From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([2001:4830:134:3::10]:50062)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <nikunj@linux.vnet.ibm.com>) id 1YuZV0-0007c0-JU
	for qemu-devel@nongnu.org; Tue, 19 May 2015 00:51:27 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <nikunj@linux.vnet.ibm.com>) id 1YuZUx-0006Op-9P
	for qemu-devel@nongnu.org; Tue, 19 May 2015 00:51:26 -0400
Received: from e28smtp09.in.ibm.com ([122.248.162.9]:44420)
	by eggs.gnu.org with esmtp (Exim 4.71)
	(envelope-from <nikunj@linux.vnet.ibm.com>) id 1YuZUw-0006OE-C4
	for qemu-devel@nongnu.org; Tue, 19 May 2015 00:51:23 -0400
Received: from /spool/local
	by e28smtp09.in.ibm.com with IBM ESMTP SMTP Gateway: Authorized Use
	Only! Violators will be prosecuted
	for <qemu-devel@nongnu.org> from <nikunj@linux.vnet.ibm.com>;
	Tue, 19 May 2015 10:21:17 +0530
From: Nikunj A Dadhania <nikunj@linux.vnet.ibm.com>
In-Reply-To: <554C68DC.4020604@ozlabs.ru>
References: <1430983314-5009-1-git-send-email-nikunj@linux.vnet.ibm.com>
	<1430983314-5009-5-git-send-email-nikunj@linux.vnet.ibm.com>
	<554C68DC.4020604@ozlabs.ru>
Date: Tue, 19 May 2015 10:21:10 +0530
Message-ID: <87vbfpf9vl.fsf@abhimanyu.in.ibm.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
Subject: Re: [Qemu-devel] [PATCH v4 4/4] spapr_pci: populate ibm,loc-code
List-Id: <qemu-devel.nongnu.org>
List-Unsubscribe: <https://lists.nongnu.org/mailman/options/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=unsubscribe>
List-Archive: <http://lists.nongnu.org/archive/html/qemu-devel>
List-Post: <mailto:qemu-devel@nongnu.org>
List-Help: <mailto:qemu-devel-request@nongnu.org?subject=help>
List-Subscribe: <https://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=subscribe>
To: Alexey Kardashevskiy <aik@ozlabs.ru>, qemu-devel@nongnu.org, david@gibson.dropbear.id.au
Cc: nikunj.dadhania@gmail.com, qemu-ppc@nongnu.org, agraf@suse.de, mdroth@linux.vnet.ibm.com

Alexey Kardashevskiy <aik@ozlabs.ru> 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 =E2=80=9Cibm,loc-code=E2=80=9D 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 <nikunj@linux.vnet.ibm.com>
>> ---
>>   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(sPAPRPH=
BState *phb,
>>       return drck->get_index(drc);
>>   }
>>
>> +static bool spapr_phb_vfio_get_devspec_value(PCIDevice *pdev, char **va=
lue)
>
> Does it have to be a separate function?

For better readability, i would prefer it this way.

>
>
>> +{
>> +    char *host;
>> +    char path[PATH_MAX];
>> +
>> +    host =3D object_property_get_str(OBJECT(pdev), "host", NULL);
>> +    if (!host) {
>> +        return false;
>> +    }
>> +
>> +    snprintf(path, sizeof(path), "/sys/bus/pci/devices/%s/devspec", hos=
t);
>> +    g_free(host);
>> +
>> +    return g_file_get_contents(path, value, NULL, NULL);
>> +}
>> +
>> +static char *spapr_phb_vfio_get_loc_code(sPAPRPHBState *sphb,  PCIDevic=
e *pdev)
>> +{
>> +    char path[PATH_MAX], *buf =3D NULL;
>> +
>> +    /* We have a vfio host bridge lets get the path. */
>> +    if (!spapr_phb_vfio_get_devspec_value(pdev, &buf)) {
>> +        return NULL;
>> +    }
>> +
>> +    snprintf(path, sizeof(path), "/proc/device-tree%s/ibm,loc-code", bu=
f);
>> +    g_free(buf);
>> +
>> +    g_file_get_contents(path, &buf, NULL, NULL);
>> +    return buf;
>> +}
>> +
>> +static char *spapr_phb_get_loc_code(sPAPRPHBState *sphb,  PCIDevice *pd=
ev)
>> +{
>> +    char *path =3D g_malloc(PATH_MAX);
>> +
>> +    if (!path) {
>> +        return NULL;
>> +    }
>> +
>> +    /*
>> +     * For non-vfio devices make up the location code out
>> +     * of the name, slot and function.
>> +     *
>> +     *       qemu_<name>:<phb-index>:<slot>.<fn>
>> +     */
>> +    snprintf(path, PATH_MAX, "qemu_%s:%02d:%02d.%1d", pdev->name,
>> +             sphb->index, PCI_SLOT(pdev->devfn), PCI_FUNC(pdev->devfn));
>
>
> g_strdup_printf?

