3.12.48-stable review
PATCH 2/33 drm/radeon: fix hotplug race at startup
In reply to Jiri SlabyFrom: Dave Airlie <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
commit 7f98ca454ad373fc1b76be804fa7138ff68c1d27 upstream.
We apparantly get a hotplug irq before we've initialised modesetting,
[drm] Loading R100 Microcode
BUG: unable to handle kernel NULL pointer dereference at (null) IP: [<c125f56f>] __mutex_lock_slowpath+0x23/0x91 *pde = 00000000 Oops: 0002 [#1] Modules linked in: radeon(+) drm_kms_helper ttm drm i2c_algo_bit backlight pcspkr psmouse evdev sr_mod input_leds led_class cdrom sg parport_pc parport floppy intel_agp intel_gtt lpc_ich acpi_cpufreq processor button mfd_core agpgart uhci_hcd ehci_hcd rng_core snd_intel8x0 snd_ac97_codec ac97_bus snd_pcm usbcore usb_common i2c_i801 i2c_core snd_timer snd soundcore thermal_sys CPU: 0 PID: 15 Comm: kworker/0:1 Not tainted 4.2.0-rc7-00015-gbf67402 #111 Hardware name: MicroLink /D850MV , BIOS MV85010A.86A.0067.P24.0304081124 04/08/2003 Workqueue: events radeon_hotplug_work_func [radeon] task: f6ca5900 ti: f6d3e000 task.ti: f6d3e000 EIP: 0060:[<c125f56f>] EFLAGS: 00010282 CPU: 0 EIP is at __mutex_lock_slowpath+0x23/0x91 EAX: 00000000 EBX: f5e900fc ECX: 00000000 EDX: fffffffe ESI: f6ca5900 EDI: f5e90100 EBP: f5e90000 ESP: f6d3ff0c DS: 007b ES: 007b FS: 0000 GS: 0000 SS: 0068 CR0: 8005003b CR2: 00000000 CR3: 36f61000 CR4: 000006d0 Stack: f5e90100 00000000 c103c4c1 f6d2a5a0 f5e900fc f6df394c c125f162 f8b0faca f6d2a5a0 c138ca00 f6df394c f7395600 c1034741 00d40000 00000000 f6d2a5a0 c138ca00 f6d2a5b8 c138ca10 c1034b58 00000001 f6d40000 f6ca5900 f6d0c940 Call Trace: [<c103c4c1>] ? dequeue_task_fair+0xa4/0xb7 [<c125f162>] ? mutex_lock+0x9/0xa [<f8b0faca>] ? radeon_hotplug_work_func+0x17/0x57 [radeon] [<c1034741>] ? process_one_work+0xfc/0x194 [<c1034b58>] ? worker_thread+0x18d/0x218 [<c10349cb>] ? rescuer_thread+0x1d5/0x1d5 [<c103742a>] ? kthread+0x7b/0x80 [<c12601c0>] ? ret_from_kernel_thread+0x20/0x30 [<c10373af>] ? init_completion+0x18/0x18 Code: 42 08 e8 8e a6 dd ff c3 57 56 53 83 ec 0c 8b 35 48 f7 37 c1 8b 10 4a 74 1a 89 c3 8d 78 04 8b 40 08 89 63
--- drivers/gpu/drm/radeon/radeon_irq_kms.c | 5 +++++ 1 file changed, 5 insertions(+)
drivers/gpu/drm/radeon/radeon_irq_kms.c +5 -0
@@ -73,6 +73,11 @@ static void radeon_hotplug_work_func(struct work_struct *work)struct drm_mode_config *mode_config = &dev->mode_config;struct drm_connector *connector; +/* we can race here at startup, some boards seem to trigger +* hotplug irqs when they shouldn't. */ +if (!rdev->mode_info.mode_config_initialized) +return ; +mutex_lock (&mode_config->mutex);if (mode_config->num_connector) {list_for_each_entry (connector, &mode_config->connector_list, head)
Signature · 1 line
2.5.2
PATCH 6/33 ip_tunnel: fix ipv4 pmtu check to honor inner ip header df
In reply to Jiri SlabyFrom: Timo Teräs <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
[ Upstream commit fc24f2b2094366da8786f59f2606307e934cea17 ]
Frag needed should be sent only if the inner header asked to not fragment. Currently fragmentation is broken if the tunnel has df set, but df was not asked in the original packet. The tunnel's df needs to be still checked to update internally the pmtu cache.
Commit 23a3647bc4f93bac broke it, and this commit fixes the ipv4 df check back to the way it was.
--- net/ipv4/ip_tunnel.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-)
net/ipv4/ip_tunnel.c +5 -3
@@ -476,7 +476,8 @@ drop:EXPORT_SYMBOL_GPL (ip_tunnel_rcv);static int tnl_update_pmtu (struct net_device *dev,struct sk_buff *skb, -struct rtable *rt,__be16 df) +struct rtable *rt,__be16 df, +const struct iphdr *inner_iph) {struct ip_tunnel *tunnel =netdev_priv (dev);int pkt_size = skb->len - tunnel->hlen - dev->hard_header_len;
@@ -493,7 +494,8 @@ static int tnl_update_pmtu(struct net_device *dev, struct sk_buff *skb,if (skb->protocol ==htons (ETH_P_IP )) {if (!skb_is_gso (skb) && - (df &htons (IP_DF )) && mtu < pkt_size) { + (inner_iph->frag_off &htons (IP_DF )) && + mtu < pkt_size) {memset (IPCB (skb),0 ,sizeof (*IPCB (skb)));icmp_send (skb,ICMP_DEST_UNREACH ,ICMP_FRAG_NEEDED ,htonl (mtu));return -E2BIG ;
@@ -611,7 +613,7 @@ void ip_tunnel_xmit(struct sk_buff *skb, struct net_device *dev, goto tx_error; } -if (tnl_update_pmtu (dev, skb, rt, tnl_params->frag_off)) { +if (tnl_update_pmtu (dev, skb, rt, tnl_params->frag_off, inner_iph)) {ip_rt_put (rt); goto tx_error; }
Signature · 1 line
2.5.2
PATCH 7/33 net/tipc: initialize security state for new connection socket
In reply to Jiri SlabyFrom: Stephen Smalley <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
[ Upstream commit fdd75ea8df370f206a8163786e7470c1277a5064 ]
Calling connect() with an AF_TIPC socket would trigger a series of error messages from SELinux along the lines of:
SELinux: Invalid class 0
type=AVC msg=audit(1434126658.487:34500): avc: denied { <unprintable> }
for pid=292 comm="kworker/u16:5" scontext=system_u:system_r:kernel_t:s0 tcontext=system_u:object_r:unlabeled_t:s0 tclass=<unprintable> permissive=0
This was due to a failure to initialize the security state of the new connection sock by the tipc code, leaving it with junk in the security class field and an unlabeled secid. Add a call to security_sk_clone() to inherit the security state from the parent socket.
--- net/tipc/socket.c | 1 + 1 file changed, 1 insertion(+)
net/tipc/socket.c +1 -0
@@ -1607,6 +1607,7 @@ static int accept(struct socket *sock, struct socket *new_sock, int flags) res =tipc_sk_create (sock_net (sock->sk), new_sock,0 ,1 );if (res) goto exit; +security_sk_clone (sock->sk, new_sock->sk); new_sk = new_sock->sk; new_tsock =tipc_sk (new_sk);
Signature · 1 line
2.5.2
PATCH 9/33 net: pktgen: fix race between pktgen_thread_worker() and kthread_stop()
In reply to Jiri SlabyFrom: Oleg Nesterov <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
[ Upstream commit fecdf8be2d91e04b0a9a4f79ff06499a36f5d14f ]
pktgen_thread_worker() is obviously racy, kthread_stop() can come between the kthread_should_stop() check and set_current_state().
--- net/core/pktgen.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)
net/core/pktgen.c +3 -1
@@ -3423,8 +3423,10 @@ static int pktgen_thread_worker(void *arg)pktgen_rem_thread (t);/* Wait for kthread_stop */ -while (!kthread_should_stop ()) { +for (;;) {set_current_state (TASK_INTERRUPTIBLE ); +if (kthread_should_stop ()) +break ;schedule (); }__set_current_state (TASK_RUNNING );
Signature · 1 line
2.5.2
PATCH 11/33 net: call rcu_read_lock early in process_backlog
In reply to Jiri SlabyFrom: Julian Anastasov <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
[ Upstream commit 2c17d27c36dcce2b6bf689f41a46b9e909877c21 ]
Incoming packet should be either in backlog queue or in RCU read-side section. Otherwise, the final sequence of flush_backlog() and synchronize_net() may miss packets that can run without device reference:
CPU 1 CPU 2
skb->dev: no reference
process_backlog:__skb_dequeue
process_backlog:local_irq_enableon_each_cpu forflush_backlog => IPI(hardirq): flush_backlog
- packet not found in backlog
CPU delayed ...synchronize_net
- no ongoing RCU
read-side sections
netdev_run_todo,rcu_barrier: no
ongoing callbacks
__netif_receive_skb_core:rcu_read_lock
- too late
free dev
process packet for freed dev
--- net/core/dev.c | 29 ++++++++++++++--------------- 1 file changed, 14 insertions(+), 15 deletions(-)
net/core/dev.c +14 -15
@@ -3521,8 +3521,6 @@ static int __netif_receive_skb_core(struct sk_buff *skb, bool pfmemalloc) pt_prev =NULL ; -rcu_read_lock (); - another_round: skb->skb_iif = skb->dev->ifindex;
@@ -3532,7 +3530,7 @@ another_round: skb->protocol == cpu_to_be16(ETH_P_8021AD)) { skb =skb_vlan_untag (skb);if (unlikely (!skb)) - goto unlock; + goto out; }#ifdef CONFIG_NET_CLS_ACT
@@ -3557,7 +3555,7 @@ skip_taps:#ifdef CONFIG_NET_CLS_ACT skb =handle_ing (skb, &pt_prev, &ret, orig_dev);if (!skb) - goto unlock; + goto out; ncls:#endif
@@ -3572,7 +3570,7 @@ ncls:if (vlan_do_receive (&skb)) goto another_round;else if (unlikely (!skb)) - goto unlock; + goto out; } rx_handler =rcu_dereference (skb->dev->rx_handler);
@@ -3584,7 +3582,7 @@ ncls:switch (rx_handler (&skb)) {case RX_HANDLER_CONSUMED : ret =NET_RX_SUCCESS ; - goto unlock; + goto out;case RX_HANDLER_ANOTHER : goto another_round;case RX_HANDLER_EXACT :
@@ -3636,8 +3634,6 @@ drop: ret =NET_RX_DROP ; } -unlock: -rcu_read_unlock (); out:return ret; }
@@ -3684,29 +3680,30 @@ static int __netif_receive_skb(struct sk_buff *skb)*/ int netif_receive_skb (struct sk_buff *skb) { +int ret; +net_timestamp_check (netdev_tstamp_prequeue, skb);if (skb_defer_rx_timestamp (skb))return NET_RX_SUCCESS ; +rcu_read_lock (); +#ifdef CONFIG_RPS if (static_key_false (&rps_needed)) {struct rps_dev_flow voidflow, *rflow = &voidflow; -int cpu, ret; - -rcu_read_lock (); - - cpu =get_rps_cpu (skb->dev, skb, &rflow); +int cpu =get_rps_cpu (skb->dev, skb, &rflow);if (cpu >=0 ) { ret =enqueue_to_backlog (skb, cpu, &rflow->last_qtail);rcu_read_unlock ();return ret; } -rcu_read_unlock (); }#endif -return __netif_receive_skb (skb); + ret =__netif_receive_skb (skb); +rcu_read_unlock (); +return ret; }EXPORT_SYMBOL (netif_receive_skb);
@@ -4116,8 +4113,10 @@ static int process_backlog(struct napi_struct *napi, int quota)unsigned int qlen;while ((skb =__skb_dequeue (&sd->process_queue))) { +rcu_read_lock ();local_irq_enable ();__netif_receive_skb (skb); +rcu_read_unlock ();local_irq_disable ();input_queue_head_incr (sd);if (++work >= quota) {
Signature · 1 line
2.5.2
PATCH 13/33 net: Fix skb csum races when peeking
In reply to Jiri SlabyFrom: Herbert Xu <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
[ Upstream commit 89c22d8c3b278212eef6a8cc66b570bc840a6f5a ]
When we calculate the checksum on the recv path, we store the result in the skb as an optimisation in case we need the checksum again down the line.
