Skip to contentSkip to conversation

2.4.20-pre2-ac3 stops responding

DISCUSSION Jean-Luc Coulon 8 participants 17 messages Aug 16, 2002 – Sep 2, 2002

lore
Share & tools
lore
#0 lore raw

Hi,

I've the followings symptoms

My computer is a K6-2 500 with 384Mb SDRAM

1st time to boot:
-----------------
with 2.4.20-pre2 : 30s
with 2.4.30-pre2-ac3 : 55s

2nd while running:
------------------
If I have high disk activity, the system stops responding for a while, it does not accepts any key action nor mouse movement. It starts running normally after few seconds.

----
Regards
Jean-Luc

#1 lore raw
Quoted 5 lines · Jean-Luc Coulon

2nd while running:
------------------
If I have high disk activity, the system stops responding for a while, it does not accepts any key action nor mouse movement. It starts running normally after few seconds.

I've got a patch that might help improve this situation:

http://surriel.com/patches/2.4/2.4.20-p2ac3-rmap14

Could you please try this patch ?

kind regards,

Rik

Signature · 3 lines
Bravely reimplemented by the knights who say "NIH".

http://www.surriel.com/		http://distro.conectiva.com/
#14 lore raw
In reply to Jean-Luc Coulon
Quoted 3 lines · Jean-Luc Coulon

If I have high disk activity, the system stops responding for a while, it does not accepts any key action nor mouse movement. It starts running normally after few seconds.

Try using Robert Love's "preempt-kernel" and "sched-hints" patches in addition to -ac3, it helps a lot.

See "ftp://ftp.de.kernel.org/pub/linux/kernel/people/rml/" for the patches.

Signature · 1 line
# Heinz Diehl, 68259 Mannheim, Germany
#16 lore raw
Quoted 9 lines · Tomas Szepe
Quoted 3 lines

with 2.4.20-pre2 : 30s
with 2.4.30-pre2-ac3 : 55s

Do you get DMA on your IDE disks at all?

hdparm -iv /dev/hdX

T.

Seems to have a problem with DMA

[root@debian-f5ibh] ~ # hdparm -iv /dev/hda2

/dev/hda2:

multcount    =  1 (on)
IO_support   =  1 (32-bit)
unmaskirq    =  0 (off)
using_dma    =  0 (off)
keepsettings =  0 (off)
readonly     =  0 (off)
readahead    =  8 (on)

geometry = 3649/255/63, sectors = 29302560, start = 9767520
HDIO_GET_IDENTITY failed: Invalid argument

[root@debian-f5ibh] ~ # hdparm -d1 /dev/hda2

/dev/hda2:
setting using_dma to 1 (on)
HDIO_SET_DMA failed: Invalid argument
using_dma = 0 (off)

#2 lore raw
In reply to Rik van Riel
Quoted 15 lines · Rik van Riel
Quoted 5 lines · Jean-Luc Coulon

2nd while running:
------------------
If I have high disk activity, the system stops responding for a while, it does not accepts any key action nor mouse movement. It starts running normally after few seconds.

I've got a patch that might help improve this situation:

http://surriel.com/patches/2.4/2.4.20-p2ac3-rmap14

Could you please try this patch ?

Quoted 7 lines

kind regards,

Rik
--
Bravely reimplemented by the knights who say "NIH".

http://www.surriel.com/ http://distro.conectiva.com/

I've applied the patch, it does not improve the things. The problem seems to be related with a DMA :
[root@debian-f5ibh] ~ # hdparm -iv /dev/hda2

/dev/hda2:

multcount    =  1 (on)
IO_support   =  1 (32-bit)
unmaskirq    =  0 (off)
using_dma    =  0 (off)
keepsettings =  0 (off)
readonly     =  0 (off)
readahead    =  8 (on)

geometry = 3649/255/63, sectors = 29302560, start = 9767520
HDIO_GET_IDENTITY failed: Invalid argument

[root@debian-f5ibh] ~ # hdparm -d1 /dev/hda2

/dev/hda2:
setting using_dma to 1 (on)
HDIO_SET_DMA failed: Invalid argument
using_dma = 0 (off)

----
Regards
Jean-Luc

#3 lore raw
Quoted 13 lines · Jean-Luc Coulon
Quoted 11 lines · Rik van Riel
Quoted 5 lines · Jean-Luc Coulon

2nd while running:
------------------
If I have high disk activity, the system stops responding for a while, it does not accepts any key action nor mouse movement. It starts running normally after few seconds.

