Skip to contentSkip to conversation

IDE CD problems in 2.6.13rc6

DISCUSSION Dave Jones 4 participants 6 messages Aug 13, 2005 – Aug 14, 2005

lore
Share & tools
lore

Dave Jones

author
#0 lore raw

I've noticed this week whilst trying to encode a bunch of audio CDs to oggs that my boxes running the latest kernels are having serious issues, whereas 2.6.12 seems to cope just fine.

The symptoms vary. On some of my machines just inserting an audio CD makes the box instantly lock up.
If I boot with the same CD in the drive, sound-juicer can read it just fine. When I get to the next CD, I have to reboot again, or it locks up.

On another box, it gets stuck in a loop where it just prints out..

hdc: irq timeout: status=0xd0 { Busy } (This line sometimes has status=0xc0)
ide: failed opcode was: unknown

The net result is that I've not got a single box that will read audio CDs without doing something bad, and I've tried it on several quite diverse systems.

I'll try and narrow down over the next few days when this started happening, but IDE / CD folks may have some better ideas about which changes were suspicious.

Dave

#1 lore raw
Quoted 17 lines · Dave Jones

I've noticed this week whilst trying to encode a bunch of audio CDs to oggs that my boxes running the latest kernels are having serious issues, whereas 2.6.12 seems to cope just fine.

The symptoms vary. On some of my machines just inserting an audio CD makes the box instantly lock up.
If I boot with the same CD in the drive, sound-juicer can read it just fine. When I get to the next CD, I have to reboot again, or it locks up.

On another box, it gets stuck in a loop where it just prints out..

hdc: irq timeout: status=0xd0 { Busy } (This line sometimes has
status=0xc0) ide: failed opcode was: unknown

Quoted 8 lines

The net result is that I've not got a single box that will read audio CDs without doing something bad, and I've tried it on several quite diverse systems.

I'll try and narrow down over the next few days when this started happening, but IDE / CD folks may have some better ideas about which changes were suspicious.

Ok I just started noticing unusual things on my IDE DVD-RW as well, presumably related, on 2.6.13-rc6. Putting in a cd and trying to read it will cause huge delays and then error out with:

ide-cd: cmd 0x28 timed out
hdc: DMA timeout retry
hdc: timeout waiting for DMA
hdc: status timeout: status=0xd0 { Busy }
ide: failed opcode was: unknown
hdc: drive not ready for command
hdc: ATAPI reset complete

then it will read fine. If I start another read it goes through the same cycle. Forcing dma *off* with hdparm fixes the problem and all subsequent accesses don't have this pause, however this was never an issue on previous kernels with dma working fine.

relevant lspci:
00:1d.0 USB Controller: Intel Corp. 82801DB/DBL/DBM (ICH4/ICH4-L/ICH4-M) USB UHCI Controller #1 (rev 02)
00:1d.1 USB Controller: Intel Corp. 82801DB/DBL/DBM (ICH4/ICH4-L/ICH4-M) USB UHCI Controller #2 (rev 02)
00:1d.2 USB Controller: Intel Corp. 82801DB/DBL/DBM (ICH4/ICH4-L/ICH4-M) USB UHCI Controller #3 (rev 02)
00:1d.7 USB Controller: Intel Corp. 82801DB/DBM (ICH4/ICH4-M) USB 2.0 EHCI Controller (rev 02)

hdparm info of hdc:
/dev/hdc:

ATAPI CD-ROM, with removable media
Model Number: LITE-ON DVDRW SOHW-1653S
Serial Number:
Firmware Revision: CS02
Standards:
Used: ATAPI for CD-ROMs, SFF-8020i, r2.5
Supported: CD-ROM ATAPI-2
Configuration:
DRQ response: 50us.
Packet size: 12 bytes
Capabilities:
LBA, IORDY(cannot be disabled)
DMA: mdma0 mdma1 mdma2 udma0 udma1 *udma2 udma3 udma4
Cycle time: min=120ns recommended=120ns
PIO: pio0 pio1 pio2 pio3 pio4
Cycle time: no flow control=227ns IORDY flow control=120ns

relevant config=y info:
CONFIG_IDE=y
CONFIG_BLK_DEV_IDE=y
CONFIG_BLK_DEV_IDEDISK=y
CONFIG_IDEDISK_MULTI_MODE=y
CONFIG_BLK_DEV_IDECD=y
CONFIG_BLK_DEV_IDEPCI=y
CONFIG_IDEPCI_SHARE_IRQ=y
CONFIG_BLK_DEV_IDEDMA_PCI=y
CONFIG_IDEDMA_PCI_AUTO=y
CONFIG_BLK_DEV_PIIX=y
CONFIG_BLK_DEV_IDEDMA=y
CONFIG_IDEDMA_IVB=y
CONFIG_IDEDMA_AUTO=y

Cheers,
Con

#2 lore raw
Quoted 39 lines · Con Kolivas
Quoted 17 lines · Dave Jones

I've noticed this week whilst trying to encode a bunch of audio CDs to oggs that my boxes running the latest kernels are having serious issues, whereas 2.6.12 seems to cope just fine.

The symptoms vary. On some of my machines just inserting an audio CD makes the box instantly lock up.
If I boot with the same CD in the drive, sound-juicer can read it just fine. When I get to the next CD, I have to reboot again, or it locks up.