This is in fact bogus for the MSG_PEEK case as this is done without any locking. So multiple threads can peek and then store the result to the same skb, potentially resulting in bogus skb states.
This patch fixes this by only storing the result if the skb is not shared. This preserves the optimisations for the few cases where it can be done safely due to locking or other reasons, e.g., SIOCINQ.
--- net/core/datagram.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
net/core/datagram.c +2 -1
@@ -777,7 +777,8 @@ __sum16 __skb_checksum_complete_head(struct sk_buff *skb, int len)if (likely (!sum)) {if (unlikely (skb->ip_summed ==CHECKSUM_COMPLETE ))netdev_rx_csum_fault (skb->dev); - skb->ip_summed =CHECKSUM_UNNECESSARY ; +if (!skb_shared (skb)) + skb->ip_summed =CHECKSUM_UNNECESSARY ; }return sum; }
Signature · 1 line
2.5.2
PATCH 15/33 bridge: mdb: fix double add notification
In reply to Jiri SlabyFrom: Nikolay Aleksandrov <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
[ Upstream commit 5ebc784625ea68a9570d1f70557e7932988cd1b4 ]
Since the mdb add/del code was introduced there have been 2 br_mdb_notify calls when doing br_mdb_add() resulting in 2 notifications on each add.
Example:
Command: bridge mdb add dev br0 port eth1 grp 239.0.0.1 permanent Before patch:
root@debian:~# bridge monitor all [MDB]dev br0 port eth1 grp 239.0.0.1 permanent [MDB]dev br0 port eth1 grp 239.0.0.1 permanent
After patch:
root@debian:~# bridge monitor all [MDB]dev br0 port eth1 grp 239.0.0.1 permanent
--- net/bridge/br_mdb.c | 1 - 1 file changed, 1 deletion(-)
net/bridge/br_mdb.c +0 -1
@@ -347,7 +347,6 @@ static int br_mdb_add_group(struct net_bridge *br, struct net_bridge_port *port,return -ENOMEM ;rcu_assign_pointer (*pp, p); -br_mdb_notify (br->dev, port, group,RTM_NEWMDB );return 0 ; }
Signature · 1 line
2.5.2
cover 3.12.48-stable review
This is the start of the stable review cycle for the 3.12.48 release. There are 33 patches in this series, all will be posted as a response to this one. If anyone has any issues with these being applied, please let me know.
Responses should be made by Thu Sep 17 16:20:25 CEST 2015. Anything received after that time might be too late.
The whole patch series can be found in one patch at:
http://kernel.org/pub/linux/kernel/people/jirislaby/stable-review/patch-3.12.48-rc1.xz
and the diffstat can be found below.
thanks,
js
===============
Akinobu Mita (1):
bio: fix argument of __bio_add_page() for max_sectors > 0xffff
Andrey Vagin (1):
netfilter: nf_conntrack: fix RCU race in nf_conntrack_find_get
Angga (1):
ipv6: Make MLD packets to only be processed locally
Benjamin LaHaise (1):
aio: fix reqs_available handling
Dan Carpenter (1):
rds: fix an integer overflow test in rds_info_getsockopt()
Daniel Borkmann (1):
rtnetlink: verify IFLA_VF_INFO attributes before passing them to
driver
Dave Airlie (1):
drm/radeon: fix hotplug race at startup
David Milburn (1):
mtip32xx: dynamically allocate buffer in debugfs functions
Edward Hyunkoo Jee (1):
inet: frags: fix defragmented packet's IP header for af_packet
Eric Dumazet (2):
net: graceful exit from netif_alloc_netdev_queues()
ipv6: lock socket in ip6_datagram_connect()
Florian Westphal (1):
netlink: don't hold mutex in rcu callback when releasing mmapd ring
Heinz Mauelshagen (1):
dm cache mq: fix memory allocation failure for large cache devices
Herbert Xu (3):
net: Clone skb before setting peeked flag
net: Fix skb csum races when peeking
net: Fix skb_set_peeked use-after-free bug
Jack Morgenstein (1):
net/mlx4_core: Fix wrong index in propagating port change event to VFs
James Smart (1):
lpfc: Fix scsi prep dma buf error.
Julian Anastasov (2):
net: do not process device backlog during unregistration
net: call rcu_read_lock early in process_backlog
Mark Rustad (2):
PCI: Add dev_flags bit to access VPD through function 0
PCI: Add VPD function 0 quirk for Intel Ethernet devices
Mika Westerberg (1):
mfd: lpc_ich: Assign subdevice ids automatically
Nikolay Aleksandrov (3):
bridge: mdb: zero out the local br_ip variable before use
bridge: mdb: fix double add notification
bonding: fix destruction of bond with devices different from
arphrd_ether
Oleg Nesterov (1):
net: pktgen: fix race between pktgen_thread_worker() and
kthread_stop()
Pablo Neira Ayuso (1):
netfilter: nf_conntrack: don't release a conntrack with non-zero
refcnt
Shirish Pargaonkar (1):
cifs: Send a logoff request before removing a smb session
Stephen Smalley (1):
net/tipc: initialize security state for new connection socket
Tilman Schmidt (1):
isdn/gigaset: reset tty->receive_room when attaching ser_gigaset
Timo Teräs (1):
ip_tunnel: fix ipv4 pmtu check to honor inner ip header df
dingtianhong (1):
bonding: correct the MAC address for "follow" fail_over_mac policy
drivers/block/mtip32xx/mtip32xx.c | 47 +++++++++--- drivers/gpu/drm/radeon/radeon_irq_kms.c | 5 ++ drivers/isdn/gigaset/ser-gigaset.c | 11 ++- drivers/md/dm-cache-policy-mq.c | 4 +- drivers/mfd/lpc_ich.c | 8 +- drivers/net/bonding/bond_main.c | 20 +++++ drivers/net/ethernet/mellanox/mlx4/eq.c | 4 +- drivers/pci/access.c | 61 ++++++++++++++- drivers/pci/quirks.c | 9 +++ drivers/scsi/lpfc/lpfc_scsi.c | 2 +- fs/aio.c | 77 ++++++++++++++++++- fs/bio.c | 2 +- fs/cifs/connect.c | 25 +++++-- fs/cifs/smb2transport.c | 10 ++- fs/cifs/transport.c | 11 ++- include/linux/pci.h | 2 + include/net/ip.h | 1 + include/net/netfilter/nf_conntrack.h | 2 + net/bridge/br_mdb.c | 3 +- net/core/datagram.c | 45 ++++++++++- net/core/dev.c | 38 +++++----- net/core/pktgen.c | 4 +- net/core/rtnetlink.c | 128 ++++++++++++++++---------------- net/ipv4/datagram.c | 16 +++- net/ipv4/ip_fragment.c | 7 +- net/ipv4/ip_tunnel.c | 8 +- net/ipv6/datagram.c | 20 +++-- net/ipv6/ip6_input.c | 6 +- net/netfilter/nf_conntrack_core.c | 55 +++++++++++--- net/netfilter/nf_synproxy_core.c | 5 +- net/netfilter/xt_CT.c | 7 +- net/netlink/af_netlink.c | 79 ++++++++++++-------- net/rds/info.c | 2 +- net/tipc/socket.c | 1 + 34 files changed, 535 insertions(+), 190 deletions(-)
Signature · 1 line
2.5.2
PATCH 21/33 netlink: don't hold mutex in rcu callback when releasing mmapd ring
In reply to Jiri SlabyFrom: Florian Westphal <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
[ Upstream commit 0470eb99b4721586ccac954faac3fa4472da0845 ]
Kirill A. Shutemov says:
This simple test-case trigers few locking asserts in kernel:
int main(int argc, char **argv)
{
unsigned int block_size = 16 * 4096;
struct nl_mmap_req req = {
.nm_block_size = block_size,
.nm_block_nr = 64,
.nm_frame_size = 16384,
.nm_frame_nr = 64 * block_size / 16384,
};
unsigned int ring_size;
int fd;
fd = socket(AF_NETLINK, SOCK_RAW, NETLINK_GENERIC);
if (setsockopt(fd, SOL_NETLINK, NETLINK_RX_RING, &req, sizeof(req)) < 0)
exit(1);
if (setsockopt(fd, SOL_NETLINK, NETLINK_TX_RING, &req, sizeof(req)) < 0)
exit(1);
ring_size = req.nm_block_nr * req.nm_block_size;
mmap(NULL, 2 * ring_size, PROT_READ|PROT_WRITE, MAP_SHARED, fd, 0);
return 0;
}+++ exited with 0 +++
BUG: sleeping function called from invalid context at /home/kas/git/public/linux-mm/kernel/locking/mutex.c:616
in_atomic(): 1, irqs_disabled(): 0, pid: 1, name: init
3 locks held by init/1:
#0: (reboot_mutex){+.+...}, at: [<ffffffff81080959>] SyS_reboot+0xa9/0x220
#1: ((reboot_notifier_list).rwsem){.+.+..}, at: [<ffffffff8107f379>] __blocking_notifier_call_chain+0x39/0x70
#2: (rcu_callback){......}, at: [<ffffffff810d32e0>] rcu_do_batch.isra.49+0x160/0x10c0Preemption disabled at:[<ffffffff8145365f>] __delay+0xf/0x20
CPU: 1 PID: 1 Comm: init Not tainted 4.1.0-00009-gbddf4c4818e0 #253 Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS Debian-1.8.2-1 04/01/2014 ffff88017b3d8000 ffff88027bc03c38 ffffffff81929ceb 0000000000000102 0000000000000000 ffff88027bc03c68 ffffffff81085a9d 0000000000000002 ffffffff81ca2a20 0000000000000268 0000000000000000 ffff88027bc03c98 Call Trace: <IRQ> [<ffffffff81929ceb>] dump_stack+0x4f/0x7b [<ffffffff81085a9d>] ___might_sleep+0x16d/0x270 [<ffffffff81085bed>] __might_sleep+0x4d/0x90 [<ffffffff8192e96f>] mutex_lock_nested+0x2f/0x430 [<ffffffff81932fed>] ? _raw_spin_unlock_irqrestore+0x5d/0x80 [<ffffffff81464143>] ? __this_cpu_preempt_check+0x13/0x20 [<ffffffff8182fc3d>] netlink_set_ring+0x1ed/0x350 [<ffffffff8182e000>] ? netlink_undo_bind+0x70/0x70 [<ffffffff8182fe20>] netlink_sock_destruct+0x80/0x150 [<ffffffff817e484d>] __sk_free+0x1d/0x160 [<ffffffff817e49a9>] sk_free+0x19/0x20
[..]
Cong Wang says:
We can't hold mutex lock in a rcu callback, [..]
Thomas Graf says:
The socket should be dead at this point. It might be simpler to add a netlink_release_ring() function which doesn't require locking at all.
