Skip to contentSkip to conversation

HP server and atto driver from hell

DISCUSSION Ralf Hildebrandt 6 participants 13 messages Apr 29, 2008 – May 9, 2008

lore
Share & tools
lore
#0 lore raw

We thought HP would be a smart choice when it comes to Linux,.. Until now, because now we have to use a PCI Express SCSI card that comes with a proprietary driver that won't even compile.

When I untar the lnx_drv_epciu320_107.tar.gz and try to compile the glue between the binary blob and the kernel, I get:

c# make -k modules
Architecture is i386 (x86_regparm)
make[1]: Entering directory /usr/src/linux-2.6.25'
CC [M] /usr/src/lnx_drv_epciu320_107/src/atto_cfg.o
/usr/src/lnx_drv_epciu320_107/src/atto_cfg.c: In function ‘attocfg_read_version’: /usr/src/lnx_drv_epciu320_107/src/atto_cfg.c:258: error: ‘struct device_driver’ has no member named ‘kobj’
make[2]: *** [/usr/src/lnx_drv_epciu320_107/src/atto_cfg.o] Error 1
CC [M] /usr/src/lnx_drv_epciu320_107/src/oswrap.o
/usr/src/lnx_drv_epciu320_107/src/oswrap.c: In function ‘express2_attocfg_register’: /usr/src/lnx_drv_epciu320_107/src/oswrap.c:908: error: ‘struct device_driver’ has no member named ‘kobj’
make[2]: *** [/usr/src/lnx_drv_epciu320_107/src/oswrap.o] Error 1
make[2]: Target __build' not remade because of errors.
make[1]: *** [_module_/usr/src/lnx_drv_epciu320_107/src] Error 2
make[1]: Target modules' not remade because of errors.
make[1]: Leaving directory /usr/src/linux-2.6.25'
make: *** [modules] Error 2

in atto_cfg.c:

   258          if (this_drvr->kobj == &driver->kobj)
   259          {
   260              if (this_drvr->read_version)
   261                  return (*this_drvr->read_version) (driver, buffer);
   262              else
   263                  return 0;
   264          }

and in oswrap.o:

905 return NULL;
906 }
907
908 return &atto_pci_driver.driver.kobj;
909 }
910

Signature · 4 lines
Ralf Hildebrandt (i.A. des IT-Zentrums)         [email protected]
Charite - Universitätsmedizin Berlin            Tel.  +49 (0)30-450 570-155
Gemeinsame Einrichtung von FU- und HU-Berlin    Fax.  +49 (0)30-450 570-962
IT-Zentrum Standort CBF                 send no mail to [email protected]
#1 lore raw
Quoted 3 lines · On Tue, 29 Apr 2008 22:17:27 +0200 Ralf Hildebrandt…

We thought HP would be a smart choice when it comes to Linux,.. Until now, because now we have to use a PCI Express SCSI card that comes with a proprietary driver that won't even compile.

I would suggest you take it up with the supplier, given EU law on product being fit for the purpose for which it was sold I would hope they jump.

Seriously it isn't going to be worth anyones time fighting binary crap when there are open alternatives available. Tell HP to replace it with a working card - also if you mail ordered it remember your EU wide 14 day return option ;)

And please don't post pieces of proprietary driver code to the kernel list - I doubt you have permission to do so from HP.

Alan

#2 lore raw

* Alan Cox <[email protected]>:

Quoted 2 lines

I would suggest you take it up with the supplier, given EU law on product being fit for the purpose for which it was sold I would hope they jump.

Hah!

Quoted 4 lines

Seriously it isn't going to be worth anyones time fighting binary crap when there are open alternatives available. Tell HP to replace it with a working card - also if you mail ordered it remember your EU wide 14 day return option ;)

Will ask the guy who bought it.

Quoted 2 lines

And please don't post pieces of proprietary driver code to the kernel list - I doubt you have permission to do so from HP.

Sorry.

Signature · 4 lines
Ralf Hildebrandt (i.A. des IT-Zentrums)         [email protected]
Charite - Universitätsmedizin Berlin            Tel.  +49 (0)30-450 570-155
Gemeinsame Einrichtung von FU- und HU-Berlin    Fax.  +49 (0)30-450 570-962
IT-Zentrum Standort CBF                 send no mail to [email protected]
#3 lore raw

* Ralf Hildebrandt <[email protected]>:

Quoted 11 lines
Quoted 2 lines

I would suggest you take it up with the supplier, given EU law on product being fit for the purpose for which it was sold I would hope they jump.

