firmware/lib/vboot_api_kernel.c

/* Copyright (c) 2013 The Chromium OS Authors. All rights reserved.
 * Use of this source code is governed by a BSD-style license that can be
 * found in the LICENSE file.
 *
 * High-level firmware wrapper API - entry points for kernel selection
 */

#include "sysincludes.h"

#include "2sysincludes.h"
#include "2common.h"
#include "2misc.h"
#include "2nvstorage.h"
#include "2rsa.h"
#include "ec_sync.h"
#include "gbb_access.h"
#include "gbb_header.h"
#include "load_kernel_fw.h"
#include "rollback_index.h"
#include "utility.h"
#include "vb2_common.h"
#include "vboot_api.h"
#include "vboot_common.h"
#include "vboot_kernel.h"

/* Global variables */
static struct RollbackSpaceFwmp fwmp;
static LoadKernelParams lkp;
static struct vb2_context ctx;
static uint8_t *unaligned_workbuf;

#ifdef CHROMEOS_ENVIRONMENT
/* Global variable accessors for unit tests */

struct RollbackSpaceFwmp *VbApiKernelGetFwmp(void)
{
	return &fwmp;
}

struct LoadKernelParams *VbApiKernelGetParams(void)
{
	return &lkp;
}

#endif

/**
 * Set recovery request (called from vboot_api_kernel.c functions only)
 */
static void VbSetRecoveryRequest(struct vb2_context *ctx,
				 uint32_t recovery_request)
{
	VB2_DEBUG("VbSetRecoveryRequest(%d)\n", (int)recovery_request);
	vb2_nv_set(ctx, VB2_NV_RECOVERY_REQUEST, recovery_request);
}

void vb2_nv_commit(struct vb2_context *ctx)
{
	/* Exit if nothing has changed */
	if (!(ctx->flags & VB2_CONTEXT_NVDATA_CHANGED))
		return;

	ctx->flags &= ~VB2_CONTEXT_NVDATA_CHANGED;
	VbExNvStorageWrite(ctx->nvdata);
}

uint32_t vb2_get_fwmp_flags(void)
{
	return fwmp.flags;
}

uint32_t VbTryLoadKernel(struct vb2_context *ctx, uint32_t get_info_flags)
{
	VbError_t retval = VBERROR_UNKNOWN;
	VbDiskInfo* disk_info = NULL;
	uint32_t disk_count = 0;
	uint32_t i;

	VB2_DEBUG("VbTryLoadKernel() start, get_info_flags=0x%x\n",
		  (unsigned)get_info_flags);

	lkp.fwmp = &fwmp;
	lkp.disk_handle = NULL;

	/* Find disks */
	if (VBERROR_SUCCESS != VbExDiskGetInfo(&disk_info, &disk_count,
					       get_info_flags))
		disk_count = 0;

	VB2_DEBUG("VbTryLoadKernel() found %d disks\n", (int)disk_count);
	if (0 == disk_count) {
		VbSetRecoveryRequest(ctx, VB2_RECOVERY_RW_NO_DISK);
		return VBERROR_NO_DISK_FOUND;
	}

	/* Loop over disks */
	for (i = 0; i < disk_count; i++) {
		VB2_DEBUG("VbTryLoadKernel() trying disk %d\n", (int)i);
		/*
		 * Sanity-check what we can. FWIW, VbTryLoadKernel() is always
		 * called with only a single bit set in get_info_flags.
		 *
		 * Ensure that we got a partition with only the flags we asked
		 * for.
		 */
		if (disk_info[i].bytes_per_lba < 512 ||
			(disk_info[i].bytes_per_lba &
				(disk_info[i].bytes_per_lba  - 1)) != 0 ||
					16 > disk_info[i].lba_count ||
					get_info_flags != (disk_info[i].flags &
					~VB_DISK_FLAG_EXTERNAL_GPT)) {
			VB2_DEBUG("  skipping: bytes_per_lba=%" PRIu64
				  " lba_count=%" PRIu64 " flags=0x%x\n",
				  disk_info[i].bytes_per_lba,
				  disk_info[i].lba_count,
				  disk_info[i].flags);
			continue;
		}
		lkp.disk_handle = disk_info[i].handle;
		lkp.bytes_per_lba = disk_info[i].bytes_per_lba;
		lkp.gpt_lba_count = disk_info[i].lba_count;
		lkp.streaming_lba_count = disk_info[i].streaming_lba_count
						?: lkp.gpt_lba_count;
		lkp.boot_flags |= disk_info[i].flags & VB_DISK_FLAG_EXTERNAL_GPT
				? BOOT_FLAG_EXTERNAL_GPT : 0;
		retval = LoadKernel(ctx, &lkp);

