Discussion:
ERROR: end_request: I/O error, dev fd0, sector 0
Sedat Dilek
2011-11-25 12:53:49 UTC
Permalink
Hi,

I am watching this issue for a long while, due to my
floppy-blacklist.conf file since next-20110603.

The bad: The boot-process stalls for approx 10-13 secs when booting
into my Debian/sid i386 system.

# dmesg | grep -A1 -B1 "end_request: I/O error, dev fd0, sector 0"
[ 2.422854] sd 2:0:0:0: [sdb] Attached SCSI disk
[ 15.289939] end_request: I/O error, dev fd0, sector 0
[ 15.342914] EXT4-fs (sda5): mounted filesystem with ordered data
mode. Opts: (null)

The funny part is my IBM laptop has no physical floppy drive!

# dmesg | egrep -i 'floppy|fd0'
[ 0.726358] Floppy drive(s): fd0 is 1.44M
[ 15.289939] end_request: I/O error, dev fd0, sector 0

If I blacklist "floppy" kernel-module, the boot-process is as expected
without the above mentionned delay.

[ /etc/modprobe.d/floppy-blacklist.conf ]
# This file blacklists the floppy driver.
#
# XXX: Workaround: My IBM ThinkPad T40p (model 2374SG6) notebook has
no physical FDD
# dmesg from linux-next (next-20110603) says:
# [ 0.755294] Floppy drive(s): fd0 is 1.44M
# [ 15.238507] end_request: I/O error, dev fd0, sector 0
blacklist floppy
- EOF -

I searched for the line and found bool blk_update_request() in
block/blk-core.c file (see [1]).

I have attached kern.log and dmesg.

Hope this helps to narrow down the issue.

Kind Regards,
- Sedat -

[1] http://git.kernel.org/?p=linux/kernel/git/next/linux-next.git;a=blob;f=block/blk-core.c#l2111
Robert Hancock
2011-11-26 16:52:03 UTC
Permalink
Post by Sedat Dilek
Hi,
I am watching this issue for a long while, due to my
floppy-blacklist.conf file since next-20110603.
The bad: The boot-process stalls for approx 10-13 secs when booting
into my Debian/sid i386 system.
# dmesg | grep -A1 -B1 "end_request: I/O error, dev fd0, sector 0"
[ 2.422854] sd 2:0:0:0: [sdb] Attached SCSI disk
[ 15.289939] end_request: I/O error, dev fd0, sector 0
[ 15.342914] EXT4-fs (sda5): mounted filesystem with ordered data
mode. Opts: (null)
The funny part is my IBM laptop has no physical floppy drive!
# dmesg | egrep -i 'floppy|fd0'
[ 0.726358] Floppy drive(s): fd0 is 1.44M
[ 15.289939] end_request: I/O error, dev fd0, sector 0
If I blacklist "floppy" kernel-module, the boot-process is as expected
without the above mentionned delay.
[ /etc/modprobe.d/floppy-blacklist.conf ]
# This file blacklists the floppy driver.
#
# XXX: Workaround: My IBM ThinkPad T40p (model 2374SG6) notebook has
no physical FDD
# [ 0.755294] Floppy drive(s): fd0 is 1.44M
# [ 15.238507] end_request: I/O error, dev fd0, sector 0
blacklist floppy
- EOF -
I searched for the line and found bool blk_update_request() in
block/blk-core.c file (see [1]).
I have attached kern.log and dmesg.
Hope this helps to narrow down the issue.
It looks like your laptop's BIOS indicates it has a floppy controller
that's active:

Nov 25 11:00:32 tbox kernel: [ 0.060821] pnp 00:09: [io 0x03f0-0x03f5]
Nov 25 11:00:32 tbox kernel: [ 0.060825] pnp 00:09: [io 0x03f7]
Nov 25 11:00:32 tbox kernel: [ 0.060828] pnp 00:09: [irq 6]
Nov 25 11:00:32 tbox kernel: [ 0.060831] pnp 00:09: [dma 2]
Nov 25 11:00:32 tbox kernel: [ 0.060892] pnp 00:09: Plug and Play
ACPI device, IDs PNP0700 (active)

That's presumably why the floppy driver ends up being loaded. And
apparently it does have a floppy controller, since the floppy driver
seems able to talk to it. Maybe that model of laptop had an optional
bay-mounted floppy drive or something?

