net: usb: lan78xx: fix possible skb leak
PATCH 1/1 net: usb: lan78xx: fix possible skb leak
If skb_linearize() fails, we need to free the skb.
TSO makes skb bigger, and this bug might be the reason Raspberry Pi 3B+ users had to disable TSO.
--- drivers/net/usb/lan78xx.c | 9 +++------ 1 file changed, 3 insertions(+), 6 deletions(-)
drivers/net/usb/lan78xx.c +3 -6
@@ -2724,11 +2724,6 @@ static int lan78xx_stop(struct net_device *net)return 0 ; } -static int lan78xx_linearize (struct sk_buff *skb) -{ -return skb_linearize (skb); -} -static struct sk_buff *lan78xx_tx_prep (struct lan78xx_net *dev,struct sk_buff *skb,gfp_t flags) {
@@ -2740,8 +2735,10 @@ static struct sk_buff *lan78xx_tx_prep(struct lan78xx_net *dev,return NULL ; } -if (lan78xx_linearize (skb) <0 ) +if (skb_linearize (skb)) { +dev_kfree_skb_any (skb);return NULL ; + } tx_cmd_a = (u32)(skb->len &TX_CMD_A_LEN_MASK_ ) |TX_CMD_A_FCS_ ;
Signature · 1 line
2.24.1.735.g03f4e72817-goog
David Miller
maintainerFrom: Eric Dumazet <[email protected]>
Date: Tue, 7 Jan 2020 10:57:01 -0800
Quoted 8 lines
If
skb_linearize()fails, we need to free the skb.TSO makes skb bigger, and this bug might be the reason Raspberry Pi 3B+ users had to disable TSO.
Fixes 55d7de9de6c3 ("Microchip's LAN7800 family USB 2/3 to 10/100/1000 Ethernet device driver")Signed-off-by Eric Dumazet <[email protected]>Reported-by RENARD Pierre-Francois <[email protected]>
Applied and queued up for -stable, thanks.
I tried with last rawhide kernel 5.5.0-0.rc5.git0.1.local.fc32.aarch64 I compiled it this night. (I check it includes the patch for lan78xx.c )
Both tests (scp and nfs ) are failing the same way as before.
Fox
Quoted 43 lines · Eric Dumazet
If
skb_linearize()fails, we need to free the skb.TSO makes skb bigger, and this bug might be the reason Raspberry Pi 3B+ users had to disable TSO.
Fixes 55d7de9de6c3 ("Microchip's LAN7800 family USB 2/3 to 10/100/1000 Ethernet device driver")Signed-off-by Eric Dumazet <[email protected]>Reported-by RENARD Pierre-Francois <[email protected]>Cc Stefan Wahren <[email protected]>Cc Woojung Huh <[email protected]>Cc Microchip Linux Driver Support <[email protected]>--- drivers/net/usb/lan78xx.c | 9 +++------ 1 file changed, 3 insertions(+), 6 deletions(-)drivers/net/usb/lan78xx.c +3
-6@@ -2724,11 +2724,6 @@ static int lan78xx_stop(struct net_device *net)return 0 ; } -static int lan78xx_linearize (struct sk_buff *skb) -{ -return skb_linearize (skb); -} -static struct sk_buff *lan78xx_tx_prep (struct lan78xx_net *dev,struct sk_buff *skb,gfp_t flags) {@@ -2740,8 +2735,10 @@ static struct sk_buff *lan78xx_tx_prep(struct lan78xx_net *dev,return NULL ; } -if (lan78xx_linearize (skb) <0 ) +if (skb_linearize (skb)) { +dev_kfree_skb_any (skb);return NULL ; + } tx_cmd_a = (u32)(skb->len &TX_CMD_A_LEN_MASK_ ) |TX_CMD_A_FCS_ ;
Quoted 8 lines · RENARD Pierre-Francois
I tried with last rawhide kernel 5.5.0-0.rc5.git0.1.local.fc32.aarch64 I compiled it this night. (I check it includes the patch for
lan78xx.c)Both tests (scp and nfs ) are failing the same way as before.
Fox
Please report the output of " grep skb /proc/slabinfo"
before and after your test.
The symptoms (of retransmit being not attempted by TCP) match the fact that skb(s) is(are) not freed by a driver (or some layer)
When TCP detects this (function skb_still_in_host_queue()), one SNMP counter is incremented
nstat -a | grep TCPSpuriousRtxHostQueues
Thanks.
