Skip to contentSkip to conversation

mount hfs on SCSI cdrom = segfault

DISCUSSION Nehal 4 participants 12 messages Mar 30, 2003 – Apr 3, 2003

lore
Share & tools
lore

Nehal

author
#0 lore raw

i have a hybrid cd (both HFS, ISO9660) , i have two CD drives, one IDE CD-Rom (actima 32x), and one SCSI CD-burner (yamaha 6416) on an advansys cfg-510 ISA scsi card

when i try to mount on IDE using hfs with:

mount -v -r -t hfs /dev/hdc /cdrom

it works fine, yet when i try on scsi with:

mount -v -r -t hfs /dev/scd0 /cdrom

i get a "Segmentation fault" error, no more output given, it also locks the drive, and sometimes i can use the 'eject' command to eject it, sometimes i cant and i gotta reboot

once though i did get a memory dump, unfortunately i didnt get all the info, i do remember it saying this (and im sure of 2555):

kernel bug in buffer.c:2555

im using kernel 2.4.20 with no patches

note: when i try to mount the cd using regular iso9660 fs, it works perfectly on both cd drives,
also i have tried 2 hybrid cd's, both times i have trouble mounting hfs on the scsi drive only

Nehal

author
#1 lore raw
Quoted 16 lines

i have a hybrid cd (both HFS, ISO9660) , i have two CD drives, one IDE CD-Rom (actima 32x), and one SCSI CD-burner (yamaha 6416) on an advansys cfg-510 ISA scsi card

when i try to mount on IDE using hfs with:

mount -v -r -t hfs /dev/hdc /cdrom

it works fine, yet when i try on scsi with:

mount -v -r -t hfs /dev/scd0 /cdrom

i get a "Segmentation fault" error, no more output given, it also locks the drive, and sometimes i can use the 'eject' command to eject it, sometimes i cant and i gotta reboot

Quoted 6 lines

note: when i try to mount the cd using regular iso9660 fs, it works perfectly on both cd drives,
also i have tried 2 hybrid cd's, both times i have trouble mounting hfs on the scsi drive only

Nehal

ok i updated firmware of writer from 1.0c to 1.0d with no help, but i found when i do 'dmesg' after mounting i get this error:
========
kernel BUG at buffer.c:2518!

invalid operand: 0000
CPU:    0
EIP:    0010:[<c013c329>]    Not tainted
EFLAGS: 00013206
eax: 000007ff   ebx: 00000b00   ecx: 00000800   edx: c11ee640
esi: 00000b00   edi: 00000200   ebp: 00000b00   esp: c3425db4
ds: 0018   es: 0018   ss: 0018
Process mount (pid: 514, stackpage=c3425000)
Stack: c6d0d760 c3425e48 c0257a59 c7f1c574 00000000 00000b00 00000200 00000000
      c0139f66 00000b00 00000000 00000200 00000000 00000001 c7568400 
00000000
      c013a1e0 00000b00 00000000 00000200 00000000 c019280a 00000b00 
00000000
Call Trace:    [<c0257a59>] [<c0139f66>] [<c013a1e0>] [<c019280a>] [<c019188a>]
 [<c01925ff>] [<c0285c30>] [<c013cdca>] [<c013e908>] [<c013d64b>]

[<c013cd3c>]
[<c013d9a1>] [<c014fcf3>] [<c0150020>] [<c014fe69>] [<c0150441>]
[<c01090ff>]

Code: 0f 0b d6 09 9a 2b 33 c0 8d 87 00 fe ff ff 3d 00 0e 00 00 76

root@Nehal:~#
========
then when i try it again it doesnt give this message, it locks up my drive

can someone please help debug this problem,
thx, Nehal

#2 lore raw

First post said Linux 2.4.20... that's good info. That info has been deleted from subsequent postings.

Would it make sense for the BUG() message to include a kernel version number??? I'm wondering since people do omit that data.

