Subject: Re: Odd data faults on U5 with Promise U66 IDE controller
To: matthew green <mrg@eterna.com.au>
From: Rafal Boni <rafal@pobox.com>
List: tech-kern
Date: 09/09/2003 04:46:27
In message <4007.1062075182@splode.eterna.com.au>, Matt writes: 

-> i thought i had a promise U100 card in my U10 for most of this year,
-> but it turns out it's the el-cheapo ST labs one i have as well.  the
-> promise is in my mac.  i'm now reminded that the reason i swapped them
-> was because of one system being unstable, and while i don't remember
-> exactly, i think it was the promise controller that was failing so
-> it must have been failing in the U10.  the $50AU card has been working
-> fine since.... looks like march:

Thanks for the suggestion -- the card I picked up (also a SilImg 680)
cost me $20USB *including* $5USB shipping and is sold under the "Syba"
brand.

-> 	pciide1 at pci2 dev 3 function 0: Silicon Image 0680 (rev. 0x02)
-> 	pciide1: bus-master DMA support present
-> 	pciide1: primary channel configured to native-PCI mode
-> 	pciide1: using ivec 18 for native-PCI interrupt
-> 
-> so - it may be some promise-specific problem.  the card that was failing
-> probed like this:

Indeed in my case it appears to be.  I've now copied all the data from one
disk to the other a few times, set up RAIDFrame to mirror the two drives on
the SilImg680 controller and performed a few reconstructions (to initially
populate the second drive and then to test failure recovery by marking one
of the components as failed and then reconstructing).  The current recon-
struction is plodding away with vmstat claiming 25MB/s from both disks.

The issue with the Promise definitely seems to be specific to that card or
line of cards and it *does* look like some sort of bug in our drivers.
Anyway, the key is that I'n now up and running and happy.

Thanks for the hints,
--rafal

----
Rafal Boni                                                     rafal@pobox.com
  We are all worms.  But I do believe I am a glowworm.  -- Winston Churchill