(unknown charset) Re: system lockup - ata spurious interrupts - somehow updated

Søren Schmidt writes:
| Doug Ambrisko wrote:
| > Soren Schmidt writes:
| > | That said, the ICH5 support is written "blindfolded" as I do not have
| > | any such HW here in the lab. So if there is a problem with ATA on those,
| > | someone with the HW and enough kernel clue should look into it (or
| > | someone could land the needed HW in my lab, which could also bring
| > | support for the SW RAID on intel's)...
| > 
| > I have some patches for ICH5 for the IDE and SATA parts of it based
| > on stable.  I haven't ported them to -current yet but I should be
| > starting to work on that soon.  I have patches for the 6300ESB ICH
| > as well.  They are basically the same SW wise except for pci ids.
| > 
| > There is nothing major in the patches except for PCI id's and
| > reading the DMA setup.  I do have tweaks that read the SATA status
| > registers and which drives are mapped to which ports.  This prevents
| > system lock ups when the SATA drives go away.
| > 
| > I just got my hands a system that looks like it might have the Intel
| > RAID in BIOS stuff to play with. 
| 
| If you get anything done let me know so we can get it reviewed and 
| possibly committed to -current. I dont have patches for the 6300 here 
| locally as well, thats just PCI ids...

Some of the issues I was seeing with the ICH5 in stable wasn't an
issue in -current due to your re-design.  Here's the simple ID changes 
for the 6300ESB.  I've tested with the SATA mode not the RAID to IDE
controller.  My test system is a Dell that is hard wired to do what
they want.  I have to hack the chipset during boot to make it 
go into the modes I want it to do :-(  The SATA/RAID thing is really 
just the same thing but flipping a bit on the chip switches the device ID.
everything else stays the same.  I need to build a wedge in something
to pre-configure the 6300 in the way that I want.  So far with my
initial testing on the ICH5 it just works in -current.

I have to deal with a "feature" of the Promise SATA card in my -stable
patches.  If it hits a media error is locks up the system.  Oh joy.
As soon as I figure out the Promise issue then I should be looking at 
-current to deal with SATA issues.

Index: ata-chipset.c
===================================================================
RCS file: /usr/local/cvsroot/freebsd/src/sys/dev/ata/ata-chipset.c,v
retrieving revision 1.67
diff -u -r1.67 ata-chipset.c
--- ata-chipset.c	17 Mar 2004 17:50:27 -0000	1.67
+++ ata-chipset.c	25 Mar 2004 18:55:35 -0000
_at__at_ -813,7 +813,10 _at__at_
      { ATA_I82801DB,   0, 0, 0x00, ATA_UDMA5, "Intel ICH4" },
      { ATA_I82801DB_1, 0, 0, 0x00, ATA_UDMA5, "Intel ICH4" },
      { ATA_I82801EB,   0, 0, 0x00, ATA_UDMA5, "Intel ICH5" },
-     { ATA_I82801EB_1, 0, 0, 0x00, ATA_SA150, "Intel ICH5" },
+     { ATA_I82801EB_SATA, 0, 0, 0x00, ATA_SA150, "Intel ICH5" },
+     { ATA_I6300ESB,   0, 0, 0x00, ATA_UDMA5, "Intel 6300ESB ICH" },
+     { ATA_I6300ESB_SATA, 0, 0, 0x00, ATA_SA150, "Intel 6300ESB ICH" },
+     { ATA_I6300ESB_SATA_RAID, 0, 0, 0x00, ATA_SA150, "Intel 6300ESB ICH RAID" },
      { 0, 0, 0, 0, 0, 0}};
     char buffer[64]; 
 
Index: ata-pci.h
===================================================================
RCS file: /usr/local/cvsroot/freebsd/src/sys/dev/ata/ata-pci.h,v
retrieving revision 1.28
diff -u -r1.28 ata-pci.h
--- ata-pci.h	16 Mar 2004 16:23:28 -0000	1.28
+++ ata-pci.h	25 Mar 2004 18:55:35 -0000
_at__at_ -118,7 +118,10 _at__at_
 #define ATA_I82801DB		0x24cb8086
 #define ATA_I82801DB_1		0x24ca8086
 #define ATA_I82801EB		0x24db8086
-#define ATA_I82801EB_1		0x24d18086
+#define ATA_I82801EB_SATA	0x24d18086
+#define ATA_I6300ESB		0x25a28086
+#define ATA_I6300ESB_SATA	0x25a38086
+#define ATA_I6300ESB_SATA_RAID	0x25b08086
 
 #define ATA_NATIONAL_ID		0x100b
 #define ATA_SC1100		0x0502100b