Skip to contentSkip to conversation

sata_promise KERNEL_BUG on 2.6.12

DISCUSSION Marcel Naziri 3 participants 6 messages Jun 20, 2005

lore
Share & tools
lore
#0 lore raw

Hi,

this is my first posting to the LKML and I have to report a problem with the sata_promise driver:
I've installed a Promise SATAII-150 TX4 with two Samsung drives. One drive is native SATA, the other is adapted via some Silicon Image SATA-PATA-bridge.

$ cat /proc/version
Linux version 2.6.12 (root@Carter) (gcc version 4.0.1 20050522 (prerelease) 
(Debian 4.0.0-9)) #1 Mon Jun 20 01:27:35 CEST 2005

Now, when I connect the drives to port 1 & 2 of the controller, booting up stops with this:
ACPI: PCI Interrupt 0000:00:0a.0[A] -> GSI 17 (level, low) -> IRQ 17
ata1: SATA max UDMA/133 cmd 0xF8802200 ctl 0xF8802238 bdma 0x0 irq 17
ata2: SATA max UDMA/133 cmd 0xF8802280 ctl 0xF88022B8 bdma 0x0 irq 17
ata3: SATA max UDMA/133 cmd 0xF8802300 ctl 0xF8802338 bdma 0x0 irq 17
ata4: SATA max UDMA/133 cmd 0xF8802380 ctl 0xF88023B8 bdma 0x0 irq 17
ata1: no device found (phy stat 00000000)
scsi0 : sata_promise
ata2: dev 0 ATA, max UDMA/100, 312581808 sectors: lba48

------------[ cut here ]------------
kernel BUG at drivers/scsi/libata-core.c:2077!
invalid operand: 0000 [#1]
PREEMPT
Modules linked in:
CPU:    0
EIP:    0060:[<c025f60f>]     Not tainted VLI
EFLAGS: 00010246   (2.6.12)

EIP is at ata_dev_set_xfermode+0xcf/0xf0
eax: 00000000 ebx: c1bdea7c ecx: 00000000 edx: 00000000
esi: c1bdea7c edi: c1bdea00 ebp: c193fe2c esp: c193fe24
ds: 007b es: 007b ss: 0068
Process swapper (pid: 1, threadinfo=c193e000 task=c18f6a20)

Stack: c1bdea00 c1bdea7c 00000000 c193fe30 c193fe30 c1bdea7c c1bdea7c c1bdea00
       00000001 c025e813 c1bdea00 c1bdea7c 0000003f c1bdea00 c1bdea00 c1bdea7c
       00000440 c025eabd c1bdea00 c1bdea7c 00000000 00000000 45000001 c1bdea00
Call Trace:
 [<c025e813>] ata_dev_set_mode+0x53/0x110
 [<c025eabd>] ata_set_mode+0x7d/0xd0
 [<c025e5f8>] ata_bus_probe+0xb8/0xd0
 [<c0261605>] ata_device_add+0x215/0x2d0
 [<c02644e3>] pdc_ata_init_one+0x223/0x2e0
 [<c01cf605>] pci_device_probe_static+0x45/0x50
 [<c01cf64b>] __pci_device_probe+0x3b/0x60
 [<c01cf69c>] pci_device_probe+0x2c/0x50
 [<c022638f>] driver_probe_device+0x2f/0x80
 [<c0226509>] driver_attach+0x59/0x90
 [<c02269e8>] bus_add_driver+0x98/0xe0
 [<c01cf915>] pci_register_driver+0x55/0x90
 [<c03da70f>] pdc_ata_init+0xf/0x20
 [<c03bc948>] do_initcalls+0x58/0xd0
 [<c01002b0>] init+0x0/0x120
 [<c01002df>] init+0x2f/0x120
 [<c0101334>] kernel_thread_helper+0x0/0xc
 [<c0101339>] kernel_thread_helper+0x5/0xc
Code: 8b 74 24 18 8b 7c 24 1c 8b 6c 24 20 83 c4 24 c3 89 e8 e8 05 c5 08 00 8b 

5c 24 14 8b 74 24 18 8b 7c 24 1c 8b 6c 24 20 83 c4 24 c3 <0f> 0b 1d 08 16 fc 30 c0
e9 6d ff ff ff e8 0f c4 08 00 eb ad 8d
<0>Kernel panic - not syncing: Attempted to kill init!