Quoted 48 lines · Eric Dumazet
Quoted 43 lines
If
skb_linearize()fails, we need to free the skb.TSO makes skb bigger, and this bug might be the reason Raspberry Pi 3B+ users had to disable TSO.
Fixes 55d7de9de6c3 ("Microchip's LAN7800 family USB 2/3 to 10/100/1000 Ethernet device driver")Signed-off-by Eric Dumazet <[email protected]>Reported-by RENARD Pierre-Francois <[email protected]>Cc Stefan Wahren <[email protected]>Cc Woojung Huh <[email protected]>Cc Microchip Linux Driver Support <[email protected]>--- drivers/net/usb/lan78xx.c | 9 +++------ 1 file changed, 3 insertions(+), 6 deletions(-)drivers/net/usb/lan78xx.c +3
-6@@ -2724,11 +2724,6 @@ static int lan78xx_stop(struct net_device *net)return 0 ; } -static int lan78xx_linearize (struct sk_buff *skb) -{ -return skb_linearize (skb); -} -static struct sk_buff *lan78xx_tx_prep (struct lan78xx_net *dev,struct sk_buff *skb,gfp_t flags) {@@ -2740,8 +2735,10 @@ static struct sk_buff *lan78xx_tx_prep(struct lan78xx_net *dev,return NULL ; } -if (lan78xx_linearize (skb) <0 ) +if (skb_linearize (skb)) { +dev_kfree_skb_any (skb);return NULL ; + } tx_cmd_a = (u32)(skb->len &TX_CMD_A_LEN_MASK_ ) |TX_CMD_A_FCS_ ;
OK
Before scp command ( and after a fresh reboot)
---------------------------------
skbuff_ext_cache 378 378 192 21 1 : tunables 0 0 0 : slabdata 18 18 0 skbuff_fclone_cache 112 112 512 16 2 : tunables 0 0 0 : slabdata 7 7 0 skbuff_head_cache 1936 2160 256 16 1 : tunables 0 0 0 : slabdata 135 135 0
---------------------------------
---------------------------------
After the hang of scp (hanged at 203 MB)
---------------------------------
skbuff_ext_cache 693 693 192 21 1 : tunables 0 0 0 : slabdata 33 33 0 skbuff_fclone_cache 128 128 512 16 2 : tunables 0 0 0 : slabdata 8 8 0 skbuff_head_cache 2032 2176 256 16 1 : tunables 0 0 0 : slabdata 136 136 0
---------------------------------
TcpExtTCPSpuriousRtxHostQueues 120 0.0
---------------------------------
After CTRL-C of scp
---------------------------------
skbuff_ext_cache 693 693 192 21 1 : tunables 0 0 0 : slabdata 33 33 0 skbuff_fclone_cache 128 128 512 16 2 : tunables 0 0 0 : slabdata 8 8 0 skbuff_head_cache 2112 2336 256 16 1 : tunables 0 0 0 : slabdata 146 146 0
---------------------------------
TcpExtTCPSpuriousRtxHostQueues 124 0.0
---------------------------------
After the hang of a second attempt (hanged at 1214 MB)
---------------------------------
skbuff_ext_cache 735 735 192 21 1 : tunables 0 0 0 : slabdata 35 35 0 skbuff_fclone_cache 160 160 512 16 2 : tunables 0 0 0 : slabdata 10 10 0 skbuff_head_cache 2096 2240 256 16 1 : tunables 0 0 0 : slabdata 140 140 0
---------------------------------
TcpExtTCPSpuriousRtxHostQueues 248 0.0
---------------------------------
After a third attempt (hanged at 55 MB)
---------------------------------
skbuff_ext_cache 735 735 192 21 1 : tunables 0 0 0 : slabdata 35 35 0 skbuff_fclone_cache 176 176 512 16 2 : tunables 0 0 0 : slabdata 11 11 0 skbuff_head_cache 2000 2144 256 16 1 : tunables 0 0 0 : slabdata 134 134 0
---------------------------------
TcpExtTCPSpuriousRtxHostQueues 365 0.0
---------------------------------
Quoted 69 lines · Eric Dumazet
Quoted 7 lines · RENARD Pierre-Francois
I tried with last rawhide kernel 5.5.0-0.rc5.git0.1.local.fc32.aarch64 I compiled it this night. (I check it includes the patch for
lan78xx.c)Both tests (scp and nfs ) are failing the same way as before.