Hah!

Quoted 4 lines

Seriously it isn't going to be worth anyones time fighting binary crap when there are open alternatives available. Tell HP to replace it with a working card - also if you mail ordered it remember your EU wide 14 day return option ;)

Will ask the guy who bought it.

The thing works with the mptspi driver just like a charm. I wonder why anybody would need the other driver...

Signature · 4 lines
Ralf Hildebrandt (i.A. des IT-Zentrums)         [email protected]
Charite - Universitätsmedizin Berlin            Tel.  +49 (0)30-450 570-155
Gemeinsame Einrichtung von FU- und HU-Berlin    Fax.  +49 (0)30-450 570-962
IT-Zentrum Standort CBF                 send no mail to [email protected]
#4 lore raw
In reply to Ralf Hildebrandt

Hello Ralf , Fyi , the fusion/mpt driver supports some of the ATTO
devices , if your lspci -v for the ATTO controllers looks simular to this one it may just work with the fusion drivers .
Hth , JimL

# lspci -v -s 0d:04.0
0d:04.0 SCSI storage controller: Atto Technology Ultra320 SCSI Host Adapter (rev 08)
        Subsystem: Atto Technology Unknown device 8027
        Flags: bus master, 66MHz, medium devsel, latency 128, IRQ 16
        I/O ports at 5000 [size=256]
        Memory at d8640000 (64-bit, non-prefetchable) [size=256K]
        Memory at d8600000 (64-bit, non-prefetchable) [size=256K]
        Capabilities: [50] Power Management version 2
        Capabilities: [58] Message Signalled Interrupts: 64bit+ Queue=0/0 Enable-
        Capabilities: [68] PCI-X non-bridge device
Quoted 43 lines · Ralf Hildebrandt

We thought HP would be a smart choice when it comes to Linux,.. Until now, because now we have to use a PCI Express SCSI card that comes with a proprietary driver that won't even compile.

When I untar the lnx_drv_epciu320_107.tar.gz and try to compile the glue between the binary blob and the kernel, I get:

c# make -k modules
Architecture is i386 (x86_regparm)
make[1]: Entering directory /usr/src/linux-2.6.25'
CC [M] /usr/src/lnx_drv_epciu320_107/src/atto_cfg.o
/usr/src/lnx_drv_epciu320_107/src/atto_cfg.c: In function ‘attocfg_read_version’: /usr/src/lnx_drv_epciu320_107/src/atto_cfg.c:258: error: ‘struct device_driver’ has no member named ‘kobj’
make[2]: *** [/usr/src/lnx_drv_epciu320_107/src/atto_cfg.o] Error 1
CC [M] /usr/src/lnx_drv_epciu320_107/src/oswrap.o
/usr/src/lnx_drv_epciu320_107/src/oswrap.c: In function ‘express2_attocfg_register’: /usr/src/lnx_drv_epciu320_107/src/oswrap.c:908: error: ‘struct device_driver’ has no member named ‘kobj’
make[2]: *** [/usr/src/lnx_drv_epciu320_107/src/oswrap.o] Error 1
make[2]: Target __build' not remade because of errors.
make[1]: *** [_module_/usr/src/lnx_drv_epciu320_107/src] Error 2
make[1]: Target modules' not remade because of errors.
make[1]: Leaving directory /usr/src/linux-2.6.25'
make: *** [modules] Error 2

in atto_cfg.c:

  258          if (this_drvr->kobj == &driver->kobj)
  259          {
  260              if (this_drvr->read_version)
  261                  return (*this_drvr->read_version) (driver, buffer);
  262              else
  263                  return 0;
  264          }

and in oswrap.o:

905 return NULL;
906 }
907
908 return &atto_pci_driver.driver.kobj;
909 }
910

Signature · 5 lines
+------------------------------------------------------------------+
| James   W.   Laferriere | System    Techniques | Give me VMS     |
| Network&System Engineer | 2133    McCullam Ave |  Give me Linux  |
| [email protected] | Fairbanks, AK. 99701 |   only  on  AXP |
+------------------------------------------------------------------+
#5 lore raw

* Mr. James W. Laferriere <[email protected]>:

Quoted 3 lines

Hello Ralf , Fyi , the fusion/mpt driver supports some of the ATTO devices , if your lspci -v for the ATTO controllers looks simular to this one it may just work with the fusion drivers . Hth , JimL

I will try.

