Re: [Qemu-devel] [PATCH v6 11/16] hw/arm/dyn_sysbus_devtree: enable vfio-calxeda-xgmac dynamic instantiation

[Eric Auger <eric.auger@linaro.org>]

On 09/10/2014 03:12 PM, Alexander Graf wrote:
> 
> 
> On 09.09.14 09:31, Eric Auger wrote:
>> vfio-calxeda-xgmac now can be instantiated using the -device option
>>
>> Signed-off-by: Eric Auger <eric.auger@linaro.org>
>>
>> ---
>>
>> v2 -> v3:
>> - correct bug of reg_attr[2*i] in vfio_fdt_add_device_node
>> - fix a bug related to compat_str_len computed on original compat
>>   instead of corrected compat
>> - wrap_vfio_fdt_add_node take a node creation function: this function
>>   needs to be specialized for each VFIO device. wrap function must be
>>   called in sysbus_device_create_devtree
>> ---
>>  hw/arm/dyn_sysbus_devtree.c | 141 ++++++++++++++++++++++++++++++++++++++++++++
>>  1 file changed, 141 insertions(+)
>>
>> diff --git a/hw/arm/dyn_sysbus_devtree.c b/hw/arm/dyn_sysbus_devtree.c
>> index 61e5b5f..3ef9430 100644
>> --- a/hw/arm/dyn_sysbus_devtree.c
>> +++ b/hw/arm/dyn_sysbus_devtree.c
>> @@ -20,6 +20,141 @@
>>  #include "hw/arm/dyn_sysbus_devtree.h"
>>  #include "qemu/error-report.h"
>>  #include "sysemu/device_tree.h"
>> +#include "hw/vfio/vfio-platform.h"
>> +#include "hw/vfio/vfio-calxeda-xgmac.h"
>> +
>> +typedef void (*vfio_fdt_add_device_node_t)(SysBusDevice *sbdev, void *opaque);
>> +
>> +static char *format_compat(char * compat)
>> +{
>> +    char *str_ptr, *corrected_compat;
>> +    /*
>> +     * process compatibility property string passed by end-user
>> +     * replaces / by , and ; by NUL character
>> +     */
>> +    corrected_compat = g_strdup(compat);
>> +
>> +    str_ptr = corrected_compat;
>> +    while ((str_ptr = strchr(str_ptr, '/')) != NULL) {
>> +        *str_ptr = ',';
>> +    }
>> +
>> +    /* substitute ";" with the NUL char */
>> +    str_ptr = corrected_compat;
>> +    while ((str_ptr = strchr(str_ptr, ';')) != NULL) {
>> +        *str_ptr = '\0';
>> +    }
>> +
>> +    /*
>> +     * corrected compat includes a "\0" before or at the same location
>> +     * as compat's one
>> +     */
>> +    return corrected_compat;
>> +}
>> +
>> +static void wrap_vfio_fdt_add_node(SysBusDevice *sbdev, void *opaque,
>> +                                   vfio_fdt_add_device_node_t add_node_fn)
>> +{
>> +    PlatformDevtreeData *data = opaque;
>> +    VFIOPlatformDevice *vdev = VFIO_PLATFORM_DEVICE(sbdev);
>> +    VFIODevice *vbasedev = &vdev->vbasedev;
>> +    gchar irq_number_prop[8];
>> +    Object *obj = OBJECT(sbdev);
>> +    char *corrected_compat;
>> +    uint64_t irq_number;
>> +    int corrected_compat_str_len, i;
>> +
>> +    corrected_compat = format_compat(vdev->compat);
>> +    corrected_compat_str_len = strlen(corrected_compat) + 1;
>> +    /* we copy the corrected_compat string + its "\0" */
>> +    snprintf(vdev->compat, corrected_compat_str_len, "%s", corrected_compat);
>> +    g_free(corrected_compat);
>> +
>> +    add_node_fn(sbdev, opaque);
>> +
>> +    for (i = 0; i < vbasedev->num_irqs; i++) {
>> +        snprintf(irq_number_prop, sizeof(irq_number_prop), "irq[%d]", i);
>> +        irq_number = object_property_get_int(obj, irq_number_prop, NULL)
>> +                                                 + data->irq_start;
>> +        /*
>> +         * for setting irqfd up we must provide the virtual IRQ number
>> +         * which is the sum of irq_start and actual platform bus irq
>> +         * index. At realize point we do not have this info.
>> +         */
>> +        vfio_start_irq_injection(sbdev, i, irq_number);
> 
> Does this really have anything to do with fdt?

No it doesn't, I aknowledge ;-)
 Also, don't we have
> notifiers that call IRQ holders when an IRQ gets connected?
Do we? I was not able to identify such modality. the notifier would be
triggered in qemu_allocate_irq right?
 That would
> probably be the cleaner approach here.
If it is and if the functionality does not exist yet, I can it, sure.

Besides I do not get how VFIO-PCI does handle the problem of late irq
binding. If someone can share some knowledge on this, it would be much
appreciated.

Best Regards

Eric

> 
> 
> Alex
>