No UDMA with BLK_DEV_IT821X with 2.6.30.2

No UDMA with BLK_DEV_IT821X with 2.6.30.2

[Dwayne Rightler]

Hello,

I have a ITE8212 PCI card and it does not seem to allow UDMA on the 
drives attached.  The drives run at ATA100 under other controllers.

Dmesg:

[    3.029612] it821x 0000:01:07.0: IDE controller (0x1283:0x8212 rev 0x13)
[    3.029835] ACPI: PCI Interrupt Link [APC4] enabled at IRQ 19
[    3.029840] pci 0000:01:07.0: PCI INT A -> Link[APC4] -> GSI 19 
(level, low) -> IRQ 19
[    3.029847] it821x 0000:01:07.0: controller in smart mode
[    3.029848] it821x 0000:01:07.0: 100% native mode on irq 19
[    3.029852]     ide1: BM-DMA at 0xcb00-0xcb07
[    3.029857]     ide2: BM-DMA at 0xcb08-0xcb0f
[    3.029859] Probing IDE interface ide1...
[    3.316151] hdc: ST3400633A, ATA DISK drive
[    3.652003] hdd: ST3400633A, ATA DISK drive
[    3.652018] hdc: Performing identify fixups.
[    3.652019] hdd: Performing identify fixups.
[    3.652022] hdc: MWDMA2 mode selected
[    3.652026] hdd: MWDMA2 mode selected
[    3.652028] Probing IDE interface ide2...
[    3.940151] hde: ST3400633A, ATA DISK drive
[    4.276003] hdf: ST3400633A, ATA DISK drive
[    4.276015] hde: Performing identify fixups.
[    4.276016] hdf: Performing identify fixups.
[    4.276019] hde: MWDMA2 mode selected
[    4.276023] hdf: MWDMA2 mode selected
[    4.276037] ide1 at 0xcf00-0xcf07,0xce02 on irq 19
[    4.276052] ide2 at 0xcd00-0xcd07,0xcc02 on irq 19
[    4.276187] ide_generic: please use "probe_mask=0x3f" module 
parameter for probing all legacy ISA IDE ports
[    4.276195] ide-gd driver 1.18
[    4.276210] hdc: max request size: 128KiB
[    4.276212] hdc: 781422768 sectors (400088 MB) w/16384KiB Cache, 
CHS=48641/255/63
[    4.276233]  hdc:<3>hdc: task_no_data_intr: status=0x51 { DriveReady 
SeekComplete Error }
[    4.280086] hdc: task_no_data_intr: error=0x04 { DriveStatusError }
[    4.280323] hdc: possibly failed opcode: 0x10
[    5.445905]  hdc1
[    5.445972] hdd: max request size: 128KiB
[    5.445974] hdd: 781422768 sectors (400088 MB) w/16384KiB Cache, 
CHS=48641/255/63
[    5.445989]  hdd:<3>hdd: task_no_data_intr: status=0x51 { DriveReady 
SeekComplete Error }
[    5.448080] hdd: task_no_data_intr: error=0x04 { DriveStatusError }
[    5.448314] hdd: possibly failed opcode: 0x10
[    6.609738]  hdd1
[    6.609795] hde: max request size: 128KiB
[    6.609797] hde: 781422768 sectors (400088 MB) w/16384KiB Cache, 
CHS=48641/255/63
[    6.609813]  hde:<3>hde: task_no_data_intr: status=0x51 { DriveReady 
SeekComplete Error }
[    6.612080] hde: task_no_data_intr: error=0x04 { DriveStatusError }
[    6.612315] hde: possibly failed opcode: 0x10
[    7.772545]  hde1
[    7.772600] hdf: max request size: 128KiB
[    7.772602] hdf: 781422768 sectors (400088 MB) w/16384KiB Cache, 
CHS=48641/255/63
[    7.772617]  hdf:<3>hdf: task_no_data_intr: status=0x51 { DriveReady 
SeekComplete Error }
[    7.776081] hdf: task_no_data_intr: error=0x04 { DriveStatusError }
[    7.776316] hdf: possibly failed opcode: 0x10
[    8.937987]  hdf1


hdparm:

