mail archive of the barebox mailing list
 help / color / mirror / Atom feed
From: Juergen Beisert <jbe@pengutronix.de>
To: barebox@lists.infradead.org
Subject: [PATCHv2 01/14] Add a tool to activate barebox as a boot loader on x86 architectures
Date: Tue, 12 Jan 2010 11:15:35 +0100	[thread overview]
Message-ID: <20100112102053.822036694@pengutronix.de> (raw)
In-Reply-To: <20100112101534.868276907@pengutronix.de>

[-- Attachment #1: add_install_tool.diff --]
[-- Type: text/plain, Size: 26044 bytes --]

To use barebox as a BIOS based bootloader for x86 architectures, the binary
must be patched to get it bootstrapped at runtime. The 'setupmbr' tool installs
the barebox-binary to the given device node or image file and patch it in
accordance to the needed sector information at runtime.

Signed-off by: Juergen Beisert <jbe@pengutronix.de>

---
 Doxyfile                    |    3 
 scripts/Makefile            |    4 
 scripts/setupmbr/Makefile   |    4 
 scripts/setupmbr/arch.h     |   55 +++
 scripts/setupmbr/setupmbr.c |  705 ++++++++++++++++++++++++++++++++++++++++++++
 5 files changed, 769 insertions(+), 2 deletions(-)

Index: barebox-2009.12.0/scripts/Makefile
===================================================================
--- barebox-2009.12.0.orig/scripts/Makefile
+++ barebox-2009.12.0/scripts/Makefile
@@ -14,5 +14,7 @@ always		:= $(hostprogs-y) $(hostprogs-m)
 
 subdir-y                     += mod
 
+subdir-$(CONFIG_X86)             += setupmbr
+
 # Let clean descend into subdirs
-subdir-	+= basic kconfig
+subdir-	+= basic kconfig setupmbr
Index: barebox-2009.12.0/scripts/setupmbr/Makefile
===================================================================
--- /dev/null
+++ barebox-2009.12.0/scripts/setupmbr/Makefile
@@ -0,0 +1,4 @@
+HOST_EXTRACFLAGS=-I$(srctree)
+
+hostprogs-y  := setupmbr
+always       := $(hostprogs-y)
Index: barebox-2009.12.0/scripts/setupmbr/arch.h
===================================================================
--- /dev/null
+++ barebox-2009.12.0/scripts/setupmbr/arch.h
@@ -0,0 +1,55 @@
+
+/* we need the one from the host */
+#include <endian.h>
+#include <stdint.h>
+
+/* Byte-orders.  */
+#define swap16(x)	\
+({ \
+   uint16_t _x = (x); \
+   (uint16_t) ((_x << 8) | (_x >> 8)); \
+})
+
+#define swap32(x)	\
+({ \
+   uint32_t _x = (x); \
+   (uint32_t) ((_x << 24) \
+                    | ((_x & (uint32_t) 0xFF00UL) << 8) \
+                    | ((_x & (uint32_t) 0xFF0000UL) >> 8) \
+                    | (_x >> 24)); \
+})
+
+#define swap64(x)	\
+({ \
+   uint64_t _x = (x); \
+   (uint64_t) ((_x << 56) \
+                    | ((_x & (uint64_t) 0xFF00ULL) << 40) \
+                    | ((_x & (uint64_t) 0xFF0000ULL) << 24) \
+                    | ((_x & (uint64_t) 0xFF000000ULL) << 8) \
+                    | ((_x & (uint64_t) 0xFF00000000ULL) >> 8) \
+                    | ((_x & (uint64_t) 0xFF0000000000ULL) >> 24) \
+                    | ((_x & (uint64_t) 0xFF000000000000ULL) >> 40) \
+                    | (_x >> 56)); \
+})
+
+#if __BYTE_ORDER == __BIG_ENDIAN
+
+/* Our target is a ia32 machine, always little endian */
+
+# define host2target_16(x)	swap16(x)
+# define host2target_32(x)	swap32(x)
+# define host2target_64(x)	swap64(x)
+# define target2host_16(x)	swap16(x)
+# define target2host_32(x)	swap32(x)
+# define target2host_64(x)	swap64(x)
+
+#else
+
+# define host2target_16(x)	(x)
+# define host2target_32(x)	(x)
+# define host2target_64(x)	(x)
+# define target2host_16(x)	(x)
