net: Add TCP_FORCE_LINGER2 to TCP setsockopt
Subsystems (4)
PATCH 1/1 net: Add TCP_FORCE_LINGER2 to TCP setsockopt
This patch adds a new TCP socket option named TCP_FORCE_LINGER2. The option has same behavior as TCP_LINGER2, except the tp->linger2 value can be greater than sysctl_tcp_fin_timeout if the user_ns is capable with CAP_NET_ADMIN.
As a server, different sockets may need different FIN-WAIT timeout and in most cases the system default value will be used. The timeout can be adjusted by setting TCP_LINGER2 but cannot be greater than the system default value. If one socket needs a timeout greater than the default, we have to adjust the sysctl which affects all sockets using the system default value. And if we want to adjust it for just one socket and keep the original value for others, all the other sockets have to set TCP_LINGER2. But with TCP_FORCE_LINGER2, the net admin can set greater tp->linger2 than the default for one socket and keep the sysctl_tcp_fin_timeout unchanged.
--- include/uapi/linux/capability.h | 1 + include/uapi/linux/tcp.h | 1 + net/ipv4/tcp.c | 9 +++++++++ 3 files changed, 11 insertions(+)
include/uapi/linux/capability.h +1 -0
@@ -199,6 +199,7 @@ struct vfs_ns_cap_data {/* Allow multicasting */ /* Allow read/write of device-specific registers */ /* Allow activation of ATM control sockets */ +/* Allow setting TCP_LINGER2 regardless of sysctl_tcp_fin_timeout */ #define CAP_NET_ADMIN 12
include/uapi/linux/tcp.h +1 -0
@@ -128,6 +128,7 @@ enum {#define TCP_CM_INQ TCP_INQ#define TCP_TX_DELAY 37/* delay outgoing packets by XX usec */ +#define TCP_FORCE_LINGER2 38/* Set TCP_LINGER2 regardless of sysctl_tcp_fin_timeout */ #define TCP_REPAIR_ON 1
net/ipv4/tcp.c +9 -0
@@ -3149,6 +3149,15 @@ static int do_tcp_setsockopt(struct sock *sk, int level,tcp_enable_tx_delay (); tp->tcp_tx_delay = val;break ; +case TCP_FORCE_LINGER2 : +if (val <0 ) + tp->linger2 =-1 ; +else if (val > net->ipv4.sysctl_tcp_fin_timeout /HZ && + !ns_capable (sock_net (sk)->user_ns,CAP_NET_ADMIN )) + tp->linger2 =0 ; +else + tp->linger2 = val *HZ ; +break ;default : err = -ENOPROTOOPT ;break ;
Signature · 1 line
2.16.6
David Miller
maintainerFrom: Cambda Zhu <[email protected]>
Date: Tue, 21 Apr 2020 20:17:37 +0800
Quoted 16 lines
This patch adds a new TCP socket option named
TCP_FORCE_LINGER2. The option has same behavior asTCP_LINGER2, except thetp->linger2value can be greater thansysctl_tcp_fin_timeoutif theuser_nsis capable withCAP_NET_ADMIN.As a server, different sockets may need different FIN-WAIT timeout and in most cases the system default value will be used. The timeout can be adjusted by setting
TCP_LINGER2but cannot be greater than the system default value. If one socket needs a timeout greater than the default, we have to adjust the sysctl which affects all sockets using the system default value. And if we want to adjust it for just one socket and keep the original value for others, all the other sockets have to setTCP_LINGER2. But withTCP_FORCE_LINGER2, the net admin can set greatertp->linger2than the default for one socket and keep thesysctl_tcp_fin_timeoutunchanged.
Quoted 1 line
Signed-off-by Cambda Zhu <[email protected]>
Eric, please review.
Quoted 53 lines · Cambda Zhu
This patch adds a new TCP socket option named
TCP_FORCE_LINGER2. The option has same behavior asTCP_LINGER2, except thetp->linger2value can be greater thansysctl_tcp_fin_timeoutif theuser_nsis capable withCAP_NET_ADMIN.As a server, different sockets may need different FIN-WAIT timeout and in most cases the system default value will be used. The timeout can be adjusted by setting
TCP_LINGER2but cannot be greater than the system default value. If one socket needs a timeout greater than the default, we have to adjust the sysctl which affects all sockets using the system default value. And if we want to adjust it for just one socket and keep the original value for others, all the other sockets have to setTCP_LINGER2. But withTCP_FORCE_LINGER2, the net admin can set greatertp->linger2than the default for one socket and keep thesysctl_tcp_fin_timeoutunchanged.Signed-off-by Cambda Zhu <[email protected]>--- include/uapi/linux/capability.h | 1 + include/uapi/linux/tcp.h | 1 + net/ipv4/tcp.c | 9 +++++++++ 3 files changed, 11 insertions(+)include/uapi/linux/capability.h +1
-0@@ -199,6 +199,7 @@ struct vfs_ns_cap_data {/* Allow multicasting */ /* Allow read/write of device-specific registers */ /* Allow activation of ATM control sockets */ +/* Allow setting TCP_LINGER2 regardless of sysctl_tcp_fin_timeout */ #define CAP_NET_ADMIN 12include/uapi/linux/tcp.h +1
-0@@ -128,6 +128,7 @@ enum {#define TCP_CM_INQ TCP_INQ#define TCP_TX_DELAY 37/* delay outgoing packets by XX usec */ +#define TCP_FORCE_LINGER2 38/* Set TCP_LINGER2 regardless of sysctl_tcp_fin_timeout */ #define TCP_REPAIR_ON 1net/ipv4/tcp.c +0
-0@@ -3149,6 +3149,15 @@ static int do_tcp_setsockopt(struct sock *sk, int level,tcp_enable_tx_delay ();
net/ipv4/tcp.c static int do_tcp_setsockopt(struct sock *sk, int level, · line 3159
Quoted 10 lines
net/ipv4/tcp.c +8
-0tp->tcp_tx_delay = val;break ; +case TCP_FORCE_LINGER2 : +if (val <0 ) + tp->linger2 =-1 ; +else if (val > net->ipv4.sysctl_tcp_fin_timeout /HZ && + !ns_capable (sock_net (sk)->user_ns,CAP_NET_ADMIN )) + tp->linger2 =0 ; +else + tp->linger2 = val *HZ ;
This multiply could overflow.
