xen/pvh: Bootstrap PVH guest

Start PVH guest at XEN_ELFNOTE_PHYS32_ENTRY address. Setup hypercall
page, initialize boot_params, enable early page tables.

Since this stub is executed before kernel entry point we cannot use
variables in .bss which is cleared by kernel. We explicitly place
variables that are initialized here into .data.

While adjusting xen_hvm_init_shared_info() make it use cpuid_e?x()
instead of cpuid() (wherever possible).

Signed-off-by: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Reviewed-by: Juergen Gross <jgross@suse.com>

4 files changed, 277 insertions, 11 deletions

diff --git a/arch/x86/xen/Kconfig b/arch/x86/xen/Kconfig
index c7b15f3e2cf3..76b6dbd627df 100644
--- a/arch/x86/xen/Kconfig
+++ b/arch/x86/xen/Kconfig
@@ -53,5 +53,5 @@ config XEN_DEBUG_FS
 
 config XEN_PVH
 	bool "Support for running as a PVH guest"
-	depends on X86_64 && XEN && XEN_PVHVM
+	depends on XEN && XEN_PVHVM && ACPI
 	def_bool n
diff --git a/arch/x86/xen/Makefile b/arch/x86/xen/Makefile
index e47e52787d32..cb0164aee156 100644
--- a/arch/x86/xen/Makefile
+++ b/arch/x86/xen/Makefile
@@ -23,3 +23,4 @@ obj-$(CONFIG_XEN_DEBUG_FS)	+= debugfs.o
 obj-$(CONFIG_XEN_DOM0)		+= vga.o
 obj-$(CONFIG_SWIOTLB_XEN)	+= pci-swiotlb-xen.o
 obj-$(CONFIG_XEN_EFI)		+= efi.o
+obj-$(CONFIG_XEN_PVH)	 	+= xen-pvh.o
diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c
index 828f1b226f56..d2144f7c8fab 100644
--- a/arch/x86/xen/enlighten.c
+++ b/arch/x86/xen/enlighten.c
@@ -45,6 +45,7 @@
 #include <xen/interface/memory.h>
 #include <xen/interface/nmi.h>
 #include <xen/interface/xen-mca.h>
+#include <xen/interface/hvm/start_info.h>
 #include <xen/features.h>
 #include <xen/page.h>
 #include <xen/hvm.h>
@@ -176,6 +177,20 @@ struct tls_descs {
  */
 static DEFINE_PER_CPU(struct tls_descs, shadow_tls_desc);
 
+#ifdef CONFIG_XEN_PVH
+/*
+ * PVH variables.
+ *
+ * xen_pvh and pvh_bootparams need to live in data segment since they
+ * are used after startup_{32|64}, which clear .bss, are invoked.
+ */
+bool xen_pvh __attribute__((section(".data"))) = 0;
+struct boot_params pvh_bootparams __attribute__((section(".data")));
+
+struct hvm_start_info pvh_start_info;
+unsigned int pvh_start_info_sz = sizeof(pvh_start_info);
+#endif
+
 static void clamp_max_cpus(void)
 {
 #ifdef CONFIG_SMP
@@ -1656,6 +1671,90 @@ asmlinkage __visible void __init xen_start_kernel(void)
 #endif
 }
 
