8 files changed, 798 insertions, 6 deletions

diff --git a/arch/x86/include/asm/xen/hypercall.h b/arch/x86/include/asm/xen/hypercall.h
index 5728852fb90f..59c226d120cd 100644
--- a/arch/x86/include/asm/xen/hypercall.h
+++ b/arch/x86/include/asm/xen/hypercall.h
@@ -48,6 +48,7 @@
 #include <xen/interface/sched.h>
 #include <xen/interface/physdev.h>
 #include <xen/interface/platform.h>
+#include <xen/interface/xen-mca.h>
 
 /*
  * The hypercall asms have to meet several constraints:
@@ -302,6 +303,13 @@ HYPERVISOR_set_timer_op(u64 timeout)
 }
 
 static inline int
+HYPERVISOR_mca(struct xen_mc *mc_op)
+{
+        mc_op->interface_version = XEN_MCA_INTERFACE_VERSION;
+        return _hypercall1(int, mca, mc_op);
+}
+
+static inline int
 HYPERVISOR_dom0_op(struct xen_platform_op *platform_op)
 {
         platform_op->interface_version = XENPF_INTERFACE_VERSION;
diff --git a/arch/x86/kernel/cpu/mcheck/mce.c b/arch/x86/kernel/cpu/mcheck/mce.c
index da27c5d2168a..aa7548799af4 100644
--- a/arch/x86/kernel/cpu/mcheck/mce.c
+++ b/arch/x86/kernel/cpu/mcheck/mce.c
@@ -57,8 +57,6 @@ static DEFINE_MUTEX(mce_chrdev_read_mutex);
 
 int mce_disabled __read_mostly;
 
-#define MISC_MCELOG_MINOR       227
-
 #define SPINUNIT 100    /* 100ns */
 
 atomic_t mce_entry;
@@ -2342,7 +2340,7 @@ static __init int mcheck_init_device(void)
 
         return err;
 }
-device_initcall(mcheck_init_device);
+device_initcall_sync(mcheck_init_device);
 
 /*
  * Old style boot options parsing. Only for compatibility.
diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c
index ff962d4b821e..9a6346865c49 100644
--- a/arch/x86/xen/enlighten.c
+++ b/arch/x86/xen/enlighten.c
@@ -38,6 +38,7 @@
 #include <xen/interface/physdev.h>
 #include <xen/interface/vcpu.h>
 #include <xen/interface/memory.h>
+#include <xen/interface/xen-mca.h>
 #include <xen/features.h>
 #include <xen/page.h>
 #include <xen/hvm.h>
@@ -341,9 +342,7 @@ static void __init xen_init_cpuid_mask(void)
         unsigned int xsave_mask;
 
         cpuid_leaf1_edx_mask =
-                ~((1 << X86_FEATURE_MCE)  |  /* disable MCE */
-                  (1 << X86_FEATURE_MCA)  |  /* disable MCA */
-                  (1 << X86_FEATURE_MTRR) |  /* disable MTRR */
+                ~((1 << X86_FEATURE_MTRR) |  /* disable MTRR */
                   (1 << X86_FEATURE_ACC));   /* thermal monitoring */
 
         if (!xen_initial_domain())
diff --git a/drivers/xen/Kconfig b/drivers/xen/Kconfig
index 8d2501e604dd..d4dffcd52873 100644
--- a/drivers/xen/Kconfig
+++ b/drivers/xen/Kconfig
@@ -196,4 +196,12 @@ config XEN_ACPI_PROCESSOR
           called xen_acpi_processor  If you do not know what to choose, select
           M here. If the CPUFREQ drivers are built in, select Y here.