--- net/netlink/af_netlink.c | 79 ++++++++++++++++++++++++++++-------------------- 1 file changed, 47 insertions(+), 32 deletions(-)
net/netlink/af_netlink.c +47 -32
@@ -342,25 +342,52 @@ err1:return NULL ; } + +static void +__netlink_set_ring (struct sock *sk,struct nl_mmap_req *req,bool tx_ring,void **pg_vec, +unsigned int order) +{ +struct netlink_sock *nlk =nlk_sk (sk); +struct sk_buff_head *queue; +struct netlink_ring *ring; + + queue = tx_ring ? &sk->sk_write_queue : &sk->sk_receive_queue; + ring = tx_ring ? &nlk->tx_ring : &nlk->rx_ring; + +spin_lock_bh (&queue->lock); + + ring->frame_max = req->nm_frame_nr -1 ; + ring->head =0 ; + ring->frame_size = req->nm_frame_size; + ring->pg_vec_pages = req->nm_block_size /PAGE_SIZE ; + +swap (ring->pg_vec_len, req->nm_block_nr); +swap (ring->pg_vec_order, order); +swap (ring->pg_vec, pg_vec); + +__skb_queue_purge (queue); +spin_unlock_bh (&queue->lock); + +WARN_ON (atomic_read (&nlk->mapped)); + +if (pg_vec) +free_pg_vec (pg_vec, order, req->nm_block_nr); +} +static int netlink_set_ring (struct sock *sk,struct nl_mmap_req *req, -bool closing,bool tx_ring) +bool tx_ring) {struct netlink_sock *nlk =nlk_sk (sk);struct netlink_ring *ring; -struct sk_buff_head *queue;void **pg_vec =NULL ;unsigned int order =0 ; -int err; ring = tx_ring ? &nlk->tx_ring : &nlk->rx_ring; - queue = tx_ring ? &sk->sk_write_queue : &sk->sk_receive_queue; -if (!closing) { -if (atomic_read (&nlk->mapped)) -return -EBUSY ; -if (atomic_read (&ring->pending)) -return -EBUSY ; - } +if (atomic_read (&nlk->mapped)) +return -EBUSY ; +if (atomic_read (&ring->pending)) +return -EBUSY ;if (req->nm_block_nr) {if (ring->pg_vec !=NULL )
@@ -392,31 +419,19 @@ static int netlink_set_ring(struct sock *sk, struct nl_mmap_req *req,return -EINVAL ; } - err = -EBUSY ;mutex_lock (&nlk->pg_vec_lock); -if (closing ||atomic_read (&nlk->mapped) ==0 ) { - err =0 ; -spin_lock_bh (&queue->lock); - - ring->frame_max = req->nm_frame_nr -1 ; - ring->head =0 ; - ring->frame_size = req->nm_frame_size; - ring->pg_vec_pages = req->nm_block_size /PAGE_SIZE ; - -swap (ring->pg_vec_len, req->nm_block_nr); -swap (ring->pg_vec_order, order); -swap (ring->pg_vec, pg_vec); - -__skb_queue_purge (queue); -spin_unlock_bh (&queue->lock); - -WARN_ON (atomic_read (&nlk->mapped)); +if (atomic_read (&nlk->mapped) ==0 ) { +__netlink_set_ring (sk, req, tx_ring, pg_vec, order); +mutex_unlock (&nlk->pg_vec_lock); +return 0 ; } +mutex_unlock (&nlk->pg_vec_lock);if (pg_vec)free_pg_vec (pg_vec, order, req->nm_block_nr); -return err; + +return -EBUSY ; }static void netlink_mm_open (struct vm_area_struct *vma)
@@ -885,10 +900,10 @@ static void netlink_sock_destruct(struct sock *sk)memset (&req,0 ,sizeof (req));if (nlk->rx_ring.pg_vec) -netlink_set_ring (sk, &req, true, false); +__netlink_set_ring (sk, &req, false,NULL ,0 );memset (&req,0 ,sizeof (req));if (nlk->tx_ring.pg_vec) - netlink_set_ring(sk, &req, true, true); + __netlink_set_ring(sk, &req, true, NULL, 0); }#endif /* CONFIG_NETLINK_MMAP */
@@ -2182,7 +2197,7 @@ static int netlink_setsockopt(struct socket *sock, int level, int optname,return -EINVAL ;if (copy_from_user (&req, optval,sizeof (req)))return -EFAULT ; - err =netlink_set_ring (sk, &req, false, + err =netlink_set_ring (sk, &req, optname ==NETLINK_TX_RING );break ; }
Signature · 1 line
2.5.2
PATCH 22/33 net/mlx4_core: Fix wrong index in propagating port change event to VFs
In reply to Jiri SlabyFrom: Jack Morgenstein <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
[ Upstream commit 1c1bf34951e8d17941bf708d1901c47e81b15d55 ]
The port-change event processing in procedure mlx4_eq_int() uses "slave" as the vf_oper array index. Since the value of "slave" is the PF function index, the result is that the PF link state is used for deciding to propagate the event for all the VFs. The VF link state should be used, so the VF function index should be used here.
--- drivers/net/ethernet/mellanox/mlx4/eq.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/net/ethernet/mellanox/mlx4/eq.c +2 -2
@@ -558,7 +558,7 @@ static int mlx4_eq_int(struct mlx4_dev *dev, struct mlx4_eq *eq)mlx4_dbg (dev,"%s: Sending MLX4_PORT_CHANGE_SUBTYPE_DOWN" " to slave: %d, port:%d\n" , __func__, i, port); - s_info = &priv->mfunc.master.vf_oper[slave].vport[port].state; + s_info = &priv->mfunc.master.vf_oper[i].vport[port].state;if (IFLA_VF_LINK_STATE_AUTO == s_info->link_state)mlx4_slave_event (dev, i, eqe); }else {/* IB port */
@@ -584,7 +584,7 @@ static int mlx4_eq_int(struct mlx4_dev *dev, struct mlx4_eq *eq)for (i =0 ; i < dev->num_slaves; i++) {if (i ==mlx4_master_func_num (dev))continue ; - s_info = &priv->mfunc.master.vf_oper[slave].vport[port].state; + s_info = &priv->mfunc.master.vf_oper[i].vport[port].state;if (IFLA_VF_LINK_STATE_AUTO == s_info->link_state)mlx4_slave_event (dev, i, eqe); }
Signature · 1 line
2.5.2
PATCH 24/33 mtip32xx: dynamically allocate buffer in debugfs functions
In reply to Jiri SlabyFrom: David Milburn <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
commit c8afd0dcbd14e2352258f2e2d359b36d0edd459f upstream.
Dynamically allocate buf to prevent warnings:
drivers/block/mtip32xx/mtip32xx.c: In function ‘mtip_hw_read_device_status’:drivers/block/mtip32xx/mtip32xx.c:2823: warning: the frame size of 1056 bytes is larger than 1024 bytes drivers/block/mtip32xx/mtip32xx.c: In function ‘mtip_hw_read_registers’:drivers/block/mtip32xx/mtip32xx.c:2894: warning: the frame size of 1056 bytes is larger than 1024 bytes drivers/block/mtip32xx/mtip32xx.c: In function ‘mtip_hw_read_flags’:drivers/block/mtip32xx/mtip32xx.c:2917: warning: the frame size of 1056 bytes is larger than 1024 bytes
--- drivers/block/mtip32xx/mtip32xx.c | 47 ++++++++++++++++++++++++++++++--------- 1 file changed, 37 insertions(+), 10 deletions(-)
drivers/block/mtip32xx/mtip32xx.c +37 -10
@@ -2810,34 +2810,51 @@ static ssize_t show_device_status(struct device_driver *drv, char *buf)static ssize_t mtip_hw_read_device_status (struct file *f,char __user *ubuf,size_t len,loff_t *offset) { +struct driver_data *dd = (struct driver_data *)f->private_data;int size = *offset; -char buf[MTIP_DFS_MAX_BUF_SIZE ]; +char *buf; +int rv =0 ;if (!len || *offset)return 0 ; + buf =kzalloc (MTIP_DFS_MAX_BUF_SIZE ,GFP_KERNEL ); +if (!buf) { +dev_err (&dd->pdev->dev, +"Memory allocation: status buffer\n" ); +return -ENOMEM ; + } + size +=show_device_status (NULL , buf); *offset = size <= len ? size : len; size =copy_to_user (ubuf, buf, *offset);if (size) -return -EFAULT ; + rv = -EFAULT ; -return *offset; +kfree (buf); +return rv ? rv : *offset; }static ssize_t mtip_hw_read_registers (struct file *f,char __user *ubuf,size_t len,loff_t *offset) {struct driver_data *dd = (struct driver_data *)f->private_data; -char buf[MTIP_DFS_MAX_BUF_SIZE ]; +char *buf;u32 group_allocated;int size = *offset; -int n; +int n, rv =0 ;if (!len || size)return 0 ; + buf =kzalloc (MTIP_DFS_MAX_BUF_SIZE ,GFP_KERNEL ); +if (!buf) { +dev_err (&dd->pdev->dev, +"Memory allocation: register buffer\n" ); +return -ENOMEM ; + } + size +=sprintf (&buf[size],"H/ S ACTive : [ 0x" );for (n = dd->slot_groups-1 ; n >=0 ; n--)
@@ -2892,21 +2909,30 @@ static ssize_t mtip_hw_read_registers(struct file *f, char __user *ubuf, *offset = size <= len ? size : len; size =copy_to_user (ubuf, buf, *offset);if (size) -return -EFAULT ; + rv = -EFAULT ; -return *offset; +kfree (buf); +return rv ? rv : *offset; }static ssize_t mtip_hw_read_flags (struct file *f,char __user *ubuf,size_t len,loff_t *offset) {struct driver_data *dd = (struct driver_data *)f->private_data; -char buf[MTIP_DFS_MAX_BUF_SIZE ]; +char *buf;int size = *offset; +int rv =0 ;if (!len || size)return 0 ; + buf =kzalloc (MTIP_DFS_MAX_BUF_SIZE ,GFP_KERNEL ); +if (!buf) { +dev_err (&dd->pdev->dev, +"Memory allocation: flag buffer\n" ); +return -ENOMEM ; + } + size +=sprintf (&buf[size],"Flag-port : [ %08lX ]\n" , dd->port->flags); size += sprintf(&buf[size], "Flag-dd : [ %08lX ]\n",
@@ -2915,9 +2941,10 @@ static ssize_t mtip_hw_read_flags(struct file *f, char __user *ubuf, *offset = size <= len ? size : len; size =copy_to_user (ubuf, buf, *offset);if (size) -return -EFAULT ; + rv = -EFAULT ; -return *offset; +kfree (buf); +return rv ? rv : *offset; }static const struct file_operations mtip_device_status_fops = {
Signature · 1 line
2.5.2
PATCH 25/33 cifs: Send a logoff request before removing a smb session
In reply to Jiri SlabyFrom: Shirish Pargaonkar <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
commit 7f48558e6489d032b1584b0cc9ac4bb11072c034 upstream.
Send a smb session logoff request before removing smb session off of the list. On a signed smb session, remvoing a session off of the list before sending a logoff request results in server returning an error for lack of smb signature.
Never seen an error during smb logoff, so as per MS-SMB2 3.2.5.1, not sure how an error during logoff should be retried. So for now, if a server returns an error to a logoff request, log the error and remove the session off of the list.
--- fs/cifs/connect.c | 25 ++++++++++++++++++++----- fs/cifs/smb2transport.c | 10 ++++++++-- fs/cifs/transport.c | 11 +++++++++-- 3 files changed, 37 insertions(+), 9 deletions(-)
fs/cifs/connect.c +20 -5
@@ -2245,6 +2245,8 @@ cifs_find_smb_ses(struct TCP_Server_Info *server, struct smb_vol *vol)spin_lock (&cifs_tcp_ses_lock);list_for_each_entry (ses, &server->smb_ses_list, smb_ses_list) { +if (ses->status == CifsExiting) +continue ;if (!match_session (ses, vol))continue ; ++ses->ses_count;
@@ -2258,24 +2260,37 @@ cifs_find_smb_ses(struct TCP_Server_Info *server, struct smb_vol *vol)static void cifs_put_smb_ses (struct cifs_ses *ses) { -unsigned int xid; +unsigned int rc, xid;struct TCP_Server_Info *server = ses->server;cifs_dbg (FYI ,"%s: ses_count=%d\n" , __func__, ses->ses_count); +spin_lock (&cifs_tcp_ses_lock); +if (ses->status == CifsExiting) { +spin_unlock (&cifs_tcp_ses_lock); +return ; + }if (--ses->ses_count >0 ) {spin_unlock (&cifs_tcp_ses_lock);return ; } - -list_del_init (&ses->smb_ses_list); +if (ses->status == CifsGood) + ses->status = CifsExiting;spin_unlock (&cifs_tcp_ses_lock); -if (ses->status == CifsGood && server->ops->logoff) { +if (ses->status == CifsExiting && server->ops->logoff) { xid =get_xid (); - server->ops->logoff (xid, ses); + rc = server->ops->logoff (xid, ses); +if (rc) +cifs_dbg (VFS ,"%s: Session Logoff failure rc=%d\n" , + __func__, rc);_free_xid (xid); } + +spin_lock (&cifs_tcp_ses_lock); +list_del_init (&ses->smb_ses_list); +spin_unlock (&cifs_tcp_ses_lock); +sesInfoFree (ses);cifs_put_tcp_session (server); }
fs/cifs/smb2transport.c +8 -2
@@ -516,13 +516,19 @@ smb2_get_mid_entry(struct cifs_ses *ses, struct smb2_hdr *buf,return -EAGAIN ; } -if (ses->status != CifsGood) { -/* check if SMB2 session is bad because we are setting it up */ +if (ses->status == CifsNew) {if ((buf->Command !=SMB2_SESSION_SETUP ) && (buf->Command !=SMB2_NEGOTIATE ))return -EAGAIN ;/* else ok - we are setting up session */ } + +if (ses->status == CifsExiting) { +if (buf->Command !=SMB2_LOGOFF ) +return -EAGAIN ; +/* else ok - we are shutting down the session */ + } + *mid =smb2_mid_entry_alloc (buf, ses->server);if (*mid ==NULL )return -ENOMEM ;
fs/cifs/transport.c +9 -2
@@ -431,13 +431,20 @@ static int allocate_mid(struct cifs_ses *ses, struct smb_hdr *in_buf,return -EAGAIN ; } -if (ses->status != CifsGood) { -/* check if SMB session is bad because we are setting it up */ +if (ses->status == CifsNew) {if ((in_buf->Command !=SMB_COM_SESSION_SETUP_ANDX ) && (in_buf->Command !=SMB_COM_NEGOTIATE ))return -EAGAIN ;/* else ok - we are setting up session */ } + +if (ses->status == CifsExiting) { +/* check if SMB session is bad because we are setting it up */ +if (in_buf->Command !=SMB_COM_LOGOFF_ANDX ) +return -EAGAIN ; +/* else ok - we are shutting down session */ + } + *ppmidQ =AllocMidQEntry (in_buf, ses->server);if (*ppmidQ ==NULL )return -ENOMEM ;
Signature · 1 line
2.5.2
PATCH 1/33 mfd: lpc_ich: Assign subdevice ids automatically
In reply to Jiri SlabyFrom: Mika Westerberg <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
commit 1abf25a25b86dcfe28d243a5af71bd1c9d6de1ef upstream.
