Re: [Qemu-devel] [PULL 48/49] i386: populate floppy drive information in DSDT

[Roman Kagan]

On Tue, Feb 09, 2016 at 11:22:01AM -0500, John Snow wrote:
> > I don't.  At the time the patch was developed there basically were no
> > mechanisms to update the geometry at all (and this was what you patchset
> > addressed, in particular, wasn't it?) so I didn't care.
> 
> That's not true.
> 
> You could swap different 1.44MB-class diskettes for other geometries,
> check this out:
> 
> static const FDFormat fd_formats[] = {
>     /* First entry is default format */
>     /* 1.44 MB 3"1/2 floppy disks */
>     { FDRIVE_DRV_144, 18, 80, 1, FDRIVE_RATE_500K, },
>     { FDRIVE_DRV_144, 20, 80, 1, FDRIVE_RATE_500K, },
>     { FDRIVE_DRV_144, 21, 80, 1, FDRIVE_RATE_500K, },
>     { FDRIVE_DRV_144, 21, 82, 1, FDRIVE_RATE_500K, },
>     { FDRIVE_DRV_144, 21, 83, 1, FDRIVE_RATE_500K, },
>     { FDRIVE_DRV_144, 22, 80, 1, FDRIVE_RATE_500K, },
>     { FDRIVE_DRV_144, 23, 80, 1, FDRIVE_RATE_500K, },
>     { FDRIVE_DRV_144, 24, 80, 1, FDRIVE_RATE_500K, },
> ...
> 
> You absolutely could get different sector and track counts before my
> patchset.

Indeed (sorry the patch was developed a couple of months ago so I had to
look at the code to refresh my memory).

However, I tried to implement the part of ACPI spec that read

> 9.9.2 _FDI (Floppy Disk Information)
> ====================================
> This object returns information about a floppy disk drive. This
> information is the same as that returned by the INT 13 Function 08H on
> IA-PCs.

so I went ahead and looked into what SeaBIOS did for int 0x13/0x08.  And
what it did was read the CMOS at 0x10 and obtain the drive type, and
then return a hardcoded set of parameters (including geometry)
associated to that drive type.  So this was what I basically did here,
too.  (As a matter of fact the first patch I submitted was just pure ASL
which mimicked exactly the SeaBIOS behavior: read the CMOS and return
the corresponding Package with parameters.)

And IIRC the drive type couldn't change at runtime so I thought I wasn't
doing worse than it was.

As for what to do now, I'll try to check how tolerant the guests are of
changing the floppy geometry under them without updating _FDI, and then
decide.

Roman.