		VB2_DEBUG("VbTryLoadKernel() LoadKernel() = %d\n", retval);

		/*
		 * Stop now if we found a kernel.
		 *
		 * TODO: If recovery requested, should track the farthest we
		 * get, instead of just returning the value from the last disk
		 * attempted.
		 */
		if (VBERROR_SUCCESS == retval)
			break;
	}

	/* If we didn't find any good kernels, don't return a disk handle. */
	if (VBERROR_SUCCESS != retval) {
		VbSetRecoveryRequest(ctx, VB2_RECOVERY_RW_NO_KERNEL);
		lkp.disk_handle = NULL;
	}

	VbExDiskFreeInfo(disk_info, lkp.disk_handle);

	/*
	 * Pass through return code.  Recovery reason (if any) has already been
	 * set by LoadKernel().
	 */
	return retval;
}

VbError_t VbBootNormal(struct vb2_context *ctx)
{
	struct vb2_shared_data *sd = vb2_get_sd(ctx);
	VbSharedDataHeader *shared = sd->vbsd;
	uint32_t max_rollforward = vb2_nv_get(ctx,
					      VB2_NV_KERNEL_MAX_ROLLFORWARD);

	/* Boot from fixed disk only */
	VB2_DEBUG("Entering\n");

	VbError_t rv = VbTryLoadKernel(ctx, VB_DISK_FLAG_FIXED);

	VB2_DEBUG("Checking if TPM kernel version needs advancing\n");

	if ((1 == shared->firmware_index) && (shared->flags & VBSD_FWB_TRIED)) {
		/*
		 * Special cases for when we're trying a new firmware B.  These
		 * are needed because firmware updates also usually change the
		 * kernel key, which means that the B firmware can only boot a
		 * new kernel, and the old firmware in A can only boot the
		 * previous kernel.
		 *
		 * Don't advance the TPM if we're trying a new firmware B,
		 * because we don't yet know if the new kernel will
		 * successfully boot.  We still want to be able to fall back to
		 * the previous firmware+kernel if the new firmware+kernel
		 * fails.
		 *
		 * If we found only invalid kernels, reboot and try again.
		 * This allows us to fall back to the previous firmware+kernel
		 * instead of giving up and going to recovery mode right away.
		 * We'll still go to recovery mode if we run out of tries and
		 * the old firmware can't find a kernel it likes.
		 */
		if (rv == VBERROR_INVALID_KERNEL_FOUND) {
			VB2_DEBUG("Trying FW B; only found invalid kernels.\n");
			VbSetRecoveryRequest(ctx, VB2_RECOVERY_NOT_REQUESTED);
		}

		return rv;
	}

	/*
	 * Limit kernel version rollforward if needed.  Can't limit kernel
	 * version to less than the version currently in the TPM.  That is,
	 * we're limiting rollforward, not allowing rollback.
	 */
	if (max_rollforward < shared->kernel_version_tpm_start)
		max_rollforward = shared->kernel_version_tpm_start;

	if (shared->kernel_version_tpm > max_rollforward) {
		VB2_DEBUG("Limiting TPM kernel version roll-forward "
			  "to 0x%x < 0x%x\n",
			  max_rollforward, shared->kernel_version_tpm);

		shared->kernel_version_tpm = max_rollforward;
	}

	if ((shared->kernel_version_tpm > shared->kernel_version_tpm_start) &&
	    RollbackKernelWrite(shared->kernel_version_tpm)) {
		VB2_DEBUG("Error writing kernel versions to TPM.\n");
		VbSetRecoveryRequest(ctx, VB2_RECOVERY_RW_TPM_W_ERROR);
		return VBERROR_TPM_WRITE_KERNEL;
	}

	return rv;
}

static VbError_t vb2_kernel_setup(VbCommonParams *cparams,
				  VbSelectAndLoadKernelParams *kparams)
{
	VbSharedDataHeader *shared =
		(VbSharedDataHeader *)cparams->shared_data_blob;

