Re: [PATCH] Uprobes: Fix deadlock between delayed_uprobe_lock and fs_reclaim
From: Oleg Nesterov
Date: Tue Feb 26 2019 - 11:22:01 EST
On 02/26, Ravi Bangoria wrote:
> On 2/8/19 2:03 PM, Ravi Bangoria wrote:
> > On 2/6/19 7:06 PM, Oleg Nesterov wrote:
> >> Ravi, I am on vacation till the end of this week, can't read your patch
> >> carefully.
> >> I am not sure I fully understand the problem, but shouldn't we change
> >> binder_alloc_free_page() to use mmput_async() ? Like it does if trylock
> >> fails.
> > I don't understand binderfs code much so I'll let Sherry comment on this.
> Sherry, Can you please comment on this.
Yes, please. I don't even know what android/binder does, but I think we need
the trivial patch below.
If nothing else, any reason why err_down_write_mmap_sem_failed needs _async
should equally apply to the case when down_write_trylock() succeeds, no?
And. This lockdep report basically means that (without this change) kmalloc
is not safe under any lock which can be taken in __mmput() path, I don't think
we should blame or change uprobes.c.
@@ -981,7 +981,7 @@ enum lru_status binder_alloc_free_page(s