HPT372N Re: hpt366 driver oops or panic with HighPoint RocketRAID 1520SATA
Hello,
I recently bought
Highpoint RocketRaid 1520 pci - dual channel - sata cards - chip is HPT372N -
to use as simple block devices for linux software raid purposes. I thought
they were supported on Linux, because of highpoint web site and my previous experience with highpoint pata controllers was ok.
However I get the same kernel OOPS reported by Johan Palmqvist on Mon Nov 07
2005 with the latest stable kernel 2.6.14.4 .
I also see a posting from Mar 17 2003 by Henning Schroeder reporting success
with the highpoint driver hpt3xx-opensource-v131.tgz.
I tried to compile the hpt3xx-opensource-v2.0.tgz against latest stable
kernel 2.6.14.4. After minor corrections I have failure to compile their
driver due to scsi_cmnd structure "has no member abort_reason". Has their been a change in scsi subsystem?
Which kernel can I use? Any advice? Sorry to interrupt your amazing work with
my question. I am off list. I will read list but cc is appreciated.
Thanks Gurus!
Mitchell Laks
Quoted 3 lines · Mitchell Laks
to use as simple block devices for linux software raid purposes. I thought
they were supported on Linux, because of highpoint web site and my previous experience with highpoint pata controllers was ok.
Some are, although there are problems remaining in things like clock timing and PLL stabilization. In addition the ide/pci/hpt366.c driver never supported the SATA bridge variants of the chipsets at all. These require certain modes are used.
What may work if the HPT card has the onboard BIOS enabled and fitted is to compile in the generic IDE PCI driver, say N to the HPT driver and boot with the option "all-generic-ide".
I've been working on a new HPT372N/302N driver for libata/SATA but it isn't yet production ready and it also does not know about the SATA bridge rules for that chip.
Quoted 4 lines
I tried to compile the hpt3xx-opensource-v2.0.tgz against latest stable
kernel 2.6.14.4. After minor corrections I have failure to compile their
driver due toscsi_cmndstructure "has no memberabort_reason". Has their been a change in scsi subsystem?
Several
Alan