From: Sergey Miroshnichenko <s.miroshnichenko@yadro.com>
To: <linuxppc-dev@lists.ozlabs.org>, <linux-pci@vger.kernel.org>
Cc: Oliver O'Halloran <oohall@gmail.com>,
Stewart Smith <stewart@linux.vnet.ibm.com>,
Alexey Kardashevskiy <aik@ozlabs.ru>,
Benjamin Herrenschmidt <benh@kernel.crashing.org>,
Russell Currey <ruscur@russell.cc>, <linux@yadro.com>,
Sergey Miroshnichenko <s.miroshnichenko@yadro.com>
Subject: [PATCH v5 7/8] powerpc/powernv/pci: Hook up the writes to PCI_SECONDARY_BUS register
Date: Mon, 11 Mar 2019 14:52:32 +0300 [thread overview]
Message-ID: <20190311115233.6514-8-s.miroshnichenko@yadro.com> (raw)
In-Reply-To: <20190311115233.6514-1-s.miroshnichenko@yadro.com>
Writing a new value to the PCI_SECONDARY_BUS register of the bridge means
that its children will become addressable on another address (new B in BDF)
or even un-addressable if the secondary bus is set to zero.
On PowerNV, device PEs are heavily BDF-dependent, so they must be updated
on every such change of its address.
Signed-off-by: Sergey Miroshnichenko <s.miroshnichenko@yadro.com>
---
arch/powerpc/platforms/powernv/pci.c | 118 ++++++++++++++++++++++++++-
1 file changed, 116 insertions(+), 2 deletions(-)
diff --git a/arch/powerpc/platforms/powernv/pci.c b/arch/powerpc/platforms/powernv/pci.c
index 8cc6661781e2..40f68955f34f 100644
--- a/arch/powerpc/platforms/powernv/pci.c
+++ b/arch/powerpc/platforms/powernv/pci.c
@@ -722,13 +722,127 @@ int pnv_pci_cfg_read(struct pci_dn *pdn,
where, size, val);
}
+static void invalidate_children_pes(struct pci_dn *pdn)
+{
+ struct pnv_phb *phb = pdn->phb->private_data;
+ struct pci_dn *child;
+ bool found_pe = false;
+ int pe_num;
+ int pe_bus;
+
+ list_for_each_entry(child, &pdn->child_list, list) {
+ struct pnv_ioda_pe *pe = (child->pe_number != IODA_INVALID_PE) ?
+ &phb->ioda.pe_array[child->pe_number] :
+ NULL;
+
+ if (!child->busno)
+ continue;
+
+ if ((child->class_code >> 8) == PCI_CLASS_BRIDGE_PCI)
+ invalidate_children_pes(child);
+
+ if (pe) {
+ u8 rid_bus = (pe->rid >> 8) & 0xff;
+
+ if (rid_bus) {
+ pe_num = child->pe_number;
+ pe_bus = rid_bus;
+ found_pe = true;
+ }
+
+ pe->rid &= 0xff;
+ }
+
+ child->busno = 0;
+ }
+
+ if (found_pe) {
+ u16 rid = pe_bus << 8;
+
+ opal_pci_set_pe(phb->opal_id, pe_num, rid, 7, 0, 0, OPAL_UNMAP_PE);
+ }
+}
+
+static u8 pre_hook_new_sec_bus(struct pci_dn *pdn, u8 new_secondary_bus)
+{
+ u32 old_secondary_bus = 0;
+
+ if ((pdn->class_code >> 8) != PCI_CLASS_BRIDGE_PCI)
+ return 0;
+
+ pnv_pci_cfg_read(pdn, PCI_SECONDARY_BUS, 1, &old_secondary_bus);
+ old_secondary_bus &= 0xff;
+
+ if (old_secondary_bus != new_secondary_bus)
+ invalidate_children_pes(pdn);
+
+ return old_secondary_bus;
+}
+
+static void update_children_pes(struct pci_dn *pdn, u8 new_secondary_bus)
+{
+ struct pnv_phb *phb = pdn->phb->private_data;
+ struct pci_dn *child;
+ bool found_pe = false;
+ int pe_num;
+
+ if (!new_secondary_bus)
+ return;
+
+ list_for_each_entry(child, &pdn->child_list, list) {
+ struct pnv_ioda_pe *pe = (child->pe_number != IODA_INVALID_PE) ?
