Skip to contentSkip to conversation

net-sysfs: check device is present when showing paths

PATCH Superseded Jamie Bainbridge 4 participants 10 messages Jul 24, 2024 – Jul 25, 2024

Subsystems (2)

v1 of 4 Superseded Latest is v4 →

lore
Share & tools
lore
#0 lore raw

cover net-sysfs: check device is present when showing paths

A sysfs reader can race with a device reset or removal.

This was fixed for speed_show with commit 4224cfd7fb65 ("net-sysfs: add check for netdevice being present to speed_show") so add the same check to carrier, duplex, testing, and dormant paths.

Submitting as separate patches to make Fixes reporting accurate.

Jamie Bainbridge (4):
net-sysfs: check device is present when showing carrier
net-sysfs: check device is present when showing duplex
net-sysfs: check device is present when showing testing
net-sysfs: check device is present when showing dormant

 net/core/net-sysfs.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

--
2.39.2

#1 lore raw

PATCH 1/4 net-sysfs: check device is present when showing carrier

A sysfs reader can race with a device reset or removal.

This was fixed for speed_show with commit 4224cfd7fb65 ("net-sysfs: add check for netdevice being present to speed_show") so add the same check to carrier_show.

Fixes facd15dfd691 ("net: core: synchronize link-watch when carrier is queried")
---
 net/core/net-sysfs.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
net/core/net-sysfs.c +1 -1
index 0e2084ce7b7572bff458ed7e02358d9258c74628..7fabe5afa3012ecad6551e12f478b755952933b8 100644
@@ -206,7 +206,7 @@ static ssize_t carrier_show(struct device *dev,
 	if (!rtnl_trylock())
 		return restart_syscall();
 
-	if (netif_running(netdev)) {
+	if (netif_running(netdev) && netif_device_present(netdev)) {
 		/* Synchronize carrier state with link watch,
 		 * see also rtnl_getlink().
 		 */
Signature · 1 line
2.39.2
#6 lore raw

PATCH 2/4 net-sysfs: check device is present when showing duplex

In reply to Jamie Bainbridge

A sysfs reader can race with a device reset or removal.

This was fixed for speed_show with commit 4224cfd7fb65 ("net-sysfs: add check for netdevice being present to speed_show") so add the same check to duplex_show.

Fixes 8ae6daca85c8 ("ethtool: Call ethtool's get/set_settings callbacks with cleaned data")
---
 net/core/net-sysfs.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
net/core/net-sysfs.c +1 -1
index 7fabe5afa3012ecad6551e12f478b755952933b8..3a539a2bd4d11c5f5d7b6f15a23d61439f178c3b 100644
@@ -261,7 +261,7 @@ static ssize_t duplex_show(struct device *dev,
 	if (!rtnl_trylock())
 		return restart_syscall();
 
-	if (netif_running(netdev)) {
+	if (netif_running(netdev) && netif_device_present(netdev)) {
 		struct ethtool_link_ksettings cmd;
 
 		if (!__ethtool_get_link_ksettings(netdev, &cmd)) {
Signature · 1 line
2.39.2
#7 lore raw

PATCH 3/4 net-sysfs: check device is present when showing testing

In reply to Jamie Bainbridge

A sysfs reader can race with a device reset or removal.

This was fixed for speed_show with commit 4224cfd7fb65 ("net-sysfs: add check for netdevice being present to speed_show") so add the same check to testing_show.

Fixes db30a57779b1 ("net: Add testing sysfs attribute")
---
 net/core/net-sysfs.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
net/core/net-sysfs.c +1 -1
index 3a539a2bd4d11c5f5d7b6f15a23d61439f178c3b..17927832a4fbb56d3e1dfbed29c567d70ab944be 100644
@@ -291,7 +291,7 @@ static ssize_t testing_show(struct device *dev,
 {
 	struct net_device *netdev = to_net_dev(dev);
 
-	if (netif_running(netdev))
+	if (netif_running(netdev) && netif_device_present(netdev))
 		return sysfs_emit(buf, fmt_dec, !!netif_testing(netdev));
 
 	return -EINVAL;
Signature · 1 line
2.39.2
#9 lore raw

PATCH 4/4 net-sysfs: check device is present when showing dormant

In reply to Jamie Bainbridge

A sysfs reader can race with a device reset or removal.