Fox
Please report the output of " grep skb
/proc/slabinfo"before and after your test.
The symptoms (of retransmit being not attempted by TCP) match the fact that skb(s) is(are) not freed by a driver (or some layer)
When TCP detects this (function
skb_still_in_host_queue()), one SNMP counter is incrementednstat -a | grep TCPSpuriousRtxHostQueues
Thanks.
Quoted 46 lines · Eric Dumazet
Quoted 43 lines
If
skb_linearize()fails, we need to free the skb.TSO makes skb bigger, and this bug might be the reason Raspberry Pi 3B+ users had to disable TSO.
Fixes 55d7de9de6c3 ("Microchip's LAN7800 family USB 2/3 to 10/100/1000 Ethernet device driver")Signed-off-by Eric Dumazet <[email protected]>Reported-by RENARD Pierre-Francois <[email protected]>Cc Stefan Wahren <[email protected]>Cc Woojung Huh <[email protected]>Cc Microchip Linux Driver Support <[email protected]>--- drivers/net/usb/lan78xx.c | 9 +++------ 1 file changed, 3 insertions(+), 6 deletions(-)drivers/net/usb/lan78xx.c +3
-6@@ -2724,11 +2724,6 @@ static int lan78xx_stop(struct net_device *net)return 0 ; } -static int lan78xx_linearize (struct sk_buff *skb) -{ -return skb_linearize (skb); -} -static struct sk_buff *lan78xx_tx_prep (struct lan78xx_net *dev,struct sk_buff *skb,gfp_t flags) {@@ -2740,8 +2735,10 @@ static struct sk_buff *lan78xx_tx_prep(struct lan78xx_net *dev,return NULL ; } -if (lan78xx_linearize (skb) <0 ) +if (skb_linearize (skb)) { +dev_kfree_skb_any (skb);return NULL ; + } tx_cmd_a = (u32)(skb->len &TX_CMD_A_LEN_MASK_ ) |TX_CMD_A_FCS_ ;
Quoted 40 lines · RENARD Pierre-Francois
OK
Before scp command ( and after a fresh reboot)
---------------------------------
skbuff_ext_cache 378 378 192 21 1 : tunables 0 0 0 : slabdata 18 18 0 skbuff_fclone_cache 112 112 512 16 2 : tunables 0 0 0 : slabdata 7 7 0 skbuff_head_cache 1936 2160 256 16 1 : tunables 0 0 0 : slabdata 135 135 0
---------------------------------
---------------------------------After the hang of scp (hanged at 203 MB)
---------------------------------
skbuff_ext_cache 693 693 192 21 1 : tunables 0 0 0 : slabdata 33 33 0 skbuff_fclone_cache 128 128 512 16 2 : tunables 0 0 0 : slabdata 8 8 0 skbuff_head_cache 2032 2176 256 16 1 : tunables 0 0 0 : slabdata 136 136 0
---------------------------------
TcpExtTCPSpuriousRtxHostQueues 120 0.0
---------------------------------After CTRL-C of scp
---------------------------------
skbuff_ext_cache 693 693 192 21 1 : tunables 0 0 0 : slabdata 33 33 0 skbuff_fclone_cache 128 128 512 16 2 : tunables 0 0 0 : slabdata 8 8 0 skbuff_head_cache 2112 2336 256 16 1 : tunables 0 0 0 : slabdata 146 146 0
---------------------------------
TcpExtTCPSpuriousRtxHostQueues 124 0.0
---------------------------------After the hang of a second attempt (hanged at 1214 MB)
---------------------------------
skbuff_ext_cache 735 735 192 21 1 : tunables 0 0 0 : slabdata 35 35 0 skbuff_fclone_cache 160 160 512 16 2 : tunables 0 0 0 : slabdata 10 10 0 skbuff_head_cache 2096 2240 256 16 1 : tunables 0 0 0 : slabdata 140 140 0
---------------------------------
TcpExtTCPSpuriousRtxHostQueues 248 0.0
---------------------------------
Quoted 10 lines
After a third attempt (hanged at 55 MB)
---------------------------------
skbuff_ext_cache 735 735 192 21 1 : tunables 0 0 0 : slabdata 35 35 0 skbuff_fclone_cache 176 176 512 16 2 : tunables 0 0 0 : slabdata 11 11 0 skbuff_head_cache 2000 2144 256 16 1 : tunables 0 0 0 : slabdata 134 134 0
---------------------------------
TcpExtTCPSpuriousRtxHostQueues 365 0.0
---------------------------------
Thanks for testing.
