From: Johannes Berg johannes.berg@intel.com
[ Upstream commit 80bf6ceaf9310b3f61934c69b382d4912deee049 ]
When we get into activate_mm(), lockdep complains that we're doing something strange:
WARNING: possible circular locking dependency detected 5.1.0-10252-gb00152307319-dirty #121 Not tainted ------------------------------------------------------ inside.sh/366 is trying to acquire lock: (____ptrval____) (&(&p->alloc_lock)->rlock){+.+.}, at: flush_old_exec+0x703/0x8d7
but task is already holding lock: (____ptrval____) (&mm->mmap_sem){++++}, at: flush_old_exec+0x6c5/0x8d7
which lock already depends on the new lock.
the existing dependency chain (in reverse order) is:
-> #1 (&mm->mmap_sem){++++}: [...] __lock_acquire+0x12ab/0x139f lock_acquire+0x155/0x18e down_write+0x3f/0x98 flush_old_exec+0x748/0x8d7 load_elf_binary+0x2ca/0xddb [...]
-> #0 (&(&p->alloc_lock)->rlock){+.+.}: [...] __lock_acquire+0x12ab/0x139f lock_acquire+0x155/0x18e _raw_spin_lock+0x30/0x83 flush_old_exec+0x703/0x8d7 load_elf_binary+0x2ca/0xddb [...]
other info that might help us debug this:
Possible unsafe locking scenario:
CPU0 CPU1 ---- ---- lock(&mm->mmap_sem); lock(&(&p->alloc_lock)->rlock); lock(&mm->mmap_sem); lock(&(&p->alloc_lock)->rlock);
*** DEADLOCK ***
2 locks held by inside.sh/366: #0: (____ptrval____) (&sig->cred_guard_mutex){+.+.}, at: __do_execve_file+0x12d/0x869 #1: (____ptrval____) (&mm->mmap_sem){++++}, at: flush_old_exec+0x6c5/0x8d7
stack backtrace: CPU: 0 PID: 366 Comm: inside.sh Not tainted 5.1.0-10252-gb00152307319-dirty #121 Stack: [...] Call Trace: [<600420de>] show_stack+0x13b/0x155 [<6048906b>] dump_stack+0x2a/0x2c [<6009ae64>] print_circular_bug+0x332/0x343 [<6009c5c6>] check_prev_add+0x669/0xdad [<600a06b4>] __lock_acquire+0x12ab/0x139f [<6009f3d0>] lock_acquire+0x155/0x18e [<604a07e0>] _raw_spin_lock+0x30/0x83 [<60151e6a>] flush_old_exec+0x703/0x8d7 [<601a8eb8>] load_elf_binary+0x2ca/0xddb [...]
I think it's because in exec_mmap() we have
down_read(&old_mm->mmap_sem); ... task_lock(tsk); ... activate_mm(active_mm, mm); (which does down_write(&mm->mmap_sem))
I'm not really sure why lockdep throws in the whole knowledge about the task lock, but it seems that old_mm and mm shouldn't ever be the same (and it doesn't deadlock) so tell lockdep that they're different.
Signed-off-by: Johannes Berg johannes.berg@intel.com Signed-off-by: Richard Weinberger richard@nod.at Signed-off-by: Sasha Levin sashal@kernel.org --- arch/um/include/asm/mmu_context.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/arch/um/include/asm/mmu_context.h b/arch/um/include/asm/mmu_context.h index 1a60e1328e2f..6aca4c90aa1a 100644 --- a/arch/um/include/asm/mmu_context.h +++ b/arch/um/include/asm/mmu_context.h @@ -56,7 +56,7 @@ static inline void activate_mm(struct mm_struct *old, struct mm_struct *new) * when the new ->mm is used for the first time. */ __switch_mm(&new->context.id); - down_write(&new->mmap_sem); + down_write_nested(&new->mmap_sem, 1); uml_setup_stubs(new); up_write(&new->mmap_sem); }