From: Jakub Kicinski <kuba@kernel.org>
To: Aurelien Aptel <aaptel@nvidia.com>
Cc: linux-nvme@lists.infradead.org, netdev@vger.kernel.org,
sagi@grimberg.me, hch@lst.de, kbusch@kernel.org, axboe@fb.com,
chaitanyak@nvidia.com, davem@davemloft.net,
aurelien.aptel@gmail.com, smalin@nvidia.com, malin1024@gmail.com,
ogerlitz@nvidia.com, yorayz@nvidia.com, borisp@nvidia.com
Subject: Re: [PATCH v10 03/25] net/ethtool: add ULP_DDP_{GET,SET} operations for caps and stats
Date: Tue, 31 Jan 2023 20:53:27 -0800 [thread overview]
Message-ID: <20230131205327.67adfc1f@kernel.org> (raw)
In-Reply-To: <20230126162136.13003-4-aaptel@nvidia.com>
On Thu, 26 Jan 2023 18:21:14 +0200 Aurelien Aptel wrote:
> diff --git a/include/uapi/linux/ethtool_netlink.h b/include/uapi/linux/ethtool_netlink.h
> index 14a1858fd106..35805d8d12a3 100644
> --- a/include/uapi/linux/ethtool_netlink.h
> +++ b/include/uapi/linux/ethtool_netlink.h
> @@ -57,6 +57,8 @@ enum {
> ETHTOOL_MSG_PLCA_GET_STATUS,
> ETHTOOL_MSG_MM_GET,
> ETHTOOL_MSG_MM_SET,
> + ETHTOOL_MSG_ULP_DDP_GET,
> + ETHTOOL_MSG_ULP_DDP_SET,
Please add the definition of the command to
Documentation/netlink/specs/ethtool.yaml
> /* add new constants above here */
> __ETHTOOL_MSG_USER_CNT,
> @@ -109,6 +111,8 @@ enum {
> ETHTOOL_MSG_PLCA_NTF,
> ETHTOOL_MSG_MM_GET_REPLY,
> ETHTOOL_MSG_MM_NTF,
> + ETHTOOL_MSG_ULP_DDP_GET_REPLY,
> + ETHTOOL_MSG_ULP_DDP_SET_REPLY,
What about notifications?
> +#include "netlink.h"
> +#include "common.h"
> +#include "bitset.h"
alphabetic order?
> +static int ulp_ddp_stats64_size(unsigned int count)
> +{
> + unsigned int len = 0;
> + unsigned int i;
> +
> + for (i = 0; i < count; i++)
> + len += nla_total_size(sizeof(u64));
len = nla_total_size(sizeof(u64)) * count
?
but it's not correct. You need nla_total_size_64bit() here
> + /* outermost nest */
> + return nla_total_size(len);
nla_total_size(0) is more common for nests.
> +}
> +
> +static int ulp_ddp_put_stats64(struct sk_buff *skb, int attrtype, const u64 *val,
> + unsigned int count)
> +{
> + struct nlattr *nest;
> + unsigned int i;
> +
> + nest = nla_nest_start(skb, attrtype);
> + if (!nest)
> + return -EMSGSIZE;
> +
> + /* skip attribute 0 (unspec) */
> + for (i = 0 ; i < count; i++)
> + if (nla_put_64bit(skb, i+1, sizeof(u64), &val[i], 0))
nla_put_u64_64bit()
And you'll need to add an attr for padding.
> + goto nla_put_failure;
> +
> + nla_nest_end(skb, nest);
> + return 0;
> +
> +nla_put_failure:
> + nla_nest_cancel(skb, nest);
> + return -EMSGSIZE;
> +}
> +const struct nla_policy ethnl_ulp_ddp_set_policy[] = {
> + [ETHTOOL_A_ULP_DDP_HEADER] =
> + NLA_POLICY_NESTED(ethnl_header_policy),
> + [ETHTOOL_A_ULP_DDP_WANTED] = { .type = NLA_NESTED },
> +};
Let's link the policy here: NLA_POLICY_NESTED(bitset_policy).
