[<prev] [next>] [day] [month] [year] [list]
Message-Id: <20241028084644.3778081-1-peng.fan@oss.nxp.com>
Date: Mon, 28 Oct 2024 16:46:43 +0800
From: "Peng Fan (OSS)" <peng.fan@....nxp.com>
To: Will Deacon <will@...nel.org>,
Lorenzo Pieralisi <lpieralisi@...nel.org>,
Krzysztof Wilczyński <kw@...ux.com>,
Manivannan Sadhasivam <manivannan.sadhasivam@...aro.org>,
Rob Herring <robh@...nel.org>,
Bjorn Helgaas <bhelgaas@...gle.com>,
Pali Rohár <pali@...nel.org>,
linux-pci@...r.kernel.org (open list:PCI DRIVER FOR GENERIC OF HOSTS),
linux-arm-kernel@...ts.infradead.org (moderated list:PCI DRIVER FOR GENERIC OF HOSTS),
linux-kernel@...r.kernel.org (open list)
Cc: Peng Fan <peng.fan@....com>
Subject: [PATCH] PCI: check bridge->bus in pci_host_common_remove
From: Peng Fan <peng.fan@....com>
When PCI node was created using an overlay and the overlay is
reverted/destroyed, the "linux,pci-domain" property no longer exists,
so of_get_pci_domain_nr will return failure. Then
of_pci_bus_release_domain_nr will actually use the dynamic IDA, even
if the IDA was allocated in static IDA. So the flow is as below:
A: of_changeset_revert
pci_host_common_remove
pci_bus_release_domain_nr
of_pci_bus_release_domain_nr
of_get_pci_domain_nr # fails because overlay is gone
ida_free(&pci_domain_nr_dynamic_ida)
With driver calls pci_host_common_remove explicity, the flow becomes:
B pci_host_common_remove
pci_bus_release_domain_nr
of_pci_bus_release_domain_nr
of_get_pci_domain_nr # succeeds in this order
ida_free(&pci_domain_nr_static_ida)
A of_changeset_revert
pci_host_common_remove
With updated flow, the pci_host_common_remove will be called twice,
so need to check 'bridge->bus' to avoid accessing invalid pointer.
Fixes: c14f7ccc9f5d ("PCI: Assign PCI domain IDs by ida_alloc()")
Signed-off-by: Peng Fan <peng.fan@....com>
---
V1:
Not sure to keep the fixes here. I could drop the Fixes tag if it is
improper.
This is to revisit the patch [1] which was rejected last year. This
new flow is using the suggest flow following Bjorn's suggestion.
But of_changeset_revert will still invoke plaform_remove and then
pci_host_common_remove. So worked out this patch together with a patch
to jailhouse driver as below:
static void destroy_vpci_of_overlay(void)
{
+ struct device_node *vpci_node = NULL;
+
if (overlay_applied) {
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(6,6,0)
+ vpci_node = of_find_node_by_path("/pci@0");
+ if (vpci_node) {
+ struct platform_device *pdev = of_find_device_by_node(vpci_node);
+ if (!pdev)
+ printk("Not found device for /pci@0\n");
+ else {
+ pci_host_common_remove(pdev);
+ platform_device_put(pdev);
+ }
+ }
+ of_node_put(vpci_node);
+#endif
+
of_changeset_revert(&overlay_changeset);
[1] https://lore.kernel.org/lkml/20230908224858.GA306960@bhelgaas/T/#md12e6097d91a012ede78c997fc5abf460029a569
drivers/pci/controller/pci-host-common.c | 6 ++++--
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/drivers/pci/controller/pci-host-common.c b/drivers/pci/controller/pci-host-common.c
index cf5f59a745b3..5a9c29fc57cd 100644
--- a/drivers/pci/controller/pci-host-common.c
+++ b/drivers/pci/controller/pci-host-common.c
@@ -86,8 +86,10 @@ void pci_host_common_remove(struct platform_device *pdev)
struct pci_host_bridge *bridge = platform_get_drvdata(pdev);
pci_lock_rescan_remove();
- pci_stop_root_bus(bridge->bus);
- pci_remove_root_bus(bridge->bus);
+ if (bridge->bus) {
+ pci_stop_root_bus(bridge->bus);
+ pci_remove_root_bus(bridge->bus);
+ }
pci_unlock_rescan_remove();
}
EXPORT_SYMBOL_GPL(pci_host_common_remove);
--
2.37.1
Powered by blists - more mailing lists