From aedde5fe6b9fc426d8df569bc1df1f38a45cad2c Mon Sep 17 00:00:00 2001 From: espie <> Date: Thu, 24 Sep 1998 17:24:37 +0000 Subject: [PATCH] Proof-reading, plus Linux kernel patch for handling OpenBSD' disklabel. --- src/etc/etc.i386/INSTALL.linux | 417 +++++++++++++++++++++++---------- 1 file changed, 298 insertions(+), 119 deletions(-) diff --git a/src/etc/etc.i386/INSTALL.linux b/src/etc/etc.i386/INSTALL.linux index 7d082723..d16b6b11 100644 --- a/src/etc/etc.i386/INSTALL.linux +++ b/src/etc/etc.i386/INSTALL.linux @@ -1,29 +1,32 @@ -$OpenBSD: INSTALL.linux,v 1.2 1998/09/22 17:21:22 espie Exp $ +$OpenBSD: INSTALL.linux,v 1.3 1998/09/24 17:24:37 espie Exp $ Warning: this document is currently being reviewed. It's not yet complete, and probably contains loads of errors. As an example, I can't figure out -why linux doesn't need mkswap as it shares the exact same blocks with +why Linux doesn't need mkswap as it shares the exact same blocks with OpenBSD. - Linux + OpenBSD: it's possible + Linux + OpenBSD: it's possible - by Marc Espie -- Marc.Espie@openbsd.org + by Marc Espie -- Marc.Espie@OpenBSD.org -It is perfectly possible to have linux and openbsd on the same disk. -As of this writing, OpenBSD can read and write linux' partitions, whereas +It is perfectly possible to have Linux and OpenBSD on the same disk. +As of this writing, OpenBSD can read and write Linux' partitions, whereas Linux cannot access OpenBSD partitions (they differ from NetBSD partitions). You can even install OpenBSD from an ext2fs partition (choose install from disk... ext2fs does not appear in the choices, but `default' it is). -First, make a bootable floppy of linux. Then, you have to find some room +First, make a bootable floppy of Linux. Then, you have to find some room for OpenBSD. Don't worry about swap space: you can share Linux' swap partition. If you want to grab space from a Windows/DOS partition, use fips. Fips20 knows all about FAT32, so windows 95 is no longer a problem. +Other sources of information, especially concerning other BSD systems, +must be taken with a healthy does of scepticism. + First principles ---------------- OpenBSD does not use the DOS partitions for more than booting. You just @@ -43,7 +46,7 @@ partition and as a DOS partition is the root partition: it MUST begin on the same sector for the boot process to work. One way to visualize things is to picture OpenBSD embedded inside DOS -partitions: the DOS partition(s) used by OpenBSD may each contain several +partitions: the DOS partitions used by OpenBSD may each contain several OpenBSD partitions. As long as the DOS partition table has the right start and length for each partition it will be kept happy. @@ -51,22 +54,22 @@ The OpenBSD disklabel is just another mechanism that yield another description of the disk. It is vitally important that the BSD root partition start precisely where the corresponding DOS partition is supposed to start, and it is better when all BSD partitions stay inside their -DOS partitions boundary. Apart from OpenBSD partitions proper, the -BSD disklabel can yield a BSD description of other DOS partitions, but +DOS partitions boundaries. Apart from OpenBSD partitions proper, the +BSD disklabel can contain a BSD description of other DOS partitions, but this is not mandatory. If you don't have any constraints, having a correct -description of all partition is better, but with bigger disks, keep in mind +description of all partitions is better, but with bigger disks, keep in mind that OpenBSD disklabels can't hold more than 16 partitions. Contrarily to popular belief, OpenBSD does *NOT* need one contiguous chunk -of the disk. It is probably the simplest setup, but other considerations -(such as the need to boot several OS, and have several small primary -partitions that all start before cylinder 1024) may lead you to use -two chunks for OpenBSD. +of the disk (a `slice' in FreeBSD lingo). It is probably the simplest setup, +but other considerations (such as the need to boot several OS, and to have +several small primary partitions that all within the first 1024 BIOS +cylinders) may lead you to use two