🚀 go-pugleaf

On 29 Oct 1996 14:18:08 GMT, Manuel Kolp <manuel@I_should_put_my_domain_in_etc_NNTP_INEWS_DOMAIN> wrote:

> Linux boots on /dev/hda2
> Disks are declared in the BIOS as follows 

>                   Type  Size   Cyl   Head  Precom  Landz Sector Mode                
> Primary Master:   User  1083   2100   16   65535   2039   63    Normal
> Primary Slave:    None   0      0      0    0       0      0    
> Secondary Master: User  640    1240   16   65535   1239   63    Normal
> Secondary Slave   None  0      0      0    0        0     0     
                                                                  ^^^^^^

> Linux detects our drives as follows: 

> hda: WDC AC21000H, 1033MB w/128kB Cache, LBA, CHS=2100/16/63
> hdb: FX300_01, ATAPI CDROM drive
> hdc: WDC AC2635F, 610MB w/64kB Cache, LBA, CHS=1240/16/63   
                                        ^^^
                                           \
Just a guess: This looks like a discrepancy here. Maybe you could try
switching the mode in the BIOS to LBA, or toggle the "use PCI bios"
option in the kernel.

Ed

-- 
Ed Doolittle <dolittle@math.toronto.edu>
 "Being on time is like being on fire."

Manuel Kolp (manuel@I_should_put_my_domain_in_etc_NNTP_INEWS_DOMAIN) wrote:
: Hi world, 

: We have a big hardisk problem on our Linux 2.0.21 box (P120 32M)

: Our fdisk repartition is the following

: Disk /dev/hda: 16 heads, 63 sectors, 2100 cylinders
: Units = cylinders of 1008 * 512 bytes

:    Device Boot   Begin    Start      End   Blocks   Id  System
: /dev/hda1            1        1      204   102784+  82  Linux swap
: /dev/hda2          205      205      611   205128   83  Linux native
: /dev/hda3          612      612     2100   750456   83  Linux native   

: Disk /dev/hdc: 16 heads, 63 sectors, 1240 cylinders
: Units = cylinders of 1008 * 512 bytes

:    Device Boot   Begin    Start      End   Blocks   Id  System
: /dev/hdc1            1        1      305   153688+   6  DOS 16-bit >=32M
: /dev/hdc2          306      306     1240   471240   83  Linux native  

: Linux boots on /dev/hda2
: Disks are declared in the BIOS as follows 

:                   Type  Size   Cyl   Head  Precom  Landz Sector Mode                
: Primary Master:   User  1083   2100   16   65535   2039   63    Normal
: Primary Slave:    None   0      0      0    0       0      0    
: Secondary Master: User  640    1240   16   65535   1239   63    Normal
: Secondary Slave   None  0      0      0    0        0     0     

: Linux detects our drives as follows: 

: hda: WDC AC21000H, 1033MB w/128kB Cache, LBA, CHS=2100/16/63
: hdb: FX300_01, ATAPI CDROM drive
: hdc: WDC AC2635F, 610MB w/64kB Cache, LBA, CHS=1240/16/63   

: We have also tried to configure our system following the large IDE Howto
: to fix the kernel's problem for IDE drives with over 1024 cylinders
: but we have also error messages :

: current->tss.cr3 = 00b32000, |r3 = 00b32000
: *pde = 00000000
: Oops: 0000
: CPU:    0
: EIP:    0010:[ext2_new_inode+1179/1612]
: EFLAGS: 00010286
: eax: 01e95c00   ebx: 0000aea1   ecx: ffff1602   edx: 01e94000
: esi: 001d98ec   edi: 21b37b32   ebp: 01ddf6c8   esp: 01c43f14
: ds: 0018   es: 0018   fs: 002b   gs: 002b   ss: 0018
: Process tar (pid: 159, process nr: 25, stackpage=01c43000)
: Stack: 00000000 01ddfe88 01ddfe88 01c43f70 0001ed90 01e98c7c 000007f0 01e96b20
: 009caaa0 01e95c00 01e966c0 00000016 009caaa0 01e98d48 0015a7f0 01ddfe88
: 00004000 01c43f70 00000000 000001ed 080619de bffffa7c 00000000 ffffffe4
: Call Trace: [ext2_mkdir+192/808] [do_mkdir+290/328] [sys_mkdir+54/80] [system_call+82/128]
: Code: 66 8b bf 1e 01 00 00 66 89 7d 0c f6 86 f0 00 00 00 04 74 0d             

: general protection: 0000
: CPU:    0
: EIP:    0010:[wake_up+53/228]
: EFLAGS: 00010206
: eax: 01db0001   ebx: c3f000ef   ecx: 001b833c   edx: 6f000000
: esi: 00000004   edi: 001b8338   ebp: 018bdf84   esp: 018bdf78  
: Unable to handle kernel paging request at virtual address e06b63aa
: current->tss.cr3 = 01dd5000, |r3 = 01dd5000
: *pde = 00000000
: Oops: 0000
: CPU:    0
: EIP:    0010:[do_mmap+107/788]
: EFLAGS: 00010287
: eax: 00002000   ebx: 00000005   ecx: 206b6361   edx: 01e1d81a
: esi: 00001812   edi: 00002000   ebp: 08000000   esp: 01dd7cc8
: ds: 0018   es: 0018   fs: 0018   gs: 002b   ss: 0018
: Process cat (pid: 49, process nr: 11, stackpage=01dd7000)
: Stack: 00000005 00000000 08000000 00088b2c 206b6361 0012e483 01ffbe40 08000000
: 000018aa 00000005 00001812 00000000 fffffff8 01dd7e6c 01df7000 01dd7e3c
: 00000000 00000000 00000000 00000000 0001ff14 00000000 01ebb4d8 01ebb4d8
: Call Trace: [load_elf_binary+1723/2860] [load_aout_binary+908/1540] 
: [search_binary_handler+63/336] [do_execve+325/440] [do_execve+362/440] [sys_execve+50/80] [s
: ystem_call+82/128]
: Code: f6 41 49 20 74 1f 8b 41 44 c1 e0 0c 01 f8 39 82 e0 01 00 00       