Quoted 10 lines
# lspci -v -s 0d:04.0
0d:04.0 SCSI storage controller: Atto Technology Ultra320 SCSI Host Adapter (rev 08)
        Subsystem: Atto Technology Unknown device 8027
        Flags: bus master, 66MHz, medium devsel, latency 128, IRQ 16
        I/O ports at 5000 [size=256]
        Memory at d8640000 (64-bit, non-prefetchable) [size=256K]
        Memory at d8600000 (64-bit, non-prefetchable) [size=256K]
        Capabilities: [50] Power Management version 2
        Capabilities: [58] Message Signalled Interrupts: 64bit+ Queue=0/0 Enable-
        Capabilities: [68] PCI-X non-bridge device

The device is 802F, not 8027, but I'll try.

Signature · 4 lines
Ralf Hildebrandt (i.A. des IT-Zentrums)         [email protected]
Charite - Universitätsmedizin Berlin            Tel.  +49 (0)30-450 570-155
Gemeinsame Einrichtung von FU- und HU-Berlin    Fax.  +49 (0)30-450 570-962
IT-Zentrum Standort CBF                 send no mail to [email protected]
#6 lore raw
In reply to Mr. James W. Laferriere

* Mr. James W. Laferriere <[email protected]>:

Quoted 5 lines

Hello Ralf , Fyi , the fusion/mpt driver supports some of the ATTO
devices , if your lspci -v for the ATTO controllers looks simular to this one it may just work with the fusion drivers .
Hth , JimL

Quoted 10 lines
# lspci -v -s 0d:04.0
0d:04.0 SCSI storage controller: Atto Technology Ultra320 SCSI Host Adapter (rev 08)
        Subsystem: Atto Technology Unknown device 8027
        Flags: bus master, 66MHz, medium devsel, latency 128, IRQ 16
        I/O ports at 5000 [size=256]
        Memory at d8640000 (64-bit, non-prefetchable) [size=256K]
        Memory at d8600000 (64-bit, non-prefetchable) [size=256K]
        Capabilities: [50] Power Management version 2
        Capabilities: [58] Message Signalled Interrupts: 64bit+ Queue=0/0 Enable-
        Capabilities: [68] PCI-X non-bridge device

Well, something seems to work:

[    3.571216] Fusion MPT base driver 3.04.06
[    3.571274] Copyright (c) 1999-2007 LSI Corporation
[    3.577292] Fusion MPT SPI Host driver 3.04.06
[    3.577389] ACPI: PCI Interrupt 0000:14:04.0[A] -> GSI 17 (level, low) -> IRQ 17
[    3.577512] mptbase: ioc0: Initiating bringup
[    3.836271] ioc0: LSI53C1030 C0: Capabilities={Initiator,Target}
[    3.992277] scsi0 : ioc0: LSI53C1030 C0, FwRev=01032700h, Ports=1, MaxQ=255, IRQ=17
[    3.992277] ACPI: PCI Interrupt 0000:14:04.1[B] -> GSI 18 (level, low) -> IRQ 18
[    3.992277] mptbase: ioc1: Initiating bringup
[    4.244297] ioc1: LSI53C1030 C0: Capabilities={Initiator,Target}
[    4.408304] scsi1 : ioc1: LSI53C1030 C0, FwRev=01032700h, Ports=1, MaxQ=255, IRQ=18
# lspci -v -s 14:04.0
14:04.0 SCSI storage controller: Atto Technology Ultra320 SCSI Host Adapter (rev 08)
        Subsystem: Atto Technology Unknown device 802f
        Flags: bus master, 66MHz, medium devsel, latency 64, IRQ 17
        I/O ports at 5000 [size=256]
        Memory at fdfc0000 (64-bit, non-prefetchable) [size=256K]
        Memory at fdf80000 (64-bit, non-prefetchable) [size=256K]
        Capabilities: [50] Power Management version 2
        Capabilities: [58] Message Signalled Interrupts: Mask- 64bit+ Queue=0/0 Enable-
	Capabilities: [68] PCI-X non-bridge device
	Kernel driver in use: mptspi
	Kernel modules: mptspi
Signature · 4 lines
Ralf Hildebrandt (i.A. des IT-Zentrums)         [email protected]
Charite - Universitätsmedizin Berlin            Tel.  +49 (0)30-450 570-155
Gemeinsame Einrichtung von FU- und HU-Berlin    Fax.  +49 (0)30-450 570-962
IT-Zentrum Standort CBF                 send no mail to [email protected]
#7 lore raw