This was fixed for speed_show with commit 4224cfd7fb65 ("net-sysfs: add check for netdevice being present to speed_show") so add the same check to dormant_show.

Fixes b00055aacdb1 ("[NET] core: add RFC2863 operstate")
---
 net/core/net-sysfs.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
net/core/net-sysfs.c +1 -1
index 17927832a4fbb56d3e1dfbed29c567d70ab944be..ff2a1f6ef7e18be56c2de51902519066431e47a8 100644
@@ -303,7 +303,7 @@ static ssize_t dormant_show(struct device *dev,
 {
 	struct net_device *netdev = to_net_dev(dev);
 
-	if (netif_running(netdev))
+	if (netif_running(netdev) && netif_device_present(netdev))
 		return sysfs_emit(buf, fmt_dec, !!netif_dormant(netdev));
 
 	return -EINVAL;
Signature · 1 line
2.39.2
#2 lore raw

Re: [PATCH net 1/4] net-sysfs: check device is present when showing carrier

In reply to Jamie Bainbridge
Quoted 1 line · Jamie Bainbridge

A sysfs reader can race with a device reset or removal.

Kind of, yes, but please check what the race actually is.

Quoted 3 lines

This was fixed for speed_show with commit 4224cfd7fb65 ("net-sysfs: add check for netdevice being present to speed_show") so add the same check to carrier_show.

You didn't say why it's needed here, so ... why is it?

FWIW, I don't think it actually _is_ needed, since the netdev struct itself is still around, linkwatch_sync_dev() will not do anything that's not still needed anyway (the removal from list must clearly either still happen or nothing happens in the function). This will not call into the driver (which would be the problematic part).

So while I don't think this is _wrong_ per se, I also don't think it's necessary, nor are you demonstrating that it is.

And for userspace it should be pretty much immaterial whether it gets a real value or -EINVAL in the race, or -ENOENT because the file disappeared anyway?

johannes

#3 lore raw

Re: [PATCH net 1/4] net-sysfs: check device is present when showing carrier

Quoted 17 lines · Johannes Berg
Quoted 1 line · Jamie Bainbridge

A sysfs reader can race with a device reset or removal.

Kind of, yes, but please check what the race actually is.

Quoted 3 lines

This was fixed for speed_show with commit 4224cfd7fb65 ("net-sysfs: add check for netdevice being present to speed_show") so add the same check to carrier_show.

You didn't say why it's needed here, so ... why is it?

FWIW, I don't think it actually _is_ needed, since the netdev struct itself is still around, linkwatch_sync_dev() will not do anything that's not still needed anyway (the removal from list must clearly either still happen or nothing happens in the function). This will not call into the driver (which would be the problematic part).

Quoted 7 lines

So while I don't think this is _wrong_ per se, I also don't think it's necessary, nor are you demonstrating that it is.

And for userspace it should be pretty much immaterial whether it gets a real value or -EINVAL in the race, or -ENOENT because the file disappeared anyway?

All of which, btw, is also true for patches 3 and 4 in this set.

For patch 2 it seems applicable.

I do wonder if ethtool itself, at least ethtool netlink, doesn't have a similar problem though, since it just uses netdev_get_by_name() / netdev_get_by_index()?

johannes

#8 lore raw

Re: [PATCH net 3/4] net-sysfs: check device is present when showing testing

In reply to Jamie Bainbridge
Quoted 18 lines · Jamie Bainbridge

A sysfs reader can race with a device reset or removal.

This was fixed for speed_show with commit 4224cfd7fb65 ("net-sysfs: add check for netdevice being present to speed_show") so add the same check to testing_show.