~Randy

On Wed, 02 Apr 2003 10:15:21 -0800 Nehal <[email protected]> wrote:

| > i have a hybrid cd (both HFS, ISO9660) , i have two CD drives,
| > one IDE CD-Rom (actima 32x), and one SCSI CD-burner (yamaha 6416)
| > on an advansys cfg-510 ISA scsi card
| >
| > when i try to mount on IDE using hfs with:
| >
| >    mount -v -r -t hfs /dev/hdc /cdrom
| >
| > it works fine, yet when i try on scsi with:
| >
| >    mount -v -r -t hfs /dev/scd0 /cdrom
| >
| > i get a "Segmentation fault" error, no more output given,
| > it also locks the drive, and sometimes i can use the
| > 'eject' command to eject it, sometimes i cant and i gotta reboot
| >
| > note: when i try to mount the cd using regular iso9660 fs, it
| > works perfectly on both cd drives,
| > also i have tried 2 hybrid cd's, both times i have trouble mounting
| > hfs on the scsi drive only
| >
| > Nehal 
| 
| ok i updated firmware of writer from 1.0c to 1.0d with no help,
| but i found when i do 'dmesg' after mounting i get this error:
| ========
| kernel BUG at buffer.c:2518!
| invalid operand: 0000
| CPU:    0
| EIP:    0010:[<c013c329>]    Not tainted
| EFLAGS: 00013206
| eax: 000007ff   ebx: 00000b00   ecx: 00000800   edx: c11ee640
| esi: 00000b00   edi: 00000200   ebp: 00000b00   esp: c3425db4
| ds: 0018   es: 0018   ss: 0018
| Process mount (pid: 514, stackpage=c3425000)
| Stack: c6d0d760 c3425e48 c0257a59 c7f1c574 00000000 00000b00 00000200 
| 00000000
|        c0139f66 00000b00 00000000 00000200 00000000 00000001 c7568400 
| 00000000
|        c013a1e0 00000b00 00000000 00000200 00000000 c019280a 00000b00 
| 00000000
| Call Trace:    [<c0257a59>] [<c0139f66>] [<c013a1e0>] [<c019280a>] 
| [<c019188a>]
|   [<c01925ff>] [<c0285c30>] [<c013cdca>] [<c013e908>] [<c013d64b>] 
| [<c013cd3c>]
|   [<c013d9a1>] [<c014fcf3>] [<c0150020>] [<c014fe69>] [<c0150441>] 
| [<c01090ff>]
| 
| Code: 0f 0b d6 09 9a 2b 33 c0 8d 87 00 fe ff ff 3d 00 0e 00 00 76
|  
| root@Nehal:~#
| ========
| then when i try it again it doesnt give this message, it locks up my drive
| 
| can someone please help debug this problem,
| thx, Nehal
#3 lore raw

Hi,

is this the same problem as this one?
http://marc.theaimsgroup.com/?l=linux-kernel&m=102890250915062&w=2

Nobody ever answered on that, though.

Quoted 72 lines · Randy.Dunlap

First post said Linux 2.4.20... that's good info. That info has been deleted from subsequent postings.

Would it make sense for the BUG() message to include a kernel version number??? I'm wondering since people do omit that data.

~Randy

On Wed, 02 Apr 2003 10:15:21 -0800 Nehal <[email protected]> wrote:

| > i have a hybrid cd (both HFS, ISO9660) , i have two CD drives,
| > one IDE CD-Rom (actima 32x), and one SCSI CD-burner (yamaha 6416)
| > on an advansys cfg-510 ISA scsi card
| >
| > when i try to mount on IDE using hfs with:
| >
| >    mount -v -r -t hfs /dev/hdc /cdrom
| >
| > it works fine, yet when i try on scsi with:
| >
| >    mount -v -r -t hfs /dev/scd0 /cdrom
| >
| > i get a "Segmentation fault" error, no more output given,
| > it also locks the drive, and sometimes i can use the
| > 'eject' command to eject it, sometimes i cant and i gotta reboot
| >
| > note: when i try to mount the cd using regular iso9660 fs, it
| > works perfectly on both cd drives,
| > also i have tried 2 hybrid cd's, both times i have trouble mounting
| > hfs on the scsi drive only
| >
| > Nehal
|
| ok i updated firmware of writer from 1.0c to 1.0d with no help,
| but i found when i do 'dmesg' after mounting i get this error:
| ========
| kernel BUG at buffer.c:2518!
| invalid operand: 0000
| CPU:    0
| EIP:    0010:[<c013c329>]    Not tainted
| EFLAGS: 00013206
| eax: 000007ff   ebx: 00000b00   ecx: 00000800   edx: c11ee640
| esi: 00000b00   edi: 00000200   ebp: 00000b00   esp: c3425db4
| ds: 0018   es: 0018   ss: 0018
| Process mount (pid: 514, stackpage=c3425000)
| Stack: c6d0d760 c3425e48 c0257a59 c7f1c574 00000000 00000b00 00000200
| 00000000
|        c0139f66 00000b00 00000000 00000200 00000000 00000001 c7568400
| 00000000
|        c013a1e0 00000b00 00000000 00000200 00000000 c019280a 00000b00
| 00000000
| Call Trace:    [<c0257a59>] [<c0139f66>] [<c013a1e0>] [<c019280a>]
| [<c019188a>]
|   [<c01925ff>] [<c0285c30>] [<c013cdca>] [<c013e908>] [<c013d64b>]
| [<c013cd3c>]
|   [<c013d9a1>] [<c014fcf3>] [<c0150020>] [<c014fe69>] [<c0150441>]
| [<c01090ff>]
|
| Code: 0f 0b d6 09 9a 2b 33 c0 8d 87 00 fe ff ff 3d 00 0e 00 00 76
|
| root@Nehal:~#
| ========
| then when i try it again it doesnt give this message, it locks up my
| drive
|
| can someone please help debug this problem,
| thx, Nehal