> +static int ulp_ddp_send_reply(struct net_device *dev, struct genl_info *info,
> + const unsigned long *wanted,
> + const unsigned long *wanted_mask,
> + const unsigned long *active,
> + const unsigned long *active_mask, bool compact)
> +{
> + struct sk_buff *rskb;
> + void *reply_payload;
> + int reply_len = 0;
> + int ret;
> +
> + reply_len = ethnl_reply_header_size();
> + ret = ethnl_bitset_size(wanted, wanted_mask, ULP_DDP_C_COUNT,
> + ulp_ddp_caps_names, compact);
> + if (ret < 0)
> + goto err;
> + reply_len += ret;
> + ret = ethnl_bitset_size(active, active_mask, ULP_DDP_C_COUNT,
> + ulp_ddp_caps_names, compact);
> + if (ret < 0)
> + goto err;
> + reply_len += ret;
> +
> + ret = -ENOMEM;
> + rskb = ethnl_reply_init(reply_len, dev, ETHTOOL_MSG_ULP_DDP_SET_REPLY,
> + ETHTOOL_A_ULP_DDP_HEADER, info,
> + &reply_payload);
> + if (!rskb)
> + goto err;
> +
> + ret = ethnl_put_bitset(rskb, ETHTOOL_A_ULP_DDP_WANTED, wanted,
> + wanted_mask, ULP_DDP_C_COUNT,
> + ulp_ddp_caps_names, compact);
> + if (ret < 0)
> + goto nla_put_failure;
> + ret = ethnl_put_bitset(rskb, ETHTOOL_A_ULP_DDP_ACTIVE, active,
> + active_mask, ULP_DDP_C_COUNT,
> + ulp_ddp_caps_names, compact);
> + if (ret < 0)
> + goto nla_put_failure;
> +
> + genlmsg_end(rskb, reply_payload);
> + ret = genlmsg_reply(rskb, info);
> + return ret;
> +
> +nla_put_failure:
> + nlmsg_free(rskb);
> + WARN_ONCE(1, "calculated message payload length (%d) not sufficient\n",
> + reply_len);
> +err:
> + GENL_SET_ERR_MSG(info, "failed to send reply message");
Don't overwrite the message, the message should be set close to
the error, if needed.
> + return ret;
> +}
> +
> +int ethnl_set_ulp_ddp(struct sk_buff *skb, struct genl_info *info)
> +{
> + DECLARE_BITMAP(old_active, ULP_DDP_C_COUNT);
> + DECLARE_BITMAP(new_active, ULP_DDP_C_COUNT);
> + DECLARE_BITMAP(req_wanted, ULP_DDP_C_COUNT);
> + DECLARE_BITMAP(req_mask, ULP_DDP_C_COUNT);
> + DECLARE_BITMAP(all_bits, ULP_DDP_C_COUNT);
> + DECLARE_BITMAP(tmp, ULP_DDP_C_COUNT);
> + struct ethnl_req_info req_info = {};
> + const struct ulp_ddp_dev_ops *ops;
> + struct nlattr **tb = info->attrs;
> + struct ulp_ddp_netdev_caps *caps;
> + struct net_device *dev;
> + int ret;
> +
> + if (!tb[ETHTOOL_A_ULP_DDP_WANTED])
GENL_REQ_ATTR_CHECK()
> + return -EINVAL;
> + ret = ethnl_parse_header_dev_get(&req_info,
> + tb[ETHTOOL_A_ULP_DDP_HEADER],
> + genl_info_net(info), info->extack,
> + true);
> + if (ret < 0)
> + return ret;
> +
> + dev = req_info.dev;
> + rtnl_lock();
> + caps = netdev_ulp_ddp_caps(dev);
> + ops = netdev_ulp_ddp_ops(dev);
> + if (!caps || !ops || !ops->set_caps) {
> + ret = -EOPNOTSUPP;
> + goto out_rtnl;
> + }
> +
> + ret = ethnl_parse_bitset(req_wanted, req_mask, ULP_DDP_C_COUNT,
> + tb[ETHTOOL_A_ULP_DDP_WANTED],
> + ulp_ddp_caps_names, info->extack);
> + if (ret < 0)
> + goto out_rtnl;
> +
> + /* if (req_mask & ~all_bits) */