+# define target2host_32(x)	(x)
+# define target2host_64(x)	(x)
+
+#endif
Index: barebox-2009.12.0/scripts/setupmbr/setupmbr.c
===================================================================
--- /dev/null
+++ barebox-2009.12.0/scripts/setupmbr/setupmbr.c
@@ -0,0 +1,705 @@
+/*
+ * Copyright (C) 2009 Juergen Beisert, Pengutronix
+ *
+ * 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, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston,
+ * MA 02111-1307 USA
+ *
+ */
+
+/**
+ * @file
+ * @brief Write the barebox binary to the MBR and the following disk sectors
+ *
+ * Also patch dedicated locations in the image to make it work at runtime
+ *
+ * Current restrictions are:
+ * - only installs into MBR and the sectors after it
+ * - tested only with QEMU
+ * - and maybe some others
+ */
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <stdint.h>
+#include <sys/mman.h>
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <unistd.h>
+#include <fcntl.h>
+#include <string.h>
+#include <assert.h>
+
+/* include the info from this barebox release */
+#include "include/linux/utsrelease.h"
+#include "arch/x86/include/asm/barebox.lds.h"
+
+/** define to disable integrity tests and debug messages */
+#define NDEBUG
+
+/* some infos about our target architecture */
+#include "arch.h"
+
+/**
+ * "Disk Address Packet Structure" to be used when calling int13,
+ * function 0x42
+ *
+ * @note All entries are in target endianess
+ */
+struct DAPS
+{
+	uint8_t size;		/**< size of this data set, 0 marks it as invalid */
+	uint8_t res1;		/**< always 0 */
+	int8_t count;		/**< number of sectors 0...127 to handle */
+	uint8_t res2;		/**< always 0 */
+	uint16_t offset;	/**< store address: offset */
+	uint16_t segment;	/**< store address: segment */
+	uint64_t lba;		/**< start sector number in LBA */
+} __attribute__ ((packed));
+
+/**
+ * Description of one partition table entry (D*S type)
+ *
+ * @note All entries are in target endianess
+ */
+struct partition_entry {
+	uint8_t boot_indicator;
+	uint8_t chs_begin[3];
+	uint8_t type;
+	uint8_t chs_end[3];
+	uint32_t partition_start;	/* LE */
+	uint32_t partition_size;	/* LE */
+} __attribute__ ((packed));
+
+#ifndef NDEBUG
+static void debugout(const struct DAPS *entry, int supress_entry)
+{
+	if (supress_entry)
+		printf("DAPS entry: ");
+	else
+		printf("DAPS entry % 3u: ", ((unsigned)entry & ( SECTOR_SIZE - 1)) / sizeof(struct DAPS));
+
+	printf("Size: % 2u, Count: % 3d, Offset: 0x%04hX, Segment: 0x%04hX, LBA: %llu\n",
+		entry->size, entry->count,
+		target2host_16(entry->offset), target2host_16(entry->segment),
+		target2host_64(entry->lba));
+}
+#else
+# define debugout(x,y) (__ASSERT_VOID_CAST(0))
+#endif
+
+/**
+ * Fill *one* DAPS
+ * @param sector The DAPS to fill
+ * @param count Sector count
+ * @param offset Realmode offset in the segment
+ * @param segment Real mode segment
+ * @param lba LBA of the first sector to read
+ * @return 0 on success
+ */
+static int fill_daps(struct DAPS *sector, unsigned count, unsigned offset, unsigned segment, uint64_t lba)
+{
+	assert(sector != NULL);
+	assert(count < 128);
+	assert(offset < 0x10000);
+	assert(segment < 0x10000);
+
+	sector->size = sizeof(struct DAPS);
+	sector->res1 = 0;
+	sector->count = (int8_t)count;
+	sector->res2 = 0;
+	sector->offset = host2target_16(offset);
+	sector->segment = host2target_16(segment);
+	sector->lba = host2target_64(lba);
+
+	return 0;
+}
+
+/**
+ * Mark a DAPS invalid to let the boot loader code stop at this entry.