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Signature · 4 lines
Oliver Feiler  <kiza@(kcore.de|lionking.org|gmx[pro].net)>
http://kiza.kcore.de/    <--    homepage
PGP-key      -->    /pgpkey.shtml
http://kiza.kcore.de/journal/
author
#11 lore raw
In reply to Randy.Dunlap

Hello Randy,

thx for responding, yes i took out 2.4.20 this was unintented, sorry (sometimes i clip, ill be more careful now).... but yes 2.4.20 unmodified

the message i pasted is unmodified, its a middle-click paste from 'dmesg' output so what u c is what i c :)

Nehal

Quoted 60 lines

First post said Linux 2.4.20... that's good info. That info has been deleted from subsequent postings.

Would it make sense for the BUG() message to include a kernel version number??? I'm wondering since people do omit that data.

~Randy

On Wed, 02 Apr 2003 10:15:21 -0800 Nehal <[email protected]> wrote:

| > i have a hybrid cd (both HFS, ISO9660) , i have two CD drives,
| > one IDE CD-Rom (actima 32x), and one SCSI CD-burner (yamaha 6416)
| > on an advansys cfg-510 ISA scsi card
| >
| > when i try to mount on IDE using hfs with:
| >
| >    mount -v -r -t hfs /dev/hdc /cdrom
| >
| > it works fine, yet when i try on scsi with:
| >
| >    mount -v -r -t hfs /dev/scd0 /cdrom
| >
| > i get a "Segmentation fault" error, no more output given,
| > it also locks the drive, and sometimes i can use the
| > 'eject' command to eject it, sometimes i cant and i gotta reboot
| >
| > note: when i try to mount the cd using regular iso9660 fs, it
| > works perfectly on both cd drives,
| > also i have tried 2 hybrid cd's, both times i have trouble mounting
| > hfs on the scsi drive only
| >
| > Nehal | | ok i updated firmware of writer from 1.0c to 1.0d with no help,
| but i found when i do 'dmesg' after mounting i get this error:
| ========
| kernel BUG at buffer.c:2518!
| invalid operand: 0000
| CPU:    0
| EIP:    0010:[<c013c329>]    Not tainted
| EFLAGS: 00013206
| eax: 000007ff   ebx: 00000b00   ecx: 00000800   edx: c11ee640
| esi: 00000b00   edi: 00000200   ebp: 00000b00   esp: c3425db4
| ds: 0018   es: 0018   ss: 0018
| Process mount (pid: 514, stackpage=c3425000)
| Stack: c6d0d760 c3425e48 c0257a59 c7f1c574 00000000 00000b00 00000200 | 00000000
|        c0139f66 00000b00 00000000 00000200 00000000 00000001 c7568400 | 00000000
|        c013a1e0 00000b00 00000000 00000200 00000000 c019280a 00000b00 | 00000000
| Call Trace:    [<c0257a59>] [<c0139f66>] [<c013a1e0>] [<c019280a>] | [<c019188a>]
|   [<c01925ff>] [<c0285c30>] [<c013cdca>] [<c013e908>] [<c013d64b>] | [<c013cd3c>]
|   [<c013d9a1>] [<c014fcf3>] [<c0150020>] [<c014fe69>] [<c0150441>] | [<c01090ff>]
| | Code: 0f 0b d6 09 9a 2b 33 c0 8d 87 00 fe ff ff 3d 00 0e 00 00 76
|  | root@Nehal:~#
| ========
| then when i try it again it doesnt give this message, it locks up my drive
| | can someone please help debug this problem,
| thx, Nehal
author
#4 lore raw
In reply to Oliver Feiler

yes, that's the one! its in grow_buffers ... there are 3 BUG() 's in this function, but my guess he is getting the same BUG()

hmm, he didnt specify his cd-rom drive... but could it be part of the problem?

