Messages in this thread | | | Date | Mon, 27 Mar 2023 04:07:43 -0700 | From | Guenter Roeck <> | Subject | Re: [PATCH v2] net: tls: fix possible race condition between do_tls_getsockopt_conf() and do_tls_setsockopt_conf() |
| |
On Mon, Mar 27, 2023 at 05:05:37PM +0800, Hangyu Hua wrote: > On 26/3/2023 22:12, Guenter Roeck wrote: > > Hi, > > > > On Tue, Feb 28, 2023 at 10:33:44AM +0800, Hangyu Hua wrote: > > > ctx->crypto_send.info is not protected by lock_sock in > > > do_tls_getsockopt_conf(). A race condition between do_tls_getsockopt_conf() > > > and do_tls_setsockopt_conf() can cause a NULL point dereference or > > > use-after-free read when memcpy. > > > > > > Please check the following link for pre-information: > > > https://lore.kernel.org/all/Y/ht6gQL+u6fj3dG@hog/ > > > > > > Fixes: 3c4d7559159b ("tls: kernel TLS support") > > > Signed-off-by: Hangyu Hua <hbh25y@gmail.com> > > > > This patch has been applied to v6.1.y. Should it be applied to older kernel > > branches as well ? I know it doesn't apply cleanly, but the conflicts > > should be easy to resolve. I'll be happy to send backports to stable@ if > > needed. > > > > Thanks, > > Guenter > > Look like Meena Shanmugam is doing this. Please check this: > > https://lore.kernel.org/all/20230323005440.518172-2-meenashanmugam@google.com/ >
Excellent. Thanks!
Guenter
| |