+ * @param sector The DAPS to be marked as invalid
+ *
+ * Marking as invalid must be done in accordance to the detection method
+ * the assembler routine in the boot loader uses:
+ * The current code tests for zero in the first two bytes of the DAPS.
+ */
+static void invalidate_daps(struct DAPS *sector)
+{
+	sector->size = MARK_DAPS_INVALID;
+	sector->res1 = 0;
+}
+
+/**
+ * Create the indirect sector with the DAPS entries
+ * @param daps_table Where to store the entries
+ * @param size Size of the whole image in bytes
+ * @param pers_sector_count Count of sectors to skip after MBR for the persistant environment storage
+ * @return 0 on success
+ *
+ * This routine calculates the DAPS entries for the case the whole
+ * barebox images fits into the MBR itself and the sectors after it.
+ * This means the start of the first partition must keep enough sectors
+ * unused.
+ * It also skips 'pers_sector_count' sectors after the MBR for special
+ * usage if given.
+ */
+static int barebox_linear_image(struct DAPS *daps_table, off_t size, long pers_sector_count)
+{
+	unsigned offset = LOAD_AREA, next_offset;
+	unsigned segment = LOAD_SEGMENT;
+	unsigned chunk_size, i = 0;
+	uint64_t lba = 2 + pers_sector_count;
+	int rc;
+
+	/*
+	 * We can load up to 127 sectors in one chunk. What a bad number...
+	 * So, we will load in chunks of 32 kiB.
+	 */
+
+	/* at runtime two sectors from the image are already loaded: MBR and indirect */
+	size -= 2 * SECTOR_SIZE;
+	/* and now round up to multiple of sector size */
+	size = (size + SECTOR_SIZE - 1) & ~(SECTOR_SIZE - 1);
+
+	/*
+	 * The largest image we can load with this method is:
+	 * (SECTOR_SIZE / sizeof(DAPS) - 1) * 32 kiB
+	 * For a 512 byte sector and a 16 byte DAPS:
+	 * (512 / 16 - 1) * 32 kiB = 992 kiB
+	 * Note: '- 1' to consider one entry is required to pad to a 32 kiB boundary
+	 */
+
+	if (size >= (SECTOR_SIZE / sizeof(struct DAPS) - 1) * 32 * 1024) {
+		fprintf(stderr, "Image too large to boot. Max size is %zu kiB, image size is %lu kiB\n",
+			(SECTOR_SIZE / sizeof(struct DAPS) - 1) * 32, size / 1024);
+		return -1;
+	}
+
+	if (size > 32 * 1024) {
+		/* first fill up until the next 32 k boundary */
+		next_offset = (offset + 32 * 1024 -1) & ~0x7fff;
+		chunk_size = next_offset - offset;
+		if (chunk_size & (SECTOR_SIZE-1)) {
+			fprintf(stderr, "Unable to pad from %X to %X in multiple of sectors\n", offset, next_offset);
+			return -1;
+		}
+
+		rc = fill_daps(&daps_table[i], chunk_size / SECTOR_SIZE, offset, segment, lba);
+		if (rc != 0)
+			return -1;
+		debugout(&daps_table[i], 0);
+
+		/* calculate values to enter the loop for the other entries */
+		size -= chunk_size;
+		i++;
+		lba += chunk_size / SECTOR_SIZE;
+		offset += chunk_size;
+		if (offset >= 0x10000) {
+			segment += 4096;
+			offset = 0;
+		}
+
+		/*
+		 * Now load the remaining image part in 32 kiB chunks
+		 */
+		while (size) {
+			if (size >= 32 * 1024 ) {
+				if (i >= (SECTOR_SIZE / sizeof(struct DAPS))) {
