Felix thanks for your first implementation supporting metadata v1.x raid
disks. I tried it on my RAID5 with v1.0 superblock. But it needs some
bugfixes:
Looking into the kernel code, obviously v0.9 superblocks and v1.x
superblocks save some values in a different dimension unit:
v0.9 chunk_size in byte
v1.x chunk_size in sectors of 512 byte
v0.9 disk_size in blocks of 1 kbyte = 1024 byte
v1.x disk_size in sectors of 512 byte
Without any additional recalculation this two fixes gave me a first
chance to boot my system, using a boot partition on a LVM2 volume inside
that RAID5 of 3 disks inside a standard PC partition (you should move
the data part of the 1st partition by some additional sectors, because
the core.img needs more that the default reserved 31 sectors here).
But unfortunately the grub-probe claimed that it requires the raid5rec
module and booting works only with the first 2 disks assembled in my
system (I used the grub-install --modules=ext2 /dev/md0 for this
first trail). Looking on the RAID superblock, especially the "Array
Slot" gave me the final hint:
> /sbin/mdadm -E /dev/sdc2
> /dev/sdc2:
> Magic : a92b4efc
> Version : 1.0
> Feature Map : 0x0
> Array UUID : 4e929199:313b3aee:c12d54f0:e74b3a0b
> Name : rheinberg:raid5
> Creation Time : Sat Oct 31 20:43:47 2009
> Raid Level : raid5
> Raid Devices : 3
>
> Avail Dev Size : 1937872480 (924.05 GiB 992.19 GB)
> Array Size : 3875744768 (1848.10 GiB 1984.38 GB)
> Used Dev Size : 1937872384 (924.05 GiB 992.19 GB)
> Super Offset : 1937872736 sectors
> State : clean
> Device UUID : 064806c1:0fc61a78:b3df4fed:840743e4
>
> Update Time : Thu Dec 17 01:59:54 2009
> Checksum : 654ad2e3 - correct
> Events : 114
>
> Layout : left-symmetric
> Chunk Size : 256K
>
> Array Slot : 3 (0, 1, failed, 2)
> Array State : uuU 1 failed
The index in the Array Slot with dev_number=2 is marked as failed and
does not exists on my RAID. My disk uses a disk-ID with value 3, which
can obviously grub2 not use as its array->index, because it expects here
disk 0, 1 or 2. Using the given dev_role array, which is appended at the
end of the superblock v1.x, solved this issue just be using it as a
translation table.
So the correct part of maraid_linux.c should be something like:
> array->total_devs = grub_le_to_cpu32 (sb_1x->raid_disks);
> array->disk_size = grub_le_to_cpu64 (sb_1x->size);
> array->chunk_size = grub_le_to_cpu32 (sb_1x->chunksize);
> if (grub_le_to_cpu32(sb_1x->dev_number) < grub_le_to_cpu32(sb_1x->max_dev))
> array->index = grub_le_to_cpu16 (sb_1x->dev_roles[
> grub_le_to_cpu32(sb_1x->dev_number) ]);
> else
> array->index = 0xffff; // disk will be later not used !
> array->uuid_len = 16;
With all these changes I could boot my machine also with anyone of the 3
disks removed.
Additionally I attached a corrected patch for using meta data 1.0, which
should work with
- grub2_1.97+20091125-1
- grub2_1.97+20091130-1
- grub2_1.98~20091210-1
by just adding that file into the /debian/patches directory.
Note that I haven't tested it with md meta data version > 1.0, e.g. 1.1
or 1.2 and also not with other RAID levels!
Peter
diff -Nur grub2-1.97+20091125/ChangeLog.raid grub2-1.97+20091129/ChangeLog.raid
--- ChangeLog.raid 1970-01-01 01:00:00.000000000 +0100
+++ ChangeLog.raid 2009-11-29 00:35:06.000000000 +0100
@@ -0,0 +1,30 @@
+2009-12-15 Peter Henn <peter.h...@web.de>
+
+ * bugfix the calculation of the RAID superblock version 1.x
+ chunk size and disk size, which are already given as number of
+ sectors and therefore not comparable to the RAID superblock
+ version 0.9x values
+ * bugfix fetching the correct device number from role table
+ instead of using the table index
+
+2009-11-16 Felix Zielcke <fziel...@z-51.de>
+
+ * disk/mdraid_linux.c (grub_mdraid_detect): Remove a wrong call
+ of free().
