'IRQ timeout' message and 'DMA disabled' with 'ls -l'

'IRQ timeout' message and 'DMA disabled' with 'ls -l'

Post by John Harri » Mon, 05 May 1997 04:00:00



I have a DOS partition on /dev/hda1 and a LINUX partition on
/dev/hda2.  There is also /dev/hdb with a Windows NT partition.

After booting, the first time I give the 'ls -l' command to the DOS
partition, /dev/hda1, the following message appears on the screen
after a 15 second stall:

hda: irq timeout: status=0xd0 { Busy }
hda: disabled DMA
hdb: disabled DMA
ide0: reset: success,

then the output to the 'ls -l' command finally appears.

Subsequent 'ls -l' commands to /dev/hda1 do not yield the message
during the same LINUX session, nor do I receive the message if I give
the 'ls -l' command to any other partition; i.e., /dev/hda2 (the LINUX
partition itself) or /dev/hdb1.  Sometimes, but rarely, the message
appears at boot time, just after the partition check, or immediately
prior to the appearance of the login prompt.

So far I have not noticed any other problems which are unambiguously
related to the phenomenon.

The problem started after I had to change from a Microsoft busmouse
(IRQ was 5) to a PS/2 mouse (IRQ is 12) because Windows NT would not
work with the former.  The new PS/2 protocol works with X-Windows.
Due to the timing, it looked like a hardware problem to me, but my
dealer says that everything is set up properly.

I have installed an new kernel (2.0.30) and have reinstalled some
library files from setup, but the problem still occurs.  It also
occurs with two other Linux systems I have installed on a zip drive
and on /dev/hdc.  

I am wondering if it is some type of memory conflict.  

I would appreciate any advice someone can give on the matter. Should it
help, I am including the current boot-time output from 'dmesg.'

John

Console: 16 point font, 400 scans
Console: colour VGA+ 80x25, 1 virtual console (max 63)
pcibios_init : BIOS32 Service Directory structure at 0x000f8440
pcibios_init : BIOS32 Service Directory entry at 0xf7e90
pcibios_init : PCI BIOS revision 2.10 entry at 0xf7ec0
Probing PCI hardware.
Calibrating delay loop.. ok - 66.36 BogoMIPS
Memory: 63220k/65536k available (716k kernel code, 384k reserved, 1216k data)
Swansea University Computer Society NET3.035 for Linux 2.0
NET3: Unix domain sockets 0.13 for Linux NET3.035.
Swansea University Computer Society TCP/IP for NET3.034
IP Protocols: ICMP, UDP, TCP
Checking 386/387 coupling... Ok, fpu using exception 16 error reporting.
Checking 'hlt' instruction... Ok.

Serial driver version 4.13 with no serial options enabled
tty00 at 0x03f8 (irq = 4) is a 16550A
tty01 at 0x02f8 (irq = 3) is a 16550A
PS/2 auxiliary pointing device detected -- driver installed.
Sound initialization started
<Sound Blaster 16 (4.11)> at 0x220 irq 10 dma 1,5
Sound initialization complete
loop: registered device at major 7
ide: i82371 PIIX (Triton) on PCI bus 0 function 57
    ide0: BM-DMA at 0xe800-0xe807
    ide1: BM-DMA at 0xe808-0xe80f
hda: FUJITSU M1638TAU, 2452MB w/128kB Cache, LBA, CHS=622/128/63, DMA
hdb: QUANTUM FIREBALL_TM2110A, 2014MB w/76kB Cache, LBA, CHS=1023/64/63, DMA
hdc: QUANTUM ELS170A, 162MB w/32kB Cache, CHS=1011/15/22
hdd: TOSHIBA CD-ROM XM-5302TA, ATAPI CDROM drive
ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
ide1 at 0x170-0x177,0x376 on irq 15
Floppy drive(s): fd0 is 1.44M, fd1 is 1.2M
Started kswapd v 1.4.2.2
FDC 0 is a post-1991 82077
scsi : 0 hosts.
scsi : detected total.
PPP: version 2.2.0 (dynamic channel allocation)
TCP compression code copyright 1989 Regents of the University of California
PPP Dynamic channel allocation code copyright 1995 Caldera, Inc.
PPP line discipline registered.
SLIP: version 0.8.4-NET3.019-NEWTTY (dynamic channels, max=256).
CSLIP: code copyright 1989 Regents of the University of California.
SLIP linefill/keepalive option.
Partition check:
 hda: hda1 hda2
 hdb: hdb1
 hdc: hdc1 hdc2
VFS: Mounted root (ext2 filesystem) readonly.
Adding Swap: 16484k swap-space (priority 2120)
Max size:328458   Log zone size:2048
First datazone:60   Root inode number 122880
ISO9660 Extensions: RRIP_1991A
registered device ppp0