When I connect the drives on port 3 & 4 of the controller, kernel gets running:
libata version 1.11 loaded.
sata_promise version 1.01
ACPI: PCI Interrupt 0000:00:0a.0[A] -> GSI 17 (level, low) -> IRQ 17
ata1: SATA max UDMA/133 cmd 0xF8802200 ctl 0xF8802238 bmdma 0x0 irq 17
ata2: SATA max UDMA/133 cmd 0xF8802280 ctl 0xF88022B8 bmdma 0x0 irq 17
ata3: SATA max UDMA/133 cmd 0xF8802300 ctl 0xF8802338 bmdma 0x0 irq 17
ata4: SATA max UDMA/133 cmd 0xF8802380 ctl 0xF88023B8 bmdma 0x0 irq 17
ata1: dev 0 cfg 49:2f00 82:346b 83:7f01 84:4003 85:3c69 86:3c01 87:4003 88:80ff
ata1: dev 0 ATA, max UDMA7, 312581808 sectors: lba48
ata1: dev 0 configured for UDMA/133
scsi0 : sata_promise
ata2: no device found (phy stat 00000000)
scsi1 : sata_promise
ata3: dev 0 cfg 49:2f00 82:346b 83:7f01 84:4003 85:3c69 86:3e01 87:4003 88:203f
ata3: dev 0 ATA, max UDMA/100, 312581808 sectors: lba48
ata3: dev 0 configured for UDMA/100
scsi2 : sata_promise
ata4: no device found (phy stat 00000000)
scsi3 : sata_promise

  Vendor: ATA       Model: SAMSUNG SP1614C   Rev: SW10
  Type:   Direct-Access                      ANSI SCSI revision: 05
  Vendor: ATA       Model: SAMSUNG SV1604N   Rev: TR10
  Type:   Direct-Access                      ANSI SCSI revision: 05

SCSI device sda: 312581808 512-byte hdwr sectors (160042 MB) SCSI device sda: drive cache: write back
SCSI device sda: 312581808 512-byte hdwr sectors (160042 MB) SCSI device sda: drive cache: write back
sda: sda1 sda2 sda3 sda4 < sda5 sda6 sda7 sda8 sda9 sda10 >
Attached scsi disk sda at scsi0, channel 0, id 0, lun 0 SCSI device sdb: 312581808 512-byte hdwr sectors (160042 MB) SCSI device sdb: drive cache: write back
SCSI device sdb: 312581808 512-byte hdwr sectors (160042 MB) SCSI device sdb: drive cache: write back
sdb: sdb1 sdb2 sdb3 sdb4 < sdb5 sdb6 sdb7 sdb8 sdb9 sdb10 >
Attached scsi disk sdb at scsi2, channel 0, id 0, lun 0

Can it deal with the fact, that the drives are not scanned in port order of the controller? They seem to be mapped like
port 1 > ata4
port 2 > ata2
port 3 > ata1
port 4 > ata3

With Promise ulsata driver on 2.4.x kernels i have no problems with drives connected on port 1 & 2.
What can I do? If more info is needed, please bug me... :)

Thanks for help!
zwobbl

#1 lore raw
Quoted 17 lines · Marcel Naziri

Now, when I connect the drives to port 1 & 2 of the controller, booting up stops with this:
ACPI: PCI Interrupt 0000:00:0a.0[A] -> GSI 17 (level, low) -> IRQ 17
ata1: SATA max UDMA/133 cmd 0xF8802200 ctl 0xF8802238 bdma 0x0 irq 17
ata2: SATA max UDMA/133 cmd 0xF8802280 ctl 0xF88022B8 bdma 0x0 irq 17
ata3: SATA max UDMA/133 cmd 0xF8802300 ctl 0xF8802338 bdma 0x0 irq 17
ata4: SATA max UDMA/133 cmd 0xF8802380 ctl 0xF88023B8 bdma 0x0 irq 17
ata1: no device found (phy stat 00000000)
scsi0 : sata_promise
ata2: dev 0 ATA, max UDMA/100, 312581808 sectors: lba48

------------[ cut here ]------------
kernel BUG at drivers/scsi/libata-core.c:2077!
invalid operand: 0000 [#1]
PREEMPT
Modules linked in:
CPU:    0
EIP:    0060:[<c025f60f>]     Not tainted VLI
EFLAGS: 00010246   (2.6.12)
Quoted 1 line

EIP is at ata_dev_set_xfermode+0xcf/0xf0

This is highly strange. Do you have any patches applied, or is this vanilla 2.6.12 kernel?