Using -1 as platform device id means that the platform driver core will not assign any id to the device (the device name will not have id at all). This results problems on systems that have multiple PCHs (Platform Controller HUBs) because all of them also include their own copy of LPC device.
All the subsequent device creations will fail because there already exists platform device with the same name.
Fix this by passing PLATFORM_DEVID_AUTO as platform device id. This makes the platform device core to allocate new ids automatically.
--- drivers/mfd/lpc_ich.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)
drivers/mfd/lpc_ich.c +4 -4
@@ -872,8 +872,8 @@ gpe0_done:lpc_ich_enable_gpio_space (dev);lpc_ich_finalize_cell (dev, &lpc_ich_cells[LPC_GPIO ]); - ret =mfd_add_devices (&dev->dev,-1 , &lpc_ich_cells[LPC_GPIO ], -1 ,NULL ,0 ,NULL ); + ret =mfd_add_devices (&dev->dev,PLATFORM_DEVID_AUTO , + &lpc_ich_cells[LPC_GPIO ],1 ,NULL ,0 ,NULL ); gpio_done:if (acpi_conflict)
@@ -932,8 +932,8 @@ static int lpc_ich_init_wdt(struct pci_dev *dev) }lpc_ich_finalize_cell (dev, &lpc_ich_cells[LPC_WDT ]); - ret =mfd_add_devices (&dev->dev,-1 , &lpc_ich_cells[LPC_WDT ], -1 ,NULL ,0 ,NULL ); + ret =mfd_add_devices (&dev->dev,PLATFORM_DEVID_AUTO , + &lpc_ich_cells[LPC_WDT ],1 ,NULL ,0 ,NULL ); wdt_done:return ret;
Signature · 1 line
2.5.2
PATCH 3/33 ipv6: Make MLD packets to only be processed locally
From: Angga <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
[ Upstream commit 4c938d22c88a9ddccc8c55a85e0430e9c62b1ac5 ]
Before commit daad151263cf ("ipv6: Make ipv6_is_mld() inline and use it from ip6_mc_input().") MLD packets were only processed locally. After the change, a copy of MLD packet goes through ip6_mr_input, causing MRT6MSG_NOCACHE message to be generated to user space.
Make MLD packet only processed locally.
--- net/ipv6/ip6_input.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-)
net/ipv6/ip6_input.c +3 -3
@@ -329,10 +329,10 @@ int ip6_mc_input(struct sk_buff *skb)if (offset <0 ) goto out; -if (!ipv6_is_mld (skb, nexthdr, offset)) - goto out; +if (ipv6_is_mld (skb, nexthdr, offset)) + deliver = true; - deliver = true; + goto out; }/* unknown RA - process it normally */ }
Signature · 1 line
2.5.2
PATCH 4/33 net: graceful exit from netif_alloc_netdev_queues()
In reply to Jiri SlabyFrom: Eric Dumazet <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
[ Upstream commit d339727c2b1a10f25e6636670ab6e1841170e328 ]
User space can crash kernel with
ip link add ifb10 numtxqueues 100000 type ifb
We must replace a BUG_ON() by proper test and return -EINVAL for crazy values.
--- net/core/dev.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
net/core/dev.c +2 -1
@@ -5559,7 +5559,8 @@ static int netif_alloc_netdev_queues(struct net_device *dev)struct netdev_queue *tx;size_t sz = count *sizeof (*tx); -BUG_ON (count <1 || count >0xffff ); +if (count <1 || count >0xffff ) +return -EINVAL ; tx =kzalloc (sz,GFP_KERNEL | __GFP_NOWARN | __GFP_REPEAT);if (!tx) {
Signature · 1 line
2.5.2
PATCH 5/33 rtnetlink: verify IFLA_VF_INFO attributes before passing them to driver
In reply to Jiri SlabyFrom: Daniel Borkmann <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
[ Upstream commit 4f7d2cdfdde71ffe962399b7020c674050329423 ]
Jason Gunthorpe reported that since commit c02db8c6290b ("rtnetlink: make SR-IOV VF interface symmetric"), we don't verify IFLA_VF_INFO attributes anymore with respect to their policy, that is, ifla_vfinfo_policy[].
Before, they were part of ifla_policy[], but they have been nested since placed under IFLA_VFINFO_LIST, that contains the attribute IFLA_VF_INFO, which is another nested attribute for the actual VF attributes such as IFLA_VF_MAC, IFLA_VF_VLAN, etc.
Despite the policy being split out from ifla_policy[] in this commit, it's never applied anywhere. nla_for_each_nested() only does basic nla_ok() testing for struct nlattr, but it doesn't know about the data context and their requirements.
Fix, on top of Jason's initial work, does 1) parsing of the attributes with the right policy, and 2) using the resulting parsed attribute table from 1) instead of the nla_for_each_nested() loop (just like we used to do when still part of ifla_policy[]).
Reference: http://thread.gmane.org/gmane.linux.network/368913
--- net/core/rtnetlink.c | 128 ++++++++++++++++++++++++++------------------------- 1 file changed, 65 insertions(+), 63 deletions(-)
net/core/rtnetlink.c +65 -63
@@ -1197,10 +1197,6 @@ static const struct nla_policy ifla_info_policy[IFLA_INFO_MAX+1] = { [IFLA_INFO_DATA] = { .type = NLA_NESTED }, }; -static const struct nla_policy ifla_vfinfo_policy[IFLA_VF_INFO_MAX +1 ] = { - [IFLA_VF_INFO ] = { .type =NLA_NESTED }, -}; -static const struct nla_policy ifla_vf_policy[IFLA_VF_MAX +1 ] = { [IFLA_VF_MAC ] = { .len =sizeof (struct ifla_vf_mac ) }, [IFLA_VF_VLAN ] = { .len =sizeof (struct ifla_vf_vlan ) },
@@ -1274,67 +1270,66 @@ static int validate_linkmsg(struct net_device *dev, struct nlattr *tb[])return 0 ; } -static int do_setvfinfo (struct net_device *dev,struct nlattr *attr) +static int do_setvfinfo (struct net_device *dev,struct nlattr **tb) { -int rem, err = -EINVAL ; -struct nlattr *vf;const struct net_device_ops *ops = dev->netdev_ops; +int err = -EINVAL ; -nla_for_each_nested (vf, attr, rem) { -switch (nla_type (vf)) { -case IFLA_VF_MAC : { -struct ifla_vf_mac *ivm; - ivm =nla_data (vf); - err = -EOPNOTSUPP ; -if (ops->ndo_set_vf_mac) - err = ops->ndo_set_vf_mac (dev, ivm->vf, - ivm->mac); -break ; - } -case IFLA_VF_VLAN : { -struct ifla_vf_vlan *ivv; - ivv =nla_data (vf); - err = -EOPNOTSUPP ; -if (ops->ndo_set_vf_vlan) - err = ops->ndo_set_vf_vlan (dev, ivv->vf, - ivv->vlan, - ivv->qos); -break ; - } -case IFLA_VF_TX_RATE : { -struct ifla_vf_tx_rate *ivt; - ivt =nla_data (vf); - err = -EOPNOTSUPP ; -if (ops->ndo_set_vf_tx_rate) - err = ops->ndo_set_vf_tx_rate (dev, ivt->vf, - ivt->rate); -break ; - } -case IFLA_VF_SPOOFCHK : { -struct ifla_vf_spoofchk *ivs; - ivs =nla_data (vf); - err = -EOPNOTSUPP ; -if (ops->ndo_set_vf_spoofchk) - err = ops->ndo_set_vf_spoofchk (dev, ivs->vf, - ivs->setting); -break ; - } -case IFLA_VF_LINK_STATE : { -struct ifla_vf_link_state *ivl; - ivl =nla_data (vf); - err = -EOPNOTSUPP ; -if (ops->ndo_set_vf_link_state) - err = ops->ndo_set_vf_link_state (dev, ivl->vf, - ivl->link_state); -break ; - } -default : - err = -EINVAL ; -break ; - } -if (err) -break ; +if (tb[IFLA_VF_MAC ]) { +struct ifla_vf_mac *ivm =nla_data (tb[IFLA_VF_MAC ]); + + err = -EOPNOTSUPP ; +if (ops->ndo_set_vf_mac) + err = ops->ndo_set_vf_mac (dev, ivm->vf, + ivm->mac); +if (err <0 ) +return err; + } + +if (tb[IFLA_VF_VLAN ]) { +struct ifla_vf_vlan *ivv =nla_data (tb[IFLA_VF_VLAN ]); + + err = -EOPNOTSUPP ; +if (ops->ndo_set_vf_vlan) + err = ops->ndo_set_vf_vlan (dev, ivv->vf, ivv->vlan, + ivv->qos); +if (err <0 ) +return err; + } + +if (tb[IFLA_VF_TX_RATE ]) { +struct ifla_vf_tx_rate *ivt =nla_data (tb[IFLA_VF_TX_RATE ]); + + err = -EOPNOTSUPP ; +if (ops->ndo_set_vf_tx_rate) + err = ops->ndo_set_vf_tx_rate (dev, ivt->vf, + ivt->rate); +if (err <0 ) +return err; } + +if (tb[IFLA_VF_SPOOFCHK ]) { +struct ifla_vf_spoofchk *ivs =nla_data (tb[IFLA_VF_SPOOFCHK ]); + + err = -EOPNOTSUPP ; +if (ops->ndo_set_vf_spoofchk) + err = ops->ndo_set_vf_spoofchk (dev, ivs->vf, + ivs->setting); +if (err <0 ) +return err; + } + +if (tb[IFLA_VF_LINK_STATE ]) { +struct ifla_vf_link_state *ivl =nla_data (tb[IFLA_VF_LINK_STATE ]); + + err = -EOPNOTSUPP ; +if (ops->ndo_set_vf_link_state) + err = ops->ndo_set_vf_link_state (dev, ivl->vf, + ivl->link_state); +if (err <0 ) +return err; + } +return err; }
@@ -1517,14 +1512,21 @@ static int do_setlink(const struct sk_buff *skb, }if (tb[IFLA_VFINFO_LIST ]) { +struct nlattr *vfinfo[IFLA_VF_MAX +1 ];struct nlattr *attr;int rem; +nla_for_each_nested (attr, tb[IFLA_VFINFO_LIST ], rem) { -if (nla_type (attr) !=IFLA_VF_INFO ) { +if (nla_type (attr) !=IFLA_VF_INFO || +nla_len (attr) <NLA_HDRLEN ) { err = -EINVAL ; goto errout; } - err =do_setvfinfo (dev, attr); + err =nla_parse_nested (vfinfo,IFLA_VF_MAX , attr, + ifla_vf_policy); +if (err <0 ) + goto errout; + err =do_setvfinfo (dev, vfinfo);if (err <0 ) goto errout; modified =1 ;
Signature · 1 line
2.5.2
PATCH 8/33 bridge: mdb: zero out the local br_ip variable before use
In reply to Jiri SlabyFrom: Nikolay Aleksandrov <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
[ Upstream commit f1158b74e54f2e2462ba5e2f45a118246d9d5b43 ]
Since commit b0e9a30dd669 ("bridge: Add vlan id to multicast groups") there's a check in br_ip_equal() for a matching vlan id, but the mdb functions were not modified to use (or at least zero it) so when an entry was added it would have a garbage vlan id (from the local br_ip variable in __br_mdb_add/del) and this would prevent it from being matched and also deleted. So zero out the whole local ip var to protect ourselves from future changes and also to fix the current bug, since there's no vlan id support in the mdb uapi - use always vlan id 0. Example before patch:
root@debian:~# bridge mdb add dev br0 port eth1 grp 239.0.0.1 permanent root@debian:~# bridge mdb dev br0 port eth1 grp 239.0.0.1 permanent root@debian:~# bridge mdb del dev br0 port eth1 grp 239.0.0.1 permanent RTNETLINK answers: Invalid argument
After patch:
root@debian:~# bridge mdb add dev br0 port eth1 grp 239.0.0.1 permanent root@debian:~# bridge mdb dev br0 port eth1 grp 239.0.0.1 permanent root@debian:~# bridge mdb del dev br0 port eth1 grp 239.0.0.1 permanent root@debian:~# bridge mdb
--- net/bridge/br_mdb.c | 2 ++ 1 file changed, 2 insertions(+)
net/bridge/br_mdb.c +2 -0
@@ -370,6 +370,7 @@ static int __br_mdb_add(struct net *net, struct net_bridge *br,if (!p || p->br != br || p->state ==BR_STATE_DISABLED )return -EINVAL ; +memset (&ip,0 ,sizeof (ip)); ip.proto = entry->addr.proto;if (ip.proto ==htons (ETH_P_IP )) ip.u.ip4 = entry->addr.u.ip4;
@@ -416,6 +417,7 @@ static int __br_mdb_del(struct net_bridge *br, struct br_mdb_entry *entry)if (!netif_running (br->dev) || br->multicast_disabled)return -EINVAL ; +memset (&ip,0 ,sizeof (ip)); ip.proto = entry->addr.proto;if (ip.proto ==htons (ETH_P_IP )) {if (timer_pending (&br->ip4_querier.timer))
Signature · 1 line
2.5.2
PATCH 10/33 net: do not process device backlog during unregistration
In reply to Jiri SlabyFrom: Julian Anastasov <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
[ Upstream commit e9e4dd3267d0c5234c5c0f47440456b10875dec9 ]
commit 381c759d9916 ("ipv4: Avoid crashing in ip_error") fixes a problem where processed packet comes from device with destroyed inetdev (dev->ip_ptr). This is not expected because inetdev_destroy is called in NETDEV_UNREGISTER phase and packets should not be processed after dev_close_many() and synchronize_net(). Above fix is still required because inetdev_destroy can be called for other reasons. But it shows the real problem: backlog can keep packets for long time and they do not hold reference to device. Such packets are then delivered to upper levels at the same time when device is unregistered.