Sure, with this I can get rid of this function.

>
>
>
>> +    return path;
>> +}
>> +
>> +
>> +static char *spapr_ibm_get_loc_code(sPAPRPHBState *sphb, PCIDevice *pde=
v)
>
> s/spapr_ibm_get_loc_code/spapr_phb_get_loc_code/

I can change this.

> Strange to see "ibm" in a function name, so far we have only used "ibm" i=
n=20
> RTAS handler names :)
>
>
>> +{
>> +    if (object_dynamic_cast(OBJECT(pdev), "vfio-pci") !=3D NULL) {
>
> QEMU does not compare object_dynamic_cast with NULL anywhere, so:
>
> if (object_dynamic_cast(OBJECT(pdev), "vfio-pci")) {

Sure

>
>
>
>
>
>> +        char *buf =3D 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 =3D g_malloc(PATH_MAX);
>> +            if (!buf) {
>> +                return NULL;
>> +            }
>> +            snprintf(buf, PATH_MAX, "vfio_%s:%02d:%02d.%1d", pdev->name,
>> +                     sphb->index, PCI_SLOT(pdev->devfn), PCI_FUNC(pdev-=
>devfn));
>
>
> g_strdup_printf?
>
> Also, "vfio_%s:%02d:%02d.%1d" looks very similar to=20
> "qemu_%s:%02d:%02d.%1d", you could extend spapr_phb_get_loc_code() to tak=
e=20
> "vfio"/"qemu" as a parameter.

Sure, let me refactor this with g_strdup_printf

>
>
>> +        }
>> +        return buf;
>> +    } else {
>> +        return spapr_phb_get_loc_code(sphb, pdev);
>> +    }
>> +}
>> +
>>   /* Macros to operate with address in OF binding to PCI */
>>   #define b_x(x, p, l)    (((x) & ((1<<(l))-1)) << (p))
>>   #define b_n(x)          b_x((x), 31, 1) /* 0 if relocatable */
>> @@ -906,12 +981,12 @@ static void populate_resource_props(PCIDevice *d, =
ResourceProps *rp)
>>   }
>>
>>   static int spapr_populate_pci_child_dt(PCIDevice *dev, void *fdt, int =
offset,
>> -                                       sPAPRPHBState *sphb,
>> -                                       const char *drc_name)
>> +                                       sPAPRPHBState *sphb)
>>   {
>>       ResourceProps rp;
>>       bool is_bridge =3D false;
>>       int pci_status;
>> +    char *buf =3D NULL;
>>       uint32_t drc_index =3D spapr_phb_get_pci_drc_index(sphb, dev);
>>
>>       if (pci_default_read_config(dev, PCI_HEADER_TYPE, 1) =3D=3D
>> @@ -973,9 +1048,10 @@ static int spapr_populate_pci_child_dt(PCIDevice *=
dev, void *fdt, int offset,
>>        * processed by OF beforehand
>>        */
>>       _FDT(fdt_setprop_string(fdt, offset, "name", "pci"));
>> -    if (drc_name) {
>> -        _FDT(fdt_setprop(fdt, offset, "ibm,loc-code", drc_name,
>> -                         strlen(drc_name)));
>> +    buf =3D spapr_ibm_get_loc_code(sphb, dev);
>> +    if (buf) {
>
> It is rather:
>
> if (!buf) {
> 	error_report("Bad thing just happened");
> 	return -1;
> }
>
> OR
>
> g_assert(!buf);

As discussed in the previous post, g_assert can be compiled out. So the
above one looks good.