I've got a patch that might help improve this situation:

http://surriel.com/patches/2.4/2.4.20-p2ac3-rmap14

Quoted 7 lines

I've applied the patch, it does not improve the things. The problem seems to be related with a DMA :

/dev/hda2:
setting using_dma to 1 (on)
HDIO_SET_DMA failed: Invalid argument
using_dma = 0 (off)

Heh indeed. There's no way to make your system fast if the CPU is tied up waiting for the disk all the time ;)

regards,

Rik

Signature · 3 lines
Bravely reimplemented by the knights who say "NIH".

http://www.surriel.com/		http://distro.conectiva.com/
#12 lore raw
In reply to Jean-Luc Coulon
Quoted 20 lines · On Fri, 16 Aug 2002 18:29:48 +0200 Jean-Luc Coulon…
Quoted 8 lines · Rik van Riel
Quoted 5 lines · Jean-Luc Coulon

2nd while running:
------------------
If I have high disk activity, the system stops responding for a while, it does not accepts any key action nor mouse movement. It starts running normally after few seconds.

/dev/hda2:

multcount    =  1 (on)
IO_support   =  1 (32-bit)
unmaskirq    =  0 (off)
using_dma    =  0 (off)
keepsettings =  0 (off)
readonly     =  0 (off)
readahead    =  8 (on)

geometry = 3649/255/63, sectors = 29302560, start = 9767520
HDIO_GET_IDENTITY failed: Invalid argument

Quoted 7 lines

[root@debian-f5ibh] ~ # hdparm -d1 /dev/hda2

/dev/hda2:
setting using_dma to 1 (on)
HDIO_SET_DMA failed: Invalid argument
using_dma = 0 (off)

how about umaskirq (hdparm -u1) ? that should help quite a bit, if it works.

oh, and by the way, does it matter that you're doing hdparm on a partition rather than the wholedisk?

#4 lore raw
In reply to Rik van Riel
Quoted 26 lines · Rik van Riel
Quoted 13 lines · Jean-Luc Coulon
Quoted 11 lines · Rik van Riel
Quoted 5 lines · Jean-Luc Coulon

2nd while running:
------------------
If I have high disk activity, the system stops responding for a while, it does not accepts any key action nor mouse movement. It starts running normally after few seconds.

I've got a patch that might help improve this situation:

http://surriel.com/patches/2.4/2.4.20-p2ac3-rmap14

Quoted 7 lines

I've applied the patch, it does not improve the things. The problem seems to be related with a DMA :

/dev/hda2:
setting using_dma to 1 (on)
HDIO_SET_DMA failed: Invalid argument
using_dma = 0 (off)

Heh indeed. There's no way to make your system fast if the CPU is tied up waiting for the disk all the time ;)

Quoted 7 lines

regards,

Rik
--
Bravely reimplemented by the knights who say "NIH".

http://www.surriel.com/ http://distro.conectiva.com/

Of course, but I've not yet found the way to setup DMA

At boot time, I get the messages :

Aug 16 11:34:19 f5ibh kernel: ALI15X3: simplex device: DMA disabled
Aug 16 11:34:19 f5ibh kernel: ide0: ALI15X3 Bus-Master DMA disabled

(BIOS)

-----
Regards
Jean-Luc

#13 lore raw
In reply to Andrew Rodland
Quoted 26 lines · Andrew Rodland
Quoted 22 lines · On Fri, 16 Aug 2002 18:29:48 +0200 Jean-Luc Coulon…
Quoted 8 lines · Rik van Riel
Quoted 5 lines · Jean-Luc Coulon

2nd while running:
------------------
If I have high disk activity, the system stops responding for a while, it does not accepts any key action nor mouse movement. It starts running normally after few seconds.