chunks for OpenBSD. If you can, it is MUCH better to devote a full disk to OpenBSD: this limits the number of mistakes you can do. Admittedly there are some cases where this isn't a option (my machine is a laptop... I have to cope with the -harddisk I have). +hard disk I have). Mapping your disk ----------------- @@ -74,25 +77,24 @@ Starting from Linux, get a grasp of your partitions. Use df to check which is what, then fdisk to get the actual setup of the disk. Here is my disk: -- - Disk /dev/hda: 128 heads, 63 sectors, 993 cylinders Units = cylinders of 8064 * 512 bytes - Device Boot Begin Start End Blocks Id System - /dev/hda1 1 1 260 1048288+ 6 DOS 16-bit >=32M - /dev/hda2 261 261 273 52416 83 Linux native - /dev/hda4 287 287 600 1270080 5 Extended - /dev/hda5 287 287 303 68512 82 Linux swap - /dev/hda6 304 304 456 616864+ 83 Linux native - /dev/hda7 457 457 520 258016+ 83 Linux native - /dev/hda8 521 521 537 68512+ 83 Linux native - /dev/hda9 538 538 601 258016+ 83 Linux native - -(In case you're wondering, yes this is a big disk. The linux playground is -large, the OpenBSD area will be huge. As a developper, I usually have loads + Device Boot Begin Start End Blocks Id System +/dev/hda1 1 1 260 1048288+ 6 DOS 16-bit >=32M +/dev/hda2 261 261 273 52416 83 Linux native +/dev/hda4 287 287 601 1270080 5 Extended +/dev/hda5 287 287 303 68512 82 Linux swap +/dev/hda6 304 304 456 616864+ 83 Linux native +/dev/hda7 457 457 520 258016+ 83 Linux native +/dev/hda8 521 521 537 68512+ 83 Linux native +/dev/hda9 538 538 601 258016+ 83 Linux native + +(In case you're wondering, yes this is a big disk. The Linux playground is +large, the OpenBSD area will be huge. As a developer, I usually have loads of source & binaries lying around... a simple OpenBSD installation can fit -within 300 Mb with room to spare). -In my setup hda2 is /, hda6 is /usr, hda7 is /usr/local hda8 is /var, and +within 300 Mb with room to spare.) +In my setup hda2 is /, hda6 is /usr, hda7 is /usr/local, hda8 is /var, and hda9 is /home. Get the display to sectors with u, and jot down the corresponding @@ -101,15 +103,15 @@ information as well: Disk /dev/hda: 128 heads, 63 sectors, 993 cylinders Units = sectors of 1 * 512 bytes - Device Boot Begin Start End Blocks Id System - /dev/hda1 63 63 2096639 1048288+ 6 DOS 16-bit >=32M - /dev/hda2 2096640 2096640 2201471 52416 83 Linux native - /dev/hda4 2306304 2306304 4846464 1270080 5 Extended - /dev/hda5 2306368 2306368 2443391 68512 82 Linux swap - /dev/hda6 2443455 2443455 3677183 616864+ 83 Linux native - /dev/hda7 3677247 3677247 4193279 258016+ 83 Linux native - /dev/hda8 4193343 4193343 4330367 68512+ 83 Linux native - /dev/hda9 4330431 4330431 4846463 258016+ 83 Linux native + Device Boot Begin Start End Blocks Id System +/dev/hda1 63 63 2096639 1048288+ 6 DOS 16-bit >=32M +/dev/hda2 2096640 2096640 2201471 52416 83 Linux native +/dev/hda4 2306304 2306304 4846463 1270080 5 Extended +/dev/hda5 2306368 2306368 2443391 68512 82 Linux swap +/dev/hda6 2443455 2443455 3677183 616864+ 83 Linux native +/dev/hda7 3677247 3677247 4193279 258016+ 83 Linux native +/dev/hda8 4193343 4193343 4330367 68512+ 83 Linux native +/dev/hda9 4330431 4330431 4846463 258016+ 83 Linux native Okay, finally switch to expert mode, and note the corresponding data. Disk /dev/hda: 128 heads, 63 sectors, 993 cylinders @@ -127,77 +129,99 @@ Nr AF Hd Sec Cyl Hd Sec Cyl Start Size ID Note that this is STILL the same data. The good point about this last display is that it is almost what you're going to see in OpenBSD fdisk ! -The only difference is that: -- OpenBSD fdisk displays cyl/hd/sec, which is a more sensible order. -- OpenBSD fdisk displays `real' offsets from the beginning of the disk. + +There are some differences though, mostly because Linux fdisk has made +some rather confusing choices: +- in simple mode it starts numbering cylinders at 1... whereas +everything else starts from 0. +- in simple mode it shows blocks of 1024 bytes, which makes for half-blocks +(marked with a +) and sizes halved from the real block size. +- in expert mode it shows extended partitions offset from the start +of the extended partition. +- the hd/sec/cyl is a confusing order, as the sector number is computed +from cyl/hd/sec, in that order. +- it never shows and doesn't care about the real disk geometry. You will notice that the 3rd primary partition is empty... this is where I intend to stick my OpenBSD root partition (both DOS and BSD partitions), -and that I have cylinders 602-993 empty... this is where I intend to +and that I have left cylinders 601--992 empty... this is where I intend to stick the rest of OpenBSD. +After OpenBSD is installed, partition 3 will show up as: +/dev/hda3 274 274 286 52416 a6 Unknown +/dev/hda3 2201472 2201472 2306303 52416 a6 Unknown + +(recent Linux fdisk will display a6 as OpenBSD, but they still don't know +how to deal with the disklabel. Important: NEVER use Linux fdisk to fiddle +with OpenBSD disklabels.) + Before starting to install OpenBSD, now would be a good time to check the INSTALL.pt document... Especially note the alignment restriction of partitions (first sector of a partition must be at head 0, sector 1 of a -cylinder). This is enforced by linux' fdisk. +cylinder). This is enforced by Linux' fdisk. The other point to note is that extended partitions are actually linked -lists. This will also show up in OpenBSD' fdisk. +lists. This will show up in OpenBSD' fdisk. Your clock and OpenBSD ---------------------- OpenBSD expects your hardware clock to be in universal time, and uses -timezones to give you local time. With linux, this depends... +time zones to give you local time. With Linux, this depends... most distributions use a small program called hwclock to set up the system time from the hardware clock when booting... there is a --utc option if your hardware clock is in universal time, but this is not always what happens by default. -Check your distribution, on redhat 5.1, this occurs in /etc/rc.d/rc.sysinit -which loads /etc/sysconfig/clock which defines a variable called UTC before -calling hwclock. -- ensure UTC is set to true +For instance, on a redhat 5.1 system, this happens in /etc/rc.d/rc.sysinit +which loads an /etc/sysconfig/clock that defines a variable called UTC, and +then proceeds calling hwclock. +- ensure UTC is set to true, - adjust your hardware clock from the system time if necessary, e.g., -hwclock --systohc --utc - -Linux fs and OpenBSD --------------------- -There is a problem with many linux rc that do mount all filesystems even -in single-user mode. After you've installed OpenBSD, if your linux kernel -knows about bsd disklabels, it may insert lots of bsd partitions in its -list. Then at the next reboot, you're in trouble. The simplest way around -that is probably to make sure you have a linux kernel around that doesn't -know about disklabels. Otherwise, you may wish to check your inittab -and your rc to make deadly sure that single-user boot will work. +hwclock --systohc --utc. + +The Linux partition table and OpenBSD +------------------------------------- +There is a problem with many Linux rc that do mount all file systems even +in single-user mode. After you've installed OpenBSD, if your Linux kernel +knows about BSD disklabels, it may insert lots of BSD partitions in its +list. Then, at the next reboot, you're in trouble. The simplest way around +this problem is probably to make sure you can boot from a Linux kernel +that doesn't know about disklabels. Otherwise, you may wish to check +your inittab and your rc to make deadly sure that single-user boot +will work. The OpenBSD installation ------------------------ If you've got the space, you can install from your ext2fs partitions. This -is what I did, as I have a slip connexion to the rest of the world, and +is what I did, as I have a slip connection to the rest of the world, and the OpenBSD install floppy does not include slip. REMEMBER TO BACKUP ALL IMPORTANT DATA ON YOUR DISK BEFORE DOING THE INSTALLATION !!! So you cp floppy*.fs /dev/fd0, then reboot... -After a while, you will see your disk configuration scroll up. -I got: -wd0 at wdc0 drive 0: -wd0: 3909MB, 7944 cyl, 16 head, 63 sec, 512 bytes/sec, 8007552 sec total -wd0: using 16-sector 16-bit pio transfers, lba addressing -at which point I got somewhat confused, especially as this is a config with -more than 1024 cylinders. In fact, this is the real disk geometry, when -you inquire about it, but the geometry that the BIOS sees IS the fdisk + +First, the BSD kernel + ram disk loads, then there is a boot prompt, and +five seconds later, the boot proper starts. + +After a while, you will see your disk configuration scroll by. I got: + wd0 at wdc0 drive 0: + wd0: 3909MB, 7944 cyl, 16 head, 63 sec, 512 bytes/sec, 8007552 sec total + wd0: using 16-sector 16-bit pio transfers, lba addressing + +at which point I got somewhat confused, especially as this is a configuration with +more than 1024 cylinders. In fact, this is the actual disk geometry, when +you inquire about it, but the geometry that the BIOS does see IS the fdisk geometry, with 993 cylinders. As the 1024 cylinders is purely a BIOS limitation, there is no actual trouble. -If you want to be sure, right after the machine loads the kernel, there is -a boot prompt: -boot> -at which point you can enter: machine diskconfig + +If you want to be sure, enquire at the boot prompt: +boot>machine diskconfig before the automatic boot sequence continues. + Then I got into fdisk, and I proceeded to enter my new OpenBSD partition. -This is what the fdisk dump looked after I added the OpenBSD partition: +This is what the fdisk dump looked after my changes: Disk: wd0 geometry: 992/128/63 [7999488 sectors] Offset: 0 Signatures: 0xAA55,0x0 @@ -269,18 +293,17 @@ jot down whatever you need, add the OpenBSD partition to look like you want it to, and save everything. After you leave fdisk, you get to the interesting part: the disklabel -itself. If all goes well, OpenBSD synthetizes a nice disklabel out of what +itself. If all goes well, OpenBSD synthesizes a nice disklabel out of what it can deduce from the disk, including the ext2fs partitions. There are only a few subtleties to take care of: - initially, you can ONLY edit the disklabel part that matches the OpenBSD -DOS partition (a `slice' in FreeBSD linguo). You can use b 0 * to be able -to edit the whole disk, -- the real disk geometry becomes relevant. OpenBSD file system can't use -partial cylinder groups, hence bsd partitions should start on cylinder -boundaries, any remaining sectors will be lost anyway. -- units for size and offset can be given as sectors (default) or cylinders -(suffix c). +DOS partition (a `slice' in FreeBSD lingo). You have to use b 0 * before +you can edit the whole disk. +- the real disk geometry becomes relevant. The Berkeley fast file system +can't use partial cylinder groups, hence BSD partitions should start +on cylinder boundaries, as any remaining sectors will be lost anyway. +- units for size and offset can be given as sectors (default) or cylinders. After edition, this is what my disklabel looks like: @@ -322,30 +345,33 @@ drivedata: 0 o: 516033 4330431 ext2fs # (Cyl. 4296*- 4807*) Things to check: -- this disklabel is saved in MBR2 (basic dos partition 2), as expected. -- all the bsd partitions proper are aligned on a cylinder boundary. +- this disklabel is saved in MBR2 (basic DOS partition 2), as expected. +- all the BSD partitions proper are aligned on a cylinder boundary. the root partition begins at the precise same offset the corresponding DOS partition begins, and it extends for the same length. -Other bsd partitions don't show up in the dos partition setup, hence they +Other BSD partitions don't show up in the DOS partition setup, hence they begin precisely on cylinder 601/4808. - the ext2fs partitions have the exact same layout under the OpenBSD disklabel. One point that is somewhat laborious is that the disklabel -E mode (which you are currently using) tends to move partitions around to ensure -that ALL partitions fit are contiguous. Hence, you may need some fiddling -and printing around to ensure that Linux partitions do show up where they +that ALL defined partitions are contiguous. Hence, you may need some fiddling +around and printing to ensure that Linux partitions do show up where they should. In my case, disklabel moved the swap and all the ext2fs partitions slightly, and I add to adjust them manually... Once the disklabel is written to disk, the installation proceeds as usual. +ext2fs partitions are perfectly usable from OpenBSD. My /home partition +is ext2fs, I have been using it for a week now without any trouble. + Booting ------- -First time I booted my system back, I did not get into linux... I forgot -I had installed lilo in the master boot block, and lilo does not heed -the active partition flag. Fix was rather simple though. Once I was into -linux, I just had to edit lilo.conf to read: +First time I booted my system back, I did not get into OpenBSD as expected... +I plain forgot I had installed lilo in the master boot block, and lilo +does not heed the active partition flag. The fix was rather simple: from +the Linux system, I just had to edit lilo.conf to add the obsd entry: boot=/dev/hda map=/boot/map @@ -353,46 +379,199 @@ install=/boot/boot.b prompt timeout=50 image=/boot/myvmlinuz - label=linux - root=/dev/hda2 - vga=4 - read-only + label=linux + root=/dev/hda2 + vga=4 + read-only image=/boot/vmlinuz-2.0.34-1 - label=redhat - root=/dev/hda2 - read-only + label=redhat + root=/dev/hda2 + read-only other=/dev/hda1 - label=dos - table=/dev/hda + label=dos + table=/dev/hda other=/dev/hda3 - label=obsd - table=/dev/hda + label=obsd + table=/dev/hda -(changed were the last 3 lines) and rerun lilo, and voila, -OpenBSD was able to boot ! +rerun lilo, and voila, OpenBSD was able to boot ! Linux and OpenBSD partitions ---------------------------- -As of 2.0.35/2.1.122, linux does not support OpenBSD partitions. The way -the linux kernel works is reasonably straightforward, code in -drivers/block/genhd.c: when it detects a BSD disklabel, it does read it -and instantiates some new partitions, that will show up in that drive's -partition list. Unfortunately, that check is hardcoded for type A5 -(NetBSD/FreeBSD) disklabel, and it recognizes only 8 partitions disklabels +As of 2.0.35/2.1.122, Linux does not support OpenBSD partitions. The way +the Linux kernel works is reasonably straightforward (code in +drivers/block/genhd.c): when detecting a BSD partition (type A5), +it checks and parses the first block for a disklabel, and use that +information to instantiate some new partitions, +that will show up in that drive's partition list, exactly like extended +partitions do show up. Unfortunately, that check is hard coded for a type A5 +(NetBSD/FreeBSD) partition, and it recognizes 8 partitions disklabels only anyway. -Just hacking the linux code to recognize A6 is not quite enough. You also +Just hacking the Linux code to recognize A6 is not quite enough. You also need to trim down the resulting list of partitions, as your ext2fs partitions will now show up twice ! -I'm currently trying to implement some smart recognition scheme, where -the bsd disklabel would appear AFTER linux partitions, and where native -linux partitions wouldn't show up as duplicates in the disklabel. +I'm experimenting with the following patch, which should make it to linux-kernel +soon: +-------------------------------------------------------------- +diff -ur linux.orig/drivers/block/genhd.c linux/drivers/block/genhd.c +--- linux.orig/drivers/block/genhd.c Mon Aug 4 20:45:35 1997 ++++ linux/drivers/block/genhd.c Thu Sep 24 18:55:31 1998 +@@ -205,11 +206,46 @@ + } + + #ifdef CONFIG_BSD_DISKLABEL ++static void check_and_add_bsd_partition(struct gendisk *hd, struct bsd_partition *bsd_p) ++{ ++ struct hd_struct *lin_p; ++ /* check relative position of partitions. */ ++ for (lin_p = hd->part + 1; lin_p - hd->part < current_minor; lin_p++) { ++ /* no relationship -> try again */ ++ if (lin_p->start_sect + lin_p->nr_sects <= bsd_p->p_offset ++ || lin_p->start_sect >= bsd_p->p_offset + bsd_p->p_size) ++ continue; ++ /* equal -> no need to add */ ++ if (lin_p->start_sect == bsd_p->p_offset && ++ lin_p->nr_sects == bsd_p->p_size) ++ return; ++ /* bsd living within dos partition */ ++ if (lin_p->start_sect <= bsd_p->p_offset && lin_p->start_sect ++ + lin_p->nr_sects >= bsd_p->p_offset + bsd_p->p_size) { ++#ifdef DEBUG_BSD_DISKLABEL ++ printk("w: %d %ld+%ld,%d+%d", ++ lin_p - hd->part, lin_p->start_sect, lin_p->nr_sects, ++ bsd_p->p_offset, bsd_p->p_size); ++#endif ++ break; ++ } ++ /* ouch: bsd and linux overlap. Don't even try for that partition */ ++#ifdef DEBUG_BSD_DISKLABEL ++ printk("???: %d %ld+%ld,%d+%d", ++ lin_p - hd->part, lin_p->start_sect, lin_p->nr_sects, ++ bsd_p->p_offset, bsd_p->p_size); ++#endif ++ printk("???"); ++ return; ++ } /* if the bsd partition is not described by DOS, we end up there */ ++ add_partition(hd, current_minor, bsd_p->p_offset, bsd_p->p_size); ++ current_minor++; ++} + /* + * Create devices for BSD partitions listed in a disklabel, under a + * dos-like partition. See extended_partition() for more information. + */ +-static void bsd_disklabel_partition(struct gendisk *hd, kdev_t dev) ++static void bsd_disklabel_partition(struct gendisk *hd, kdev_t dev, int max_partitions) + { + struct buffer_head *bh; + struct bsd_disklabel *l; +@@ -225,19 +261,15 @@ + return; + } + +- p = &l->d_partitions[0]; +- while (p - &l->d_partitions[0] <= BSD_MAXPARTITIONS) { ++ if (l->d_npartitions < max_partitions) ++ max_partitions = l->d_npartitions; ++ for (p = l->d_partitions; p - l->d_partitions < max_partitions; p++) { + if ((current_minor & mask) >= (4 + hd->max_p)) + break; +- +- if (p->p_fstype != BSD_FS_UNUSED) { +- add_partition(hd, current_minor, p->p_offset, p->p_size); +- current_minor++; +- } +- p++; ++ if (p->p_fstype != BSD_FS_UNUSED) ++ check_and_add_bsd_partition(hd, p); + } + brelse(bh); +- + } + #endif + +@@ -248,6 +280,11 @@ + struct partition *p; + unsigned char *data; + int mask = (1 << hd->minor_shift) - 1; ++#ifdef CONFIG_BSD_DISKLABEL ++ /* no bsd disklabel as a default */ ++ kdev_t bsd_kdev = 0; ++ int bsd_maxpart; ++#endif + #ifdef CONFIG_BLK_DEV_IDE + int tested_for_xlate = 0; + +@@ -365,13 +402,29 @@ + hd->part[minor].nr_sects = 2; + } + #ifdef CONFIG_BSD_DISKLABEL ++ /* tag first disklabel for late recognition */ + if (SYS_IND(p) == BSD_PARTITION) { +- printk(" <"); +- bsd_disklabel_partition(hd, MKDEV(hd->major, minor)); +- printk(" >"); ++ printk("!"); ++ if (!bsd_kdev) { ++ bsd_kdev = MKDEV(hd->major, minor); ++ bsd_maxpart = BSD_MAXPARTITIONS; ++ } ++ } else if (SYS_IND(p) == OPENBSD_PARTITION) { ++ printk("!"); ++ if (!bsd_kdev) { ++ bsd_kdev = MKDEV(hd->major, minor); ++ bsd_maxpart = OPENBSD_MAXPARTITIONS; ++ } + } + #endif + } ++#ifdef CONFIG_BSD_DISKLABEL ++ if (bsd_kdev) { ++ printk(" <"); ++ bsd_disklabel_partition(hd, bsd_kdev, bsd_maxpart); ++ printk(" >"); ++ } ++#endif + /* + * Check for old-style Disk Manager partition table + */ +diff -ur linux.orig/include/linux/genhd.h linux/include/linux/genhd.h +--- linux.orig/include/linux/genhd.h Wed Mar 18 20:25:30 1998 ++++ linux/include/linux/genhd.h Thu Sep 24 18:55:43 1998 +@@ -69,12 +70,19 @@ + #ifdef CONFIG_BSD_DISKLABEL + /* + * BSD disklabel support by Yossi Gottlieb ++ * ++ * updated by Marc Espie + */ + ++/* check against BSD src/sys/sys/disklabel.h for consistency */ ++ + #define BSD_PARTITION 0xa5 /* Partition ID */ ++#define OPENBSD_PARTITION 0xa6 + + #define BSD_DISKMAGIC (0x82564557UL) /* The disk magic number */ + #define BSD_MAXPARTITIONS 8 ++#define OPENBSD_MAXPARTITIONS 16 ++ + #define BSD_FS_UNUSED 0 /* disklabel unused partition entry ID */ + struct bsd_disklabel { + __u32 d_magic; /* the magic number */ +---------------------------------------------------------------------- +It does: +- trigger the recognition of OpenBSD disklabels, +- force bsd partitions to appear at the end of the partition list, +- check the bsd disklabel for redundancy and consistency with the dos +partition table, so that the same partition does not show up twice. You also need a working ufs, such as -Usually, the patch must be done manually so that the new module will -compile. +This package yields a new module, plus a small patch to tie the module +with the rest of the Linux kernel. I would recommend checking that patch +manually, as Linux module information tends to vary widely, and it is +pretty trivial to add by hand anyway. Binary compatibility -------------------- -not yet written... +To be checked and written.