On 13/10/2026 17:26, Peter Maydell wrote:
On Wed, 16 Sept 2026 at 11:49, Mark Cave-Ayland
<[email protected]> wrote:
The isa_pic variable exists to allow the output to be fetched or read (fetched
and acknowledged) from the x86 PIC. Since this can also contain a cascaded
interrupt from a slave i8259, switch isa_pic from I8259CommonState to
I8259PICState and update both pic_get_output() and pic_read_irq() accordingly.
This has the advantage that a i8259-based PIC device contains its own internal
reference to both the primary and secondary i8259s, and therefore it becomes
possible to completely remove the slave_pic variable.
Signed-off-by: Mark Cave-Ayland <[email protected]>
---
Reviewed-by: Peter Maydell <[email protected]>
(That isa_pic global is kind of ugly; in an ideal world the callers
would have a pointer to the PIC they cared about rather than relying
on there being exactly one in the system...)
That was actually one of the ultimate aims of this series, however I
submitted the 8259 changes separately to avoid sending a much larger
series :)
ATB,
Mark.
