Fix the netlink type for hardware timestamp flags, which are represented
as a bitset of flags. Although only one flag is supported currently, the
correct netlink bitset type should be used instead of u32. Address this
by adding a new named string set description for the hwtstamp flag
structure.

Signed-off-by: Kory Maincent <kory.maincent@bootlin.com>
Fixes: 6e9e2eed4f39 ("net: ethtool: Add support for tsconfig command to get/set hwtstamp config")
---
 Documentation/netlink/specs/ethtool.yaml |  3 ++-
 include/uapi/linux/ethtool.h             |  2 ++
 net/ethtool/common.c                     |  5 +++++
 net/ethtool/common.h                     |  2 ++
 net/ethtool/strset.c                     |  5 +++++
 net/ethtool/tsconfig.c                   | 33 ++++++++++++++++++++++----------
 6 files changed, 39 insertions(+), 11 deletions(-)

diff --git a/Documentation/netlink/specs/ethtool.yaml b/Documentation/netlink/specs/ethtool.yaml
index 259cb211a338..655d8d10fe24 100644
--- a/Documentation/netlink/specs/ethtool.yaml
+++ b/Documentation/netlink/specs/ethtool.yaml
@@ -1524,7 +1524,8 @@ attribute-sets:
         nested-attributes: bitset
       -
         name: hwtstamp-flags
-        type: u32
+        type: nest
+        nested-attributes: bitset
 
 operations:
   enum-model: directional
diff --git a/include/uapi/linux/ethtool.h b/include/uapi/linux/ethtool.h
index d1089b88efc7..9b18c4cfe56f 100644
--- a/include/uapi/linux/ethtool.h
+++ b/include/uapi/linux/ethtool.h
@@ -682,6 +682,7 @@ enum ethtool_link_ext_substate_module {
  * @ETH_SS_STATS_ETH_CTRL: names of IEEE 802.3 MAC Control statistics
  * @ETH_SS_STATS_RMON: names of RMON statistics
  * @ETH_SS_STATS_PHY: names of PHY(dev) statistics
+ * @ETH_SS_TS_FLAGS: hardware timestamping flags
  *
  * @ETH_SS_COUNT: number of defined string sets
  */
@@ -708,6 +709,7 @@ enum ethtool_stringset {
 	ETH_SS_STATS_ETH_CTRL,
 	ETH_SS_STATS_RMON,
 	ETH_SS_STATS_PHY,
+	ETH_SS_TS_FLAGS,
 
 	/* add new constants above here */
 	ETH_SS_COUNT
diff --git a/net/ethtool/common.c b/net/ethtool/common.c
index 2bd77c94f9f1..d88e9080643b 100644
--- a/net/ethtool/common.c
+++ b/net/ethtool/common.c
@@ -462,6 +462,11 @@ const char ts_rx_filter_names[][ETH_GSTRING_LEN] = {
 };
 static_assert(ARRAY_SIZE(ts_rx_filter_names) == __HWTSTAMP_FILTER_CNT);
 
+const char ts_flags_names[][ETH_GSTRING_LEN] = {
+	[const_ilog2(HWTSTAMP_FLAG_BONDED_PHC_INDEX)] = "bonded-phc-index",
+};
+static_assert(ARRAY_SIZE(ts_flags_names) == __HWTSTAMP_FLAG_CNT);
+
 const char udp_tunnel_type_names[][ETH_GSTRING_LEN] = {
 	[ETHTOOL_UDP_TUNNEL_TYPE_VXLAN]		= "vxlan",
 	[ETHTOOL_UDP_TUNNEL_TYPE_GENEVE]	= "geneve",
diff --git a/net/ethtool/common.h b/net/ethtool/common.h
index 850eadde4bfc..58e9e7db06f9 100644
--- a/net/ethtool/common.h
+++ b/net/ethtool/common.h
@@ -13,6 +13,7 @@
 	ETHTOOL_LINK_MODE_ ## speed ## base ## type ## _ ## duplex ## _BIT
 
