xen-devel.lists.xenproject.org archive mirror
 help / color / mirror / Atom feed
From: Suravee Suthikulpanit <suravee.suthikulpanit@amd.com>
To: xen-devel@lists.xen.org
Cc: Jan Beulich <JBeulich@suse.com>,
	andrew.cooper3@citrix.com,
	Suravee Suthikulpanit <suravee.suthikulpanit@amd.com>,
	sherry.hurwitz@amd.com, boris.ostrovsky@oracle.com
Subject: [PATCH v2 08/10] x86/SVM: Add interrupt management code via AVIC
Date: Fri, 30 Dec 2016 23:45:59 -0600	[thread overview]
Message-ID: <1483163161-2402-9-git-send-email-suravee.suthikulpanit@amd.com> (raw)
In-Reply-To: <1483163161-2402-1-git-send-email-suravee.suthikulpanit@amd.com>

Enabling AVIC implicitly disables the V_IRQ, V_INTR_PRIO, V_IGN_TPR,
and V_INTR_VECTOR fields in the VMCB Control Word. Therefore, this patch
introduces new interrupt injection code via AVIC backing page.

Signed-off-by: Suravee Suthikulpanit <suravee.suthikulpanit@amd.com>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: Jan Beulich <JBeulich@suse.com>
Cc: Boris Ostrovsky <boris.ostrovsky@oracle.com>
---
 xen/arch/x86/hvm/svm/avic.c        | 28 ++++++++++++++++++++++++++++
 xen/arch/x86/hvm/svm/intr.c        |  4 ++++
 xen/arch/x86/hvm/svm/svm.c         | 12 ++++++++++--
 xen/include/asm-x86/hvm/svm/avic.h |  1 +
 4 files changed, 43 insertions(+), 2 deletions(-)

diff --git a/xen/arch/x86/hvm/svm/avic.c b/xen/arch/x86/hvm/svm/avic.c
index 6351c8e..faa5e45 100644
--- a/xen/arch/x86/hvm/svm/avic.c
+++ b/xen/arch/x86/hvm/svm/avic.c
@@ -636,6 +636,34 @@ void svm_avic_vmexit_do_noaccel(struct cpu_user_regs *regs)
     return;
 }
 
+void svm_avic_deliver_posted_intr(struct vcpu *v, u8 vec)
+{
+    struct vlapic *vlapic = vcpu_vlapic(v);
+
+    /* Fallback to use non-AVIC if vcpu is not enabled with AVIC. */
+    if ( !svm_avic_vcpu_enabled(v) )
+    {
+        if ( !vlapic_test_and_set_vector(vec, &vlapic->regs->data[APIC_IRR]) )
+            vcpu_kick(v);
+        return;
+    }
+
+    if ( !(guest_cpu_user_regs()->eflags & X86_EFLAGS_IF) )
+        return;
+
+    if ( vlapic_test_and_set_vector(vec, &vlapic->regs->data[APIC_IRR]) )
+        return;
+
+    /*
+     * If vcpu is running on another cpu, hit the doorbell to signal
+     * it to process interrupt. Otherwise, kick it.
+     */
+    if ( v->is_running && (v != current) )
+        wrmsrl(AVIC_DOORBELL, cpu_data[v->processor].apicid);
+    else
+        vcpu_kick(v);
+}
+
 /*
  * Local variables:
  * mode: C
diff --git a/xen/arch/x86/hvm/svm/intr.c b/xen/arch/x86/hvm/svm/intr.c
index bd94731..876d2ad 100644
--- a/xen/arch/x86/hvm/svm/intr.c
+++ b/xen/arch/x86/hvm/svm/intr.c
@@ -30,6 +30,7 @@
 #include <asm/hvm/io.h>
 #include <asm/hvm/support.h>
 #include <asm/hvm/vlapic.h>
+#include <asm/hvm/svm/avic.h>
 #include <asm/hvm/svm/svm.h>
 #include <asm/hvm/svm/intr.h>
 #include <asm/hvm/nestedhvm.h> /* for nestedhvm_vcpu_in_guestmode */
@@ -101,6 +102,9 @@ static void svm_enable_intr_window(struct vcpu *v, struct hvm_intack intack)
     HVMTRACE_3D(INTR_WINDOW, intack.vector, intack.source,
                 vmcb->eventinj.fields.v?vmcb->eventinj.fields.vector:-1);
 
+    if ( svm_avic_vcpu_enabled(v) )
+        return;
+
     /*
      * Create a dummy virtual interrupt to intercept as soon as the
      * guest can accept the real interrupt.
diff --git a/xen/arch/x86/hvm/svm/svm.c b/xen/arch/x86/hvm/svm/svm.c
index df59b8d..922f48f 100644
--- a/xen/arch/x86/hvm/svm/svm.c
+++ b/xen/arch/x86/hvm/svm/svm.c
@@ -1089,7 +1089,8 @@ static void noreturn svm_do_resume(struct vcpu *v)
         hvm_asid_flush_vcpu(v);
     }
 
-    if ( !vcpu_guestmode && !vlapic_hw_disabled(vlapic) )
+    if ( !vcpu_guestmode && !vlapic_hw_disabled(vlapic) &&
+         !svm_avic_vcpu_enabled(v) )
     {
         vintr_t intr;
 
@@ -2309,7 +2310,8 @@ void svm_vmexit_handler(struct cpu_user_regs *regs)
      * NB. We need to preserve the low bits of the TPR to make checked builds
      * of Windows work, even though they don't actually do anything.
      */
-    if ( !vcpu_guestmode && !vlapic_hw_disabled(vlapic) )
+    if ( !vcpu_guestmode && !vlapic_hw_disabled(vlapic) &&
+         !svm_avic_vcpu_enabled(v) )
     {
         intr = vmcb_get_vintr(vmcb);
         vlapic_set_reg(vlapic, APIC_TASKPRI,
@@ -2502,6 +2504,12 @@ void svm_vmexit_handler(struct cpu_user_regs *regs)
         u32 general1_intercepts = vmcb_get_general1_intercepts(vmcb);
         intr = vmcb_get_vintr(vmcb);
 