 
+config XEN_MCE_LOG
+        bool "Xen platform mcelog"
+        depends on XEN_DOM0 && X86_64 && X86_MCE
+        default n
+        help
+          Allow kernel fetching MCE error from Xen platform and
+          converting it into Linux mcelog format for mcelog tools
+
 endmenu
diff --git a/drivers/xen/Makefile b/drivers/xen/Makefile
index fc3488631136..a7870292bc75 100644
--- a/drivers/xen/Makefile
+++ b/drivers/xen/Makefile
@@ -18,6 +18,7 @@ obj-$(CONFIG_XEN_PVHVM)			+= platform-pci.o
 obj-$(CONFIG_XEN_TMEM)                  += tmem.o
 obj-$(CONFIG_SWIOTLB_XEN)               += swiotlb-xen.o
 obj-$(CONFIG_XEN_DOM0)                  += pci.o acpi.o
+obj-$(CONFIG_XEN_MCE_LOG)               += mcelog.o
 obj-$(CONFIG_XEN_PCIDEV_BACKEND)        += xen-pciback/
 obj-$(CONFIG_XEN_PRIVCMD)               += xen-privcmd.o
 obj-$(CONFIG_XEN_ACPI_PROCESSOR)        += xen-acpi-processor.o
diff --git a/drivers/xen/mcelog.c b/drivers/xen/mcelog.c
new file mode 100644
index 000000000000..72e87d2f1929
--- /dev/null
+++ b/drivers/xen/mcelog.c
@@ -0,0 +1,392 @@
+/******************************************************************************
+ * mcelog.c
+ * Driver for receiving and transferring machine check error infomation
+ *
+ * Copyright (c) 2012 Intel Corporation
+ * Author: Liu, Jinsong <jinsong.liu@intel.com>
+ * Author: Jiang, Yunhong <yunhong.jiang@intel.com>
+ * Author: Ke, Liping <liping.ke@intel.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation; or, when distributed
+ * separately from the Linux kernel or incorporated into other
+ * software packages, subject to the following license:
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this source file (the "Software"), to deal in the Software without
+ * restriction, including without limitation the rights to use, copy, modify,
+ * merge, publish, distribute, sublicense, and/or sell copies of the Software,
+ * and to permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ */
+
+#include <linux/init.h>
+#include <linux/types.h>
+#include <linux/kernel.h>
+#include <linux/slab.h>
+#include <linux/fs.h>
+#include <linux/device.h>
+#include <linux/miscdevice.h>
+#include <linux/uaccess.h>
+#include <linux/capability.h>
+
+#include <xen/interface/xen.h>
+#include <xen/events.h>
+#include <xen/interface/vcpu.h>
+#include <xen/xen.h>
+#include <asm/xen/hypercall.h>
+#include <asm/xen/hypervisor.h>
+
+#define XEN_MCELOG "xen_mcelog: "
+
+static struct mc_info g_mi;
+static struct mcinfo_logical_cpu *g_physinfo;
+static uint32_t ncpus;
+
+static DEFINE_SPINLOCK(mcelog_lock);
+
+static struct xen_mce_log xen_mcelog = {
+        .signature      = XEN_MCE_LOG_SIGNATURE,
+        .len            = XEN_MCE_LOG_LEN,
+        .recordlen      = sizeof(struct xen_mce),
+};
+
+static DEFINE_SPINLOCK(xen_mce_chrdev_state_lock);
+static int xen_mce_chrdev_open_count;   /* #times opened */
+static int xen_mce_chrdev_open_exclu;   /* already open exclusive? */
+
+static int xen_mce_chrdev_open(struct inode *inode, struct file *file)
+{
+        spin_lock(&xen_mce_chrdev_state_lock);
+
+        if (xen_mce_chrdev_open_exclu ||
+            (xen_mce_chrdev_open_count && (file->f_flags & O_EXCL))) {
+                spin_unlock(&xen_mce_chrdev_state_lock);
+
+                return -EBUSY;
+        }
+
+        if (file->f_flags & O_EXCL)
+                xen_mce_chrdev_open_exclu = 1;
+        xen_mce_chrdev_open_count++;
+
+        spin_unlock(&xen_mce_chrdev_state_lock);
+
+        return nonseekable_open(inode, file);
+}
+
+static int xen_mce_chrdev_release(struct inode *inode, struct file *file)
+{
+        spin_lock(&xen_mce_chrdev_state_lock);
+
+        xen_mce_chrdev_open_count--;
+        xen_mce_chrdev_open_exclu = 0;
+
+        spin_unlock(&xen_mce_chrdev_state_lock);
+
+        return 0;
