuresets
tools/libxc: Wire a featureset through to cpuid policy logic
tools/libxc: Use featuresets rather than guesswork
tools/libxc: Calculate xstate cpuid leaf from guest information
.gitignore | 1 +
tools/flask/policy/policy/modules/xen/xen.te | 2 +
tools/libxc/Makefile | 9 +
tools/libxc/include/xenctrl.h | 22 +-
tools/libxc/xc_bitops.h | 37 +-
tools/libxc/xc_cpufeature.h | 151 -------
tools/libxc/xc_cpuid_x86.c | 639 +++++++++++++++++----------
tools/libxl/libxl_cpuid.c | 2 +-
tools/misc/Makefile | 4 +
tools/misc/xen-cpuid.c | 394 +++++++++++++++++
tools/ocaml/libs/xc/xenctrl.ml | 3 +
tools/ocaml/libs/xc/xenctrl.mli | 4 +
tools/ocaml/libs/xc/xenctrl_stubs.c | 37 +-
tools/python/xen/lowlevel/xc/xc.c | 2 +-
xen/arch/x86/apic.c | 2 +-
xen/arch/x86/cpu/amd.c | 292 ++++++++----
xen/arch/x86/cpu/common.c | 41 +-
xen/arch/x86/cpu/intel.c | 278 ++++++++----
xen/arch/x86/cpuid.c | 218 +++++++++
xen/arch/x86/crash.c | 3 +
xen/arch/x86/domain.c | 18 +-
xen/arch/x86/domctl.c | 138 ++++++
xen/arch/x86/hvm/hvm.c | 125 ++++--
xen/arch/x86/setup.c | 3 +
xen/arch/x86/sysctl.c | 57 +++
xen/arch/x86/traps.c | 254 +++++++----
xen/arch/x86/xstate.c | 6 +-
xen/include/asm-x86/cpufeature.h | 4 +
xen/include/asm-x86/cpuid.h | 51 +++
xen/include/asm-x86/domain.h | 2 +
xen/include/asm-x86/processor.h | 2 +-
xen/include/public/arch-x86/cpufeatureset.h | 190 ++++----
xen/include/public/sysctl.h | 50 +++
xen/tools/gen-cpuid.py | 185 +++++++-
xen/xsm/flask/hooks.c | 6 +
xen/xsm/flask/policy/access_vectors | 4 +
36 files changed, 2400 insertions(+), 836 deletions(-)
delete mode 100644 tools/libxc/xc_cpufeature.h
create mode 100644 tools/misc/xen-cpuid.c
--
2.1.4
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel
next reply other threads:[~2016-04-07 11:57 UTC|newest]
Thread overview: 50+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-04-07 11:57 Andrew Cooper [this message]
2016-04-07 11:57 ` [PATCH v5 01/21] xen/x86: Annotate VM applicability in featureset Andrew Cooper
2016-04-07 23:01 ` Jan Beulich
2016-04-07 11:57 ` [PATCH v5 02/21] xen/x86: Calculate maximum host and guest featuresets Andrew Cooper
2016-04-07 23:04 ` Jan Beulich
2016-04-07 11:57 ` [PATCH v5 03/21] xen/x86: Generate deep dependencies of features Andrew Cooper
2016-04-07 23:18 ` Jan Beulich
2016-04-07 23:36 ` Andrew Cooper
2016-04-08 15:17 ` Jan Beulich
2016-04-08 15:18 ` Andrew Cooper
2016-04-07 11:57 ` [PATCH v5 04/21] xen/x86: Clear dependent features when clearing a cpu cap Andrew Cooper
2016-04-08 15:36 ` Konrad Rzeszutek Wilk
2016-04-07 11:57 ` [PATCH v5 05/21] xen/x86: Improve disabling of features which have dependencies Andrew Cooper
2016-04-08 15:04 ` Konrad Rzeszutek Wilk
2016-04-07 11:57 ` [PATCH v5 06/21] xen/x86: Improvements to in-hypervisor cpuid sanity checks Andrew Cooper
2016-04-08 16:10 ` Konrad Rzeszutek Wilk
2016-04-08 18:06 ` Jan Beulich
2016-04-07 11:57 ` [PATCH v5 07/21] x86/cpu: Move set_cpumask() calls into c_early_init() Andrew Cooper
2016-04-08 18:09 ` Jan Beulich
2016-04-07 11:57 ` [PATCH v5 08/21] x86/cpu: Sysctl and common infrastructure for levelling context switching Andrew Cooper
2016-04-07 16:54 ` Daniel De Graaf
2016-04-08 16:12 ` Konrad Rzeszutek Wilk
2016-04-07 11:57 ` [PATCH v5 09/21] x86/cpu: Rework AMD masking MSR setup Andrew Cooper
2016-04-08 16:13 ` Konrad Rzeszutek Wilk
2016-04-07 11:57 ` [PATCH v5 10/21] x86/cpu: Rework Intel masking/faulting setup Andrew Cooper
2016-04-08 16:14 ` Konrad Rzeszutek Wilk
2016-04-07 11:57 ` [PATCH v5 11/21] x86/cpu: Context switch cpuid masks and faulting state in context_switch() Andrew Cooper
2016-04-08 16:15 ` Konrad Rzeszutek Wilk
2016-04-07 11:57 ` [PATCH v5 12/21] x86/pv: Provide custom cpumasks for PV domains Andrew Cooper
2016-04-08 16:17 ` Konrad Rzeszutek Wilk
2016-04-07 11:57 ` [PATCH v5 13/21] x86/domctl: Update PV domain cpumasks when setting cpuid policy Andrew Cooper
2016-04-08 16:26 ` Konrad Rzeszutek Wilk
2016-04-07 11:57 ` [PATCH v5 14/21] xen+tools: Export maximum host and guest cpu featuresets via SYSCTL Andrew Cooper
2016-04-07 16:54 ` Daniel De Graaf
2016-04-08 16:32 ` Konrad Rzeszutek Wilk
2016-04-07 11:57 ` [PATCH v5 15/21] tools/libxc: Modify bitmap operations to take void pointers Andrew Cooper
2016-04-07 13:00 ` Wei Liu
2016-04-08 16:34 ` Konrad Rzeszutek Wilk
2016-04-07 11:57 ` [PATCH v5 16/21] tools/libxc: Use public/featureset.h for cpuid policy generation Andrew Cooper
2016-04-08 16:37 ` Konrad Rzeszutek Wilk
2016-04-07 11:57 ` [PATCH v5 17/21] tools/libxc: Expose the automatically generated cpu featuremask information Andrew Cooper
2016-04-08 16:38 ` Konrad Rzeszutek Wilk
2016-04-07 11:57 ` [PATCH v5 18/21] tools: Utility for dealing with featuresets Andrew Cooper
2016-04-07 11:57 ` [PATCH v5 19/21] tools/libxc: Wire a featureset through to cpuid policy logic Andrew Cooper
2016-04-07 11:57 ` [PATCH v5 20/21] tools/libxc: Use featuresets rather than guesswork Andrew Cooper
2016-04-07 11:57 ` [PATCH v5 21/21] tools/libxc: Calculate xstate cpuid leaf from guest information Andrew Cooper
2016-04-07 12:58 ` Wei Liu
2016-04-08 21:00 ` Jan Beulich
2016-04-08 21:45 ` Andrew Cooper
2016-04-08 22:38 ` Jan Beulich
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1460030246-30153-1-git-send-email-andrew.cooper3@citrix.com \
--to=andrew.cooper3@citrix.com \
--cc=xen-devel@lists.xen.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.