Fixes db30a57779b1 ("net: Add testing sysfs attribute")
---
 net/core/net-sysfs.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
net/core/net-sysfs.c +0 -0
index 3a539a2bd4d11c5f5d7b6f15a23d61439f178c3b..17927832a4fbb56d3e1dfbed29c567d70ab944be 100644

net/core/net-sysfs.c static ssize_t testing_show(struct device *dev, · line 294

Quoted 5 lines
net/core/net-sysfs.c +1 -1
@@ -291,7 +291,7 @@ static ssize_t testing_show(struct device *dev,
 {
 	struct net_device *netdev = to_net_dev(dev);
 
-	if (netif_running(netdev))
+	if (netif_running(netdev) && netif_device_present(netdev))

Maybe a dumb observation, but how can it be running if it is not present?

And if we are not holding any locks, it might be gone by the time you call netif_testing()?

net/core/net-sysfs.c static ssize_t testing_show(struct device *dev, · line 295

Quoted 1 line
net/core/net-sysfs.c +0 -0
 		return sysfs_emit(buf, fmt_dec, !!netif_testing(netdev));

Andrew

#4 lore raw

Re: [PATCH net 1/4] net-sysfs: check device is present when showing carrier

In reply to Johannes Berg
Quoted 29 lines · Johannes Berg
Quoted 17 lines · Johannes Berg
Quoted 1 line · Jamie Bainbridge

A sysfs reader can race with a device reset or removal.

Kind of, yes, but please check what the race actually is.

Quoted 3 lines

This was fixed for speed_show with commit 4224cfd7fb65 ("net-sysfs: add check for netdevice being present to speed_show") so add the same check to carrier_show.

You didn't say why it's needed here, so ... why is it?

FWIW, I don't think it actually _is_ needed, since the netdev struct itself is still around, linkwatch_sync_dev() will not do anything that's not still needed anyway (the removal from list must clearly either still happen or nothing happens in the function). This will not call into the driver (which would be the problematic part).

Quoted 7 lines

So while I don't think this is _wrong_ per se, I also don't think it's necessary, nor are you demonstrating that it is.

And for userspace it should be pretty much immaterial whether it gets a real value or -EINVAL in the race, or -ENOENT because the file disappeared anyway?

All of which, btw, is also true for patches 3 and 4 in this set.

Quoted 7 lines

For patch 2 it seems applicable.

I do wonder if ethtool itself, at least ethtool netlink, doesn't have a similar problem though, since it just uses netdev_get_by_name() / netdev_get_by_index()?

johannes

You are correct, patch 2 (duplex) is the one where we panicked during device reset. I thought to fix the other "show" functions in advance while I was there.

I will revise this and re-submit with only the necessary patch.

Thanks for the review, it is appreciated.

Jamie

#5 lore raw

Re: [PATCH net 1/4] net-sysfs: check device is present when showing carrier

In reply to Jamie Bainbridge

Hi Jamie,

Quoted 9 lines · Jamie Bainbridge

A sysfs reader can race with a device reset or removal.

This was fixed for speed_show with commit 4224cfd7fb65 ("net-sysfs: add check for netdevice being present to speed_show") so add the same check to carrier_show.

Fixes facd15dfd691 ("net: core: synchronize link-watch when carrier is queried")

nit: When resubmitting your fix, there is no need for a blank line
between the Fixes and SOB tags.

Thanks,
Shigeru

Quoted 21 lines
---
 net/core/net-sysfs.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
net/core/net-sysfs.c +1 -1
index 0e2084ce7b7572bff458ed7e02358d9258c74628..7fabe5afa3012ecad6551e12f478b755952933b8 100644
@@ -206,7 +206,7 @@ static ssize_t carrier_show(struct device *dev,
 	if (!rtnl_trylock())
 		return restart_syscall();
 
-	if (netif_running(netdev)) {
+	if (netif_running(netdev) && netif_device_present(netdev)) {
 		/* Synchronize carrier state with link watch,
 		 * see also rtnl_getlink().
 		 */
Signature · 1 line
2.39.2