I'm not sure there's a workaround for this at the kernel level, since
the kernel probably can't determine if there's actually a floppy drive
installed without trying to read from it (which is likely being
triggered by partition probing or something). The read error will be
what is triggering the message you saw.

Is there a BIOS option to disable the floppy controller?
Sedat Dilek
2011-11-27 13:52:18 UTC
Permalink
Post by Sedat Dilek
Hi,
I am watching this issue for a long while, due to my
floppy-blacklist.conf file since next-20110603.
The bad: The boot-process stalls for approx 10-13 secs when booting
into my Debian/sid i386 system.
# dmesg | grep -A1 -B1 "end_request: I/O error, dev fd0, sector 0"
[    2.422854] sd 2:0:0:0: [sdb] Attached SCSI disk
[   15.289939] end_request: I/O error, dev fd0, sector 0
[   15.342914] EXT4-fs (sda5): mounted filesystem with ordered data
mode. Opts: (null)
The funny part is my IBM laptop has no physical floppy drive!
# dmesg | egrep -i 'floppy|fd0'
[    0.726358] Floppy drive(s): fd0 is 1.44M
[   15.289939] end_request: I/O error, dev fd0, sector 0
If I blacklist "floppy" kernel-module, the boot-process is as expected
without the above mentionned delay.
[ /etc/modprobe.d/floppy-blacklist.conf ]
# This file blacklists the floppy driver.
#
# XXX: Workaround: My IBM ThinkPad T40p (model 2374SG6) notebook has
no physical FDD
# [    0.755294] Floppy drive(s): fd0 is 1.44M
# [   15.238507] end_request: I/O error, dev fd0, sector 0
blacklist floppy
- EOF -
I searched for the line and found bool blk_update_request() in
block/blk-core.c file (see [1]).
I have attached kern.log and dmesg.
Hope this helps to narrow down the issue.
It looks like your laptop's BIOS indicates it has a floppy controller that's
Nov 25 11:00:32 tbox kernel: [    0.060821] pnp 00:09: [io  0x03f0-0x03f5]
Nov 25 11:00:32 tbox kernel: [    0.060825] pnp 00:09: [io  0x03f7]
Nov 25 11:00:32 tbox kernel: [    0.060828] pnp 00:09: [irq 6]
Nov 25 11:00:32 tbox kernel: [    0.060831] pnp 00:09: [dma 2]
Nov 25 11:00:32 tbox kernel: [    0.060892] pnp 00:09: Plug and Play ACPI
device, IDs PNP0700 (active)
That's presumably why the floppy driver ends up being loaded. And apparently
it does have a floppy controller, since the floppy driver seems able to talk
to it. Maybe that model of laptop had an optional bay-mounted floppy drive
or something?
Could be... next to CD_ROM bay maybe.
I'm not sure there's a workaround for this at the kernel level, since the
kernel probably can't determine if there's actually a floppy drive installed
without trying to read from it (which is likely being triggered by partition
probing or something). The read error will be what is triggering the message
you saw.
Is there a BIOS option to disable the floppy controller?
I have disabled "Floppy-Drive" in BIOS setup now (was set to
"enabled"), but I have still the modprobe backlist for floppy
activated.

Can you explain the stalling of ***10-12 seconds*** with a wrong FDD
BIOS setting?
That's really a long duration, with systemd my boot-process takes
approx 15-18 seconds.

Attached is the new dmesg output (kernel-config was not changed).

