Merge branch 'devlink-locking'

Jakub Kicinski says:

====================
improve ethtool/rtnl vs devlink locking

During ethtool netlink development we decided to move some of
the commmands to devlink. Since we don't want drivers to implement
both devlink and ethtool version of the commands ethtool ioctl
falls back to calling devlink. Unfortunately devlink locks must
be taken before rtnl_lock. This results in a questionable
dev_hold() / rtnl_unlock() / devlink / rtnl_lock() / dev_put()
pattern.

This method "works" but it working depends on drivers in question
not doing much in ethtool_ops->begin / complete, and on the netdev
not having needs_free_netdev set.

Since commit 437ebfd90a25 ("devlink: Count struct devlink consumers")
we can hold a reference on a devlink instance and prevent it from
going away (sort of like netdev with dev_hold()). We can use this
to create a more natural reference nesting where we get a ref on
the devlink instance and make the devlink call entirely outside
of the rtnl_lock section.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

1 files changed, 108 insertions, 40 deletions

diff --git a/net/ethtool/ioctl.c b/net/ethtool/ioctl.c
index 44430b6ab843..65e9bc1058b5 100644
--- a/net/ethtool/ioctl.c
+++ b/net/ethtool/ioctl.c
@@ -32,6 +32,29 @@
 #include <generated/utsrelease.h>
 #include "common.h"
 
+/* State held across locks and calls for commands which have devlink fallback */
+struct ethtool_devlink_compat {
+	struct devlink *devlink;
+	union {
+		struct ethtool_flash efl;
+		struct ethtool_drvinfo info;
+	};
+};
+
+static struct devlink *netdev_to_devlink_get(struct net_device *dev)
+{
+	struct devlink_port *devlink_port;
+
+	if (!dev->netdev_ops->ndo_get_devlink_port)
+		return NULL;
+
+	devlink_port = dev->netdev_ops->ndo_get_devlink_port(dev);
+	if (!devlink_port)
+		return NULL;
+
+	return devlink_try_get(devlink_port->devlink);
+}
+
 /*
  * Some useful ethtool_ops methods that're device independent.
  * If we find that all drivers want to do the same thing here,
@@ -697,22 +720,20 @@ static int ethtool_set_settings(struct net_device *dev, void __user *useraddr)
 	return ret;
 }
 
-static noinline_for_stack int ethtool_get_drvinfo(struct net_device *dev,
-						  void __user *useraddr)
+static int
+ethtool_get_drvinfo(struct net_device *dev, struct ethtool_devlink_compat *rsp)
 {
-	struct ethtool_drvinfo info;
 	const struct ethtool_ops *ops = dev->ethtool_ops;
 
-	memset(&info, 0, sizeof(info));
-	info.cmd = ETHTOOL_GDRVINFO;
-	strlcpy(info.version, UTS_RELEASE, sizeof(info.version));
+	rsp->info.cmd = ETHTOOL_GDRVINFO;
+	strlcpy(rsp->info.version, UTS_RELEASE, sizeof(rsp->info.version));
 	if (ops->get_drvinfo) {
-		ops->get_drvinfo(dev, &info);
+		ops->get_drvinfo(dev, &rsp->info);
 	} else if (dev->dev.parent && dev->dev.parent->driver) {
-		strlcpy(info.bus_info, dev_name(dev->dev.parent),
-			sizeof(info.bus_info));
-		strlcpy(info.driver, dev->dev.parent->driver->name,
-			sizeof(info.driver));
+		strlcpy(rsp->info.bus_info, dev_name(dev->dev.parent),
+			sizeof(rsp->info.bus_info));
+		strlcpy(rsp->info.driver, dev->dev.parent->driver->name,
+			sizeof(rsp->info.driver));
 	} else {
 		return -EOPNOTSUPP;
 	}
@@ -726,30 +747,27 @@ static noinline_for_stack int ethtool_get_drvinfo(struct net_device *dev,
 
 		rc = ops->get_sset_count(dev, ETH_SS_TEST);
 		if (rc >= 0)
-			info.testinfo_len = rc;
+			rsp->info.testinfo_len = rc;
 		rc = ops->get_sset_count(dev, ETH_SS_STATS);
 		if (rc >= 0)
-			info.n_stats = rc;
+			rsp->info.n_stats = rc;
 		rc = ops->get_sset_count(dev, ETH_SS_PRIV_FLAGS);
 		if (rc >= 0)
-			info.n_priv_flags = rc;
+			rsp->info.n_priv_flags = rc;
 	}
 	if (ops->get_regs_len) {
 		int ret = ops->get_regs_len(dev);
 
 		if (ret > 0)
-			info.regdump_len = ret;
+			rsp->info.regdump_len = ret;
 	}
 
 	if (ops->get_eeprom_len)
-		info.eedump_len = ops->get_eeprom_len(dev);
+		rsp->info.eedump_len = ops->get_eeprom_len(dev);
 
-	if (!info.fw_version[0])
-		devlink_compat_running_version(dev, info.fw_version,
-					       sizeof(info.fw_version));
+	if (!rsp->info.fw_version[0])
+		rsp->devlink = netdev_to_devlink_get(dev);
 
-	if (copy_to_user(useraddr, &info, sizeof(info)))
-		return -EFAULT;
 	return 0;
 }
 