Since tp->linger2 is an int, and a negative value has a specific meaning, you probably should have some sanity checks.
Even if the old TCP_LINGER2 silently put a 0,
maybe a new option should return an error if val*HZ would overflow.
Quoted 5 lines
net/ipv4/tcp.c +1
-0+break ;default : err = -ENOPROTOOPT ;break ;
PATCH v2 1/1 net: Add TCP_FORCE_LINGER2 to TCP setsockopt
This patch adds a new TCP socket option named TCP_FORCE_LINGER2. The option has same behavior as TCP_LINGER2, except the tp->linger2 value can be greater than sysctl_tcp_fin_timeout if the user_ns is capable with CAP_NET_ADMIN.
As a server, different sockets may need different FIN-WAIT timeout and in most cases the system default value will be used. The timeout can be adjusted by setting TCP_LINGER2 but cannot be greater than the system default value. If one socket needs a timeout greater than the default, we have to adjust the sysctl which affects all sockets using the system default value. And if we want to adjust it for just one socket and keep the original value for others, all the other sockets have to set TCP_LINGER2. But with TCP_FORCE_LINGER2, the net admin can set greater tp->linger2 than the default for one socket and keep the sysctl_tcp_fin_timeout unchanged.
---
Changes in v2:
- Add int overflow check.
include/uapi/linux/capability.h | 1 + include/uapi/linux/tcp.h | 1 + net/ipv4/tcp.c | 11 +++++++++++ 3 files changed, 13 insertions(+)
include/uapi/linux/capability.h +1 -0
@@ -199,6 +199,7 @@ struct vfs_ns_cap_data {/* Allow multicasting */ /* Allow read/write of device-specific registers */ /* Allow activation of ATM control sockets */ +/* Allow setting TCP_LINGER2 regardless of sysctl_tcp_fin_timeout */ #define CAP_NET_ADMIN 12
include/uapi/linux/tcp.h +1 -0
@@ -128,6 +128,7 @@ enum {#define TCP_CM_INQ TCP_INQ#define TCP_TX_DELAY 37/* delay outgoing packets by XX usec */ +#define TCP_FORCE_LINGER2 38/* Set TCP_LINGER2 regardless of sysctl_tcp_fin_timeout */ #define TCP_REPAIR_ON 1
net/ipv4/tcp.c +11 -0
@@ -3149,6 +3149,17 @@ static int do_tcp_setsockopt(struct sock *sk, int level,tcp_enable_tx_delay (); tp->tcp_tx_delay = val;break ; +case TCP_FORCE_LINGER2 : +if (val <0 ) + tp->linger2 =-1 ; +else if (val >INT_MAX /HZ ) + err = -EINVAL ; +else if (val > net->ipv4.sysctl_tcp_fin_timeout /HZ && + !ns_capable (sock_net (sk)->user_ns,CAP_NET_ADMIN )) + tp->linger2 =0 ; +else + tp->linger2 = val *HZ ; +break ;default : err = -ENOPROTOOPT ;break ;
Signature · 1 line
2.16.6
Quoted 63 lines · Cambda Zhu
This patch adds a new TCP socket option named
TCP_FORCE_LINGER2. The option has same behavior asTCP_LINGER2, except thetp->linger2value can be greater thansysctl_tcp_fin_timeoutif theuser_nsis capable withCAP_NET_ADMIN.As a server, different sockets may need different FIN-WAIT timeout and in most cases the system default value will be used. The timeout can be adjusted by setting
TCP_LINGER2but cannot be greater than the system default value. If one socket needs a timeout greater than the default, we have to adjust the sysctl which affects all sockets using the system default value. And if we want to adjust it for just one socket and keep the original value for others, all the other sockets have to setTCP_LINGER2. But withTCP_FORCE_LINGER2, the net admin can set greatertp->linger2than the default for one socket and keep thesysctl_tcp_fin_timeoutunchanged.Signed-off-by Cambda Zhu <[email protected]>---
Changes in v2:
- Add int overflow check.include/uapi/linux/capability.h | 1 + include/uapi/linux/tcp.h | 1 + net/ipv4/tcp.c | 11 +++++++++++ 3 files changed, 13 insertions(+)include/uapi/linux/capability.h +1
-0@@ -199,6 +199,7 @@ struct vfs_ns_cap_data {/* Allow multicasting */ /* Allow read/write of device-specific registers */ /* Allow activation of ATM control sockets */ +/* Allow setting TCP_LINGER2 regardless of sysctl_tcp_fin_timeout */ #define CAP_NET_ADMIN 12include/uapi/linux/tcp.h +1
-0@@ -128,6 +128,7 @@ enum {#define TCP_CM_INQ TCP_INQ#define TCP_TX_DELAY 37/* delay outgoing packets by XX usec */ +#define TCP_FORCE_LINGER2 38/* Set TCP_LINGER2 regardless of sysctl_tcp_fin_timeout */ #define TCP_REPAIR_ON 1net/ipv4/tcp.c +4