+#ifdef CONFIG_XEN_PVH
+static void __init init_pvh_bootparams(void)
+{
+	struct xen_memory_map memmap;
+	unsigned int i;
+	int rc;
+
+	memset(&pvh_bootparams, 0, sizeof(pvh_bootparams));
+
+	memmap.nr_entries = ARRAY_SIZE(pvh_bootparams.e820_map);
+	set_xen_guest_handle(memmap.buffer, pvh_bootparams.e820_map);
+	rc = HYPERVISOR_memory_op(XENMEM_memory_map, &memmap);
+	if (rc) {
+		xen_raw_printk("XENMEM_memory_map failed (%d)\n", rc);
+		BUG();
+	}
+
+	if (memmap.nr_entries < E820MAX - 1) {
+		pvh_bootparams.e820_map[memmap.nr_entries].addr =
+			ISA_START_ADDRESS;
+		pvh_bootparams.e820_map[memmap.nr_entries].size =
+			ISA_END_ADDRESS - ISA_START_ADDRESS;
+		pvh_bootparams.e820_map[memmap.nr_entries].type =
+			E820_RESERVED;
+		memmap.nr_entries++;
+	} else
+		xen_raw_printk("Warning: Can fit ISA range into e820\n");
+
+	sanitize_e820_map(pvh_bootparams.e820_map,
+			  ARRAY_SIZE(pvh_bootparams.e820_map),
+			  &memmap.nr_entries);
+
+	pvh_bootparams.e820_entries = memmap.nr_entries;
+	for (i = 0; i < pvh_bootparams.e820_entries; i++)
+		e820_add_region(pvh_bootparams.e820_map[i].addr,
+				pvh_bootparams.e820_map[i].size,
+				pvh_bootparams.e820_map[i].type);
+
+	pvh_bootparams.hdr.cmd_line_ptr =
+		pvh_start_info.cmdline_paddr;
+
+	/* The first module is always ramdisk. */
+	if (pvh_start_info.nr_modules) {
+		struct hvm_modlist_entry *modaddr =
+			__va(pvh_start_info.modlist_paddr);
+		pvh_bootparams.hdr.ramdisk_image = modaddr->paddr;
+		pvh_bootparams.hdr.ramdisk_size = modaddr->size;
+	}
+
+	/*
+	 * See Documentation/x86/boot.txt.
+	 *
+	 * Version 2.12 supports Xen entry point but we will use default x86/PC
+	 * environment (i.e. hardware_subarch 0).
+	 */
+	pvh_bootparams.hdr.version = 0x212;
+	pvh_bootparams.hdr.type_of_loader = (9 << 4) | 0; /* Xen loader */
+}
+
+/*
+ * This routine (and those that it might call) should not use
+ * anything that lives in .bss since that segment will be cleared later.
+ */
+void __init xen_prepare_pvh(void)
+{
+	u32 msr;
+	u64 pfn;
+
+	if (pvh_start_info.magic != XEN_HVM_START_MAGIC_VALUE) {
+		xen_raw_printk("Error: Unexpected magic value (0x%08x)\n",
+				pvh_start_info.magic);
+		BUG();
+	}
+
+	xen_pvh = 1;
+
+	msr = cpuid_ebx(xen_cpuid_base() + 2);
+	pfn = __pa(hypercall_page);
+	wrmsr_safe(msr, (u32)pfn, (u32)(pfn >> 32));
+
+	init_pvh_bootparams();
+}
+#endif
+
 void __ref xen_hvm_init_shared_info(void)
 {
 	int cpu;
@@ -1695,20 +1794,29 @@ void __ref xen_hvm_init_shared_info(void)
 static void __init init_hvm_pv_info(void)
 {
 	int major, minor;
-	uint32_t eax, ebx, ecx, edx, pages, msr, base;
-	u64 pfn;
+	uint32_t eax, ebx, ecx, edx, base;
 
 	base = xen_cpuid_base();
-	cpuid(base + 1, &eax, &ebx, &ecx, &edx);
+	eax = cpuid_eax(base + 1);
 
 	major = eax >> 16;
 	minor = eax & 0xffff;
 	printk(KERN_INFO "Xen version %d.%d.\n", major, minor);
 
-	cpuid(base + 2, &pages, &msr, &ecx, &edx);
+	xen_domain_type = XEN_HVM_DOMAIN;
 
-	pfn = __pa(hypercall_page);
-	wrmsr_safe(msr, (u32)pfn, (u32)(pfn >> 32));
+	/* PVH set up hypercall page in xen_prepare_pvh(). */
+	if (xen_pvh_domain())
+		pv_info.name = "Xen PVH";
+	else {
+		u64 pfn;
+		uint32_t msr;
+
+		pv_info.name = "Xen HVM";
+		msr = cpuid_ebx(base + 2);
+		pfn = __pa(hypercall_page);
+		wrmsr_safe(msr, (u32)pfn, (u32)(pfn >> 32));
+	}
 
 	xen_setup_features();
 
@@ -1717,10 +1825,6 @@ static void __init init_hvm_pv_info(void)
 		this_cpu_write(xen_vcpu_id, ebx);
 	else
 		this_cpu_write(xen_vcpu_id, smp_processor_id());
-
-	pv_info.name = "Xen HVM";
-
-	xen_domain_type = XEN_HVM_DOMAIN;
 }
 #endif
 
