Lockup after loading cdfs.ifs in /APIC mode

[Chuck McKinnis]

Tonight I upgraded one of my D975XBX2 machines to ACPI 3.08. This one has no SCSI - it's a Core2Quad with 2GB RAM, a SATA 160GB disk attached to one of the onboard ICH7 SATA controllers, a PATA DVD=-RW drive on the IDE port and a Netcell SR5000 SATA RAID controller in one of the two PCI slots. ATI X300SE graphics with SNAP 3..1.8.

I upgraded from 3.06 to 3.08 and rebooted, Alt+F2 shows the last driver loaded is CDFS.IFS (level 14.103). Ctrl-Alt-Del does not work but Ctrl-Numlock-Numlock usually does.

I downgraded to 3.07 which does work.

DaniS506.add 1.8.1, daniatapi.flt 0.40, os2cdrom.dmd 10.163

Hmm, looks like PATA shares an IRQ with USB.

[Chuck McKinnis]

pci.exe output

[Chuck McKinnis]

rmview /irq output

[pasha]

can you remove os2logo, install acpi psd from debug dir, boot, when lockup - do photo and attach here?

[Chuck McKinnis]

Done. Last things to appear on screen are the messages UNMASK set as APIC UNMASK IRQ6 mask 0xFFFF94B9 Real: FFEBB8F4 just above the 'The current hard disk drive is: i:' line even though these are buried in the other messages on screen

[pasha]

Have you floopy drive? If no, remming all basedev=flop*

[Chuck McKinnis]

REMing ibm1flpy.add does allow it to get further. It loads udf.ifs then a bunch more drivers and last, amouse.sys. The last message this time is
UNMASK IRQ 12 mask 0xFFFF94B9 Real: FFEBB8F4
Amouse.sys is the latest 2.80 release and its basedev has the SMP parameter specified. This system works perfectly with 3.07 so something broke between then and now.

[pasha]

Remming amouse.sys or replace to drover mouse.sys from hobbes.nmsu.edu

[Chuck McKinnis]

Sorry, amouse.sys works too well for me to dump it. I'll just skip 3.08 until you can get this fixed in 3.09. This is something that's broken in the last release as amouse.sys has worked for me since before acpi 2..20

[pasha]

Ask this problem to autor of amouse.sys or use /PIC

[Chuck McKinnis]

e-co:

OK, we have sent all info to the developer of Amouse. He is alive, so let's wait for update..

[Chuck McKinnis]

I don't know how you can blame the author of amouse for this. The thing that changed was acpi.psd not amouse!

[Chuck McKinnis]

8)

don't worry, it's turn of the developer of AMouse now.

we can't keep crutches and temporary bugs for compatibility with other drivers and applications.

ACPI is improved, now let's lift the device drivers.

[Chuck McKinnis]

This has broken IBM1FLPY.ADD too and that developer is not active any more.

[Steven Levine]

To long since last feedback.