netdev - Re: [PATCH xfrm 1/2] xfrm: release all offloaded policy memory

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CANn89iJ1q5+uyy=L_v2szsLg2D12=HZtnqo7j3KFeoH_O2j1aA@mail.gmail.com>
Date:   Thu, 20 Apr 2023 19:33:14 +0200
From:   Eric Dumazet <edumazet@...gle.com>
To:     Leon Romanovsky <leon@...nel.org>
Cc:     Steffen Klassert <steffen.klassert@...unet.com>,
        "David S. Miller" <davem@...emloft.net>,
        Herbert Xu <herbert@...dor.apana.org.au>,
        Jakub Kicinski <kuba@...nel.org>, netdev@...r.kernel.org,
        Paolo Abeni <pabeni@...hat.com>, Raed Salem <raeds@...dia.com>
Subject: Re: [PATCH xfrm 1/2] xfrm: release all offloaded policy memory

On Thu, Apr 20, 2023 at 7:05 PM Leon Romanovsky <leon@...nel.org> wrote:
>
> On Thu, Apr 20, 2023 at 06:51:52PM +0200, Eric Dumazet wrote:
> > On Wed, Apr 19, 2023 at 2:19 PM Leon Romanovsky <leon@...nel.org> wrote:
> > >
> > > From: Leon Romanovsky <leonro@...dia.com>
> > >
> > > Failure to add offloaded policy will cause to the following
> > > error once user will try to reload driver.
> > >
> > > Unregister_netdevice: waiting for eth3 to become free. Usage count = 2
> > >
> > > This was caused by xfrm_dev_policy_add() which increments reference
> > > to net_device. That reference was supposed to be decremented
> > > in xfrm_dev_policy_free(). However the latter wasn't called.
> > >
> > >  unregister_netdevice: waiting for eth3 to become free. Usage count = 2
> > >  leaked reference.
> > >   xfrm_dev_policy_add+0xff/0x3d0
> > >   xfrm_policy_construct+0x352/0x420
> > >   xfrm_add_policy+0x179/0x320
> > >   xfrm_user_rcv_msg+0x1d2/0x3d0
> > >   netlink_rcv_skb+0xe0/0x210
> > >   xfrm_netlink_rcv+0x45/0x50
> > >   netlink_unicast+0x346/0x490
> > >   netlink_sendmsg+0x3b0/0x6c0
> > >   sock_sendmsg+0x73/0xc0
> > >   sock_write_iter+0x13b/0x1f0
> > >   vfs_write+0x528/0x5d0
> > >   ksys_write+0x120/0x150
> > >   do_syscall_64+0x3d/0x90
> > >   entry_SYSCALL_64_after_hwframe+0x46/0xb0
> > >
> > > Fixes: 919e43fad516 ("xfrm: add an interface to offload policy")
> > > Signed-off-by: Leon Romanovsky <leonro@...dia.com>
> > > ---
> >
> > Reviewed-by: Eric Dumazet <edumazet@...gle.com>
> >
> > While reviewing this patch, I also saw xfrm_dev_policy_add() could use
> > GFP_KERNEL ?
>
> netdev_tracker_alloc(...) line was copied from commit e1b539bd73a7
> ("xfrm: add net device refcount tracker to struct xfrm_state_offload")
>
> Thanks

Then I guess Steffen can fix both call sites...