-0@@ -3149,6 +3149,17 @@ static int do_tcp_setsockopt(struct sock *sk, int level,tcp_enable_tx_delay (); tp->tcp_tx_delay = val;break ; +case TCP_FORCE_LINGER2 : +if (val <0 ) + tp->linger2 =-1 ; +else if (val >INT_MAX /HZ )
net/ipv4/tcp.c static int do_tcp_setsockopt(struct sock *sk, int level, · line 3165
Quoted 10 lines
net/ipv4/tcp.c +7
-0+ err = -EINVAL ; + else if (val > net->ipv4.sysctl_tcp_fin_timeout / HZ && + !ns_capable(sock_net(sk)->user_ns, CAP_NET_ADMIN)) + tp->linger2 = 0; + else + tp->linger2 = val * HZ; +break ;default : err = -ENOPROTOOPT ;break ;
INT_MAX looks quite
Anyway, I do not think we need a new socket option, since really it will need documentation and add more confusion.
net->ipv4.sysctl_tcp_fin_timeout is the default value for sockets which have tp->linger2 cleared.
Fact that it has been used to cap TCP_LINGER2 was probably a mistake.
What about adding a new define and simply let TCP_LINGER2 use it ?
Really there is no point trying to allow hours or even days for FIN timeout,
and no point limiting a socket from having a value between net->ipv4.sysctl_tcp_fin_timeout and 2 minutes, at least from security perspective, these values seem legal as far as TCP specs are concerned.
include/net/tcp.h +1 -0
@@ -126,6 +126,7 @@ void tcp_time_wait(struct sock *sk, int state, int timeo);* to combine FIN-WAIT-2 timeout with * TIME-WAIT timer. */ +#define TCP_FIN_TIMEOUT_MAX (120 * HZ)/* max TCP_LINGER2 value (two minutes) */ #define TCP_DELACK_MAX ((unsigned)(HZ/5))/* maximal time to delay before sending an ACK */ #if HZ >=100
net/ipv4/tcp.c +1 -1
@@ -3035,7 +3035,7 @@ static int do_tcp_setsockopt(struct sock *sk, int level,case TCP_LINGER2 :if (val <0 ) tp->linger2 =-1 ; -else if (val > net->ipv4.sysctl_tcp_fin_timeout /HZ ) +else if (val >TCP_FIN_TIMEOUT_MAX /HZ ) tp->linger2 =0 ;else tp->linger2 = val *HZ ;
Quoted 87 lines · Eric Dumazet
Quoted 63 lines · Cambda Zhu
This patch adds a new TCP socket option named
TCP_FORCE_LINGER2. The option has same behavior asTCP_LINGER2, except thetp->linger2value can be greater thansysctl_tcp_fin_timeoutif theuser_nsis capable withCAP_NET_ADMIN.As a server, different sockets may need different FIN-WAIT timeout and in most cases the system default value will be used. The timeout can be adjusted by setting
TCP_LINGER2but cannot be greater than the system default value. If one socket needs a timeout greater than the default, we have to adjust the sysctl which affects all sockets using the system default value. And if we want to adjust it for just one socket and keep the original value for others, all the other sockets have to setTCP_LINGER2. But withTCP_FORCE_LINGER2, the net admin can set greatertp->linger2than the default for one socket and keep thesysctl_tcp_fin_timeoutunchanged.Signed-off-by Cambda Zhu <[email protected]>---
Changes in v2:
- Add int overflow check.include/uapi/linux/capability.h | 1 + include/uapi/linux/tcp.h | 1 + net/ipv4/tcp.c | 11 +++++++++++ 3 files changed, 13 insertions(+)include/uapi/linux/capability.h +1
-0@@ -199,6 +199,7 @@ struct vfs_ns_cap_data {/* Allow multicasting */ /* Allow read/write of device-specific registers */ /* Allow activation of ATM control sockets */ +/* Allow setting TCP_LINGER2 regardless of sysctl_tcp_fin_timeout */ #define CAP_NET_ADMIN 12include/uapi/linux/tcp.h +1
-0@@ -128,6 +128,7 @@ enum {#define TCP_CM_INQ TCP_INQ#define TCP_TX_DELAY 37/* delay outgoing packets by XX usec */ +#define TCP_FORCE_LINGER2 38/* Set TCP_LINGER2 regardless of sysctl_tcp_fin_timeout */ #define TCP_REPAIR_ON 1net/ipv4/tcp.c +4
-0@@ -3149,6 +3149,17 @@ static int do_tcp_setsockopt(struct sock *sk, int level,tcp_enable_tx_delay (); tp->tcp_tx_delay = val;break ; +case TCP_FORCE_LINGER2 : +if (val <0 ) + tp->linger2 =-1 ; +else if (val >INT_MAX /HZ )net/ipv4/tcp.c static int do_tcp_setsockopt(struct sock *sk, int level, · line 3165
Quoted 10 lines
net/ipv4/tcp.c +7
-0+ err = -EINVAL ; + else if (val > net->ipv4.sysctl_tcp_fin_timeout / HZ && + !ns_capable(sock_net(sk)->user_ns, CAP_NET_ADMIN)) + tp->linger2 = 0; + else + tp->linger2 = val * HZ; +break ;default : err = -ENOPROTOOPT ;break ;
INT_MAXlooks quiteAnyway, I do not think we need a new socket option, since really it will need documentation and add more confusion.
