Re: [PATCH kvmtool 02/21] hw/serial: Use device abstraction for FDT generator function

[Andre Przywara <andre.przywara@arm.com>]

On Thu, 11 Feb 2021 12:05:35 +0000
Alexandru Elisei <alexandru.elisei@arm.com> wrote:

Hi,

> On 12/10/20 2:28 PM, Andre Przywara wrote:
> > At the moment we use the .generate_fdt_node member of the ioport ops
> > structure to store the function pointer for the FDT node generator
> > function. ioport__register() will then put a wrapper and this pointer
> > into the device header.
> > The serial device is the only device making use of this special ioport
> > feature, so let's move this over to using the device header directly.
> >
> > This will allow us to get rid of this .generate_fdt_node member in the
> > ops and simplify the code.
> >
> > Signed-off-by: Andre Przywara <andre.przywara@arm.com>
> > ---
> >  hw/serial.c       | 49 +++++++++++++++++++++++++++++++++++++----------
> >  include/kvm/kvm.h |  2 ++
> >  2 files changed, 41 insertions(+), 10 deletions(-)
> >
> > diff --git a/hw/serial.c b/hw/serial.c
> > index 13c4663e..b0465d99 100644
> > --- a/hw/serial.c
> > +++ b/hw/serial.c
> > @@ -23,6 +23,7 @@
> >  #define UART_IIR_TYPE_BITS	0xc0
> >  
> >  struct serial8250_device {
> > +	struct device_header	dev_hdr;
> >  	struct mutex		mutex;
> >  	u8			id;
> >  
> > @@ -53,9 +54,20 @@ struct serial8250_device {
> >  	.msr			= UART_MSR_DCD | UART_MSR_DSR | UART_MSR_CTS, \
> >  	.mcr			= UART_MCR_OUT2,
> >  
> > +#ifdef CONFIG_HAS_LIBFDT
> > +static
> > +void serial8250_generate_fdt_node(void *fdt, struct device_header *dev_hdr,
> > +				  fdt_irq_fn irq_fn);
> > +#else
> > +#define serial8250_generate_fdt_node	NULL
> > +#endif
> >  static struct serial8250_device devices[] = {
> >  	/* ttyS0 */
> >  	[0]	= {
> > +		.dev_hdr = {
> > +			.bus_type	= DEVICE_BUS_IOPORT,
> > +			.data		= serial8250_generate_fdt_node,
> > +		},
> >  		.mutex			= MUTEX_INITIALIZER,
> >  
> >  		.id			= 0,
> > @@ -66,6 +78,10 @@ static struct serial8250_device devices[] = {
> >  	},
> >  	/* ttyS1 */
> >  	[1]	= {
> > +		.dev_hdr = {
> > +			.bus_type	= DEVICE_BUS_IOPORT,
> > +			.data		= serial8250_generate_fdt_node,
> > +		},
> >  		.mutex			= MUTEX_INITIALIZER,
> >  
> >  		.id			= 1,
> > @@ -76,6 +92,10 @@ static struct serial8250_device devices[] = {
> >  	},
> >  	/* ttyS2 */
> >  	[2]	= {
> > +		.dev_hdr = {
> > +			.bus_type	= DEVICE_BUS_IOPORT,
> > +			.data		= serial8250_generate_fdt_node,
> > +		},
> >  		.mutex			= MUTEX_INITIALIZER,
> >  
> >  		.id			= 2,
> > @@ -86,6 +106,10 @@ static struct serial8250_device devices[] = {
> >  	},
> >  	/* ttyS3 */
> >  	[3]	= {
> > +		.dev_hdr = {
> > +			.bus_type	= DEVICE_BUS_IOPORT,
> > +			.data		= serial8250_generate_fdt_node,
> > +		},
> >  		.mutex			= MUTEX_INITIALIZER,
> >  
> >  		.id			= 3,
> > @@ -371,13 +395,14 @@ char *fdt_stdout_path = NULL;
> >  
> >  #define DEVICE_NAME_MAX_LEN 32
> >  static
> > -void serial8250_generate_fdt_node(struct ioport *ioport, void *fdt,
> > -				  void (*generate_irq_prop)(void *fdt,
> > -							    u8 irq,
> > -							    enum irq_type))
> > +void serial8250_generate_fdt_node(void *fdt, struct device_header *dev_hdr,
> > +				  fdt_irq_fn irq_fn)
> >  {
> >  	char dev_name[DEVICE_NAME_MAX_LEN];
> > -	struct serial8250_device *dev = ioport->priv;
> > +	struct serial8250_device *dev = container_of(dev_hdr,
> > +						     struct serial8250_device,
> > +						     dev_hdr);
> > +
> >  	u64 addr = KVM_IOPORT_AREA + dev->iobase;
> >  	u64 reg_prop[] = {
> >  		cpu_to_fdt64(addr),
> > @@ -395,24 +420,26 @@ void serial8250_generate_fdt_node(struct ioport *ioport, void *fdt,
> >  	_FDT(fdt_begin_node(fdt, dev_name));
> >  	_FDT(fdt_property_string(fdt, "compatible", "ns16550a"));
> >  	_FDT(fdt_property(fdt, "reg", reg_prop, sizeof(reg_prop)));
> > -	generate_irq_prop(fdt, dev->irq, IRQ_TYPE_LEVEL_HIGH);
> > +	irq_fn(fdt, dev->irq, IRQ_TYPE_LEVEL_HIGH);
> >  	_FDT(fdt_property_cell(fdt, "clock-frequency", 1843200));
> >  	_FDT(fdt_end_node(fdt));
> >  }
> > -#else
> > -#define serial8250_generate_fdt_node	NULL
> >  #endif
> >  
> >  static struct ioport_operations serial8250_ops = {
> >  	.io_in			= serial8250_in,
> >  	.io_out			= serial8250_out,
> > -	.generate_fdt_node	= serial8250_generate_fdt_node,
> >  };
> >  
> > -static int serial8250__device_init(struct kvm *kvm, struct serial8250_device *dev)
> > +static int serial8250__device_init(struct kvm *kvm,
> > +				   struct serial8250_device *dev)
> >  {
> >  	int r;
> >  
> > +	r = device__register(&dev->dev_hdr);
> > +	if (r < 0)
> > +		return r;
> > +
> >  	ioport__map_irq(&dev->irq);
> >  	r = ioport__register(kvm, dev->iobase, &serial8250_ops, 8, dev);  
> 
> It's unfortunate that we now create two devices for one serial instance: one here,
> and one created by ioport__register(). But I guess that's unavoidable in this
> patch, and later patches will remove it.