+}
+
+static ssize_t xen_mce_chrdev_read(struct file *filp, char __user *ubuf,
+                                size_t usize, loff_t *off)
+{
+        char __user *buf = ubuf;
+        unsigned num;
+        int i, err;
+
+        spin_lock(&mcelog_lock);
+
+        num = xen_mcelog.next;
+
+        /* Only supports full reads right now */
+        err = -EINVAL;
+        if (*off != 0 || usize < XEN_MCE_LOG_LEN*sizeof(struct xen_mce))
+                goto out;
+
+        err = 0;
+        for (i = 0; i < num; i++) {
+                struct xen_mce *m = &xen_mcelog.entry[i];
+
+                err |= copy_to_user(buf, m, sizeof(*m));
+                buf += sizeof(*m);
+        }
+
+        memset(xen_mcelog.entry, 0, num * sizeof(struct xen_mce));
+        xen_mcelog.next = 0;
+
+        if (err)
+                err = -EFAULT;
+
+out:
+        spin_unlock(&mcelog_lock);
+
+        return err ? err : buf - ubuf;
+}
+
+static long xen_mce_chrdev_ioctl(struct file *f, unsigned int cmd,
+                                unsigned long arg)
+{
+        int __user *p = (int __user *)arg;
+
+        if (!capable(CAP_SYS_ADMIN))
+                return -EPERM;
+
+        switch (cmd) {
+        case MCE_GET_RECORD_LEN:
+                return put_user(sizeof(struct xen_mce), p);
+        case MCE_GET_LOG_LEN:
+                return put_user(XEN_MCE_LOG_LEN, p);
+        case MCE_GETCLEAR_FLAGS: {
+                unsigned flags;
+
+                do {
+                        flags = xen_mcelog.flags;
+                } while (cmpxchg(&xen_mcelog.flags, flags, 0) != flags);
+
+                return put_user(flags, p);
+        }
+        default:
+                return -ENOTTY;
+        }
+}
+
+static const struct file_operations xen_mce_chrdev_ops = {
+        .open                   = xen_mce_chrdev_open,
+        .release                = xen_mce_chrdev_release,
+        .read                   = xen_mce_chrdev_read,
+        .unlocked_ioctl         = xen_mce_chrdev_ioctl,
+        .llseek                 = no_llseek,
+};
+
+static struct miscdevice xen_mce_chrdev_device = {
+        MISC_MCELOG_MINOR,
+        "mcelog",
+        &xen_mce_chrdev_ops,
+};
+
+/*
+ * Caller should hold the mcelog_lock
+ */
+static void xen_mce_log(struct xen_mce *mce)
+{
+        unsigned entry;
+
+        entry = xen_mcelog.next;
+
+        /*
+         * When the buffer fills up discard new entries.
+         * Assume that the earlier errors are the more
+         * interesting ones:
+         */
+        if (entry >= XEN_MCE_LOG_LEN) {
+                set_bit(XEN_MCE_OVERFLOW,
+                        (unsigned long *)&xen_mcelog.flags);
+                return;
+        }
+
+        memcpy(xen_mcelog.entry + entry, mce, sizeof(struct xen_mce));
+
+        xen_mcelog.next++;
+}
+
+static int convert_log(struct mc_info *mi)
+{
+        struct mcinfo_common *mic;
+        struct mcinfo_global *mc_global;
+        struct mcinfo_bank *mc_bank;
+        struct xen_mce m;
+        uint32_t i;
+
+        mic = NULL;
+        x86_mcinfo_lookup(&mic, mi, MC_TYPE_GLOBAL);
+        if (unlikely(!mic)) {
+                pr_warning(XEN_MCELOG "Failed to find global error info\n");
+                return -ENODEV;
+        }
+
+        memset(&m, 0, sizeof(struct xen_mce));
+
+        mc_global = (struct mcinfo_global *)mic;
+        m.mcgstatus = mc_global->mc_gstatus;
+        m.apicid = mc_global->mc_apicid;
+
+        for (i = 0; i < ncpus; i++)
+                if (g_physinfo[i].mc_apicid == m.apicid)
+                        break;
+        if (unlikely(i == ncpus)) {
+                pr_warning(XEN_MCELOG "Failed to match cpu with apicid %d\n",
+                           m.apicid);
+                return -ENODEV;
+        }
+
