1 files changed, 143 insertions, 85 deletions

diff --git a/arch/powerpc/kvm/book3s_hv_builtin.c b/arch/powerpc/kvm/book3s_hv_builtin.c
index a795a13f4a7..7cde8a66520 100644
--- a/arch/powerpc/kvm/book3s_hv_builtin.c
+++ b/arch/powerpc/kvm/book3s_hv_builtin.c
@@ -6,6 +6,7 @@
  * published by the Free Software Foundation.
  */
 
+#include <linux/cpu.h>
 #include <linux/kvm_host.h>
 #include <linux/preempt.h>
 #include <linux/export.h>
@@ -13,47 +14,34 @@
 #include <linux/spinlock.h>
 #include <linux/bootmem.h>
 #include <linux/init.h>
+#include <linux/memblock.h>
+#include <linux/sizes.h>
 
 #include <asm/cputable.h>
 #include <asm/kvm_ppc.h>
 #include <asm/kvm_book3s.h>
 
+#include "book3s_hv_cma.h"
 /*
- * This maintains a list of RMAs (real mode areas) for KVM guests to use.
+ * Hash page table alignment on newer cpus(CPU_FTR_ARCH_206)
+ * should be power of 2.
+ */
+#define HPT_ALIGN_PAGES		((1 << 18) >> PAGE_SHIFT) /* 256k */
+/*
+ * By default we reserve 5% of memory for hash pagetable allocation.
+ */
+static unsigned long kvm_cma_resv_ratio = 5;
+/*
+ * We allocate RMAs (real mode areas) for KVM guests from the KVM CMA area.
  * Each RMA has to be physically contiguous and of a size that the
  * hardware supports.  PPC970 and POWER7 support 64MB, 128MB and 256MB,
  * and other larger sizes.  Since we are unlikely to be allocate that
  * much physically contiguous memory after the system is up and running,
- * we preallocate a set of RMAs in early boot for KVM to use.
+ * we preallocate a set of RMAs in early boot using CMA.
+ * should be power of 2.
  */
-static unsigned long kvm_rma_size = 64 << 20;	/* 64MB */
-static unsigned long kvm_rma_count;
-
-static int __init early_parse_rma_size(char *p)
-{
-	if (!p)
-		return 1;
-
-	kvm_rma_size = memparse(p, &p);
-
-	return 0;
-}
-early_param("kvm_rma_size", early_parse_rma_size);
-
-static int __init early_parse_rma_count(char *p)
-{
-	if (!p)
-		return 1;
-
-	kvm_rma_count = simple_strtoul(p, NULL, 0);
-
-	return 0;
-}
-early_param("kvm_rma_count", early_parse_rma_count);
-
-static struct kvmppc_rma_info *rma_info;
-static LIST_HEAD(free_rmas);
-static DEFINE_SPINLOCK(rma_lock);
+unsigned long kvm_rma_pages = (1 << 27) >> PAGE_SHIFT;	/* 128MB */
+EXPORT_SYMBOL_GPL(kvm_rma_pages);
 
 /* Work out RMLS (real mode limit selector) field value for a given RMA size.
    Assumes POWER7 or PPC970. */
@@ -81,76 +69,146 @@ static inline int lpcr_rmls(unsigned long rma_size)
 	}
 }
 
-/*
- * Called at boot time while the bootmem allocator is active,
- * to allocate contiguous physical memory for the real memory
- * areas for guests.
- */
-void __init kvm_rma_init(void)
+static int __init early_parse_rma_size(char *p)
 {
-	unsigned long i;
-	unsigned long j, npages;
-	void *rma;
-	struct page *pg;
-
-	/* Only do this on PPC970 in HV mode */
-	if (!cpu_has_feature(CPU_FTR_HVMODE) ||
-	    !cpu_has_feature(CPU_FTR_ARCH_201))
-		return;
-
-	if (!kvm_rma_size || !kvm_rma_count)
-		return;
+	unsigned long kvm_rma_size;
 
-	/* Check that the requested size is one supported in hardware */
+	pr_debug("%s(%s)\n", __func__, p);
+	if (!p)
+		return -EINVAL;
+	kvm_rma_size = memparse(p, &p);
+	/*
+	 * Check that the requested size is one supported in hardware
+	 */
 	if (lpcr_rmls(kvm_rma_size) < 0) {
 		pr_err("RMA size of 0x%lx not supported\n", kvm_rma_size);
-		return;
-	}
-
-	npages = kvm_rma_size >> PAGE_SHIFT;
-	rma_info = alloc_bootmem(kvm_rma_count * sizeof(struct kvmppc_rma_info));
-	for (i = 0; i < kvm_rma_count; ++i) {
-		rma = alloc_bootmem_align(kvm_rma_size, kvm_rma_size);
-		pr_info("Allocated KVM RMA at %p (%ld MB)\n", rma,
-			kvm_rma_size >> 20);
-		rma_info[i].base_virt = rma;
-		rma_info[i].base_pfn = __pa(rma) >> PAGE_SHIFT;
-		rma_info[i].npages = npages;
-		list_add_tail(&rma_info[i].list, &free_rmas);
-		atomic_set(&rma_info[i].use_count, 0);
-
-		pg = pfn_to_page(rma_info[i].base_pfn);
-		for (j = 0; j < npages; ++j) {
-			atomic_inc(&pg->_count);
-			++pg;
-		}
+		return -EINVAL;
 	}
+	kvm_rma_pages = kvm_rma_size >> PAGE_SHIFT;
+	return 0;
 }
+early_param("kvm_rma_size", early_parse_rma_size);
 
