Add support for FreeBSD/i386 guests under bhyve.

- Similar to the hack for bootinfo32.c in userboot, define
  _MACHINE_ELF_WANT_32BIT in the load_elf32 file handlers in userboot.
  This allows userboot to load 32-bit kernels and modules.
- Copy the SMAP generation code out of bootinfo64.c and into its own
  file so it can be shared with bootinfo32.c to pass an SMAP to the i386
  kernel.
- Use uint32_t instead of u_long when aligning module metadata in
  bootinfo32.c in userboot, as otherwise the metadata used 64-bit
  alignment which corrupted the layout.
- Populate the basemem and extmem members of the bootinfo struct passed
  to 32-bit kernels.
- Fix the 32-bit stack in userboot to start at the top of the stack
  instead of the bottom so that there is room to grow before the
  kernel switches to its own stack.
- Push a fake return address onto the 32-bit stack in addition to the
  arguments normally passed to exec() in the loader.  This return
  address is needed to convince recover_bootinfo() in the 32-bit
  locore code that it is being invoked from a "new" boot block.
- Add a routine to libvmmapi to setup a 32-bit flat mode register state
  including a GDT and TSS that is able to start the i386 kernel and
  update bhyveload to use it when booting an i386 kernel.
- Use the guest register state to determine the CPU's current instruction
  mode (32-bit vs 64-bit) and paging mode (flat, 32-bit, PAE, or long
  mode) in the instruction emulation code.  Update the gla2gpa() routine
  used when fetching instructions to handle flat mode, 32-bit paging, and
  PAE paging in addition to long mode paging.  Don't look for a REX
  prefix when the CPU is in 32-bit mode, and use the detected mode to
  enable the existing 32-bit mode code when decoding the mod r/m byte.

Reviewed by:	grehan, neel
MFC after:	1 month

2 files changed, 167 insertions, 2 deletions

diff --git a/lib/libvmmapi/vmmapi.h b/lib/libvmmapi/vmmapi.h
index 8428987..4a3db5d 100644
--- a/lib/libvmmapi/vmmapi.h
+++ b/lib/libvmmapi/vmmapi.h
@@ -111,5 +111,8 @@ int	vcpu_reset(struct vmctx *ctx, int vcpu);
 int	vm_setup_freebsd_registers(struct vmctx *ctx, int vcpu,
 				uint64_t rip, uint64_t cr3, uint64_t gdtbase,
 				uint64_t rsp);
+int	vm_setup_freebsd_registers_i386(struct vmctx *vmctx, int vcpu,
+					uint32_t eip, uint32_t gdtbase,
+					uint32_t esp);
 void	vm_setup_freebsd_gdt(uint64_t *gdtr);
 #endif	/* _VMMAPI_H_ */
diff --git a/lib/libvmmapi/vmmapi_freebsd.c b/lib/libvmmapi/vmmapi_freebsd.c
index 9bd2988..d801184 100644
--- a/lib/libvmmapi/vmmapi_freebsd.c
+++ b/lib/libvmmapi/vmmapi_freebsd.c
@@ -35,14 +35,176 @@ __FBSDID("$FreeBSD$");
 #include <machine/segments.h>
 #include <machine/vmm.h>
 
+#include <errno.h>
+#include <string.h>
+
 #include "vmmapi.h"
 
+#define	I386_TSS_SIZE		104
+
+#define	DESC_PRESENT		0x00000080
+#define	DESC_LONGMODE		0x00002000
+#define	DESC_DEF32		0x00004000
+#define	DESC_GRAN		0x00008000
 #define	DESC_UNUSABLE		0x00010000
 