+					fprintf(stderr, "Internal tool error: Too many DAPS entries!\n");
+					return -1;
+				}
+				rc = fill_daps(&daps_table[i], 64, offset, segment, lba);
+				if (rc != 0)
+					return -1;
+				debugout(&daps_table[i], 0);
+
+				size -= 32 * 1024;
+				lba += 64;
+				offset += 32 * 1024;
+				if (offset >= 0x10000) {
+					segment += 4096;
+					offset = 0;
+				}
+				i++;
+			} else {
+				if (i >= (SECTOR_SIZE / sizeof(struct DAPS))) {
+					fprintf(stderr, "Internal tool error: Too many DAPS entries!\n");
+					return -1;
+				}
+				rc = fill_daps(&daps_table[i], size / SECTOR_SIZE, offset, segment, lba);
+				if (rc != 0)
+					return -1;
+				debugout(&daps_table[i], 0);
+				size = 0;	/* finished */
+				i++;
+			}
+		};
+	} else {
+		/* less than 32 kiB. Very small image... */
+		rc = fill_daps(&daps_table[i], size / SECTOR_SIZE, offset, segment, lba);
+		if (rc != 0)
+			return -1;
+		debugout(&daps_table[i], 0);
+		i++;
+	}
+
+	/*
+	 * Do not mark an entry as invalid if the buffer is full. The
+	 * boot code stops if all entries of a buffer are read.
+	 */
+	if (i >= (SECTOR_SIZE / sizeof(struct DAPS)))
+		return 0;
+
+	/* mark the last DAPS invalid */
+	invalidate_daps(&daps_table[i]);
+	debugout(&daps_table[i], 0);
+
+	return 0;
+}
+
+/**
+ * Do some simple sanity checks if this sector could be an MBR
+ * @param sector Sector with data to check
+ * @param size Size of the buffer
+ * @return 0 if successfull
+ */
+static int check_for_valid_mbr(const uint8_t *sector, off_t size)
+{
+	if (size < SECTOR_SIZE) {
+		fprintf(stderr, "MBR too small to be valid\n");
+		return -1;
+	}
+
+	if ((sector[OFFSET_OF_SIGNATURE] != 0x55) ||
+		(sector[OFFSET_OF_SIGNATURE + 1] != 0xAA)) {
+		fprintf(stderr, "No MBR signature found\n");
+		return -1;
+	}
+
+	/* FIXME: try to check if there is a valid partition table */
+	return 0;
+}
+
+/**
+ * Check space between start of the image and the start of the first partition
+ * @param hd_image HD image to examine
+ * @param size Size of the barebox image
+ * @return 0 on success, -1 if the barebox image is too large
+ */
+static int check_for_space(const void *hd_image, off_t size)
+{
+	struct partition_entry *partition;
+	uint8_t *mbr_disk_sector = (uint8_t*)hd_image;
+	off_t spare_sector_count;
+
+	assert(hd_image != NULL);
+	assert(size > 0);
+
+	if (check_for_valid_mbr(hd_image, size) != 0)
+		return -1;
+
+	/* where to read */
+	partition = (struct partition_entry*) &mbr_disk_sector[OFFSET_OF_PARTITION_TABLE];
+
+	/* TODO describes the first entry always the first partition? */
+	spare_sector_count = target2host_32(partition->partition_start);
+
+#ifdef DEBUG
+	printf("Debug: Required free sectors for barebox prior first partition: %lu, hd image provides: %lu\n",
+		(size + SECTOR_SIZE - 1) / SECTOR_SIZE, spare_sector_count);
+#endif
+	spare_sector_count *= SECTOR_SIZE;
+	if (spare_sector_count < size) {
+		fprintf(stderr, "Not enough space after MBR to store barebox\n");