> + bitmap_fill(all_bits, ULP_DDP_C_COUNT);
> + bitmap_andnot(tmp, req_mask, all_bits, ULP_DDP_C_COUNT);
> + if (!bitmap_empty(tmp, ULP_DDP_C_COUNT)) {
> + ret = -EINVAL;
> + goto out_rtnl;
> + }
> +
> + /* new_active = (old_active & ~req_mask) | (wanted & req_mask)
> + * new_active &= caps_hw
> + */
> + bitmap_copy(old_active, caps->active, ULP_DDP_C_COUNT);
> + bitmap_and(req_wanted, req_wanted, req_mask, ULP_DDP_C_COUNT);
> + bitmap_andnot(new_active, old_active, req_mask, ULP_DDP_C_COUNT);
> + bitmap_or(new_active, new_active, req_wanted, ULP_DDP_C_COUNT);
> + bitmap_and(new_active, new_active, caps->hw, ULP_DDP_C_COUNT);
> + if (!bitmap_equal(old_active, new_active, ULP_DDP_C_COUNT)) {
> + ret = ops->set_caps(dev, new_active);
We should pass extack to the driver, so that the driver can report a
meaningful error
> + if (ret)
> + netdev_err(dev, "set_ulp_ddp_capabilities() returned error %d\n", ret);
and drop this
> + bitmap_copy(new_active, caps->active, ULP_DDP_C_COUNT);
> + }
> +
> + ret = 0;
> + if (!(req_info.flags & ETHTOOL_FLAG_OMIT_REPLY)) {
> + DECLARE_BITMAP(wanted_diff_mask, ULP_DDP_C_COUNT);
> + DECLARE_BITMAP(active_diff_mask, ULP_DDP_C_COUNT);
> + bool compact = req_info.flags & ETHTOOL_FLAG_COMPACT_BITSETS;
> +
> + /* wanted_diff_mask = req_wanted ^ new_active
> + * active_diff_mask = old_active ^ new_active -> mask of bits that have changed
> + * wanted_diff_mask &= req_mask -> mask of bits that have diff value than wanted
> + * req_wanted &= wanted_diff_mask -> bits that have diff value than wanted
> + * new_active &= active_diff_mask -> bits that have changed
> + */
> + bitmap_xor(wanted_diff_mask, req_wanted, new_active, ULP_DDP_C_COUNT);
> + bitmap_xor(active_diff_mask, old_active, new_active, ULP_DDP_C_COUNT);
> + bitmap_and(wanted_diff_mask, wanted_diff_mask, req_mask, ULP_DDP_C_COUNT);
> + bitmap_and(req_wanted, req_wanted, wanted_diff_mask, ULP_DDP_C_COUNT);
> + bitmap_and(new_active, new_active, active_diff_mask, ULP_DDP_C_COUNT);
> + ret = ulp_ddp_send_reply(dev, info,
> + req_wanted, wanted_diff_mask,
> + new_active, active_diff_mask,
> + compact);
> + }
> +
> +out_rtnl:
> + rtnl_unlock();
> + ethnl_parse_header_dev_put(&req_info);
> + return ret;
> +}
next prev parent reply other threads:[~2023-02-01 4:53 UTC|newest]
Thread overview: 30+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-01-26 16:21 [PATCH v10 00/25] nvme-tcp receive offloads Aurelien Aptel
2023-01-26 16:21 ` [PATCH v10 01/25] net: Introduce direct data placement tcp offload Aurelien Aptel
2023-01-26 16:21 ` [PATCH v10 02/25] net/ethtool: add new stringset ETH_SS_ULP_DDP_{CAPS,STATS} Aurelien Aptel
2023-01-26 16:21 ` [PATCH v10 03/25] net/ethtool: add ULP_DDP_{GET,SET} operations for caps and stats Aurelien Aptel
2023-02-01 4:53 ` Jakub Kicinski [this message]
2023-02-01 19:18 ` Aurelien Aptel
