qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Qemu-devel] [PATCH v4 1/3] spapr: introduce a fixed IRQ number spac


From: Greg Kurz
Subject: Re: [Qemu-devel] [PATCH v4 1/3] spapr: introduce a fixed IRQ number space
Date: Fri, 6 Jul 2018 16:16:38 +0200

On Fri, 6 Jul 2018 15:27:17 +0200
Cédric Le Goater <address@hidden> wrote:

>   > +/*
> > + * The register property of a VIO device is defined in livirt using a
> > + * base number + 0x1000 increment and in QEMU by incrementing the base
> > + * register number 0x71000000.
> > + *
> > + * The formula below tries to compute a unique index number from the
> > + * register value that will be used to define the IRQ number of the
> > + * VIO device. A maximum of 256 (0x100) VIO devices is covered.
> > + *
> > + * To minimize collisions, we define two distinct ranges depending on
> > + * the "reg" value definition:
> > + *
> > + *     [0x00 - 0x7f]    user/libvirt
> > + *     [0x80 - 0xff]    QEMU VIO model
> > + *
> > + * Collisions will be detected when the IRQ is claimed.
> > + */
> > +static inline uint32_t spapr_vio_reg_to_irq(uint32_t reg)
> > +{
> > +    if (reg >= SPAPR_VIO_REG_BASE) {
> > +        return SPAPR_IRQ_VIO | (reg & 0x7f) | 0x80;
> > +    } else {
> > +        return SPAPR_IRQ_VIO | ((reg & 0x7f) ^ ((reg >> 12) & 0x7f));  
> 
> This is not good enough for the VIO devices with a user defined "reg".
> 
> I propose to extract bits [29-28] and [16-12] to compose a 7bit index :
> 
>       (((reg >> 28) & 0x3) << 5) | ((reg >> 12) & 0x1f)
> 

Looks good.

> That would give us a mapping :
> 
>       0x00000000  -> 0
>       0x00001000  -> 1
>               ...
>       0x0001F000  -> 31
>       0x10000000  -> 32
>       0x10001000  -> 33
>               ...
>       0x1001F000  -> 63
>       0x20000000  -> 64
>       0x20001000  -> 65
>               ...
>       0x2001F000  -> 95
>       0x30000000  -> 96
>       0x30001000  -> 97
>               ...
>       0x3001F000  -> 127
> 
> C.
> 
> 




reply via email to

[Prev in Thread] Current Thread [Next in Thread]