xen: move per-cpu area management into common code

Centralize per-cpu area management to reduce code duplication and
enhance maintainability across architectures.

The per-cpu area management code, which is largely common among
architectures, is moved to a shared implementation in
xen/common/percpu.c. This change includes:
 * Remove percpu.c from the X86 and Arm architectures.
 * For x86, define INVALID_PERCPU_AREAS and PARK_OFFLINE_CPUS_VAR.
 * Drop the declaration of __per_cpu_offset[] from stubs.c in
   PPC and RISC-V to facilitate the build of the common per-cpu code.

No functional changes for x86.

For Arm add support of CPU_RESUME_FAILED, CPU_REMOVE and freeing of
percpu in the case when system_state != SYS_STATE_suspend, however,
there is no change in behavior for Arm at this time.

Move the asm-generic/percpu.h definitions to xen/percpu.h, except for
__per_cpu_start[] and __per_cpu_data_end[], which are moved to
common/percpu.c as they are only used in common/percpu.c.

Signed-off-by: Oleksii Kurochko <oleksii.kurochko@gmail.com>
Reviewed-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Michal Orzel <michal.orzel@amd.com>

xen/arch/arm/Makefile

@@ -39,7 +39,6 @@ obj-$(CONFIG_MEM_ACCESS) += mem_access.o
 obj-y += mm.o
 obj-y += monitor.o
 obj-y += p2m.o
-obj-y += percpu.o
 obj-y += platform.o
 obj-y += platform_hypercall.o
 obj-y += physdev.o

xen/arch/arm/include/asm/smp.h

@@ -2,6 +2,7 @@
 #define __ASM_SMP_H
 
 #ifndef __ASSEMBLY__
+#include <xen/percpu.h>
 #include <xen/cpumask.h>
 #include <asm/current.h>
 #endif
@@ -12,12 +13,6 @@ extern unsigned long smp_up_cpu;
 DECLARE_PER_CPU(cpumask_var_t, cpu_sibling_mask);
 DECLARE_PER_CPU(cpumask_var_t, cpu_core_mask);
 
-/*
- * Do we, for platform reasons, need to actually keep CPUs online when we
- * would otherwise prefer them to be off?
- */
-#define park_offline_cpus false
-
 extern void noreturn stop_cpu(void);
 
 extern int arch_smp_init(void);

xen/arch/arm/percpu.c

@@ -1,85 +0,0 @@
-/* SPDX-License-Identifier: GPL-2.0 */
-#include <xen/percpu.h>
-#include <xen/cpu.h>
-#include <xen/init.h>
-#include <xen/mm.h>
-#include <xen/rcupdate.h>
-
-unsigned long __per_cpu_offset[NR_CPUS];
-#define INVALID_PERCPU_AREA (-(long)__per_cpu_start)
-#define PERCPU_ORDER (get_order_from_bytes(__per_cpu_data_end-__per_cpu_start))
-
-void __init percpu_init_areas(void)
-{
-    unsigned int cpu;
-    for ( cpu = 1; cpu < NR_CPUS; cpu++ )
-        __per_cpu_offset[cpu] = INVALID_PERCPU_AREA;
-}
-
-static int init_percpu_area(unsigned int cpu)
-{
-    char *p;
-    if ( __per_cpu_offset[cpu] != INVALID_PERCPU_AREA )
-        return -EBUSY;
-    if ( (p = alloc_xenheap_pages(PERCPU_ORDER, 0)) == NULL )
-        return -ENOMEM;
-    memset(p, 0, __per_cpu_data_end - __per_cpu_start);
-    __per_cpu_offset[cpu] = p - __per_cpu_start;
-    return 0;
-}
-
-struct free_info {
-    unsigned int cpu;
-    struct rcu_head rcu;
-};
-static DEFINE_PER_CPU(struct free_info, free_info);
-
-static void _free_percpu_area(struct rcu_head *head)
-{
-    struct free_info *info = container_of(head, struct free_info, rcu);
-    unsigned int cpu = info->cpu;
-    char *p = __per_cpu_start + __per_cpu_offset[cpu];
-    free_xenheap_pages(p, PERCPU_ORDER);
-    __per_cpu_offset[cpu] = INVALID_PERCPU_AREA;
-}
-
-static void free_percpu_area(unsigned int cpu)
-{
-    struct free_info *info = &per_cpu(free_info, cpu);
-    info->cpu = cpu;
-    call_rcu(&info->rcu, _free_percpu_area);
-}
-
-static int cpu_percpu_callback(
-    struct notifier_block *nfb, unsigned long action, void *hcpu)
-{
-    unsigned int cpu = (unsigned long)hcpu;
-    int rc = 0;
-
-    switch ( action )
-    {
-    case CPU_UP_PREPARE:
-        rc = init_percpu_area(cpu);
-        break;
-    case CPU_UP_CANCELED:
-    case CPU_DEAD:
-        free_percpu_area(cpu);
-        break;
-    default:
-        break;
-    }
-
-    return notifier_from_errno(rc);
-}
-
-static struct notifier_block cpu_percpu_nfb = {
-    .notifier_call = cpu_percpu_callback,
-    .priority = 100 /* highest priority */
-};
-
-static int __init percpu_presmp_init(void)
-{
-    register_cpu_notifier(&cpu_percpu_nfb);
-    return 0;
-}
-presmp_initcall(percpu_presmp_init);