+
+2009-11-16 Felix Zielcke <fziel...@z-51.de>
+
+ * disk/mdraid_linux.c (grub_mdraid_detect): Fix the unsupported
+ RAID version error with metadata 1.x.
+
+2009-11-06 Felix Zielcke <fziel...@z-51.de>
+
+ * disk/dmraid_nvidia.c (grub_dmraid_nv_detect): Set array->name to NULL.
+ * disk/mdraid_linux.c (grub_raid_super_1x): New structure.
+ (WriteMostly1): New macro.
+ Set array->name to NULL for metadata format 0.90. Add support for
+ metadata 1.x. Fix some comments.
+ * disk/raid.c (): Add support for name based RAID arrays. Fix a
+ few comments.
+ * util/getroot.c (grub_util_get_grub_dev): Add support for
+ /dev/md/name style devices.
diff -Nur grub2-1.97+20091125/disk/dmraid_nvidia.c
grub2-1.97+20091129/disk/dmraid_nvidia.c
--- disk/dmraid_nvidia.c 2009-11-25 19:21:55.000000000 +0100
+++ disk/dmraid_nvidia.c 2009-11-29 01:05:07.000000000 +0100
@@ -1,7 +1,7 @@
/* dmraid_nvidia.c - module to handle Nvidia fakeraid. */
/*
* GRUB -- GRand Unified Bootloader
- * Copyright (C) 2006,2007,2008 Free Software Foundation, Inc.
+ * Copyright (C) 2006,2007,2008,2009 Free Software Foundation, Inc.
*
* GRUB is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
@@ -132,6 +132,7 @@
"Unsupported RAID level: %d", sb.array.raid_level);
}
+ array->name = NULL;
array->number = 0;
array->total_devs = sb.array.total_volumes;
array->chunk_size = sb.array.stripe_block_size;
diff -Nur grub2-1.97+20091125/disk/mdraid_linux.c
grub2-1.97+20091129/disk/mdraid_linux.c
--- disk/mdraid_linux.c 2009-11-25 19:21:55.000000000 +0100
+++ disk/mdraid_linux.c 2009-11-29 01:07:19.000000000 +0100
@@ -1,7 +1,7 @@
-/* mdraid_linux.c - module to handle linux softraid. */
+/* mdraid_linux.c - module to handle Linux Software RAID. */
/*
* GRUB -- GRand Unified Bootloader
- * Copyright (C) 2008 Free Software Foundation, Inc.
+ * Copyright (C) 2008,2009 Free Software Foundation, Inc.
*
* GRUB is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
@@ -159,32 +159,146 @@
struct grub_raid_disk_09 this_disk;
} __attribute__ ((packed));
+/*
+ * The version-1 superblock :
+ * All numeric fields are little-endian.
+ *
+ * Total size: 256 bytes plus 2 per device.
+ * 1K allows 384 devices.
+ */
+
+struct grub_raid_super_1x
+{
+ /* Constant array information - 128 bytes. */
+ grub_uint32_t magic; /* MD_SB_MAGIC: 0xa92b4efc - little endian. */
+ grub_uint32_t major_version; /* 1. */
+ grub_uint32_t feature_map; /* Bit 0 set if 'bitmap_offset' is meaningful.
*/
+ grub_uint32_t pad0; /* Always set to 0 when writing. */
+
+ grub_uint8_t set_uuid[16]; /* User-space generated. */
+ char set_name[32]; /* Set and interpreted by user-space. */
+
+ grub_uint64_t ctime; /* Lo 40 bits are seconds, top 24 are
microseconds or 0. */
+ grub_uint32_t level; /* -4 (multipath), -1 (linear), 0,1,4,5. */
+ grub_uint32_t layout; /* only for raid5 and raid10 currently.