/dev/hda2:

multcount    =  1 (on)
IO_support   =  1 (32-bit)
unmaskirq    =  0 (off)
using_dma    =  0 (off)
keepsettings =  0 (off)
readonly     =  0 (off)
readahead    =  8 (on)

geometry = 3649/255/63, sectors = 29302560, start = 9767520
HDIO_GET_IDENTITY failed: Invalid argument

[root@debian-f5ibh] ~ # hdparm -d1 /dev/hda2

Quoted 7 lines
Quoted 5 lines

/dev/hda2:
setting using_dma to 1 (on)
HDIO_SET_DMA failed: Invalid argument
using_dma = 0 (off)

how about umaskirq (hdparm -u1) ? that should help quite a bit, if it works.

I can set it up, but I cannot set up the DMA

Quoted 3 lines

oh, and by the way, does it matter that you're doing hdparm on a partition rather than the wholedisk?

There is no difference

----
regards
jean-Luc

#5 lore raw
In reply to Jean-Luc Coulon
Quoted 5 lines · Jean-Luc Coulon

At boot time, I get the messages :

Aug 16 11:34:19 f5ibh kernel: ALI15X3: simplex device: DMA disabled
Aug 16 11:34:19 f5ibh kernel: ide0: ALI15X3 Bus-Master DMA disabled

(BIOS)

Linux did the simplex device check. Your ALi controller only permits DMA on one of the devices at a time. What is attached to the ALi controller ? Also does 2.4.19 base enable DMA correctly ?

If so then my guess is there is a bug in the changing of the pci setup code in 2.4.20pre2-ac3, which shouldnt be hard to figure out

#6 lore raw

There is a problem with IOPS assignment and I am still working on how make each host carry its own fixup.

Quoted 11 lines · Alan Cox
Quoted 5 lines · Jean-Luc Coulon

At boot time, I get the messages :

Aug 16 11:34:19 f5ibh kernel: ALI15X3: simplex device: DMA disabled
Aug 16 11:34:19 f5ibh kernel: ide0: ALI15X3 Bus-Master DMA disabled

(BIOS)

Linux did the simplex device check. Your ALi controller only permits DMA on one of the devices at a time. What is attached to the ALi controller ? Also does 2.4.19 base enable DMA correctly ?

Quoted 9 lines

If so then my guess is there is a bug in the changing of the pci setup code in 2.4.20pre2-ac3, which shouldnt be hard to figure out

-
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/

Andre Hedrick
LAD Storage Consulting Group

#7 lore raw
In reply to Alan Cox
Quoted 8 lines · Alan Cox
Quoted 5 lines · Jean-Luc Coulon

At boot time, I get the messages :

Aug 16 11:34:19 f5ibh kernel: ALI15X3: simplex device: DMA disabled
Aug 16 11:34:19 f5ibh kernel: ide0: ALI15X3 Bus-Master DMA disabled

(BIOS)

Quoted 6 lines

Linux did the simplex device check. Your ALi controller only permits DMA on one of the devices at a time. What is attached to the ALi controller ? Also does 2.4.19 base enable DMA correctly ?

If so then my guess is there is a bug in the changing of the pci setup code in 2.4.20pre2-ac3, which shouldnt be hard to figure out

sorry for the late answer, I was not @home for one week.

Attached to my ide ports are :
on the primary channel : 1 ide disk, QUANTUM FIREBALLP LM30

                         1 ide disk,  ST3491A

on the secondary channel : 1 cdrom reader, CREATIVECD3621E
1 cdrom burner, GoldStar CD-RW CED-8083B

The second HDD is not DMA

-------
regards
Jean-Luc

#8 lore raw
In reply to Alan Cox
Quoted 8 lines · Alan Cox
Quoted 5 lines · Jean-Luc Coulon

At boot time, I get the messages :

Aug 16 11:34:19 f5ibh kernel: ALI15X3: simplex device: DMA disabled
Aug 16 11:34:19 f5ibh kernel: ide0: ALI15X3 Bus-Master DMA disabled

(BIOS)

Quoted 6 lines

Linux did the simplex device check. Your ALi controller only permits DMA on one of the devices at a time. What is attached to the ALi controller ? Also does 2.4.19 base enable DMA correctly ?

