linux-stable.git/net/netfilter, branch v5.14.7

netfilter: nft_ct: protect nft_ct_pcpu_template_refcnt with mutex

2021-09-22T10:39:28+00:00

[ Upstream commit e3245a7b7b34bd2e97f744fd79463add6e9d41f4 ]

Syzbot hit use-after-free in nf_tables_dump_sets. The problem was in
missing lock protection for nft_ct_pcpu_template_refcnt.

Before commit f102d66b335a ("netfilter: nf_tables: use dedicated
mutex to guard transactions") all transactions were serialized by global
mutex, but then global mutex was changed to local per netnamespace
commit_mutex.

This change causes use-after-free bug, when 2 netnamespaces concurently
changing nft_ct_pcpu_template_refcnt without proper locking. Fix it by
adding nft_ct_pcpu_mutex and protect all nft_ct_pcpu_template_refcnt
changes with it.

Fixes: f102d66b335a ("netfilter: nf_tables: use dedicated mutex to guard transactions")
Reported-and-tested-by: syzbot+649e339fa6658ee623d3@syzkaller.appspotmail.com
Signed-off-by: Pavel Skripkin 
Acked-by: Florian Westphal 
Signed-off-by: Pablo Neira Ayuso 
Signed-off-by: Sasha Levin

net: Fix offloading indirect devices dependency on qdisc order creation

2021-09-18T11:43:53+00:00

[ Upstream commit 74fc4f828769cca1c3be89ea92cb88feaa27ef52 ]

Currently, when creating an ingress qdisc on an indirect device before
the driver registered for callbacks, the driver will not have a chance
to register its filter configuration callbacks.

To fix that, modify the code such that it keeps track of all the ingress
qdiscs that call flow_indr_dev_setup_offload(). When a driver calls
flow_indr_dev_register(),  go through the list of tracked ingress qdiscs
and call the driver callback entry point so as to give it a chance to
register its callback.

Reviewed-by: Jiri Pirko 
Signed-off-by: Eli Cohen 
Signed-off-by: David S. Miller 
Signed-off-by: Sasha Levin

netfilter: nft_compat: use nfnetlink_unicast()

2021-09-18T11:43:44+00:00

[ Upstream commit 241d1af4c11a75d4c17ecc0193a6ab60553efbfc ]

Use nfnetlink_unicast() which already translates EAGAIN to ENOBUFS,
since EAGAIN is reserved to report missing module dependencies to the
nfnetlink core.

e0241ae6ac59 ("netfilter: use nfnetlink_unicast() forgot to update
this spot.

Reported-by: Yajun Deng 
Signed-off-by: Pablo Neira Ayuso 
Signed-off-by: Sasha Levin

netfilter: nfnetlink_hook: translate inet ingress to netdev

2021-08-06T15:07:41+00:00

The NFPROTO_INET pseudofamily is not exposed through this new netlink
interface. The netlink dump either shows NFPROTO_IPV4 or NFPROTO_IPV6
for NFPROTO_INET prerouting/input/forward/output/postrouting hooks.
The NFNLA_CHAIN_FAMILY attribute provides the family chain, which
specifies if this hook applies to inet traffic only (either IPv4 or
IPv6).

Translate the inet/ingress hook to netdev/ingress to fully hide the
NFPROTO_INET implementation details.

Fixes: e2cf17d3774c ("netfilter: add new hook nfnl subsystem")
Signed-off-by: Pablo Neira Ayuso

netfilter: conntrack: remove offload_pickup sysctl again

2021-08-06T15:07:41+00:00

These two sysctls were added because the hardcoded defaults (2 minutes,
tcp, 30 seconds, udp) turned out to be too low for some setups.

They appeared in 5.14-rc1 so it should be fine to remove it again.

Marcelo convinced me that there should be no difference between a flow
that was offloaded vs. a flow that was not wrt. timeout handling.
Thus the default is changed to those for TCP established and UDP stream,
5 days and 120 seconds, respectively.