*/
+ grub_uint64_t size; /* Used size of component devices, in 512byte
sectors. */
+
+ grub_uint32_t chunksize; /* In 512byte sectors. */
+ grub_uint32_t raid_disks;
+ grub_uint32_t bitmap_offset; /* Sectors after start of superblock that
bitmap starts
+ * NOTE: signed, so bitmap can be before
superblock
+ * only meaningful of feature_map[0] is set.
+ */
+
+ /* These are only valid with feature bit '4'. */
+ grub_uint32_t new_level; /* New level we are reshaping to. */
+ grub_uint64_t reshape_position; /* Next address in array-space for
reshape. */
+ grub_uint32_t delta_disks; /* Change in number of raid_disks. */
+ grub_uint32_t new_layout; /* New layout. */
+ grub_uint32_t new_chunk; /* New chunk size (512byte sectors). */
+ grub_uint8_t pad1[128 - 124]; /* Set to 0 when written. */
+
+ /* Constant this-device information - 64 bytes. */
+ grub_uint64_t data_offset; /* Sector start of data, often 0. */
+ grub_uint64_t data_size; /* Sectors in this device that can be used for
data. */
+ grub_uint64_t super_offset; /* Sector start of this superblock. */
+ grub_uint64_t recovery_offset; /* Sectors before this offset (from
data_offset) have been recovered. */
+ grub_uint32_t dev_number; /* Permanent identifier of this device - not
role in raid. */
+ grub_uint32_t cnt_corrected_read; /* Number of read errors that were
corrected by re-writing. */
+ grub_uint8_t device_uuid[16]; /* User-space setable, ignored by
kernel. */
+ grub_uint8_t devflags; /* Per-device flags. Only one defined... */
+ grub_uint8_t pad2[64 - 57]; /* Set to 0 when writing. */
+
+ /* Array state information - 64 bytes. */
+ grub_uint64_t utime; /* 40 bits second, 24 btes microseconds. */
+ grub_uint64_t events; /* Incremented when superblock updated.
*/
+ grub_uint64_t resync_offset; /* Data before this offset (from data_offset)
known to be in sync. */
+ grub_uint32_t sb_csum; /* Checksum upto devs[max_dev]. */
+ grub_uint32_t max_dev; /* Size of devs[] array to consider. */
+ grub_uint8_t pad3[64 - 32]; /* Set to 0 when writing. */
+
+ /* Device state information. Indexed by dev_number.
+ * 2 bytes per device.
+ * Note there are no per-device state flags. State information is rolled
+ * into the 'roles' value. If a device is spare or faulty, then it doesn't
+ * have a meaningful role.
+ */
+ grub_uint16_t dev_roles[0]; /* Role in array, or 0xffff for a spare, or
0xfffe for faulty. */
+} __attribute__ ((packed));
+
+#define WriteMostly1 1 /* Mask for writemostly flag in above devflags.
*/
+
static grub_err_t
grub_mdraid_detect (grub_disk_t disk, struct grub_raid_array *array)
{
grub_disk_addr_t sector;
- grub_uint64_t size;
+ grub_uint64_t size, sb_size;
struct grub_raid_super_09 sb;
+ struct grub_raid_super_1x *sb_1x;
grub_uint32_t *uuid;
+ grub_uint8_t minor_version;
- /* The sector where the RAID superblock is stored, if available. */
+ /* The sector where the mdraid 0.90 superblock is stored, if available. */
size = grub_disk_get_size (disk);
sector = NEW_SIZE_SECTORS (size);
if (grub_disk_read (disk, sector, 0, SB_BYTES, &sb))
return grub_errno;
- /* Look whether there is a RAID superblock. */
- if (sb.md_magic != SB_MAGIC)
+ /* Look whether there is a mdraid 0.90 superblock. */
+ if (sb.md_magic == SB_MAGIC)
+ goto superblock_0_90;
+
+ /* Check for an 1.x superblock.
+ * It's always aligned to a 4K boundary
+ * and depending on the minor version it can be:
+ * 0: At least 8K, but less than 12K, from end of device
+ * 1: At start of device
+ * 2: 4K from start of device.