	/* Start timer */
	shared->timer_vb_select_and_load_kernel_enter = VbExGetTimer();

	/*
	 * Set up vboot context.
	 *
	 * TODO: Propagate this up to higher API levels, and use more of the
	 * context fields (e.g. secdatak) and flags.
	 */
	memset(&ctx, 0, sizeof(ctx));

	/* Translate vboot1 flags back to vboot2 */
	if (shared->recovery_reason)
		ctx.flags |= VB2_CONTEXT_RECOVERY_MODE;
	if (shared->flags & VBSD_BOOT_DEV_SWITCH_ON)
		ctx.flags |= VB2_CONTEXT_DEVELOPER_MODE;

	/*
	 * The following flags are set by depthcharge.
	 *
	 * TODO: Some of these are set at compile-time, so could be #defines
	 * instead of flags.  That would save on firmware image size because
	 * features that won't be used in an image could be compiled out.
	 */
	if (shared->flags & VBSD_EC_SOFTWARE_SYNC)
		ctx.flags |= VB2_CONTEXT_EC_SYNC_SUPPORTED;
	if (shared->flags & VBSD_EC_SLOW_UPDATE)
		ctx.flags |= VB2_CONTEXT_EC_SYNC_SLOW;
	if (shared->flags & VBSD_EC_EFS)
		ctx.flags |= VB2_CONTEXT_EC_EFS;
	if (shared->flags & VBSD_NVDATA_V2)
		ctx.flags |= VB2_CONTEXT_NVDATA_V2;

	VbExNvStorageRead(ctx.nvdata);
	vb2_nv_init(&ctx);

	ctx.workbuf_size = VB2_KERNEL_WORKBUF_RECOMMENDED_SIZE +
			VB2_WORKBUF_ALIGN;

	unaligned_workbuf = ctx.workbuf = malloc(ctx.workbuf_size);
	if (!unaligned_workbuf) {
		VB2_DEBUG("Can't allocate work buffer\n");
		VbSetRecoveryRequest(&ctx, VB2_RECOVERY_RW_SHARED_DATA);
		return VBERROR_INIT_SHARED_DATA;
	}

	if (VB2_SUCCESS != vb2_align(&ctx.workbuf, &ctx.workbuf_size,
				     VB2_WORKBUF_ALIGN,
				     VB2_KERNEL_WORKBUF_RECOMMENDED_SIZE)) {
		VB2_DEBUG("Can't align work buffer\n");
		VbSetRecoveryRequest(&ctx, VB2_RECOVERY_RW_SHARED_DATA);
		return VBERROR_INIT_SHARED_DATA;
	}

	if (VB2_SUCCESS != vb2_init_context(&ctx)) {
		VB2_DEBUG("Can't init vb2_context\n");
		free(unaligned_workbuf);
		VbSetRecoveryRequest(&ctx, VB2_RECOVERY_RW_SHARED_DATA);
		return VBERROR_INIT_SHARED_DATA;
	}

	struct vb2_shared_data *sd = vb2_get_sd(&ctx);
	sd->recovery_reason = shared->recovery_reason;

	/*
	 * Save a pointer to the old vboot1 shared data, since we haven't
	 * finished porting the library to use the new vb2 context and shared
	 * data.
	 *
	 * TODO: replace this with fields directly in vb2 shared data.
	 */
	sd->vbsd = shared;

	/*
	 * If we're in recovery mode just to do memory retraining, all we
	 * need to do is reboot.
	 */
	if (sd->recovery_reason == VB2_RECOVERY_TRAIN_AND_REBOOT) {
		VB2_DEBUG("Reboot after retraining in recovery.\n");
		return VBERROR_REBOOT_REQUIRED;
	}

	/* Fill in params for calls to LoadKernel() */
	memset(&lkp, 0, sizeof(lkp));
	lkp.kernel_buffer = kparams->kernel_buffer;
	lkp.kernel_buffer_size = kparams->kernel_buffer_size;

	/* Clear output params in case we fail */
	kparams->disk_handle = NULL;
	kparams->partition_number = 0;
	kparams->bootloader_address = 0;
	kparams->bootloader_size = 0;
	kparams->flags = 0;
	memset(kparams->partition_guid, 0, sizeof(kparams->partition_guid));