* Ralf Hildebrandt <[email protected]>:

Quoted 14 lines

Well, something seems to work:

[    3.571216] Fusion MPT base driver 3.04.06
[    3.571274] Copyright (c) 1999-2007 LSI Corporation
[    3.577292] Fusion MPT SPI Host driver 3.04.06
[    3.577389] ACPI: PCI Interrupt 0000:14:04.0[A] -> GSI 17 (level, low) -> IRQ 17
[    3.577512] mptbase: ioc0: Initiating bringup
[    3.836271] ioc0: LSI53C1030 C0: Capabilities={Initiator,Target}
[    3.992277] scsi0 : ioc0: LSI53C1030 C0, FwRev=01032700h, Ports=1, MaxQ=255, IRQ=17
[    3.992277] ACPI: PCI Interrupt 0000:14:04.1[B] -> GSI 18 (level, low) -> IRQ 18
[    3.992277] mptbase: ioc1: Initiating bringup
[    4.244297] ioc1: LSI53C1030 C0: Capabilities={Initiator,Target}
[    4.408304] scsi1 : ioc1: LSI53C1030 C0, FwRev=01032700h, Ports=1, MaxQ=255, IRQ=18
Quoted 12 lines
# lspci -v -s 14:04.0
14:04.0 SCSI storage controller: Atto Technology Ultra320 SCSI Host Adapter (rev 08)
        Subsystem: Atto Technology Unknown device 802f
        Flags: bus master, 66MHz, medium devsel, latency 64, IRQ 17
        I/O ports at 5000 [size=256]
        Memory at fdfc0000 (64-bit, non-prefetchable) [size=256K]
        Memory at fdf80000 (64-bit, non-prefetchable) [size=256K]
        Capabilities: [50] Power Management version 2
        Capabilities: [58] Message Signalled Interrupts: Mask- 64bit+ Queue=0/0 Enable-
	Capabilities: [68] PCI-X non-bridge device
	Kernel driver in use: mptspi
	Kernel modules: mptspi

According to http://lwn.net/Articles/254394, Eric Moore added the ATTO 4LD support:

mpt fusion: Add support for ATTO 4LD: Rebranded LSI

And the HP U320e SCSI Bus Adapter seesm to be a rebranded ATTO 4LD. But I cannot be sure.

Signature · 4 lines
Ralf Hildebrandt (i.A. des IT-Zentrums)         [email protected]
Charite - Universitätsmedizin Berlin            Tel.  +49 (0)30-450 570-155
Gemeinsame Einrichtung von FU- und HU-Berlin    Fax.  +49 (0)30-450 570-962
IT-Zentrum Standort CBF                 send no mail to [email protected]
#8 lore raw
Quoted 35 lines · Ralf Hildebrandt

* Ralf Hildebrandt <[email protected]>:

Quoted 6 lines

Well, something seems to work:

[    3.571216] Fusion MPT base driver 3.04.06
[    3.571274] Copyright (c) 1999-2007 LSI Corporation
[    3.577292] Fusion MPT SPI Host driver 3.04.06
[    3.577389] ACPI: PCI Interrupt 0000:14:04.0[A] -> GSI 

17 (level, low) -> IRQ 17

Quoted 3 lines
[    3.577512] mptbase: ioc0: Initiating bringup
[    3.836271] ioc0: LSI53C1030 C0: Capabilities={Initiator,Target}
[    3.992277] scsi0 : ioc0: LSI53C1030 C0, 