 
 
 

'IRQ timeout' message and 'DMA disabled' with 'ls -l'

Post by Robert Hell » Mon, 05 May 1997 04:00:00



  In a message on Sun, 4 May 1997 09:22:53 -0700, wrote :

JH> I have a DOS partition on /dev/hda1 and a LINUX partition on
JH> /dev/hda2.  There is also /dev/hdb with a Windows NT partition.
JH>
JH> After booting, the first time I give the 'ls -l' command to the DOS
JH> partition, /dev/hda1, the following message appears on the screen
JH> after a 15 second stall:
JH>
JH> hda: irq timeout: status=0xd0 { Busy }
JH> hda: disabled DMA
JH> hdb: disabled DMA
JH> ide0: reset: success,
JH>
JH> then the output to the 'ls -l' command finally appears.
JH>
JH> Subsequent 'ls -l' commands to /dev/hda1 do not yield the message
JH> during the same LINUX session, nor do I receive the message if I give
JH> the 'ls -l' command to any other partition; i.e., /dev/hda2 (the LINUX
JH> partition itself) or /dev/hdb1.  Sometimes, but rarely, the message
JH> appears at boot time, just after the partition check, or immediately
JH> prior to the appearance of the login prompt.
JH>
JH> So far I have not noticed any other problems which are unambiguously
JH> related to the phenomenon.
JH>
JH> The problem started after I had to change from a Microsoft busmouse
JH> (IRQ was 5) to a PS/2 mouse (IRQ is 12) because Windows NT would not
JH> work with the former.  The new PS/2 protocol works with X-Windows.
JH> Due to the timing, it looked like a hardware problem to me, but my
JH> dealer says that everything is set up properly.

What IRQ is the ide controller using?  If 12, then that is the problem.  Change
the IRQ on the PS/2 mouse to a *unused* IRQ.  It looks like Linux is trying to
use DMA on your ide controller, but there is an IRQ problem, so the DMA is being
disabled and Linux is reverting to programmed I/O for your IDE disks, which
probably tosses performance down the drain.

JH>
JH> I have installed an new kernel (2.0.30) and have reinstalled some
JH> library files from setup, but the problem still occurs.  It also
JH> occurs with two other Linux systems I have installed on a zip drive
JH> and on /dev/hdc.  
JH>
JH> I am wondering if it is some type of memory conflict.  
JH>
JH> I would appreciate any advice someone can give on the matter. Should it
JH> help, I am including the current boot-time output from 'dmesg.'
JH>
JH>
JH> John
JH>
JH> Console: 16 point font, 400 scans
JH> Console: colour VGA+ 80x25, 1 virtual console (max 63)
JH> pcibios_init : BIOS32 Service Directory structure at 0x000f8440
JH> pcibios_init : BIOS32 Service Directory entry at 0xf7e90
JH> pcibios_init : PCI BIOS revision 2.10 entry at 0xf7ec0
JH> Probing PCI hardware.
JH> Calibrating delay loop.. ok - 66.36 BogoMIPS
JH> Memory: 63220k/65536k available (716k kernel code, 384k reserved, 1216k data)
JH> Swansea University Computer Society NET3.035 for Linux 2.0
JH> NET3: Unix domain sockets 0.13 for Linux NET3.035.
JH> Swansea University Computer Society TCP/IP for NET3.034
JH> IP Protocols: ICMP, UDP, TCP
JH> Checking 386/387 coupling... Ok, fpu using exception 16 error reporting.
JH> Checking 'hlt' instruction... Ok.

