From: muta...@gmail.com   
      
   On Tuesday, October 26, 2021 at 4:56:11 AM UTC+11, JJ wrote:   
      
   > > Let me try again.   
   > >   
   > > Let's assume we have a real CDROM and a real hard disk   
   > > and a real BIOS, but I'm writing the real BIOS and I'm making   
   > > up my own specs for the BIOS, which OS vendors will need   
   > > to comply to.   
   > >   
   > > Hard disks and CDROMs will be less than 2 GiB in size.   
   > >   
   > > Rather than expose the OS to the concept of sectors, I'm   
   > > going to make everything byte-oriented. If you want to read   
   > > 512 bytes, then do a bios->fread of 512 bytes, don't ask for   
   > > a sector. Sector size may not be 512 bytes anyway.   
   > >   
   > > With bios->fopen, bios->fseek and bios->fread I am confident   
   > > that an OS will be able to call my BIOS and implement a FAT   
   > > file system.   
   > >   
   > > With those same functions above, I am not sure whether an   
   > > OS will be able to read a normal/real CDROM.   
   > >   
   > > I'm willing to add/subtract metadata on the fly in my BIOS,   
   > > but I'm still not sure a file system can be implemented. I'm   
   > > not sure if the OS authors will insist that they need a   
   > > bios->skip_to_track and bios->read_variable_sector or who   
   > > knows what else functions.   
      
   > Either way, a code would still be needed to translate sector addressing to   
   > byte addressing; because disk storages are all sector based at hardware   
   > level.   
      
   Sure. There would be a burden on the BIOS to remember   
   the current location (after either an fread or an fseek) so   
   that it knows which sector(s) and starting point to read   
   from. Is that a problem? UEFI put an entire OS into firmware.   
   I'm suggesting putting a partial C library into firmware.   
      
   BTW, I mentioned that hard disks would be limited to 2 GiB   
   in size, but that would be the 32-bit version of the BIOS. A   
   64-bit version would be limited to 2**63 bytes.   
      
   Anyway, my question remains - I can see that it is possible   
   to put a partial C library into the BIOS to cope with hard   
   disks - but can the same be done for CDROMs or does the   
   C90 standard need to be expanded on to include new calls   
   to deal with CDROMs and/or mainframe CKD disks?   
      
   BFN. Paul.   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)