Can you turn off preempt and try to reproduce ?
Can you provide your full .config ?

Quoted 5 lines

Can it deal with the fact, that the drives are not scanned in port order of the controller? They seem to be mapped like port 1 > ata4
port 2 > ata2
port 3 > ata1
port 4 > ata3

The driver scans the ports in the order presented internally in the hardware.

Jeff

#2 lore raw

Also sprach Jeff Garzik am Montag 20 Juni 2005 04:25:

Quoted 21 lines
Quoted 18 lines · Marcel Naziri

Now, when I connect the drives to port 1 & 2 of the controller, booting up stops with this:
ACPI: PCI Interrupt 0000:00:0a.0[A] -> GSI 17 (level, low) -> IRQ 17
ata1: SATA max UDMA/133 cmd 0xF8802200 ctl 0xF8802238 bdma 0x0 irq 17
ata2: SATA max UDMA/133 cmd 0xF8802280 ctl 0xF88022B8 bdma 0x0 irq 17
ata3: SATA max UDMA/133 cmd 0xF8802300 ctl 0xF8802338 bdma 0x0 irq 17
ata4: SATA max UDMA/133 cmd 0xF8802380 ctl 0xF88023B8 bdma 0x0 irq 17
ata1: no device found (phy stat 00000000)
scsi0 : sata_promise
ata2: dev 0 ATA, max UDMA/100, 312581808 sectors: lba48

------------[ cut here ]------------
kernel BUG at drivers/scsi/libata-core.c:2077!
invalid operand: 0000 [#1]
PREEMPT
Modules linked in:
CPU:    0
EIP:    0060:[<c025f60f>]     Not tainted VLI
EFLAGS: 00010246   (2.6.12)
Quoted 1 line

EIP is at ata_dev_set_xfermode+0xcf/0xf0

Quoted 2 lines

This is highly strange. Do you have any patches applied, or is this vanilla 2.6.12 kernel?

it's vanilla. and I have this problem since 2.6.10 (that is the first one i tried with the controller)

Quoted 1 line

Can you turn off preempt and try to reproduce ?

I can try this tomorrow...

Quoted 1 line

Can you provide your full .config ?

Is attached...

Quoted 9 lines
Quoted 6 lines

Can it deal with the fact, that the drives are not scanned in port order of the controller? They seem to be mapped like
port 1 > ata4
port 2 > ata2
port 3 > ata1
port 4 > ata3

The driver scans the ports in the order presented internally in the hardware.

Could the driver "remap" it? It's confusing that the boot loader sees the drives in another way than the kernel do.

Greets
zwobbl

#3 lore raw
Quoted 1 line · Marcel Naziri

Could the driver "remap" it? It's confusing that the boot loader sees the drives in another way than the kernel do.

Unfortunately it largely depends on how the board maker decided to wire up the ports. If the Promise driver gets it right, I probably need to poke into BIOS somewhere...

Jeff

#4 lore raw
In reply to Jeff Garzik

Also sprach Jeff Garzik am Montag 20 Juni 2005 04:25:

Quoted 21 lines
Quoted 18 lines · Marcel Naziri

Now, when I connect the drives to port 1 & 2 of the controller, booting up stops with this:
ACPI: PCI Interrupt 0000:00:0a.0[A] -> GSI 17 (level, low) -> IRQ 17
ata1: SATA max UDMA/133 cmd 0xF8802200 ctl 0xF8802238 bdma 0x0 irq 17
ata2: SATA max UDMA/133 cmd 0xF8802280 ctl 0xF88022B8 bdma 0x0 irq 17
ata3: SATA max UDMA/133 cmd 0xF8802300 ctl 0xF8802338 bdma 0x0 irq 17
ata4: SATA max UDMA/133 cmd 0xF8802380 ctl 0xF88023B8 bdma 0x0 irq 17
ata1: no device found (phy stat 00000000)
scsi0 : sata_promise
ata2: dev 0 ATA, max UDMA/100, 312581808 sectors: lba48

------------[ cut here ]------------
kernel BUG at drivers/scsi/libata-core.c:2077!
invalid operand: 0000 [#1]
PREEMPT
Modules linked in:
CPU:    0
EIP:    0060:[<c025f60f>]     Not tainted VLI
EFLAGS: 00010246   (2.6.12)
Quoted 1 line

EIP is at ata_dev_set_xfermode+0xcf/0xf0

Quoted 4 lines

This is highly strange. Do you have any patches applied, or is this vanilla 2.6.12 kernel?