	/* Point to GBB data from cparams */
	sd->gbb = cparams->gbb_data;
	sd->gbb_size = cparams->gbb_size;
	sd->gbb_flags = sd->gbb->flags;

	/* Disable rollback TPM calls if necessary. */
	if (kparams->inflags & VB_SALK_INFLAGS_ALTFW_NO_ROLLBACKS) {
		/* Double-check that we're in developer mode. There shouldn't be any
		   way to run the custom Fuchsia firmware outside of developer mode,
		   but add this just to be safe. If we ever upstream Fuchsia support
		   into the real depthcharge build, disabling rollbacks won't be
		   necessary as we won't need altfw anymore. */
		if (ctx.flags & VB2_CONTEXT_DEVELOPER_MODE) {
			RollbackDisable();
		} else {
			VB2_DEBUG("Must be in dev mode to disable rollbacks\n");
			return VBERROR_DEV_FIRMWARE_SWITCH_MISMATCH;
		}
	}

	/* Read kernel version from the TPM.  Ignore errors in recovery mode. */
	if (RollbackKernelRead(&shared->kernel_version_tpm)) {
		VB2_DEBUG("Unable to get kernel versions from TPM\n");
		if (!(ctx.flags & VB2_CONTEXT_RECOVERY_MODE)) {
			VbSetRecoveryRequest(&ctx, VB2_RECOVERY_RW_TPM_R_ERROR);
			return VBERROR_TPM_READ_KERNEL;
		}
	}

	shared->kernel_version_tpm_start = shared->kernel_version_tpm;

	/* Read FWMP.  Ignore errors in recovery mode. */
	if (sd->gbb_flags & VB2_GBB_FLAG_DISABLE_FWMP) {
		memset(&fwmp, 0, sizeof(fwmp));
	} else if (RollbackFwmpRead(&fwmp)) {
		VB2_DEBUG("Unable to get FWMP from TPM\n");
		if (!(ctx.flags & VB2_CONTEXT_RECOVERY_MODE)) {
			VbSetRecoveryRequest(&ctx, VB2_RECOVERY_RW_TPM_R_ERROR);
			return VBERROR_TPM_READ_FWMP;
		}
	}

	return VBERROR_SUCCESS;
}

static VbError_t vb2_kernel_phase4(VbSelectAndLoadKernelParams *kparams)
{
	struct vb2_shared_data *sd = vb2_get_sd(&ctx);

	/* Save disk parameters */
	kparams->disk_handle = lkp.disk_handle;
	kparams->partition_number = lkp.partition_number;
	kparams->bootloader_address = lkp.bootloader_address;
	kparams->bootloader_size = lkp.bootloader_size;
	kparams->flags = lkp.flags;
	kparams->kernel_buffer = lkp.kernel_buffer;
	kparams->kernel_buffer_size = lkp.kernel_buffer_size;
	memcpy(kparams->partition_guid, lkp.partition_guid,
	       sizeof(kparams->partition_guid));

	/* Lock the kernel versions if not in recovery mode */
	if (!(ctx.flags & VB2_CONTEXT_RECOVERY_MODE) &&
	    RollbackKernelLock(sd->recovery_reason)) {
		VB2_DEBUG("Error locking kernel versions.\n");
		VbSetRecoveryRequest(&ctx, VB2_RECOVERY_RW_TPM_L_ERROR);
		return VBERROR_TPM_LOCK_KERNEL;
	}

	return VBERROR_SUCCESS;
}

static void vb2_kernel_cleanup(struct vb2_context *ctx, VbCommonParams *cparams)
{
	/*
	 * This must directly access cparams for now because we could have had
	 * an error setting up the vboot2 context.  In that case
	 * vb2_shared_data is not available.
	 */
	VbSharedDataHeader *shared =
		(VbSharedDataHeader *)cparams->shared_data_blob;

	/*
	 * Clean up vboot context.
	 *
	 * TODO: This should propagate up to higher levels
	 */

	/* Free buffers */
	free(unaligned_workbuf);

	vb2_nv_commit(ctx);

	/* Stop timer */
	shared->timer_vb_select_and_load_kernel_exit = VbExGetTimer();

	/* Store how much shared data we used, if any */
	cparams->shared_data_size = shared->data_used;
}

