Date: Mon, 17 Jun 2024 21:44:58 +0100
From: Ignat Korchagin <ignat@...udflare.com>
To: Kuniyuki Iwashima <kuniyu@...zon.com>
Cc: davem@...emloft.net, dsahern@...nel.org, edumazet@...gle.com, 
	kernel-team@...udflare.com, kuba@...nel.org, linux-kernel@...r.kernel.org, 
	netdev@...r.kernel.org, pabeni@...hat.com, revest@...omium.org, 
	stable@...r.kernel.org
Subject: Re: [PATCH net v2] net: do not leave a dangling sk pointer, when
 socket creation fails

On Mon, Jun 17, 2024 at 9:31 PM Kuniyuki Iwashima <kuniyu@...zon.com> wrote:
>
> From: Ignat Korchagin <ignat@...udflare.com>
> Date: Mon, 17 Jun 2024 20:59:34 +0100
> > A KASAN enabled kernel will log something like below (decoded and stripped):
> > [   78.328507][  T299] ==================================================================
> > [ 78.329018][ T299] BUG: KASAN: slab-use-after-free in __sock_gen_cookie (./arch/x86/include/asm/atomic64_64.h:15 ./include/linux/atomic/atomic-arch-fallback.h:2583 ./include/linux/atomic/atomic-instrumented.h:1611 net/core/sock_diag.c:29)
> > [   78.329366][  T299] Read of size 8 at addr ffff888007110dd8 by task traceroute/299
> > [   78.329366][  T299]
> > [   78.329366][  T299] CPU: 2 PID: 299 Comm: traceroute Tainted: G            E      6.10.0-rc2+ #2
> > [   78.329366][  T299] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.16.2-debian-1.16.2-1 04/01/2014
> > [   78.329366][  T299] Call Trace:
> > [   78.329366][  T299]  <TASK>
> > [ 78.329366][ T299] dump_stack_lvl (lib/dump_stack.c:117 (discriminator 1))
> > [ 78.329366][ T299] print_report (mm/kasan/report.c:378 mm/kasan/report.c:488)
> > [ 78.329366][ T299] ? __sock_gen_cookie (./arch/x86/include/asm/atomic64_64.h:15 ./include/linux/atomic/atomic-arch-fallback.h:2583 ./include/linux/atomic/atomic-instrumented.h:1611 net/core/sock_diag.c:29)
> > [ 78.329366][ T299] kasan_report (mm/kasan/report.c:603)
> > [ 78.329366][ T299] ? __sock_gen_cookie (./arch/x86/include/asm/atomic64_64.h:15 ./include/linux/atomic/atomic-arch-fallback.h:2583 ./include/linux/atomic/atomic-instrumented.h:1611 net/core/sock_diag.c:29)
> > [ 78.329366][ T299] kasan_check_range (mm/kasan/generic.c:183 mm/kasan/generic.c:189)
> > [ 78.329366][ T299] __sock_gen_cookie (./arch/x86/include/asm/atomic64_64.h:15 ./include/linux/atomic/atomic-arch-fallback.h:2583 ./include/linux/atomic/atomic-instrumented.h:1611 net/core/sock_diag.c:29)
> > [ 78.329366][ T299] bpf_get_socket_ptr_cookie (./arch/x86/include/asm/preempt.h:94 ./include/linux/sock_diag.h:42 net/core/filter.c:5094 net/core/filter.c:5092)
> > [ 78.329366][ T299] bpf_prog_875642cf11f1d139___sock_release+0x6e/0x8e
> > [ 78.329366][ T299] bpf_trampoline_6442506592+0x47/0xaf
> > [ 78.329366][ T299] __sock_release (net/socket.c:652)
> > [ 78.329366][ T299] __sock_create (net/socket.c:1601)
> > ...
> > [   78.329366][  T299] Allocated by task 299 on cpu 2 at 78.328492s:
> > [ 78.329366][ T299] kasan_save_stack (mm/kasan/common.c:48)
> > [ 78.329366][ T299] kasan_save_track (mm/kasan/common.c:68)
> > [ 78.329366][ T299] __kasan_slab_alloc (mm/kasan/common.c:312 mm/kasan/common.c:338)
> > [ 78.329366][ T299] kmem_cache_alloc_noprof (mm/slub.c:3941 mm/slub.c:4000 mm/slub.c:4007)
> > [ 78.329366][ T299] sk_prot_alloc (net/core/sock.c:2075)
> > [ 78.329366][ T299] sk_alloc (net/core/sock.c:2134)
> > [ 78.329366][ T299] inet_create (net/ipv4/af_inet.c:327 net/ipv4/af_inet.c:252)
> > [ 78.329366][ T299] __sock_create (net/socket.c:1572)
> > [ 78.329366][ T299] __sys_socket (net/socket.c:1660 net/socket.c:1644 net/socket.c:1706)
> > [ 78.329366][ T299] __x64_sys_socket (net/socket.c:1718)
> > [ 78.329366][ T299] do_syscall_64 (arch/x86/entry/common.c:52 arch/x86/entry/common.c:83)
> > [ 78.329366][ T299] entry_SYSCALL_64_after_hwframe (arch/x86/entry/entry_64.S:130)
> > [   78.329366][  T299]
> > [   78.329366][  T299] Freed by task 299 on cpu 2 at 78.328502s:
> > [ 78.329366][ T299] kasan_save_stack (mm/kasan/common.c:48)
> > [ 78.329366][ T299] kasan_save_track (mm/kasan/common.c:68)
> > [ 78.329366][ T299] kasan_save_free_info (mm/kasan/generic.c:582)
> > [ 78.329366][ T299] poison_slab_object (mm/kasan/common.c:242)
> > [ 78.329366][ T299] __kasan_slab_free (mm/kasan/common.c:256)
> > [ 78.329366][ T299] kmem_cache_free (mm/slub.c:4437 mm/slub.c:4511)
> > [ 78.329366][ T299] __sk_destruct (net/core/sock.c:2117 net/core/sock.c:2208)
> > [ 78.329366][ T299] inet_create (net/ipv4/af_inet.c:397 net/ipv4/af_inet.c:252)
> > [ 78.329366][ T299] __sock_create (net/socket.c:1572)
> > [ 78.329366][ T299] __sys_socket (net/socket.c:1660 net/socket.c:1644 net/socket.c:1706)
> > [ 78.329366][ T299] __x64_sys_socket (net/socket.c:1718)
> > [ 78.329366][ T299] do_syscall_64 (arch/x86/entry/common.c:52 arch/x86/entry/common.c:83)
> > [ 78.329366][ T299] entry_SYSCALL_64_after_hwframe (arch/x86/entry/entry_64.S:130)
> >
> > Fix this by clearing the struct socket reference in sk_common_release() to cover
> > all protocol families create functions.
> >
> > Fixes: c5dbb89fc2ac ("bpf: Expose bpf_get_socket_cookie to tracing programs")
> > Suggested-by: Kuniyuki Iwashima <kuniyu@...zon.com>
> > Signed-off-by: Ignat Korchagin <ignat@...udflare.com>
> > Cc: stable@...r.kernel.org
> > Link: https://lore.kernel.org/netdev/20240613194047.36478-1-kuniyu@amazon.com/T/
> > ---
> > Changes in v2:
> >   * moved the NULL-ing of the socket reference to sk_common_release() (as
> >     suggested by Kuniyuki Iwashima)
> >   * trimmed down the KASAN report in the commit message to show only relevant
> >     info
>
> It seems the most important repro was lost.  I'd like to keep that
> in the commit message so that we can easily understand the Fixes:
> tag and how the issue happens.