FwRev=01032700h, Ports=1, MaxQ=255, IRQ=17

Quoted 1 line
[    3.992277] ACPI: PCI Interrupt 0000:14:04.1[B] -> GSI 

18 (level, low) -> IRQ 18

Quoted 3 lines
[    3.992277] mptbase: ioc1: Initiating bringup
[    4.244297] ioc1: LSI53C1030 C0: Capabilities={Initiator,Target}
[    4.408304] scsi1 : ioc1: LSI53C1030 C0, 

FwRev=01032700h, Ports=1, MaxQ=255, IRQ=18

Quoted 3 lines
# lspci -v -s 14:04.0
14:04.0 SCSI storage controller: Atto Technology Ultra320 

SCSI Host Adapter (rev 08)

Quoted 7 lines

Subsystem: Atto Technology Unknown device 802f
Flags: bus master, 66MHz, medium devsel, latency 64, IRQ 17 I/O ports at 5000 [size=256]
Memory at fdfc0000 (64-bit, non-prefetchable) [size=256K] Memory at fdf80000 (64-bit, non-prefetchable) [size=256K]
Capabilities: [50] Power Management version 2
Capabilities: [58] Message Signalled Interrupts:

Mask- 64bit+ Queue=0/0 Enable-

Quoted 3 lines

Capabilities: [68] PCI-X non-bridge device Kernel driver in use: mptspi Kernel modules: mptspi

Quoted 8 lines

According to http://lwn.net/Articles/254394, Eric Moore added the ATTO 4LD support:

mpt fusion: Add support for ATTO 4LD: Rebranded LSI

And the HP U320e SCSI Bus Adapter seesm to be a rebranded ATTO 4LD. But I cannot be sure.

Yeah, ATTO support was added about a year ago, thanks to Sven Rudolph. However I doubt HP card is *not* a rebranded ATTO card. They both have the same scsi controller, but the bios and firmware on the HP card are directly from LSI, however I'm pretty sure that ATTO did everything them selves.

Eric

#9 lore raw

* Moore, Eric <[email protected]>:

Quoted 9 lines
Quoted 8 lines

According to http://lwn.net/Articles/254394, Eric Moore added the ATTO 4LD support:

mpt fusion: Add support for ATTO 4LD: Rebranded LSI

And the HP U320e SCSI Bus Adapter seesm to be a rebranded ATTO 4LD. But I cannot be sure.

Quoted 5 lines

Yeah, ATTO support was added about a year ago, thanks to Sven Rudolph. However I doubt HP card is *not* a rebranded ATTO card. They both have the same scsi controller, but the bios and firmware on the HP card are directly from LSI, however I'm pretty sure that ATTO did everything themselves.

You mean it probably won't work at all with the drive mpt spi driver?

Signature · 4 lines
Ralf Hildebrandt (i.A. des IT-Zentrums)         [email protected]
Charite - Universitätsmedizin Berlin            Tel.  +49 (0)30-450 570-155
Gemeinsame Einrichtung von FU- und HU-Berlin    Fax.  +49 (0)30-450 570-962
IT-Zentrum Standort CBF                 send no mail to [email protected]
#10 lore raw
In reply to Moore, Eric
Quoted 45 lines · Eric
Quoted 35 lines · Ralf Hildebrandt

* Ralf Hildebrandt <[email protected]>:

Quoted 6 lines

Well, something seems to work:

[    3.571216] Fusion MPT base driver 3.04.06
[    3.571274] Copyright (c) 1999-2007 LSI Corporation
[    3.577292] Fusion MPT SPI Host driver 3.04.06
[    3.577389] ACPI: PCI Interrupt 0000:14:04.0[A] -> GSI 

17 (level, low) -> IRQ 17

Quoted 3 lines
[    3.577512] mptbase: ioc0: Initiating bringup
[    3.836271] ioc0: LSI53C1030 C0: Capabilities={Initiator,Target}
[    3.992277] scsi0 : ioc0: LSI53C1030 C0, 

FwRev=01032700h, Ports=1, MaxQ=255, IRQ=17

Quoted 1 line
[    3.992277] ACPI: PCI Interrupt 0000:14:04.1[B] -> GSI 

18 (level, low) -> IRQ 18

Quoted 3 lines
[    3.992277] mptbase: ioc1: Initiating bringup
[    4.244297] ioc1: LSI53C1030 C0: Capabilities={Initiator,Target}
[    4.408304] scsi1 : ioc1: LSI53C1030 C0, 

