net: fix possible overflow in __sk_mem_raise_allocated()
Subsystems (3)
PATCH 1/1 net: fix possible overflow in __sk_mem_raise_allocated()
With many active TCP sockets, fat TCP sockets could fool __sk_mem_raise_allocated() thanks to an overflow.
They would increase their share of the memory, instead of decreasing it.
Signed-off-by Eric Dumazet <[email protected]>
--- include/net/sock.h | 2 +- net/core/sock.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-)
include/net/sock.h +1 -1
@@ -1277,7 +1277,7 @@ static inline void sk_sockets_allocated_inc(struct sock *sk)percpu_counter_inc (sk->sk_prot->sockets_allocated); } -static inline int +static inline u64 sk_sockets_allocated_read_positive (struct sock *sk) {return percpu_counter_read_positive (sk->sk_prot->sockets_allocated);
net/core/sock.c +1 -1
@@ -2380,7 +2380,7 @@ int __sk_mem_raise_allocated(struct sock *sk, int size, int amt, int kind) }if (sk_has_memory_pressure (sk)) { -int alloc; +u64 alloc;if (!sk_under_memory_pressure (sk))return 1 ;
Signature · 1 line
2.20.1.791.gb4d0f1c61a-goog
From: Eric Dumazet <[email protected]>
Date: Tue, 12 Feb 2019 12:26:27 -0800
Quoted 7 lines
With many active TCP sockets, fat TCP sockets could fool
__sk_mem_raise_allocated()thanks to an overflow.They would increase their share of the memory, instead of decreasing it.
Signed-off-by Eric Dumazet <[email protected]>
Applied, thanks Eric.