Building modules
Rob Coleman (489) 12 posts |
I’ve recently setup the DDE on my Pandaboard and, having downloaded the latest OMAP4 sources, was able to successfully build a ROM image. However, I’m not sure how to build individual modules to soft-load onto my system. I’d like to modify CDFSSoftSCSI as the current version doesn’t work with my USB CD-ROM drive – I think I know why but want to try various changes to the existing code. I’ve tried running the relevant !MkRAM and !MkRAMDebug files but I just end up with a hanging call to the InstViaRG Perl script when amu_machine is invoked. (Perl$Dir has been set and Build: seems to be referencing the correct directory.) Can anyone work out what I’m doing wrong and/or point me in the right direction? Thanks in advance, Rob |
Colin (478) 2433 posts |
Run !builder and set it up with the build directory of your Omap sources and environment ROOL.OMAP4- like you did when you compiled the sources. In the CDFSSoftSCSI directory use !MkClean then !MKRam the module will be in the rm directory. You only need to use !MkClean once. |
Colin (478) 2433 posts |
I should add you don’t need to modify the version in the rom source tree you can copy CDFSSoftSCSI elsewhere and modify the copy. |
Sprow (202) 1113 posts |
I’m curious to know if it previously worked, say with RISC OS 5.24, or whether it simply never worked in the past and now. If it used to work then I’ve a progression of 3 Pandaboard ROMs to try: number 1 number 2 number 3 and would love to hear if there’s any observable difference. A power cycle is best between each to make sure the CD-ROM drive gets reset. If it’s a drive that has never worked on any version, then I’ll leave you to your experiments – have fun! |
Rob Coleman (489) 12 posts |
Thanks both – I’ll try building it again tonight.
It has never worked – even though it’s been installed in my Panda system for years, I’ve lived without it. I tried using the debug version of CDFSSoftSCSI that was posted on here a while back and the drive seems to be upset by the attempt to set the inactivity timer multiplier. Everything before this works but everything afterwards fails – even re-running an inquiry command that previously worked. Looking at the standard, I think the drive may only accept requests to set fields that are changeable. From a brief look at the code, CDFSSoftSCSI doesn’t appear to ascertain which fields are changeable but instead changes them and accepts that this might cause an error. Unfortunately, my drive doesn’t seem to recover from the error… |
Rob Coleman (489) 12 posts |
Many thanks again – that worked. And my quick hack to remove anything that attempted to change a field has fixed the issue with my drive. So, I finally have the CDROM drive working with my Pandaboard. |
Doug Webb (190) 1129 posts |
Hi Rob, Rob Sprwoson was looking at an issue with my Pandaboard and the latest ROM’s where the Cd Drive would hang. I fixed that issue by changing the CD Rom from the 8 year old Sony one to a newer LG drive. Could you make your version of CDFSSoftSCSI available so I could see if it fixes the issue with my old drive? It most likely will not as it’s behaviour changed to totally not working once the firmware was updated but it would be good to see if it did. Thanks Doug |
Rob Coleman (489) 12 posts |
Hi Doug, I’ve put it on Dropbox . (It’s currently a quick and dirty hack – I should really have got it to find out which fields the drive permits changes to instead of just commenting out the code.) If you want me to talk you through the steps I used to debug my issue, just say. Best wishes, Rob |
Doug Webb (190) 1129 posts |
Thanks Rob I have downloaded it and will try it when I dig out my Pandaboard sometime hopefully on Thursday or Friday. |
Doug Webb (190) 1129 posts |
Hi Rob, I’ve tried your module on the Pandaboard with the old Sony drive and it still doesn’t work. I have also now tried the drive and USB/Sata bridge it is attached to in a Windows PC and it gives the same symptoms so it looks like the drive is faulty and most likley due to the stress of transit in the post that gave it a final push. The new LG drive works OK and as it is a spare machine for testing then thats OK. Thanks for making your fix available and great to hear you now have a working drive as well. Doug |