From: Liam Howlett <liam.howlett@xxxxxxxxxx> [ Upstream commit f829b4b212a315b912cb23fd10aaf30534bb5ce9 ] When the superuser flushes the entire cache, the mmap_read_lock() is not taken, but mmap_read_unlock() is called. Add the missing mmap_read_lock() call. Fixes: cd2567b6850b1648 ("m68k: call find_vma with the mmap_sem held in sys_cacheflush()") Signed-off-by: Liam R. Howlett <Liam.Howlett@xxxxxxxxxx> Reviewed-by: Matthew Wilcox (Oracle) <willy@xxxxxxxxxxxxx> Link: https://lore.kernel.org/r/20210407200032.764445-1-Liam.Howlett@xxxxxxxxxx Signed-off-by: Geert Uytterhoeven <geert@xxxxxxxxxxxxxx> [ mmap_read_lock() open-coded using down_read() as was done prior to v5.8 ] Signed-off-by: Finn Thain <fthain@xxxxxxxxxxxxxx> --- arch/m68k/kernel/sys_m68k.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/arch/m68k/kernel/sys_m68k.c b/arch/m68k/kernel/sys_m68k.c index 6363ec83a290..38dcc1a2097d 100644 --- a/arch/m68k/kernel/sys_m68k.c +++ b/arch/m68k/kernel/sys_m68k.c @@ -388,6 +388,8 @@ sys_cacheflush (unsigned long addr, int scope, int cache, unsigned long len) ret = -EPERM; if (!capable(CAP_SYS_ADMIN)) goto out; + + down_read(¤t->mm->mmap_sem); } else { struct vm_area_struct *vma;