/dev/hdd:

  Model=ST3400633A, FwRev=3.AAH, SerialNo=3NF1QK7K
  Config={ HardSect NotMFM HdSw>15uSec Fixed DTR>10Mbs RotSpdTol>.5% }
  RawCHS=16383/16/63, TrkSize=0, SectSize=0, ECCbytes=4
  BuffType=unknown, BuffSize=16384kB, MaxMultSect=16, MultSect=16
  CurCHS=16383/16/63, CurSects=16514064, LBA=yes, LBAsects=781422768
  IORDY=on/off, tPIO={min:240,w/IORDY:120}, tDMA={min:120,rec:120}
  PIO modes:  pio0 pio1 pio2 pio3 pio4
  DMA modes:  *mdma0 mdma1 mdma2
  AdvancedPM=no
  Drive conforms to: Unspecified:  ATA/ATAPI-1,2,3,4,5,6,7

  * signifies the current active mode

Re: No UDMA with BLK_DEV_IT821X with 2.6.30.2

[Alan Cox]

On Mon, 20 Jul 2009 18:03:22 -0500
Dwayne Rightler <drightler@technicalogic.com> wrote:

> Hello,
> 
> I have a ITE8212 PCI card and it does not seem to allow UDMA on the 
> drives attached.  The drives run at ATA100 under other controllers.

The old IDE driver support for IT821x has been somewhat broken for some
time. The drivers/ata (libata) driver should handle it fully. In smart
mode it's "IDE-ish" and the old IDE code really isn't flexible enough to
cope. With the "ATAPI BIOS" its a fairly normal controller so both drivers
ought to work.

Re: No UDMA with BLK_DEV_IT821X with 2.6.30.2

[Dwayne Rightler]

With the libata driver and "ATAPI BIOS" I get this sort of garbage (note 
I don't get these drive errors on other controllers):

[  604.314837] ata8.01: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x6
[  604.314895] ata8.01: BMDMA stat 0x4
[  604.314951] ata8.01: cmd c8/00:00:e0:20:e4/00:00:00:00:00/f1 tag 0 
dma 131072 in
[  604.314952]          res 00/00:00:80:00:80/00:00:00:00:00/80 Emask 
0x2 (HSM violation)
[  604.315097] ata8: soft resetting link
[  605.462588] ata8.00: configured for UDMA/100
[  605.508491] ata8.01: configured for UDMA/100
[  605.508501] ata8: EH complete
[  979.174939] ata8.01: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x6
[  979.174949] ata8.01: BMDMA stat 0x4
[  979.174958] ata8.01: cmd c8/00:00:e0:4e:89/00:00:00:00:00/f4 tag 0 
dma 131072 in
[  979.174959]          res 00/00:00:80:00:80/00:00:00:00:00/80 Emask 
0x2 (HSM violation)
[  979.175008] ata8: soft resetting link
[  979.944287] ata8.01: NODEV after polling detection
[  979.944290] ata8.01: revalidation failed (errno=-2)
[  984.944026] ata8: soft resetting link
[  985.100384] ata8.01: NODEV after polling detection
[  985.100388] ata8.01: revalidation failed (errno=-2)
[  990.100032] ata8: soft resetting link
[  990.256322] ata8.01: NODEV after polling detection
[  990.256325] ata8.01: revalidation failed (errno=-2)
[  990.256334] ata8.01: disabled
[  990.256338] ata8.00: failed to IDENTIFY (I/O error, err_mask=0x40)
[  990.256340] ata8.00: revalidation failed (errno=-5)
[  995.256024] ata8: soft resetting link
[  995.539133] ata8.00: configured for UDMA/100
[  995.539158] ata8: EH complete
[  995.539185] sd 7:0:1:0: [sdg] Unhandled error code
[  995.539187] sd 7:0:1:0: [sdg] Result: hostbyte=DID_BAD_TARGET 
driverbyte=DRIVER_OK
[  995.539192] end_request: I/O error, dev sdg, sector 76107488
[  995.539206] Buffer I/O error on device sdg, logical block 9513436
[  995.539215] Buffer I/O error on device sdg, logical block 9513437
[  995.539222] Buffer I/O error on device sdg, logical block 9513438
[  995.539229] Buffer I/O error on device sdg, logical block 9513439
[  995.539244] Buffer I/O error on device sdg, logical block 9513440
[  995.539260] Buffer I/O error on device sdg, logical block 9513441
[  995.539276] Buffer I/O error on device sdg, logical block 9513442
[  995.539292] Buffer I/O error on device sdg, logical block 9513443
[  995.539308] Buffer I/O error on device sdg, logical block 9513444
[  995.539323] Buffer I/O error on device sdg, logical block 9513445
[  995.539359] sd 7:0:1:0: [sdg] Unhandled error code
[  995.539361] sd 7:0:1:0: [sdg] Result: hostbyte=DID_BAD_TARGET 
driverbyte=DRIVER_OK
[  995.539364] end_request: I/O error, dev sdg, sector 76107744
[  995.539411] sd 7:0:1:0: [sdg] Unhandled error code
[  995.539413] sd 7:0:1:0: [sdg] Result: hostbyte=DID_BAD_TARGET 
driverbyte=DRIVER_OK
[  995.539415] end_request: I/O error, dev sdg, sector 76107488