Well, we do this already, wherever we register multiple I/O ports per
device, RTC and i8042 are an example (because we register one device
per I/O port). This really isn't a particular problem, it's just ugly,
so just get's a tiny bit more uglier for a few patches that
hopefully nobody will actually use (in anger).

> 
> >  
> > @@ -438,6 +465,7 @@ cleanup:
> >  		struct serial8250_device *dev = &devices[j];
> >  
> >  		ioport__unregister(kvm, dev->iobase);
> > +		device__unregister(&dev->dev_hdr);
> >  	}
> >  
> >  	return r;
> > @@ -455,6 +483,7 @@ int serial8250__exit(struct kvm *kvm)
> >  		r = ioport__unregister(kvm, dev->iobase);
> >  		if (r < 0)
> >  			return r;
> > +		device__unregister(&dev->dev_hdr);
> >  	}
> >  
> >  	return 0;
> > diff --git a/include/kvm/kvm.h b/include/kvm/kvm.h
> > index 53373b08..ee99c28e 100644
> > --- a/include/kvm/kvm.h
> > +++ b/include/kvm/kvm.h
> > @@ -31,6 +31,8 @@
> >  	.name = #ext,			\
> >  	.code = ext
> >  
> > +typedef void (*fdt_irq_fn)(void *fdt, u8 irq, enum irq_type);  
> 
> Shouldn't that last parameter be enum irq_type irq_type? Would you consider moving
> the typedef to include/kvm/fdt.h?

I think where I came from was that typedef's don't need parameter names,
just the type. But just having (void *, u8, enum irq_type) is really
hard to understand, so I added parameter names where this helps. "enum
irq_type irq_type" looks redundant, but I can certainly add this for
consistency.

> Other than the nitpicks above, the patch looks sound to me. I also did a quick
> test by running kvm-unit-tests:
> 
> Reviewed-by: Alexandru Elisei <alexandru.elisei@arm.com>

Thanks!

Andre

> 
> Thanks,
> Alex
> > +
> >  enum {
> >  	KVM_VMSTATE_RUNNING,
> >  	KVM_VMSTATE_PAUSED,