+		fprintf(stderr, "Move begin of the first partition beyond sector %lu\n", (size + SECTOR_SIZE - 1) / SECTOR_SIZE);
+		return -1;
+	}
+
+	return 0;
+}
+
+/**
+ * Setup the persistant environment storage information
+ * @param patch_area Where to patch
+ * @param pers_sector_start Start sector of the persistant environment storage
+ * @param pers_sector_count Count of sectors for the persistant environment storage
+ * @return 0 on success
+ */
+static int store_pers_env_info(void *patch_area, uint64_t pers_sector_start, long pers_sector_count)
+{
+	uint64_t *start_lba = (uint64_t*)(patch_area + PATCH_AREA_PERS_START);
+	uint16_t *count_lba = (uint16_t*)(patch_area + PATCH_AREA_PERS_SIZE);
+
+	assert(patch_area != NULL);
+	assert(pers_sector_count >= 0);
+
+	if (pers_sector_count == 0) {
+		*count_lba = host2target_16(PATCH_AREA_PERS_SIZE_UNUSED);
+		return 0;
+	}
+
+	*start_lba = host2target_64(pers_sector_start);
+	*count_lba = host2target_16(pers_sector_count);
+
+	return 0;
+}
+
+/**
+ * Prepare the MBR and indirect sector for runtime
+ * @param fd_barebox barebox image to use
+ * @param fd_hd Hard disk image to prepare
+ * @param pers_sector_count Count of sectors to skip after MBR for the persistant environment storage
+ * @return 0 on success
+ *
+ * This routine expects a prepared hard disk image file with a partition table
+ * in its first sector. This method only is currently supported.
+ */
+static int barebox_overlay_mbr(int fd_barebox, int fd_hd, long pers_sector_count)
+{
+	const void *barebox_image;
+	void *hd_image;
+	int rc;
+	struct stat sb;
+	struct DAPS *embed;	/* part of the MBR */
+	struct DAPS *indirect;	/* sector with indirect DAPS */
+	off_t required_size;
+
+	if (fstat(fd_barebox, &sb) == -1) {
+		perror("fstat");
+		return -1;
+	}
+
+	/* the barebox image won't be touched */
+	barebox_image = mmap(NULL, sb.st_size,  PROT_READ, MAP_SHARED, fd_barebox, 0);
+	if (barebox_image == MAP_FAILED) {
+		perror("mmap");
+		return -1;
+	}
+
+	rc = check_for_valid_mbr(barebox_image, sb.st_size);
+	if (rc != 0) {
+		fprintf(stderr, "barebox image seems not valid: Bad MBR signature\n");
+		goto on_error_hd;
+	}
+
+	/*
+	 * the persistant environment storage is in front of the main
+	 * barebox image. To handle both, we need more space in front of the
+	 * the first partition.
+	 */
+	required_size = sb.st_size + pers_sector_count * SECTOR_SIZE;
+
+	/* the hd image will be modified */
+	hd_image = mmap(NULL, required_size,  PROT_READ | PROT_WRITE,
+					MAP_SHARED, fd_hd, 0);
+	if (hd_image == MAP_FAILED) {
+		perror("mmap");
+		rc = -1;
+		goto on_error_hd;
+	}
+
+	/* check for space */
+	rc = check_for_space(hd_image, required_size);
+	if (rc != 0)
+		goto on_error_space;
+
+	/* embed barebox's boot code into the disk drive image */
+	memcpy(hd_image, barebox_image, OFFSET_OF_PARTITION_TABLE);
+
+	/*
+	 * embed the barebox main image into the disk drive image,
+	 * but keep the persistant environment storage untouched
+	 * (if defined), e.g. store the main image behind this special area.