On another box, it gets stuck in a loop where it just prints out..

hdc: irq timeout: status=0xd0 { Busy } (This line sometimes has
status=0xc0) ide: failed opcode was: unknown

Quoted 8 lines

The net result is that I've not got a single box that will read audio CDs without doing something bad, and I've tried it on several quite diverse systems.

I'll try and narrow down over the next few days when this started happening, but IDE / CD folks may have some better ideas about which changes were suspicious.

Ok I just started noticing unusual things on my IDE DVD-RW as well, presumably related, on 2.6.13-rc6. Putting in a cd and trying to read it will cause huge delays and then error out with:

ide-cd: cmd 0x28 timed out
hdc: DMA timeout retry
hdc: timeout waiting for DMA
hdc: status timeout: status=0xd0 { Busy }
ide: failed opcode was: unknown
hdc: drive not ready for command
hdc: ATAPI reset complete

Quoted 8 lines

then it will read fine. If I start another read it goes through the same cycle. Forcing dma *off* with hdparm fixes the problem and all subsequent accesses don't have this pause, however this was never an issue on previous kernels with dma working fine.

hdparm info of hdc:

DMA: mdma0 mdma1 mdma2 udma0 udma1 *udma2 udma3 udma4

I should also point out the above *udma2 is interesting because:
hdparm /dev/hdc

/dev/hdc:
HDIO_GET_MULTCOUNT failed: Invalid argument

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

readahead = 256 (on)

This *thinks* that dma is off yet querying hdc says it's in udma2 mode and the actual performance very much suggests to me that dma is enabled as it doesn't have those awful lags that dma off dvd access usually has.

Even though it apparently works fine I do get these instead though:
hdc: drive_cmd: status=0x51 { DriveReady SeekComplete Error }
hdc: drive_cmd: error=0x04 { AbortedCommand }
ide: failed opcode was: 0xec
hdc: drive_cmd: status=0x51 { DriveReady SeekComplete Error }
hdc: drive_cmd: error=0x04 { AbortedCommand }
ide: failed opcode was: 0xec

blah...

Con

#3 lore raw
Quoted 4 lines · Con Kolivas
Quoted 4 lines

presumably related, on 2.6.13-rc6. Putting in a cd and trying to read it will cause huge delays and then error out with:

ide-cd: cmd 0x28 timed out

(Thats READ_10)

Quoted 6 lines
Quoted 6 lines

hdc: DMA timeout retry
hdc: timeout waiting for DMA
hdc: status timeout: status=0xd0 { Busy }
ide: failed opcode was: unknown
hdc: drive not ready for command
hdc: ATAPI reset complete

So we asked the drive to do a read and it stayed busy - possibly from some previous command. In a lot of cases you see the command *after* the one that broke it. The reset then recovered the drive.

Quoted 4 lines

Even though it apparently works fine I do get these instead though:
hdc: drive_cmd: status=0x51 { DriveReady SeekComplete Error }
hdc: drive_cmd: error=0x04 { AbortedCommand }
ide: failed opcode was: 0xec

Thats somewhere between bizarre and fascinating.

#4 lore raw
In reply to Dave Jones

Hi Dave,

Quoted 17 lines · Dave Jones

I've noticed this week whilst trying to encode a bunch of audio CDs to oggs that my boxes running the latest kernels are having serious issues, whereas 2.6.12 seems to cope just fine.

The symptoms vary. On some of my machines just inserting an audio CD makes the box instantly lock up.
If I boot with the same CD in the drive, sound-juicer can read it just fine. When I get to the next CD, I have to reboot again, or it locks up.

On another box, it gets stuck in a loop where it just prints out..

hdc: irq timeout: status=0xd0 { Busy } (This line sometimes has status=0xc0)
ide: failed opcode was: unknown

Quoted 8 lines

The net result is that I've not got a single box that will read audio CDs without doing something bad, and I've tried it on several quite diverse systems.

I'll try and narrow down over the next few days when this started happening, but IDE / CD folks may have some better ideas about which changes were suspicious.

I checked 2.6.13-rc6 patch and I see no IDE / CD changes which could be responsible for this regression. You can try reverting ide-cd changes and see if this helps. IRQ routing changes?

Bartlomiej

#5 lore raw
Quoted 2 lines · Bartlomiej Zolnierkiewicz
Quoted 2 lines

The symptoms vary. On some of my machines just inserting an audio CD makes the box instantly lock up.

You've got all the gnome cruft running. Start by turning that off. Then try inserting/removing audio discs, playing them with "cdplay". I'd expect that to be ok

Next try ripping an audio cd without all the gnome crud loaded. If that works (it may or may not) then its practical to try and pin down the bug and see if its error handling (the locking patterns all changed in the base kernel recently hence all those new PIIX error reports) or something else.

Quoted 3 lines

I checked 2.6.13-rc6 patch and I see no IDE / CD changes which could be responsible for this regression. You can try reverting ide-cd changes and see if this helps. IRQ routing changes?

Then you'd expect to see "drive confused" methods. It could be IRQ related - missing IRQ might do it although I'd expect that to cause crashes during probe if we got extra IRQs (because the code to handle it was removed from the base kernel even though it is required).

Another thing that will do this is if you bomb the drive with commands from somewhere (say GNOME) or crash the firmware, which is why you need to isolate all the gnome gunge for testing, and if that makes the difference see if its a kernel or higher level change and if so what its doing that causes problems.

Alan