+        m.socketid = g_physinfo[i].mc_chipid;
+        m.cpu = m.extcpu = g_physinfo[i].mc_cpunr;
+        m.cpuvendor = (__u8)g_physinfo[i].mc_vendor;
+        m.mcgcap = g_physinfo[i].mc_msrvalues[__MC_MSR_MCGCAP].value;
+
+        mic = NULL;
+        x86_mcinfo_lookup(&mic, mi, MC_TYPE_BANK);
+        if (unlikely(!mic)) {
+                pr_warning(XEN_MCELOG "Fail to find bank error info\n");
+                return -ENODEV;
+        }
+
+        do {
+                if ((!mic) || (mic->size == 0) ||
+                    (mic->type != MC_TYPE_GLOBAL   &&
+                     mic->type != MC_TYPE_BANK     &&
+                     mic->type != MC_TYPE_EXTENDED &&
+                     mic->type != MC_TYPE_RECOVERY))
+                        break;
+
+                if (mic->type == MC_TYPE_BANK) {
+                        mc_bank = (struct mcinfo_bank *)mic;
+                        m.misc = mc_bank->mc_misc;
+                        m.status = mc_bank->mc_status;
+                        m.addr = mc_bank->mc_addr;
+                        m.tsc = mc_bank->mc_tsc;
+                        m.bank = mc_bank->mc_bank;
+                        m.finished = 1;
+                        /*log this record*/
+                        xen_mce_log(&m);
+                }
+                mic = x86_mcinfo_next(mic);
+        } while (1);
+
+        return 0;
+}
+
+static int mc_queue_handle(uint32_t flags)
+{
+        struct xen_mc mc_op;
+        int ret = 0;
+
+        mc_op.cmd = XEN_MC_fetch;
+        mc_op.interface_version = XEN_MCA_INTERFACE_VERSION;
+        set_xen_guest_handle(mc_op.u.mc_fetch.data, &g_mi);
+        do {
+                mc_op.u.mc_fetch.flags = flags;
+                ret = HYPERVISOR_mca(&mc_op);
+                if (ret) {
+                        pr_err(XEN_MCELOG "Failed to fetch %s error log\n",
+                               (flags == XEN_MC_URGENT) ?
+                               "urgnet" : "nonurgent");
+                        break;
+                }
+
+                if (mc_op.u.mc_fetch.flags & XEN_MC_NODATA ||
+                    mc_op.u.mc_fetch.flags & XEN_MC_FETCHFAILED)
+                        break;
+                else {
+                        ret = convert_log(&g_mi);
+                        if (ret)
+                                pr_warning(XEN_MCELOG
+                                           "Failed to convert this error log, "
+                                           "continue acking it anyway\n");
+
+                        mc_op.u.mc_fetch.flags = flags | XEN_MC_ACK;
+                        ret = HYPERVISOR_mca(&mc_op);
+                        if (ret) {
+                                pr_err(XEN_MCELOG
+                                       "Failed to ack previous error log\n");
+                                break;
+                        }
+                }
+        } while (1);
+
+        return ret;
+}
+
+/* virq handler for machine check error info*/
+static irqreturn_t xen_mce_interrupt(int irq, void *dev_id)
+{
+        int err;
+        unsigned long tmp;
+
+        spin_lock_irqsave(&mcelog_lock, tmp);
+
+        /* urgent mc_info */
+        err = mc_queue_handle(XEN_MC_URGENT);
+        if (err)
+                pr_err(XEN_MCELOG
+                       "Failed to handle urgent mc_info queue, "
+                       "continue handling nonurgent mc_info queue anyway.\n");
+
+        /* nonurgent mc_info */
+        err = mc_queue_handle(XEN_MC_NONURGENT);
+        if (err)
+                pr_err(XEN_MCELOG
+                       "Failed to handle nonurgent mc_info queue.\n");
+
+        spin_unlock_irqrestore(&mcelog_lock, tmp);
+
+        return IRQ_HANDLED;
+}
+
+static int bind_virq_for_mce(void)
+{
+        int ret;
+        struct xen_mc mc_op;
+