If so then my guess is there is a bug in the changing of the pci setup code in 2.4.20pre2-ac3, which shouldnt be hard to figure out

Sorry, I forgot,

2.4.19 enables DMA correctly, as well as 2.4.20-pre1

---------
Regards
Jean-Luc

#9 lore raw
In reply to Alan Cox
Quoted 4 lines · Alan Cox
Quoted 2 lines · Jean-Luc Coulon

At boot time, I get the messages :

Quoted 8 lines
Quoted 3 lines
Aug 16 11:34:19 f5ibh kernel: ALI15X3: simplex device: DMA disabled
Aug 16 11:34:19 f5ibh kernel: ide0: ALI15X3 Bus-Master DMA disabled

(BIOS)

Linux did the simplex device check. Your ALi controller only permits DMA on one of the devices at a time. What is attached to the ALi controller ? Also does 2.4.19 base enable DMA correctly ?

What is the simplex device check ? You will find thereafter a copy of the syslog while booting 2.4.19, it seems that it accepts both with DMA enabled.

Quoted 1 line

If sSep 2 09:00:51 f5ibh kernel: ALI15X3: chipset revision 193

f5ibh kernel: ALI15X3: not 100%% native mode: will probe irqs later
f5ibh kernel: ide0: BM-DMA at 0xd000-0xd007, BIOS settings: hda:DMA,
hdb:DMA
f5ibh kernel: ide1: BM-DMA at 0xd008-0xd00f, BIOS settings: hdc:pio,
hdd:pio
f5ibh kernel: hda: QUANTUM FIREBALLP LM30, ATA DISK drive fibh kernel: hdb: ST3491A, ATA DISK drive
f5ibh kernel: hdc: GoldStar CD-RW CED-8083B, ATAPI CD/DVD-ROM drive f5ibh kernel: hdd: CREATIVECD3621E, ATAPI CD/DVD-ROM drive f5ibh kernel: ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
f5ibh kernel: ide1 at 0x170-0x177,0x376 on irq 15
f5ibh kernel: blk: queue c023d544, I/O limit 4095Mb (mask 0xffffffff) f5ibh kernel: hda: 58633344 sectors (30020 MB) w/1900KiB Cache, CHS=3649/255/63, UDMA(33)
f5ibh kernel: blk: queue c023d680, I/O limit 4095Mb (mask 0xffffffff) f5ibh kernel: hdb: 836070 sectors (428 MB) w/120KiB Cache, CHS=899/15/62, DMA

---
Regards
Jean-Luc

#10 lore raw
Quoted 27 lines · Jean-Luc Coulon
Quoted 4 lines · Alan Cox
Quoted 2 lines · Jean-Luc Coulon

At boot time, I get the messages :

Quoted 8 lines
Quoted 3 lines
Aug 16 11:34:19 f5ibh kernel: ALI15X3: simplex device: DMA disabled
Aug 16 11:34:19 f5ibh kernel: ide0: ALI15X3 Bus-Master DMA disabled

(BIOS)

Linux did the simplex device check. Your ALi controller only permits DMA on one of the devices at a time. What is attached to the ALi controller ? Also does 2.4.19 base enable DMA correctly ?

What is the simplex device check ? You will find thereafter a copy of the syslog while booting 2.4.19, it seems that it accepts both with DMA enabled.

