On 10/17/21 20:44, BALATON Zoltan wrote: > On Sun, 17 Oct 2021, Philippe Mathieu-Daudé wrote: >> On 10/15/21 03:06, BALATON Zoltan wrote: >>> Use via_isa_set_irq() which better encapsulates irq handling in the >>> vt82xx model and avoids using isa_get_irq() that has a comment saying >>> it should not be used. >>> >>> Signed-off-by: BALATON Zoltan <bala...@eik.bme.hu> >>> --- >>> hw/ide/via.c | 4 ++-- >>> 1 file changed, 2 insertions(+), 2 deletions(-) >>> >>> diff --git a/hw/ide/via.c b/hw/ide/via.c >>> index 94cc2142c7..252d18f4ac 100644 >>> --- a/hw/ide/via.c >>> +++ b/hw/ide/via.c >>> @@ -29,7 +29,7 @@ >>> #include "migration/vmstate.h" >>> #include "qemu/module.h" >>> #include "sysemu/dma.h" >>> - >>> +#include "hw/isa/vt82c686.h" >>> #include "hw/ide/pci.h" >>> #include "trace.h" >>> >>> @@ -112,7 +112,7 @@ static void via_ide_set_irq(void *opaque, int n, >>> int level) >>> d->config[0x70 + n * 8] &= ~0x80; >>> } >>> >>> - qemu_set_irq(isa_get_irq(NULL, 14 + n), level); >>> + via_isa_set_irq(pci_get_function_0(d), 14 + n, level); >> >> Since pci_get_function_0() is expensive, we should cache >> 'PCIDevice *func0' in PCIIDEState, setting the pointer in >> via_ide_realize(). Do you mind sending a follow-up patch? > > I can do that but waiting for a decision on how to proceed. Will Gerd > take my first series this is based on as is then this should be a > separate series doing the clean up using pci_get_function_0 or should > these two series be merged? I'd also squash setting user_creatable = > false into this patch (and do similar for the usb one) unless you guys > think it should be a separate patch?
I don't know what Gerd will do with the USB patches. Your VIA patches are orthogonal, so I'm queuing them (1, 2, 4 and extra user_creatable) via mips-next.