+        memset(&mc_op, 0, sizeof(struct xen_mc));
+
+        /* Fetch physical CPU Numbers */
+        mc_op.cmd = XEN_MC_physcpuinfo;
+        mc_op.interface_version = XEN_MCA_INTERFACE_VERSION;
+        set_xen_guest_handle(mc_op.u.mc_physcpuinfo.info, g_physinfo);
+        ret = HYPERVISOR_mca(&mc_op);
+        if (ret) {
+                pr_err(XEN_MCELOG "Failed to get CPU numbers\n");
+                return ret;
+        }
+
+        /* Fetch each CPU Physical Info for later reference*/
+        ncpus = mc_op.u.mc_physcpuinfo.ncpus;
+        g_physinfo = kcalloc(ncpus, sizeof(struct mcinfo_logical_cpu),
+                             GFP_KERNEL);
+        if (!g_physinfo)
+                return -ENOMEM;
+        set_xen_guest_handle(mc_op.u.mc_physcpuinfo.info, g_physinfo);
+        ret = HYPERVISOR_mca(&mc_op);
+        if (ret) {
+                pr_err(XEN_MCELOG "Failed to get CPU info\n");
+                kfree(g_physinfo);
+                return ret;
+        }
+
+        ret  = bind_virq_to_irqhandler(VIRQ_MCA, 0,
+                                       xen_mce_interrupt, 0, "mce", NULL);
+        if (ret < 0) {
+                pr_err(XEN_MCELOG "Failed to bind virq\n");
+                kfree(g_physinfo);
+                return ret;
+        }
+
+        return 0;
+}
+
+static int __init xen_late_init_mcelog(void)
+{
+        /* Only DOM0 is responsible for MCE logging */
+        if (xen_initial_domain()) {
+                /* register character device /dev/mcelog for xen mcelog */
+                if (misc_register(&xen_mce_chrdev_device))
+                        return -ENODEV;
+                return bind_virq_for_mce();
+        }
+
+        return -ENODEV;
+}
+device_initcall(xen_late_init_mcelog);
diff --git a/include/linux/miscdevice.h b/include/linux/miscdevice.h
index 0549d2115507..e0deeb2cc939 100644
--- a/include/linux/miscdevice.h
+++ b/include/linux/miscdevice.h
@@ -35,6 +35,7 @@
 #define MPT_MINOR               220
 #define MPT2SAS_MINOR           221
 #define UINPUT_MINOR            223
+#define MISC_MCELOG_MINOR       227
 #define HPET_MINOR              228
 #define FUSE_MINOR              229
 #define KVM_MINOR               232
diff --git a/include/xen/interface/xen-mca.h b/include/xen/interface/xen-mca.h
new file mode 100644
index 000000000000..73a4ea714d93
--- /dev/null
+++ b/include/xen/interface/xen-mca.h
@@ -0,0 +1,385 @@
+/******************************************************************************
+ * arch-x86/mca.h
+ * Guest OS machine check interface to x86 Xen.
+ *
+ * Contributed by Advanced Micro Devices, Inc.
+ * Author: Christoph Egger <Christoph.Egger@amd.com>
+ *
+ * Updated by Intel Corporation
+ * Author: Liu, Jinsong <jinsong.liu@intel.com>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+#ifndef __XEN_PUBLIC_ARCH_X86_MCA_H__
+#define __XEN_PUBLIC_ARCH_X86_MCA_H__
+
+/* Hypercall */
+#define __HYPERVISOR_mca __HYPERVISOR_arch_0
+
+#define XEN_MCA_INTERFACE_VERSION       0x01ecc003
+
+/* IN: Dom0 calls hypercall to retrieve nonurgent error log entry */
+#define XEN_MC_NONURGENT        0x1
+/* IN: Dom0 calls hypercall to retrieve urgent error log entry */
+#define XEN_MC_URGENT           0x2
+/* IN: Dom0 acknowledges previosly-fetched error log entry */
+#define XEN_MC_ACK              0x4
+
+/* OUT: All is ok */
+#define XEN_MC_OK               0x0
+/* OUT: Domain could not fetch data. */
+#define XEN_MC_FETCHFAILED      0x1
+/* OUT: There was no machine check data to fetch. */
+#define XEN_MC_NODATA           0x2
+
+#ifndef __ASSEMBLY__
+/* vIRQ injected to Dom0 */
+#define VIRQ_MCA VIRQ_ARCH_0
+
+/*
+ * mc_info entry types
+ * mca machine check info are recorded in mc_info entries.
+ * when fetch mca info, it can use MC_TYPE_... to distinguish
+ * different mca info.
