From: Andrew Cooper <andrew.cooper3@citrix.com>
To: Xen-devel <xen-devel@lists.xen.org>
Cc: Andrew Cooper <andrew.cooper3@citrix.com>,
Ian Jackson <Ian.Jackson@eu.citrix.com>,
Ian Campbell <Ian.Campbell@citrix.com>,
Wei Liu <wei.liu2@citrix.com>
Subject: [PATCH RFC 31/31] tools/libxc: Calculate xstate cpuid leaf from guest information
Date: Wed, 16 Dec 2015 21:24:33 +0000 [thread overview]
Message-ID: <1450301073-28191-32-git-send-email-andrew.cooper3@citrix.com> (raw)
In-Reply-To: <1450301073-28191-1-git-send-email-andrew.cooper3@citrix.com>
It is unsafe to generate the guests xstate leaves from host information, as it
prevents the differences between hosts from being hidden.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
---
CC: Ian Campbell <Ian.Campbell@citrix.com>
CC: Ian Jackson <Ian.Jackson@eu.citrix.com>
CC: Wei Liu <wei.liu2@citrix.com>
---
tools/libxc/xc_cpuid_x86.c | 44 ++++++++++++++++++++++++++++++++++----------
1 file changed, 34 insertions(+), 10 deletions(-)
diff --git a/tools/libxc/xc_cpuid_x86.c b/tools/libxc/xc_cpuid_x86.c
index 6c8995f..9c1d30d 100644
--- a/tools/libxc/xc_cpuid_x86.c
+++ b/tools/libxc/xc_cpuid_x86.c
@@ -285,40 +285,64 @@ static void intel_xc_cpuid_policy(xc_interface *xch,
}
}
+#define X86_XCR0_X87 (1ULL << 0)
+#define X86_XCR0_SSE (1ULL << 1)
+#define X86_XCR0_AVX (1ULL << 2)
+#define X86_XCR0_LWP (1ULL << 62)
+
#define XSAVEOPT (1 << 0)
/* Configure extended state enumeration leaves (0x0000000D for xsave) */
static void xc_cpuid_config_xsave(xc_interface *xch,
const struct cpuid_domain_info *info,
const unsigned int *input, unsigned int *regs)
{
- if ( info->xfeature_mask == 0 )
+ uint64_t guest_xfeature_mask;
+
+ if ( info->xfeature_mask == 0 ||
+ !test_bit(X86_FEATURE_XSAVE, info->featureset) )
{
regs[0] = regs[1] = regs[2] = regs[3] = 0;
return;
}
+ guest_xfeature_mask = X86_XCR0_SSE | X86_XCR0_X87;
+
+ if ( test_bit(X86_FEATURE_AVX, info->featureset) )
+ guest_xfeature_mask |= X86_XCR0_AVX;
+
+ if ( test_bit(X86_FEATURE_LWP, info->featureset) )
+ guest_xfeature_mask |= X86_XCR0_LWP;
+
+ /*
+ * Clamp to host mask. Should be no-op, as guest_xfeature_mask should not
+ * be able to be calculated as larger than info->xfeature_mask.
+ *
+ * TODO - see about making this a harder error.
+ */
+ guest_xfeature_mask &= info->xfeature_mask;
+
switch ( input[1] )
{
- case 0:
+ case 0:
/* EAX: low 32bits of xfeature_enabled_mask */
- regs[0] = info->xfeature_mask & 0xFFFFFFFF;
+ regs[0] = guest_xfeature_mask & 0xFFFFFFFF;
/* EDX: high 32bits of xfeature_enabled_mask */
- regs[3] = (info->xfeature_mask >> 32) & 0xFFFFFFFF;
+ regs[3] = (guest_xfeature_mask >> 32) & 0xFFFFFFFF;
/* ECX: max size required by all HW features */
{
unsigned int _input[2] = {0xd, 0x0}, _regs[4];
regs[2] = 0;
- for ( _input[1] = 2; _input[1] < 64; _input[1]++ )
+ for ( _input[1] = 2; _input[1] <= 62; _input[1]++ )
{
cpuid(_input, _regs);
if ( (_regs[0] + _regs[1]) > regs[2] )
regs[2] = _regs[0] + _regs[1];
}
}
- /* EBX: max size required by enabled features.
- * This register contains a dynamic value, which varies when a guest
- * enables or disables XSTATE features (via xsetbv). The default size
- * after reset is 576. */
+ /* EBX: max size required by enabled features.