diff --git a/arch/x86/xen/xen-pvh.S b/arch/x86/xen/xen-pvh.S
new file mode 100644
index 000000000000..5e246716d58f
--- /dev/null
+++ b/arch/x86/xen/xen-pvh.S
@@ -0,0 +1,161 @@
+/*
+ * Copyright C 2016, Oracle and/or its affiliates. All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+	.code32
+	.text
+#define _pa(x)          ((x) - __START_KERNEL_map)
+
+#include <linux/elfnote.h>
+#include <linux/init.h>
+#include <linux/linkage.h>
+#include <asm/segment.h>
+#include <asm/asm.h>
+#include <asm/boot.h>
+#include <asm/processor-flags.h>
+#include <asm/msr.h>
+#include <xen/interface/elfnote.h>
+
+	__HEAD
+
+/*
+ * Entry point for PVH guests.
+ *
+ * Xen ABI specifies the following register state when we come here:
+ *
+ * - `ebx`: contains the physical memory address where the loader has placed
+ *          the boot start info structure.
+ * - `cr0`: bit 0 (PE) must be set. All the other writeable bits are cleared.
+ * - `cr4`: all bits are cleared.
+ * - `cs `: must be a 32-bit read/execute code segment with a base of ‘0’
+ *          and a limit of ‘0xFFFFFFFF’. The selector value is unspecified.
+ * - `ds`, `es`: must be a 32-bit read/write data segment with a base of
+ *               ‘0’ and a limit of ‘0xFFFFFFFF’. The selector values are all
+ *               unspecified.
+ * - `tr`: must be a 32-bit TSS (active) with a base of '0' and a limit
+ *         of '0x67'.
+ * - `eflags`: bit 17 (VM) must be cleared. Bit 9 (IF) must be cleared.
+ *             Bit 8 (TF) must be cleared. Other bits are all unspecified.
+ *
+ * All other processor registers and flag bits are unspecified. The OS is in
+ * charge of setting up it's own stack, GDT and IDT.
+ */
+
+ENTRY(pvh_start_xen)
+	cld
+
+	lgdt (_pa(gdt))
+
+	mov $(__BOOT_DS),%eax
+	mov %eax,%ds
+	mov %eax,%es
+	mov %eax,%ss
+
+	/* Stash hvm_start_info. */
+	mov $_pa(pvh_start_info), %edi
+	mov %ebx, %esi
+	mov _pa(pvh_start_info_sz), %ecx
+	shr $2,%ecx
+	rep
+	movsl
+
+	mov $_pa(early_stack_end), %esp
+
+	/* Enable PAE mode. */
+	mov %cr4, %eax
+	orl $X86_CR4_PAE, %eax
+	mov %eax, %cr4
+
+#ifdef CONFIG_X86_64
+	/* Enable Long mode. */
+	mov $MSR_EFER, %ecx
+	rdmsr
+	btsl $_EFER_LME, %eax
+	wrmsr
+
+	/* Enable pre-constructed page tables. */
+	mov $_pa(init_level4_pgt), %eax
+	mov %eax, %cr3
+	mov $(X86_CR0_PG | X86_CR0_PE), %eax
+	mov %eax, %cr0
+
+	/* Jump to 64-bit mode. */
+	ljmp $__KERNEL_CS, $_pa(1f)
+
+	/* 64-bit entry point. */
+	.code64
+1:
+	call xen_prepare_pvh
+
+	/* startup_64 expects boot_params in %rsi. */
+	mov $_pa(pvh_bootparams), %rsi
+	mov $_pa(startup_64), %rax
+	jmp *%rax
+
+#else /* CONFIG_X86_64 */
+
+	call mk_early_pgtbl_32
+
+	mov $_pa(initial_page_table), %eax
+	mov %eax, %cr3
+
+	mov %cr0, %eax
+	or $(X86_CR0_PG | X86_CR0_PE), %eax
+	mov %eax, %cr0
+
+	ljmp $__BOOT_CS, $1f
+1:
+	call xen_prepare_pvh
+	mov $_pa(pvh_bootparams), %esi
+
+	/* startup_32 doesn't expect paging and PAE to be on. */
+	ljmp $__BOOT_CS, $_pa(2f)
+2:
+	mov %cr0, %eax
+	and $~X86_CR0_PG, %eax
+	mov %eax, %cr0
+	mov %cr4, %eax
+	and $~X86_CR4_PAE, %eax
+	mov %eax, %cr4
+
+	ljmp $__BOOT_CS, $_pa(startup_32)
+#endif
+END(pvh_start_xen)
+
+	.section ".init.data","aw"
+	.balign 8
+gdt:
+	.word gdt_end - gdt_start
+	.long _pa(gdt_start)
+	.word 0
+gdt_start:
+	.quad 0x0000000000000000            /* NULL descriptor */
+	.quad 0x0000000000000000            /* reserved */
+#ifdef CONFIG_X86_64
+	.quad GDT_ENTRY(0xa09a, 0, 0xfffff) /* __KERNEL_CS */
+#else
+	.quad GDT_ENTRY(0xc09a, 0, 0xfffff) /* __KERNEL_CS */
+#endif
+	.quad GDT_ENTRY(0xc092, 0, 0xfffff) /* __KERNEL_DS */
+gdt_end:
+
+	.balign 4
+early_stack:
+	.fill 256, 1, 0
+early_stack_end:
+
+	ELFNOTE(Xen, XEN_ELFNOTE_PHYS32_ENTRY,
+	             _ASM_PTR (pvh_start_xen - __START_KERNEL_map))