Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.34 as permitted sender) client-ip=23.128.96.34;
Date:   Sat, 7 Oct 2023 13:18:52 +0200
From:   Greg KH <gregkh@linuxfoundation.org>
To:     Kassey Li <quic_yingangl@quicinc.com>
Cc:     gregkh@google.com, cmllamas@google.com, surenb@google.com,
        arve@android.com, joel@joelfernandes.org, brauner@kernel.org,
        tkjos@android.com, maco@android.com, linux-kernel@vger.kernel.org
Subject: Re: [PATCH] binder: add mutex_lock for mmap and NULL when free
Message-ID: <2023100725-oversized-shore-c873@gregkh>
References: <20231007034046.2352124-1-quic_yingangl@quicinc.com>
 <2023100715-blade-valiant-714d@gregkh>
 <26988068-8c9f-8591-db6e-44c8105af638@quicinc.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <26988068-8c9f-8591-db6e-44c8105af638@quicinc.com>
Precedence: bulk

On Sat, Oct 07, 2023 at 07:07:40PM +0800, Kassey Li wrote:
> 
> 
> On 2023/10/7 14:44, Greg KH wrote:
> > On Sat, Oct 07, 2023 at 11:40:46AM +0800, Kassey Li wrote:
> > > Enforce alloc->mutex in binder_alloc_mmap_handler when add
> > > the entry to list.
> > > 
> > > Assign the freed pages/page_ptr to NULL to catch possible
> > > use after free with NULL pointer access.
> > > 
> > > Signed-off-by: Kassey Li <quic_yingangl@quicinc.com>
> > > ---
> > >   drivers/android/binder_alloc.c | 5 ++++-
> > >   1 file changed, 4 insertions(+), 1 deletion(-)
> > 
> > What commit id does this fix?
> 
>   there is no specific commit id this change going to fix.
> 
>   it is a follow up for commit
> 	19c987241ca1216a51118b2bd0185b8bc5081783  binder: separate out binder_alloc
> functions (mutex lock added for list access in alloc/free)
> 	f2517eb76f1f2f7f89761f9db2b202e89931738c  android: binder: Add global lru
> shrinker to binder (set page->page_ptr = NULL;)
> 
>   the background to raise this change that we are easy hit below crash in
> monkey test:
> 
> where a wrong end is passing to
> binder_update_page_range, thus calculate a weird index
> for
>   page = &alloc->pages[index]

Obviously it is a fix for some commit, please list that here.

thanks,

greg k-h