Re: [Qemu-devel] [PATCH v4 4/4] spapr_pci: populate ibm,loc-code

[Nikunj A Dadhania]

Alexander Graf <address@hidden> writes:

> On 05/19/2015 11:58 AM, Nikunj A Dadhania wrote:
>> Alexey Kardashevskiy <address@hidden> writes:
>>
>>> On 05/19/2015 06:14 PM, Nikunj A Dadhania wrote:
>>>> Alexey Kardashevskiy <address@hidden> writes:
>>>>
>>>>> On 05/19/2015 04:56 PM, Nikunj A Dadhania wrote:
>>>>>> Alexey Kardashevskiy <address@hidden> writes:
>>>>>>
>>>>>>> On 05/19/2015 02:51 PM, Nikunj A Dadhania wrote:
>>>>>>>> Alexey Kardashevskiy <address@hidden> writes:
>>>>>>>>
>>>>>>>>> On 05/07/2015 05:21 PM, Nikunj A Dadhania wrote:
>>>>>>>>>> Each hardware instance has a platform unique location code.  The OF
>>>>>>>>>> device tree that describes a part of a hardware entity must include
>>>>>>>>>> the “ibm,loc-code” property with a value that represents the location
>>>>>>>>>> code for that hardware entity.
>>>>>>>>>>
>>>>>>>>>> Populate ibm,loc-code.
>>>>>>>>>>
>>>>>>>>>> 1) PCI passthru devices need to identify with its own ibm,loc-code
>>>>>>>>>>         available on the host. In failure cases use:
>>>>>>>>>>         vfio_<name>:<phb-index>:<slot>.<fn>
>>>>>>>>>>
>>>>>>>>>> 2) Emulated devices encode as following:
>>>>>>>>>>         qemu_<name>:<phb-index>:<slot>.<fn>
>>>>>>>>>>
>>>>>>>>>> Signed-off-by: Nikunj A Dadhania <address@hidden>
>>>>>>>>>> ---
>>>>>>>>>>       hw/ppc/spapr_pci.c | 98 
>>>>>>>>>> +++++++++++++++++++++++++++++++++++++++++++++++-------
>>>>>>>>>>       1 file changed, 86 insertions(+), 12 deletions(-)
>>>>>>>>>>
>>>>>>>>>> diff --git a/hw/ppc/spapr_pci.c b/hw/ppc/spapr_pci.c
>>>>>>>>>> index 12f1b9c..d901007 100644
>>>>>>>>>> --- a/hw/ppc/spapr_pci.c
>>>>>>>>>> +++ b/hw/ppc/spapr_pci.c
>>>>>>>>>> @@ -769,6 +769,81 @@ static uint32_t 
>>>>>>>>>> spapr_phb_get_pci_drc_index(sPAPRPHBState *phb,
>>>>>>>>>>           return drck->get_index(drc);
>>>>>>>>>>       }
>>>>>>>>>>
>>>>>>>>>> +static bool spapr_phb_vfio_get_devspec_value(PCIDevice *pdev, char 
>>>>>>>>>> **value)
>>>>>>>>> Does it have to be a separate function?
>>>>>>>> For better readability, i would prefer it this way.
>>>>>>> This is why I asked - I was having problems understanding the difference
>>>>>>> between these two having 6 words names ;) Do not insist though.
>>>>>>>
>>>>>> This is what I have now, simplified:
>>>>>>
>>>>>> +static char *spapr_phb_vfio_get_loc_code(sPAPRPHBState *sphb,  
>>>>>> PCIDevice *pdev)
>>>>>> +{
>>>>>> +    char *path = NULL, *buf = NULL;
>>>>>> +    char *host = NULL;
>>>>>> +
>>>>>> +    /* Get the PCI VFIO host id */
>>>>>> +    host = object_property_get_str(OBJECT(pdev), "host", NULL);
>>>>>> +    if (!host) {
>>>>>> +        goto err_out;
>>>>>> +    }
>>>>>> +
>>>>>> +    /* Construct the path of the file that will give us the DT location 
>>>>>> */
