[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <a4f73bac-12dd-4fbf-ac56-0c704563d0c1@amd.com>
Date: Thu, 15 Aug 2024 09:13:38 -0700
From: "Nelson, Shannon" <shannon.nelson@....com>
To: Maciej Żenczykowski <maze@...gle.com>,
Maciej Żenczykowski <zenczykowski@...il.com>
Cc: Linux Network Development Mailing List <netdev@...r.kernel.org>,
"David S . Miller" <davem@...emloft.net>, Eric Dumazet
<edumazet@...gle.com>, Jakub Kicinski <kuba@...nel.org>,
Paolo Abeni <pabeni@...hat.com>,
"Kory Maincent (Dent Project)" <kory.maincent@...tlin.com>,
Ahmed Zaki <ahmed.zaki@...el.com>, Edward Cree <ecree.xilinx@...il.com>,
Yuyang Huang <yuyanghuang@...gle.com>, Lorenzo Colitti <lorenzo@...gle.com>
Subject: Re: [PATCH net-next] ethtool: add tunable api to disable various
firmware offloads
On 8/13/2024 3:33 PM, Maciej Żenczykowski wrote:
>
> In order to save power (battery), most network hardware
> designed for low power environments (ie. battery powered
> devices) supports varying types of hardware/firmware offload
> (filtering and/or generating replies) of incoming packets.
>
> The goal being to prevent device wakeups caused by ingress 'spam'.
>
> This is particularly true for wifi (especially phones/tablets),
> but isn't actually wifi specific. It can also be implemented
> in wired nics (TV) or usb ethernet dongles.
>
> For examples TVs require this to keep power consumption
> under (the EU mandated) 2 Watts while idle (display off),
> while still being discoverable on the network.
>
> This may include things like: ARP/IPv6 ND, IGMP/MLD, ping, mdns,
> but various other possibilities are also possible,
> for example:
> ethertype filtering (discarding non-IP ethertypes),
> nat-t keepalive (discarding ingress, automating periodic egress),
> tcp keepalive (generation/processing/filtering),
> tethering (forwarding) offload
>
> In many ways, in its goals, it is somewhat similar to the
> relatively standard destination mac filtering most wired nics
> have supported for ages: reduce the amount of traffic the host
> must handle.
>
> While working on Android we've discovered that there is
> no device/driver agnostic way to disable these offloads.
>
> This patch is an attempt to rectify this.
>
> It does not add an API to configure these offloads, as usually
> this isn't needed as the driver will just fetch the required
> configuration information straight from the kernel.
>
> What it does do is add a simple API to allow disabling (masking)
> them, either for debugging or for test purposes.
I can see how this would be useful for test/debug, but it seems to me it
is only half of a solution. Wouldn't there also be a need to re-enable
the offloads without having to reboot/restart the gizmo being tested?
Or even query the current status?