Can you turn off preempt and try to reproduce ?

With preempt disabled it boots up. drives attached to port 1 & 2 gets attached to ata4 & ata2:

libata version 1.11 loaded.
sata_promise version 1.01
ACPI: PCI Interrupt 0000:00:0a.0[A] -> GSI 17 (level, low) -> IRQ 17
ata1: SATA max UDMA/133 cmd 0xF8802200 ctl 0xF8802238 bmdma 0x0 irq 17
ata2: SATA max UDMA/133 cmd 0xF8802280 ctl 0xF88022B8 bmdma 0x0 irq 17
ata3: SATA max UDMA/133 cmd 0xF8802300 ctl 0xF8802338 bmdma 0x0 irq 17
ata4: SATA max UDMA/133 cmd 0xF8802380 ctl 0xF88023B8 bmdma 0x0 irq 17
ata1: no device found (phy stat 00000000)
scsi0 : sata_promise
ata2: dev 0 cfg 49:2f00 82:346b 83:7f01 84:4003 85:3c69 86:3e01 87:4003 88:203f
ata2: dev 0 ATA, max UDMA/100, 312581808 sectors: lba48
ata2: dev 0 configured for UDMA/100
scsi1 : sata_promise
ata3: no device found (phy stat 00000000)
scsi2 : sata_promise
ata4: dev 0 cfg 49:2f00 82:346b 83:7f01 84:4003 85:3c69 86:3c01 87:4003 88:80ff
ata4: dev 0 ATA, max UDMA7, 312581808 sectors: lba48
ata4: dev 0 configured for UDMA/133
scsi3 : sata_promise

  Vendor: ATA       Model: SAMSUNG SV1604N   Rev: TR10
  Type:   Direct-Access                      ANSI SCSI revision: 05
  Vendor: ATA       Model: SAMSUNG SP1614C   Rev: SW10
  Type:   Direct-Access                      ANSI SCSI revision: 05

SCSI device sda: 312581808 512-byte hdwr sectors (160042 MB) SCSI device sda: drive cache: write back
SCSI device sda: 312581808 512-byte hdwr sectors (160042 MB) SCSI device sda: drive cache: write back
sda: sda1 sda2 sda3 sda4 < sda5 sda6 sda7 sda8 sda9 sda10 >
Attached scsi disk sda at scsi1, channel 0, id 0, lun 0 SCSI device sdb: 312581808 512-byte hdwr sectors (160042 MB) SCSI device sdb: drive cache: write back
SCSI device sdb: 312581808 512-byte hdwr sectors (160042 MB) SCSI device sdb: drive cache: write back
sdb: sdb1 sdb2 sdb3 sdb4 < sdb5 sdb6 sdb7 sdb8 sdb9 sdb10 >
Attached scsi disk sdb at scsi3, channel 0, id 0, lun 0

What else can I try? What helpful information can I provide?

# lspci -v -s 0a.0
0000:00:0a.0 Unknown mass storage controller: Promise Technology, Inc.: 
Unknown device 3d18 (rev 02)
        Subsystem: Promise Technology, Inc.: Unknown device 3d18
        Flags: bus master, 66MHz, medium devsel, latency 72, IRQ 17
        I/O ports at e800 [size=128]
        I/O ports at e400 [size=256]
        Memory at dffff000 (32-bit, non-prefetchable) [size=4K]
        Memory at dffc0000 (32-bit, non-prefetchable) [size=128K]
        Expansion ROM at dffe0000 [disabled] [size=32K]
        Capabilities: [60] Power Management version 2

Hopefully with preempt disabled also these sporadic messages in syslog will disappear:

Jun 20 01:28:31 Carter kernel: ata3: status=0x51 { DriveReady SeekComplete 
Error }
Jun 20 01:28:31 Carter kernel: ata3: called with no error (51)!

