[PATCH v2 4/8] powerpc/mm: protect linear mapping modifications by a mutex
From: David Hildenbrand
Date: Wed Nov 11 2020 - 09:53:56 EST
This code currently relies on mem_hotplug_begin()/mem_hotplug_done() -
create_section_mapping()/remove_section_mapping() implementations
cannot tollerate getting called concurrently.
Let's prepare for callers (memtrace) not holding any such locks (and
don't force them to mess with memory hotplug locks).
Other parts in these functions don't seem to rely on external locking.
Cc: Michael Ellerman <mpe@xxxxxxxxxxxxxx>
Cc: Benjamin Herrenschmidt <benh@xxxxxxxxxxxxxxxxxxx>
Cc: Paul Mackerras <paulus@xxxxxxxxx>
Cc: Rashmica Gupta <rashmica.g@xxxxxxxxx>
Cc: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
Cc: Mike Rapoport <rppt@xxxxxxxxxx>
Cc: Michal Hocko <mhocko@xxxxxxxx>
Cc: Oscar Salvador <osalvador@xxxxxxx>
Cc: Wei Yang <richard.weiyang@xxxxxxxxxxxxxxxxx>
Signed-off-by: David Hildenbrand <david@xxxxxxxxxx>
---
arch/powerpc/mm/mem.c | 5 +++++
1 file changed, 5 insertions(+)
diff --git a/arch/powerpc/mm/mem.c b/arch/powerpc/mm/mem.c
index 8a86d81f8df0..ca5c4b54c366 100644
--- a/arch/powerpc/mm/mem.c
+++ b/arch/powerpc/mm/mem.c
@@ -58,6 +58,7 @@
#define CPU_FTR_NOEXECUTE 0
#endif
+static DEFINE_MUTEX(linear_mapping_mutex);
unsigned long long memory_limit;
bool init_mem_is_free;
@@ -126,8 +127,10 @@ int __ref arch_create_linear_mapping(int nid, u64 start, u64 size,
int rc;
start = (unsigned long)__va(start);
+ mutex_lock(&linear_mapping_mutex);
rc = create_section_mapping(start, start + size, nid,
params->pgprot);
+ mutex_unlock(&linear_mapping_mutex);
if (rc) {
pr_warn("Unable to create linear mapping for 0x%llx..0x%llx: %d\n",
start, start + size, rc);
@@ -144,7 +147,9 @@ void __ref arch_remove_linear_mapping(u64 start, u64 size)
start = (unsigned long)__va(start);
flush_dcache_range_chunked(start, start + size, FLUSH_CHUNK_SIZE);
+ mutex_lock(&linear_mapping_mutex);
ret = remove_section_mapping(start, start + size);
+ mutex_unlock(&linear_mapping_mutex);
WARN_ON_ONCE(ret);
/* Ensure all vmalloc mappings are flushed in case they also
--
2.26.2