softload 5.23 and CDFSSoftATAPI

I am unable to get a softload of 5.23 to complete on my Iyonix.
It aborts while trying to initialize CDFSSoftATAPI.
“init mod CDFSSoftATAPI Error abort.. FC41E7E8..”
The abort address is consistent, but is slightly different for another
5.23 dev ROM – eg FC41FA08.
(I have tried 5.23 dev ROMs of 15 December 2015 and 22 February 2016)

My machine is an Iyonix the flashed ROMs are on 5.22 stable.
!!ZeroPain is installed in PreDesk.
It may or may not be relevant but I have a CDROM drive and a DVD r/w drive in the machine.
I would appreciate any suggestion as to how to proceed.

Further information. The ROMs from 4 July 2015 load successfully. The CDFSSoftATAPI module level is 1.33; the module in the ROMs that wont load is 1.34.

Problem solved. One of the CD drives was not working but was
configured. When I configured 1 drive instead of two the latest 5.23
loads successfully.

Actually, I was about to ask if you could try running the following bit of BASIC (shouldn’t matter which OS version or CD config you use):

FOR A=0 TO 3 : SYS "ADFS_IDEDeviceInfo",0,A TO ,T,N,P : PRINT A,T,N,P : NEXT

It looks like there’s a bug in CDFSSoftATAPI where it will crash if it finds an ATAPI device which didn’t respond to the identify command that ADFS sent it (number in second column will be 2, fourth column will be 0). Which would certainly fit with what you’re saying about one of the drives not working.

If there is an entry that matches that, then this looks like a bug that’s been around for a while – you’ll only have noticed it recently because of the zero page relocation that was introduced in July (and the bug is during ROM init, when ZeroPain won’t have been loaded yet).

FOR A=0 TO 3 : SYS “ADFS_IDEDeviceInfo”,0,A TO ,T,N,P : PRINT A,T,N,P : NEXT

As you predicted:-
Program renumbered
0 1 4 537060292
1 1 5 537060804
2 2 126 537061316
3 2 126 0
*

Should I report this as a bug? and if so how?
For the moment I can live with having just one CD/DVD drive configured and will softload 5.23 to test for ZeroPain problems in my setup.

It’s OK, I should be able to check in a fix for the crash tonight.

The fix is now checked in, so expect to see it in tomorrow’s ROM. The version number will be the same (1.34), but *Help CDFSSoftATAPI or *Help Modules should show “1.34 (23 Feb 2016) 1.11.2.1” (this is just a quirk of the way branches interact with version numbers; the main branch now contains a newer version of the module which won’t build without the SATA driver sources being available, so there’s a temporary branch to allow a pre-SATA version to be built for Iyonix/RiscPC).