From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:39770) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1XMKvR-0001dn-BB for qemu-devel@nongnu.org; Tue, 26 Aug 2014 13:53:06 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1XMKvB-00018w-VH for qemu-devel@nongnu.org; Tue, 26 Aug 2014 13:52:57 -0400 Received: from e35.co.us.ibm.com ([32.97.110.153]:35633) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1XMKvB-00018e-Ng for qemu-devel@nongnu.org; Tue, 26 Aug 2014 13:52:41 -0400 Received: from /spool/local by e35.co.us.ibm.com with IBM ESMTP SMTP Gateway: Authorized Use Only! Violators will be prosecuted for from ; Tue, 26 Aug 2014 11:52:40 -0600 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable From: Michael Roth In-Reply-To: <53FC4ED0.2050004@ozlabs.ru> References: <1408407718-10835-1-git-send-email-mdroth@linux.vnet.ibm.com> <1408407718-10835-3-git-send-email-mdroth@linux.vnet.ibm.com> <53FC4ED0.2050004@ozlabs.ru> Message-ID: <20140826175235.21832.49395@loki> Date: Tue, 26 Aug 2014 12:52:35 -0500 Subject: Re: [Qemu-devel] [PATCH 02/12] spapr_pci: populate DRC dt entries for PHBs List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Alexey Kardashevskiy , qemu-devel@nongnu.org Cc: ncmike@ncultra.org, nfont@linux.vnet.ibm.com, qemu-ppc@nongnu.org, agraf@suse.de, tyreld@linux.vnet.ibm.com Quoting Alexey Kardashevskiy (2014-08-26 04:09:36) > On 08/19/2014 10:21 AM, Michael Roth wrote: > > Reserve 32 entries of type PCI in each PHB's initial FDT. This > > advertises to guests that each PHB is DR-capable device with > > physical hotpluggable slots. This is necessary for allowing > > hotplugging of devices to it later via bus rescan or guest rpaphp > > hotplug module. > > = > > Each entry is assigned a name of "Slot <*32 +1>", > > advertised as a hotpluggable PCI slot, and assigned to power domain > > -1 to indicate to the guest that power management is handled by the > > hardware. > > = > > This models a DR-capable PCI expansion device attached to a host/lpar > > via a single PHB with 32 physical hotpluggable slots (as opposed to a > > virtual bridge device with external management console). Hotplug will > > be handled by the guest via bus rescan or the rpaphp hotplug module. > > = > > Signed-off-by: Michael Roth > > --- > > hw/ppc/spapr.c | 3 +- > > hw/ppc/spapr_pci.c | 102 ++++++++++++++++++++++++++++++++++++= ++++++++ > > include/hw/pci-host/spapr.h | 1 + > > 3 files changed, 105 insertions(+), 1 deletion(-) > > = > > diff --git a/hw/ppc/spapr.c b/hw/ppc/spapr.c > > index d5e46c3..90b25b3 100644 > > --- a/hw/ppc/spapr.c > > +++ b/hw/ppc/spapr.c > > @@ -890,7 +890,8 @@ static void spapr_finalize_fdt(sPAPREnvironment *sp= apr, > > QLIST_FOREACH(phb, &spapr->phbs, list) { > > drc_entry =3D spapr_phb_to_drc_entry(phb->buid); > > g_assert(drc_entry); > > - ret =3D spapr_populate_pci_dt(phb, PHANDLE_XICP, fdt); > > + ret =3D spapr_populate_pci_dt(phb, PHANDLE_XICP, drc_entry->dr= c_index, > > + fdt); > > } > > = > > if (ret < 0) { > > diff --git a/hw/ppc/spapr_pci.c b/hw/ppc/spapr_pci.c > > index e85134f..924d488 100644 > > --- a/hw/ppc/spapr_pci.c > > +++ b/hw/ppc/spapr_pci.c > > @@ -851,8 +851,104 @@ static int spapr_phb_children_dt(Object *child, v= oid *opaque) > > return 1; > > } > > = > > +static void spapr_create_drc_phb_dt_entries(void *fdt, int bus_off, in= t phb_index) > > +{ > > + char char_buf[1024]; > > + uint32_t int_buf[SPAPR_DRC_PHB_SLOT_MAX + 1]; > > + uint32_t *entries; > > + int i, ret, offset; > > + > > + /* ibm,drc-indexes */ > > + memset(int_buf, 0 , sizeof(int_buf)); > > + int_buf[0] =3D SPAPR_DRC_PHB_SLOT_MAX; > > + > > + for (i =3D 1; i <=3D SPAPR_DRC_PHB_SLOT_MAX; i++) { > > + int_buf[i] =3D SPAPR_DRC_DEV_ID_BASE + (phb_index << 8) + ((i = - 1) << 3); > > + } > > + > > + ret =3D fdt_setprop(fdt, bus_off, "ibm,drc-indexes", int_buf, > > + sizeof(int_buf)); > > + if (ret) { > > + fprintf(stderr, "error adding 'ibm,drc-indexes' field for PHB = FDT"); > > + } > > + > > + /* ibm,drc-power-domains */ > > + memset(int_buf, 0, sizeof(int_buf)); > > + int_buf[0] =3D SPAPR_DRC_PHB_SLOT_MAX; > > + > > + for (i =3D 1; i <=3D SPAPR_DRC_PHB_SLOT_MAX; i++) { > > + int_buf[i] =3D 0xffffffff; > > + } > > + > > + ret =3D fdt_setprop(fdt, bus_off, "ibm,drc-power-domains", int_buf, > > + sizeof(int_buf)); > > + if (ret) { > > + fprintf(stderr, > > + "error adding 'ibm,drc-power-domains' field for PHB FD= T"); > > + } > > + > > + /* ibm,drc-names */ > > + memset(char_buf, 0, sizeof(char_buf)); > > + entries =3D (uint32_t *)&char_buf[0]; > > + *entries =3D SPAPR_DRC_PHB_SLOT_MAX; > > + offset =3D sizeof(*entries); > > + > > + for (i =3D 1; i <=3D SPAPR_DRC_PHB_SLOT_MAX; i++) { > > + offset +=3D sprintf(char_buf + offset, "Slot %d", > > + (phb_index * SPAPR_DRC_PHB_SLOT_MAX) + i - 1= ); > > + char_buf[offset++] =3D '\0'; > > + } > > + > > + ret =3D fdt_setprop(fdt, bus_off, "ibm,drc-names", char_buf, offse= t); > > + if (ret) { > > + fprintf(stderr, "error adding 'ibm,drc-names' field for PHB FD= T"); > > + } > > + > > + /* ibm,drc-types */ > > + memset(char_buf, 0, sizeof(char_buf)); > > + entries =3D (uint32_t *)&char_buf[0]; > > + *entries =3D SPAPR_DRC_PHB_SLOT_MAX; > > + offset =3D sizeof(*entries); > > + > > + for (i =3D 0; i < SPAPR_DRC_PHB_SLOT_MAX; i++) { > > + offset +=3D sprintf(char_buf + offset, "28"); > = > = > "28"? Is it for "PHB"? For each actual PCI slot. C.6.1 in PAPR 2.7 defines a stringified 28 as: "A PCI Express Rev 2 slot with 8x lanes." ...there may be more appropriate values to use though, such as "1": "A 32-bit, 5 Volt conventional PCI slot which accommodates cards that ope= rate up to 33 MHz Only." since we don't emulate PCIe with spapr-host-bridge. I seem to recall having= issues with some of these other values though, will test again and see. > = > = > = > = > -- = > Alexey