Skip to contentSkip to conversation

net: make for_each_netdev_dump() a little more bug-proof

PATCH Applied Jakub Kicinski 3 participants 5 messages Jun 13, 2024 – Jun 17, 2024

Subsystems (2)
lore
Share & tools
lore

Jakub Kicinski

author maintainer
#0 lore raw

PATCH 1/1 net: make for_each_netdev_dump() a little more bug-proof

I find the behavior of xa_for_each_start() slightly counter-intuitive. It doesn't end the iteration by making the index point after the last element. IOW calling xa_for_each_start() again after it "finished" will run the body of the loop for the last valid element, instead of doing nothing.

This works fine for netlink dumps if they terminate correctly (i.e. coalesce or carefully handle NLM_DONE), but as we keep getting reminded legacy dumps are unlikely to go away.

Fixing this generically at the xa_for_each_start() level seems hard - there is no index reserved for "end of iteration". ifindexes are 31b wide, tho, and iterator is ulong so for for_each_netdev_dump() it's safe to go to the next element.

---
 include/linux/netdevice.h | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
include/linux/netdevice.h +2 -1
index f148a01dd1d1..85111502cf8f 100644
@@ -3021,7 +3021,8 @@ int call_netdevice_notifiers_info(unsigned long val,
 #define net_device_entry(lh)	list_entry(lh, struct net_device, dev_list)
 
 #define for_each_netdev_dump(net, d, ifindex)				\
-	xa_for_each_start(&(net)->dev_by_index, (ifindex), (d), (ifindex))
+	for (; (d = xa_find(&(net)->dev_by_index, &ifindex,		\
+			    ULONG_MAX, XA_PRESENT)); ifindex++)
 
 static inline struct net_device *next_net_device(struct net_device *dev)
 {
Signature · 1 line
2.45.2
#1 lore raw
Quoted 6 lines · Jakub Kicinski

I find the behavior of xa_for_each_start() slightly counter-intuitive. It doesn't end the iteration by making the index point after the last element. IOW calling xa_for_each_start() again after it "finished" will run the body of the loop for the last valid element, instead of doing nothing.

Quoted 7 lines

This works fine for netlink dumps if they terminate correctly (i.e. coalesce or carefully handle NLM_DONE), but as we keep getting reminded legacy dumps are unlikely to go away.

Fixing this generically at the xa_for_each_start() level seems hard - there is no index reserved for "end of iteration". ifindexes are 31b wide,

you are right, it would be easier to go one step [of macros] up, we have

 453│ #define xa_for_each_range(xa, index, entry, start, last)         \
 454│         for (index = start,                                      \
 455│              entry = xa_find(xa, &index, last, XA_PRESENT);      \
 456│              entry;                                              \
 457│              entry = xa_find_after(xa, &index, last, XA_PRESENT))

You could simply change L456 to:
entry || (index = 0);
to achieve what you want; but that would slow down a little lot's of places, only to change value of the index that should not be used :P

For me a proper solution would be to fast forward into C11 era, and move @index allocation into the loop, so value could not be abused.

but that is a lot of usage code, so I'm not against your current patch

Quoted 22 lines

tho, and iterator is ulong so for for_each_netdev_dump() it's safe to go to the next element.

---
  include/linux/netdevice.h | 3 ++-
  1 file changed, 2 insertions(+), 1 deletion(-)
include/linux/netdevice.h +2 -1
index f148a01dd1d1..85111502cf8f 100644
@@ -3021,7 +3021,8 @@ int call_netdevice_notifiers_info(unsigned long val,
  #define net_device_entry(lh)	list_entry(lh, struct net_device, dev_list)
 