JH> Serial driver version 4.13 with no serial options enabled
JH> tty00 at 0x03f8 (irq = 4) is a 16550A
JH> tty01 at 0x02f8 (irq = 3) is a 16550A
JH> PS/2 auxiliary pointing device detected -- driver installed.
JH> Sound initialization started
JH> <Sound Blaster 16 (4.11)> at 0x220 irq 10 dma 1,5
JH> Sound initialization complete
JH> loop: registered device at major 7
JH> ide: i82371 PIIX (Triton) on PCI bus 0 function 57
JH>     ide0: BM-DMA at 0xe800-0xe807
JH>     ide1: BM-DMA at 0xe808-0xe80f
JH> hda: FUJITSU M1638TAU, 2452MB w/128kB Cache, LBA, CHS=622/128/63, DMA
JH> hdb: QUANTUM FIREBALL_TM2110A, 2014MB w/76kB Cache, LBA, CHS=1023/64/63, DMA
JH> hdc: QUANTUM ELS170A, 162MB w/32kB Cache, CHS=1011/15/22
JH> hdd: TOSHIBA CD-ROM XM-5302TA, ATAPI CDROM drive
JH> ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
JH> ide1 at 0x170-0x177,0x376 on irq 15
JH> Floppy drive(s): fd0 is 1.44M, fd1 is 1.2M
JH> Started kswapd v 1.4.2.2
JH> FDC 0 is a post-1991 82077
JH> scsi : 0 hosts.
JH> scsi : detected total.
JH> PPP: version 2.2.0 (dynamic channel allocation)
JH> TCP compression code copyright 1989 Regents of the University of California
JH> PPP Dynamic channel allocation code copyright 1995 Caldera, Inc.
JH> PPP line discipline registered.
JH> SLIP: version 0.8.4-NET3.019-NEWTTY (dynamic channels, max=256).
JH> CSLIP: code copyright 1989 Regents of the University of California.
JH> SLIP linefill/keepalive option.
JH> Partition check:
JH>  hda: hda1 hda2
JH>  hdb: hdb1
JH>  hdc: hdc1 hdc2
JH> VFS: Mounted root (ext2 filesystem) readonly.
JH> Adding Swap: 16484k swap-space (priority 2120)
JH> Max size:328458   Log zone size:2048
JH> First datazone:60   Root inode number 122880
JH> ISO9660 Extensions: RRIP_1991A
JH> registered device ppp0
JH>
JH>
JH>
JH>
JH>
JH>
JH>
JH>
JH>
JH>
JH>
JH>
JH>                                                                          

--
                                     \/

http://vis-www.cs.umass.edu/~heller  ||FidoNet:    1:321/153
http://netmar.com/mall/shops/heller  /\

 
 
 

'IRQ timeout' message and 'DMA disabled' with 'ls -l'

Post by Michael Dimmi » Wed, 14 May 1997 04:00:00





>  In a message on Sun, 4 May 1997 09:22:53 -0700, wrote :
>JH> I have a DOS partition on /dev/hda1 and a LINUX partition on
>JH> /dev/hda2.  There is also /dev/hdb with a Windows NT partition.
>JH> After booting, the first time I give the 'ls -l' command to the DOS
>JH> partition, /dev/hda1, the following message appears on the screen
>JH> after a 15 second stall:
>JH> hda: irq timeout: status=0xd0 { Busy }
>JH> hda: disabled DMA
>JH> hdb: disabled DMA
>JH> ide0: reset: success,
>JH> then the output to the 'ls -l' command finally appears.
>JH> Subsequent 'ls -l' commands to /dev/hda1 do not yield the message
>JH> during the same LINUX session, nor do I receive the message if I give
>JH> the 'ls -l' command to any other partition; i.e., /dev/hda2 (the LINUX
>JH> partition itself) or /dev/hdb1.
>JH> The problem started after I had to change from a Microsoft busmouse
>JH> (IRQ was 5) to a PS/2 mouse (IRQ is 12) because Windows NT would not
>JH> work with the former.  The new PS/2 protocol works with X-Windows.
>JH> Due to the timing, it looked like a hardware problem to me, but my
>JH> dealer says that everything is set up properly.
>What IRQ is the ide controller using?  If 12, then that is the
>problem.  Change the IRQ on the PS/2 mouse to a *unused* IRQ.  It
>looks like Linux is trying to use DMA on your ide controller, but
>there is an IRQ problem, so the DMA is being disabled and Linux is
>reverting to programmed I/O for your IDE disks, which probably tosses
>performance down the drain.

