Re: [PATCH net-next v12 00/15] net: introduce QUIC infrastructure and core subcomponents

Xin Long <[email protected]> Wed, 3 Jun 2026 10:14:24 -0400
Newsgroups dev.linux.lists.quic,dev.linux.lists.kernel-tls-handshake,org.kernel.vger.linux-cifs,org.kernel.vger.netdev
Message-ID <CADvbK_e7T=Yfb800TbO+BcpDEp=Sf6rbtiy2daLaskFAjeg=Ww@mail.gmail.com>
On Tue, Jun 2, 2026 at 2:18=E2=80=AFPM Jakub Kicinski <[email protected]> wro=
te:
>
> On Tue, 2 Jun 2026 09:12:30 -0400 Xin Long wrote:
> > Note: I've run Sashiko on this patchset many times locally over the pas=
t
> > month and addressed all the real issues it found. I don't see any new r=
eal
> > issues being reported on:
>
> Coccinelle had a report tho:
>
> The NIPA CI coccicheck run flagged a new Coccinelle warning introduced by
> this patch:
>
>   net/quic/crypto.c:1164:13-20: WARNING opportunity for kmemdup
>
> At line 1164 of the newly introduced net/quic/crypto.c, there is a
> kmalloc() call immediately followed by a memcpy() into the allocated
> buffer.  Coccinelle's memdup checker suggests replacing this two-step
> pattern with a single kmemdup() call, which is the preferred kernel idiom=
.
>
> The fix is straightforward - replacing:
>
>   p =3D kmalloc(len, gfp);
>   if (!p)
>       return -ENOMEM;
>   memcpy(p, src, len);
>
> with:
>
>   p =3D kmemdup(src, len, gfp);
>   if (!p)
>       return -ENOMEM;
>
> Please update the patch to use kmemdup() at that location.

Will address it in v13.

BTW, is there a public link to the Coccinelle check results?
I don't see them showing up on Patchwork.

Thanks.