Calling flush_backlog after NETDEV_UNREGISTER_FINAL still accounts all packets from backlog but before that some packets continue to be delivered to upper levels long after the synchronize_net call which is supposed to wait the last ones. Also, as Eric pointed out, processed packets, mostly from other devices, can continue to add new packets to backlog.
Fix the problem by moving flush_backlog early, after the device driver is stopped and before the synchronize_net() call. Then use netif_running check to make sure we do not add more packets to backlog. We have to do it in enqueue_to_backlog context when the local IRQ is disabled. As result, after the flush_backlog and synchronize_net sequence all packets should be accounted.
Thanks to Eric W. Biederman for the test script and his valuable feedback!
--- net/core/dev.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-)
net/core/dev.c +4 -2
@@ -3193,6 +3193,8 @@ static int enqueue_to_backlog(struct sk_buff *skb, int cpu,local_irq_save (flags);rps_lock (sd); +if (!netif_running (skb->dev)) + goto drop; qlen =skb_queue_len (&sd->input_pkt_queue);if (qlen <= netdev_max_backlog && !skb_flow_limit (skb, qlen)) {if (skb_queue_len (&sd->input_pkt_queue)) {
@@ -3214,6 +3216,7 @@ enqueue: goto enqueue; } +drop: sd->dropped++;rps_unlock (sd);
@@ -5302,6 +5305,7 @@ static void rollback_registered_many(struct list_head *head)unlist_netdevice (dev); dev->reg_state =NETREG_UNREGISTERING ; +on_each_cpu (flush_backlog, dev,1 ); }synchronize_net ();
@@ -5918,8 +5922,6 @@ void netdev_run_todo(void) dev->reg_state =NETREG_UNREGISTERED ; -on_each_cpu (flush_backlog, dev,1 ); -netdev_wait_allrefs (dev);/* paranoia */
Signature · 1 line
2.5.2
PATCH 12/33 net: Clone skb before setting peeked flag
In reply to Jiri SlabyFrom: Herbert Xu <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
[ Upstream commit 738ac1ebb96d02e0d23bc320302a6ea94c612dec ]
Shared skbs must not be modified and this is crucial for broadcast and/or multicast paths where we use it as an optimisation to avoid unnecessary cloning.
The function skb_recv_datagram breaks this rule by setting peeked without cloning the skb first. This causes funky races which leads to double-free.
This patch fixes this by cloning the skb and replacing the skb in the list when setting skb->peeked.
--- net/core/datagram.c | 41 ++++++++++++++++++++++++++++++++++++++--- 1 file changed, 38 insertions(+), 3 deletions(-)
net/core/datagram.c +38 -3
@@ -130,6 +130,35 @@ out_noerr: goto out; } +static int skb_set_peeked (struct sk_buff *skb) +{ +struct sk_buff *nskb; + +if (skb->peeked) +return 0 ; + +/* We have to unshare an skb before modifying it. */ +if (!skb_shared (skb)) + goto done; + + nskb =skb_clone (skb,GFP_ATOMIC ); +if (!nskb) +return -ENOMEM ; + + skb->prev->next = nskb; + skb->next->prev = nskb; + nskb->prev = skb->prev; + nskb->next = skb->next; + +consume_skb (skb); + skb = nskb; + +done: + skb->peeked =1 ; + +return 0 ; +} +/** * __skb_recv_datagram - Receive a datagram skbuff * @sk: socket
@@ -164,7 +193,9 @@ out_noerr:struct sk_buff *__skb_recv_datagram (struct sock *sk,unsigned int flags,int *peeked,int *off,int *err) { +struct sk_buff_head *queue = &sk->sk_receive_queue;struct sk_buff *skb, *last; +unsigned long cpu_flags;long timeo;/* * Caller is allowed not to check sk->sk_err before skb_recv_datagram()
@@ -183,8 +214,6 @@ struct sk_buff *__skb_recv_datagram(struct sock *sk, unsigned int flags,* Look at current nfs client by the way... * However, this function was correct in any case. 8) */ -unsigned long cpu_flags; -struct sk_buff_head *queue = &sk->sk_receive_queue;int _off = *off; last = (struct sk_buff *)queue;
@@ -198,7 +227,11 @@ struct sk_buff *__skb_recv_datagram(struct sock *sk, unsigned int flags, _off -= skb->len;continue ; } - skb->peeked =1 ; + + error =skb_set_peeked (skb); +if (error) + goto unlock_err; +atomic_inc (&skb->users); }else __skb_unlink (skb, queue);
@@ -222,6 +255,8 @@ struct sk_buff *__skb_recv_datagram(struct sock *sk, unsigned int flags,return NULL ; +unlock_err: +spin_unlock_irqrestore (&queue->lock, cpu_flags); no_packet: *err = error;return NULL ;
Signature · 1 line
2.5.2
PATCH 14/33 net: Fix skb_set_peeked use-after-free bug
In reply to Jiri SlabyFrom: Herbert Xu <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
[ Upstream commit a0a2a6602496a45ae838a96db8b8173794b5d398 ]
The commit 738ac1ebb96d02e0d23bc320302a6ea94c612dec ("net: Clone skb before setting peeked flag") introduced a use-after-free bug in skb_recv_datagram. This is because skb_set_peeked may create a new skb and free the existing one. As it stands the caller will continue to use the old freed skb.
This patch fixes it by making skb_set_peeked return the new skb (or the old one if unchanged).
--- net/core/datagram.c | 13 +++++++------ 1 file changed, 7 insertions(+), 6 deletions(-)
net/core/datagram.c +7 -6
@@ -130,12 +130,12 @@ out_noerr: goto out; } -static int skb_set_peeked (struct sk_buff *skb) +static struct sk_buff *skb_set_peeked (struct sk_buff *skb) {struct sk_buff *nskb;if (skb->peeked) -return 0 ; +return skb;/* We have to unshare an skb before modifying it. */ if (!skb_shared (skb))
@@ -143,7 +143,7 @@ static int skb_set_peeked(struct sk_buff *skb) nskb =skb_clone (skb,GFP_ATOMIC );if (!nskb) -return -ENOMEM ; +return ERR_PTR (-ENOMEM ); skb->prev->next = nskb; skb->next->prev = nskb;
@@ -156,7 +156,7 @@ static int skb_set_peeked(struct sk_buff *skb) done: skb->peeked =1 ; -return 0 ; +return skb; }/**
@@ -228,8 +228,9 @@ struct sk_buff *__skb_recv_datagram(struct sock *sk, unsigned int flags,continue ; } - error =skb_set_peeked (skb); -if (error) + skb =skb_set_peeked (skb); + error =PTR_ERR (skb); +if (IS_ERR (skb)) goto unlock_err;atomic_inc (&skb->users);
Signature · 1 line
2.5.2
PATCH 16/33 isdn/gigaset: reset tty->receive_room when attaching ser_gigaset
In reply to Jiri SlabyFrom: Tilman Schmidt <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
[ Upstream commit fd98e9419d8d622a4de91f76b306af6aa627aa9c ]
Commit 79901317ce80 ("n_tty: Don't flush buffer when closing ldisc"), first merged in kernel release 3.10, caused the following regression in the Gigaset M101 driver:
Before that commit, when closing the N_TTY line discipline in preparation to switching to N_GIGASET_M101, receive_room would be reset to a non-zero value by the call to n_tty_flush_buffer() in n_tty's close method. With the removal of that call, receive_room might be left at zero, blocking data reception on the serial line.
The present patch fixes that regression by setting receive_room to an appropriate value in the ldisc open method.
--- drivers/isdn/gigaset/ser-gigaset.c | 11 ++++++++++- 1 file changed, 10 insertions(+), 1 deletion(-)
drivers/isdn/gigaset/ser-gigaset.c +10 -1
@@ -524,9 +524,18 @@ gigaset_tty_open(struct tty_struct *tty) cs->hw.ser->tty = tty;atomic_set (&cs->hw.ser->refcnt,1 );init_completion (&cs->hw.ser->dead_cmp); - tty->disc_data = cs; +/* Set the amount of data we're willing to receive per call +* from the hardware driver to half of the input buffer size +* to leave some reserve. +* Note: We don't do flow control towards the hardware driver. +* If more data is received than will fit into the input buffer, +* it will be dropped and an error will be logged. This should +* never happen as the device is slow and the buffer size ample. +*/ + tty->receive_room =RBUFSIZE /2 ; +/* OK.. Initialization of the datastructures and the HW is done.. Now * startup system and notify the LL that we are ready to run */
Signature · 1 line
2.5.2
PATCH 17/33 ipv6: lock socket in ip6_datagram_connect()
In reply to Jiri SlabyFrom: Eric Dumazet <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
[ Upstream commit 03645a11a570d52e70631838cb786eb4253eb463 ]
ip6_datagram_connect() is doing a lot of socket changes without socket being locked.
This looks wrong, at least for udp_lib_rehash() which could corrupt lists because of concurrent udp_sk(sk)->udp_portaddr_hash accesses.