Marcelo also suggested to account for the timeout value used for the
offloading, this avoids increase beyond the value in the conntrack-sysctl
and will also instantly expire the conntrack entry with altered sysctls.

Example:
   nf_conntrack_udp_timeout_stream=60
   nf_flowtable_udp_timeout=60

This will remove offloaded udp flows after one minute, rather than two.

An earlier version of this patch also cleared the ASSURED bit to
allow nf_conntrack to evict the entry via early_drop (i.e., table full).
However, it looks like we can safely assume that connection timed out
via HW is still in established state, so this isn't needed.

Quoting Oz:
 [..] the hardware sends all packets with a set FIN flags to sw.
 [..] Connections that are aged in hardware are expected to be in the
 established state.

In case it turns out that back-to-sw-path transition can occur for
'dodgy' connections too (e.g., one side disappeared while software-path
would have been in RETRANS timeout), we can adjust this later.

Cc: Oz Shlomo 
Cc: Paul Blakey 
Suggested-by: Marcelo Ricardo Leitner 
Signed-off-by: Florian Westphal 
Reviewed-by: Marcelo Ricardo Leitner 
Reviewed-by: Oz Shlomo 
Signed-off-by: Pablo Neira Ayuso

netfilter: nfnetlink_hook: Use same family as request message

2021-08-06T15:07:41+00:00

Use the same family as the request message, for consistency. The
netlink payload provides sufficient information to describe the hook
object, including the family.

This makes it easier to userspace to correlate the hooks are that
visited by the packets for a certain family.

Fixes: e2cf17d3774c ("netfilter: add new hook nfnl subsystem")
Signed-off-by: Pablo Neira Ayuso

netfilter: nfnetlink_hook: use the sequence number of the request message

2021-08-06T15:07:40+00:00

The sequence number allows to correlate the netlink reply message (as
part of the dump) with the original request message.

The cb->seq field is internally used to detect an interference (update)
of the hook list during the netlink dump, do not use it as sequence
number in the netlink dump header.

Fixes: e2cf17d3774c ("netfilter: add new hook nfnl subsystem")
Signed-off-by: Pablo Neira Ayuso

netfilter: nfnetlink_hook: missing chain family

2021-08-06T15:07:40+00:00

The family is relevant for pseudo-families like NFPROTO_INET
otherwise the user needs to rely on the hook function name to
differentiate it from NFPROTO_IPV4 and NFPROTO_IPV6 names.

Add nfnl_hook_chain_desc_attributes instead of using the existing
NFTA_CHAIN_* attributes, since these do not provide a family number.

Fixes: e2cf17d3774c ("netfilter: add new hook nfnl subsystem")
Signed-off-by: Pablo Neira Ayuso

netfilter: nfnetlink_hook: strip off module name from hookfn

2021-08-06T15:07:40+00:00

NFNLA_HOOK_FUNCTION_NAME should include the hook function name only,
the module name is already provided by NFNLA_HOOK_MODULE_NAME.

Fixes: e2cf17d3774c ("netfilter: add new hook nfnl subsystem")
Signed-off-by: Pablo Neira Ayuso

netfilter: conntrack: collect all entries in one cycle

2021-08-06T15:07:35+00:00

Michal Kubecek reports that conntrack gc is responsible for frequent
wakeups (every 125ms) on idle systems.

On busy systems, timed out entries are evicted during lookup.
The gc worker is only needed to remove entries after system becomes idle
after a busy period.

To resolve this, always scan the entire table.
If the scan is taking too long, reschedule so other work_structs can run
and resume from next bucket.

After a completed scan, wait for 2 minutes before the next cycle.
Heuristics for faster re-schedule are removed.

GC_SCAN_INTERVAL could be exposed as a sysctl in the future to allow
tuning this as-needed or even turn the gc worker off.

Reported-by: Michal Kubecek 
Signed-off-by: Florian Westphal 
Signed-off-by: Pablo Neira Ayuso