2023-01-26 16:21 ` [PATCH v10 04/25] Documentation: document netlink ULP_DDP_GET/SET messages Aurelien Aptel
2023-01-26 16:21 ` [PATCH v10 05/25] iov_iter: skip copy if src == dst for direct data placement Aurelien Aptel
2023-01-26 16:21 ` [PATCH v10 06/25] net/tls,core: export get_netdev_for_sock Aurelien Aptel
2023-01-26 16:21 ` [PATCH v10 07/25] nvme-tcp: Add DDP offload control path Aurelien Aptel
2023-01-26 16:21 ` [PATCH v10 08/25] nvme-tcp: Add DDP data-path Aurelien Aptel
2023-01-26 16:21 ` [PATCH v10 09/25] nvme-tcp: RX DDGST offload Aurelien Aptel
2023-01-26 16:21 ` [PATCH v10 10/25] nvme-tcp: Deal with netdevice DOWN events Aurelien Aptel
2023-01-26 16:21 ` [PATCH v10 11/25] nvme-tcp: Add modparam to control the ULP offload enablement Aurelien Aptel
2023-01-26 16:21 ` [PATCH v10 12/25] Documentation: add ULP DDP offload documentation Aurelien Aptel
2023-01-26 16:21 ` [PATCH v10 13/25] net/mlx5e: Rename from tls to transport static params Aurelien Aptel
2023-01-26 16:21 ` [PATCH v10 14/25] net/mlx5e: Refactor ico sq polling to get budget Aurelien Aptel
2023-01-26 16:21 ` [PATCH v10 15/25] net/mlx5e: Have mdev pointer directly on the icosq structure Aurelien Aptel
2023-01-26 16:21 ` [PATCH v10 16/25] net/mlx5e: Refactor doorbell function to allow avoiding a completion Aurelien Aptel
2023-01-26 16:21 ` [PATCH v10 17/25] net/mlx5: Add NVMEoTCP caps, HW bits, 128B CQE and enumerations Aurelien Aptel
2023-01-26 16:21 ` [PATCH v10 18/25] net/mlx5e: NVMEoTCP, offload initialization Aurelien Aptel
2023-01-26 16:21 ` [PATCH v10 19/25] net/mlx5e: TCP flow steering for nvme-tcp acceleration Aurelien Aptel
2023-01-26 16:21 ` [PATCH v10 20/25] net/mlx5e: NVMEoTCP, use KLM UMRs for buffer registration Aurelien Aptel
2023-01-26 16:21 ` [PATCH v10 21/25] net/mlx5e: NVMEoTCP, queue init/teardown Aurelien Aptel
2023-01-26 16:21 ` [PATCH v10 22/25] net/mlx5e: NVMEoTCP, ddp setup and resync Aurelien Aptel
2023-01-26 16:21 ` [PATCH v10 23/25] net/mlx5e: NVMEoTCP, async ddp invalidation Aurelien Aptel
2023-01-26 16:21 ` [PATCH v10 24/25] net/mlx5e: NVMEoTCP, data-path for DDP+DDGST offload Aurelien Aptel
2023-01-26 16:21 ` [PATCH v10 25/25] net/mlx5e: NVMEoTCP, statistics Aurelien Aptel
2023-01-30 6:45 ` [PATCH v10 00/25] nvme-tcp receive offloads Christoph Hellwig
2023-02-06 5:12 ` Chaitanya Kulkarni
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20230131205327.67adfc1f@kernel.org \
--to=kuba@kernel.org \
--cc=aaptel@nvidia.com \
--cc=aurelien.aptel@gmail.com \
--cc=axboe@fb.com \
--cc=borisp@nvidia.com \
--cc=chaitanyak@nvidia.com \
--cc=davem@davemloft.net \
--cc=hch@lst.de \
--cc=kbusch@kernel.org \
--cc=linux-nvme@lists.infradead.org \
--cc=malin1024@gmail.com \
--cc=netdev@vger.kernel.org \
--cc=ogerlitz@nvidia.com \
--cc=sagi@grimberg.me \
--cc=smalin@nvidia.com \
--cc=yorayz@nvidia.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).