-struct kvmppc_rma_info *kvm_alloc_rma(void)
+struct kvm_rma_info *kvm_alloc_rma()
 {
-	struct kvmppc_rma_info *ri;
-
-	ri = NULL;
-	spin_lock(&rma_lock);
-	if (!list_empty(&free_rmas)) {
-		ri = list_first_entry(&free_rmas, struct kvmppc_rma_info, list);
-		list_del(&ri->list);
-		atomic_inc(&ri->use_count);
-	}
-	spin_unlock(&rma_lock);
+	struct page *page;
+	struct kvm_rma_info *ri;
+
+	ri = kmalloc(sizeof(struct kvm_rma_info), GFP_KERNEL);
+	if (!ri)
+		return NULL;
+	page = kvm_alloc_cma(kvm_rma_pages, kvm_rma_pages);
+	if (!page)
+		goto err_out;
+	atomic_set(&ri->use_count, 1);
+	ri->base_pfn = page_to_pfn(page);
 	return ri;
+err_out:
+	kfree(ri);
+	return NULL;
 }
 EXPORT_SYMBOL_GPL(kvm_alloc_rma);
 
-void kvm_release_rma(struct kvmppc_rma_info *ri)
+void kvm_release_rma(struct kvm_rma_info *ri)
 {
 	if (atomic_dec_and_test(&ri->use_count)) {
-		spin_lock(&rma_lock);
-		list_add_tail(&ri->list, &free_rmas);
-		spin_unlock(&rma_lock);
-
+		kvm_release_cma(pfn_to_page(ri->base_pfn), kvm_rma_pages);
+		kfree(ri);
 	}
 }
 EXPORT_SYMBOL_GPL(kvm_release_rma);
 
+static int __init early_parse_kvm_cma_resv(char *p)
+{
+	pr_debug("%s(%s)\n", __func__, p);
+	if (!p)
+		return -EINVAL;
+	return kstrtoul(p, 0, &kvm_cma_resv_ratio);
+}
+early_param("kvm_cma_resv_ratio", early_parse_kvm_cma_resv);
+
+struct page *kvm_alloc_hpt(unsigned long nr_pages)
+{
+	unsigned long align_pages = HPT_ALIGN_PAGES;
+
+	/* Old CPUs require HPT aligned on a multiple of its size */
+	if (!cpu_has_feature(CPU_FTR_ARCH_206))
+		align_pages = nr_pages;
+	return kvm_alloc_cma(nr_pages, align_pages);
+}
+EXPORT_SYMBOL_GPL(kvm_alloc_hpt);
+
+void kvm_release_hpt(struct page *page, unsigned long nr_pages)
+{
+	kvm_release_cma(page, nr_pages);
+}
+EXPORT_SYMBOL_GPL(kvm_release_hpt);
+
+/**
+ * kvm_cma_reserve() - reserve area for kvm hash pagetable
+ *
+ * This function reserves memory from early allocator. It should be
+ * called by arch specific code once the early allocator (memblock or bootmem)
+ * has been activated and all other subsystems have already allocated/reserved
+ * memory.
+ */
+void __init kvm_cma_reserve(void)
+{
+	unsigned long align_size;
+	struct memblock_region *reg;
+	phys_addr_t selected_size = 0;
+	/*
+	 * We cannot use memblock_phys_mem_size() here, because
+	 * memblock_analyze() has not been called yet.
+	 */
+	for_each_memblock(memory, reg)
+		selected_size += memblock_region_memory_end_pfn(reg) -
+				 memblock_region_memory_base_pfn(reg);
+
+	selected_size = (selected_size * kvm_cma_resv_ratio / 100) << PAGE_SHIFT;
+	if (selected_size) {
+		pr_debug("%s: reserving %ld MiB for global area\n", __func__,
+			 (unsigned long)selected_size / SZ_1M);
+		/*
+		 * Old CPUs require HPT aligned on a multiple of its size. So for them
+		 * make the alignment as max size we could request.
+		 */
+		if (!cpu_has_feature(CPU_FTR_ARCH_206))
+			align_size = __rounddown_pow_of_two(selected_size);
+		else
+			align_size = HPT_ALIGN_PAGES << PAGE_SHIFT;
+
+		align_size = max(kvm_rma_pages << PAGE_SHIFT, align_size);
+		kvm_cma_declare_contiguous(selected_size, align_size);
+	}
+}
+
+/*
+ * When running HV mode KVM we need to block certain operations while KVM VMs
+ * exist in the system. We use a counter of VMs to track this.
+ *
+ * One of the operations we need to block is onlining of secondaries, so we
+ * protect hv_vm_count with get/put_online_cpus().
+ */
+static atomic_t hv_vm_count;
+
+void kvm_hv_vm_activated(void)
+{
+	get_online_cpus();
+	atomic_inc(&hv_vm_count);
+	put_online_cpus();
+}
+EXPORT_SYMBOL_GPL(kvm_hv_vm_activated);
+
+void kvm_hv_vm_deactivated(void)
+{
+	get_online_cpus();
+	atomic_dec(&hv_vm_count);
+	put_online_cpus();
+}
+EXPORT_SYMBOL_GPL(kvm_hv_vm_deactivated);
+
+bool kvm_hv_mode_active(void)
+{
+	return atomic_read(&hv_vm_count) != 0;
+}