net->ipv4.sysctl_tcp_fin_timeoutis the default value for sockets which havetp->linger2cleared.
Quoted 9 lines
Fact that it has been used to cap
TCP_LINGER2was probably a mistake.What about adding a new define and simply let
TCP_LINGER2use it ?Really there is no point trying to allow hours or even days for FIN timeout,
and no point limiting a socket from having a value betweennet->ipv4.sysctl_tcp_fin_timeoutand 2 minutes, at least from security perspective, these values seem legal as far as TCP specs are concerned.
I also think using sysctl_tcp_fin_timeout to cap TCP_LINGER2 is probably a mistake, and adding a new define for TCP_LINGER2 is a good idea. I have considered the solution and found it may have some compatibility issues. Whatever it’s a mistake or not, a system administrator may have used it to limit the max timeout for all sockets. And when I think about the behavior of TCP_LINGER2, I'm afraid the server may also rely on the behavior so the server can set any value greater than sysctl_tcp_fin_timeout and result in the same timeout.
Maybe my worry is unnecessary. If so, I think your suggestion is better.
Regards,
Cambda
Quoted 26 lines
include/net/tcp.h +1
-0@@ -126,6 +126,7 @@ void tcp_time_wait(struct sock *sk, int state, int timeo);* to combine FIN-WAIT-2 timeout with * TIME-WAIT timer. */ +#define TCP_FIN_TIMEOUT_MAX (120 * HZ)/* max TCP_LINGER2 value (two minutes) */ #define TCP_DELACK_MAX ((unsigned)(HZ/5))/* maximal time to delay before sending an ACK */ #if HZ >=100 net/ipv4/tcp.c +1
-1@@ -3035,7 +3035,7 @@ static int do_tcp_setsockopt(struct sock *sk, int level,case TCP_LINGER2 :if (val <0 ) tp->linger2 =-1 ; -else if (val > net->ipv4.sysctl_tcp_fin_timeout /HZ ) +else if (val >TCP_FIN_TIMEOUT_MAX /HZ ) tp->linger2 =0 ;else tp->linger2 = val *HZ ;
Quoted 105 lines · Cambda Zhu
Quoted 87 lines · Eric Dumazet
Quoted 63 lines · Cambda Zhu
This patch adds a new TCP socket option named
TCP_FORCE_LINGER2. The option has same behavior asTCP_LINGER2, except thetp->linger2value can be greater thansysctl_tcp_fin_timeoutif theuser_nsis capable withCAP_NET_ADMIN.As a server, different sockets may need different FIN-WAIT timeout and in most cases the system default value will be used. The timeout can be adjusted by setting
TCP_LINGER2but cannot be greater than the system default value. If one socket needs a timeout greater than the default, we have to adjust the sysctl which affects all sockets using the system default value. And if we want to adjust it for just one socket and keep the original value for others, all the other sockets have to setTCP_LINGER2. But withTCP_FORCE_LINGER2, the net admin can set greatertp->linger2than the default for one socket and keep thesysctl_tcp_fin_timeoutunchanged.Signed-off-by Cambda Zhu <[email protected]>---
Changes in v2:
- Add int overflow check.include/uapi/linux/capability.h | 1 + include/uapi/linux/tcp.h | 1 + net/ipv4/tcp.c | 11 +++++++++++ 3 files changed, 13 insertions(+)include/uapi/linux/capability.h +1
-0@@ -199,6 +199,7 @@ struct vfs_ns_cap_data {/* Allow multicasting */ /* Allow read/write of device-specific registers */ /* Allow activation of ATM control sockets */ +/* Allow setting TCP_LINGER2 regardless of sysctl_tcp_fin_timeout */ #define CAP_NET_ADMIN 12include/uapi/linux/tcp.h +1
-0@@ -128,6 +128,7 @@ enum {#define TCP_CM_INQ TCP_INQ#define TCP_TX_DELAY 37/* delay outgoing packets by XX usec */ +#define TCP_FORCE_LINGER2 38/* Set TCP_LINGER2 regardless of sysctl_tcp_fin_timeout */ #define TCP_REPAIR_ON 1net/ipv4/tcp.c +4
-0@@ -3149,6 +3149,17 @@ static int do_tcp_setsockopt(struct sock *sk, int level,tcp_enable_tx_delay (); tp->tcp_tx_delay = val;break ; +case TCP_FORCE_LINGER2 : +if (val <0 ) + tp->linger2 =-1 ; +else if (val >INT_MAX /HZ )net/ipv4/tcp.c static int do_tcp_setsockopt(struct sock *sk, int level, · line 3165
Quoted 10 lines
net/ipv4/tcp.c +7
-0+ err = -EINVAL ; + else if (val > net->ipv4.sysctl_tcp_fin_timeout / HZ && + !ns_capable(sock_net(sk)->user_ns, CAP_NET_ADMIN)) + tp->linger2 = 0; + else + tp->linger2 = val * HZ; +break ;default : err = -ENOPROTOOPT ;break ;
INT_MAXlooks quiteAnyway, I do not think we need a new socket option, since really it will need documentation and add more confusion.