Nehal

Quoted 88 lines

Hi,

is this the same problem as this one?
http://marc.theaimsgroup.com/?l=linux-kernel&m=102890250915062&w=2

Nobody ever answered on that, though.

Quoted 74 lines · Randy.Dunlap

First post said Linux 2.4.20... that's good info. That info has been deleted from subsequent postings.

Would it make sense for the BUG() message to include a kernel version number??? I'm wondering since people do omit that data.

~Randy

On Wed, 02 Apr 2003 10:15:21 -0800 Nehal <[email protected]> wrote:

| > i have a hybrid cd (both HFS, ISO9660) , i have two CD drives,
| > one IDE CD-Rom (actima 32x), and one SCSI CD-burner (yamaha 6416)
| > on an advansys cfg-510 ISA scsi card
| >
| > when i try to mount on IDE using hfs with:
| >
| >    mount -v -r -t hfs /dev/hdc /cdrom
| >
| > it works fine, yet when i try on scsi with:
| >
| >    mount -v -r -t hfs /dev/scd0 /cdrom
| >
| > i get a "Segmentation fault" error, no more output given,
| > it also locks the drive, and sometimes i can use the
| > 'eject' command to eject it, sometimes i cant and i gotta reboot
| >
| > note: when i try to mount the cd using regular iso9660 fs, it
| > works perfectly on both cd drives,
| > also i have tried 2 hybrid cd's, both times i have trouble mounting
| > hfs on the scsi drive only
| >
| > Nehal
|
| ok i updated firmware of writer from 1.0c to 1.0d with no help,
| but i found when i do 'dmesg' after mounting i get this error:
| ========
| kernel BUG at buffer.c:2518!
| invalid operand: 0000
| CPU:    0
| EIP:    0010:[<c013c329>]    Not tainted
| EFLAGS: 00013206
| eax: 000007ff   ebx: 00000b00   ecx: 00000800   edx: c11ee640
| esi: 00000b00   edi: 00000200   ebp: 00000b00   esp: c3425db4
| ds: 0018   es: 0018   ss: 0018
| Process mount (pid: 514, stackpage=c3425000)
| Stack: c6d0d760 c3425e48 c0257a59 c7f1c574 00000000 00000b00 00000200
| 00000000
|        c0139f66 00000b00 00000000 00000200 00000000 00000001 c7568400
| 00000000
|        c013a1e0 00000b00 00000000 00000200 00000000 c019280a 00000b00
| 00000000
| Call Trace:    [<c0257a59>] [<c0139f66>] [<c013a1e0>] [<c019280a>]
| [<c019188a>]
|   [<c01925ff>] [<c0285c30>] [<c013cdca>] [<c013e908>] [<c013d64b>]
| [<c013cd3c>]
|   [<c013d9a1>] [<c014fcf3>] [<c0150020>] [<c014fe69>] [<c0150441>]
| [<c01090ff>]
|
| Code: 0f 0b d6 09 9a 2b 33 c0 8d 87 00 fe ff ff 3d 00 0e 00 00 76
|
| root@Nehal:~#
| ========
| then when i try it again it doesnt give this message, it locks up my
| drive
|
| can someone please help debug this problem,
| thx, Nehal

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

#5 lore raw
Quoted 5 lines · Nehal

yes, that's the one! its in grow_buffers ... there are 3 BUG() 's in this function, but my guess he is getting the same BUG()

hmm, he didnt specify his cd-rom drive... but could it be part of the problem?

It was me who posted this previous message. :)

I don't think it is drive dependent. But it may have something to do with SCSI since it works for you with IDE and oopses with SCSI. I just reproduced the oops with a
hdc: HL-DT-ST RW/DVD GCC-4120B, ATAPI CD/DVD-ROM drive
IDE drive running via ide-scsi. Same oops again, see below. Yes I know with NVdriver, but the oops is the same with and without. Now the disc isn't mounted, the drive door is locked and 'eject /dev/cdrom' gives

eject: CDROMEJECT ioctl failed for `/dev/scsi/host1/bus0/target0/lun0/cd':
Device or resource busy

kernel BUG at buffer.c:2497!

invalid operand: 0000
CPU:    0
EIP:    0010:[<c0136e5e>]    Tainted: P

Using defaults from ksymoops -t elf32-i386 -a i386
EFLAGS: 00010206
eax: 000007ff ebx: 0000000b ecx: 00000800 edx: c15dd700
esi: 00000002 edi: 00000b02 ebp: 00000000 esp: ccf01dbc
ds: 0018 es: 0018 ss: 0018
Process mount (pid: 4975, stackpage=ccf01000)