+ */
+
+ sb_1x = grub_malloc (sizeof (struct grub_raid_super_1x));
+ if (!sb_1x)
+ return grub_errno;
+
+ for (minor_version = 0; minor_version < 3; ++minor_version)
+ {
+ switch (minor_version)
+ {
+ case 0:
+ sector = (size - 8 * 2) & ~(4 * 2 - 1);
+ break;
+ case 1:
+ sector = 0;
+ break;
+ case 2:
+ sector = 4 * 2;
+ break;
+ }
+
+ if (grub_disk_read
+ (disk, sector, 0, sizeof (struct grub_raid_super_1x), sb_1x))
+ {
+ grub_free (sb_1x);
+ return grub_errno;
+ }
+
+ if (sb_1x->magic == SB_MAGIC)
+ goto superblock_1_x;
+ }
+
+ /* Neither 0.90 nor 1.x. */
+ if (grub_le_to_cpu32 (sb_1x->magic) != SB_MAGIC)
return grub_error (GRUB_ERR_OUT_OF_RANGE, "not raid");
- /* FIXME: Also support version 1.0. */
+superblock_0_90:
+
if (sb.major_version != 0 || sb.minor_version != 90)
return grub_error (GRUB_ERR_NOT_IMPLEMENTED_YET,
"Unsupported RAID version: %d.%d",
sb.major_version, sb.minor_version);
- /* FIXME: Check the checksum. */
+ /* FIXME: Check the checksum. */
/* Multipath. */
if ((int) sb.level == -4)
@@ -195,6 +309,7 @@
return grub_error (GRUB_ERR_NOT_IMPLEMENTED_YET,
"Unsupported RAID level: %d", sb.level);
+ array->name = NULL;
array->number = sb.md_minor;
array->level = sb.level;
array->layout = sb.layout;
@@ -205,7 +320,7 @@
array->uuid_len = 16;
array->uuid = grub_malloc (16);
if (!array->uuid)
- return grub_errno;
+ return grub_errno;
uuid = (grub_uint32_t *) array->uuid;
uuid[0] = sb.set_uuid0;
@@ -214,6 +329,66 @@
uuid[3] = sb.set_uuid3;
return 0;
+
+ superblock_1_x:
+
+ if (sb_1x->major_version != 1)
+ return grub_error (GRUB_ERR_NOT_IMPLEMENTED_YET,
+ "Unsupported RAID version: %d",
+ sb_1x->major_version);
+ /* Multipath. */
+ if ((int) sb_1x->level == -4)
+ sb_1x->level = 1;
+
+ if (sb_1x->level != 0 && sb_1x->level != 1 && sb_1x->level != 4 &&
+ sb_1x->level != 5 && sb_1x->level != 6 && sb_1x->level != 10)
+ {
+ return grub_error (GRUB_ERR_NOT_IMPLEMENTED_YET,
+ "Unsupported RAID level: %d", sb_1x->level);
+ grub_free (sb_1x);
+ }
+ /* 1.x superblocks don't have a fixed size on disk. So we have to
+ read it again now that we now the max device count. */
+ sb_size = sizeof (struct grub_raid_super_1x) + 2 * grub_le_to_cpu32
(sb_1x->max_dev);
+ sb_1x = grub_realloc (sb_1x, sb_size);
+ if (! sb_1x)
+ return grub_errno;
+
+ if (grub_disk_read (disk, sector, 0, sb_size, sb_1x))
+ {
+ grub_free (sb_1x);
+ return grub_errno;
+ }
+
+ array->name = grub_strdup (sb_1x->set_name);
+ if (! array->name)
+ {
+ grub_free (sb_1x);
+ return grub_errno;
+ }
+
+ array->number = 0;
+ array->level = grub_le_to_cpu32 (sb_1x->level);
+ array->layout = grub_le_to_cpu32 (sb_1x->layout);
+ array->total_devs = grub_le_to_cpu32 (sb_1x->raid_disks);
+ array->disk_size = grub_le_to_cpu64 (sb_1x->size);
+ array->chunk_size = grub_le_to_cpu32 (sb_1x->chunksize);
+ if (grub_le_to_cpu32(sb_1x->dev_number) < grub_le_to_cpu32(sb_1x->max_dev))
+ array->index = grub_le_to_cpu16 (sb_1x->dev_roles[
grub_le_to_cpu32(sb_1x->dev_number) ]);
+ else
+ array->index = 0xffff; // disk will be later not used !
