Two-channel ide controllers + 3rd ide drive?

Two-channel ide controllers + 3rd ide drive?

Post by The Linux Advoca » Tue, 21 Mar 1995 02:04:58

I'm going to be getting a 3rd ide drive in about a week, and I would like
to know if the dual-channel ide controllers are supported. (ie, those that
have two masters and two slaves.)  Also, if they are indeed supported, are
ALL of them supported, or just some?  The one I'm looking at getting is a cheap
no-name clone one.

Lastly, how do I make the /hdc devices?  I know how to use mknod, but I don't
konw how to determine the major and minor numbers of the 3rd drive.

(Please respond via email, as our news disk tends to overflow, and news posts
are only kept for a day.)


1. atdisk patch with a two-channel IDE controller

Has anyone had any experience getting the atdisk patch to work with one of those
generic IDE controllers that have two IDE channels on board? I already determined,
by tracing the lines on my "S1345 VESA LOCAL BUS VL IDE & I/O Controller" that the
second IDE channel uses IRQ 15, and as far as I know the controller is at the
"standard" second IDE ports (170-177,370-377).  Nevertheless the atdisk patch
does not find the single HD that I've attached to the second IDE channel.  The
kernel goes into an infinite loop on startup, trying to reset the controller
and then timing out.  (I tried #define-ing HD1_DontReset; no luck.)

Unfortunately the one page of "documentation" that comes with this
no-name IDE controller is too sketchy to be of much help: it notes that
"most S1345 installations only turn on Primary IDE channel at power up.  To
use secondary channel, you need to run the supplied driver.  Contact your
dealer to configure both channels on at power up." There is, of course,
no "supplied driver" and the "dealer" knows nothing about it.  I suspect, from
this sentence, that the second IDE channel  on this board is normally disabled
upon powerup and Linux's kernel initialization doesn't enable it.  Presumably
either a trace must be cut, or something soldered, somewhere on the board (I've
accounted for all the "real" jumpers and even blindly cut and then resoldered the
undocumented ones, with no success), or else some software command must be sent
during kernel initialization to the board's logic in order to enable the channel --
I'd be able to test that under MS-DOS, I suppose, if I had the "supplied driver,"
but I don't.

If anyone has any suggestions (or a "supplied driver" for this kind of board!) I'd
be grateful to hear.  I suspect that this might be of interest to others as well,
since the VL-bus multi-IO controller with two IDE channels seems to be fairly
'standard issue' nowadays -- I've seen it in several new PCs.  It would be nice
to get it working under Linux.

Adam Jacobs


2. Linux and Teles-card

3. Added a second drive to second IDE channel.

4. Memory tester for linux somewhere?

5. 2.5.69-mm9, CMD649 IDE controller 2nd IDE channel doesn't work

6. Red Hat 5.2 and AMD

7. Help Mounting 3rd IDE drive on a single KT410A EIDE controller

8. PATCH: usb-uhci: interrupt out with urb->interval 0 [linux-usb-devel]

9. 3rd IDE hard drive setup problems & Controller recommendations

10. Detection of IDE disk at second IDE controller

11. EIDE drive on second IDE channel not found by fdisk

12. Linux on 2nd IDE drive on 2nd IDE controller?

13. Tested: Journaling Filesystems on Linux software RAID-5 using IDE-drives and 2 IDE controller cards