Re: [Qemu-devel] [PATCH 2/6] [RFC] Emulation of GRLIB IRQMP as defined in GRLIB IP Core User's Manual.

[Fabien Chouteau <chouteau@adacore.com>]

On 12/11/2010 11:31 AM, Blue Swirl wrote:
> On Tue, Dec 7, 2010 at 10:43 AM, Fabien Chouteau<chouteau@adacore.com>  wrote:
>> On 12/06/2010 06:25 PM, Blue Swirl wrote:
>>>
>>> On Mon, Dec 6, 2010 at 9:26 AM, Fabien Chouteau<chouteau@adacore.com>
>>>   wrote:
>>>>
>>>> Signed-off-by: Fabien Chouteau<chouteau@adacore.com>
>>>> ---
>>>>   hw/grlib_irqmp.c |  416
>>>> ++++++++++++++++++++++++++++++++++++++++++++++++++++++
>>>>   1 files changed, 416 insertions(+), 0 deletions(-)
>>>>
>>>> diff --git a/hw/grlib_irqmp.c b/hw/grlib_irqmp.c
>>>> new file mode 100644
>>>> index 0000000..69e1553
>>>> --- /dev/null
>>>> +++ b/hw/grlib_irqmp.c
>>>> @@ -0,0 +1,416 @@
>>>> +/*
>>>> + * QEMU GRLIB IRQMP Emulator
>>>> + *
>>>> + * (Multiprocessor and extended interrupt not supported)
>>>> + *
>>>> + * Copyright (c) 2010 AdaCore
>>>> + *
>>>> + * Permission is hereby granted, free of charge, to any person obtaining
>>>> a copy
>>>> + * of this software and associated documentation files (the "Software"),
>>>> to deal
>>>> + * in the Software without restriction, including without limitation the
>>>> rights
>>>> + * to use, copy, modify, merge, publish, distribute, sublicense, and/or
>>>> sell
>>>> + * copies of the Software, and to permit persons to whom the Software is
>>>> + * furnished to do so, subject to the following conditions:
>>>> + *
>>>> + * The above copyright notice and this permission notice shall be
>>>> included in
>>>> + * all copies or substantial portions of the Software.
>>>> + *
>>>> + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
>>>> EXPRESS OR
>>>> + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
>>>> MERCHANTABILITY,
>>>> + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT
>>>> SHALL
>>>> + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR
>>>> OTHER
>>>> + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
>>>> ARISING FROM,
>>>> + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
>>>> DEALINGS IN
>>>> + * THE SOFTWARE.
>>>> + */
>>>> +
>>>> +#include "sysbus.h"
>>>> +#include "cpu.h"
>>>> +
>>>> +#include "grlib.h"
>>>> +
>>>> +/* #define DEBUG_IRQ */
>>>> +
>>>> +#ifdef DEBUG_IRQ
>>>> +#define DPRINTF(fmt, ...)                                       \
>>>> +    do { printf("IRQMP: " fmt , ## __VA_ARGS__); } while (0)
>>>> +#else
>>>> +#define DPRINTF(fmt, ...)
>>>> +#endif
>>>> +
>>>> +#define IRQMP_MAX_CPU 16
>>>> +#define IRQMP_REG_SIZE 256      /* Size of memory mapped registers */
>>>> +
>>>> +/* Memory mapped register offsets */
>>>> +#define LEVEL_OFFSET     0x00
>>>> +#define PENDING_OFFSET   0x04
>>>> +#define FORCE0_OFFSET    0x08
>>>> +#define CLEAR_OFFSET     0x0C
>>>> +#define MP_STATUS_OFFSET 0x10
>>>> +#define BROADCAST_OFFSET 0x14
>>>> +#define MASK_OFFSET      0x40
>>>> +#define FORCE_OFFSET     0x80
>>>> +#define EXTENDED_OFFSET  0xC0
>>>> +
>>>> +typedef struct IRQMP
>>>> +{
>>>> +    SysBusDevice busdev;
>>>> +
>>>> +    CPUSPARCState *env;
>>>
>>> Devices should never access CPUState directly. Instead, board level
>>> should create CPU irqs and these should then be passed here.
>>>
>>
>> This case is special, Leon3 is a System-On-Chip and some of the components
>> are very close to the processor.
>> IRQMP is not really a peripheral nor a part of the CPU, it's both...
>
> It's not a special case, it could be easily implemented separately.
> MMUs, FPUs or co-processors could be special even if they have been
> implemented as separate chips with real hardware. But we are actually
> not looking at the (historical or current) chip boundaries but more
> like what makes sense from QEMU architecture point of view.

OK then, let's go back to your first comment, why a device can't access
CPUState directly? And why Leon3.c would be better to do that.


-- 
Fabien Chouteau