+ */
+#define MC_TYPE_GLOBAL          0
+#define MC_TYPE_BANK            1
+#define MC_TYPE_EXTENDED        2
+#define MC_TYPE_RECOVERY        3
+
+struct mcinfo_common {
+        uint16_t type; /* structure type */
+        uint16_t size; /* size of this struct in bytes */
+};
+
+#define MC_FLAG_CORRECTABLE     (1 << 0)
+#define MC_FLAG_UNCORRECTABLE   (1 << 1)
+#define MC_FLAG_RECOVERABLE     (1 << 2)
+#define MC_FLAG_POLLED          (1 << 3)
+#define MC_FLAG_RESET           (1 << 4)
+#define MC_FLAG_CMCI            (1 << 5)
+#define MC_FLAG_MCE             (1 << 6)
+
+/* contains x86 global mc information */
+struct mcinfo_global {
+        struct mcinfo_common common;
+
+        uint16_t mc_domid; /* running domain at the time in error */
+        uint16_t mc_vcpuid; /* virtual cpu scheduled for mc_domid */
+        uint32_t mc_socketid; /* physical socket of the physical core */
+        uint16_t mc_coreid; /* physical impacted core */
+        uint16_t mc_core_threadid; /* core thread of physical core */
+        uint32_t mc_apicid;
+        uint32_t mc_flags;
+        uint64_t mc_gstatus; /* global status */
+};
+
+/* contains x86 bank mc information */
+struct mcinfo_bank {
+        struct mcinfo_common common;
+
+        uint16_t mc_bank; /* bank nr */
+        uint16_t mc_domid; /* domain referenced by mc_addr if valid */
+        uint64_t mc_status; /* bank status */
+        uint64_t mc_addr; /* bank address */
+        uint64_t mc_misc;
+        uint64_t mc_ctrl2;
+        uint64_t mc_tsc;
+};
+
+struct mcinfo_msr {
+        uint64_t reg; /* MSR */
+        uint64_t value; /* MSR value */
+};
+
+/* contains mc information from other or additional mc MSRs */
+struct mcinfo_extended {
+        struct mcinfo_common common;
+        uint32_t mc_msrs; /* Number of msr with valid values. */
+        /*
+         * Currently Intel extended MSR (32/64) include all gp registers
+         * and E(R)FLAGS, E(R)IP, E(R)MISC, up to 11/19 of them might be
+         * useful at present. So expand this array to 16/32 to leave room.
+         */
+        struct mcinfo_msr mc_msr[sizeof(void *) * 4];
+};
+
+/* Recovery Action flags. Giving recovery result information to DOM0 */
+
+/* Xen takes successful recovery action, the error is recovered */
+#define REC_ACTION_RECOVERED (0x1 << 0)
+/* No action is performed by XEN */
+#define REC_ACTION_NONE (0x1 << 1)
+/* It's possible DOM0 might take action ownership in some case */
+#define REC_ACTION_NEED_RESET (0x1 << 2)
+
+/*
+ * Different Recovery Action types, if the action is performed successfully,
+ * REC_ACTION_RECOVERED flag will be returned.
+ */
+
+/* Page Offline Action */
+#define MC_ACTION_PAGE_OFFLINE (0x1 << 0)
+/* CPU offline Action */
+#define MC_ACTION_CPU_OFFLINE (0x1 << 1)
+/* L3 cache disable Action */
+#define MC_ACTION_CACHE_SHRINK (0x1 << 2)
+
+/*
+ * Below interface used between XEN/DOM0 for passing XEN's recovery action
+ * information to DOM0.