+ array->uuid_len = 16;
+ array->uuid = grub_malloc (16);
+ if (!array->uuid)
+ {
+ grub_free (sb_1x);
+ return grub_errno;
+ }
+
+ grub_memcpy (array->uuid, sb_1x->set_uuid, 16);
+
+ grub_free (sb_1x);
+ return 0;
}
static struct grub_raid grub_mdraid_dev = {
diff -Nur grub2-1.97+20091125/disk/raid.c grub2-1.97+20091129/disk/raid.c
--- disk/raid.c 2009-11-25 19:21:55.000000000 +0100
+++ disk/raid.c 2009-11-29 01:11:41.000000000 +0100
@@ -1,7 +1,7 @@
/* raid.c - module to read RAID arrays. */
/*
* GRUB -- GRand Unified Bootloader
- * Copyright (C) 2006,2007,2008 Free Software Foundation, Inc.
+ * Copyright (C) 2006,2007,2008,2009 Free Software Foundation, Inc.
*
* GRUB is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
@@ -525,6 +525,8 @@
array->nr_devs = 0;
grub_memset (&array->device, 0, sizeof (array->device));
+ if (array->name)
+ goto skip_duplicate_check;
/* Check whether we don't have multiple arrays with the same number. */
for (p = array_list; p != NULL; p = p->next)
{
@@ -534,38 +536,44 @@
if (p)
{
- /* The number is already in use, so we need to find an new number. */
+ /* The number is already in use, so we need to find a new one. */
int i = 0;
- while (1)
- {
- for (p = array_list; p != NULL; p = p->next)
- {
- if (p->number == i)
- break;
- }
-
- if (!p)
- {
- /* We found an unused number. */
- array->number = i;
- break;
- }
-
- i++;
- }
- }
-
- array->name = grub_malloc (13);
+ while (1)
+ {
+ for (p = array_list; p != NULL; p = p->next)
+ {
+ if (p->number == i)
+ break;
+ }
+
+ if (! p)
+ {
+ /* We found an unused number. */
+ array->number = i;
+ break;
+ }
+
+ i++;
+ }
+ }
+ skip_duplicate_check:
+ /* mdraid 1.x superblocks have only a name stored not a number.
+ Use it directly as GRUB device. */
if (! array->name)
- {
- grub_free (array->uuid);
- grub_free (array);
-
- return grub_errno;
- }
+ {
+ array->name = grub_malloc (13);
+ if (! array->name)
+ {
+ grub_free (array->uuid);
+ grub_free (array);
- grub_sprintf (array->name, "md%d", array->number);
+ return grub_errno;
+ }
+ grub_sprintf (array->name, "md%d", array->number);
+ }
+ else
+ grub_sprintf (array->name, "%s", array->name);
grub_dprintf ("raid", "Found array %s (%s)\n", array->name,
scanner_name);
diff -Nur grub2-1.97+20091125/util/getroot.c grub2-1.97+20091129/util/getroot.c
--- util/getroot.c 2009-11-25 19:21:55.000000000 +0100
+++ util/getroot.c 2009-11-29 01:03:52.000000000 +0100
@@ -590,6 +590,20 @@
asprintf (&grub_dev, "md%s", p);
free (p);
}
+ else if (os_dev[7] == '/')
+ {
+ /* mdraid 1.x with a free name. */
+ char *p , *q;
+
+ p = strdup (os_dev + sizeof ("/dev/md/") - 1);
+
+ q = strchr (p, 'p');
+ if (q)
+ *q = ',';
+
+ asprintf (&grub_dev, "%s", p);
+ free (p);
+ }
else
grub_util_error ("Unknown kind of RAID device `%s'", os_dev);
