Phonet: remove dangling pipe if an endpoint is closed early
Subsystems (3)
PATCH 1/1 Phonet: remove dangling pipe if an endpoint is closed early
From: Rémi Denis-Courmont <[email protected]>
Closing a pipe endpoint is not normally allowed by the Phonet pipe, other than as a side after-effect of removing the pipe between two endpoints. But there is no way to prevent Linux userspace processes from being killed or suffering from bugs, so this can still happen. We might as well forcefully close Phonet pipe endpoints then.
The cellular modem supports only a few existing pipes at a time. So we really should not leak them. This change instructs the modem to destroy the pipe if either of the pipe's endpoint (Linux socket) is closed too early.
--- include/net/phonet/pep.h | 5 +++++ net/phonet/pep.c | 27 ++++++++++++++++++++++++++- 2 files changed, 31 insertions(+), 1 deletions(-)
include/net/phonet/pep.h +5 -0
@@ -77,6 +77,11 @@ static inline struct pnpipehdr *pnp_hdr(struct sk_buff *skb)#define MAX_PNPIPE_HEADER (MAX_PHONET_HEADER + 4)enum { +PNS_PIPE_CREATE_REQ =0x00 , +PNS_PIPE_CREATE_RESP , +PNS_PIPE_REMOVE_REQ , +PNS_PIPE_REMOVE_RESP , +PNS_PIPE_DATA =0x20 ,PNS_PIPE_ALIGNED_DATA ,
net/phonet/pep.c +26 -1
@@ -620,6 +620,28 @@ drop:return err; } +static int pipe_do_remove (struct sock *sk) +{ +struct pep_sock *pn =pep_sk (sk); +struct pnpipehdr *ph; +struct sk_buff *skb; + + skb =alloc_skb (MAX_PNPIPE_HEADER ,GFP_KERNEL ); +if (!skb) +return -ENOMEM ; + +skb_reserve (skb,MAX_PNPIPE_HEADER ); +__skb_push (skb,sizeof (*ph)); +skb_reset_transport_header (skb); + ph =pnp_hdr (skb); + ph->utid =0 ; + ph->message_id =PNS_PIPE_REMOVE_REQ ; + ph->pipe_handle = pn->pipe_handle; + ph->data[0 ] =PAD ; + +return pn_skb_send (sk, skb, &pipe_srv); +} +/* associated socket ceases to exist */ static void pep_sock_close (struct sock *sk,long timeout) {
@@ -638,7 +660,10 @@ static void pep_sock_close(struct sock *sk, long timeout)sk_for_each_safe (sknode, p, n, &pn->ackq)sk_del_node_init (sknode); sk->sk_state =TCP_CLOSE ; - } + }else if ((1 << sk->sk_state) & (TCPF_SYN_RECV |TCPF_ESTABLISHED )) +/* Forcefully remove dangling Phonet pipe */ +pipe_do_remove (sk); + ifindex = pn->ifindex; pn->ifindex =0 ;release_sock (sk);
Signature · 1 line
1.7.0.4
From: Rémi Denis-Courmont <[email protected]>
Date: Thu, 16 Sep 2010 01:19:53 +0300
Quoted 8 lines
From: Rémi Denis-Courmont <[email protected]>
Closing a pipe endpoint is not normally allowed by the Phonet pipe, other than as a side after-effect of removing the pipe between two endpoints. But there is no way to prevent Linux userspace processes from being killed or suffering from bugs, so this can still happen. We might as well forcefully close Phonet pipe endpoints then.
Quoted 6 lines
The cellular modem supports only a few existing pipes at a time. So we really should not leak them. This change instructs the modem to destroy the pipe if either of the pipe's endpoint (Linux socket) is closed too early.
Signed-off-by Rémi Denis-Courmont <[email protected]>
Applied.
David Miller
maintainerFrom: David Miller <[email protected]>
Date: Wed, 15 Sep 2010 21:32:58 -0700 (PDT)
Quoted 1 line
Applied.
Btw, I had to apply the following warning fix against your other set of changes.
--------------------
phonet: Fix build warning.
net/phonet/socket.c: In function ‘pn_res_seq_show’:net/phonet/socket.c:726: warning: format ‘%02X’ expects type ‘unsigned int’, but argument 3 has type ‘long int’
--- net/phonet/socket.c | 3 ++- 1 files changed, 2 insertions(+), 1 deletions(-)
net/phonet/socket.c +2 -1
@@ -723,7 +723,8 @@ static int pn_res_seq_show(struct seq_file *seq, void *v)struct sock *sk = *psk;seq_printf (seq,"%02X %5d %lu%n" , - psk - pnres.sk,sock_i_uid (sk),sock_i_ino (sk), &len); + (int ) (psk - pnres.sk),sock_i_uid (sk), +sock_i_ino (sk), &len); }seq_printf (seq,"%*s\n" ,63 - len,"" );return 0 ;
Signature · 1 line
1.7.2.2
Hello,
Quoted 7 lines · David Miller
From: David Miller <[email protected]>
Date: Wed, 15 Sep 2010 21:32:58 -0700 (PDT)Quoted 1 line
Applied.
Btw, I had to apply the following warning fix against your other set of changes.
Oh yeah, I only have 32-bits systems at hand so ptrdiff_t == int for me :-( Sorry.
Signature · 2 lines
Rémi Denis-Courmont http://www.remlab.net/