On Tue, 5 Dec 2000, D. Hugh Redelmeier wrote:
> | From: Andre Hedrick <[EMAIL PROTECTED]>
>
> Thanks for responding. I'm very interested in what you post, but
> first I must say that it really doesn't relate to what I intended to
> say in my message.
Oh you missed my warning on the complete GOAT-SCREW of the PIRQ routing
tables some time back, I see.
> My message was warning about a design "feature" of the board that
> limits the total number of bus mastering boards in PCI slots 4 and 5
> to one. This was not explained at all in my manual, and needs to be
> made very clear to all who might try to sonfigure a BP6 system.
It is worse than you know or have explained.
AGP partially shared PCI 1 (only A, B)
PCI 2
PCI 3 shared w/ HPT366
PCI 4 shared PCI 5
> I think that your message shows that your IDE code makes the HPT366
> perform wonderfully. Even with shared interrupts. Even with 2.2.18
> (with your patches, I presume).
>
> It is interesting that hd[abc] are on the HPT366; I thought that these
> would normally be on the Intel 440BX chipset's IDE controler. Your
> timings don't include anything on the BX's controler, so we don't see
> how the HPT366 compares.
Trust that they are faster, but you will have the tools in hand soon to
verify. Also note that the numbers reported are about 35-40% slower than
reality because of the memcpy and context switching in the CR3's
> Now, I have some questions.
>
> - is there a "blessed" version of your IDE patches for 2.2.18?
>
> (My recollection was that you've handed off maintenance of the 2.2
> version of the patch to someone else so that you can focus on 2.4. I
> don't know how that is working out.)
There were two new chipsets and thus I needed to include them.
> - is there any chance that your IDE patches will make it into the
> standard 2.2.x kernel?
Do pigs fly? Get Alan Cox Drunk and then take advantage of him.
> (My impression was that AC has left the door open to this, but not yet
> invited you in :-)
>
> - While studying the mailing list to glean ideas about my problem, I
> came across a posting that raised questions in my mind about
> interrupt sharing. The message was posted by Eberhard Moenkeberg
> <[EMAIL PROTECTED]> on Tue, 8 Feb 2000 09:43:28 +0100 (MET).
> He suggested that SA_INTERRUPT|SA_SHIRQ should be replaced by
> just SA_SHIRQ in several drivers of 2.2.14:
> block/ide-probe.c
> char/bttv.c
> scsi/ncr53c8xx.c
> scsi/sym53c8xx.c
> Do you think that this is correct? I noticed that 2.2.16 (what I'm
> running in my RHL6.2) has no such change.
Reasonable, try it and be careful that you parse the legacy against
add-on.
> - The BP6 manual suggests that the HPT366 was designed for hard disks
> and recommends not using it for other ATA/ATAPI devices such as CD
> drives. Do you feel that this recommendation applies to LINUX (i.e.
> is it a feature of dubious BIOS or MSWindows drivers, or is it a
> limitation of the hardware)?
Exactly, this was my suggestion, they put this warning in the manual.
These new funny chipsets do ATAPI-DMA in goofy manners and I have not
deployed or gotten tanked (do not drink any more for 11 years) to go on a
vision quest. The about of caffine it takes to get into an alterstate
makes it impossible to use a keyboard. The reality is that it is nasty
and I have not created a safe method of bouncong the DMA-Engine locations
because the are different.
> Thanks for all you've done for LINUX IDE!
Cool! I mostly get flames.
Cheers,
Andre Hedrick
CTO Timpanogas Research Group
EVP Linux Development, TRG
Linux ATA Development
--
=- To unsubscribe, email [EMAIL PROTECTED] with the -=
=- body of "unsubscribe linux-abit". -=
