drivers/xen/pcpu.c

   1 /******************************************************************************
   2  * pcpu.c
   3  * Management physical cpu in dom0, get pcpu info and provide sys interface
   4  *
   5  * Copyright (c) 2012 Intel Corporation
   6  * Author: Liu, Jinsong <jinsong.liu@intel.com>
   7  * Author: Jiang, Yunhong <yunhong.jiang@intel.com>
   8  *
   9  * This program is free software; you can redistribute it and/or
  10  * modify it under the terms of the GNU General Public License version 2
  11  * as published by the Free Software Foundation; or, when distributed
  12  * separately from the Linux kernel or incorporated into other
  13  * software packages, subject to the following license:
  14  *
  15  * Permission is hereby granted, free of charge, to any person obtaining a copy
  16  * of this source file (the "Software"), to deal in the Software without
  17  * restriction, including without limitation the rights to use, copy, modify,
  18  * merge, publish, distribute, sublicense, and/or sell copies of the Software,
  19  * and to permit persons to whom the Software is furnished to do so, subject to
  20  * the following conditions:
  21  *
  22  * The above copyright notice and this permission notice shall be included in
  23  * all copies or substantial portions of the Software.
  24  *
  25  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
  26  * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
  27  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
  28  * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
  29  * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
  30  * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
  31  * IN THE SOFTWARE.
  32  */
  33
  34 #define pr_fmt(fmt) "xen_cpu: " fmt
  35
  36 #include <linux/interrupt.h>
  37 #include <linux/spinlock.h>
  38 #include <linux/cpu.h>
  39 #include <linux/stat.h>
  40 #include <linux/capability.h>
  41
  42 #include <xen/xen.h>
  43 #include <xen/acpi.h>
  44 #include <xen/xenbus.h>
  45 #include <xen/events.h>
  46 #include <xen/interface/platform.h>
  47 #include <asm/xen/hypervisor.h>
  48 #include <asm/xen/hypercall.h>
  49
  50
  51 /*
  52  * @cpu_id: Xen physical cpu logic number
  53  * @flags: Xen physical cpu status flag
  54  * - XEN_PCPU_FLAGS_ONLINE: cpu is online
  55  * - XEN_PCPU_FLAGS_INVALID: cpu is not present
  56  */
  57 struct pcpu {
  58         struct list_head list;
  59         struct device dev;
  60         uint32_t cpu_id;
  61         uint32_t flags;
  62 };
  63
  64 static struct bus_type xen_pcpu_subsys = {
  65         .name = "xen_cpu",
  66         .dev_name = "xen_cpu",
  67 };
  68
  69 static DEFINE_MUTEX(xen_pcpu_lock);
  70
  71 static LIST_HEAD(xen_pcpus);
  72
  73 static int xen_pcpu_down(uint32_t cpu_id)
  74 {
  75         struct xen_platform_op op = {
  76                 .cmd                    = XENPF_cpu_offline,
  77                 .interface_version      = XENPF_INTERFACE_VERSION,
  78                 .u.cpu_ol.cpuid         = cpu_id,
  79         };
  80
  81         return HYPERVISOR_dom0_op(&op);
  82 }
  83
  84 static int xen_pcpu_up(uint32_t cpu_id)
  85 {
  86         struct xen_platform_op op = {
  87                 .cmd                    = XENPF_cpu_online,
  88                 .interface_version      = XENPF_INTERFACE_VERSION,
  89                 .u.cpu_ol.cpuid         = cpu_id,
  90         };
  91
  92         return HYPERVISOR_dom0_op(&op);
  93 }
  94
  95 static ssize_t show_online(struct device *dev,
  96                            struct device_attribute *attr,
  97                            char *buf)
  98 {
  99         struct pcpu *cpu = container_of(dev, struct pcpu, dev);
 100
 101         return sprintf(buf, "%u\n", !!(cpu->flags & XEN_PCPU_FLAGS_ONLINE));
 102 }
 103
 104 static ssize_t __ref store_online(struct device *dev,
 105                                   struct device_attribute *attr,
 106                                   const char *buf, size_t count)
 107 {
 108         struct pcpu *pcpu = container_of(dev, struct pcpu, dev);
 109         unsigned long long val;
 110         ssize_t ret;
 111
 112         if (!capable(CAP_SYS_ADMIN))
 113                 return -EPERM;
 114
 115         if (kstrtoull(buf, 0, &val) < 0)
 116                 return -EINVAL;
 117
 118         switch (val) {
 119         case 0:
 120                 ret = xen_pcpu_down(pcpu->cpu_id);
 121                 break;
 122         case 1:
 123                 ret = xen_pcpu_up(pcpu->cpu_id);
 124                 break;
 125         default:
 126                 ret = -EINVAL;
 127         }
 128
 129         if (ret >= 0)
 130                 ret = count;
 131         return ret;
 132 }
 133 static DEVICE_ATTR(online, S_IRUGO | S_IWUSR, show_online, store_online);
 134
 135 static struct attribute *pcpu_dev_attrs[] = {
 136         &dev_attr_online.attr,
 137         NULL
 138 };
 139
 140 static umode_t pcpu_dev_is_visible(struct kobject *kobj,
 141                                    struct attribute *attr, int idx)
 142 {
 143         struct device *dev = kobj_to_dev(kobj);
 144         /*
 145          * Xen never offline cpu0 due to several restrictions
 146          * and assumptions. This basically doesn't add a sys control
 147          * to user, one cannot attempt to offline BSP.
