On 05.06.2023 11:23, Simon Horman wrote:
On Thu, Jun 01, 2023 at 03:19:25PM +0200, Marcin Szycik wrote:
quoted
From: Alexander Lobakin <redacted>
Historically, tunnel flags like TUNNEL_CSUM or TUNNEL_ERSPAN_OPT
have been defined as __be16. Now all of those 16 bits are occupied
and there's no more free space for new flags.
It can't be simply switched to a bigger container with no
adjustments to the values, since it's an explicit Endian storage,
and on LE systems (__be16)0x0001 equals to
(__be64)0x0001000000000000.
We could probably define new 64-bit flags depending on the
Endianness, i.e. (__be64)0x0001 on BE and (__be64)0x00010000... on
LE, but that would introduce an Endianness dependency and spawn a
ton of Sparse warnings. To mitigate them, all of those places which
were adjusted with this change would be touched anyway, so why not
define stuff properly if there's no choice.
Hi Marcin,
a few nits from my side, that you can take or leave.
Overall this looks good to me.
Reviewed-by: Simon Horman <redacted>
Hi Simon,
Thank you for review, I will fix the minor issues in v2.
quoted
Define IP_TUNNEL_*_BIT counterparts as a bit number instead of the
value already coded and a fistful of <16 <-> bitmap> converters and
helpers. The two flags which have a different bit position are
SIT_ISATAP_64 and VTI_ISVTI_64, as they were defined not as
__cpu_to_be16(), but as (__force __be16), i.e. had different
positions on LE and BE. Now they have a strongly defined place.
Change all __be16 fields which were used to store those flags, to
IP_TUNNEL_DECLARE_FLAGS() -> DECLARE_BITMAP(__IP_TUNNEL_FLAG_NUM) ->
unsigned long[1] for now, and replace all TUNNEL_* occurencies to
their 64-bit bitmap counterparts. Use the converters in the places
which talk to the userspace, hardware (NFP) or other hosts (GRE
header). The rest must explicitly use the new flags only. This must
be done at once, otherwise there will be too much conversions
throughout the code in the intermediate commits.
Finally, disable the old __be16 flags for use in the kernel code
(except for the two 'irregular' flags mentioned above), to prevent
any accidential (mis)use of them. For the userspace, nothing is
nit: s/accidential/accidental/
quoted
changed, only additions were made.
Signed-off-by: Alexander Lobakin <redacted>
Signed-off-by: Marcin Szycik <redacted>
quoted
---
drivers/net/bareudp.c | 19 ++-
.../ethernet/mellanox/mlx5/core/en/tc_tun.h | 2 +-
.../mellanox/mlx5/core/en/tc_tun_encap.c | 6 +-
.../mellanox/mlx5/core/en/tc_tun_geneve.c | 12 +-
.../mellanox/mlx5/core/en/tc_tun_gre.c | 9 +-
.../mellanox/mlx5/core/en/tc_tun_vxlan.c | 9 +-
.../net/ethernet/mellanox/mlx5/core/en_tc.c | 15 +-
.../ethernet/mellanox/mlxsw/spectrum_ipip.c | 32 +++--
.../ethernet/mellanox/mlxsw/spectrum_span.c | 6 +-
.../ethernet/netronome/nfp/flower/action.c | 26 +++-
drivers/net/geneve.c | 46 +++---
drivers/net/vxlan/vxlan_core.c | 14 +-
include/net/dst_metadata.h | 10 +-
include/net/flow_dissector.h | 2 +-
include/net/gre.h | 59 ++++----
include/net/ip6_tunnel.h | 4 +-
include/net/ip_tunnels.h | 90 ++++++++++--
include/net/udp_tunnel.h | 4 +-
include/uapi/linux/if_tunnel.h | 33 +++++
net/bridge/br_vlan_tunnel.c | 5 +-
net/core/filter.c | 20 +--
net/core/flow_dissector.c | 12 +-
net/ipv4/fou_bpf.c | 2 +-
net/ipv4/gre_demux.c | 2 +-
net/ipv4/ip_gre.c | 131 +++++++++++-------
net/ipv4/ip_tunnel.c | 51 ++++---
net/ipv4/ip_tunnel_core.c | 81 +++++++----
net/ipv4/ip_vti.c | 31 +++--
net/ipv4/ipip.c | 21 ++-
net/ipv4/udp_tunnel_core.c | 5 +-
net/ipv6/ip6_gre.c | 87 +++++++-----
net/ipv6/ip6_tunnel.c | 14 +-
net/ipv6/sit.c | 9 +-
net/netfilter/ipvs/ip_vs_core.c | 6 +-
net/netfilter/ipvs/ip_vs_xmit.c | 20 +--
net/netfilter/nft_tunnel.c | 45 +++---
net/openvswitch/flow_netlink.c | 55 ++++----
net/psample/psample.c | 26 ++--
net/sched/act_tunnel_key.c | 39 +++---
net/sched/cls_flower.c | 27 ++--
40 files changed, 695 insertions(+), 392 deletions(-)
nit: this is a rather long patch
Yes, but most of it comes from the fact that tunnel flags are simply used
in many places, and we need to touch all of them.
Alex, do you see a way to logically split this patch into smaller ones?
...
quoted
diff --git a/include/uapi/linux/if_tunnel.h b/include/uapi/linux/if_tunnel.h
index 102119628ff5..d222e70d8621 100644
--- a/include/uapi/linux/if_tunnel.h
+++ b/include/uapi/linux/if_tunnel.h
@@ -161,6 +161,14 @@ enum {
#define IFLA_VTI_MAX (__IFLA_VTI_MAX - 1)
+#ifndef __KERNEL__
+/* Historically, tunnel flags have been defined as __be16 and now there are
+ * no free bits left. It is strongly advised to switch the already existing
+ * userspace code to u32/BIGINT and the new *_BIT definitions from down below,
+ * as __be16 can't be simply casted to a wider type on LE systems. All new
nit: s/casted/cast/
...