xen/arch/ppc/include/asm/smp.h

@@ -7,10 +7,4 @@
 DECLARE_PER_CPU(cpumask_var_t, cpu_sibling_mask);
 DECLARE_PER_CPU(cpumask_var_t, cpu_core_mask);
 
-/*
- * Do we, for platform reasons, need to actually keep CPUs online when we
- * would otherwise prefer them to be off?
- */
-#define park_offline_cpus false
-
 #endif

xen/arch/ppc/stubs.c

@@ -141,7 +141,6 @@ void smp_send_state_dump(unsigned int cpu)
 /* domain.c */
 
 DEFINE_PER_CPU(struct vcpu *, curr_vcpu);
-unsigned long __per_cpu_offset[NR_CPUS];
 
 void context_switch(struct vcpu *prev, struct vcpu *next)
 {

xen/arch/riscv/include/asm/smp.h

@@ -10,12 +10,6 @@
 DECLARE_PER_CPU(cpumask_var_t, cpu_sibling_mask);
 DECLARE_PER_CPU(cpumask_var_t, cpu_core_mask);
 
-/*
- * Do we, for platform reasons, need to actually keep CPUs online when we
- * would otherwise prefer them to be off?
- */
-#define park_offline_cpus false
-
 /*
  * Mapping between Xen logical cpu index and hartid.
  */

xen/arch/riscv/stubs.c

@@ -133,7 +133,6 @@ void smp_send_state_dump(unsigned int cpu)
 /* domain.c */
 
 DEFINE_PER_CPU(struct vcpu *, curr_vcpu);
-unsigned long __per_cpu_offset[NR_CPUS];
 
 void context_switch(struct vcpu *prev, struct vcpu *next)
 {

xen/arch/x86/Makefile

@@ -54,7 +54,6 @@ obj-y += mpparse.o
 obj-y += nmi.o
 obj-y += numa.o
 obj-y += pci.o
-obj-y += percpu.o
 obj-y += physdev.o
 obj-$(CONFIG_COMPAT) += x86_64/physdev.o
 obj-$(CONFIG_X86_PSR) += psr.o

xen/arch/x86/include/asm/Makefile

@@ -1,3 +1,2 @@
 # SPDX-License-Identifier: GPL-2.0-only
 generic-y += div64.h
-generic-y += percpu.h

xen/arch/x86/include/asm/percpu.h

@@ -0,0 +1,14 @@
+#ifndef __X86_PERCPU_H__
+#define __X86_PERCPU_H__
+
+#define PARK_OFFLINE_CPUS_VAR
+
+/*
+ * Force uses of per_cpu() with an invalid area to attempt to access the
+ * middle of the non-canonical address space resulting in a #GP, rather than a
+ * possible #PF at (NULL + a little) which has security implications in the
+ * context of PV guests.
+ */
+#define INVALID_PERCPU_AREA (0x8000000000000000UL - (unsigned long)__per_cpu_start)
+
+#endif /* __X86_PERCPU_H__ */

xen/common/Makefile

@@ -31,6 +31,7 @@ obj-y += notifier.o
 obj-$(CONFIG_NUMA) += numa.o
 obj-y += page_alloc.o
 obj-y += pdx.o
+obj-y += percpu.o
 obj-$(CONFIG_PERF_COUNTERS) += perfc.o
 obj-bin-$(CONFIG_HAS_PMAP) += pmap.init.o
 obj-y += preempt.o

xen/common/percpu.c

@@ -1,20 +1,22 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
 #include <xen/percpu.h>
 #include <xen/cpu.h>
 #include <xen/init.h>
 #include <xen/mm.h>
 #include <xen/rcupdate.h>
+#include <xen/sections.h>
 
