EIO AP-1680 ATA133 RAID PCI Controller

I was under the impression that these EIDE RAID cards were "hardware RAID" as I see it this seems to just be a 2-Channel IDE controller That I will have to set up as MD (Software RAID).

Has anyone got one of these working as a hardware RAID controller under LINUX ?

Does anyone have any hints you can give me to get this working without using "Software RAID" ?

Here are the details of this controller:

EIO (Extreme IO) is a division of InnoVISION (www.ivmm.com)

They have a ATA RAID solution: "EIO AP-1680 ATA133 RAID PCI Controller" It can be viewed here: http://www.ivmm.com/eio/products_ap1680.html

Which they claim is LINUX compatible

It is based on Silicon Image's "Sil 0680 Ultra ATA133" chipset.

Silicon Image "Sil 0680" chipset link is:
http://siimage.cubik.com/products/sii0680.asp

The Driver for this chipset is in the kernel source:
CONFIG_BLK_DEV_CMD64X and CONFIG_BLK_DEV_CMD680

Output from lspci -vvv is this:
---------------------------------------------------------------------------
00:09.0 RAID bus controller: CMD Technology Inc: Unknown device 0680 (rev 02)
Subsystem: Unknown device 1771:1680
Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr-
Stepping- SERR- FastB2B-
Status: Cap+ 66Mhz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort-
<TAbort- <MAbort- >SERR- <PERR-
Latency: 32, cache line size 01
Interrupt: pin A routed to IRQ 12
Region 0: I/O ports at a000 [size=8]
Region 1: I/O ports at a400 [size=4]
Region 2: I/O ports at a800 [size=8]
Region 3: I/O ports at ac00 [size=4]
Region 4: I/O ports at b000 [size=16]
Region 5: Memory at df001000 (32-bit, non-prefetchable) [size=256] Expansion ROM at <unassigned> [disabled] [size=512K]
Capabilities: [60] Power Management version 2
Flags: PMEClk- DSI+ D1+ D2+ AuxCurrent=0mA
PME(D0-,D1-,D2-,D3hot-,D3cold-)
Status: D0 PME-Enable- DSel=0 DScale=2 PME-
-----------------------------------------------------------------

I have attached two(2) 40GB drives to this card one on each ide channel both set as Master. I went into the controllers BIOS and made configuration changes to reflect RAID 1 and built the set adding both drives to it. Everything looks fine a boot time. The controllers BIOS states two drives found combined as RAID 1 as 0x80. Under Window 98 I only see one drive. Which is what I would expect.

But after compiling these drivers into my LINUX kernel the system sees my drives as 2 separate disks (I have an ATAPI CDROM drive on Motherboard's 2nd-IDE channel) Here are my messages at boot time:
----------------------------------------------------------
VP_IDE: IDE controller on PCI bus 00 dev 39
VP_IDE: chipset revision 6
VP_IDE: not 100% native mode: will probe irqs later
VP_IDE: VIA vt82c686b (rev 40) IDE UDMA100 controller on pci00:07.1
ide0: BM-DMA at 0x9000-0x9007, BIOS settings: hda:pio, hdb:pio
ide1: BM-DMA at 0x9008-0x900f, BIOS settings: hdc:pio, hdd:DMA
PCI: Found IRQ 12 for device 00:09.0
CMD680: chipset revision 2
CMD680: not 100% native mode: will probe irqs later
ide2: BM-DMA at 0xb000-0xb007, BIOS settings: hde:pio, hdf:pio
ide3: BM-DMA at 0xb008-0xb00f, BIOS settings: hdg:pio, hdh:pio
hdd: CD-ROM 56X/AKH, ATAPI CD/DVD-ROM drive
hde: WDC WD400AB-00BVA0, ATA DISK drive
hdg: WDC WD400AB-00BVA0, ATA DISK drive
ide1 at 0x170-0x177,0x376 on irq 15
ide2 at 0xa000-0xa007,0xa402 on irq 12
ide3 at 0xa800-0xa807,0xac02 on irq 12
blk: queue c02c766c, I/O limit 4095Mb (mask 0xffffffff)
hde: 78165360 sectors (40021 MB) w/2048KiB Cache, CHS=77545/16/63, UDMA(100)
blk: queue c02c79d0, I/O limit 4095Mb (mask 0xffffffff)
hdg: 78165360 sectors (40021 MB) w/2048KiB Cache, CHS=77545/16/63, UDMA(100)
Partition check:
hde: [PTBL] [4865/255/63] hde1 hde2 hde3
hdg: [PTBL] [4865/255/63] hdg1 hdg2 hdg3
--------------------------------------------------------------------

Thanks

Peter

Tk,

Most of these "hardware" raid cards do not have an I/O processor, which (which is the definition of "hardware" raid). They have a BIOS which translates and allows the RAID to be booted into the target OS. Once in the target OS, their specific driver takes over the RAID task (so RAID is part of the driver instead of some higher level algorithm).

You might try the ataraid drivers and see if your's is supported. The only advantage you get with these types of cards is that you can boot multiple OS'es with the same layout on the disk.

Thanks
-steve

Thanks for the reply.

I thought that $19 delivered was too good to be true!!!

Are HP and Promise cards ture "HARDWARE" cards with there own I/O processor?

Peter

I don't know the specifics of those cards, but if they include an i/o processor, they are a hardware RAID.

Keep in mind that unless you want to boot or use a RAID 5, or want to share your RAID volumes between different operating systems, there is little reason to purchase a hardware RAID card. Software RAID is quite sufficient for most needs and the host processor running at 2+ ghz is much better at RAID then a 60 mhz i960 IOP. Keep in mind, though, that hardware RAID adaptors have built in xor accelerators which can do multiple xors in one instruction in hardware, allowing for much better xor performance then the intel or ppc processor can do.

