3.16.40-rc1 review patch. If anyone has any objections, please let me know.
------------------
From: Furquan Shaikh <furquan@xxxxxxxxxx>
commit 7e75678d23167c2527e655658a8ef36a36c8b4d9 upstream.
persistent_ram_update uses vmap / iomap based on whether the buffer is in
memory region or reserved region. However, both map it as non-cacheable
memory. For armv8 specifically, non-cacheable mapping requests use a
memory type that has to be accessed aligned to the request size. memcpy()
doesn't guarantee that.
Signed-off-by: Furquan Shaikh <furquan@xxxxxxxxxx>
Signed-off-by: Enric Balletbo Serra <enric.balletbo@xxxxxxxxxxxxx>
Reviewed-by: Aaron Durbin <adurbin@xxxxxxxxxxxx>
Reviewed-by: Olof Johansson <olofj@xxxxxxxxxxxx>
Tested-by: Furquan Shaikh <furquan@xxxxxxxxxxxx>
Signed-off-by: Kees Cook <keescook@xxxxxxxxxxxx>
Signed-off-by: Ben Hutchings <ben@xxxxxxxxxxxxxxx>
---
fs/pstore/ram_core.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
--- a/fs/pstore/ram_core.c
+++ b/fs/pstore/ram_core.c
@@ -263,7 +263,7 @@ static void notrace persistent_ram_updat
const void *s, unsigned int start, unsigned int count)
{
struct persistent_ram_buffer *buffer = prz->buffer;
- memcpy(buffer->data + start, s, count);
+ memcpy_toio(buffer->data + start, s, count);
persistent_ram_update_ecc(prz, start, count);
}