+ &phb->ioda.pe_array[child->pe_number] :
+ NULL;
+
+ if (child->busno)
+ continue;
+
+ child->busno = new_secondary_bus;
+
+ if (pe) {
+ pe->rid |= (child->busno << 8);
+ pe_num = child->pe_number;
+ found_pe = true;
+ }
+ }
+
+ if (found_pe) {
+ u16 rid = new_secondary_bus << 8;
+
+ opal_pci_set_pe(phb->opal_id, pe_num, rid, 7, 0, 0, OPAL_MAP_PE);
+ }
+}
+
+static void post_hook_new_sec_bus(struct pci_dn *pdn, u8 new_secondary_bus)
+{
+ if ((pdn->class_code >> 8) != PCI_CLASS_BRIDGE_PCI)
+ return;
+
+ update_children_pes(pdn, new_secondary_bus);
+}
+
int pnv_pci_cfg_write(struct pci_dn *pdn,
int where, int size, u32 val)
{
struct pnv_phb *phb = pdn->phb->private_data;
+ u8 old_secondary_bus = 0, new_secondary_bus = 0;
+ int rc;
+
+ if (where == PCI_SECONDARY_BUS) {
+ new_secondary_bus = val & 0xff;
+ old_secondary_bus = pre_hook_new_sec_bus(pdn, new_secondary_bus);
+ } else if (where == PCI_PRIMARY_BUS && size > 1) {
+ new_secondary_bus = (val >> 8) & 0xff;
+ old_secondary_bus = pre_hook_new_sec_bus(pdn, new_secondary_bus);
+ }
- return pnv_pci_cfg_write_raw(phb->opal_id, pdn->busno, pdn->devfn,
- where, size, val);
+ rc = pnv_pci_cfg_write_raw(phb->opal_id, pdn->busno, pdn->devfn,
+ where, size, val);
+
+ if (new_secondary_bus && old_secondary_bus != new_secondary_bus)
+ post_hook_new_sec_bus(pdn, new_secondary_bus);
+
+ return rc;
}
#if CONFIG_EEH
--
2.20.1
next prev parent reply other threads:[~2019-03-11 11:52 UTC|newest]
Thread overview: 15+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-03-11 11:52 [PATCH v5 0/8] powerpc/powernv/pci: Make hotplug self-sufficient, independent of FW and DT Sergey Miroshnichenko
2019-03-11 11:52 ` [PATCH v5 1/8] powerpc/pci: Access PCI config space directly w/o pci_dn Sergey Miroshnichenko
2019-04-30 4:22 ` Oliver
2019-03-11 11:52 ` [PATCH v5 2/8] powerpc/powernv/pci: Suppress an EEH error when reading an empty slot Sergey Miroshnichenko
2019-04-30 4:26 ` Oliver
2019-03-11 11:52 ` [PATCH v5 3/8] powerpc/pci: Create pci_dn on demand Sergey Miroshnichenko
2019-03-11 11:52 ` [PATCH v5 4/8] powerpc/pci: Reduce code duplication in pci_add_device_node_info Sergey Miroshnichenko
2019-03-11 11:52 ` [PATCH v5 5/8] powerpc/pci/IOV: Add support for runtime enabling the VFs Sergey Miroshnichenko
2019-04-30 6:00 ` Oliver O'Halloran
2019-05-14 14:44 ` Sergey Miroshnichenko
2019-03-11 11:52 ` [PATCH v5 6/8] powerpc/pci: Don't rely on DT is the PCI_REASSIGN_ALL_BUS is set Sergey Miroshnichenko
2019-03-11 11:52 ` Sergey Miroshnichenko [this message]
2019-03-11 11:52 ` [PATCH v5 8/8] powerpc/powernv/pci: Enable reassigning the bus numbers Sergey Miroshnichenko
2019-03-27 14:10 ` [PATCH v5 0/8] powerpc/powernv/pci: Make hotplug self-sufficient, independent of FW and DT Bjorn Helgaas
2019-03-28 12:44 ` Oliver
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20190311115233.6514-8-s.miroshnichenko@yadro.com \
--to=s.miroshnichenko@yadro.com \
--cc=aik@ozlabs.ru \
--cc=benh@kernel.crashing.org \
--cc=linux-pci@vger.kernel.org \
--cc=linux@yadro.com \
--cc=linuxppc-dev@lists.ozlabs.org \
--cc=oohall@gmail.com \
--cc=ruscur@russell.cc \
--cc=stewart@linux.vnet.ibm.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).