[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20200305192416.GA23804@tuxdriver.com>
Date: Thu, 5 Mar 2020 14:24:16 -0500
From: "John W. Linville" <linville@...driver.com>
To: Michal Kubecek <mkubecek@...e.cz>
Cc: netdev@...r.kernel.org, Andrew Lunn <andrew@...n.ch>,
Florian Fainelli <f.fainelli@...il.com>
Subject: Re: [PATCH ethtool v2 00/25] initial netlink interface
implementation for 5.6 release
On Wed, Mar 04, 2020 at 09:24:35PM +0100, Michal Kubecek wrote:
> This series adds initial support for ethtool netlink interface provided by
> kernel since 5.6-rc1. The traditional ioctl interface is still supported
> for compatibility with older kernels. The netlink interface and message
> formats are documented in Documentation/networking/ethtool-netlink.rst file
> in kernel source tree.
>
> Netlink interface is preferred but ethtool falls back to ioctl if netlink
> interface is not available (i.e. the "ethtool" genetlink family is not
> registered). It also falls back if a particular command is not implemented
> in netlink (kernel returns -EOPNOTSUPP). This allows new ethtool versions
> to work with older kernel versions while support for ethool commands is
> added in steps.
>
> The series aims to touch existing ioctl code as little as possible in the
> first phase to minimize the risk of introducing regressions. It is also
> possible to build ethtool without netlink support if --disable-netlink is
> passed to configure script. The most visible changes to existing code are
>
> - UAPI header copies are moved to uapi/ under original names
> - some variables and functions which are going to be shared with netlink
> code are moved from ethtool.c to common.c and common.h
> - args[] array in ethtool.c was rewritten to use named initializers
>
> Except for changes to main(), all netlink specific code is in a separate
> directory netlink/ and is divided into multiple files.
>
> Changes in v2:
> - add support for permanent hardware addres ("ethtool -P", patch 20)
> - add support for pretty printing of netlink messages (patches 21-25)
> - make output of "ethtool <dev>" closer to ioctl implementation
> - two more kernel uapi header copies (patch 5)
> - support for rtnetlink socket and requests (needed for "ethtool -P")
> - some kerneldoc style comments
>
> Michal Kubecek (25):
> move UAPI header copies to a separate directory
> update UAPI header copies
> add --debug option to control debugging messages
> use named initializers in command line option list
> netlink: add netlink related UAPI header files
> netlink: introduce the netlink interface
> netlink: message buffer and composition helpers
> netlink: netlink socket wrapper and helpers
> netlink: initialize ethtool netlink socket
> netlink: add support for string sets
> netlink: add notification monitor
> move shared code into a common file
> netlink: add bitset helpers
> netlink: partial netlink handler for gset (no option)
> netlink: support getting wake-on-lan and debugging settings
> netlink: add basic command line parsing helpers
> netlink: add bitset command line parser handlers
> netlink: add netlink handler for sset (-s)
> netlink: support tests with netlink enabled
> netlink: add handler for permaddr (-P)
> netlink: support for pretty printing netlink messages
> netlink: message format description for ethtool netlink
> netlink: message format descriptions for genetlink control
> netlink: message format descriptions for rtnetlink
> netlink: use pretty printing for ethtool netlink messages
>
> Makefile.am | 31 +-
> common.c | 145 +++
> common.h | 26 +
> configure.ac | 14 +-
> ethtool.8.in | 48 +-
> ethtool.c | 819 ++++++++------
> internal.h | 31 +-
> netlink/bitset.c | 218 ++++
> netlink/bitset.h | 26 +
> netlink/desc-ethtool.c | 139 +++
> netlink/desc-genlctrl.c | 56 +
> netlink/desc-rtnl.c | 96 ++
> netlink/extapi.h | 46 +
> netlink/monitor.c | 229 ++++
> netlink/msgbuff.c | 255 +++++
> netlink/msgbuff.h | 117 ++
> netlink/netlink.c | 216 ++++
> netlink/netlink.h | 87 ++
> netlink/nlsock.c | 405 +++++++
> netlink/nlsock.h | 45 +
> netlink/parser.c | 1058 ++++++++++++++++++
> netlink/parser.h | 144 +++
> netlink/permaddr.c | 114 ++
> netlink/prettymsg.c | 237 ++++
> netlink/prettymsg.h | 118 ++
> netlink/settings.c | 955 ++++++++++++++++
> netlink/strset.c | 297 +++++
> netlink/strset.h | 25 +
> test-cmdline.c | 29 +-
> test-features.c | 11 +
> ethtool-copy.h => uapi/linux/ethtool.h | 17 +
> uapi/linux/ethtool_netlink.h | 237 ++++
> uapi/linux/genetlink.h | 89 ++
> uapi/linux/if_link.h | 1051 +++++++++++++++++
> net_tstamp-copy.h => uapi/linux/net_tstamp.h | 27 +
> uapi/linux/netlink.h | 248 ++++
> uapi/linux/rtnetlink.h | 777 +++++++++++++
> 37 files changed, 8102 insertions(+), 381 deletions(-)
> create mode 100644 common.c
> create mode 100644 common.h
> create mode 100644 netlink/bitset.c
> create mode 100644 netlink/bitset.h
> create mode 100644 netlink/desc-ethtool.c
> create mode 100644 netlink/desc-genlctrl.c
> create mode 100644 netlink/desc-rtnl.c
> create mode 100644 netlink/extapi.h
> create mode 100644 netlink/monitor.c
> create mode 100644 netlink/msgbuff.c
> create mode 100644 netlink/msgbuff.h
> create mode 100644 netlink/netlink.c
> create mode 100644 netlink/netlink.h
> create mode 100644 netlink/nlsock.c
> create mode 100644 netlink/nlsock.h
> create mode 100644 netlink/parser.c
> create mode 100644 netlink/parser.h
> create mode 100644 netlink/permaddr.c
> create mode 100644 netlink/prettymsg.c
> create mode 100644 netlink/prettymsg.h
> create mode 100644 netlink/settings.c
> create mode 100644 netlink/strset.c
> create mode 100644 netlink/strset.h
> rename ethtool-copy.h => uapi/linux/ethtool.h (99%)
> create mode 100644 uapi/linux/ethtool_netlink.h
> create mode 100644 uapi/linux/genetlink.h
> create mode 100644 uapi/linux/if_link.h
> rename net_tstamp-copy.h => uapi/linux/net_tstamp.h (84%)
> create mode 100644 uapi/linux/netlink.h
> create mode 100644 uapi/linux/rtnetlink.h
>
> --
> 2.25.1
Just a quick check -- executing "./autogen.sh ; ./configure ; make
distcheck" fails with the attached log output.
John
--
John W. Linville Someday the world will need a hero, and you
linville@...driver.com might be all we have. Be ready.
View attachment "make-distcheck.log" of type "text/plain" (7589 bytes)
Powered by blists - more mailing lists