sh: Ensure active regions have a backing PMB entry.

In the NUMA or memory hot-add case where system memory has been
partitioned up, we immediately run in to a situation where the existing
PMB entry doesn't cover the new range (primarily as a result of the entry
size being shrunk to match the node size early in the initialization). In
order to fix this up it's necessary to preload a PMB mapping for the new
range prior to activation in order to circumvent reset by MMU.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>

1 files changed, 15 insertions, 2 deletions

diff --git a/arch/sh/kernel/setup.c b/arch/sh/kernel/setup.c
index 61404ed..57bd938 100644
--- a/arch/sh/kernel/setup.c
+++ b/arch/sh/kernel/setup.c
@@ -191,13 +191,18 @@ void __init __add_active_range(unsigned int nid, unsigned long start_pfn,
 						unsigned long end_pfn)
 {
 	struct resource *res = &mem_resources[nid];
+	unsigned long start, end;
 
 	WARN_ON(res->name); /* max one active range per node for now */
 
+	start = start_pfn << PAGE_SHIFT;
+	end = end_pfn << PAGE_SHIFT;
+
 	res->name = "System RAM";
-	res->start = start_pfn << PAGE_SHIFT;
-	res->end = (end_pfn << PAGE_SHIFT) - 1;
+	res->start = start;
+	res->end = end - 1;
 	res->flags = IORESOURCE_MEM | IORESOURCE_BUSY;
+
 	if (request_resource(&iomem_resource, res)) {
 		pr_err("unable to request memory_resource 0x%lx 0x%lx\n",
 		       start_pfn, end_pfn);
@@ -213,6 +218,14 @@ void __init __add_active_range(unsigned int nid, unsigned long start_pfn,
 	request_resource(res, &data_resource);
 	request_resource(res, &bss_resource);
 
+	/*
+	 * Also make sure that there is a PMB mapping that covers this
+	 * range before we attempt to activate it, to avoid reset by MMU.
+	 * We can hit this path with NUMA or memory hot-add.
+	 */
+	pmb_bolt_mapping((unsigned long)__va(start), start, end - start,
+			 PAGE_KERNEL);
+
 	add_active_range(nid, start_pfn, end_pfn);
 }