--- include/net/ip.h | 1 + net/ipv4/datagram.c | 16 ++++++++++++---- net/ipv6/datagram.c | 20 +++++++++++++++----- 3 files changed, 28 insertions(+), 9 deletions(-)
include/net/ip.h +1 -0
@@ -141,6 +141,7 @@ static inline struct sk_buff *ip_finish_skb(struct sock *sk, struct flowi4 *fl4) }/* datagram.c */ +int __ip4_datagram_connect (struct sock *sk,struct sockaddr *uaddr,int addr_len);extern int ip4_datagram_connect (struct sock *sk,struct sockaddr *uaddr,int addr_len);
net/ipv4/datagram.c +12 -4
@@ -20,7 +20,7 @@#include <net/route.h> #include <net/tcp_states.h> -int ip4_datagram_connect (struct sock *sk,struct sockaddr *uaddr,int addr_len) +int __ip4_datagram_connect (struct sock *sk,struct sockaddr *uaddr,int addr_len) {struct inet_sock *inet =inet_sk (sk);struct sockaddr_in *usin = (struct sockaddr_in *) uaddr;
@@ -39,8 +39,6 @@ int ip4_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len)sk_dst_reset (sk); -lock_sock (sk); - oif = sk->sk_bound_dev_if; saddr = inet->inet_saddr;if (ipv4_is_multicast (usin->sin_addr.s_addr)) {
@@ -81,9 +79,19 @@ int ip4_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len)sk_dst_set (sk, &rt->dst); err =0 ; out: -release_sock (sk);return err; } +EXPORT_SYMBOL (__ip4_datagram_connect); + +int ip4_datagram_connect (struct sock *sk,struct sockaddr *uaddr,int addr_len) +{ +int res; + +lock_sock (sk); + res =__ip4_datagram_connect (sk, uaddr, addr_len); +release_sock (sk); +return res; +}EXPORT_SYMBOL (ip4_datagram_connect);/* Because UDP xmit path can manipulate sk_dst_cache without holding
net/ipv6/datagram.c +15 -5
@@ -40,7 +40,7 @@ static bool ipv6_mapped_addr_any(const struct in6_addr *a)return ipv6_addr_v4mapped (a) && (a->s6_addr32[3 ] ==0 ); } -int ip6_datagram_connect (struct sock *sk,struct sockaddr *uaddr,int addr_len) +static int __ip6_datagram_connect (struct sock *sk,struct sockaddr *uaddr,int addr_len) {struct sockaddr_in6 *usin = (struct sockaddr_in6 *) uaddr;struct inet_sock *inet =inet_sk (sk);
@@ -56,7 +56,7 @@ int ip6_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len)if (usin->sin6_family ==AF_INET ) {if (__ipv6_only_sock (sk))return -EAFNOSUPPORT ; - err =ip4_datagram_connect (sk, uaddr, addr_len); + err =__ip4_datagram_connect (sk, uaddr, addr_len); goto ipv4_connected; }
@@ -99,9 +99,9 @@ int ip6_datagram_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len) sin.sin_addr.s_addr = daddr->s6_addr32[3 ]; sin.sin_port = usin->sin6_port; - err =ip4_datagram_connect (sk, - (struct sockaddr *) &sin, -sizeof (sin)); + err =__ip4_datagram_connect (sk, + (struct sockaddr *) &sin, +sizeof (sin)); ipv4_connected:if (err)
@@ -204,6 +204,16 @@ out:fl6_sock_release (flowlabel);return err; } + +int ip6_datagram_connect (struct sock *sk,struct sockaddr *uaddr,int addr_len) +{ +int res; + +lock_sock (sk); + res =__ip6_datagram_connect (sk, uaddr, addr_len); +release_sock (sk); +return res; +}EXPORT_SYMBOL_GPL (ip6_datagram_connect); void ipv6_icmp_error(struct sock *sk, struct sk_buff *skb, int err,
Signature · 1 line
2.5.2
PATCH 18/33 bonding: fix destruction of bond with devices different from arphrd_ether
In reply to Jiri SlabyFrom: Nikolay Aleksandrov <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
[ Upstream commit 06f6d1094aa0992432b1e2a0920b0ee86ccd83bf ]
When the bonding is being unloaded and the netdevice notifier is unregistered it executes NETDEV_UNREGISTER for each device which should remove the bond's proc entry but if the device enslaved is not of ARPHRD_ETHER type and is in front of the bonding, it may execute bond_release_and_destroy() first which would release the last slave and destroy the bond device leaving the proc entry and thus we will get the following error (with dynamic debug on for bond_netdev_event to see the events order):
[ 908.963051] eql: event: 9 [ 908.963052] eql: IFF_SLAVE [ 908.963054] eql: event: 2 [ 908.963056] eql: IFF_SLAVE [ 908.963058] eql: event: 6 [ 908.963059] eql: IFF_SLAVE [ 908.963110] bond0: Releasing active interface eql [ 908.976168] bond0: Destroying bond bond0 [ 908.976266] bond0 (unregistering): Released all slaves [ 908.984097] ------------[ cut here ]------------ [ 908.984107] WARNING: CPU: 0 PID: 1787 at fs/proc/generic.c:575 remove_proc_entry+0x112/0x160() [ 908.984110] remove_proc_entry: removing non-empty directory 'net/bonding', leaking at least 'bond0' [ 908.984111] Modules linked in: bonding(-) eql(O) 9p nfsd auth_rpcgss
oid_registry nfs_acl nfs lockd grace fscache sunrpc crct10dif_pclmul crc32_pclmul crc32c_intel ghash_clmulni_intel ppdev qxl drm_kms_helper snd_hda_codec_generic aesni_intel ttm aes_x86_64 glue_helper pcspkr lrw gf128mul ablk_helper cryptd snd_hda_intel virtio_console snd_hda_codec psmouse serio_raw snd_hwdep snd_hda_core 9pnet_virtio 9pnet evdev joydev drm virtio_balloon snd_pcm snd_timer snd soundcore i2c_piix4 i2c_core pvpanic acpi_cpufreq parport_pc parport processor thermal_sys button autofs4 ext4 crc16 mbcache jbd2 hid_generic usbhid hid sg sr_mod cdrom ata_generic virtio_blk virtio_net floppy ata_piix e1000 libata ehci_pci virtio_pci scsi_mod uhci_hcd ehci_hcd virtio_ring virtio usbcore usb_common [last unloaded: bonding]
[ 908.984168] CPU: 0 PID: 1787 Comm: rmmod Tainted: G W O 4.2.0-rc2+ #8 [ 908.984170] Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011 [ 908.984172] 0000000000000000 ffffffff81732d41 ffffffff81525b34 ffff8800358dfda8 [ 908.984175] ffffffff8106c521 ffff88003595af78 ffff88003595af40 ffff88003e3a4280 [ 908.984178] ffffffffa058d040 0000000000000000 ffffffff8106c59a ffffffff8172ebd0 [ 908.984181] Call Trace: [ 908.984188] [<ffffffff81525b34>] ? dump_stack+0x40/0x50 [ 908.984193] [<ffffffff8106c521>] ? warn_slowpath_common+0x81/0xb0 [ 908.984196] [<ffffffff8106c59a>] ? warn_slowpath_fmt+0x4a/0x50 [ 908.984199] [<ffffffff81218352>] ? remove_proc_entry+0x112/0x160 [ 908.984205] [<ffffffffa05850e6>] ? bond_destroy_proc_dir+0x26/0x30 [bonding] [ 908.984208] [<ffffffffa057540e>] ? bond_net_exit+0x8e/0xa0 [bonding] [ 908.984217] [<ffffffff8142f407>] ? ops_exit_list.isra.4+0x37/0x70 [ 908.984225] [<ffffffff8142f52d>] ? unregister_pernet_operations+0x8d/0xd0 [ 908.984228] [<ffffffff8142f58d>] ? unregister_pernet_subsys+0x1d/0x30 [ 908.984232] [<ffffffffa0585269>] ? bonding_exit+0x23/0xdba [bonding] [ 908.984236] [<ffffffff810e28ba>] ? SyS_delete_module+0x18a/0x250 [ 908.984241] [<ffffffff81086f99>] ? task_work_run+0x89/0xc0 [ 908.984244] [<ffffffff8152b732>] ? entry_SYSCALL_64_fastpath+0x16/0x75 [ 908.984247] ---[ end trace 7c006ed4abbef24b ]---
Thus remove the proc entry manually if bond_release_and_destroy() is used. Because of the checks in bond_remove_proc_entry() it's not a problem for a bond device to change namespaces (the bug fixed by the Fixes commit) but since commitf9399814927ad ("bonding: Don't allow bond devices to change network namespaces.") that can't happen anyway.
the netdev events")
--- drivers/net/bonding/bond_main.c | 1 + 1 file changed, 1 insertion(+)
drivers/net/bonding/bond_main.c +1 -0
@@ -1917,6 +1917,7 @@ static int bond_release_and_destroy(struct net_device *bond_dev, bond_dev->priv_flags |=IFF_DISABLE_NETPOLL ;pr_info ("%s: destroying bond %s.\n" , bond_dev->name, bond_dev->name); +bond_remove_proc_entry (bond);unregister_netdevice (bond_dev); }return ret;
Signature · 1 line
2.5.2
PATCH 19/33 bonding: correct the MAC address for "follow" fail_over_mac policy
In reply to Jiri SlabyFrom: dingtianhong <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
[ Upstream commit a951bc1e6ba58f11df5ed5ddc41311e10f5fd20b ]
The "follow" fail_over_mac policy is useful for multiport devices that either become confused or incur a performance penalty when multiple ports are programmed with the same MAC address, but the same MAC address still may happened by this steps for this policy:
1) echo +eth0 > /sys/class/net/bond0/bonding/slaves bond0 has the same mac address with eth0, it is MAC1.
2) echo +eth1 > /sys/class/net/bond0/bonding/slaves eth1 is backup, eth1 has MAC2.
3) ifconfig eth0 down
eth1 became active slave, bond will swap MAC for eth0 and eth1, so eth1 has MAC1, and eth0 has MAC2.
4) ifconfig eth1 down
there is no active slave, and eth1 still has MAC1, eth2 has MAC2.
5) ifconfig eth0 up
the eth0 became active slave again, the bond set eth0 to MAC1.
Something wrong here, then if you set eth1 up, the eth0 and eth1 will have the same MAC address, it will break this policy for ACTIVE_BACKUP mode.
This patch will fix this problem by finding the old active slave and swap them MAC address before change active slave.
--- drivers/net/bonding/bond_main.c | 19 +++++++++++++++++++ 1 file changed, 19 insertions(+)
drivers/net/bonding/bond_main.c +19 -0
@@ -671,6 +671,22 @@ static void bond_set_dev_addr(struct net_device *bond_dev,call_netdevice_notifiers (NETDEV_CHANGEADDR , bond_dev); } +static struct slave *bond_get_old_active (struct bonding *bond, +struct slave *new_active) +{ +struct slave *slave; + +bond_for_each_slave (bond, slave) { +if (slave == new_active) +continue ; + +if (ether_addr_equal (bond->dev->dev_addr, slave->dev->dev_addr)) +return slave; + } + +return NULL ; +} +/* * bond_do_fail_over_mac *
@@ -712,6 +728,9 @@ static void bond_do_fail_over_mac(struct bonding *bond,write_unlock_bh (&bond->curr_slave_lock);read_unlock (&bond->lock); +if (!old_active) + old_active =bond_get_old_active (bond, new_active); +if (old_active) {memcpy (tmp_mac, new_active->dev->dev_addr,ETH_ALEN ); memcpy(saddr.sa_data, old_active->dev->dev_addr,
Signature · 1 line
2.5.2
PATCH 20/33 inet: frags: fix defragmented packet's IP header for af_packet
In reply to Jiri SlabyFrom: Edward Hyunkoo Jee <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
[ Upstream commit 0848f6428ba3a2e42db124d41ac6f548655735bf ]
When ip_frag_queue() computes positions, it assumes that the passed sk_buff does not contain L2 headers.
However, when PACKET_FANOUT_FLAG_DEFRAG is used, IP reassembly functions can be called on outgoing packets that contain L2 headers.
Also, IPv4 checksum is not corrected after reassembly.
