Re: [PATCH] net: Optimize flush calculation in inet_gro_receive()
Eric Dumazet <[email protected]>
| Newsgroups | org.kernel.vger.linux-parisc,org.kernel.vger.linux-kernel,org.kernel.vger.netdev |
|---|---|
| Message-ID | <CANn89iJv59ZEeaTCY28emRPr2LGAD42hukmLnYU-KfJJs052xQ@mail.gmail.com> |
On Tue, Apr 14, 2026 at 2:36 AM David Laight <[email protected]> wrote: > The network stack is supposed to ensure that all receive packets are > aligned to that IP header is on a 4-byte boundary. > This typically requires the ethernet receive buffer be 4n+2 aligned. > Unfortunately there is some ethernet hardware that requires 4n aligned > buffers (often on SoC devices with cpu that fault misaligned accesses). > (Just writing two bytes of garbage before the frame solves the issue.) This is mostly true, but depends on NET_IP_ALIGN, On x86, this constant was changed to 0 back in 2010. commit ea812ca1b06113597adcd8e70c0f84a413d97544 Author: Alexander Duyck <[email protected]> Date: Tue Jun 29 18:38:00 2010 +0000 x86: Align skb w/ start of cacheline on newer core 2/Xeon Arch