+	 */
+	memcpy(hd_image + ((pers_sector_count + 1) * SECTOR_SIZE),
+			barebox_image + SECTOR_SIZE, sb.st_size - SECTOR_SIZE);
+
+	/* now, prepare this hard disk image for BIOS based booting */
+	embed = hd_image + PATCH_AREA;
+	indirect = hd_image + ((pers_sector_count + 1) * SECTOR_SIZE);
+
+	/*
+	 * Fill the embedded DAPS to let the basic boot code find the
+	 * indirect sector at runtime
+	 */
+#ifdef DEBUG
+	printf("Debug: Fill in embedded DAPS\n");
+#endif
+	rc = fill_daps(embed, 1, INDIRECT_AREA, INDIRECT_SEGMENT,
+				1 + pers_sector_count);
+	if (rc != 0)
+		goto on_error_space;
+	debugout(embed, 1);
+
+#ifdef DEBUG
+	printf("Debug: Fill in indirect sector\n");
+#endif
+	/*
+	 * fill the indirect sector with the remaining DAPS to load the
+	 * whole barebox image at runtime
+	 */
+	rc = barebox_linear_image(indirect, sb.st_size, pers_sector_count);
+	if (rc != 0)
+		goto on_error_space;
+
+	/*
+	 * TODO: Replace the fixed LBA starting number by a calculated one,
+	 * to support barebox as a chained loader in a different start
+	 * sector than the MBR
+	 */
+	rc = store_pers_env_info(embed, 1, pers_sector_count);
+	if (rc != 0)
+		goto on_error_space;
+
+on_error_space:
+	munmap(hd_image, required_size);
+
+on_error_hd:
+	munmap((void*)barebox_image, sb.st_size);
+
+	return rc;
+}
+
+static void print_usage(const char *pname)
+{
+	printf("%s: Preparing a hard disk image for boot with barebox on x86.\n", pname);
+	printf("Usage is\n %s [options] -m <barebox image> -d <hd image>\n", pname);
+	printf(" [options] are:\n -s <count> sector count of the persistant environment storage\n");
+	printf(" <barebox image> barebox's boot image file\n");
+	printf(" <hd image> HD image to store the barebox image\n");
+	printf(" If no '-s <x>' was given, barebox occupies sectors 0 to n, else sector 0 and x+1 to n\n");
+}
+
+int main(int argc, char *argv[])
+{
+	int rc = 0, c;
+	char *barebox_image_filename = NULL, *hd_image_filename = NULL;
+	int fd_barebox_image = 0, fd_hd_image = 0;
+	long barebox_pers_size = -1;
+
+	if (argc == 1) {
+		print_usage(argv[0]);
+		exit(0);
+	}
+
+	/* handle command line options first */
+	while (1) {
+		c = getopt(argc, argv, "m:d:s:hv");
+		if (c == -1)
+			break;
+
+		switch (c) {
+		case 's':
+			barebox_pers_size = strtol(optarg, NULL, 0);
+			break;
+		case 'm':
+			barebox_image_filename = strdup(optarg);
+			break;
+		case 'd':
+			hd_image_filename = strdup(optarg);
+			break;
+		case 'h':
+			print_usage(argv[0]);
+			rc = 0;
+			goto on_error;
+		case 'v':
+			printf("setupmbr for u-boot-v%s\n", UTS_RELEASE);
+			printf("Send bug reports to 'barebox@lists.infradead.org'\n");
+			rc = 0;
+			goto on_error;
+		}
+	}
+
+	if (barebox_image_filename == NULL) {
+		print_usage(argv[0]);
+		rc = -1;
+		goto on_error;
+	}
+
+	fd_barebox_image = open(barebox_image_filename, O_RDONLY);
+	if (fd_barebox_image == -1) {
+		fprintf(stderr, "Cannot open '%s' for reading\n",
+				barebox_image_filename);
+		rc = -1;
+		goto on_error;
+	}
+
+	fd_hd_image = open(hd_image_filename, O_RDWR);
+	if (fd_hd_image == -1) {
+		fprintf(stderr, "Cannot open '%s'\n", hd_image_filename);
+		rc = -1;
+		goto on_error;
+	}
+
+	if (barebox_pers_size < 0)
+		barebox_pers_size = 0;
+
+	rc = barebox_overlay_mbr(fd_barebox_image, fd_hd_image, barebox_pers_size);
+
+on_error:
+	if (fd_barebox_image != -1)
+		close(fd_barebox_image);
+	if (fd_hd_image != -1)
+		close(fd_hd_image);
+
+	if (barebox_image_filename != NULL)
+		free(barebox_image_filename);
+	if (hd_image_filename != NULL)
+		free(hd_image_filename);
+
+	return rc;
+}
+
+/** @page x86_bootloader barebox acting as PC bootloader
+
+@section x86_bootloader_features Features
+
+@a barebox can act as a bootloader for PC based systems. In this case a special
+binary layout will be created to be able to store it on some media the PC
+BIOS can boot from. It can boot Linux kernels stored also on the same boot
+media and be configured at runtime, with the possibility to store the changed
+configuration on the boot media.
