On 12/22/13 18:14, Gabriel L. Somlo wrote: > On Sun, Dec 22, 2013 at 04:58:58PM +0100, Laszlo Ersek wrote: >> After this patch, ISA interrupt 6 is used by both "SMC" and "FDC0". The >> latter depends on the FDEN object, but FDEN is currently constant 1. >> >> Probably not a problem in practice (ie. most users won't try to specify >> both a floppy disk controller and an AppleSMC device), but you might >> want to handle that case nonetheless (exit with an error or some such). > > I couldn't find a command line option to prevent QEMU from starting > with a floppy controller, so unless I missed it, we'd always detect > a "conflict". > > According to the applesmc.c source, the emulated Apple SMC doesn't > support IRQ, so the number itself should be irrelevant. IRQ #6 is > what's used on real Apple hardware, but when I tried with a different > number (e.g. #5), OS X booted fine in QEMU (it does fail to boot if > we leave out IRQNoFlags entirely from the SMC DSDT node, though). > > I could patch the value of FDEN to 0 whenever I enable the SMC _STA > method (i.e, when I patch its value to 0x0B), but that still wouldn't > take care of the fact that the emulated FDC is still present. > > So, my preferred course of action would be, in this order: > > 1. Do nothing :) > > or > > 2. Use "IRQNoFlags() { 5 }" with the SMC (or any other > number that isn't already allocated.
I don't think there's anything left free: 0 - system timer (not listed explicitly) 1 - KBD (PNP0303) 2 - cascade / PIC (PNP0000) (listed only in OVMF's builtin DSDT) 3 - COM2 (PNP0501) 4 - COM1 (PNP0501) 5 - LNK[ABCD] (PNP0C0F) 6 - FDC0 (PNP0700) 7 - LPT (PNP0400) 8 - RTC (PNP0B00) 9 - LNKS (PNP0C0F) 10 - LNK[ABCD] (PNP0C0F) 11 - LNK[ABCD] (PNP0C0F) 12 - MOU (PNP0F13) 13 - FPU (PNP0C04) (listed only in OVMF's builtin DSDT) 14 - primary IDE (not listed explicitly) 15 - secondary IDE (not listed explicitly) See also http://www.plasma-online.de/english/identify/serial/pnp_id_pnp.html http://www.webopedia.com/quick_ref/IRQnumbers.asp You could reuse eg. #5, but then you'd have to distribute PCI LNK[ABCD] over #10 and #11 only, which I guess is too high a price (both patch-wise and at runtime). Rather don't touch that :) > Any other suggestions or ideas would be welcome ! I guess the "by the book" solution would be to really stop the FDC from being emulated when the AppleSMC is present, but I mention that idea only because I like to waste bandwidth. Option 1 ("Do nothing") sounds appropriate to me. Sorry for taking up some of your time... Thanks, Laszlo