Err, if you look at the dmesg listing, you'll see that the ide
controllers are on IRQ14 & 15, as per spec.  (See below)

However, it does look like an IRQ problem.  Do you have a network
card, and if so, where's that one's IRQ?  Check the IRQs in your
system (have a look at /proc/interrupt) to see if there's a problem.

May 3 22:09:42 PDT 1997

Have you tried recompiling with gcc 2.7.2.1 or 2.7.2.2?  ISTR there's
a bug in 2.7.2, though whether that has anything to do with it I don't
know.

Did you build the kernel by patching, or from a new source tree?

Quote:>JH> PS/2 auxiliary pointing device detected -- driver installed.
>JH> Sound initialization started
>JH> <Sound Blaster 16 (4.11)> at 0x220 irq 10 dma 1,5
>JH> Sound initialization complete

Funny, I've got an AWE32 and don't get this bit.  Mind you, that could
be as I've compiled the sound stuff as a module.

Quote:>JH> ide: i82371 PIIX (Triton) on PCI bus 0 function 57
>JH>     ide0: BM-DMA at 0xe800-0xe807
>JH>     ide1: BM-DMA at 0xe808-0xe80f

Again, a slightly funny address, but I'll assume it's right: mine are
0x3000-0x3007 and 0x3008-0x300F

Quote:>JH> hda: FUJITSU M1638TAU, 2452MB w/128kB Cache, LBA, CHS=622/128/63, DMA
>JH> hdb: QUANTUM FIREBALL_TM2110A, 2014MB w/76kB Cache, LBA, CHS=1023/64/63, DMA
>JH> hdc: QUANTUM ELS170A, 162MB w/32kB Cache, CHS=1011/15/22
>JH> hdd: TOSHIBA CD-ROM XM-5302TA, ATAPI CDROM drive

That's a lot of disks!

Quote:>JH> ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
>JH> ide1 at 0x170-0x177,0x376 on irq 15

Yup, looks right.

Quote:>JH> Floppy drive(s): fd0 is 1.44M, fd1 is 1.2M
>JH> Started kswapd v 1.4.2.2
>JH> FDC 0 is a post-1991 82077
>JH> scsi : 0 hosts.
>JH> scsi : detected total.

If you don't have SCSI, switch it off!  There's no point having unused
code in the kernel.

<snip remaining stuff>

If you recompile the kernel with a) latest gcc/libc, b) from a new
tree and c) with only necessary options switched on (do you really
have a CMD640 chipset?  Do you really have a PIIX chipset?) this may
solve your problem!

--
Michael Dimmick           | ".sig, .sig, wherefore art thou .sig?"

Follow the instructions!  | Don't spam me!  I bite back...

 
 
 

1. 64k files in dir - 'ls' slow but 'strace ls' fast?

        OK, riddle me this...

        I have a directory with 64,000 files in it.  They're all zero-
length regular files in an ext2 filesystem on a SCSI hard drive, nothing
unusual.  But check this:

chindi:/tmp/manyfiles$ time ls
(snip list of file)
real    7m37.655s
user    0m2.640s
sys     6m20.450s

        Ouch!  That's /slow/.  Let's see why it's slow:

chindi:/tmp/manyfiles$ strace -c ls
(snip)
% time     seconds  usecs/call     calls    errors syscall
------ ----------- ----------- --------- --------- ----------------
 85.15    0.321233          50      6400           write
 13.65    0.051490         107       482           getdents
  0.44    0.001653           8       217           brk
(snip)
------ ----------- ----------- --------- --------- ----------------
100.00    0.377256                  7380         2 total

        0.4 SECONDS?!  Something is obviously very different here, but
what?  Amusingly enough, the strace manpage says "A traced process runs
slowly"...

JDW

--
If mail to me bounces, try removing the "+STRING" part.

2. DGA

3. 'ls -e' vs 'ls -l'

4. which Linux distro has most applications in it?

5. strange 'DMA disabled' message in 2.4.20-pre5-ac6

6. sed -newbie

7. wu-ftp server problem, unable to use 'ls' or 'dir'

8. diff/patch

9. Why does 'ls' give '/' as the output?

10. why would 'cp' or 'ls -al' command hangs on Solaris 2.7 ?

11. Why does 'ls' give '/' as the output?

12. ?? ftp 'ls'-command with 'pg' ??

13. color ls and 'more' or 'less'