Subject: Re: [linux-audio-user] hdparm and SATA
From: Malcolm Baldridge (linux-audio_AT_paypc.com)
Date: Sun Jul 18 2004 - 02:13:30 EEST
> I have one Seagate Barracuda SATA drive that is seen by the system BIOS
> on the primary master IDE channel. I suspect different boards and BIOSs
> will handle SATA differently - some will handle it as some kind of SCSI,
> and some as some kind of IDE, but I could be wrong. I have used hdparm
> on my SATA drive as though it were an IDE, but it didn't change much.
Really, what I'm referring to are the "IDE madness" things about hdparm: the
X (chipset timings, which are very dangerous), the -m settings, the
bus-mastering DMA settings (-d1), etc.
IDE needed to have abandoned this whole Cylinder/Head/Sector fiction a long
long time ago, and you'd have thought with all of the re-invented schemes
and shifting "size barriers", the standards bodies would have just embraced
an LBA scheme which has served SCSI without a single incident of drama since
1985, but noooooo...
Anyway. The long and short of it is, with SATA interfaces and drives, the
bus mastering DMA/multi-sector/chipset timing stuff should be setup
appropriately without hand-waving. For modern Linux kernels built with
chipset-specific IDE drivers internally, it should also be setup optimally.
The chaos factor is for these wingnut distros which build "generic" kernels
and/or leave things to the whims of the gods and BIOS.
-- A focus on Quality.
This archive was generated by hypermail 2b28 : Sun Jul 18 2004 - 02:17:10 EEST