 #define __SOF_TIMESTAMPING_CNT (const_ilog2(SOF_TIMESTAMPING_LAST) + 1)
+#define __HWTSTAMP_FLAG_CNT (const_ilog2(HWTSTAMP_FLAG_LAST) + 1)
 
 struct link_mode_info {
 	int				speed;
@@ -38,6 +39,7 @@ extern const char wol_mode_names[][ETH_GSTRING_LEN];
 extern const char sof_timestamping_names[][ETH_GSTRING_LEN];
 extern const char ts_tx_type_names[][ETH_GSTRING_LEN];
 extern const char ts_rx_filter_names[][ETH_GSTRING_LEN];
+extern const char ts_flags_names[][ETH_GSTRING_LEN];
 extern const char udp_tunnel_type_names[][ETH_GSTRING_LEN];
 
 int __ethtool_get_link(struct net_device *dev);
diff --git a/net/ethtool/strset.c b/net/ethtool/strset.c
index 818cf01f0911..6b76c05caba4 100644
--- a/net/ethtool/strset.c
+++ b/net/ethtool/strset.c
@@ -75,6 +75,11 @@ static const struct strset_info info_template[] = {
 		.count		= __HWTSTAMP_FILTER_CNT,
 		.strings	= ts_rx_filter_names,
 	},
+	[ETH_SS_TS_FLAGS] = {
+		.per_dev	= false,
+		.count		= __HWTSTAMP_FLAG_CNT,
+		.strings	= ts_flags_names,
+	},
 	[ETH_SS_UDP_TUNNEL_TYPES] = {
 		.per_dev	= false,
 		.count		= __ETHTOOL_UDP_TUNNEL_TYPE_CNT,
diff --git a/net/ethtool/tsconfig.c b/net/ethtool/tsconfig.c
index 2dc3a3b76615..4e8fd4f5c27b 100644
--- a/net/ethtool/tsconfig.c
+++ b/net/ethtool/tsconfig.c
@@ -54,7 +54,7 @@ static int tsconfig_prepare_data(const struct ethnl_req_info *req_base,
 
 	data->hwtst_config.tx_type = BIT(cfg.tx_type);
 	data->hwtst_config.rx_filter = BIT(cfg.rx_filter);
-	data->hwtst_config.flags = BIT(cfg.flags);
+	data->hwtst_config.flags = cfg.flags;
 
 	data->hwprov_desc.index = -1;
 	hwprov = rtnl_dereference(dev->hwprov);
@@ -91,10 +91,16 @@ static int tsconfig_reply_size(const struct ethnl_req_info *req_base,
 
 	BUILD_BUG_ON(__HWTSTAMP_TX_CNT > 32);
 	BUILD_BUG_ON(__HWTSTAMP_FILTER_CNT > 32);
+	BUILD_BUG_ON(__HWTSTAMP_FLAG_CNT > 32);
 
-	if (data->hwtst_config.flags)
-		/* _TSCONFIG_HWTSTAMP_FLAGS */
-		len += nla_total_size(sizeof(u32));
+	if (data->hwtst_config.flags) {
+		ret = ethnl_bitset32_size(&data->hwtst_config.flags,
+					  NULL, __HWTSTAMP_FLAG_CNT,
+					  ts_flags_names, compact);
+		if (ret < 0)
+			return ret;
+		len += ret;	/* _TSCONFIG_HWTSTAMP_FLAGS */
+	}
 
 	if (data->hwtst_config.tx_type) {
 		ret = ethnl_bitset32_size(&data->hwtst_config.tx_type,
@@ -130,8 +136,10 @@ static int tsconfig_fill_reply(struct sk_buff *skb,
 	int ret;
 