After this happens hdparm fails:

root@darkomen-desktop:~# hdparm -i /dev/sdg

/dev/sdg:
  HDIO_GET_IDENTITY failed: No message of desired type


As well as smartctl:

root@darkomen-desktop:~# smartctl -H /dev/sdg
smartctl version 5.38 [x86_64-unknown-linux-gnu] Copyright (C) 2002-8 
Bruce Allen
Home page is http://smartmontools.sourceforge.net/

Short INQUIRY response, skip product id
A mandatory SMART command failed: exiting. To continue, add one or more 
'-T permissive' options.



Alan Cox wrote:
> On Mon, 20 Jul 2009 18:03:22 -0500
> Dwayne Rightler <drightler@technicalogic.com> wrote:
> 
>> Hello,
>>
>> I have a ITE8212 PCI card and it does not seem to allow UDMA on the 
>> drives attached.  The drives run at ATA100 under other controllers.
> 
> The old IDE driver support for IT821x has been somewhat broken for some
> time. The drivers/ata (libata) driver should handle it fully. In smart
> mode it's "IDE-ish" and the old IDE code really isn't flexible enough to
> cope. With the "ATAPI BIOS" its a fairly normal controller so both drivers
> ought to work.

Re: No UDMA with BLK_DEV_IT821X with 2.6.30.2

[Bartlomiej Zolnierkiewicz]

On Tuesday 21 July 2009 01:03:22 Dwayne Rightler wrote:
> Hello,
> 
> I have a ITE8212 PCI card and it does not seem to allow UDMA on the 
> drives attached.  The drives run at ATA100 under other controllers.
> 
> Dmesg:
> 
> [    3.029612] it821x 0000:01:07.0: IDE controller (0x1283:0x8212 rev 0x13)
> [    3.029835] ACPI: PCI Interrupt Link [APC4] enabled at IRQ 19
> [    3.029840] pci 0000:01:07.0: PCI INT A -> Link[APC4] -> GSI 19 
> (level, low) -> IRQ 19
> [    3.029847] it821x 0000:01:07.0: controller in smart mode

The controller is running in the "smart mode" in which it takes care
of setting the transfer speed (the driver reports standard MWDMA2 mode
to higher-layers for compatibility reasons).