Stack: 00000b02 00000200 00000000 00000001 00004480 c0134fb7 00000b02 00000000 
       00000200 00000b02 d01cac00 00000000 c01351d8 00000b02 00000000 00000200 
       00000000 e1cb45c3 00000b02 00000000 00000200 00000b02 00000001 00000000 
Call Trace:    [<c0134fb7>] [<c01351d8>] [<e1cb45c3>] [<e1cb3839>]

[<e1cb4443>]
[<c01ef6ed>] [<c01380e0>] [<e1cb7fe4>] [<e1cb7fe4>] [<c01382bb>]
[<e1cb7fe4>]
[<c0148059>] [<c0148332>] [<c014817d>] [<c01486b4>] [<c010870b>]

Code: 0f 0b c1 09 00 e2 25 c0 8b 44 24 20 05 00 fe ff ff 3d 00 0e
Quoted 7 lines
Quoted 7 lines

EIP; c0136e5e <grow_buffers+3e/110> <=====

eax; 000007ff Before first symbol
ecx; 00000800 Before first symbol
edx; c15dd700 <_end+12b4ae4/2051c3e4>
edi; 00000b02 Before first symbol
esp; ccf01dbc <_end+cbd91a0/2051c3e4>

Trace; c0134fb7 <getblk+27/40>
Trace; c01351d8 <bread+18/70>
Trace; e1cb45c3 <[hfs]hfs_buffer_get+23/80>
Trace; e1cb3839 <[hfs]hfs_part_find+19/170>
Trace; e1cb4443 <[hfs]hfs_read_super+73/190>
Trace; c01ef6ed <media_changed+3d/70>
Trace; c01380e0 <get_sb_bdev+210/280>
Trace; e1cb7fe4 <[hfs]hfs_fs+0/1c>
Trace; e1cb7fe4 <[hfs]hfs_fs+0/1c>
Trace; c01382bb <do_kern_mount+5b/110>
Trace; e1cb7fe4 <[hfs]hfs_fs+0/1c>
Trace; c0148059 <do_add_mount+69/140>
Trace; c0148332 <do_mount+162/180>
Trace; c014817d <copy_mount_options+4d/a0>
Trace; c01486b4 <sys_mount+84/d0>
Trace; c010870b <system_call+33/38>

Code; c0136e5e <grow_buffers+3e/110>
00000000 <_EIP>:
Code; c0136e5e <grow_buffers+3e/110> <=====
0: 0f 0b ud2a <=====
Code; c0136e60 <grow_buffers+40/110>
2: c1 09 00 rorl $0x0,(%ecx)
Code; c0136e63 <grow_buffers+43/110>
5: e2 25 loop 2c <_EIP+0x2c> c0136e8a
<grow_buffers+6a/110>
Code; c0136e65 <grow_buffers+45/110>
7: c0 8b 44 24 20 05 00 rorb $0x0,0x5202444(%ebx)
Code; c0136e6c <grow_buffers+4c/110>
e: fe (bad)
Code; c0136e6d <grow_buffers+4d/110>
f: ff (bad)
Code; c0136e6e <grow_buffers+4e/110>
10: ff (bad)
Code; c0136e6f <grow_buffers+4f/110>
11: 3d 00 0e 00 00 cmp $0xe00,%eax

2 warnings issued. Results may not be reliable.

Quoted 83 lines

Nehal

Quoted 80 lines

Hi,

is this the same problem as this one?
http://marc.theaimsgroup.com/?l=linux-kernel&m=102890250915062&w=2

Nobody ever answered on that, though.

Quoted 72 lines · Randy.Dunlap

First post said Linux 2.4.20... that's good info. That info has been deleted from subsequent postings.

Would it make sense for the BUG() message to include a kernel version number??? I'm wondering since people do omit that data.