 #define	GUEST_NULL_SEL		0
 #define	GUEST_CODE_SEL		1
 #define	GUEST_DATA_SEL		2
-#define	GUEST_GDTR_LIMIT	(3 * 8 - 1)
+#define	GUEST_TSS_SEL		3
+#define	GUEST_GDTR_LIMIT64	(3 * 8 - 1)
+
+static struct segment_descriptor i386_gdt[] = {
+	{},						/* NULL */
+	{ .sd_lolimit = 0xffff, .sd_type = SDT_MEMER,	/* CODE */
+	  .sd_p = 1, .sd_hilimit = 0xf, .sd_def32 = 1, .sd_gran = 1 }, 
+	{ .sd_lolimit = 0xffff, .sd_type = SDT_MEMRW,	/* DATA */
+	  .sd_p = 1, .sd_hilimit = 0xf, .sd_def32 = 1, .sd_gran = 1 },
+	{ .sd_lolimit = I386_TSS_SIZE - 1,		/* TSS */
+	  .sd_type = SDT_SYS386TSS, .sd_p = 1 }
+};
+
+/*
+ * Setup the 'vcpu' register set such that it will begin execution at
+ * 'eip' in flat mode.
+ */
+int
+vm_setup_freebsd_registers_i386(struct vmctx *vmctx, int vcpu, uint32_t eip,
+				uint32_t gdtbase, uint32_t esp)
+{
+	uint64_t cr0, rflags, desc_base;
+	uint32_t desc_access, desc_limit, tssbase;
+	uint16_t gsel;
+	struct segment_descriptor *gdt;
+	int error, tmp;
+
+	/* A 32-bit guest requires unrestricted mode. */	
+	error = vm_get_capability(vmctx, vcpu, VM_CAP_UNRESTRICTED_GUEST, &tmp);
+	if (error)
+		goto done;
+	error = vm_set_capability(vmctx, vcpu, VM_CAP_UNRESTRICTED_GUEST, 1);
+	if (error)
+		goto done;
+
+	cr0 = CR0_PE | CR0_NE;
+	if ((error = vm_set_register(vmctx, vcpu, VM_REG_GUEST_CR0, cr0)) != 0)
+		goto done;
+
+	if ((error = vm_set_register(vmctx, vcpu, VM_REG_GUEST_CR4, 0)) != 0)
+		goto done;
+
+	/*
+	 * Forcing EFER to 0 causes bhyve to clear the "IA-32e guest
+	 * mode" entry control.
+	 */
+	if ((error = vm_set_register(vmctx, vcpu, VM_REG_GUEST_EFER, 0)))
+		goto done;
+
+	gdt = vm_map_gpa(vmctx, gdtbase, 0x1000);
+	if (gdt == NULL)
+		return (EFAULT);
+	memcpy(gdt, i386_gdt, sizeof(i386_gdt));
+	desc_base = gdtbase;
+	desc_limit = sizeof(i386_gdt) - 1;
+	error = vm_set_desc(vmctx, vcpu, VM_REG_GUEST_GDTR,
+			    desc_base, desc_limit, 0);
+	if (error != 0)
+		goto done;
+
+	/* Place the TSS one page above the GDT. */
+	tssbase = gdtbase + 0x1000;
+	gdt[3].sd_lobase = tssbase;	
+
+	rflags = 0x2;
+	error = vm_set_register(vmctx, vcpu, VM_REG_GUEST_RFLAGS, rflags);
+	if (error)
+		goto done;
+
+	desc_base = 0;
+	desc_limit = 0xffffffff;
+	desc_access = DESC_GRAN | DESC_DEF32 | DESC_PRESENT | SDT_MEMERA;
+	error = vm_set_desc(vmctx, vcpu, VM_REG_GUEST_CS,
+			    desc_base, desc_limit, desc_access);
+
+	desc_access = DESC_GRAN | DESC_DEF32 | DESC_PRESENT | SDT_MEMRWA;
+	error = vm_set_desc(vmctx, vcpu, VM_REG_GUEST_DS,
+			    desc_base, desc_limit, desc_access);
+	if (error)
+		goto done;
+
+	error = vm_set_desc(vmctx, vcpu, VM_REG_GUEST_ES,
+			    desc_base, desc_limit, desc_access);
+	if (error)
+		goto done;
+
+	error = vm_set_desc(vmctx, vcpu, VM_REG_GUEST_FS,
+			    desc_base, desc_limit, desc_access);
+	if (error)
+		goto done;
+
+	error = vm_set_desc(vmctx, vcpu, VM_REG_GUEST_GS,
+			    desc_base, desc_limit, desc_access);
+	if (error)
+		goto done;
+
+	error = vm_set_desc(vmctx, vcpu, VM_REG_GUEST_SS,
+			    desc_base, desc_limit, desc_access);
+	if (error)
+		goto done;
+
+	desc_base = tssbase;
+	desc_limit = I386_TSS_SIZE - 1;
+	desc_access = DESC_PRESENT | SDT_SYS386BSY;
+	error = vm_set_desc(vmctx, vcpu, VM_REG_GUEST_TR,
+			    desc_base, desc_limit, desc_access);
+	if (error)
+		goto done;
+
+	
+	error = vm_set_desc(vmctx, vcpu, VM_REG_GUEST_LDTR, 0, 0,
+			    DESC_UNUSABLE);
+	if (error)
+		goto done;
+
+	gsel = GSEL(GUEST_CODE_SEL, SEL_KPL);
+	if ((error = vm_set_register(vmctx, vcpu, VM_REG_GUEST_CS, gsel)) != 0)
+		goto done;
+	
+	gsel = GSEL(GUEST_DATA_SEL, SEL_KPL);
+	if ((error = vm_set_register(vmctx, vcpu, VM_REG_GUEST_DS, gsel)) != 0)
+		goto done;
+	
+	if ((error = vm_set_register(vmctx, vcpu, VM_REG_GUEST_ES, gsel)) != 0)
+		goto done;
+
+	if ((error = vm_set_register(vmctx, vcpu, VM_REG_GUEST_FS, gsel)) != 0)
+		goto done;
+	
+	if ((error = vm_set_register(vmctx, vcpu, VM_REG_GUEST_GS, gsel)) != 0)
+		goto done;
+	
+	if ((error = vm_set_register(vmctx, vcpu, VM_REG_GUEST_SS, gsel)) != 0)
+		goto done;
+
+	gsel = GSEL(GUEST_TSS_SEL, SEL_KPL);
+	if ((error = vm_set_register(vmctx, vcpu, VM_REG_GUEST_TR, gsel)) != 0)
+		goto done;
+
+	/* LDTR is pointing to the null selector */
+	if ((error = vm_set_register(vmctx, vcpu, VM_REG_GUEST_LDTR, 0)) != 0)
+		goto done;
+
+	/* entry point */
+	if ((error = vm_set_register(vmctx, vcpu, VM_REG_GUEST_RIP, eip)) != 0)
+		goto done;
+
+	if ((error = vm_set_register(vmctx, vcpu, VM_REG_GUEST_RSP, esp)) != 0)
+		goto done;
+
+	error = 0;
+done:
+	return (error);
+}
 
 void     
 vm_setup_freebsd_gdt(uint64_t *gdtr)
@@ -168,7 +330,7 @@ vm_setup_freebsd_registers(struct vmctx *vmctx, int vcpu,
 		goto done;
 
 	desc_base = gdtbase;
-	desc_limit = GUEST_GDTR_LIMIT;
+	desc_limit = GUEST_GDTR_LIMIT64;
 	error = vm_set_desc(vmctx, vcpu, VM_REG_GUEST_GDTR,
 			    desc_base, desc_limit, 0);
 	if (error != 0)