 	if (data->hwtst_config.flags) {
-		ret = nla_put_u32(skb, ETHTOOL_A_TSCONFIG_HWTSTAMP_FLAGS,
-				  data->hwtst_config.flags);
+		ret = ethnl_put_bitset32(skb, ETHTOOL_A_TSCONFIG_HWTSTAMP_FLAGS,
+					 &data->hwtst_config.flags, NULL,
+					 __HWTSTAMP_FLAG_CNT,
+					 ts_flags_names, compact);
 		if (ret < 0)
 			return ret;
 	}
@@ -180,7 +188,7 @@ const struct nla_policy ethnl_tsconfig_set_policy[ETHTOOL_A_TSCONFIG_MAX + 1] =
 	[ETHTOOL_A_TSCONFIG_HEADER] = NLA_POLICY_NESTED(ethnl_header_policy),
 	[ETHTOOL_A_TSCONFIG_HWTSTAMP_PROVIDER] =
 		NLA_POLICY_NESTED(ethnl_ts_hwtst_prov_policy),
-	[ETHTOOL_A_TSCONFIG_HWTSTAMP_FLAGS] = { .type = NLA_U32 },
+	[ETHTOOL_A_TSCONFIG_HWTSTAMP_FLAGS] = { .type = NLA_NESTED },
 	[ETHTOOL_A_TSCONFIG_RX_FILTERS] = { .type = NLA_NESTED },
 	[ETHTOOL_A_TSCONFIG_TX_TYPES] = { .type = NLA_NESTED },
 };
@@ -296,6 +304,7 @@ static int ethnl_set_tsconfig(struct ethnl_req_info *req_base,
 
 	BUILD_BUG_ON(__HWTSTAMP_TX_CNT > 32);
 	BUILD_BUG_ON(__HWTSTAMP_FILTER_CNT > 32);
+	BUILD_BUG_ON(__HWTSTAMP_FLAG_CNT > 32);
 
 	if (!netif_device_present(dev))
 		return -ENODEV;
@@ -377,9 +386,13 @@ static int ethnl_set_tsconfig(struct ethnl_req_info *req_base,
 	}
 
 	if (tb[ETHTOOL_A_TSCONFIG_HWTSTAMP_FLAGS]) {
-		ethnl_update_u32(&hwtst_config.flags,
-				 tb[ETHTOOL_A_TSCONFIG_HWTSTAMP_FLAGS],
-				 &config_mod);
+		ret = ethnl_update_bitset32(&hwtst_config.flags,
+					    __HWTSTAMP_FLAG_CNT,
+					    tb[ETHTOOL_A_TSCONFIG_HWTSTAMP_FLAGS],
+					    ts_flags_names, info->extack,
+					    &config_mod);
+		if (ret < 0)
+			goto err_free_hwprov;
 	}
 
 	ret = net_hwtstamp_validate(&hwtst_config);

-- 
2.34.1

On Tue, 28 Jan 2025 16:35:48 +0100 Kory Maincent wrote:
> Fix the netlink type for hardware timestamp flags, which are represented
> as a bitset of flags. Although only one flag is supported currently, the
> correct netlink bitset type should be used instead of u32. Address this
> by adding a new named string set description for the hwtstamp flag
> structure.

Makes sense, please mention explicitly in the commit message that the
code has been introduced in the current release so the uAPI change is
still okay.

In general IMHO YNL makes the bitset functionality less important.
But in this case consistency with other fields seems worth it.
The patch LGTM.

On Wed, 29 Jan 2025 16:49:07 -0800
Jakub Kicinski <kuba@kernel.org> wrote:

> On Tue, 28 Jan 2025 16:35:48 +0100 Kory Maincent wrote:
> > Fix the netlink type for hardware timestamp flags, which are represented
> > as a bitset of flags. Although only one flag is supported currently, the
> > correct netlink bitset type should be used instead of u32. Address this
> > by adding a new named string set description for the hwtstamp flag
> > structure.  
> 
> Makes sense, please mention explicitly in the commit message that the
> code has been introduced in the current release so the uAPI change is
> still okay.

Ack.

> In general IMHO YNL makes the bitset functionality less important.

Do you mean you prefer u32 for bitfield instead of the bitset type? Why?

-- 
Köry Maincent, Bootlin
Embedded Linux and kernel engineering
https://bootlin.com

On Thu, 30 Jan 2025 10:27:16 +0100 Kory Maincent wrote:
> > In general IMHO YNL makes the bitset functionality less important.  
> 
> Do you mean you prefer u32 for bitfield instead of the bitset type? Why?

Not in this case, but in general uint can carry up to 64 bits, 
and the names of the fields are in the spec (both Python and 
C code automatically decode them based on the spec). I see less
of a need for the complex "forward compatibility" handling
in the kernel because of that.