sln
>
> Cc: "Kory Maincent (Dent Project)" <kory.maincent@...tlin.com>
> Cc: Ahmed Zaki <ahmed.zaki@...el.com>
> Cc: Edward Cree <ecree.xilinx@...il.com>
> Cc: "David S. Miller" <davem@...emloft.net>
> Cc: Eric Dumazet <edumazet@...gle.com>
> Cc: Jakub Kicinski <kuba@...nel.org>
> Cc: Paolo Abeni <pabeni@...hat.com>
> Cc: Yuyang Huang <yuyanghuang@...gle.com>
> Cc: Lorenzo Colitti <lorenzo@...gle.com>
> Signed-off-by: Maciej Żenczykowski <maze@...gle.com>
> ---
> include/uapi/linux/ethtool.h | 15 +++++++++++++++
> net/ethtool/common.c | 1 +
> net/ethtool/ioctl.c | 5 +++++
> 3 files changed, 21 insertions(+)
>
> diff --git a/include/uapi/linux/ethtool.h b/include/uapi/linux/ethtool.h
> index 4a0a6e703483..7b58860c3731 100644
> --- a/include/uapi/linux/ethtool.h
> +++ b/include/uapi/linux/ethtool.h
> @@ -224,12 +224,27 @@ struct ethtool_value {
> #define PFC_STORM_PREVENTION_AUTO 0xffff
> #define PFC_STORM_PREVENTION_DISABLE 0
>
> +/* For power/wakeup (*not* performance) related offloads */
> +enum tunable_fw_offload_disable {
> + ETHTOOL_TUNABLE_FW_OFFLOAD_DISABLE_ALL,
> + ETHTOOL_TUNABLE_FW_OFFLOAD_DISABLE_IPV4_ARP,
> + ETHTOOL_TUNABLE_FW_OFFLOAD_DISABLE_IPV6_ND,
> + ETHTOOL_TUNABLE_FW_OFFLOAD_DISABLE_IPV4_PING,
> + ETHTOOL_TUNABLE_FW_OFFLOAD_DISABLE_IPV6_PING,
> + ETHTOOL_TUNABLE_FW_OFFLOAD_DISABLE_IPV4_IGMP,
> + ETHTOOL_TUNABLE_FW_OFFLOAD_DISABLE_IPV6_MLD,
> + ETHTOOL_TUNABLE_FW_OFFLOAD_DISABLE_IPV4_MDNS,
> + ETHTOOL_TUNABLE_FW_OFFLOAD_DISABLE_IPV6_MDNS,
> + /* 55 bits remaining for future use */
> +};
> +
> enum tunable_id {
> ETHTOOL_ID_UNSPEC,
> ETHTOOL_RX_COPYBREAK,
> ETHTOOL_TX_COPYBREAK,
> ETHTOOL_PFC_PREVENTION_TOUT, /* timeout in msecs */
> ETHTOOL_TX_COPYBREAK_BUF_SIZE,
> + ETHTOOL_FW_OFFLOAD_DISABLE, /* u64 bits numbered from LSB per tunable_fw_offload_disable */
> /*
> * Add your fresh new tunable attribute above and remember to update
> * tunable_strings[] in net/ethtool/common.c
> diff --git a/net/ethtool/common.c b/net/ethtool/common.c
> index 7257ae272296..8dc0c084fce5 100644
> --- a/net/ethtool/common.c
> +++ b/net/ethtool/common.c
> @@ -91,6 +91,7 @@ tunable_strings[__ETHTOOL_TUNABLE_COUNT][ETH_GSTRING_LEN] = {
> [ETHTOOL_TX_COPYBREAK] = "tx-copybreak",
> [ETHTOOL_PFC_PREVENTION_TOUT] = "pfc-prevention-tout",
> [ETHTOOL_TX_COPYBREAK_BUF_SIZE] = "tx-copybreak-buf-size",
> + [ETHTOOL_FW_OFFLOAD_DISABLE] = "fw-offload-disable",
> };
>
> const char
> diff --git a/net/ethtool/ioctl.c b/net/ethtool/ioctl.c
> index 18cf9fa32ae3..31318ded17a7 100644
> --- a/net/ethtool/ioctl.c
> +++ b/net/ethtool/ioctl.c
> @@ -2733,6 +2733,11 @@ static int ethtool_get_module_eeprom(struct net_device *dev,
> static int ethtool_tunable_valid(const struct ethtool_tunable *tuna)
> {
> switch (tuna->id) {
> + case ETHTOOL_FW_OFFLOAD_DISABLE:
> + if (tuna->len != sizeof(u64) ||
> + tuna->type_id != ETHTOOL_TUNABLE_U64)
> + return -EINVAL;
> + break;
> case ETHTOOL_RX_COPYBREAK:
> case ETHTOOL_TX_COPYBREAK:
> case ETHTOOL_TX_COPYBREAK_BUF_SIZE:
> --
> 2.46.0.76.ge559c4bf1a-goog
>
>
Powered by blists - more mailing lists