linux.git/net/core/netpoll.c, branch v6.12

move asm/unaligned.h to linux/unaligned.h

2024-10-02T21:23:23+00:00

asm/unaligned.h is always an include of asm-generic/unaligned.h;
might as well move that thing to linux/unaligned.h and include
that - there's nothing arch-specific in that header.

auto-generated by the following:

for i in `git grep -l -w asm/unaligned.h`; do
	sed -i -e "s/asm\/unaligned.h/linux\/unaligned.h/" $i
done
for i in `git grep -l -w asm-generic/unaligned.h`; do
	sed -i -e "s/asm-generic\/unaligned.h/linux\/unaligned.h/" $i
done
git mv include/asm-generic/unaligned.h include/linux/unaligned.h
git mv tools/include/asm-generic/unaligned.h tools/include/linux/unaligned.h
sed -i -e "/unaligned.h/d" include/asm-generic/Kbuild
sed -i -e "s/__ASM_GENERIC/__LINUX/" include/linux/unaligned.h tools/include/linux/unaligned.h

netpoll: remove netpoll_srcu

2024-09-07T01:24:59+00:00

netpoll_srcu is currently used from netpoll_poll_disable() and
__netpoll_cleanup()

Both functions run under RTNL, using netpoll_srcu adds confusion
and no additional protection.

Moreover the synchronize_srcu() call in __netpoll_cleanup() is
performed before clearing np->dev->npinfo, which violates RCU rules.

After this patch, netpoll_poll_disable() and netpoll_poll_enable()
simply use rtnl_dereference().

This saves a big chunk of memory (more than 192KB on platforms
with 512 cpus)

Signed-off-by: Eric Dumazet 
Reviewed-by: Breno Leitao 
Link: https://patch.msgid.link/20240905084909.2082486-1-edumazet@google.com
Signed-off-by: Jakub Kicinski

net: Correct spelling in net/core

2024-08-26T16:37:23+00:00

Correct spelling in net/core.
As reported by codespell.

Signed-off-by: Simon Horman 
Link: https://patch.msgid.link/20240822-net-spell-v1-13-3a98971ce2d2@kernel.org
Signed-off-by: Jakub Kicinski

netpoll: Ensure clean state on setup failures

2024-08-26T16:25:44+00:00

Modify netpoll_setup() and __netpoll_setup() to ensure that the netpoll
structure (np) is left in a clean state if setup fails for any reason.
This prevents carrying over misconfigured fields in case of partial
setup success.

Key changes:
- np->dev is now set only after successful setup, ensuring it's always
  NULL if netpoll is not configured or if netpoll_setup() fails.
- np->local_ip is zeroed if netpoll setup doesn't complete successfully.
- Added DEBUG_NET_WARN_ON_ONCE() checks to catch unexpected states.
- Reordered some operations in __netpoll_setup() for better logical flow.

These changes improve the reliability of netpoll configuration, since it
assures that the structure is fully initialized or totally unset.

Suggested-by: Paolo Abeni 
Signed-off-by: Breno Leitao 
Link: https://patch.msgid.link/20240822111051.179850-2-leitao@debian.org
Signed-off-by: Jakub Kicinski

Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net

2024-08-23T00:06:18+00:00

Cross-merge networking fixes after downstream PR.

No conflicts.

Adjacent changes:

drivers/net/ethernet/broadcom/bnxt/bnxt.h
  c948c0973df5 ("bnxt_en: Don't clear ntuple filters and rss contexts during ethtool ops")
  f2878cdeb754 ("bnxt_en: Add support to call FW to update a VNIC")

Link: https://patch.msgid.link/20240822210125.1542769-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski

netpoll: do not export netpoll_poll_[disable|enable]()

2024-08-22T00:31:28+00:00

netpoll_poll_disable() and netpoll_poll_enable() are only used
from core networking code, there is no need to export them.