>
> Your code can only return NULL if g_malloc() failed (otherwise it will be=
=20
> at least "(qemu"vfio)_%s:%02d:%02d.%1d") and if this happened, something=
=20
> went horribly bad.

Right. So in this case QEMU will exit.

>
>
>> +        _FDT(fdt_setprop_string(fdt, offset, "ibm,loc-code", buf));
>> +        g_free(buf);
>>       }
>>       if (drc_index) {
>>           _FDT(fdt_setprop_cell(fdt, offset, "ibm,my-drc-index", drc_ind=
ex));
>> @@ -1003,8 +1079,7 @@ typedef struct sPAPRFDT {
>>   } sPAPRFDT;
>>
>>   /* create OF node for pci device and required OF DT properties */
>> -static int spapr_create_pci_child_dt(PCIDevice *pdev, sPAPRFDT *p,
>> -                                     const char *drc_name)
>> +static int spapr_create_pci_child_dt(PCIDevice *pdev, sPAPRFDT *p)
>>   {
>>       int offset, ret;
>>       char nodename[64];
>> @@ -1017,8 +1092,8 @@ static int spapr_create_pci_child_dt(PCIDevice *pd=
ev, sPAPRFDT *p,
>>           sprintf(nodename, "pci@%d", slot);
>>       }
>>       offset =3D fdt_add_subnode(p->fdt, p->node_off, nodename);
>> -    ret =3D spapr_populate_pci_child_dt(pdev, p->fdt, offset, p->sphb,
>> -                                      drc_name);
>> +
>> +    ret =3D spapr_populate_pci_child_dt(pdev, p->fdt, offset, p->sphb);
>>       g_assert(!ret);
>>       if (ret) {
>>           return 0;
>> @@ -1033,7 +1108,6 @@ static void spapr_phb_add_pci_device(sPAPRDRConnec=
tor *drc,
>>   {
>>       sPAPRDRConnectorClass *drck =3D SPAPR_DR_CONNECTOR_GET_CLASS(drc);
>>       DeviceState *dev =3D DEVICE(pdev);
>> -    const char *drc_name =3D drck->get_name(drc);
>>       int fdt_start_offset =3D 0, fdt_size;
>>       sPAPRFDT s_fdt =3D {NULL, 0, NULL};
>>
>> @@ -1041,7 +1115,7 @@ static void spapr_phb_add_pci_device(sPAPRDRConnec=
tor *drc,
>>           s_fdt.fdt =3D create_device_tree(&fdt_size);
>>           s_fdt.sphb =3D phb;
>>           s_fdt.node_off =3D 0;
>> -        fdt_start_offset =3D spapr_create_pci_child_dt(pdev, &s_fdt, dr=
c_name);
>> +        fdt_start_offset =3D spapr_create_pci_child_dt(pdev, &s_fdt);
>>           if (!fdt_start_offset) {
>>               error_setg(errp, "Failed to create pci child device tree n=
ode");
>>               goto out;
>> @@ -1519,7 +1593,7 @@ static void spapr_populate_pci_devices_dt(PCIBus *=
bus, PCIDevice *pdev,
>>       int offset;
>>       sPAPRFDT s_fdt;
>>
>> -    offset =3D spapr_create_pci_child_dt(pdev, p, NULL);
>> +    offset =3D spapr_create_pci_child_dt(pdev, p);
>>       if (!offset) {
>>           error_report("Failed to create pci child device tree node");
>>           return;
>>
>

Thanks for the review.

Regards
Nikunj