- Sedat -
Robert Hancock
2011-12-01 02:25:03 UTC
Permalink
Post by Sedat Dilek
Post by Sedat Dilek
Hi,
I am watching this issue for a long while, due to my
floppy-blacklist.conf file since next-20110603.
The bad: The boot-process stalls for approx 10-13 secs when booting
into my Debian/sid i386 system.
# dmesg | grep -A1 -B1 "end_request: I/O error, dev fd0, sector 0"
[ =A0 =A02.422854] sd 2:0:0:0: [sdb] Attached SCSI disk
[ =A0 15.289939] end_request: I/O error, dev fd0, sector 0
[ =A0 15.342914] EXT4-fs (sda5): mounted filesystem with ordered da=
ta
Post by Sedat Dilek
Post by Sedat Dilek
mode. Opts: (null)
The funny part is my IBM laptop has no physical floppy drive!
# dmesg | egrep -i 'floppy|fd0'
[ =A0 =A00.726358] Floppy drive(s): fd0 is 1.44M
[ =A0 15.289939] end_request: I/O error, dev fd0, sector 0
If I blacklist "floppy" kernel-module, the boot-process is as expec=
ted
Post by Sedat Dilek
Post by Sedat Dilek
without the above mentionned delay.
[ /etc/modprobe.d/floppy-blacklist.conf ]
# This file blacklists the floppy driver.
#
# XXX: Workaround: My IBM ThinkPad T40p (model 2374SG6) notebook ha=
s
Post by Sedat Dilek
Post by Sedat Dilek
no physical FDD
# [ =A0 =A00.755294] Floppy drive(s): fd0 is 1.44M
# [ =A0 15.238507] end_request: I/O error, dev fd0, sector 0
blacklist floppy
- EOF -
I searched for the line and found bool blk_update_request() in
block/blk-core.c file (see [1]).
I have attached kern.log and dmesg.
Hope this helps to narrow down the issue.
It looks like your laptop's BIOS indicates it has a floppy controlle=
r that's
Post by Sedat Dilek
Nov 25 11:00:32 tbox kernel: [ =A0 =A00.060821] pnp 00:09: [io =A00x=
03f0-0x03f5]
Post by Sedat Dilek
Nov 25 11:00:32 tbox kernel: [ =A0 =A00.060825] pnp 00:09: [io =A00x=
03f7]
Post by Sedat Dilek
Nov 25 11:00:32 tbox kernel: [ =A0 =A00.060828] pnp 00:09: [irq 6]
Nov 25 11:00:32 tbox kernel: [ =A0 =A00.060831] pnp 00:09: [dma 2]
Nov 25 11:00:32 tbox kernel: [ =A0 =A00.060892] pnp 00:09: Plug and =
Play ACPI
Post by Sedat Dilek
device, IDs PNP0700 (active)
That's presumably why the floppy driver ends up being loaded. And ap=
parently
Post by Sedat Dilek
it does have a floppy controller, since the floppy driver seems able=
to talk
Post by Sedat Dilek
to it. Maybe that model of laptop had an optional bay-mounted floppy=
drive
Post by Sedat Dilek
or something?
Could be... next to CD_ROM bay maybe.
I'm not sure there's a workaround for this at the kernel level, sinc=
e the
Post by Sedat Dilek
kernel probably can't determine if there's actually a floppy drive i=
nstalled
Post by Sedat Dilek
without trying to read from it (which is likely being triggered by p=
artition
Post by Sedat Dilek
probing or something). The read error will be what is triggering the=
message
Post by Sedat Dilek
you saw.
Is there a BIOS option to disable the floppy controller?
I have disabled "Floppy-Drive" in BIOS setup now (was set to
"enabled"), but I have still the modprobe backlist for floppy
activated.
Can you explain the stalling of ***10-12 seconds*** with a wrong FDD
BIOS setting?
That's really a long duration, with systemd my boot-process takes
approx 15-18 seconds.
Well, I can imagine that the floppy request takes that long to time
out or fail. I'm not sure why this would be delaying the boot process,
however.
Post by Sedat Dilek
Attached is the new dmesg output (kernel-config was not changed).
It looks like the BIOS is no longer reporting the floppy controller as
present, so hopefully you shouldn't need the workaround.

Loading...