+ * This register contains a dynamic value, which varies when a guest
+ * enables or disables XSTATE features (via xsetbv). The default size
+ * after reset is 576. */
regs[1] = 512 + 64; /* FP/SSE + XSAVE.HEADER */
break;
case 1: /* leaf 1 */
@@ -326,7 +350,7 @@ static void xc_cpuid_config_xsave(xc_interface *xch,
regs[1] = regs[2] = regs[3] = 0;
break;
case 2 ... 63: /* sub-leaves */
- if ( !(info->xfeature_mask & (1ULL << input[1])) )
+ if ( !(guest_xfeature_mask & (1ULL << input[1])) )
{
regs[0] = regs[1] = regs[2] = regs[3] = 0;
break;
--
2.1.4
prev parent reply other threads:[~2015-12-16 21:24 UTC|newest]
Thread overview: 123+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-12-16 21:24 [PATCH RFC 00/31] x86: Improvements to cpuid handling for guests Andrew Cooper
2015-12-16 21:24 ` [PATCH RFC 01/31] xen/public: Export featureset information in the public API Andrew Cooper
2015-12-22 16:28 ` Jan Beulich
2015-12-22 16:42 ` Andrew Cooper
2015-12-22 16:59 ` Jan Beulich
2015-12-23 10:05 ` Andrew Cooper
2015-12-23 10:24 ` Jan Beulich
2015-12-23 11:26 ` Andrew Cooper
2016-01-06 7:43 ` Jan Beulich
2015-12-16 21:24 ` [PATCH RFC 02/31] tools/libxc: Use public/featureset.h for cpuid policy generation Andrew Cooper
2015-12-22 16:29 ` Jan Beulich
2016-01-05 14:13 ` Ian Campbell
2016-01-05 14:17 ` Andrew Cooper
2016-01-05 14:18 ` Ian Campbell
2016-01-05 14:23 ` Andrew Cooper
2016-01-05 15:02 ` Ian Campbell
2016-01-05 15:42 ` Andrew Cooper
2016-01-05 16:09 ` Ian Campbell
2015-12-16 21:24 ` [PATCH RFC 03/31] xen/x86: Store antifeatures inverted in a featureset Andrew Cooper
2015-12-22 16:32 ` Jan Beulich
2015-12-22 17:03 ` Andrew Cooper
2016-01-05 14:19 ` Ian Campbell
2016-01-05 14:24 ` Andrew Cooper
2015-12-16 21:24 ` [PATCH RFC 04/31] xen/x86: Mask out unknown features from Xen's capabilities Andrew Cooper
2015-12-22 16:42 ` Jan Beulich
2015-12-22 17:01 ` Andrew Cooper
2015-12-16 21:24 ` [PATCH RFC 05/31] xen/x86: Collect more CPUID feature words Andrew Cooper
2015-12-22 16:46 ` Jan Beulich
2015-12-22 17:17 ` Andrew Cooper
2015-12-16 21:24 ` [PATCH RFC 06/31] xen/x86: Infrastructure to calculate guest featuresets Andrew Cooper
2015-12-22 16:50 ` Jan Beulich
2015-12-16 21:24 ` [PATCH RFC 07/31] xen/x86: Export host featureset via SYSCTL Andrew Cooper
2015-12-22 16:57 ` Jan Beulich
2015-12-16 21:24 ` [PATCH RFC 08/31] tools/stubs: Expose host featureset to userspace Andrew Cooper
2016-01-05 15:36 ` Ian Campbell
2016-01-05 15:59 ` Andrew Cooper
2016-01-05 16:09 ` Ian Campbell
2016-01-05 16:19 ` Andrew Cooper
2016-01-05 16:38 ` Ian Campbell
2015-12-16 21:24 ` [PATCH RFC 09/31] xen/x86: Calculate PV featureset Andrew Cooper
2015-12-22 17:07 ` Jan Beulich
2015-12-22 17:13 ` Andrew Cooper
2015-12-22 17:18 ` Jan Beulich
2015-12-16 21:24 ` [PATCH RFC 10/31] xen/x86: Calculate HVM featureset Andrew Cooper
2015-12-22 17:11 ` Jan Beulich
2015-12-22 17:21 ` Andrew Cooper
2015-12-16 21:24 ` [PATCH RFC 11/31] xen/x86: Calculate Raw featureset Andrew Cooper
2015-12-22 17:14 ` Jan Beulich
2015-12-22 17:27 ` Andrew Cooper
2015-12-16 21:24 ` [PATCH RFC 12/31] tools: Utility for dealing with featuresets Andrew Cooper
2016-01-05 15:17 ` Ian Campbell
2016-01-05 16:14 ` Andrew Cooper
2016-01-05 16:34 ` Ian Campbell
2016-01-05 17:13 ` Andrew Cooper
2016-01-05 17:37 ` Ian Campbell
2016-01-05 18:04 ` Andrew Cooper
2016-01-06 10:38 ` Ian Campbell
2016-01-06 10:40 ` Ian Campbell
2016-01-06 10:42 ` Ian Campbell
2015-12-16 21:24 ` [PATCH RFC 13/31] tools/libxc: Wire a featureset through to cpuid policy logic Andrew Cooper
2016-01-05 15:42 ` Ian Campbell