  #define for_each_netdev_dump(net, d, ifindex)				\
-	xa_for_each_start(&(net)->dev_by_index, (ifindex), (d), (ifindex))
+	for (; (d = xa_find(&(net)->dev_by_index, &ifindex,		\
+			    ULONG_MAX, XA_PRESENT)); ifindex++)
 
  static inline struct net_device *next_net_device(struct net_device *dev)
  {
author maintainer
#2 lore raw
Quoted 7 lines · On Fri, 14 Jun 2024 03:45:47 +0200 Przemek Kitszel wrote:

you are right, it would be easier to go one step [of macros] up, we have

  453│ #define xa_for_each_range(xa, index, entry, start, last)         \
  454│         for (index = start,                                      \
  455│              entry = xa_find(xa, &index, last, XA_PRESENT);      \
  456│              entry;                                              \
  457│              entry = xa_find_after(xa, &index, last, XA_PRESENT))
Quoted 7 lines

You could simply change L456 to:
entry || (index = 0);
to achieve what you want; but that would slow down a little lot's of places, only to change value of the index that should not be used :P

For me a proper solution would be to fast forward into C11 era, and move @index allocation into the loop, so value could not be abused.

I think we're already in C11 era for that exact reason 🤐️ But please don't take this as an invitation to do crazy things!

In netlink dumps, tho, we keep the index in persistent socket context, because the iteration is split into multiple recvmsg() calls, each one continuing where the previous one left off.

Reviewed-by on v1 1/1
#3 lore raw
In reply to Jakub Kicinski
Quoted 25 lines · Jakub Kicinski

I find the behavior of xa_for_each_start() slightly counter-intuitive. It doesn't end the iteration by making the index point after the last element. IOW calling xa_for_each_start() again after it "finished" will run the body of the loop for the last valid element, instead of doing nothing.

This works fine for netlink dumps if they terminate correctly (i.e. coalesce or carefully handle NLM_DONE), but as we keep getting reminded legacy dumps are unlikely to go away.

Fixing this generically at the xa_for_each_start() level seems hard - there is no index reserved for "end of iteration". ifindexes are 31b wide, tho, and iterator is ulong so for for_each_netdev_dump() it's safe to go to the next element.

---
  include/linux/netdevice.h | 3 ++-
  1 file changed, 2 insertions(+), 1 deletion(-)
include/linux/netdevice.h +0 -0
index f148a01dd1d1..85111502cf8f 100644

include/linux/netdevice.h int call_netdevice_notifiers_info(unsigned long val, · line 3028

Quoted 9 lines
include/linux/netdevice.h +2 -1
@@ -3021,7 +3021,8 @@ int call_netdevice_notifiers_info(unsigned long val,
  #define net_device_entry(lh)	list_entry(lh, struct net_device, dev_list)
 
  #define for_each_netdev_dump(net, d, ifindex)				\
-	xa_for_each_start(&(net)->dev_by_index, (ifindex), (d), (ifindex))
+	for (; (d = xa_find(&(net)->dev_by_index, &ifindex,		\
+			    ULONG_MAX, XA_PRESENT)); ifindex++)
 
  static inline struct net_device *next_net_device(struct net_device *dev)
  {

Makes sense,

#4 lore raw
In reply to Jakub Kicinski
This patch was applied to netdev/net-next.git (main) by David S. Miller <[email protected]>:

Hello:

This patch was applied to netdev/net-next.git (main) by David S. Miller <[email protected]>:

Quoted 11 lines · On Thu, 13 Jun 2024 14:33:16 -0700 you wrote:

I find the behavior of xa_for_each_start() slightly counter-intuitive. It doesn't end the iteration by making the index point after the last element. IOW calling xa_for_each_start() again after it "finished" will run the body of the loop for the last valid element, instead of doing nothing.

This works fine for netlink dumps if they terminate correctly (i.e. coalesce or carefully handle NLM_DONE), but as we keep getting reminded legacy dumps are unlikely to go away.

[...]

Here is the summary with links:
- [net-next] net: make for_each_netdev_dump() a little more bug-proof

    https://git.kernel.org/netdev/net-next/c/f22b4b55edb5

You are awesome, thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html