~Randy

On Wed, 02 Apr 2003 10:15:21 -0800 Nehal <[email protected]> wrote:

| > i have a hybrid cd (both HFS, ISO9660) , i have two CD drives,
| > one IDE CD-Rom (actima 32x), and one SCSI CD-burner (yamaha 6416)
| > on an advansys cfg-510 ISA scsi card
| >
| > when i try to mount on IDE using hfs with:
| >
| >    mount -v -r -t hfs /dev/hdc /cdrom
| >
| > it works fine, yet when i try on scsi with:
| >
| >    mount -v -r -t hfs /dev/scd0 /cdrom
| >
| > i get a "Segmentation fault" error, no more output given,
| > it also locks the drive, and sometimes i can use the
| > 'eject' command to eject it, sometimes i cant and i gotta reboot
| >
| > note: when i try to mount the cd using regular iso9660 fs, it
| > works perfectly on both cd drives,
| > also i have tried 2 hybrid cd's, both times i have trouble mounting
| > hfs on the scsi drive only
| >
| > Nehal
|
| ok i updated firmware of writer from 1.0c to 1.0d with no help,
| but i found when i do 'dmesg' after mounting i get this error:
| ========
| kernel BUG at buffer.c:2518!
| invalid operand: 0000
| CPU:    0
| EIP:    0010:[<c013c329>]    Not tainted
| EFLAGS: 00013206
| eax: 000007ff   ebx: 00000b00   ecx: 00000800   edx: c11ee640
| esi: 00000b00   edi: 00000200   ebp: 00000b00   esp: c3425db4
| ds: 0018   es: 0018   ss: 0018
| Process mount (pid: 514, stackpage=c3425000)
| Stack: c6d0d760 c3425e48 c0257a59 c7f1c574 00000000 00000b00 00000200
| 00000000
|        c0139f66 00000b00 00000000 00000200 00000000 00000001 c7568400
| 00000000
|        c013a1e0 00000b00 00000000 00000200 00000000 c019280a 00000b00
| 00000000
| Call Trace:    [<c0257a59>] [<c0139f66>] [<c013a1e0>] [<c019280a>]
| [<c019188a>]
|   [<c01925ff>] [<c0285c30>] [<c013cdca>] [<c013e908>] [<c013d64b>]
| [<c013cd3c>]
|   [<c013d9a1>] [<c014fcf3>] [<c0150020>] [<c014fe69>] [<c0150441>]
| [<c01090ff>]
|
| Code: 0f 0b d6 09 9a 2b 33 c0 8d 87 00 fe ff ff 3d 00 0e 00 00 76
|
| root@Nehal:~#
| ========
| then when i try it again it doesnt give this message, it locks up my
| drive
|
| can someone please help debug this problem,
| thx, Nehal

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Signature · 4 lines
Oliver Feiler  <kiza@(kcore.de|lionking.org|gmx[pro].net)>
http://kiza.kcore.de/    <--    homepage
PGP-key      -->    /pgpkey.shtml
http://kiza.kcore.de/journal/
author
#6 lore raw

ok you are right! i set hdc=scsi, reboot, and mounted my ide cdrom drive (device /dev/scd1) as hfs, and boom it crashed with same message :) this is good, at least i know its not my scsi drive/controller at fault

so definitely a bug in scsi somehow or somewhere, so i guess anyone with a hfs cd can reproduce, so somebody please fix :)

btw Oliver, once u type eject, if it fails, u can just press the eject button on your drive and it should open.... although the drive is still unusable until u reboot

thx, Nehal

Quoted 190 lines · Nehal
Quoted 9 lines

yes, that's the one! its in grow_buffers ... there are 3 BUG() 's in this function, but my guess he is getting the same BUG()

hmm, he didnt specify his cd-rom drive... but could it be part of the problem?

It was me who posted this previous message. :)

I don't think it is drive dependent. But it may have something to do with SCSI since it works for you with IDE and oopses with SCSI. I just reproduced the oops with a
hdc: HL-DT-ST RW/DVD GCC-4120B, ATAPI CD/DVD-ROM drive
IDE drive running via ide-scsi. Same oops again, see below. Yes I know with NVdriver, but the oops is the same with and without. Now the disc isn't mounted, the drive door is locked and 'eject /dev/cdrom' gives

eject: CDROMEJECT ioctl failed for `/dev/scsi/host1/bus0/target0/lun0/cd': Device or resource busy

kernel BUG at buffer.c:2497!

invalid operand: 0000
CPU:    0
EIP:    0010:[<c0136e5e>]    Tainted: P Using defaults from ksymoops -t elf32-i386 -a i386
EFLAGS: 00010206
eax: 000007ff   ebx: 0000000b   ecx: 00000800   edx: c15dd700
esi: 00000002   edi: 00000b02   ebp: 00000000   esp: ccf01dbc
ds: 0018   es: 0018   ss: 0018
Process mount (pid: 4975, stackpage=ccf01000)
Stack: 00000b02 00000200 00000000 00000001 00004480 c0134fb7 00000b02 00000000 
      00000200 00000b02 d01cac00 00000000 c01351d8 00000b02 00000000 00000200 
      00000000 e1cb45c3 00000b02 00000000 00000200 00000b02 00000001 00000000 
Call Trace:    [<c0134fb7>] [<c01351d8>] [<e1cb45c3>] [<e1cb3839>] [<e1cb4443>]
 [<c01ef6ed>] [<c01380e0>] [<e1cb7fe4>] [<e1cb7fe4>] [<c01382bb>] 
[<e1cb7fe4>]
 [<c0148059>] [<c0148332>] [<c014817d>] [<c01486b4>] [<c010870b>]
Code: 0f 0b c1 09 00 e2 25 c0 8b 44 24 20 05 00 fe ff ff 3d 00 0e
Quoted 6 lines
Quoted 3 lines