--- net/ipv4/ip_fragment.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-)
net/ipv4/ip_fragment.c +5 -2
@@ -356,7 +356,7 @@ static int ip_frag_queue(struct ipq *qp, struct sk_buff *skb) ihl =ip_hdrlen (skb);/* Determine the position of this fragment. */ - end = offset + skb->len - ihl; + end = offset + skb->len -skb_network_offset (skb) - ihl; err = -EINVAL ;/* Is this the final fragment? */
@@ -386,7 +386,7 @@ static int ip_frag_queue(struct ipq *qp, struct sk_buff *skb) goto err; err = -ENOMEM ; -if (pskb_pull (skb, ihl) ==NULL ) +if (!pskb_pull (skb,skb_network_offset (skb) + ihl)) goto err; err =pskb_trim_rcsum (skb, end - offset);
@@ -627,6 +627,9 @@ static int ip_frag_reasm(struct ipq *qp, struct sk_buff *prev, iph->frag_off = qp->q.max_size ?htons (IP_DF ) :0 ; iph->tot_len =htons (len); iph->tos |= ecn; + +ip_send_check (iph); +IP_INC_STATS_BH (net,IPSTATS_MIB_REASMOKS ); qp->q.fragments =NULL ; qp->q.fragments_tail =NULL ;
Signature · 1 line
2.5.2
PATCH 23/33 rds: fix an integer overflow test in rds_info_getsockopt()
In reply to Jiri SlabyFrom: Dan Carpenter <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
[ Upstream commit 468b732b6f76b138c0926eadf38ac88467dcd271 ]
"len" is a signed integer. We check that len is not negative, so it goes from zero to INT_MAX. PAGE_SIZE is unsigned long so the comparison is type promoted to unsigned long. ULONG_MAX - 4095 is a higher than INT_MAX so the condition can never be true.
I don't know if this is harmful but it seems safe to limit "len" to INT_MAX - 4095.
--- net/rds/info.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
net/rds/info.c +1 -1
@@ -176,7 +176,7 @@ int rds_info_getsockopt(struct socket *sock, int optname, char __user *optval,/* check for all kinds of wrapping and the like */ start = (unsigned long )optval; -if (len <0 || len +PAGE_SIZE -1 < len || start + len < start) { +if (len <0 || len >INT_MAX -PAGE_SIZE +1 || start + len < start) { ret = -EINVAL ; goto out; }
Signature · 1 line
2.5.2
PATCH 26/33 lpfc: Fix scsi prep dma buf error.
In reply to Jiri SlabyFrom: James Smart <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
commit 5116fbf136ea21b8678a85eee5c03508736ada9f upstream.
Didn't check for less-than-or-equal zero. Means we may later call scsi_dma_unmap() even though we don't have valid mappings.
--- drivers/scsi/lpfc/lpfc_scsi.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/scsi/lpfc/lpfc_scsi.c +1 -1
@@ -3423,7 +3423,7 @@ lpfc_scsi_prep_dma_buf_s4(struct lpfc_hba *phba, struct lpfc_scsi_buf *lpfc_cmd)*/ nseg =scsi_dma_map (scsi_cmnd); -if (unlikely (!nseg)) +if (unlikely (nseg <=0 ))return 1 ; sgl +=1 ;/* clear the last flag in the fcp_rsp map entry */
Signature · 1 line
2.5.2
PATCH 27/33 bio: fix argument of __bio_add_page() for max_sectors > 0xffff
In reply to Jiri SlabyFrom: Akinobu Mita <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
commit 34f2fd8dfe6185b0eaaf7d661281713a6170b077 upstream.
The data type of max_sectors and max_hw_sectors in queue settings are unsigned int. But these values are passed to __bio_add_page() as an argument whose data type is unsigned short. In the worst case such as max_sectors is 0x10000, bio_add_page() can't add a page and IOs can't proceed.
--- fs/bio.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
fs/bio.c +1 -1
@@ -601,7 +601,7 @@ EXPORT_SYMBOL(bio_get_nr_vecs);static int __bio_add_page (struct request_queue *q,struct bio *bio,struct page *page,unsigned int len,unsigned int offset, -unsigned short max_sectors) +unsigned int max_sectors) {int retried_segments =0 ;struct bio_vec *bvec;
Signature · 1 line
2.5.2
PATCH 28/33 dm cache mq: fix memory allocation failure for large cache devices
In reply to Jiri SlabyFrom: Heinz Mauelshagen <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
commit 14f398ca2f26a2ed6236aec54395e0fa06ec8a82 upstream.
The memory allocated for the multiqueue policy's hash table doesn't need to be physically contiguous. Use vzalloc() instead of kzalloc(). Fedora has been carrying this fix since 10/10/2013.
Failure seen during creation of a 10TB cached device with a 2048 sector block size and 411GB cache size:
dmsetup: page allocation failure: order:9, mode:0x10c0d0
CPU: 11 PID: 29235 Comm: dmsetup Not tainted 3.10.4 #3 Hardware name: Supermicro X8DTL/X8DTL, BIOS 2.1a 12/30/2011 000000000010c0d0 ffff880090941898 ffffffff81387ab4 ffff880090941928 ffffffff810bb26f 0000000000000009 000000000010c0d0 ffff880090941928 ffffffff81385dbc ffffffff815f3840 ffffffff00000000 000002000010c0d0 Call Trace: [<ffffffff81387ab4>] dump_stack+0x19/0x1b [<ffffffff810bb26f>] warn_alloc_failed+0x110/0x124 [<ffffffff81385dbc>] ? __alloc_pages_direct_compact+0x17c/0x18e [<ffffffff810bda2e>] __alloc_pages_nodemask+0x6c7/0x75e [<ffffffff810bdad7>] __get_free_pages+0x12/0x3f [<ffffffff810ea148>] kmalloc_order_trace+0x29/0x88 [<ffffffff810ec1fd>] __kmalloc+0x36/0x11b [<ffffffffa031eeed>] ? mq_create+0x1dc/0x2cf [dm_cache_mq] [<ffffffffa031efc0>] mq_create+0x2af/0x2cf [dm_cache_mq] [<ffffffffa0314605>] dm_cache_policy_create+0xa7/0xd2 [dm_cache] [<ffffffffa0312530>] ? cache_ctr+0x245/0xa13 [dm_cache] [<ffffffffa031263e>] cache_ctr+0x353/0xa13 [dm_cache] [<ffffffffa012b916>] dm_table_add_target+0x227/0x2ce [dm_mod] [<ffffffffa012e8e4>] table_load+0x286/0x2ac [dm_mod] [<ffffffffa012e65e>] ? dev_wait+0x8a/0x8a [dm_mod] [<ffffffffa012e324>] ctl_ioctl+0x39a/0x3c2 [dm_mod] [<ffffffffa012e35a>] dm_ctl_ioctl+0xe/0x12 [dm_mod] [<ffffffff81101181>] vfs_ioctl+0x21/0x34 [<ffffffff811019d3>] do_vfs_ioctl+0x3b1/0x3f4 [<ffffffff810f4d2e>] ? ____fput+0x9/0xb [<ffffffff81050b6c>] ? task_work_run+0x7e/0x92 [<ffffffff81101a68>] SyS_ioctl+0x52/0x82 [<ffffffff81391d92>] system_call_fastpath+0x16/0x1b
--- drivers/md/dm-cache-policy-mq.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/md/dm-cache-policy-mq.c +2 -2
@@ -855,7 +855,7 @@ static void mq_destroy(struct dm_cache_policy *p)struct mq_policy *mq =to_mq_policy (p);free_bitset (mq->allocation_bitset); -kfree (mq->table); +vfree (mq->table);free_entries (mq);kfree (mq); }
@@ -1106,7 +1106,7 @@ static struct dm_cache_policy *mq_create(dm_cblock_t cache_size, mq->nr_buckets =next_power (from_cblock (cache_size) /2 ,16 ); mq->hash_bits =ffs (mq->nr_buckets) -1 ; - mq->table =kzalloc (sizeof (*mq->table) * mq->nr_buckets,GFP_KERNEL ); + mq->table =vzalloc (sizeof (*mq->table) * mq->nr_buckets);if (!mq->table) goto bad_alloc_table;
Signature · 1 line
2.5.2
PATCH 29/33 aio: fix reqs_available handling
In reply to Jiri SlabyFrom: Benjamin LaHaise <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
commit d856f32a86b2b015ab180ab7a55e455ed8d3ccc5 upstream.
As reported by Dan Aloni, commit f8567a3845ac ("aio: fix aio request leak when events are reaped by userspace") introduces a regression when user code attempts to perform io_submit() with more events than are available in the ring buffer. Reverting that commit would reintroduce a regression when user space event reaping is used.
Fixing this bug is a bit more involved than the previous attempts to fix this regression. Since we do not have a single point at which we can count events as being reaped by user space and io_getevents(), we have to track event completion by looking at the number of events left in the event ring. So long as there are as many events in the ring buffer as there have been completion events generate, we cannot call put_reqs_available(). The code to check for this is now placed in refill_reqs_available().
A test program from Dan and modified by me for verifying this bug is available at http://www.kvack.org/~bcrl/20140824-aio_bug.c .
--- fs/aio.c | 77 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++---- 1 file changed, 73 insertions(+), 4 deletions(-)
fs/aio.c +73 -4
@@ -146,6 +146,7 @@ struct kioctx {struct {unsigned tail; +unsigned completed_events;spinlock_t completion_lock; }____cacheline_aligned_in_smp ;
@@ -899,6 +900,68 @@ out:return ret; } +/* refill_reqs_available +* Updates the reqs_available reference counts used for tracking the +* number of free slots in the completion ring. This can be called +* from aio_complete() (to optimistically update reqs_available) or +* from aio_get_req() (the we're out of events case). It must be +* called holding ctx->completion_lock. +*/ +static void refill_reqs_available (struct kioctx *ctx,unsigned head, +unsigned tail) +{ +unsigned events_in_ring, completed; + +/* Clamp head since userland can write to it. */ + head %= ctx->nr_events; +if (head <= tail) + events_in_ring = tail - head; +else + events_in_ring = ctx->nr_events - (head - tail); + + completed = ctx->completed_events; +if (events_in_ring < completed) + completed -= events_in_ring; +else + completed =0 ; + +if (!completed) +return ; + + ctx->completed_events -= completed; +put_reqs_available (ctx, completed); +} + +/* user_refill_reqs_available +* Called to refill reqs_available when aio_get_req() encounters an +* out of space in the completion ring. +*/ +static void user_refill_reqs_available (struct kioctx *ctx) +{ +spin_lock_irq (&ctx->completion_lock); +if (ctx->completed_events) { +struct aio_ring *ring; +unsigned head; + +/* Access of ring->head may race with aio_read_events_ring() +* here, but that's okay since whether we read the old version +* or the new version, and either will be valid. The important +* part is that head cannot pass tail since we prevent +* aio_complete() from updating tail by holding +* ctx->completion_lock. Even if head is invalid, the check +* against ctx->completed_events below will make sure we do the +* safe/right thing. +*/ + ring =kmap_atomic (ctx->ring_pages[0 ]); + head = ring->head; +kunmap_atomic (ring); + +refill_reqs_available (ctx, head, ctx->tail); + } + +spin_unlock_irq (&ctx->completion_lock); +} +/* aio_get_req * Allocate a slot for an aio request. * Returns NULL if no requests are free.
@@ -907,8 +970,11 @@ static inline struct kiocb *aio_get_req(struct kioctx *ctx) {struct kiocb *req; -if (!get_reqs_available (ctx)) -return NULL ; +if (!get_reqs_available (ctx)) { +user_refill_reqs_available (ctx); +if (!get_reqs_available (ctx)) +return NULL ; + } req =kmem_cache_alloc (kiocb_cachep,GFP_KERNEL |__GFP_ZERO);if (unlikely (!req))
@@ -967,8 +1033,8 @@ void aio_complete(struct kiocb *iocb, long res, long res2)struct kioctx *ctx = iocb->ki_ctx;struct aio_ring *ring;struct io_event *ev_page, *event; +unsigned tail, pos, head;unsigned long flags; -unsigned tail, pos;/* * Special case handling for sync iocbs:
@@ -1029,10 +1095,14 @@ void aio_complete(struct kiocb *iocb, long res, long res2) ctx->tail = tail; ring =kmap_atomic (ctx->ring_pages[0 ]); + head = ring->head; ring->tail = tail;kunmap_atomic (ring);flush_dcache_page (ctx->ring_pages[0 ]); + ctx->completed_events++; +if (ctx->completed_events >1 ) +refill_reqs_available (ctx, head, tail);spin_unlock_irqrestore (&ctx->completion_lock, flags);pr_debug ("added to ring %p at [%u]\n" , iocb, tail);
@@ -1047,7 +1117,6 @@ void aio_complete(struct kiocb *iocb, long res, long res2)/* everything turned out well, dispose of the aiocb. */ kiocb_free (iocb); -put_reqs_available (ctx,1 );/* * We have to order our ring_info tail store above and test
Signature · 1 line
2.5.2
PATCH 30/33 netfilter: nf_conntrack: fix RCU race in nf_conntrack_find_get
In reply to Jiri SlabyFrom: Andrey Vagin <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
[ Upstream commit c6825c0976fa7893692e0e43b09740b419b23c09 ]
Lets look at destroy_conntrack:
hlist_nulls_del_rcu(&ct->tuplehash[IP_CT_DIR_ORIGINAL].hnnode);
...nf_conntrack_free(ct)
kmem_cache_free(net->ct.nf_conntrack_cachep, ct);
net->ct.nf_conntrack_cachep is created with SLAB_DESTROY_BY_RCU.
