From: Tetsu Yamamoto <tyamamot@redhat.com>
Date: Mon, 13 Oct 2008 16:05:57 -0400
Subject: [xen] ia64: fix INIT injection
Message-id: 48F3AA25.8030608@redhat.com
O-Subject: [RHEL5.3 PATCH] xen-ia64: fix INIT injection
Bugzilla: 464445
RH-Acked-by: Bill Burns <bburns@redhat.com>
RH-Acked-by: Don Dutile <ddutile@redhat.com>
RH-Acked-by: Prarit Bhargava <prarit@redhat.com>

BZ464445
https://bugzilla.redhat.com/show_bug.cgi?id=464445

[Description]
On ia64/xen, 'xm trigger init' causes domain panic if userland
application is running on the guest.  This means that there is no way to
dump core file for windows guest by INIT signal, which is the only way
to get dump file from dom0 (xm dump-core does not work for windows guest).

This is fixed by the patch in the upstream.
- [IA64] fix INIT injection.
   http://xenbits.xensource.com/xen-unstable.hg?rev/3d96f88fb220

[Upstream Status]
The patch was applied in the upstream 4 weeks ago.

[brew ID]
http://brewweb.devel.redhat.com/brew/taskinfo?taskID=1517066

[Test Status]
I've tested this patch with kernel-119 on ia64 box and confirmed that
windows guest domain does not panic and can make dump file by 'xm
trigger init' command.

Please review and ACK.

Regards,

Tetsu Yamamoto

# HG changeset patch
# User Isaku Yamahata <yamahata@valinux.co.jp>
# Date 1221728055 -32400
# Node ID 3d96f88fb220dd1b29f4873deca378913b3db3ff
# Parent  4a381ddc764a635e9242686ef8cefb5af363c873
[IA64] fix INIT injection.

xm trigger command sometimes causes an unexpected domain panic.
There are several symptoms:
 * Guest nested fault (INIT handler runs with vpsr.cpl != 0)
 * Interrupt when IC=0
 * Unexpected virtual <--> physical mode transition

Signed-off-by: Kouya Shimura <kouya@jp.fujitsu.com>
Signed-off-by: Kazuhiro Suzuki <kaz@jp.fujitsu.com>

diff --git a/arch/ia64/vmx/vlsapic.c b/arch/ia64/vmx/vlsapic.c
index b262a80..70ef86f 100644
--- a/arch/ia64/vmx/vlsapic.c
+++ b/arch/ia64/vmx/vlsapic.c
@@ -38,6 +38,7 @@
 #include <asm/vmx_platform.h>
 #include <asm/viosapic.h>
 #include <asm/vlsapic.h>
+#include <asm/vmx_phy_mode.h>
 #include <asm/linux/jiffies.h>
 #include <xen/domain.h>
 
@@ -607,9 +608,8 @@ struct vcpu * vlsapic_lid_to_vcpu(struct domain *d, uint16_t dest)
  * To inject INIT to guest, we must set the PAL_INIT entry 
  * and set psr to switch to physical mode
  */
-#define PAL_INIT_ENTRY 0x80000000ffffffa0
 #define PSR_SET_BITS (IA64_PSR_DT | IA64_PSR_IT | IA64_PSR_RT | \
-                      IA64_PSR_IC | IA64_PSR_RI)
+                      IA64_PSR_IC | IA64_PSR_RI | IA64_PSR_I | IA64_PSR_CPL)
 
 static void vmx_inject_guest_pal_init(VCPU *vcpu)
 {
diff --git a/arch/ia64/vmx/vmx_phy_mode.c b/arch/ia64/vmx/vmx_phy_mode.c
index b6155aa..ee070c1 100644
--- a/arch/ia64/vmx/vmx_phy_mode.c
+++ b/arch/ia64/vmx/vmx_phy_mode.c
@@ -237,7 +237,12 @@ void
 switch_mm_mode(VCPU *vcpu, IA64_PSR old_psr, IA64_PSR new_psr)
 {
     int act;
-    act = mm_switch_action(old_psr, new_psr);
+    /* Switch to physical mode when injecting PAL_INIT */
+    if (unlikely(MODE_IND(new_psr) == 0 &&
+                 vcpu_regs(vcpu)->cr_iip == PAL_INIT_ENTRY))
+        act = SW_V2P;
+    else
+        act = mm_switch_action(old_psr, new_psr);
     perfc_incra(vmx_switch_mm_mode, act);
     switch (act) {
     case SW_V2P:
diff --git a/include/asm-ia64/vmx_phy_mode.h b/include/asm-ia64/vmx_phy_mode.h
index fcc6e5b..ee7b71b 100644
--- a/include/asm-ia64/vmx_phy_mode.h
+++ b/include/asm-ia64/vmx_phy_mode.h
@@ -120,4 +120,6 @@ extern void physical_tlb_miss(VCPU *vcpu, u64 vadr, int type);
 #define GUEST_VIRT  1   /* Guest in virtual mode */
 #define GUEST_PHYS  2   /* Guest in physical mode, requiring emulation */
 
+#define PAL_INIT_ENTRY 0x80000000ffffffa0
+
 #endif /* _PHY_MODE_H_ */