EIP; c0136e5e <grow_buffers+3e/110> <=====

Quoted 7 lines
Quoted 7 lines

eax; 000007ff Before first symbol
ecx; 00000800 Before first symbol
edx; c15dd700 <_end+12b4ae4/2051c3e4>
edi; 00000b02 Before first symbol
esp; ccf01dbc <_end+cbd91a0/2051c3e4>

Trace; c0134fb7 <getblk+27/40>
Trace; c01351d8 <bread+18/70>
Trace; e1cb45c3 <[hfs]hfs_buffer_get+23/80>
Trace; e1cb3839 <[hfs]hfs_part_find+19/170>
Trace; e1cb4443 <[hfs]hfs_read_super+73/190>
Trace; c01ef6ed <media_changed+3d/70>
Trace; c01380e0 <get_sb_bdev+210/280>
Trace; e1cb7fe4 <[hfs]hfs_fs+0/1c>
Trace; e1cb7fe4 <[hfs]hfs_fs+0/1c>
Trace; c01382bb <do_kern_mount+5b/110>
Trace; e1cb7fe4 <[hfs]hfs_fs+0/1c>
Trace; c0148059 <do_add_mount+69/140>
Trace; c0148332 <do_mount+162/180>
Trace; c014817d <copy_mount_options+4d/a0>
Trace; c01486b4 <sys_mount+84/d0>
Trace; c010870b <system_call+33/38>

Code; c0136e5e <grow_buffers+3e/110>
00000000 <_EIP>:

Code;  c0136e5e <grow_buffers+3e/110>   <=====
  0:   0f 0b                     ud2a      <=====
Code;  c0136e60 <grow_buffers+40/110>
  2:   c1 09 00                  rorl   $0x0,(%ecx)
Code;  c0136e63 <grow_buffers+43/110>
  5:   e2 25                     loop   2c <_EIP+0x2c> c0136e8a 
<grow_buffers+6a/110>
Code;  c0136e65 <grow_buffers+45/110>
  7:   c0 8b 44 24 20 05 00      rorb   $0x0,0x5202444(%ebx)
Code;  c0136e6c <grow_buffers+4c/110>
  e:   fe                        (bad)  
Code;  c0136e6d <grow_buffers+4d/110>
  f:   ff                        (bad)  
Code;  c0136e6e <grow_buffers+4e/110>
 10:   ff                        (bad)  
Code;  c0136e6f <grow_buffers+4f/110>
 11:   3d 00 0e 00 00            cmp    $0xe00,%eax

2 warnings issued. Results may not be reliable.

Quoted 88 lines

Nehal

Quoted 84 lines

Hi,

is this the same problem as this one?
http://marc.theaimsgroup.com/?l=linux-kernel&m=102890250915062&w=2

Nobody ever answered on that, though.

Quoted 74 lines · Randy.Dunlap

First post said Linux 2.4.20... that's good info. That info has been deleted from subsequent postings.

Would it make sense for the BUG() message to include a kernel version number??? I'm wondering since people do omit that data.

~Randy

On Wed, 02 Apr 2003 10:15:21 -0800 Nehal <[email protected]> wrote:

| > i have a hybrid cd (both HFS, ISO9660) , i have two CD drives,
| > one IDE CD-Rom (actima 32x), and one SCSI CD-burner (yamaha 6416)
| > on an advansys cfg-510 ISA scsi card
| >
| > when i try to mount on IDE using hfs with:
| >
| >    mount -v -r -t hfs /dev/hdc /cdrom
| >
| > it works fine, yet when i try on scsi with:
| >
| >    mount -v -r -t hfs /dev/scd0 /cdrom
| >
| > i get a "Segmentation fault" error, no more output given,
| > it also locks the drive, and sometimes i can use the
| > 'eject' command to eject it, sometimes i cant and i gotta reboot
| >
| > note: when i try to mount the cd using regular iso9660 fs, it
| > works perfectly on both cd drives,
| > also i have tried 2 hybrid cd's, both times i have trouble mounting
| > hfs on the scsi drive only
| >
| > Nehal
|
| ok i updated firmware of writer from 1.0c to 1.0d with no help,
| but i found when i do 'dmesg' after mounting i get this error:
| ========
| kernel BUG at buffer.c:2518!
| invalid operand: 0000
| CPU:    0
| EIP:    0010:[<c013c329>]    Not tainted
| EFLAGS: 00013206
| eax: 000007ff   ebx: 00000b00   ecx: 00000800   edx: c11ee640
| esi: 00000b00   edi: 00000200   ebp: 00000b00   esp: c3425db4
| ds: 0018   es: 0018   ss: 0018
| Process mount (pid: 514, stackpage=c3425000)
| Stack: c6d0d760 c3425e48 c0257a59 c7f1c574 00000000 00000b00 00000200
| 00000000
|        c0139f66 00000b00 00000000 00000200 00000000 00000001 c7568400
| 00000000
|        c013a1e0 00000b00 00000000 00000200 00000000 c019280a 00000b00
| 00000000
| Call Trace:    [<c0257a59>] [<c0139f66>] [<c013a1e0>] [<c019280a>]
| [<c019188a>]
|   [<c01925ff>] [<c0285c30>] [<c013cdca>] [<c013e908>] [<c013d64b>]
| [<c013cd3c>]
|   [<c013d9a1>] [<c014fcf3>] [<c0150020>] [<c014fe69>] [<c0150441>]
| [<c01090ff>]
|
| Code: 0f 0b d6 09 9a 2b 33 c0 8d 87 00 fe ff ff 3d 00 0e 00 00 76
|
| root@Nehal:~#
| ========
| then when i try it again it doesnt give this message, it locks up my
| drive
|
| can someone please help debug this problem,
| thx, Nehal

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

#7 lore raw
Quoted 8 lines · On 02-Apr-2003 Nehal wrote:

ok you are right! i set hdc=scsi, reboot, and mounted my ide cdrom drive (device /dev/scd1) as hfs, and boom it crashed with same message :) this is good, at least i know its not my scsi drive/controller at fault

so definitely a bug in scsi somehow or somewhere, so i guess anyone with a hfs cd can reproduce, so somebody please fix :)

Once upon a time it worked just fine. Then someone removed support for !=512 bytes sectors...
To workaround, use loopback.

Bye.

author
#8 lore raw
Quoted 14 lines · Giuliano Pochini
Quoted 10 lines · On 02-Apr-2003 Nehal wrote:

ok you are right! i set hdc=scsi, reboot, and mounted my ide cdrom drive (device /dev/scd1) as hfs, and boom it crashed with same message :) this is good, at least i know its not my scsi drive/controller at fault

so definitely a bug in scsi somehow or somewhere, so i guess anyone with a hfs cd can reproduce, so somebody please fix :)

Quoted 4 lines

Once upon a time it worked just fine. Then someone removed support for !=512 bytes sectors...
To workaround, use loopback.

(yes Oliver has told me about this workaround)

1. do u know why it was removed?
2. is there a reason why can't support for it be put back?

thx, Nehal

#9 lore raw
Quoted 9 lines
Quoted 5 lines

Once upon a time it worked just fine. Then someone removed support for !=512 bytes sectors...
To workaround, use loopback.

(yes Oliver has told me about this workaround)

1. do u know why it was removed?
2. is there a reason why can't support for it be put back?

I don't know why, and I don't know if it was removed from the hfs code or if hfs relied on some features of a lower layer that has been modified.

Bye.

author
#10 lore raw
Quoted 8 lines · Giuliano Pochini
Quoted 8 lines
Quoted 6 lines

Once upon a time it worked just fine. Then someone removed support for !=512 bytes sectors...
To workaround, use loopback.

(yes Oliver has told me about this workaround)

Quoted 10 lines
Quoted 4 lines

1. do u know why it was removed?
2. is there a reason why can't support for it be put back?

I don't know why, and I don't know if it was removed from the hfs code or if hfs relied on some features of a lower layer that has been modified.

i did some searching around and only found that this problem started during 2.3.x because of a scsi rewrite

anyways, i did try the loop method, it did not work, i got a segmentation fault when i did the following:

losetup /dev/loop0 /dev/scd0
mount -r -t hfs /dev/loop0 /cdrom

oh well, i'll just use my IDE drive for now..

thx anyways, Nehal