+ */
+struct page_offline_action {
+        /* Params for passing the offlined page number to DOM0 */
+        uint64_t mfn;
+        uint64_t status;
+};
+
+struct cpu_offline_action {
+        /* Params for passing the identity of the offlined CPU to DOM0 */
+        uint32_t mc_socketid;
+        uint16_t mc_coreid;
+        uint16_t mc_core_threadid;
+};
+
+#define MAX_UNION_SIZE 16
+struct mcinfo_recovery {
+        struct mcinfo_common common;
+        uint16_t mc_bank; /* bank nr */
+        uint8_t action_flags;
+        uint8_t action_types;
+        union {
+                struct page_offline_action page_retire;
+                struct cpu_offline_action cpu_offline;
+                uint8_t pad[MAX_UNION_SIZE];
+        } action_info;
+};
+
+
+#define MCINFO_MAXSIZE 768
+struct mc_info {
+        /* Number of mcinfo_* entries in mi_data */
+        uint32_t mi_nentries;
+        uint32_t flags;
+        uint64_t mi_data[(MCINFO_MAXSIZE - 1) / 8];
+};
+DEFINE_GUEST_HANDLE_STRUCT(mc_info);
+
+#define __MC_MSR_ARRAYSIZE 8
+#define __MC_MSR_MCGCAP 0
+#define __MC_NMSRS 1
+#define MC_NCAPS 7
+struct mcinfo_logical_cpu {
+        uint32_t mc_cpunr;
+        uint32_t mc_chipid;
+        uint16_t mc_coreid;
+        uint16_t mc_threadid;
+        uint32_t mc_apicid;
+        uint32_t mc_clusterid;
+        uint32_t mc_ncores;
+        uint32_t mc_ncores_active;
+        uint32_t mc_nthreads;
+        uint32_t mc_cpuid_level;
+        uint32_t mc_family;
+        uint32_t mc_vendor;
+        uint32_t mc_model;
+        uint32_t mc_step;
+        char mc_vendorid[16];
+        char mc_brandid[64];
+        uint32_t mc_cpu_caps[MC_NCAPS];
+        uint32_t mc_cache_size;
+        uint32_t mc_cache_alignment;
+        uint32_t mc_nmsrvals;
+        struct mcinfo_msr mc_msrvalues[__MC_MSR_ARRAYSIZE];
+};
+DEFINE_GUEST_HANDLE_STRUCT(mcinfo_logical_cpu);
+
+/*
+ * Prototype:
+ *    uint32_t x86_mcinfo_nentries(struct mc_info *mi);
+ */
+#define x86_mcinfo_nentries(_mi)    \
+        ((_mi)->mi_nentries)
+/*
+ * Prototype:
+ *    struct mcinfo_common *x86_mcinfo_first(struct mc_info *mi);
+ */
+#define x86_mcinfo_first(_mi)       \
+        ((struct mcinfo_common *)(_mi)->mi_data)
+/*
+ * Prototype:
+ *    struct mcinfo_common *x86_mcinfo_next(struct mcinfo_common *mic);
+ */
+#define x86_mcinfo_next(_mic)       \
+        ((struct mcinfo_common *)((uint8_t *)(_mic) + (_mic)->size))
+
+/*
+ * Prototype:
+ *    void x86_mcinfo_lookup(void *ret, struct mc_info *mi, uint16_t type);
+ */
+static inline void x86_mcinfo_lookup(struct mcinfo_common **ret,
+                                     struct mc_info *mi, uint16_t type)
+{
+        uint32_t i;
+        struct mcinfo_common *mic;
+        bool found = 0;
+
+        if (!ret || !mi)
+                return;
+
+        mic = x86_mcinfo_first(mi);
+        for (i = 0; i < x86_mcinfo_nentries(mi); i++) {
+                if (mic->type == type) {
+                        found = 1;
+                        break;
+                }
+                mic = x86_mcinfo_next(mic);
+        }
+
+        *ret = found ? mic : NULL;
+}
+
+/*
+ * Fetch machine check data from hypervisor.