Signed-off-by: Eric Dumazet 
Reviewed-by: Simon Horman 
Link: https://patch.msgid.link/20240820162053.3870927-1-edumazet@google.com
Signed-off-by: Jakub Kicinski

net: netpoll: extract core of netpoll_cleanup

2024-08-13T08:58:58+00:00

Extract the core part of netpoll_cleanup(), so, it could be called from
a caller that has the rtnl lock already.

Netconsole uses this in a weird way right now:

	__netpoll_cleanup(&nt->np);
	spin_lock_irqsave(&target_list_lock, flags);
	netdev_put(nt->np.dev, &nt->np.dev_tracker);
	nt->np.dev = NULL;
	nt->enabled = false;

This will be replaced by do_netpoll_cleanup() as the locking situation
is overhauled.

Signed-off-by: Breno Leitao 
Reviewed-by: Rik van Riel 
Signed-off-by: Paolo Abeni

netpoll: Fix race condition in netpoll_owner_active

2024-05-01T02:03:47+00:00

KCSAN detected a race condition in netpoll:

	BUG: KCSAN: data-race in net_rx_action / netpoll_send_skb
	write (marked) to 0xffff8881164168b0 of 4 bytes by interrupt on cpu 10:
	net_rx_action (./include/linux/netpoll.h:90 net/core/dev.c:6712 net/core/dev.c:6822)

	read to 0xffff8881164168b0 of 4 bytes by task 1 on cpu 2:
	netpoll_send_skb (net/core/netpoll.c:319 net/core/netpoll.c:345 net/core/netpoll.c:393)
	netpoll_send_udp (net/core/netpoll.c:?)

	value changed: 0x0000000a -> 0xffffffff

This happens because netpoll_owner_active() needs to check if the
current CPU is the owner of the lock, touching napi->poll_owner
non atomically. The ->poll_owner field contains the current CPU holding
the lock.

Use an atomic read to check if the poll owner is the current CPU.

Signed-off-by: Breno Leitao 
Link: https://lore.kernel.org/r/20240429100437.3487432-1-leitao@debian.org
Signed-off-by: Jakub Kicinski

netpoll: allocate netdev tracker right away

2023-06-15T07:21:11+00:00

Commit 5fa5ae605821 ("netpoll: add net device refcount tracker to struct netpoll")
was part of one of the initial netdev tracker introduction patches.
It added an explicit netdev_tracker_alloc() for netpoll, presumably
because the flow of the function is somewhat odd.
After most of the core networking stack was converted to use
the tracking hold() variants, netpoll's call to old dev_hold()
stands out a bit.

np is allocated by the caller and ready to use, we can use
netdev_hold() here, even tho np->ndev will only be set to
ndev inside __netpoll_setup().

Reviewed-by: Eric Dumazet 
Signed-off-by: Jakub Kicinski 
Signed-off-by: David S. Miller

net: don't let netpoll invoke NAPI if in xmit context

2023-04-02T12:26:21+00:00

Commit 0db3dc73f7a3 ("[NETPOLL]: tx lock deadlock fix") narrowed
down the region under netif_tx_trylock() inside netpoll_send_skb().
(At that point in time netif_tx_trylock() would lock all queues of
the device.) Taking the tx lock was problematic because driver's
cleanup method may take the same lock. So the change made us hold
the xmit lock only around xmit, and expected the driver to take
care of locking within ->ndo_poll_controller().

Unfortunately this only works if netpoll isn't itself called with
the xmit lock already held. Netpoll code is careful and uses
trylock(). The drivers, however, may be using plain lock().
Printing while holding the xmit lock is going to result in rare
deadlocks.

Luckily we record the xmit lock owners, so we can scan all the queues,
the same way we scan NAPI owners. If any of the xmit locks is held
by the local CPU we better not attempt any polling.

It would be nice if we could narrow down the check to only the NAPIs
and the queue we're trying to use. I don't see a way to do that now.

Reported-by: Roman Gushchin 
Fixes: 0db3dc73f7a3 ("[NETPOLL]: tx lock deadlock fix")
Signed-off-by: Jakub Kicinski 
Reviewed-by: Eric Dumazet 
Signed-off-by: David S. Miller