netdev - Re: [PATCH net v2 2/2] net: clarify SO_DEVMEM

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: <Zy1_IG9v1KK8u2X4@mini-arch>
Date: Thu, 7 Nov 2024 19:01:52 -0800
From: Stanislav Fomichev <stfomichev@...il.com>
To: Mina Almasry <almasrymina@...gle.com>
Cc: netdev@...r.kernel.org, Jakub Kicinski <kuba@...nel.org>,
	Eric Dumazet <edumazet@...gle.com>,
	Willem de Bruijn <willemb@...gle.com>,
	"David S. Miller" <davem@...emloft.net>, linux-doc@...r.kernel.org,
	linux-kernel@...r.kernel.org, Paolo Abeni <pabeni@...hat.com>,
	Simon Horman <horms@...nel.org>, Jonathan Corbet <corbet@....net>,
	Yi Lai <yi1.lai@...ux.intel.com>,
	Stanislav Fomichev <sdf@...ichev.me>
Subject: Re: [PATCH net v2 2/2] net: clarify SO_DEVMEM_DONTNEED behavior in
 documentation

On 11/07, Mina Almasry wrote:
> On Thu, Nov 7, 2024 at 5:30 PM Stanislav Fomichev <stfomichev@...il.com> wrote:
> >
> > On 11/07, Mina Almasry wrote:
> > > Document new behavior when the number of frags passed is too big.
> > >
> > > Signed-off-by: Mina Almasry <almasrymina@...gle.com>
> > > ---
> > >  Documentation/networking/devmem.rst | 9 +++++++++
> > >  1 file changed, 9 insertions(+)
> > >
> > > diff --git a/Documentation/networking/devmem.rst b/Documentation/networking/devmem.rst
> > > index a55bf21f671c..d95363645331 100644
> > > --- a/Documentation/networking/devmem.rst
> > > +++ b/Documentation/networking/devmem.rst
> > > @@ -225,6 +225,15 @@ The user must ensure the tokens are returned to the kernel in a timely manner.
> > >  Failure to do so will exhaust the limited dmabuf that is bound to the RX queue
> > >  and will lead to packet drops.
> > >
> > > +The user must pass no more than 128 tokens, with no more than 1024 total frags
> > > +among the token->token_count across all the tokens. If the user provides more
> > > +than 1024 frags, the kernel will free up to 1024 frags and return early.
> > > +
> > > +The kernel returns the number of actual frags freed. The number of frags freed
> > > +can be less than the tokens provided by the user in case of:
> > > +
> >
> > [..]
> >
> > > +(a) an internal kernel leak bug.
> >
> > If you're gonna respin, might be worth mentioning that the dmesg
> > will contain a warning in case of a leak?
> 
> We will not actually warn in the likely cases of leak.
> 
> We warn when we find an entry in the xarray that is not a net_iov, or
> if napi_pp_put_page fails on that net_iov. Both are very unlikely to
> happen honestly.
> 
> The likely 'leaks' are when we don't find the frag_id in the xarray.
> We do not warn on that because the user can intentionally trigger the
> warning with invalid input. If the user is actually giving valid input
> and the warn still happens, likely a kernel bug like I mentioned in
> another thread, but we still don't warn.

In this case, maybe don't mention the leaks at all? If it's not
actionable, not sure how it helps?