+        if ( svm_avic_vcpu_enabled(v) )
+        {
+            gdprintk(XENLOG_ERR, "AVIC VINTR:\n");
+            domain_crash(v->domain);
+        }
+
         intr.fields.irq = 0;
         general1_intercepts &= ~GENERAL1_INTERCEPT_VINTR;
 
diff --git a/xen/include/asm-x86/hvm/svm/avic.h b/xen/include/asm-x86/hvm/svm/avic.h
index 7e0abcb..1676e01 100644
--- a/xen/include/asm-x86/hvm/svm/avic.h
+++ b/xen/include/asm-x86/hvm/svm/avic.h
@@ -40,4 +40,5 @@ int svm_avic_init_vmcb(struct vcpu *v);
 void svm_avic_vmexit_do_incomp_ipi(struct cpu_user_regs *regs);
 void svm_avic_vmexit_do_noaccel(struct cpu_user_regs *regs);
 
+void svm_avic_deliver_posted_intr(struct vcpu *v, u8 vector);
 #endif /* _SVM_AVIC_H_ */
-- 
1.9.1


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
https://lists.xen.org/xen-devel

  parent reply	other threads:[~2016-12-31  5:45 UTC|newest]

Thread overview: 48+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-12-31  5:45 [PATCH v2 00/10] Introduce AMD SVM AVIC Suravee Suthikulpanit
2016-12-31  5:45 ` [PATCH v2 01/10] x86/HVM: Introduce struct hvm_pi_ops Suravee Suthikulpanit
2017-01-05  2:54   ` Tian, Kevin
2017-01-05  7:57     ` Jan Beulich
2017-01-05 15:51   ` Jan Beulich
2017-01-10  6:51     ` Suravee Suthikulpanit
2017-01-10  8:24       ` Jan Beulich
2017-01-10  9:45         ` Suravee Suthikulpanit
2016-12-31  5:45 ` [PATCH v2 02/10] x86/vLAPIC: Declare vlapic_read_aligned() and vlapic_reg_write() as non-static Suravee Suthikulpanit
2017-01-05 15:53   ` Jan Beulich
2017-01-10  6:57     ` Suravee Suthikulpanit
2017-01-10  8:25       ` Jan Beulich
2016-12-31  5:45 ` [PATCH v2 03/10] x86/HVM: Call vlapic_destroy after vcpu_destroy Suravee Suthikulpanit
2017-01-05  2:56   ` Tian, Kevin
2017-01-05 15:56   ` Jan Beulich
2017-01-10  8:18     ` Suravee Suthikulpanit
2016-12-31  5:45 ` [PATCH v2 04/10] x86/SVM: Modify VMCB fields to add AVIC support Suravee Suthikulpanit
2016-12-31  5:45 ` [PATCH v2 05/10] x86/HVM/SVM: Add AVIC initialization code Suravee Suthikulpanit
2017-01-02 16:37   ` Andrew Cooper
2017-01-04 17:24     ` Suravee Suthikulpanit
2017-01-04 17:59       ` Andrew Cooper
2017-01-10  3:06     ` Suravee Suthikulpanit
2017-01-03 14:54   ` Boris Ostrovsky
2016-12-31  5:45 ` [PATCH v2 06/10] x86/SVM: Add AVIC vmexit handlers Suravee Suthikulpanit
2017-01-02 17:28   ` Andrew Cooper
2017-01-05  4:07     ` Suravee Suthikulpanit
2017-01-03 15:34   ` Boris Ostrovsky
2017-01-05  6:41     ` Suravee Suthikulpanit
2016-12-31  5:45 ` [PATCH v2 07/10] x86/SVM: Add vcpu scheduling support for AVIC Suravee Suthikulpanit
2017-01-02 17:35   ` Andrew Cooper
2017-01-03 15:43   ` Boris Ostrovsky
2016-12-31  5:45 ` Suravee Suthikulpanit [this message]
2017-01-02 17:45   ` [PATCH v2 08/10] x86/SVM: Add interrupt management code via AVIC Andrew Cooper
2017-02-28 12:01     ` George Dunlap
2017-01-05 16:01   ` Jan Beulich
2016-12-31  5:46 ` [PATCH v2 09/10] x86/SVM: Hook up miscellaneous AVIC functions Suravee Suthikulpanit
2017-01-02 17:49   ` Andrew Cooper
2017-01-05 16:05   ` Jan Beulich
2017-01-10  8:35     ` Suravee Suthikulpanit
2017-01-10  9:00       ` Jan Beulich
2017-01-10 10:28         ` Suravee Suthikulpanit
2016-12-31  5:46 ` [PATCH v2 10/10] x86/SVM: Add AMD AVIC key handler Suravee Suthikulpanit
2017-01-03 16:01   ` Boris Ostrovsky
2017-01-03 16:04     ` Andrew Cooper
2017-01-05  8:00       ` Suravee Suthikulpanit
2017-01-05 16:07   ` Jan Beulich
2017-01-10 11:14     ` Suravee Suthikulpanit
2017-01-10 12:55       ` 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=1483163161-2402-9-git-send-email-suravee.suthikulpanit@amd.com \
    --to=suravee.suthikulpanit@amd.com \
    --cc=JBeulich@suse.com \
    --cc=andrew.cooper3@citrix.com \
    --cc=boris.ostrovsky@oracle.com \
    --cc=sherry.hurwitz@amd.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).