net->ipv4.sysctl_tcp_fin_timeoutis the default value for sockets which havetp->linger2cleared.Quoted 9 lines
Fact that it has been used to cap
TCP_LINGER2was probably a mistake.What about adding a new define and simply let
TCP_LINGER2use it ?Really there is no point trying to allow hours or even days for FIN timeout,
and no point limiting a socket from having a value betweennet->ipv4.sysctl_tcp_fin_timeoutand 2 minutes, at least from security perspective, these values seem legal as far as TCP specs are concerned.I also think using
sysctl_tcp_fin_timeoutto capTCP_LINGER2is probably a mistake, and adding a new define forTCP_LINGER2is a good idea. I have considered the solution and found it may have some compatibility issues. Whatever it’s a mistake or not, a system administrator may have used it to limit the max timeout for all sockets. And when I think about the behavior ofTCP_LINGER2, I'm afraid the server may also rely on the behavior so the server can set any value greater thansysctl_tcp_fin_timeoutand result in the same timeout.
Quoted 2 lines
Maybe my worry is unnecessary. If so, I think your suggestion is better.
Ṕlease send a v3 with a proper changelog then, thanks !
Quoted 112 lines · Eric Dumazet
Quoted 105 lines · Cambda Zhu
Quoted 87 lines · Eric Dumazet
Quoted 63 lines · Cambda Zhu
This patch adds a new TCP socket option named
TCP_FORCE_LINGER2. The option has same behavior asTCP_LINGER2, except thetp->linger2value can be greater thansysctl_tcp_fin_timeoutif theuser_nsis capable withCAP_NET_ADMIN.As a server, different sockets may need different FIN-WAIT timeout and in most cases the system default value will be used. The timeout can be adjusted by setting
TCP_LINGER2but cannot be greater than the system default value. If one socket needs a timeout greater than the default, we have to adjust the sysctl which affects all sockets using the system default value. And if we want to adjust it for just one socket and keep the original value for others, all the other sockets have to setTCP_LINGER2. But withTCP_FORCE_LINGER2, the net admin can set greatertp->linger2than the default for one socket and keep thesysctl_tcp_fin_timeoutunchanged.Signed-off-by Cambda Zhu <[email protected]>---
Changes in v2:
- Add int overflow check.include/uapi/linux/capability.h | 1 + include/uapi/linux/tcp.h | 1 + net/ipv4/tcp.c | 11 +++++++++++ 3 files changed, 13 insertions(+)include/uapi/linux/capability.h +1
-0@@ -199,6 +199,7 @@ struct vfs_ns_cap_data {/* Allow multicasting */ /* Allow read/write of device-specific registers */ /* Allow activation of ATM control sockets */ +/* Allow setting TCP_LINGER2 regardless of sysctl_tcp_fin_timeout */ #define CAP_NET_ADMIN 12include/uapi/linux/tcp.h +1
-0@@ -128,6 +128,7 @@ enum {#define TCP_CM_INQ TCP_INQ#define TCP_TX_DELAY 37/* delay outgoing packets by XX usec */ +#define TCP_FORCE_LINGER2 38/* Set TCP_LINGER2 regardless of sysctl_tcp_fin_timeout */ #define TCP_REPAIR_ON 1net/ipv4/tcp.c +4
-0@@ -3149,6 +3149,17 @@ static int do_tcp_setsockopt(struct sock *sk, int level,tcp_enable_tx_delay (); tp->tcp_tx_delay = val;break ; +case TCP_FORCE_LINGER2 : +if (val <0 ) + tp->linger2 =-1 ; +else if (val >INT_MAX /HZ )net/ipv4/tcp.c static int do_tcp_setsockopt(struct sock *sk, int level, · line 3165
Quoted 10 lines
net/ipv4/tcp.c +7
-0+ err = -EINVAL ; + else if (val > net->ipv4.sysctl_tcp_fin_timeout / HZ && + !ns_capable(sock_net(sk)->user_ns, CAP_NET_ADMIN)) + tp->linger2 = 0; + else + tp->linger2 = val * HZ; +break ;default : err = -ENOPROTOOPT ;break ;
INT_MAXlooks quiteAnyway, I do not think we need a new socket option, since really it will need documentation and add more confusion.