+
+@section x86_bootloader_restrictions Restrictions
+
+Due to some BIOS and @a barebox restrictions the boot media must be
+prepared in some special way:
+
+@li @a barebox must be stored in the MBR (Master Boot Record) of the boot media.
+    Currently its not possible to store and boot it in one of the partition
+    sectors (to use it as a second stage loader). This is no eternal
+    restriction. It only needs further effort to add this feature.
+@li @a barebox currently cannot run a chained boot loader. Also, this is no
+    eternal restriction, only further effort needed.
+@li @a barebox comes with limited filesystem support. There is currently no
+    support for the most common and popular filesystems used in the *NIX world.
+    This restricts locations where to store a kernel and other runtime
+    information
+@li @a barebox must be stored to the first n sectors of the boot media. To ensure
+    this does not collide with partitions on the boot media, the first
+    partition must start at a sector behind the ones @a barebox occupies.
+@li @a barebox handles its runtime configuration in a special way: It stores it
+    in a binary way into some reserved sectors ("persistant storage").
+
+@section x86_bootloader_preparations Boot Preparations
+
+To store the @a barebox image to a boot media, it comes with the tool
+@p setupmbr in the directory @p scripts/setupmbr/ . To be able to use it on
+the boot media of your choice, some preparations are required:
+
+@subsection x86_bootloader_preparations_partition Keep Sectors free
+
+Build the @a barebox image and check its size. At least this amount of
+sectors must be kept free after the MBR prior the first partition. Do this
+simple calulation:
+
+	sectors = (\<size of barebox image\> + 511) / 512
+
+To be able to store the runtime configuration, further free sectors are
+required. Its up to you and your requirements, how large this persistant
+storage must be. If you need 16 kiB for this purpose, you need to keep
+additional 32 sectors free.
+
+For this example we are reserving 300 sectors for the @a barebox image and
+additionaly 32 sectors for the persistant storage. So, the first partition on
+the boot media must start at sector 333 or later.
+
+Run the @p fdisk tool to setup such a partition table:
+
+@verbatim
+[jb@host]~> fdisk /dev/sda
+Command (m for help): p
+
+Disk /dev/sda: 20.7 MB, 212680704 bytes
+16 heads, 63 sectors/track, 406 cylinders
+Units = cylinders of 1008 * 512 = 516096 bytes
+
+   Device Boot      Start         End      Blocks   Id  System
+@endverbatim
+
+Change the used units to @p sectors for easier handling.
+
+@verbatim
+Command (m for help): u
+Changing display/entry units to sectors
+
+Command (m for help): p
+
+Disk /dev/sda: 20.7 MB, 212680704 bytes
+16 heads, 63 sectors/track, 406 cylinders, total 409248 sectors
+Units = sectors of 1 * 512 = 512 bytes
+
+   Device Boot      Start         End      Blocks   Id  System
+@endverbatim
+
+Now its possible to create the first partition with the required offset:
+
+@verbatim
+Command (m for help): n
+Command action
+   e   extended
+   p   primary partition (1-4)
+p
+Partition number (1-4): 1
+First sector (63-409247, default 63): 333
+Last sector or +size or +sizeM or +sizeK (333-409247, default 409247): +18M
+Command (m for help): p
+
+Disk /dev/sda: 20.7 MB, 212680704 bytes
+16 heads, 63 sectors/track, 406 cylinders, total 409248 sectors
+Units = sectors of 1 * 512 = 512 bytes
+
+        Device Boot      Start         End      Blocks   Id  System
+/dev/sda                   333       35489       17578+  83  Linux
+@endverbatim
+
+That's all. Do whatever is required now with the new partition (formatting
+and populating the root filesystem for example) to make it useful.
