From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:58807) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1Z0Ced-0008DL-Oh for qemu-devel@nongnu.org; Wed, 03 Jun 2015 13:40:40 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1Z0Cea-000453-IZ for qemu-devel@nongnu.org; Wed, 03 Jun 2015 13:40:39 -0400 Date: Wed, 3 Jun 2015 19:40:28 +0200 From: Thomas Huth Message-ID: <20150603194028.1ea14040@thh440s> In-Reply-To: <1433330757-6043-6-git-send-email-nikunj@linux.vnet.ibm.com> References: <1433330757-6043-1-git-send-email-nikunj@linux.vnet.ibm.com> <1433330757-6043-6-git-send-email-nikunj@linux.vnet.ibm.com> MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable Subject: Re: [Qemu-devel] [PATCH v6 5/6] spapr_pci: populate ibm,loc-code List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Nikunj A Dadhania Cc: qemu-devel@nongnu.org, aik@ozlabs.ru, agraf@suse.de, mdroth@linux.vnet.ibm.com, qemu-ppc@nongnu.org, david@gibson.dropbear.id.au On Wed, 3 Jun 2015 16:55:56 +0530 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. >=20 > Populate ibm,loc-code. >=20 > 1) PCI passthru devices need to identify with its own ibm,loc-code > available on the host. In failure cases use: > vfio_:::. >=20 > 2) Emulated devices encode as following: > qemu_:::. >=20 > Signed-off-by: Nikunj A Dadhania > --- > hw/ppc/spapr_pci.c | 78 +++++++++++++++++++++++++++++++++++++++++++++++-= ------ > 1 file changed, 68 insertions(+), 10 deletions(-) >=20 > diff --git a/hw/ppc/spapr_pci.c b/hw/ppc/spapr_pci.c > index 4226468..986bb21 100644 > --- a/hw/ppc/spapr_pci.c > +++ b/hw/ppc/spapr_pci.c > @@ -746,6 +746,60 @@ static AddressSpace *spapr_pci_dma_iommu(PCIBus *bus= , void *opaque, int devfn) > return &phb->iommu_as; > } > =20 > +static char *spapr_phb_vfio_get_loc_code(sPAPRPHBState *sphb, PCIDevice= *pdev) > +{ > + char *path =3D NULL, *buf =3D NULL, *host =3D NULL; > + > + /* Get the PCI VFIO host id */ > + host =3D 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 =3D 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 =3D 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; I'd maybe change the above 4 lines into: if (path && g_file_get_contents(path, &buf, NULL, NULL)) { return buf; } so that you can get rid of one goto here. > +err_out: > + g_free(path); > + return NULL; > +} > + > +static char *spapr_phb_get_loc_code(sPAPRPHBState *sphb, PCIDevice *pdev) > +{ > + char *buf; > + const char *devtype =3D "qemu"; > + uint32_t busnr =3D pci_bus_num(PCI_BUS(qdev_get_parent_bus(DEVICE(pd= ev)))); > + > + if (object_dynamic_cast(OBJECT(pdev), "vfio-pci")) { > + buf =3D spapr_phb_vfio_get_loc_code(sphb, pdev); > + if (buf) { > + return buf; > + } > + devtype =3D "vfio"; > + } > + /* > + * For emulated devices and VFIO-failure case, make up > + * the loc-code. > + */ > + buf =3D g_strdup_printf("%s_%s:%04x:%02x:%02x.%x", > + devtype, pdev->name, sphb->index, busnr, > + PCI_SLOT(pdev->devfn), PCI_FUNC(pdev->devfn)); > + return buf; > +} > + > /* 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 */ > @@ -884,11 +938,12 @@ static void populate_resource_props(PCIDevice *d, R= esourceProps *rp) > =20 > static int spapr_populate_pci_child_dt(PCIDevice *dev, void *fdt, int of= fset, > int phb_index, int drc_index, > - const char *drc_name) > + sPAPRPHBState *sphb) > { > ResourceProps rp; > bool is_bridge =3D false; > int pci_status; > + char *buf =3D NULL; Is the "=3D NULL" required here? If not, please remove, newer version of gcc tend to complain otherwise. > if (pci_default_read_config(dev, PCI_HEADER_TYPE, 1) =3D=3D > PCI_HEADER_TYPE_BRIDGE) { > @@ -949,10 +1004,15 @@ 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_phb_get_loc_code(sphb, dev); > + if (!buf) { > + error_report("Failed setting the ibm,loc-code"); > + return -1; > } > + > + _FDT(fdt_setprop_string(fdt, offset, "ibm,loc-code", buf)); I wonder whether this will cause some Coverity warnings later ... the _FDT macro can return immediately (ugh, return in a macro ... IMHO a bad idea...). buf is not freed in that case, and that might trigger a warning... > + g_free(buf); > + > if (drc_index) { > _FDT(fdt_setprop_cell(fdt, offset, "ibm,my-drc-index", drc_index= )); > } [...] Thomas