This seems to suggest there is another bug in the driver, leading to some skb being never freed.
Since the driver seems to limit aggregation to 9000 bytes (MAX_SINGLE_PACKET_SIZE) I wonder if gso skbs should also be limited to the same value.
drivers/net/usb/lan78xx.c +2 -0
@@ -3787,6 +3787,8 @@ static int lan78xx_probe(struct usb_interface *intf,if (ret <0 ) goto out4; +netif_set_gso_max_size (netdev,MAX_SINGLE_PACKET_SIZE -MAX_HEADER ); + ret =register_netdev (netdev);if (ret !=0 ) {netif_err (dev, probe, netdev,"couldn't register the device\n" );
Quoted 77 lines · Eric Dumazet
Quoted 69 lines
Quoted 7 lines · RENARD Pierre-Francois
I tried with last rawhide kernel 5.5.0-0.rc5.git0.1.local.fc32.aarch64 I compiled it this night. (I check it includes the patch for
lan78xx.c)Both tests (scp and nfs ) are failing the same way as before.
Fox
Please report the output of " grep skb
/proc/slabinfo"before and after your test.
The symptoms (of retransmit being not attempted by TCP) match the fact that skb(s) is(are) not freed by a driver (or some layer)
When TCP detects this (function
skb_still_in_host_queue()), one SNMP counter is incrementednstat -a | grep TCPSpuriousRtxHostQueues
Thanks.
Quoted 46 lines · Eric Dumazet
Quoted 43 lines
If
skb_linearize()fails, we need to free the skb.TSO makes skb bigger, and this bug might be the reason Raspberry Pi 3B+ users had to disable TSO.
Fixes 55d7de9de6c3 ("Microchip's LAN7800 family USB 2/3 to 10/100/1000 Ethernet device driver")Signed-off-by Eric Dumazet <[email protected]>Reported-by RENARD Pierre-Francois <[email protected]>Cc Stefan Wahren <[email protected]>Cc Woojung Huh <[email protected]>Cc Microchip Linux Driver Support <[email protected]>---
drivers/net/usb/lan78xx.c| 9 +++------1 file changed, 3 insertions(+), 6 deletions(-)drivers/net/usb/lan78xx.c +3
-6@@ -2724,11 +2724,6 @@ static int lan78xx_stop(struct net_device *net) return 0; } -static int lan78xx_linearize (struct sk_buff *skb) -{ - return skb_linearize(skb); -} - static struct sk_buff *lan78xx_tx_prep(struct lan78xx_net *dev, struct sk_buff *skb, gfp_t flags) {@@ -2740,8 +2735,10 @@ static struct sk_buff *lan78xx_tx_prep(struct lan78xx_net *dev, return NULL; } - if (lan78xx_linearize(skb) < 0) + if (skb_linearize(skb)) { + dev_kfree_skb_any(skb); return NULL; + } tx_cmd_a = (u32)(skb->len & TX_CMD_A_LEN_MASK_) | TX_CMD_A_FCS_;
Hello,
I applied patch and I am NOT able to reproduce the issue with both NFS or SCP !
Congratulations !