VbError_t VbSelectAndLoadKernel(VbCommonParams *cparams,
				VbSelectAndLoadKernelParams *kparams)
{
	VbError_t retval = vb2_kernel_setup(cparams, kparams);
	if (retval)
		goto VbSelectAndLoadKernel_exit;

	/*
	 * Do EC software sync unless we're in recovery mode. This has UI but
	 * it's just a single non-interactive WAIT screen.
	 */
	if (!(ctx.flags & VB2_CONTEXT_RECOVERY_MODE)) {
		retval = ec_sync_all(&ctx);
		if (retval)
			goto VbSelectAndLoadKernel_exit;
	}

	/* Select boot path */
	if (ctx.flags & VB2_CONTEXT_RECOVERY_MODE) {
		/* Recovery boot.  This has UI. */
		if (kparams->inflags & VB_SALK_INFLAGS_ENABLE_DETACHABLE_UI)
			retval = VbBootRecoveryMenu(&ctx);
		else
			retval = VbBootRecovery(&ctx);
		VbExEcEnteringMode(0, VB_EC_RECOVERY);
	} else if (ctx.flags & VB2_CONTEXT_DEVELOPER_MODE) {
		if (kparams->inflags & VB_SALK_INFLAGS_VENDOR_DATA_SETTABLE)
			ctx.flags |= VB2_CONTEXT_VENDOR_DATA_SETTABLE;

		/* Developer boot.  This has UI. */
		if (kparams->inflags & VB_SALK_INFLAGS_ENABLE_DETACHABLE_UI)
			retval = VbBootDeveloperMenu(&ctx);
		else
			retval = VbBootDeveloper(&ctx);
		VbExEcEnteringMode(0, VB_EC_DEVELOPER);
	} else {
		/* Normal boot */
		retval = VbBootNormal(&ctx);
		VbExEcEnteringMode(0, VB_EC_NORMAL);
	}

 VbSelectAndLoadKernel_exit:

	if (VBERROR_SUCCESS == retval)
		retval = vb2_kernel_phase4(kparams);

	vb2_kernel_cleanup(&ctx, cparams);

	/* Pass through return value from boot path */
	VB2_DEBUG("Returning %d\n", (int)retval);
	return retval;
}

VbError_t VbVerifyMemoryBootImage(VbCommonParams *cparams,
				  VbSelectAndLoadKernelParams *kparams,
				  void *boot_image,
				  size_t image_size)
{
	VbPublicKey* kernel_subkey = NULL;
	uint8_t *kbuf;
	VbKeyBlockHeader *key_block;
	VbKernelPreambleHeader *preamble;
	uint64_t body_offset;
	int hash_only = 0;
	int dev_switch;
	uint32_t allow_fastboot_full_cap = 0;
	struct vb2_workbuf wb;

	VbError_t retval = vb2_kernel_setup(cparams, kparams);
	if (retval)
		goto fail;

	struct vb2_shared_data *sd = vb2_get_sd(&ctx);
	VbSharedDataHeader *shared = sd->vbsd;

	if ((boot_image == NULL) || (image_size == 0)) {
		retval = VBERROR_INVALID_PARAMETER;
		goto fail;
	}

	kbuf = boot_image;

	/*
	 * We don't care verifying the image if:
	 * 1. dev-mode switch is on and
	 * 2a. GBB_FLAG_FORCE_DEV_BOOT_FASTBOOT_FULL_CAP is set, or
	 * 2b. DEV_BOOT_FASTBOOT_FULL_CAP flag is set in NvStorage
	 *
	 * Check only the integrity of the image.
	 */
	dev_switch = shared->flags & VBSD_BOOT_DEV_SWITCH_ON;
	allow_fastboot_full_cap =
			vb2_nv_get(&ctx, VB2_NV_DEV_BOOT_FASTBOOT_FULL_CAP);

	if (0 == allow_fastboot_full_cap) {
		allow_fastboot_full_cap = !!(sd->gbb_flags &
				VB2_GBB_FLAG_FORCE_DEV_BOOT_FASTBOOT_FULL_CAP);
	}

	if (dev_switch && allow_fastboot_full_cap) {
		VB2_DEBUG("Only performing integrity-check.\n");
		hash_only = 1;
	} else {
		/* Get recovery key. */
		retval = VbGbbReadRecoveryKey(&ctx, &kernel_subkey);
		if (VBERROR_SUCCESS != retval) {
			VB2_DEBUG("Gbb Read Recovery key failed.\n");
			goto fail;
		}
	}

