Thread (25 messages) read the whole thread 25 messages, 5 authors, 2022-05-11

Re: [PATCH v5 net-next 13/13] mlx5: support BIG TCP packets

From: Eric Dumazet <edumazet@google.com>
Date: 2022-05-10 02:00:45

On Mon, May 9, 2022 at 6:38 PM Jakub Kicinski [off-list ref] wrote:
On Mon,  9 May 2022 15:21:49 -0700 Eric Dumazet wrote:
quoted
From: Coco Li <redacted>

mlx5 supports LSOv2.

IPv6 gro/tcp stacks insert a temporary Hop-by-Hop header
with JUMBO TLV for big packets.

We need to ignore/skip this HBH header when populating TX descriptor.

Note that ipv6_has_hopopt_jumbo() only recognizes very specific packet
layout, thus mlx5e_sq_xmit_wqe() is taking care of this layout only.

v2: clear hopbyhop in mlx5e_tx_get_gso_ihs()
v4: fix compile error for CONFIG_MLX5_CORE_IPOIB=y

Signed-off-by: Coco Li <redacted>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Cc: Saeed Mahameed <saeedm@nvidia.com>
Cc: Leon Romanovsky <leon@kernel.org>
So we're leaving the warning for Kees to deal with?
I think so. I do not see an easy way to escape this, unless perhaps add some
extra obfuscation, so that gcc can not determine the memcpy() third
argument at compile time.

Alternative is to remove mlx5 patch from the upstream series.
Kees is there some form of "I know what I'm doing" cast
that you could sneak us under the table?
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help