Fox
Quoted 149 lines · Eric Dumazet
Quoted 40 lines · RENARD Pierre-Francois
OK
Before scp command ( and after a fresh reboot)
---------------------------------
skbuff_ext_cache 378 378 192 21 1 : tunables 0 0 0 : slabdata 18 18 0 skbuff_fclone_cache 112 112 512 16 2 : tunables 0 0 0 : slabdata 7 7 0 skbuff_head_cache 1936 2160 256 16 1 : tunables 0 0 0 : slabdata 135 135 0
---------------------------------
---------------------------------After the hang of scp (hanged at 203 MB)
---------------------------------
skbuff_ext_cache 693 693 192 21 1 : tunables 0 0 0 : slabdata 33 33 0 skbuff_fclone_cache 128 128 512 16 2 : tunables 0 0 0 : slabdata 8 8 0 skbuff_head_cache 2032 2176 256 16 1 : tunables 0 0 0 : slabdata 136 136 0
---------------------------------
TcpExtTCPSpuriousRtxHostQueues 120 0.0
---------------------------------After CTRL-C of scp
---------------------------------
skbuff_ext_cache 693 693 192 21 1 : tunables 0 0 0 : slabdata 33 33 0 skbuff_fclone_cache 128 128 512 16 2 : tunables 0 0 0 : slabdata 8 8 0 skbuff_head_cache 2112 2336 256 16 1 : tunables 0 0 0 : slabdata 146 146 0
---------------------------------
TcpExtTCPSpuriousRtxHostQueues 124 0.0
---------------------------------After the hang of a second attempt (hanged at 1214 MB)
---------------------------------
skbuff_ext_cache 735 735 192 21 1 : tunables 0 0 0 : slabdata 35 35 0 skbuff_fclone_cache 160 160 512 16 2 : tunables 0 0 0 : slabdata 10 10 0 skbuff_head_cache 2096 2240 256 16 1 : tunables 0 0 0 : slabdata 140 140 0
---------------------------------
TcpExtTCPSpuriousRtxHostQueues 248 0.0
---------------------------------Quoted 10 lines
After a third attempt (hanged at 55 MB)
---------------------------------
skbuff_ext_cache 735 735 192 21 1 : tunables 0 0 0 : slabdata 35 35 0 skbuff_fclone_cache 176 176 512 16 2 : tunables 0 0 0 : slabdata 11 11 0 skbuff_head_cache 2000 2144 256 16 1 : tunables 0 0 0 : slabdata 134 134 0
---------------------------------
TcpExtTCPSpuriousRtxHostQueues 365 0.0
---------------------------------Thanks for testing.
This seems to suggest there is another bug in the driver, leading to some skb being never freed.
Since the driver seems to limit aggregation to 9000 bytes (
MAX_SINGLE_PACKET_SIZE) I wonder if gso skbs should also be limited to the same value.drivers/net/usb/lan78xx.c +2
-0@@ -3787,6 +3787,8 @@ static int lan78xx_probe(struct usb_interface *intf,if (ret <0 ) goto out4; +netif_set_gso_max_size (netdev,MAX_SINGLE_PACKET_SIZE -MAX_HEADER ); + ret =register_netdev (netdev);if (ret !=0 ) {netif_err (dev, probe, netdev,"couldn't register the device\n" );Quoted 76 lines · Eric Dumazet
Quoted 67 lines
Quoted 7 lines · RENARD Pierre-Francois
I tried with last rawhide kernel 5.5.0-0.rc5.git0.1.local.fc32.aarch64 I compiled it this night. (I check it includes the patch for
lan78xx.c)Both tests (scp and nfs ) are failing the same way as before.
Fox
Please report the output of " grep skb
/proc/slabinfo"before and after your test.
The symptoms (of retransmit being not attempted by TCP) match the fact that skb(s) is(are) not freed by a driver (or some layer)
When TCP detects this (function
skb_still_in_host_queue()), one SNMP counter is incrementednstat -a | grep TCPSpuriousRtxHostQueues
Thanks.
Quoted 44 lines · Eric Dumazet
Quoted 43 lines
If
skb_linearize()fails, we need to free the skb.TSO makes skb bigger, and this bug might be the reason Raspberry Pi 3B+ users had to disable TSO.
Fixes 55d7de9de6c3 ("Microchip's LAN7800 family USB 2/3 to 10/100/1000 Ethernet device driver")Signed-off-by Eric Dumazet <[email protected]>Reported-by RENARD Pierre-Francois <[email protected]>Cc Stefan Wahren <[email protected]>Cc Woojung Huh <[email protected]>Cc Microchip Linux Driver Support <[email protected]>--- drivers/net/usb/lan78xx.c | 9 +++------ 1 file changed, 3 insertions(+), 6 deletions(-)drivers/net/usb/lan78xx.c +3
-6@@ -2724,11 +2724,6 @@ static int lan78xx_stop(struct net_device *net) return 0; } -static int lan78xx_linearize (struct sk_buff *skb) -{ - return skb_linearize(skb); -} - static struct sk_buff *lan78xx_tx_prep(struct lan78xx_net *dev, struct sk_buff *skb, gfp_t flags) {@@ -2740,8 +2735,10 @@ static struct sk_buff *lan78xx_tx_prep(struct lan78xx_net *dev, return NULL; } - if (lan78xx_linearize(skb) < 0) + if (skb_linearize(skb)) { + dev_kfree_skb_any(skb); return NULL; + } tx_cmd_a = (u32)(skb->len & TX_CMD_A_LEN_MASK_) | TX_CMD_A_FCS_;