Thanks
-steve

But what is the CPU usage per MB/sec? If software RAID5 was just as efficient as hardware RAID5, except twice as fast, you would expect the CPU usage to be double.

While the performance of RAID 5 using software RAID is generally far superior to that provided by hardware RAID, host CPU usage can be -quite- high in high-load situations. The benchmarks recently posted had cpu usage numbers at over 90% on most sequeential writes and reads.
Greg

-----BEGIN PGP MESSAGE-----
Version: PGPfreeware 7.0.3 for non-commercial use <http://www.pgp.com>

qANQR1DBwE4DpUAjANw7zLUQA/4ywsdT3nN/LPnpWKA4DXIrRXJXf34yW0CXLdKu MOZZTn5YxZgZWnKwP35j6kQS8kZM7RuwvCAYrkIt2BE3v9tNTdVM3/BZ9H+Im5XB eqcekJsx0Hv48EIOBIf+VKvvoIitaGUAn7alj/Gr1On6dfLkJlHLxeEYMd6nETsL EaVCqgQAo/EPu+aWJBTvOWMPnZ69HzfhWOoQzzh7QWi7MAWkfdPQtzhh1CYB2bUW CmauosibyUkRLXAFUR/zkVqM9oQz7ANHYG00qlVsH2+H+C/vsDylmzkKIr/V2crZ kS6wm4zcXlnm8oYNO+5mpvdnTa6SVXCDK8plPs1dmQzkLP76xFXBwU4DmGQ29uee dsAQB/4si6HCYAbKBG39H4sZ8/qrADvhJVtAZntzSJBJlvHLDTDG5K6Uk+O0ei9h S34wSKIXlqNPUTlE7s8z2b8o4Tgzgj9pP6cTj3wixIAH+DqW+l+sZukRDCkWlOH9 7e+yYhyv4Lw4+vuPHqQa2S0pYy5TFO4kYrWWrAI/IVLVTk4g0YpmWSY/TtoqsLvc IJAWSi/3M9LeG8jEnvQ6adUKZBt1usZlZcP8vPaZFLFWz0OXCNkOefIMKt8zEn0w g+JPOieCpqvthtOKsCFqmwBujD5IESpK6DSTRLWKF5gwnpdMDBQbIjLq8g37D5WJ Per6MGb+TqWEuIKRHRwbSyHppdKPB/9Bke1byB3LLr367Mn3aMYTBUYbCY+1vSJy ED20c09ERBuW+jE+ZHxaUNdQv+Er/77YXEyvMdnlMyeUDOojDvporW8ohTNOWWw+ ZPSWG//i07UvU7e3zBlO4oEo+CIJtwBuKQb3oyL1/7dX6tjUDlNEQURgo/QZYZg4 SMM/re7BHH1YOgzQklignCMbaTYmlSjNDi6nAFsTFtH0OxepJNV5W4QgaTz75JJ+ W9eWITaZ8QZUILtTsFbJVZEeElOA5J0RMQ8ljvZEtick6xwaNOQn6hiRXkHO9pqU 1ydw4FgChnCB6h2F1KndYgKPPeYsgowXAHmyIz6Bakt572Suml/dycFO4kzuzGV9 3kdEMPIqdiegNbGMqY94gVwwp2BtuvxNsl+Dasdk6uFeiQtQSvRQ2ra2tfCUiwgN D5zaUEuTjFa54BNs8hZ3OugpC1+xTjFNj4vUzT1rYE+A9qkRKXb/STvmyvoNh8yq 1g857geLE+ycTkqR9A59hnpm25CZWGiOUN6ph3YvfpSxNMrk/vBYTHep/B8iD3io 3ow5E0vCCwwbpjhDuKFa5Xbi9PrHzouzC1LR+4m1eMRUtQv4Xd8JGmA7sNR+UINO 2pG/1N7drzzGe5N8Sq/6jdxgh+/SNUKI0gIXQVvogMZuQcIgi0gnIPtNJyrCSmGY zBi2CmXXEvup0g/1Ty1rWs8cpDUJan9JJnXCLVXVu1hwULTmrMFEiC+zH4ZnDEq7 tPRZalLpanGE+5AOX9X/0yIjQpEee80p46mzNgQAyb+YNSC2NQzCM0OhMTLsc5Bv 6zAlX1B1lltnlrs6AEtcGgg5J1jJ3vovo2FKn2uLoFoqz3Ai4e0nCeV/7lJHB3tu BxqObuuYQ8pvWLRBMOKrMMCPPW7jACmeM8RL2AzFgFKmizKZwW+g0u41oR1Sbjx9 34mlS7UyjSnc7fD36kNtwN6Y3TOEtsBd6r5yoiDEKtp23FiDUd27/1HuaMLJ0z8j ICzYJkRbZ5BZMojkJUL6MTWKInVSo07KZetlh5q5YCntXPDrGYAARQ== =2bn1
-----END PGP MESSAGE-----

With our best regards,

Maurice W. Hilarius Telephone: 01-780-456-9771
Hard Data Ltd. FAX: 01-780-456-9772
11060 - 166 Avenue mailto:[email protected]
Edmonton, AB, Canada http://www.harddata.com/
T5X 1Y3

It's really impolite to send encrypted mail to the list. Anyone to whom the message is encrypted won't be able to read it

[snip]

The formatting of this message is all wonky, sorry about it being confusing. Anyway, I agree, except that I have no numbers to back up the CPU use differences, or the MB/sec differences. Without them, any claims are pretty silly to try to make.
Greg