FwRev=01032700h, Ports=1, MaxQ=255, IRQ=18

Quoted 3 lines
# lspci -v -s 14:04.0
14:04.0 SCSI storage controller: Atto Technology Ultra320 

SCSI Host Adapter (rev 08)

Quoted 7 lines

Subsystem: Atto Technology Unknown device 802f
Flags: bus master, 66MHz, medium devsel, latency 64, IRQ 17 I/O ports at 5000 [size=256]
Memory at fdfc0000 (64-bit, non-prefetchable) [size=256K] Memory at fdf80000 (64-bit, non-prefetchable) [size=256K]
Capabilities: [50] Power Management version 2
Capabilities: [58] Message Signalled Interrupts:

Mask- 64bit+ Queue=0/0 Enable-

Quoted 3 lines

Capabilities: [68] PCI-X non-bridge device Kernel driver in use: mptspi Kernel modules: mptspi

Quoted 8 lines

According to http://lwn.net/Articles/254394, Eric Moore added the ATTO 4LD support:

mpt fusion: Add support for ATTO 4LD: Rebranded LSI

And the HP U320e SCSI Bus Adapter seesm to be a rebranded ATTO 4LD. But I cannot be sure.

Quoted 5 lines

Yeah, ATTO support was added about a year ago, thanks to Sven Rudolph. However I doubt HP card is *not* a rebranded ATTO card. They both have the same scsi controller, but the bios and firmware on the HP card are directly from LSI, however I'm pretty sure that ATTO did everything them selves.

HP did two completely different PCI-Express Low-Profile SCSI HBAs:

The single-channel (HP SC11 Xe) one looks very similiar to the LSI20320IE, but the HP version appeared months before the LSI version.

The dual-channel version (AH627A; see
<http://www.hardware.com/product.asp?id=AH627A>, other Web sites show a wrong photo with a PCI-X connector) seems to be an Atto UL5D Low-profile.

Ralf's card looks a lot like my Atto UL5D Lowprofile. For the first SCSI channel:

  # lspci -v -s 0d:04.0
0d:04.0 SCSI storage controller: Atto Technology Ultra320 SCSI Host Adapter (rev 08)
        Subsystem: Atto Technology Unknown device 802f
        Flags: bus master, 66MHz, medium devsel, latency 128, IRQ 17
        I/O ports at 6000 [size=256]
        Memory at fbc40000 (64-bit, non-prefetchable) [size=256K]
        Memory at fbc00000 (64-bit, non-prefetchable) [size=256K]
        [virtual] Expansion ROM at c2800000 [disabled] [size=1M]
        Capabilities: [50] Power Management version 2
        Capabilities: [58] Message Signalled Interrupts: Mask- 64bit+ Queue=0/0 Enable-
        Capabilities: [68] PCI-X non-bridge device

I just tested mine with Linux 2.6.24.4, and it finds the attached DLT drives and they work. I missed Ralf's original problem but the output above suggests that it didn't find any devices?

Sven

#11 lore raw
In reply to Ralf Hildebrandt

Hi Ralf,

* Ralf Hildebrandt <[email protected]>:

Quoted 8 lines

According to http://lwn.net/Articles/254394, Eric Moore added the ATTO 4LD support:

mpt fusion: Add support for ATTO 4LD: Rebranded LSI

And the HP U320e SCSI Bus Adapter seesm to be a rebranded ATTO 4LD. But I cannot be sure.

I don't know anything about this card, but I can try to track down some help from other groups inside HP.

What HP server do you have? What HP part number is the card?

Thanks.

/ac

#12 lore raw
In reply to Mr. James W. Laferriere

* Mr. James W. Laferriere <[email protected]>:

Quoted 4 lines

Hello Ralf , Fyi , the fusion/mpt driver supports some of the ATTO
devices , if your lspci -v for the ATTO controllers looks simular to this one it may just work with the fusion drivers .
Hth , JimL

Yes, it just works :)

Signature · 4 lines
Ralf Hildebrandt (i.A. des IT-Zentrums)         [email protected]
Charite - Universitätsmedizin Berlin            Tel.  +49 (0)30-450 570-155
Gemeinsame Einrichtung von FU- und HU-Berlin    Fax.  +49 (0)30-450 570-962
IT-Zentrum Standort CBF                 send no mail to [email protected]