>>>>>> +    path = g_strdup_printf("/sys/bus/pci/devices/%s/devspec", host);
>>>>>> +    g_free(host);
>>>>>> +    if (path && !g_file_get_contents(path, &buf, NULL, NULL)) {
>>>>>> +        goto err_out;
>>>>>> +    }
>>>>>> +    g_free(path);
>>>>>> +
>>>>>> +    /* Construct and read from host device tree the loc-code */
>>>>>> +    path = g_strdup_printf("/proc/device-tree%s/ibm,loc-code", buf);
>>>>>> +    g_free(buf);
>>>>>> +    if (path && !g_file_get_contents(path, &buf, NULL, NULL)) {
>>>>>> +        goto err_out;
>>>>>> +    }
>>>>>> +    return buf;
>>>>>> +
>>>>>> +err_out:
>>>>>> +    g_free(path);
>>>>>> +    return NULL;
>>>>>> +}
>>>>>> +
>>>>>> +static char *spapr_phb_get_loc_code(sPAPRPHBState *sphb, PCIDevice 
>>>>>> *pdev)
>>>>>> +{
>>>>>> +    if (object_dynamic_cast(OBJECT(pdev), "vfio-pci")) {
>>>>>> +        char *buf = spapr_phb_vfio_get_loc_code(sphb, pdev);
>>>>>> +
>>>>>> +        /*
>>>>>> +         * In case of failures reading the loc-code, make it up
>>>>>> +         * indicating a vfio device
>>>>>> +         */
>>>>>> +        if (!buf) {
>>>>>> +            buf = g_strdup_printf("vfio_%s:%02d:%02d.%1d", pdev->name,
>>>>>> +                                  sphb->index, PCI_SLOT(pdev->devfn),
>>>>>> +                                  PCI_FUNC(pdev->devfn));
>>>>>> +        }
>>>>>> +        return buf;
>>>>>> +    } else {
>>>>>> +        return g_strdup_printf("qemu_%s:%02d:%02d.%1d", pdev->name,
>>>>>> +                               sphb->index, PCI_SLOT(pdev->devfn),
>>>>>> +                               PCI_FUNC(pdev->devfn));
>>>>>> +    }
>>>>>> +}
>>>>>
>>>>> I'd do this but I do not insist :)
>>>> That is fine as well.
>>>>
>>>>> static char *spapr_phb_get_loc_code(sPAPRPHBState *sphb, PCIDevice *pdev)
>>>>> {
>>>>>        char *buf;
>>>>>        char *devtype = "qemu";
>>>>>
>>>>>        if (object_dynamic_cast(OBJECT(pdev), "vfio-pci")) {
>>>>>            buf = spapr_phb_vfio_get_loc_code(sphb, pdev);
>>>>>            if (buf) {
>>>>>                return buf;
>>>>>            }
>>>>>            devtype = "vfio";
>>>> That has to be a snprintf.
>>> No it does not. g_strdup_printf() below is enough
>>    CC    ppc64-softmmu/hw/ppc/spapr_pci.o
>> hw/ppc/spapr_pci.c: In function ‘spapr_phb_get_loc_code’:
>> hw/ppc/spapr_pci.c:807:21: error: initialization discards ‘const’ qualifier 
>> from pointer target type [-Werror]
>>       char *devtype = "qemu";
>>                       ^
>> hw/ppc/spapr_pci.c:814:17: error: assignment discards ‘const’ qualifier from 
>> pointer target type [-Werror]
>>           devtype = "vfio";
>>                   ^
>> cc1: all warnings being treated as errors
>> make[1]: *** [hw/ppc/spapr_pci.o] Error 1
>> make: *** [subdir-ppc64-softmmu] Error 2
>
> Make it const char *devtype?

Cool, that works.

Regards,
Nikunj