RE: [net-next, v3, 05/10] ethtool: add a new command for getting PHC virtual clocks
From: "Y.b. Lu" <yangbo.lu@nxp.com>
Date: 2021-06-22 10:10:15
Also in:
linux-kselftest, lkml, mptcp
Hi Jakub,
-----Original Message----- From: Jakub Kicinski <kuba@kernel.org> Sent: 2021年6月16日 3:49 To: Y.b. Lu <yangbo.lu@nxp.com> Cc: netdev@vger.kernel.org; linux-kernel@vger.kernel.org; linux-kselftest@vger.kernel.org; mptcp@lists.linux.dev; Richard Cochran [off-list ref]; David S . Miller [off-list ref]; Mat Martineau [off-list ref]; Matthieu Baerts [off-list ref]; Shuah Khan [off-list ref]; Michal Kubecek [off-list ref]; Florian Fainelli [off-list ref]; Andrew Lunn [off-list ref]; Rui Sousa [off-list ref]; Sebastien Laveze [off-list ref] Subject: Re: [net-next, v3, 05/10] ethtool: add a new command for getting PHC virtual clocks On Tue, 15 Jun 2021 17:45:12 +0800 Yangbo Lu wrote:quoted
Add an interface for getting PHC (PTP Hardware Clock) virtual clocks, which are based on PHC physical clock providing hardware timestamp to network packets. Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>quoted
diff --git a/include/uapi/linux/ethtool.hb/include/uapi/linux/ethtool.h index cfef6b08169a..0fb04f945767 100644--- a/include/uapi/linux/ethtool.h +++ b/include/uapi/linux/ethtool.h@@ -17,6 +17,7 @@ #include <linux/const.h> #include <linux/types.h> #include <linux/if_ether.h> +#include <linux/ptp_clock.h> #ifndef __KERNEL__ #include <limits.h> /* for INT_MAX */ @@ -1341,6 +1342,18 @@ structethtool_ts_info { __u32 rx_reserved[3]; }; +/** + * struct ethtool_phc_vclocks - holds a device's PTP virtual clocks + * @cmd: command number = %ETHTOOL_GET_PHC_VCLOCKS + * @num: number of PTP vclocks + * @index: all index values of PTP vclocks */ struct +ethtool_phc_vclocks { + __u32 cmd; + __u8 num; + __s32 index[PTP_MAX_VCLOCKS]; +}; + /* * %ETHTOOL_SFEATURES changes features present in features[].valid tothequoted
* values of corresponding bits in features[].requested. Bits in .requested @@ -1552,6 +1565,7 @@ enum ethtool_fec_config_bits { #define ETHTOOL_PHY_STUNABLE 0x0000004f /* Set PHY tunableconfiguration */quoted
#define ETHTOOL_GFECPARAM 0x00000050 /* Get FEC settings */ #define ETHTOOL_SFECPARAM 0x00000051 /* Set FEC settings */ +#define ETHTOOL_GET_PHC_VCLOCKS 0x00000052 /* Get PHC virtualclocks info */ We don't add new IOCTL commands, only netlink API is going to be extended. Please remove the IOCTL interface & uAPI.
Will remove. Thanks.
quoted
/* compatibility with older code */ #define SPARC_ETH_GSET ETHTOOL_GSETquoted
+/* PHC VCLOCKS */ + +enum { + ETHTOOL_A_PHC_VCLOCKS_UNSPEC, + ETHTOOL_A_PHC_VCLOCKS_HEADER, /* nest - _A_HEADER_**/quoted
+ ETHTOOL_A_PHC_VCLOCKS_NUM, /* u8 */u32, no need to limit yourself, the netlink attribute is rounded up to 4B anyway.
Get it. Will use u32.
quoted
+ ETHTOOL_A_PHC_VCLOCKS_INDEX, /* s32 */This is an array, AFAICT, not a single s32.
Will fix. Thanks.
quoted
+ + /* add new constants above here */ + __ETHTOOL_A_PHC_VCLOCKS_CNT, + ETHTOOL_A_PHC_VCLOCKS_MAX = (__ETHTOOL_A_PHC_VCLOCKS_CNT -1) };quoted
+ /* CABLE TEST */ enum {quoted
+static int phc_vclocks_fill_reply(struct sk_buff *skb, + const struct ethnl_req_info *req_base, + const struct ethnl_reply_data *reply_base) { + const struct phc_vclocks_reply_data *data = + PHC_VCLOCKS_REPDATA(reply_base); + const struct ethtool_phc_vclocks *phc_vclocks = &data->phc_vclocks; + + if (phc_vclocks->num <= 0) + return 0; + + if (nla_put_u32(skb, ETHTOOL_A_PHC_VCLOCKS_NUM, phc_vclocks->num)||quoted
+ nla_put(skb, ETHTOOL_A_PHC_VCLOCKS_INDEX, + sizeof(phc_vclocks->index), phc_vclocks->index))Looks like you'll report the whole array, why not just num?
Will report just num. Thanks.
quoted
+ return -EMSGSIZE; + + return 0; +} + +const struct ethnl_request_ops ethnl_phc_vclocks_request_ops = { + .request_cmd = ETHTOOL_MSG_PHC_VCLOCKS_GET, + .reply_cmd = ETHTOOL_MSG_PHC_VCLOCKS_GET_REPLY, + .hdr_attr = ETHTOOL_A_PHC_VCLOCKS_HEADER, + .req_info_size = sizeof(struct phc_vclocks_req_info), + .reply_data_size = sizeof(struct phc_vclocks_reply_data), + + .prepare_data = phc_vclocks_prepare_data, + .reply_size = phc_vclocks_reply_size, + .fill_reply = phc_vclocks_fill_reply, +};