RO5. on Red Squirrel
Colin Ferris (399) 1818 posts |
RO 5.19 running on Red Squirrel – trying to get HostFs etc to work. |
Theo Markettos (89) 919 posts |
It sounds like the podules are being detected OK. Have you any means of running the HostFS/etc modules from other media (for example, an ADFS disc image that you format and set up in Red Squirrel running RO3.7) to check that they work OK in RO5? What does *ROMModules say? The modules aren’t unplugged are they? Do you know if the HostFS/etc modules are 32 bit compatible themselves? |
Colin Ferris (399) 1818 posts |
Changed the ‘hardware_mask% from &03FFF000 to &FFFFFC00 and the modules are now showing up under *rom. |
Colin Ferris (399) 1818 posts |
HostFS filing system now seems to be working. |
Jeffrey Lee (213) 6048 posts |
What happens if you try increasing the size of the screen memory dynamic area in the task manager? If it’s definitely limited to 1MB then it sounds like the VRAM detection code isn’t working properly, or Red Squirrel isn’t emulating the IOMD VIDCR register properly. Overall the VRAM detection code should be the same as in 3.7, but due to the way the HAL/OS has been split I had to implement a workaround for the fact that the HAL has nowhere safe to store some VRAM-related variables which it calculates in the VRAM detection code. This workaround is to read back the value of the IOMD VIDCR register once the OS has started and use that to recalculate the variables. This works because the HAL/OS programs different values into VIDCR depending on the amount of VRAM you’ve got fitted. However I’m not sure if reading back the wrong value from VIDCR would result in the OS only being able to use 1MB of VRAM, so this might be a red herring. If the task manager does show that you’ve got 2MB of VRAM fitted, it might be that the VIDC bandwidth limit hasn’t been set yet. What boot sequence (if any) are you using? If you use the ROOL boot sequence then it should work (although there may be problems on real hardware). If you’re not using any !Boot yet then you can just try ‘*VIDCBandwidthLimit 38000000 76000000 152000000’ |
Colin Ferris (399) 1818 posts |
First RO3.7 Screen memory goes up to 2048K and the *VIDCbandwidth command doesn’t seem make any difference. ( No boot is being run with either OS’s) With RO5.19 Screen memory can go up to from 1400k to just over 2000k it varies. Using *VIDCbandwidth command – seems to have a strange effect – the colour being displayed – dropping down to 16 when modes are changed. Changed the *VIDCbandwidth to Iyonix values and I could change to higher colours again. When clicking on 32thou colours in 1024 × 768 the window changes *wimpmode X1024 Y768 C32K – Works. |
Colin Ferris (399) 1818 posts |
Bit missing -:When clicking on 32thou colours in 1024 × 768 the window changes |
Colin Ferris (399) 1818 posts |
Follow-up Modified RO5.19 SA Model – now I can get – 1024×768 C16M – but need to use *VIDCbandwidth 40000000000 4000000000 4000000000 then *wimpmode X1024 Y768 C16M (Frame rate is shown ‘Unknown’)If the mode selector is used – the colour or resolution changes to lower values. Press *F12 and *rom. rtn – the listing is correct – not splitting – half way down – to appear half way across the screen – as with the original SA Model. |
Colin Ferris (399) 1818 posts |
Made a 18Hz screen mode – I can now select 1024×768 C16M from the mode selector. |