@@ -2178,19 +2196,15 @@ static int ethtool_set_value(struct net_device *dev, char __user *useraddr,
 	return actor(dev, edata.data);
 }
 
-static noinline_for_stack int ethtool_flash_device(struct net_device *dev,
-						   char __user *useraddr)
+static int
+ethtool_flash_device(struct net_device *dev, struct ethtool_devlink_compat *req)
 {
-	struct ethtool_flash efl;
-
-	if (copy_from_user(&efl, useraddr, sizeof(efl)))
-		return -EFAULT;
-	efl.data[ETHTOOL_FLASH_MAX_FILENAME - 1] = 0;
-
-	if (!dev->ethtool_ops->flash_device)
-		return devlink_compat_flash_update(dev, efl.data);
+	if (!dev->ethtool_ops->flash_device) {
+		req->devlink = netdev_to_devlink_get(dev);
+		return 0;
+	}
 
-	return dev->ethtool_ops->flash_device(dev, &efl);
+	return dev->ethtool_ops->flash_device(dev, &req->efl);
 }
 
 static int ethtool_set_dump(struct net_device *dev,
@@ -2700,19 +2714,19 @@ static int ethtool_set_fecparam(struct net_device *dev, void __user *useraddr)
 
 /* The main entry point in this file.  Called from net/core/dev_ioctl.c */
 
-int dev_ethtool(struct net *net, struct ifreq *ifr, void __user *useraddr)
+static int
+__dev_ethtool(struct net *net, struct ifreq *ifr, void __user *useraddr,
+	      u32 ethcmd, struct ethtool_devlink_compat *devlink_state)
 {
-	struct net_device *dev = __dev_get_by_name(net, ifr->ifr_name);
-	u32 ethcmd, sub_cmd;
+	struct net_device *dev;
+	u32 sub_cmd;
 	int rc;
 	netdev_features_t old_features;
 
+	dev = __dev_get_by_name(net, ifr->ifr_name);
 	if (!dev)
 		return -ENODEV;
 
-	if (copy_from_user(&ethcmd, useraddr, sizeof(ethcmd)))
-		return -EFAULT;
-
 	if (ethcmd == ETHTOOL_PERQUEUE) {
 		if (copy_from_user(&sub_cmd, useraddr + sizeof(ethcmd), sizeof(sub_cmd)))
 			return -EFAULT;
@@ -2786,7 +2800,7 @@ int dev_ethtool(struct net *net, struct ifreq *ifr, void __user *useraddr)
 		rc = ethtool_set_settings(dev, useraddr);
 		break;
 	case ETHTOOL_GDRVINFO:
-		rc = ethtool_get_drvinfo(dev, useraddr);
+		rc = ethtool_get_drvinfo(dev, devlink_state);
 		break;
 	case ETHTOOL_GREGS:
 		rc = ethtool_get_regs(dev, useraddr);
@@ -2888,7 +2902,7 @@ int dev_ethtool(struct net *net, struct ifreq *ifr, void __user *useraddr)
 		rc = ethtool_set_rxnfc(dev, ethcmd, useraddr);
 		break;
 	case ETHTOOL_FLASHDEV:
-		rc = ethtool_flash_device(dev, useraddr);
+		rc = ethtool_flash_device(dev, devlink_state);
 		break;
 	case ETHTOOL_RESET:
 		rc = ethtool_reset(dev, useraddr);
@@ -3000,6 +3014,60 @@ out:
 	return rc;
 }
 
+int dev_ethtool(struct net *net, struct ifreq *ifr, void __user *useraddr)
+{
+	struct ethtool_devlink_compat *state;
+	u32 ethcmd;
+	int rc;
+
+	if (copy_from_user(&ethcmd, useraddr, sizeof(ethcmd)))
+		return -EFAULT;
+
+	state = kzalloc(sizeof(*state), GFP_KERNEL);
+	if (!state)
+		return -ENOMEM;
+
+	switch (ethcmd) {
+	case ETHTOOL_FLASHDEV:
+		if (copy_from_user(&state->efl, useraddr, sizeof(state->efl))) {
+			rc = -EFAULT;
+			goto exit_free;
+		}
+		state->efl.data[ETHTOOL_FLASH_MAX_FILENAME - 1] = 0;
+		break;
+	}
+
+	rtnl_lock();
+	rc = __dev_ethtool(net, ifr, useraddr, ethcmd, state);
+	rtnl_unlock();
+	if (rc)
+		goto exit_free;
+
+	switch (ethcmd) {
+	case ETHTOOL_FLASHDEV:
+		if (state->devlink)
+			rc = devlink_compat_flash_update(state->devlink,
+							 state->efl.data);
+		break;
+	case ETHTOOL_GDRVINFO:
+		if (state->devlink)
+			devlink_compat_running_version(state->devlink,
+						       state->info.fw_version,
+						       sizeof(state->info.fw_version));
+		if (copy_to_user(useraddr, &state->info, sizeof(state->info))) {
+			rc = -EFAULT;
+			goto exit_free;
+		}
+		break;
+	}
+
+exit_free:
+	if (state->devlink)
+		devlink_put(state->devlink);
+	kfree(state);
+	return rc;
+}
+
 struct ethtool_rx_flow_key {
 	struct flow_dissector_key_basic			basic;
 	union {