-unsigned long __per_cpu_offset[NR_CPUS];
+#ifndef INVALID_PERCPU_AREA
+#define INVALID_PERCPU_AREA (-(long)__per_cpu_start)
+#endif
 
-/*
- * Force uses of per_cpu() with an invalid area to attempt to access the
- * middle of the non-canonical address space resulting in a #GP, rather than a
- * possible #PF at (NULL + a little) which has security implications in the
- * context of PV guests.
- */
-#define INVALID_PERCPU_AREA (0x8000000000000000UL - (unsigned long)__per_cpu_start)
 #define PERCPU_ORDER get_order_from_bytes(__per_cpu_data_end - __per_cpu_start)
 
+extern char __per_cpu_start[];
+extern const char __per_cpu_data_end[];
+
+unsigned long __read_mostly __per_cpu_offset[NR_CPUS];
+
 void __init percpu_init_areas(void)
 {
     unsigned int cpu;
@@ -28,7 +30,7 @@ static int init_percpu_area(unsigned int cpu)
     char *p;
 
     if ( __per_cpu_offset[cpu] != INVALID_PERCPU_AREA )
-        return 0;
+        return park_offline_cpus ? 0 : -EBUSY;
 
     if ( (p = alloc_xenheap_pages(PERCPU_ORDER, 0)) == NULL )
         return -ENOMEM;
@@ -74,16 +76,21 @@ static int cf_check cpu_percpu_callback(
     case CPU_UP_PREPARE:
         rc = init_percpu_area(cpu);
         break;
+
     case CPU_UP_CANCELED:
     case CPU_DEAD:
     case CPU_RESUME_FAILED:
         if ( !park_offline_cpus && system_state != SYS_STATE_suspend )
             free_percpu_area(cpu);
         break;
+
     case CPU_REMOVE:
         if ( park_offline_cpus )
             free_percpu_area(cpu);
         break;
+
+    default:
+        break;
     }
 
     return notifier_from_errno(rc);

xen/include/asm-generic/percpu.h

@@ -2,29 +2,6 @@
 #ifndef __ASM_GENERIC_PERCPU_H__
 #define __ASM_GENERIC_PERCPU_H__
 
-#ifndef __ASSEMBLY__
-
-#include <xen/types.h>
-#include <asm/current.h>
-
-extern char __per_cpu_start[];
-extern const char __per_cpu_data_end[];
-extern unsigned long __per_cpu_offset[NR_CPUS];
-void percpu_init_areas(void);
-
-#define per_cpu(var, cpu)  \
-    (*RELOC_HIDE(&per_cpu__##var, __per_cpu_offset[cpu]))
-
-#define this_cpu(var) \
-    (*RELOC_HIDE(&per_cpu__##var, get_per_cpu_offset()))
-
-#define per_cpu_ptr(var, cpu)  \
-    (*RELOC_HIDE(var, __per_cpu_offset[cpu]))
-#define this_cpu_ptr(var) \
-    (*RELOC_HIDE(var, get_per_cpu_offset()))
-
-#endif
-
 #endif /* __ASM_GENERIC_PERCPU_H__ */
 
 /*

xen/include/xen/percpu.h

@@ -29,6 +29,36 @@
 
 #include <asm/percpu.h>
 
+#ifndef __ASSEMBLY__
+
+#include <xen/types.h>
+#include <asm/current.h>
+
+#ifndef PARK_OFFLINE_CPUS_VAR
+/*
+ * Do we, for platform reasons, need to actually keep CPUs online when we
+ * would otherwise prefer them to be off?
+ */
+#define park_offline_cpus false
+#endif
+
+extern unsigned long __per_cpu_offset[];
+
+#define per_cpu(var, cpu)  \
+    (*RELOC_HIDE(&per_cpu__##var, __per_cpu_offset[cpu]))
+
+#define this_cpu(var) \
+    (*RELOC_HIDE(&per_cpu__##var, get_per_cpu_offset()))
+
+#define per_cpu_ptr(var, cpu)  \
+    (*RELOC_HIDE(var, __per_cpu_offset[cpu]))
+#define this_cpu_ptr(var) \
+    (*RELOC_HIDE(var, get_per_cpu_offset()))
+
+void percpu_init_areas(void);
+
+#endif /* __ASSEMBLY__ */
+
 /* Linux compatibility. */
 #define get_cpu_var(var) this_cpu(var)
 #define put_cpu_var(var)