[ILUG] Highpoint Sata RAID

[Rick Moen]

Quoting John Coleman (john.coleman at gmail.com):

> The LSI Sata Megaraid 150-6 and Intel SRCS16 are exactly the same
> card. The PCI Vendor and Device IDs might be different, but they
> should both still work with the Megaraid driver (The LSI definately
> does). They are identical down to the BIOS and firmware versioning on
> the cards.

Thank you, John.  I appreciate the information, very much.

The disturbing part is that I _still_ am not clear on how I could have
found this out in any straightforward fashion from Intel's 
informational pages on their card
(http://www.intel.com/design/servers/raid/srcs16/).  I can't find any
mention of LSI Logic or Megaraid on that set of page -- any more than I
could find mention of ICP Vortex on pages about the Intel SRCS14L.

Searching the support.intel.com download pages for SRCS16 via its own
search engine comes up dry -- but doing an end-run using Google finds 
ftp://aiedownload.intel.com/df-support/7624/ENG/RH_DRV_readme.txt, which 
confirms what you say:

  This download supports Intel(R) SCSI RAID controllers using software
  stack 2 (the SRCU42X, SRCU42E, SRCZCRX, SROMBU42E, & SRCS16) [...]
  The following files are included in this zip: [...]
  megaraid2-v2.10.8.1-rh90-kernels.img

Doing likewise for the SRCS14L finds
ftp://aiedownload.intel.com/df-support/8257/ENG/GDTH%20readme.txt :

  The file gdth.tgz is a tar-zipped file containg driver source code
  taken from the Linux kernel. [...]

Doing the same for SRCS28X finds only a BIOS-reflashing utility and a
Java-based management utility -- no Linux driver information.  However,
checking instead inside the Linux pci.ids database
(http://pciids.sourceforge.net/) reveals that it's (like the SRCS16) a
MegaRAID-based card.

I've corrected my SATA Linux page, accordingly.

> The LSI 150-4 is just the 150-6 board with 2 Sil3112A chips driving 2
> sata ports each.  The Intel SRCS14L is not directly related (from what
> I can see from visual inspection), though it does seem to share much
> the same hardware, even if the boards differ in layout.

As noted above, they appear to present very different programming
interfaces to the PCI bus.  All of this seems to have implications for
many of the other new entries to my page -- where I've noted that
such-and-such new card is said to use Marvell 88SX60xx chips, implying
that one might fruitfully try Jeff Garzik's new Marvell driver.  But, if 
the situation is like that of the Intel card's use of SiI 3112A chips,
that would be non-sequitur.