net->ipv4.sysctl_tcp_fin_timeoutis the default value for sockets which havetp->linger2cleared.Quoted 9 lines
Fact that it has been used to cap
TCP_LINGER2was probably a mistake.What about adding a new define and simply let
TCP_LINGER2use it ?Really there is no point trying to allow hours or even days for FIN timeout,
and no point limiting a socket from having a value betweennet->ipv4.sysctl_tcp_fin_timeoutand 2 minutes, at least from security perspective, these values seem legal as far as TCP specs are concerned.I also think using
sysctl_tcp_fin_timeoutto capTCP_LINGER2is probably a mistake, and adding a new define forTCP_LINGER2is a good idea. I have considered the solution and found it may have some compatibility issues. Whatever it’s a mistake or not, a system administrator may have used it to limit the max timeout for all sockets. And when I think about the behavior ofTCP_LINGER2, I'm afraid the server may also rely on the behavior so the server can set any value greater thansysctl_tcp_fin_timeoutand result in the same timeout.
Quoted 4 lines
Quoted 2 lines
Maybe my worry is unnecessary. If so, I think your suggestion is better.
Ṕlease send a v3 with a proper changelog then, thanks !
If the value between sysctl_tcp_fin_timeout and 2 minutes is legal, should we set tp->linger2 to TCP_FIN_TIMEOUT_MAX or return an error instead of set it to zero which means the default value?
Regards,
Cambda
Quoted 4 lines · Cambda Zhu
If the value between
sysctl_tcp_fin_timeoutand 2 minutes is legal, should we settp->linger2toTCP_FIN_TIMEOUT_MAXor return an error instead of set it to zero which means the default value?
I would not send an error.
This might break some applications that could consider an error as a serious one.
(You can reuse my patch basically)
Quoted 57 lines · Cambda Zhu
This patch adds a new TCP socket option named
TCP_FORCE_LINGER2. The option has same behavior asTCP_LINGER2, except thetp->linger2value can be greater thansysctl_tcp_fin_timeoutif theuser_nsis capable withCAP_NET_ADMIN.As a server, different sockets may need different FIN-WAIT timeout and in most cases the system default value will be used. The timeout can be adjusted by setting
TCP_LINGER2but cannot be greater than the system default value. If one socket needs a timeout greater than the default, we have to adjust the sysctl which affects all sockets using the system default value. And if we want to adjust it for just one socket and keep the original value for others, all the other sockets have to setTCP_LINGER2. But withTCP_FORCE_LINGER2, the net admin can set greatertp->linger2than the default for one socket and keep thesysctl_tcp_fin_timeoutunchanged.Signed-off-by Cambda Zhu <[email protected]>---
Changes in v2:
- Add int overflow check.include/uapi/linux/capability.h | 1 + include/uapi/linux/tcp.h | 1 + net/ipv4/tcp.c | 11 +++++++++++ 3 files changed, 13 insertions(+)include/uapi/linux/capability.h +1
-0@@ -199,6 +199,7 @@ struct vfs_ns_cap_data {/* Allow multicasting */ /* Allow read/write of device-specific registers */ /* Allow activation of ATM control sockets */ +/* Allow setting TCP_LINGER2 regardless of sysctl_tcp_fin_timeout */ #define CAP_NET_ADMIN 12include/uapi/linux/tcp.h +1
-0@@ -128,6 +128,7 @@ enum {#define TCP_CM_INQ TCP_INQ#define TCP_TX_DELAY 37/* delay outgoing packets by XX usec */ +#define TCP_FORCE_LINGER2 38/* Set TCP_LINGER2 regardless of sysctl_tcp_fin_timeout */ #define TCP_REPAIR_ON 1net/ipv4/tcp.c +0
-0@@ -3149,6 +3149,17 @@ static int do_tcp_setsockopt(struct sock *sk, int level,tcp_enable_tx_delay ();
net/ipv4/tcp.c static int do_tcp_setsockopt(struct sock *sk, int level, · line 3159
Quoted 10 lines
net/ipv4/tcp.c +8
-0tp->tcp_tx_delay = val;break ; +case TCP_FORCE_LINGER2 : +if (val <0 ) + tp->linger2 =-1 ; +else if (val >INT_MAX /HZ ) + err = -EINVAL ; +else if (val > net->ipv4.sysctl_tcp_fin_timeout /HZ && + !ns_capable (sock_net (sk)->user_ns,CAP_NET_ADMIN )) + tp->linger2 =0 ;
Instead of silently falling back to LINGER2 behavior for unprivileged users, I would fail without privileges, similar to SO_(SND|RCV)BUFFORCE.
Also, those have capable instead of ns_capable. If there is risk to system integrity, that is the right choice.