2016-01-05 16:20 ` Andrew Cooper
2015-12-16 21:24 ` [PATCH RFC 14/31] tools/libxc: Use featureset rather than guesswork Andrew Cooper
2016-01-05 15:54 ` Ian Campbell
2016-01-05 16:22 ` Andrew Cooper
2015-12-16 21:24 ` [PATCH RFC 15/31] x86: Generate deep dependencies of x86 features Andrew Cooper
2016-01-05 16:03 ` Ian Campbell
2016-01-05 16:42 ` Andrew Cooper
2016-01-05 16:54 ` Ian Campbell
2016-01-05 17:09 ` Andrew Cooper
2016-01-05 17:19 ` Ian Campbell
2015-12-16 21:24 ` [PATCH RFC 16/31] x86: Automatically generate known_features Andrew Cooper
2015-12-16 21:24 ` [PATCH RFC 17/31] xen/x86: Clear dependent features when clearing a cpu cap Andrew Cooper
2015-12-16 21:24 ` [PATCH RFC 18/31] xen/x86: Improve disabling of features which have dependencies Andrew Cooper
2016-01-21 16:48 ` Jan Beulich
2015-12-16 21:24 ` [PATCH RFC 19/31] tools/libxc: Sanitise guest featuresets Andrew Cooper
2016-01-05 16:05 ` Ian Campbell
2015-12-16 21:24 ` [PATCH RFC 20/31] x86: Improvements to in-hypervisor cpuid sanity checks Andrew Cooper
2016-01-21 17:02 ` Jan Beulich
2016-01-21 17:21 ` Andrew Cooper
2016-01-21 18:15 ` Andrew Cooper
2016-01-22 7:47 ` Jan Beulich
2016-01-22 7:45 ` Jan Beulich
2015-12-16 21:24 ` [PATCH RFC 21/31] x86/domctl: Break out logic to update domain state from cpuid information Andrew Cooper
2016-01-21 17:05 ` Jan Beulich
2016-01-21 17:08 ` Andrew Cooper
2015-12-16 21:24 ` [PATCH RFC 22/31] x86/cpu: Move set_cpumask() calls into c_early_init() Andrew Cooper
2016-01-21 17:08 ` Jan Beulich
2015-12-16 21:24 ` [PATCH RFC 23/31] xen/x86: Export cpuid levelling capabilities via SYSCTL Andrew Cooper
2016-01-21 17:23 ` Jan Beulich
2015-12-16 21:24 ` [PATCH RFC 24/31] tools/stubs: Expose host levelling capabilities to userspace Andrew Cooper
2015-12-16 21:24 ` [PATCH RFC 25/31] xen/x86: Common infrastructure for levelling context switching Andrew Cooper
2016-01-22 8:56 ` Jan Beulich
2016-01-22 10:05 ` Andrew Cooper
2015-12-16 21:24 ` [PATCH RFC 26/31] xen/x86: Rework AMD masking MSR setup Andrew Cooper
2016-01-22 9:27 ` Jan Beulich
2016-01-22 11:01 ` Andrew Cooper
2016-01-22 11:13 ` Jan Beulich
2016-01-22 13:59 ` Andrew Cooper
2016-01-22 14:12 ` Jan Beulich
2016-01-22 17:03 ` Andrew Cooper
2016-01-25 11:25 ` Jan Beulich
2015-12-16 21:24 ` [PATCH RFC 27/31] xen/x86: Rework Intel masking/faulting setup Andrew Cooper
2016-01-22 9:40 ` Jan Beulich
2016-01-22 14:09 ` Andrew Cooper
2016-01-22 14:29 ` Jan Beulich
2016-01-22 14:46 ` Andrew Cooper
2016-01-22 14:53 ` Jan Beulich
2015-12-16 21:24 ` [PATCH RFC 28/31] xen/x86: Context switch all levelling state in context_switch() Andrew Cooper
2016-01-22 9:52 ` Jan Beulich
2016-01-22 14:19 ` Andrew Cooper
2016-01-22 14:31 ` Jan Beulich
2016-01-22 14:39 ` Andrew Cooper
2015-12-16 21:24 ` [PATCH RFC 29/31] x86/pv: Provide custom cpumasks for PV domains Andrew Cooper
2016-01-22 9:56 ` Jan Beulich
2016-01-22 14:24 ` Andrew Cooper
2016-01-22 14:33 ` Jan Beulich
2016-01-22 14:42 ` Andrew Cooper
2016-01-22 14:48 ` Jan Beulich
2016-01-22 14:56 ` Andrew Cooper
2015-12-16 21:24 ` [PATCH RFC 30/31] x86/domctl: Update PV domain cpumasks when setting cpuid policy Andrew Cooper
2016-01-22 10:02 ` Jan Beulich
2015-12-16 21:24 ` Andrew Cooper [this message]
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=1450301073-28191-32-git-send-email-andrew.cooper3@citrix.com \
--to=andrew.cooper3@citrix.com \
--cc=Ian.Campbell@citrix.com \
--cc=Ian.Jackson@eu.citrix.com \
--cc=wei.liu2@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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).