Hm... Indeed... Not my intention. Somehow I stripped the repro steps
and not only the KASAN message

> While at it, could you remove the timestamp and thread id in KASAN
> splat ?

OK

>
> >
> >  net/core/sock.c | 11 +++++++++++
> >  1 file changed, 11 insertions(+)
> >
> > diff --git a/net/core/sock.c b/net/core/sock.c
> > index 8629f9aecf91..575af557c46b 100644
> > --- a/net/core/sock.c
> > +++ b/net/core/sock.c
> > @@ -3742,6 +3742,17 @@ void sk_common_release(struct sock *sk)
> >
> >       sk->sk_prot->unhash(sk);
> >
> > +     /*
> > +      * struct net_proto_family create functions like inet_create() or
>
> nit: This should be netdev style:
>
>         /* struct net_proto_family ...
>          * ...
>          */
>
>   See: Documentation/process/maintainer-netdev.rst
>
> But I think the comment is not needed here if the commit message has
>
>   * KASAN splat
>   * How the problem happens
>     * run bpf_get_socket_cookie() in __sock_release()
>   * What the problem is
>     * UAF happens if pf->create() fails after calling sock_init_data()
>
> , we can just git-blame the change below.
>
> Thanks!
>
>
> > +      * inet6_create() have an error path, which call this function. This sk
> > +      * may have already been associated with a struct socket, so ensure to
> > +      * clear this reference not to leave a dangling pointer in the
> > +      * struct socket instance.
> > +      */
> > +
> > +     if (sk->sk_socket)
> > +             sk->sk_socket->sk = NULL;
> > +
> >       /*
> >        * In this point socket cannot receive new packets, but it is possible
> >        * that some packets are in flight because some CPU runs receiver and
> > --
> > 2.39.2

Powered by blists - more mailing lists