Slight aside, if the original setsockopt had checked optval == sizeof(int), we could have added a variant of different size (say, with an additional flags field), instead of having to create a new socket option.
Quoted 9 lines
net/ipv4/tcp.c +3
-1+else + tp->linger2 = val *HZ ; +break ;default : err = -ENOPROTOOPT ;break ; --2.16.6
Quoted 68 lines · Willem de Bruijn
Quoted 63 lines · Cambda Zhu
This patch adds a new TCP socket option named
TCP_FORCE_LINGER2. The option has same behavior asTCP_LINGER2, except thetp->linger2value can be greater thansysctl_tcp_fin_timeoutif theuser_nsis capable withCAP_NET_ADMIN.As a server, different sockets may need different FIN-WAIT timeout and in most cases the system default value will be used. The timeout can be adjusted by setting
TCP_LINGER2but cannot be greater than the system default value. If one socket needs a timeout greater than the default, we have to adjust the sysctl which affects all sockets using the system default value. And if we want to adjust it for just one socket and keep the original value for others, all the other sockets have to setTCP_LINGER2. But withTCP_FORCE_LINGER2, the net admin can set greatertp->linger2than the default for one socket and keep thesysctl_tcp_fin_timeoutunchanged.Signed-off-by Cambda Zhu <[email protected]>---
Changes in v2:
- Add int overflow check.include/uapi/linux/capability.h | 1 + include/uapi/linux/tcp.h | 1 + net/ipv4/tcp.c | 11 +++++++++++ 3 files changed, 13 insertions(+)include/uapi/linux/capability.h +1
-0@@ -199,6 +199,7 @@ struct vfs_ns_cap_data {/* Allow multicasting */ /* Allow read/write of device-specific registers */ /* Allow activation of ATM control sockets */ +/* Allow setting TCP_LINGER2 regardless of sysctl_tcp_fin_timeout */ #define CAP_NET_ADMIN 12include/uapi/linux/tcp.h +1
-0@@ -128,6 +128,7 @@ enum {#define TCP_CM_INQ TCP_INQ#define TCP_TX_DELAY 37/* delay outgoing packets by XX usec */ +#define TCP_FORCE_LINGER2 38/* Set TCP_LINGER2 regardless of sysctl_tcp_fin_timeout */ #define TCP_REPAIR_ON 1net/ipv4/tcp.c +4
-0@@ -3149,6 +3149,17 @@ static int do_tcp_setsockopt(struct sock *sk, int level,tcp_enable_tx_delay (); tp->tcp_tx_delay = val;break ; +case TCP_FORCE_LINGER2 : +if (val <0 ) + tp->linger2 =-1 ; +else if (val >INT_MAX /HZ )
Quoted 9 lines
net/ipv4/tcp.c static int do_tcp_setsockopt(struct sock *sk, int level, · line 3159
Quoted 4 lines
net/ipv4/tcp.c +4
-0+ err = -EINVAL ; + else if (val > net->ipv4.sysctl_tcp_fin_timeout / HZ && + !ns_capable(sock_net(sk)->user_ns, CAP_NET_ADMIN)) + tp->linger2 = 0;Instead of silently falling back to LINGER2 behavior for unprivileged users, I would fail without privileges, similar to SO_(SND|RCV)BUFFORCE.
Yes, I have considered failing without privileges too. It’s clearer but the user may have to set both of TCP_LINGER2 and TCP_FORCE_LINGER2 to set the timeout as TCP_FORCE_LINGER2 not introduced.
Quoted 3 lines
Also, those have capable instead of
ns_capable. If there is risk to system integrity, that is the right choice.
I think both are ok, but the sysctl_tcp_fin_timeout is in net ns.
Quoted 5 lines
Slight aside, if the original setsockopt had checked optval == sizeof(int), we could have added a variant of different size (say, with an additional flags field), instead of having to create a new socket option.
Maybe it’s a little weird… I don’t know. :)
Regards,
Cambda
Quoted 8 lines
Quoted 8 lines
net/ipv4/tcp.c +3
-1+else + tp->linger2 = val *HZ ; +break ;default : err = -ENOPROTOOPT ;break ; --2.16.6
PATCH v3 1/1 net: Replace the limit of TCP_LINGER2 with TCP_FIN_TIMEOUT_MAX
In reply to Cambda ZhuThis patch changes the behavior of TCP_LINGER2 about its limit. The sysctl_tcp_fin_timeout used to be the limit of TCP_LINGER2 but now it's only the default value. A new macro named TCP_FIN_TIMEOUT_MAX is added as the limit of TCP_LINGER2, which is 2 minutes.
Since TCP_LINGER2 used sysctl_tcp_fin_timeout as the default value and the limit in the past, the system administrator cannot set the default value for most of sockets and let some sockets have a greater timeout. It might be a mistake that let the sysctl to be the limit of the TCP_LINGER2. Maybe we can add a new sysctl to set the max of TCP_LINGER2, but FIN-WAIT-2 timeout is usually no need to be too long and 2 minutes are legal considering TCP specs.
Changes in v3:
- Remove the new socket option and change the TCP_LINGER2 behavior so that the timeout can be set to value between sysctl_tcp_fin_timeout and 2 minutes.