Quoted 1 line

If sSep 2 09:00:51 f5ibh kernel: ALI15X3: chipset revision 193

f5ibh kernel: ALI15X3: not 100%% native mode: will probe irqs later
f5ibh kernel: ide0: BM-DMA at 0xd000-0xd007, BIOS settings: hda:DMA,
hdb:DMA
f5ibh kernel: ide1: BM-DMA at 0xd008-0xd00f, BIOS settings: hdc:pio,
hdd:pio
f5ibh kernel: hda: QUANTUM FIREBALLP LM30, ATA DISK drive fibh kernel: hdb: ST3491A, ATA DISK drive
f5ibh kernel: hdc: GoldStar CD-RW CED-8083B, ATAPI CD/DVD-ROM drive f5ibh kernel: hdd: CREATIVECD3621E, ATAPI CD/DVD-ROM drive

Quoted 8 lines

f5ibh kernel: ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
f5ibh kernel: ide1 at 0x170-0x177,0x376 on irq 15
f5ibh kernel: blk: queue c023d544, I/O limit 4095Mb (mask 0xffffffff) f5ibh kernel: hda: 58633344 sectors (30020 MB) w/1900KiB Cache, CHS=3649/255/63, UDMA(33)
f5ibh kernel: blk: queue c023d680, I/O limit 4095Mb (mask 0xffffffff) f5ibh kernel: hdb: 836070 sectors (428 MB) w/120KiB Cache, CHS=899/15/62, DMA

Are you sure this is a 2.4.19 boot? It's either 2.4.20-preX (X >= 2) or 2.4.19 with block-highmem

Signature · 1 line
Jens Axboe
#11 lore raw

Jens,

Ooops ! you are right, it is 2.4.20-pre5

------
Regards
Jean-Luc

Quoted 43 lines · Jens Axboe
Quoted 27 lines · Jean-Luc Coulon
Quoted 4 lines · Alan Cox
Quoted 2 lines · Jean-Luc Coulon

At boot time, I get the messages :

Quoted 8 lines
Quoted 3 lines
Aug 16 11:34:19 f5ibh kernel: ALI15X3: simplex device: DMA disabled
Aug 16 11:34:19 f5ibh kernel: ide0: ALI15X3 Bus-Master DMA disabled

(BIOS)

Linux did the simplex device check. Your ALi controller only permits DMA on one of the devices at a time. What is attached to the ALi controller ? Also does 2.4.19 base enable DMA correctly ?

What is the simplex device check ? You will find thereafter a copy of the syslog while booting 2.4.19, it seems that it accepts both with DMA enabled.

Quoted 1 line

If sSep 2 09:00:51 f5ibh kernel: ALI15X3: chipset revision 193

f5ibh kernel: ALI15X3: not 100%% native mode: will probe irqs later
f5ibh kernel: ide0: BM-DMA at 0xd000-0xd007, BIOS settings: hda:DMA,
hdb:DMA
f5ibh kernel: ide1: BM-DMA at 0xd008-0xd00f, BIOS settings: hdc:pio,
hdd:pio
f5ibh kernel: hda: QUANTUM FIREBALLP LM30, ATA DISK drive fibh kernel: hdb: ST3491A, ATA DISK drive
f5ibh kernel: hdc: GoldStar CD-RW CED-8083B, ATAPI CD/DVD-ROM drive f5ibh kernel: hdd: CREATIVECD3621E, ATAPI CD/DVD-ROM drive

Quoted 8 lines

f5ibh kernel: ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
f5ibh kernel: ide1 at 0x170-0x177,0x376 on irq 15
f5ibh kernel: blk: queue c023d544, I/O limit 4095Mb (mask 0xffffffff) f5ibh kernel: hda: 58633344 sectors (30020 MB) w/1900KiB Cache, CHS=3649/255/63, UDMA(33)
f5ibh kernel: blk: queue c023d680, I/O limit 4095Mb (mask 0xffffffff) f5ibh kernel: hdb: 836070 sectors (428 MB) w/120KiB Cache, CHS=899/15/62, DMA

Are you sure this is a 2.4.19 boot? It's either 2.4.20-preX (X >= 2) or 2.4.19 with block-highmem

--
Jens Axboe