But I think that the one bridged ATA-drive may cause this warning/error!? Is it harmful?
On very rare conditions sometimes it results in harder errors like that ending up in lockup filesystems and my box:

May 25 01:25:50 Carter kernel: ata1: command timeout
May 25 01:25:50 Carter kernel: ata1: status=0x51 { DriveReady SeekComplete 
Error }
May 25 01:25:50 Carter kernel: ata1: called with no error (51)!
May 25 01:26:20 Carter kernel: ata1: command timeout
May 25 01:26:20 Carter kernel: ATA: abnormal status 0xFF on port 0xF880221C
May 25 01:26:20 Carter kernel: ata1: status=0xff { Busy }
May 25 01:26:20 Carter kernel: SCSI error : <0 0 0 0> return code = 0x8000002
May 25 01:26:20 Carter kernel: sda: Current: sense key: Aborted Command
May 25 01:26:20 Carter kernel: Additional sense: Scsi parity error
May 25 01:26:20 Carter kernel: end_request: I/O error, dev sda, sector 9079889
May 25 01:26:50 Carter kernel: ata1: command timeout
May 25 01:26:50 Carter kernel: ATA: abnormal status 0xFF on port 0xF880221C
May 25 01:26:50 Carter kernel: ata1: status=0xff { Busy }
May 25 01:26:50 Carter kernel: SCSI error : <0 0 0 0> return code = 0x8000002
May 25 01:26:50 Carter kernel: sda: Current: sense key: Aborted Command
May 25 01:26:50 Carter kernel: Additional sense: Scsi parity error
May 25 01:26:50 Carter kernel: end_request: I/O error, dev sda, sector 9079897
May 25 01:27:20 Carter kernel: ata1: command timeout
May 25 01:27:20 Carter kernel: ATA: abnormal status 0xFF on port 0xF880221C
May 25 01:27:20 Carter kernel: ata1: status=0xff { Busy }
May 25 01:27:20 Carter kernel: SCSI error : <0 0 0 0> return code = 0x8000002
May 25 01:27:20 Carter kernel: sda: Current: sense key: Aborted Command
May 25 01:27:20 Carter kernel: Additional sense: Scsi parity error
May 25 01:27:20 Carter kernel: end_request: I/O error, dev sda, sector 9079905
May 25 01:27:50 Carter kernel: ata1: command timeout
May 25 01:27:50 Carter kernel: ATA: abnormal status 0xFF on port 0xF880221C
May 25 01:27:50 Carter kernel: ata1: status=0xff { Busy }
May 25 01:27:50 Carter kernel: SCSI error : <0 0 0 0> return code = 0x8000002
May 25 01:27:50 Carter kernel: sda: Current: sense key: Aborted Command
May 25 01:27:50 Carter kernel: Additional sense: Scsi parity error
May 25 01:27:50 Carter kernel: end_request: I/O error, dev sda, sector 9079913
May 25 01:28:20 Carter kernel: ata1: command timeout
May 25 01:28:20 Carter kernel: ATA: abnormal status 0xFF on port 0xF880221C
May 25 01:28:20 Carter kernel: ata1: status=0xff { Busy }
May 25 01:28:20 Carter kernel: SCSI error : <0 0 0 0> return code = 0x8000002
May 25 01:28:20 Carter kernel: sda: Current: sense key: Aborted Command
May 25 01:28:20 Carter kernel: Additional sense: Scsi parity error
May 25 01:28:20 Carter kernel: end_request: I/O error, dev sda, sector 9079921
May 25 01:28:46 Carter syslog-ng[2376]: STATS: dropped 0
May 25 01:28:50 Carter kernel: ata1: command timeout
May 25 01:28:50 Carter kernel: ATA: abnormal status 0xFF on port 0xF880221C
May 25 01:28:50 Carter kernel: ata1: status=0xff { Busy }
May 25 01:28:50 Carter kernel: SCSI error : <0 0 0 0> return code = 0x8000002
May 25 01:28:50 Carter kernel: sda: Current: sense key: Aborted Command
May 25 01:28:50 Carter kernel: Additional sense: Scsi parity error
May 25 01:28:50 Carter kernel: end_request: I/O error, dev sda, sector 9079929
May 25 01:30:05 Carter syslog-ng[2680]: syslog-ng version 1.6.5 starting

Greets
zwobbl>

#5 lore raw

I can confirm the the problem on my Promise SATAII-150 TX4 with the 2.6.12. However after a immediate downgrade to 2.6.11 the problem persist.
updating the controller bios to 1.00.0.34 didn't helped either.

The kernel freezes the interrupt from the satalib after a short time. strangly it occurs only after even numbers (like 200000 or 300000). Bevor the freez, all devices are working fine except for some

[kernel] irq 18: nobody cared!

messages.

Simon