	/* If we fail at any step, retval returned would be invalid kernel. */
	retval = VBERROR_INVALID_KERNEL_FOUND;

	/* Allocate work buffer */
	vb2_workbuf_from_ctx(&ctx, &wb);

	/* Verify the key block. */
	key_block = (VbKeyBlockHeader *)kbuf;
	struct vb2_keyblock *keyblock2 = (struct vb2_keyblock *)kbuf;
	int rv;
	if (hash_only) {
		rv = vb2_verify_keyblock_hash(keyblock2, image_size, &wb);
	} else {
		/* Unpack kernel subkey */
		struct vb2_public_key kernel_subkey2;
		if (VB2_SUCCESS !=
		    vb2_unpack_key(&kernel_subkey2,
				   (struct vb2_packed_key *)kernel_subkey)) {
			VB2_DEBUG("Unable to unpack kernel subkey\n");
			goto fail;
		}
		rv = vb2_verify_keyblock(keyblock2, image_size,
					 &kernel_subkey2, &wb);
	}

	if (VB2_SUCCESS != rv) {
		VB2_DEBUG("Verifying key block signature/hash failed.\n");
		goto fail;
	}

	/* Check the key block flags against the current boot mode. */
	if (!(key_block->key_block_flags &
	      (dev_switch ? KEY_BLOCK_FLAG_DEVELOPER_1 :
	       KEY_BLOCK_FLAG_DEVELOPER_0))) {
		VB2_DEBUG("Key block developer flag mismatch.\n");
		if (hash_only == 0)
			goto fail;
	}

	if (!(key_block->key_block_flags & KEY_BLOCK_FLAG_RECOVERY_1)) {
		VB2_DEBUG("Key block recovery flag mismatch.\n");
		if (hash_only == 0)
			goto fail;
	}

	/* Get key for preamble/data verification from the key block. */
	struct vb2_public_key data_key2;
	if (VB2_SUCCESS != vb2_unpack_key(&data_key2, &keyblock2->data_key)) {
		VB2_DEBUG("Unable to unpack kernel data key\n");
		goto fail;
	}

	/* Verify the preamble, which follows the key block */
	preamble = (VbKernelPreambleHeader *)(kbuf + key_block->key_block_size);
	struct vb2_kernel_preamble *preamble2 =
			(struct vb2_kernel_preamble *)
			(kbuf + key_block->key_block_size);

	if (VB2_SUCCESS != vb2_verify_kernel_preamble(
			preamble2,
			image_size - key_block->key_block_size,
			&data_key2,
			&wb)) {
		VB2_DEBUG("Preamble verification failed.\n");
		goto fail;
	}

	VB2_DEBUG("Kernel preamble is good.\n");

	/* Verify kernel data */
	body_offset = key_block->key_block_size + preamble->preamble_size;
	if (VB2_SUCCESS != vb2_verify_data(
			(const uint8_t *)(kbuf + body_offset),
			image_size - body_offset,
			(struct vb2_signature *)&preamble->body_signature,
			&data_key2, &wb)) {
		VB2_DEBUG("Kernel data verification failed.\n");
		goto fail;
	}

	VB2_DEBUG("Kernel is good.\n");

	/* Fill in output parameters. */
	kparams->kernel_buffer = kbuf + body_offset;
	kparams->kernel_buffer_size = image_size - body_offset;
	kparams->bootloader_address = preamble->bootloader_address;
	kparams->bootloader_size = preamble->bootloader_size;
	if (VbKernelHasFlags(preamble) == VBOOT_SUCCESS)
		kparams->flags = preamble->flags;

	retval = VBERROR_SUCCESS;

 fail:
	vb2_kernel_cleanup(&ctx, cparams);
	if (NULL != kernel_subkey)
		free(kernel_subkey);
	return retval;
}

VbError_t VbUnlockDevice(void)
{
	VB2_DEBUG("Enabling dev-mode...\n");
	if (TPM_SUCCESS != SetVirtualDevMode(1))
		return VBERROR_TPM_SET_BOOT_MODE_STATE;

	VB2_DEBUG("Mode change will take effect on next reboot.\n");
	return VBERROR_SUCCESS;
}