: hda2: bad access: block=345376, count=345386
: end_request: I/O error, dev 03:02, sector 345376
: hda2: bad access: block=345378, count=345384
: end_request: I/O error, dev 03:02, sector 345378
: hda2: bad access: block=345380, count=345382
: end_request: I/O error, dev 03:02, sector 345380
: hda2: bad access: block=345382, count=345380
: end_request: I/O error, dev 03:02, sector 345382
: hda2: bad access: block=345384, count=345378
: end_request: I/O error, dev 03:02, sector 345384 

: Some other messages look like :

: EXT2-fs warning (device 03:03) : ext2_unlink Deleting nonexistent
: file(145028),0
: EXT2-fs warning (device 03:03) : ext2_free_bocks:bit already 
: cleared for block 583727

: Entry 'win.dat' in /usr/local/samba has deleted/unused inode 21859

: Aie, killing interrupt handler
: kfree of non-kmalloced memory: 001b6614, next=00000000, order=0
: kfree of non-kmalloced memory: 001b604, next=00000000, order=0
: kfree of non-kmalloced memory: 001b6b18, next=00000000, order=0
: idle task may not sleep
: idle task may not sleep
: idle task may not sleep
: idle task may not sleep
: idle task may not sleep
: klogd

: When checking disks with fsck, Linux generates reports like: 

: Block bitmap differences: -17128 -17129 -17130 -17131 -17132 -41852 -90807 -9080
: 8 -90809 -90810 -90811 -90812 -90813 +90995 -91163.  FIXED
: Free blocks count wrong for group 2 (7309, counted=7314).  FIXED
: Free blocks count wrong for group 5 (7331, counted=7332).  FIXED
: Free blocks count wrong for group 11 (7267, counted=7274).  FIXED
: Free blocks count wrong (162692, counted=162705).  FIXED

: /dev/hda2: ***** FILE SYSTEM WAS MODIFIED *****
: /dev/hda2: 3656/51376 files (3.3% non-contiguous), 42423/205128 blocks
: e2fsck 1.04, 16-May-96 for EXT2 FS 0.5b, 95/08/09
: e2fsck 1.04, 16-May-96 for EXT2 FS 0.5b, 95/08/09       


: Can somebody explain us what is the problem? and give us a solution
: to fix errors?
: Are there physical damages on our hard disks?
: Or is it only a hard drive configuration problem?

: Thanks in advance

: Manuel Kolp
: kolp@qant.ucl.ac.be
: IAG-QANT
: University of Louvain
: Belgium
Hi Manuel,

these error msgs look very much like the ones I get since last week.
I have not found out yet where the problem is, so please let me know if
you do.
The weird thing for me is that I can mount the partitions manually, but
the kernel crashes if I try to make it mount them as root.
The problem lies not in the kernel, but somewhere "between LILO and the
fs".
I will let you know if I find the bug.

good luck...
Klaus. (Klaus.syttkus@student.uni-tuebingen.de)

: :

dolittle@weniac.math.toronto.edu (Ed Doolittle) wrote:

 >On 29 Oct 1996 14:18:08 GMT, Manuel Kolp <manuel@I_should_put_my_domain_in_etc_NNTP_INEWS_DOMAIN> wrote:
 >
 >>                   Type  Size   Cyl   Head  Precom  Landz Sector Mode
 >> Primary Master:   User  1083   2100   16   65535   2039   63    Normal
 >> Primary Slave:    None   0      0      0    0       0      0
 >> Secondary Master: User  640    1240   16   65535   1239   63    Normal
 >> Secondary Slave   None  0      0      0    0        0     0
 [...]                                                                  ^^^^^^
 >> Linux detects our drives as follows:
 >
 >> hda: WDC AC21000H, 1033MB w/128kB Cache, LBA, CHS=2100/16/63
 >> hdb: FX300_01, ATAPI CDROM drive
 >> hdc: WDC AC2635F, 610MB w/64kB Cache, LBA, CHS=1240/16/63
 >                                        ^^^
 >Just a guess: This looks like a discrepancy here.

No, it isn't... unfortunately.  LBA is just a way of addressing the
sectors on the drive.  For reasons completely opaque to me, it has
become synonymous with translation in BIOSes and the DOS world.  But
not so with Linux.

Besides, had there been a translation problem, fdisk would've
screamed bloody murder about the partition table.  No, it's not this
simple.  The partition table and the geometry used are fully
consistent.

I suggest this problem should be submitted to a kernel wizard who
can make sense of the crashes.  It may be related to a hardware
problem.  I can't be any more specific personally.

 - Peter

--
pieterh@sci.kun.nl                 http://thef-nym.sci.kun.nl/~pieterh/