Changes in v2:
- Add int overflow check for the new socket option.
Changes in v1:
- Add a new socket option to set timeout greater than sysctl_tcp_fin_timeout.
--- include/net/tcp.h | 1 + net/ipv4/tcp.c | 4 ++-- 2 files changed, 3 insertions(+), 2 deletions(-)
include/net/tcp.h +1 -0
@@ -126,6 +126,7 @@ void tcp_time_wait(struct sock *sk, int state, int timeo);* to combine FIN-WAIT-2 timeout with * TIME-WAIT timer. */ +#define TCP_FIN_TIMEOUT_MAX (120 * HZ)/* max TCP_LINGER2 value (two minutes) */ #define TCP_DELACK_MAX ((unsigned)(HZ/5))/* maximal time to delay before sending an ACK */ #if HZ >=100
net/ipv4/tcp.c +2 -2
@@ -3035,8 +3035,8 @@ static int do_tcp_setsockopt(struct sock *sk, int level,case TCP_LINGER2 :if (val <0 ) tp->linger2 =-1 ; -else if (val > net->ipv4.sysctl_tcp_fin_timeout /HZ ) - tp->linger2 =0 ; +else if (val >TCP_FIN_TIMEOUT_MAX /HZ ) + tp->linger2 =TCP_FIN_TIMEOUT_MAX ;else tp->linger2 = val *HZ ;break ;
Signature · 1 line
2.16.6
Re: [PATCH net-next v3] net: Replace the limit of TCP_LINGER2 with TCP_FIN_TIMEOUT_MAX
From: Cambda Zhu <[email protected]>
Date: Fri, 24 Apr 2020 16:06:16 +0800
Quoted 18 lines
This patch changes the behavior of
TCP_LINGER2about its limit. Thesysctl_tcp_fin_timeoutused to be the limit ofTCP_LINGER2but now it's only the default value. A new macro namedTCP_FIN_TIMEOUT_MAXis added as the limit ofTCP_LINGER2, which is 2 minutes.Since
TCP_LINGER2usedsysctl_tcp_fin_timeoutas the default value and the limit in the past, the system administrator cannot set the default value for most of sockets and let some sockets have a greater timeout. It might be a mistake that let the sysctl to be the limit of theTCP_LINGER2. Maybe we can add a new sysctl to set the max ofTCP_LINGER2, but FIN-WAIT-2 timeout is usually no need to be too long and 2 minutes are legal considering TCP specs.Changes in v3:
- Remove the new socket option and change theTCP_LINGER2behavior so that the timeout can be set to value betweensysctl_tcp_fin_timeoutand 2 minutes.
Quoted 8 lines
Changes in v2:
- Add int overflow check for the new socket option.Changes in v1:
- Add a new socket option to set timeout greater thansysctl_tcp_fin_timeout.Signed-off-by Cambda Zhu <[email protected]>
Applied, thank you.
From: Cambda Zhu
Quoted 6 lines
Sent: 21 April 2020 13:18
This patch adds a new TCP socket option named
TCP_FORCE_LINGER2. The option has same behavior asTCP_LINGER2, except thetp->linger2value can be greater thansysctl_tcp_fin_timeoutif theuser_nsis capable withCAP_NET_ADMIN.
Did you consider adding an extra sysctl so that the limit for TCP_LINGER2 could be greater than the default?
It might even be sensible to set the limit to a few times the default.
All users can set the socket buffer sizes to twice the default. Being able to do the same for the linger timeout wouldn't be a problem.
David
-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK Registration No: 1397386 (Wales)
Quoted 18 lines · David Laight
From: Cambda Zhu
Quoted 6 lines
Sent: 21 April 2020 13:18
This patch adds a new TCP socket option named
TCP_FORCE_LINGER2. The option has same behavior asTCP_LINGER2, except thetp->linger2value can be greater thansysctl_tcp_fin_timeoutif theuser_nsis capable withCAP_NET_ADMIN.Did you consider adding an extra sysctl so that the limit for
TCP_LINGER2could be greater than the default?It might even be sensible to set the limit to a few times the default.
Quoted 8 lines
All users can set the socket buffer sizes to twice the default. Being able to do the same for the linger timeout wouldn't be a problem.
David
-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK Registration No: 1397386 (Wales)
The default tcp_fin_timeout is 60s which may be too long for a server with tens of thousands connections or qps. To reduce the connections as many as possible, the system administer may set the sysctl_tcp_fin_timeout and most of the connections would use the default timeout. The timeout may be very sensitive. For example if sysctl_tcp_fin_timeout is 10s, 20s may cause double the FIN-WAIT connections and server load needs to be reevaluated.
Besides from my experience a longer FIN-WAIT timeout will have a better experience for clients, because there’re some clients, such as video player, needing to finish the play before closing the connection. If we add a sysctl with default/max timeout for all users, all connections should be set to the max timeout except some special requirements which can also be set with TCP_LINGER2 to use a smaller one.
Maybe my experience is not correct for your scenes, so could you describe a situation that the timeout with default/max value is better?
Regards,
Cambda