 148          */
 149         return dev->id ? attr->mode : 0;
 150 }
 151
 152 static const struct attribute_group pcpu_dev_group = {
 153         .attrs = pcpu_dev_attrs,
 154         .is_visible = pcpu_dev_is_visible,
 155 };
 156
 157 static const struct attribute_group *pcpu_dev_groups[] = {
 158         &pcpu_dev_group,
 159         NULL
 160 };
 161
 162 static bool xen_pcpu_online(uint32_t flags)
 163 {
 164         return !!(flags & XEN_PCPU_FLAGS_ONLINE);
 165 }
 166
 167 static void pcpu_online_status(struct xenpf_pcpuinfo *info,
 168                                struct pcpu *pcpu)
 169 {
 170         if (xen_pcpu_online(info->flags) &&
 171            !xen_pcpu_online(pcpu->flags)) {
 172                 /* the pcpu is onlined */
 173                 pcpu->flags |= XEN_PCPU_FLAGS_ONLINE;
 174                 kobject_uevent(&pcpu->dev.kobj, KOBJ_ONLINE);
 175         } else if (!xen_pcpu_online(info->flags) &&
 176                     xen_pcpu_online(pcpu->flags)) {
 177                 /* The pcpu is offlined */
 178                 pcpu->flags &= ~XEN_PCPU_FLAGS_ONLINE;
 179                 kobject_uevent(&pcpu->dev.kobj, KOBJ_OFFLINE);
 180         }
 181 }
 182
 183 static struct pcpu *get_pcpu(uint32_t cpu_id)
 184 {
 185         struct pcpu *pcpu;
 186
 187         list_for_each_entry(pcpu, &xen_pcpus, list) {
 188                 if (pcpu->cpu_id == cpu_id)
 189                         return pcpu;
 190         }
 191
 192         return NULL;
 193 }
 194
 195 static void pcpu_release(struct device *dev)
 196 {
 197         struct pcpu *pcpu = container_of(dev, struct pcpu, dev);
 198
 199         list_del(&pcpu->list);
 200         kfree(pcpu);
 201 }
 202
 203 static void unregister_and_remove_pcpu(struct pcpu *pcpu)
 204 {
 205         struct device *dev;
 206
 207         if (!pcpu)
 208                 return;
 209
 210         dev = &pcpu->dev;
 211         /* pcpu remove would be implicitly done */
 212         device_unregister(dev);
 213 }
 214
 215 static int register_pcpu(struct pcpu *pcpu)
 216 {
 217         struct device *dev;
 218         int err = -EINVAL;
 219
 220         if (!pcpu)
 221                 return err;
 222
 223         dev = &pcpu->dev;
 224         dev->bus = &xen_pcpu_subsys;
 225         dev->id = pcpu->cpu_id;
 226         dev->release = pcpu_release;
 227         dev->groups = pcpu_dev_groups;
 228
 229         err = device_register(dev);
 230         if (err) {
 231                 pcpu_release(dev);
 232                 return err;
 233         }
 234
 235         return 0;
 236 }
 237
 238 static struct pcpu *create_and_register_pcpu(struct xenpf_pcpuinfo *info)
 239 {
 240         struct pcpu *pcpu;
 241         int err;
 242
 243         if (info->flags & XEN_PCPU_FLAGS_INVALID)
 244                 return ERR_PTR(-ENODEV);
 245
 246         pcpu = kzalloc(sizeof(struct pcpu), GFP_KERNEL);
 247         if (!pcpu)
 248                 return ERR_PTR(-ENOMEM);
 249
 250         INIT_LIST_HEAD(&pcpu->list);
 251         pcpu->cpu_id = info->xen_cpuid;
 252         pcpu->flags = info->flags;
 253
 254         /* Need hold on xen_pcpu_lock before pcpu list manipulations */
 255         list_add_tail(&pcpu->list, &xen_pcpus);
 256
 257         err = register_pcpu(pcpu);
 258         if (err) {
 259                 pr_warn("Failed to register pcpu%u\n", info->xen_cpuid);
 260                 return ERR_PTR(-ENOENT);
 261         }
 262
 263         return pcpu;
 264 }
 265
 266 /*
 267  * Caller should hold the xen_pcpu_lock
 268  */
 269 static int sync_pcpu(uint32_t cpu, uint32_t *max_cpu)
 270 {
 271         int ret;
 272         struct pcpu *pcpu = NULL;
 273         struct xenpf_pcpuinfo *info;
 274         struct xen_platform_op op = {
 275                 .cmd                   = XENPF_get_cpuinfo,
 276                 .interface_version     = XENPF_INTERFACE_VERSION,
 277                 .u.pcpu_info.xen_cpuid = cpu,
 278         };
 279
 280         ret = HYPERVISOR_dom0_op(&op);
 281         if (ret)
 282                 return ret;
 283
 284         info = &op.u.pcpu_info;
 285         if (max_cpu)
 286                 *max_cpu = info->max_present;
 287
 288         pcpu = get_pcpu(cpu);
 289
 290         /*
 291          * Only those at cpu present map has its sys interface.