+
+In the next step, @a barebox gets installed to this boot media:
+
+@verbatim
+[jb@host]~> scripts/setupmbr/setupmbr -s 32 -m ./barebox -d /dev/sda
+@endverbatim
+
+This command writes the @a barebox image file './barebox' onto the device
+@p /dev/sda.
+
+The @p -s option will keep the persistant storage sectors free and untouched
+and set flags in the MBR to forward its existance, size and location to
+@a barebox at runtime. @p setupmbr also does not change the partition table.
+
+The @a barebox image gets stored on the boot media like this:
+
+@verbatim
+sector 0   1             33                              333
+       |---|-------------|--------------- ~~~ ------------|--------------
+      MBR    persistant              barebox                 first
+              storage               main image              partition
+@endverbatim
+
+If the @p -s option is omitted, the "persistant storage" part simply does
+not exist:
+
+@verbatim
+sector 0   1                              333
+       |---|--------------- ~~~ ------------|--------------
+      MBR               barebox                 first
+                       main image              partition
+@endverbatim
+
+@note The @p setupmbr tool is also working on real image file than on device
+      nodes only. So, there is no restriction what kind of file will be
+      modified.
+*/
Index: barebox-2009.12.0/Doxyfile
===================================================================
--- barebox-2009.12.0.orig/Doxyfile
+++ barebox-2009.12.0/Doxyfile
@@ -484,7 +484,8 @@ INPUT                  = Documentation \
                          commands \
                          common \
                          board \
-                         lib
+                         lib \
+                         scripts/setupmbr
 
 # This tag can be used to specify the character encoding of the source files that
 # doxygen parses. Internally doxygen uses the UTF-8 encoding, which is also the default

-- 

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

  reply	other threads:[~2010-01-12 10:21 UTC|newest]

Thread overview: 16+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-01-12 10:15 [PATCHv2 00/14] Bring in basic x86 support into 'barebox' Juergen Beisert
2010-01-12 10:15 ` Juergen Beisert [this message]
2010-01-12 10:15 ` [PATCHv2 02/14] Consider real and protected mode in the dump file Juergen Beisert
2010-01-12 10:15 ` [PATCHv2 03/14] Adding x86 usage documentation to the tree Juergen Beisert
2010-01-12 10:15 ` [PATCHv2 04/14] Adding required architecture header files Juergen Beisert
2010-01-12 10:15 ` [PATCHv2 05/14] Bring in the first x86 specific code Juergen Beisert
2010-01-12 10:15 ` [PATCHv2 06/14] Add some generic functions to make x86 work Juergen Beisert
2010-01-12 10:15 ` [PATCHv2 07/14] Add functions to be able to boot with BIOSs help Juergen Beisert
2010-01-12 10:15 ` [PATCHv2 08/14] Add a generic PC platform Juergen Beisert
2010-01-12 10:15 ` [PATCHv2 09/14] Bring in the time reference for the x86 platforms Juergen Beisert
2010-01-12 10:15 ` [PATCHv2 10/14] Start to add ATA support to barebox Juergen Beisert
2010-01-12 10:15 ` [PATCHv2 11/14] Add the main disk driver Juergen Beisert
2010-01-12 10:15 ` [PATCHv2 12/14] Add a low level disk drive access driver Juergen Beisert
2010-01-12 10:15 ` [PATCHv2 13/14] Add a special command to load and start a bzImage on x86 Juergen Beisert
2010-01-12 10:15 ` [PATCHv2 14/14] Add the whole x86 architecture to the build system Juergen Beisert
2010-01-14  9:29 ` [PATCHv2 00/14] Bring in basic x86 support into 'barebox' Sascha Hauer

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20100112102053.822036694@pengutronix.de \
    --to=jbe@pengutronix.de \
    --cc=barebox@lists.infradead.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox