2.6.19 cannot find root partition
Hi,
The patch to identify AIX disks and ignore them has caused at least one machine to fail to find the root partition on 2.6.19. The patch is:
http://lkml.org/lkml/2006/7/31/117
The problem is some disk formatters do not blow away the first 4 bytes of the disk. If the disk we are installing to used to have AIX on it, then the first 4 bytes will still have IBMA in EBCDIC.
The install in question was debian etch. Im not sure what the best fix is, perhaps the AIX detection code could check more than the first 4 bytes.
Anton
Quoted 3 lines · Anton Blanchard
The install in question was debian etch. Im not sure what the best fix is, perhaps the AIX detection code could check more than the first 4 bytes.
Can you send me a dump of the first block?
Quoted 8 lines · Anton Blanchard
Hi,
The patch to identify AIX disks and ignore them has caused at least one machine to fail to find the root partition on 2.6.19. The patch is:
Quoted 7 lines
The problem is some disk formatters do not blow away the first 4 bytes of the disk. If the disk we are installing to used to have AIX on it, then the first 4 bytes will still have IBMA in EBCDIC.
The install in question was debian etch. Im not sure what the best fix is, perhaps the AIX detection code could check more than the first 4 bytes.
Yeah, I had this problem on paulus' old PReP machine - I had to manually zap the AIX magic number to get it to see the perfectly good partitions.
Signature · 4 lines
David Gibson | I'll have my music baroque, and my code david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_ | _way_ _around_! http://www.ozlabs.org/~dgibson
Quoted 3 lines · David Gibson
Yeah, I had this problem on paulus' old PReP machine - I had to manually zap the AIX magic number to get it to see the perfectly good partitions.
Something like this in aix_label_present should do it, beside echo > /dev/sda
struct partition *pt = (struct partition *) (p + 0x1be);
int slot;
for (slot = 1; slot <= 4; slot++, pt++) {
if (pt->sys_ind == LINUX_SWAP_PARTITION ||
pt->sys_ind == LINUX_RAID_PARTITION ||
pt->sys_ind == 0x83 ||
pt->sys_ind == 0x8e ||
is_extended_partition(pt))
return 0;
} PATCH 1/1 relax check for AIX in msdos partition table
In reply to David GibsonQuoted 11 lines · David Gibson
Quoted 9 lines · Anton Blanchard
The patch to identify AIX disks and ignore them has caused at least one machine to fail to find the root partition on 2.6.19. The patch is:
http://lkml.org/lkml/2006/7/31/117
The problem is some disk formatters do not blow away the first 4 bytes of the disk. If the disk we are installing to used to have AIX on it, then the first 4 bytes will still have IBMA in EBCDIC.
Quoted 7 lines
Quoted 3 lines
The install in question was debian etch. Im not sure what the best fix is, perhaps the AIX detection code could check more than the first 4 bytes.
Yeah, I had this problem on paulus' old PReP machine - I had to manually zap the AIX magic number to get it to see the perfectly good partitions.
What pre-historic fdisk was used to create the partition table? At least the current version of fdisk does not let me manipulate the existing partition table before I remove the AIX marker. The patch below should fix it. Partition table entries 2,3,4 from an AIX 4.3 install have type 0x41.
Also, the whole partition info for primary partitions is in this block:
dd if=/dev/sdb count=$(( 4 * 16 )) bs=1 skip=$(( 0x1be ))
All other data do not matter, beside the 0x55aa marker at the end of the first block.
--- fs/partitions/msdos.c | 12 +++++++++++- include/linux/genhd.h | 2 ++ 2 files changed, 13 insertions(+), 1 deletion(-)
fs/partitions/msdos.c +11 -1
@@ -63,15 +63,25 @@ msdos_magic_present(unsigned char *p)#define AIX_LABEL_MAGIC4 0xC1static int aix_magic_present (unsigned char *p,struct block_device *bdev) { +struct partition *pt = (struct partition *) (p +0x1be );Sector sect;unsigned char *d; -int ret =0 ; +int slot, ret =0 ;if (p[0 ] !=AIX_LABEL_MAGIC1 && p[1 ] !=AIX_LABEL_MAGIC2 && p[2 ] !=AIX_LABEL_MAGIC3 && p[3 ] !=AIX_LABEL_MAGIC4 )return 0 ; +/* Assume the partition table is valid if Linux partitions exists */ +for (slot =1 ; slot <=4 ; slot++, pt++) { +if (pt->sys_ind ==LINUX_SWAP_PARTITION || + pt->sys_ind ==LINUX_RAID_PARTITION || + pt->sys_ind ==LINUX_DATA_PARTITION || + pt->sys_ind ==LINUX_LVM_PARTITION || +is_extended_partition (pt)) +return 0 ; + } d =read_dev_sector (bdev,7 , §);if (d) {if (d[0 ] =='_' && d[1 ] =='L' && d[2 ] =='V' && d[3 ] =='M' )
include/linux/genhd.h +2 -0
@@ -21,6 +21,8 @@ enum {WIN98_EXTENDED_PARTITION =0x0f ,LINUX_SWAP_PARTITION =0x82 , +LINUX_DATA_PARTITION =0x83 , +LINUX_LVM_PARTITION =0x8e ,LINUX_RAID_PARTITION =0xfd ,/* autodetect RAID partition */ SOLARIS_X86_PARTITION = LINUX_SWAP_PARTITION,
PATCH 1/1 fix logic error in AIX detection
Correct the AIX magic check to let 'echo > /dev/sdb' actually work.
--- fs/partitions/msdos.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)
fs/partitions/msdos.c +4 -4
@@ -68,10 +68,10 @@ static int aix_magic_present(unsigned chunsigned char *d;int slot, ret =0 ; -if (p[0 ] !=AIX_LABEL_MAGIC1 && - p[1 ] !=AIX_LABEL_MAGIC2 && - p[2 ] !=AIX_LABEL_MAGIC3 && - p[3 ] !=AIX_LABEL_MAGIC4 ) +if (!(p[0 ] ==AIX_LABEL_MAGIC1 && + p[1 ] ==AIX_LABEL_MAGIC2 && + p[2 ] ==AIX_LABEL_MAGIC3 && + p[3 ] ==AIX_LABEL_MAGIC4 ))return 0 ;/* Assume the partition table is valid if Linux partitions exists */ for (slot =1 ; slot <=4 ; slot++, pt++) {