 292          */
 293         if (info->flags & XEN_PCPU_FLAGS_INVALID) {
 294                 unregister_and_remove_pcpu(pcpu);
 295                 return 0;
 296         }
 297
 298         if (!pcpu) {
 299                 pcpu = create_and_register_pcpu(info);
 300                 if (IS_ERR_OR_NULL(pcpu))
 301                         return -ENODEV;
 302         } else
 303                 pcpu_online_status(info, pcpu);
 304
 305         return 0;
 306 }
 307
 308 /*
 309  * Sync dom0's pcpu information with xen hypervisor's
 310  */
 311 static int xen_sync_pcpus(void)
 312 {
 313         /*
 314          * Boot cpu always have cpu_id 0 in xen
 315          */
 316         uint32_t cpu = 0, max_cpu = 0;
 317         int err = 0;
 318         struct pcpu *pcpu, *tmp;
 319
 320         mutex_lock(&xen_pcpu_lock);
 321
 322         while (!err && (cpu <= max_cpu)) {
 323                 err = sync_pcpu(cpu, &max_cpu);
 324                 cpu++;
 325         }
 326
 327         if (err)
 328                 list_for_each_entry_safe(pcpu, tmp, &xen_pcpus, list)
 329                         unregister_and_remove_pcpu(pcpu);
 330
 331         mutex_unlock(&xen_pcpu_lock);
 332
 333         return err;
 334 }
 335
 336 static void xen_pcpu_work_fn(struct work_struct *work)
 337 {
 338         xen_sync_pcpus();
 339 }
 340 static DECLARE_WORK(xen_pcpu_work, xen_pcpu_work_fn);
 341
 342 static irqreturn_t xen_pcpu_interrupt(int irq, void *dev_id)
 343 {
 344         schedule_work(&xen_pcpu_work);
 345         return IRQ_HANDLED;
 346 }
 347
 348 /* Sync with Xen hypervisor after cpu hotadded */
 349 void xen_pcpu_hotplug_sync(void)
 350 {
 351         schedule_work(&xen_pcpu_work);
 352 }
 353 EXPORT_SYMBOL_GPL(xen_pcpu_hotplug_sync);
 354
 355 /*
 356  * For hypervisor presented cpu, return logic cpu id;
 357  * For hypervisor non-presented cpu, return -ENODEV.
 358  */
 359 int xen_pcpu_id(uint32_t acpi_id)
 360 {
 361         int cpu_id = 0, max_id = 0;
 362         struct xen_platform_op op;
 363
 364         op.cmd = XENPF_get_cpuinfo;
 365         while (cpu_id <= max_id) {
 366                 op.u.pcpu_info.xen_cpuid = cpu_id;
 367                 if (HYPERVISOR_dom0_op(&op)) {
 368                         cpu_id++;
 369                         continue;
 370                 }
 371
 372                 if (acpi_id == op.u.pcpu_info.acpi_id)
 373                         return cpu_id;
 374                 if (op.u.pcpu_info.max_present > max_id)
 375                         max_id = op.u.pcpu_info.max_present;
 376                 cpu_id++;
 377         }
 378
 379         return -ENODEV;
 380 }
 381 EXPORT_SYMBOL_GPL(xen_pcpu_id);
 382
 383 static int __init xen_pcpu_init(void)
 384 {
 385         int irq, ret;
 386
 387         if (!xen_initial_domain())
 388                 return -ENODEV;
 389
 390         irq = bind_virq_to_irqhandler(VIRQ_PCPU_STATE, 0,
 391                                       xen_pcpu_interrupt, 0,
 392                                       "xen-pcpu", NULL);
 393         if (irq < 0) {
 394                 pr_warn("Failed to bind pcpu virq\n");
 395                 return irq;
 396         }
 397
 398         ret = subsys_system_register(&xen_pcpu_subsys, NULL);
 399         if (ret) {
 400                 pr_warn("Failed to register pcpu subsys\n");
 401                 goto err1;
 402         }
 403
 404         ret = xen_sync_pcpus();
 405         if (ret) {
 406                 pr_warn("Failed to sync pcpu info\n");
 407                 goto err2;
 408         }
 409
 410         return 0;
 411
 412 err2:
 413         bus_unregister(&xen_pcpu_subsys);
 414 err1:
 415         unbind_from_irqhandler(irq, NULL);
 416         return ret;
 417 }
 418 arch_initcall(xen_pcpu_init);