+ */
+#define XEN_MC_fetch            1
+struct xen_mc_fetch {
+        /*
+         * IN: XEN_MC_NONURGENT, XEN_MC_URGENT,
+         * XEN_MC_ACK if ack'king an earlier fetch
+         * OUT: XEN_MC_OK, XEN_MC_FETCHAILED, XEN_MC_NODATA
+         */
+        uint32_t flags;
+        uint32_t _pad0;
+        /* OUT: id for ack, IN: id we are ack'ing */
+        uint64_t fetch_id;
+
+        /* OUT variables. */
+        GUEST_HANDLE(mc_info) data;
+};
+DEFINE_GUEST_HANDLE_STRUCT(xen_mc_fetch);
+
+
+/*
+ * This tells the hypervisor to notify a DomU about the machine check error
+ */
+#define XEN_MC_notifydomain     2
+struct xen_mc_notifydomain {
+        /* IN variables */
+        uint16_t mc_domid; /* The unprivileged domain to notify */
+        uint16_t mc_vcpuid; /* The vcpu in mc_domid to notify */
+
+        /* IN/OUT variables */
+        uint32_t flags;
+};
+DEFINE_GUEST_HANDLE_STRUCT(xen_mc_notifydomain);
+
+#define XEN_MC_physcpuinfo      3
+struct xen_mc_physcpuinfo {
+        /* IN/OUT */
+        uint32_t ncpus;
+        uint32_t _pad0;
+        /* OUT */
+        GUEST_HANDLE(mcinfo_logical_cpu) info;
+};
+
+#define XEN_MC_msrinject        4
+#define MC_MSRINJ_MAXMSRS       8
+struct xen_mc_msrinject {
+        /* IN */
+        uint32_t mcinj_cpunr; /* target processor id */
+        uint32_t mcinj_flags; /* see MC_MSRINJ_F_* below */
+        uint32_t mcinj_count; /* 0 .. count-1 in array are valid */
+        uint32_t _pad0;
+        struct mcinfo_msr mcinj_msr[MC_MSRINJ_MAXMSRS];
+};
+
+/* Flags for mcinj_flags above; bits 16-31 are reserved */
+#define MC_MSRINJ_F_INTERPOSE   0x1
+
+#define XEN_MC_mceinject        5
+struct xen_mc_mceinject {
+        unsigned int mceinj_cpunr; /* target processor id */
+};
+
+struct xen_mc {
+        uint32_t cmd;
+        uint32_t interface_version; /* XEN_MCA_INTERFACE_VERSION */
+        union {
+                struct xen_mc_fetch        mc_fetch;
+                struct xen_mc_notifydomain mc_notifydomain;
+                struct xen_mc_physcpuinfo  mc_physcpuinfo;
+                struct xen_mc_msrinject    mc_msrinject;
+                struct xen_mc_mceinject    mc_mceinject;
+        } u;
+};
+DEFINE_GUEST_HANDLE_STRUCT(xen_mc);
+
+/* Fields are zero when not available */
+struct xen_mce {
+        __u64 status;
+        __u64 misc;
+        __u64 addr;
+        __u64 mcgstatus;
+        __u64 ip;
+        __u64 tsc;      /* cpu time stamp counter */
+        __u64 time;     /* wall time_t when error was detected */
+        __u8  cpuvendor;        /* cpu vendor as encoded in system.h */
+        __u8  inject_flags;     /* software inject flags */
+        __u16  pad;
+        __u32 cpuid;    /* CPUID 1 EAX */
+        __u8  cs;               /* code segment */
+        __u8  bank;     /* machine check bank */
+        __u8  cpu;      /* cpu number; obsolete; use extcpu now */
+        __u8  finished;   /* entry is valid */
+        __u32 extcpu;   /* linux cpu number that detected the error */
+        __u32 socketid; /* CPU socket ID */
+        __u32 apicid;   /* CPU initial apic ID */
+        __u64 mcgcap;   /* MCGCAP MSR: machine check capabilities of CPU */
+};
+
+/*
+ * This structure contains all data related to the MCE log.  Also
+ * carries a signature to make it easier to find from external
+ * debugging tools.  Each entry is only valid when its finished flag
+ * is set.
+ */
+
+#define XEN_MCE_LOG_LEN 32
+
+struct xen_mce_log {
+        char signature[12]; /* "MACHINECHECK" */
+        unsigned len;       /* = XEN_MCE_LOG_LEN */
+        unsigned next;
+        unsigned flags;
+        unsigned recordlen;     /* length of struct xen_mce */
+        struct xen_mce entry[XEN_MCE_LOG_LEN];
+};
+
+#define XEN_MCE_OVERFLOW 0              /* bit 0 in flags means overflow */
+
+#define XEN_MCE_LOG_SIGNATURE   "MACHINECHECK"
+
+#define MCE_GET_RECORD_LEN   _IOR('M', 1, int)
+#define MCE_GET_LOG_LEN      _IOR('M', 2, int)
+#define MCE_GETCLEAR_FLAGS   _IOR('M', 3, int)
+
+#endif /* __ASSEMBLY__ */
+#endif /* __XEN_PUBLIC_ARCH_X86_MCA_H__ */