The hash is protected by rcu, so readers look up conntracks without locks.
A conntrack is removed from the hash, but in this moment a few readers still can use the conntrack. Then this conntrack is released and another thread creates conntrack with the same address and the equal tuple. After this a reader starts to validate the conntrack:
* It's not dying, because a new conntrack was created
* nf_ct_tuple_equal() returns true.
But this conntrack is not initialized yet, so it can not be used by two threads concurrently. In this case BUG_ON may be triggered from nf_nat_setup_info().
Florian Westphal suggested to check the confirm bit too. I think it's right.
task 1 task 2 task 3
nf_conntrack_find_get ____nf_conntrack_find
destroy_conntrack
hlist_nulls_del_rcu
nf_conntrack_free
kmem_cache_free
__nf_conntrack_alloc kmem_cache_alloc memset(&ct->tuplehash[IP_CT_DIR_MAX], if (nf_ct_is_dying(ct)) if (!nf_ct_tuple_equal()
I'm not sure, that I have ever seen this race condition in a real life. Currently we are investigating a bug, which is reproduced on a few nodes. In our case one conntrack is initialized from a few tasks concurrently, we don't have any other explanation for this.
<2>[46267.083061] kernel BUG at net/ipv4/netfilter/nf_nat_core.c:322! ... <4>[46267.083951] RIP: 0010:[<ffffffffa01e00a4>] [<ffffffffa01e00a4>] nf_nat_setup_info+0x564/0x590 [nf_nat] ... <4>[46267.085549] Call Trace: <4>[46267.085622] [<ffffffffa023421b>] alloc_null_binding+0x5b/0xa0 [iptable_nat] <4>[46267.085697] [<ffffffffa02342bc>] nf_nat_rule_find+0x5c/0x80 [iptable_nat] <4>[46267.085770] [<ffffffffa0234521>] nf_nat_fn+0x111/0x260 [iptable_nat] <4>[46267.085843] [<ffffffffa0234798>] nf_nat_out+0x48/0xd0 [iptable_nat] <4>[46267.085919] [<ffffffff814841b9>] nf_iterate+0x69/0xb0 <4>[46267.085991] [<ffffffff81494e70>] ? ip_finish_output+0x0/0x2f0 <4>[46267.086063] [<ffffffff81484374>] nf_hook_slow+0x74/0x110 <4>[46267.086133] [<ffffffff81494e70>] ? ip_finish_output+0x0/0x2f0 <4>[46267.086207] [<ffffffff814b5890>] ? dst_output+0x0/0x20 <4>[46267.086277] [<ffffffff81495204>] ip_output+0xa4/0xc0 <4>[46267.086346] [<ffffffff814b65a4>] raw_sendmsg+0x8b4/0x910 <4>[46267.086419] [<ffffffff814c10fa>] inet_sendmsg+0x4a/0xb0 <4>[46267.086491] [<ffffffff814459aa>] ? sock_update_classid+0x3a/0x50 <4>[46267.086562] [<ffffffff81444d67>] sock_sendmsg+0x117/0x140 <4>[46267.086638] [<ffffffff8151997b>] ? _spin_unlock_bh+0x1b/0x20 <4>[46267.086712] [<ffffffff8109d370>] ? autoremove_wake_function+0x0/0x40 <4>[46267.086785] [<ffffffff81495e80>] ? do_ip_setsockopt+0x90/0xd80 <4>[46267.086858] [<ffffffff8100be0e>] ? call_function_interrupt+0xe/0x20 <4>[46267.086936] [<ffffffff8118cb10>] ? ub_slab_ptr+0x20/0x90 <4>[46267.087006] [<ffffffff8118cb10>] ? ub_slab_ptr+0x20/0x90 <4>[46267.087081] [<ffffffff8118f2e8>] ? kmem_cache_alloc+0xd8/0x1e0 <4>[46267.087151] [<ffffffff81445599>] sys_sendto+0x139/0x190 <4>[46267.087229] [<ffffffff81448c0d>] ? sock_setsockopt+0x16d/0x6f0 <4>[46267.087303] [<ffffffff810efa47>] ? audit_syscall_entry+0x1d7/0x200 <4>[46267.087378] [<ffffffff810ef795>] ? __audit_syscall_exit+0x265/0x290 <4>[46267.087454] [<ffffffff81474885>] ? compat_sys_setsockopt+0x75/0x210 <4>[46267.087531] [<ffffffff81474b5f>] compat_sys_socketcall+0x13f/0x210 <4>[46267.087607] [<ffffffff8104dea3>] ia32_sysret+0x0/0x5 <4>[46267.087676] Code: 91 20 e2 01 75 29 48 89 de 4c 89 f7 e8 56 fa ff ff 85 c0 0f 84 68 fc ff ff 0f b6 4d c6 41 8b 45 00 e9 4d fb ff ff e8 7c 19 e9 e0 <0f> 0b eb fe f6 05 17 91 20 e2 80 74 ce 80 3d 5f 2e 00 00 00 74 <1>[46267.088023] RIP [<ffffffffa01e00a4>] nf_nat_setup_info+0x564/0x590
--- net/netfilter/nf_conntrack_core.c | 21 +++++++++++++++++---- 1 file changed, 17 insertions(+), 4 deletions(-)
net/netfilter/nf_conntrack_core.c +17 -4
@@ -318,6 +318,21 @@ static void death_by_timeout(unsigned long ul_conntrack)nf_ct_delete ((struct nf_conn *)ul_conntrack,0 ,0 ); } +static inline bool +nf_ct_key_equal (struct nf_conntrack_tuple_hash *h, +const struct nf_conntrack_tuple *tuple, +u16 zone) +{ +struct nf_conn *ct =nf_ct_tuplehash_to_ctrack (h); + +/* A conntrack can be recreated with the equal tuple, +* so we need to check that the conntrack is confirmed +*/ +return nf_ct_tuple_equal (tuple, &h->tuple) && +nf_ct_zone (ct) == zone && +nf_ct_is_confirmed (ct); +} +/* * Warning : * - Caller must take a reference on returned object
@@ -339,8 +354,7 @@ ____nf_conntrack_find(struct net *net, u16 zone,local_bh_disable (); begin:hlist_nulls_for_each_entry_rcu (h, n, &net->ct.hash[bucket], hnnode) { -if (nf_ct_tuple_equal (tuple, &h->tuple) && -nf_ct_zone (nf_ct_tuplehash_to_ctrack (h)) == zone) { +if (nf_ct_key_equal (h, tuple, zone)) {NF_CT_STAT_INC (net, found);local_bh_enable ();return h;
@@ -387,8 +401,7 @@ begin: !atomic_inc_not_zero(&ct->ct_general.use))) h = NULL; else { -if (unlikely (!nf_ct_tuple_equal (tuple, &h->tuple) || -nf_ct_zone (ct) != zone)) { +if (unlikely (!nf_ct_key_equal (h, tuple, zone))) {nf_ct_put (ct); goto begin; }
Signature · 1 line
2.5.2
PATCH 31/33 netfilter: nf_conntrack: don't release a conntrack with non-zero refcnt
In reply to Jiri SlabyFrom: Pablo Neira Ayuso <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
[ Upstream commit e53376bef2cd97d3e3f61fdc677fb8da7d03d0da ]
With this patch, the conntrack refcount is initially set to zero and it is bumped once it is added to any of the list, so we fulfill Eric's golden rule which is that all released objects always have a refcount that equals zero.
Andrey Vagin reports that nf_conntrack_free can't be called for a conntrack with non-zero ref-counter, because it can race with nf_conntrack_find_get().
A conntrack slab is created with SLAB_DESTROY_BY_RCU. Non-zero ref-counter says that this conntrack is used. So when we release a conntrack with non-zero counter, we break this assumption.
CPU1 CPU2____nf_conntrack_find()
nf_ct_put()
destroy_conntrack()
...
init_conntrack
__nf_conntrack_alloc (set use = 1)
atomic_inc_not_zero(&ct->use) (use = 2)
if (!l4proto->new(ct, skb, dataoff, timeouts))
nf_conntrack_free(ct); (use = 2 !!!)
...
__nf_conntrack_alloc (set use = 1)
if (!nf_ct_key_equal(h, tuple, zone))
nf_ct_put(ct); (use = 0)
destroy_conntrack()
/* continue to work with CT */After applying the path "[PATCH] netfilter: nf_conntrack: fix RCU race in nf_conntrack_find_get" another bug was triggered in destroy_conntrack():
<4>[67096.759334] ------------[ cut here ]------------ <2>[67096.759353] kernel BUG at net/netfilter/nf_conntrack_core.c:211! ... <4>[67096.759837] Pid: 498649, comm: atdd veid: 666 Tainted: G C --------------- 2.6.32-042stab084.18 #1 042stab084_18 /DQ45CB <4>[67096.759932] RIP: 0010:[<ffffffffa03d99ac>] [<ffffffffa03d99ac>] destroy_conntrack+0x15c/0x190 [nf_conntrack] <4>[67096.760255] Call Trace: <4>[67096.760255] [<ffffffff814844a7>] nf_conntrack_destroy+0x17/0x30 <4>[67096.760255] [<ffffffffa03d9bb5>] nf_conntrack_find_get+0x85/0x130 [nf_conntrack] <4>[67096.760255] [<ffffffffa03d9fb2>] nf_conntrack_in+0x352/0xb60 [nf_conntrack] <4>[67096.760255] [<ffffffffa048c771>] ipv4_conntrack_local+0x51/0x60 [nf_conntrack_ipv4] <4>[67096.760255] [<ffffffff81484419>] nf_iterate+0x69/0xb0 <4>[67096.760255] [<ffffffff814b5b00>] ? dst_output+0x0/0x20 <4>[67096.760255] [<ffffffff814845d4>] nf_hook_slow+0x74/0x110 <4>[67096.760255] [<ffffffff814b5b00>] ? dst_output+0x0/0x20 <4>[67096.760255] [<ffffffff814b66d5>] raw_sendmsg+0x775/0x910 <4>[67096.760255] [<ffffffff8104c5a8>] ? flush_tlb_others_ipi+0x128/0x130 <4>[67096.760255] [<ffffffff8100bc4e>] ? apic_timer_interrupt+0xe/0x20 <4>[67096.760255] [<ffffffff8100bc4e>] ? apic_timer_interrupt+0xe/0x20 <4>[67096.760255] [<ffffffff814c136a>] inet_sendmsg+0x4a/0xb0 <4>[67096.760255] [<ffffffff81444e93>] ? sock_sendmsg+0x13/0x140 <4>[67096.760255] [<ffffffff81444f97>] sock_sendmsg+0x117/0x140 <4>[67096.760255] [<ffffffff8102e299>] ? native_smp_send_reschedule+0x49/0x60 <4>[67096.760255] [<ffffffff81519beb>] ? _spin_unlock_bh+0x1b/0x20 <4>[67096.760255] [<ffffffff8109d930>] ? autoremove_wake_function+0x0/0x40 <4>[67096.760255] [<ffffffff814960f0>] ? do_ip_setsockopt+0x90/0xd80 <4>[67096.760255] [<ffffffff8100bc4e>] ? apic_timer_interrupt+0xe/0x20 <4>[67096.760255] [<ffffffff8100bc4e>] ? apic_timer_interrupt+0xe/0x20 <4>[67096.760255] [<ffffffff814457c9>] sys_sendto+0x139/0x190 <4>[67096.760255] [<ffffffff810efa77>] ? audit_syscall_entry+0x1d7/0x200 <4>[67096.760255] [<ffffffff810ef7c5>] ? __audit_syscall_exit+0x265/0x290 <4>[67096.760255] [<ffffffff81474daf>] compat_sys_socketcall+0x13f/0x210 <4>[67096.760255] [<ffffffff8104dea3>] ia32_sysret+0x0/0x5
I have reused the original title for the RFC patch that Andrey posted and most of the original patch description.
--- include/net/netfilter/nf_conntrack.h | 2 ++ net/netfilter/nf_conntrack_core.c | 34 +++++++++++++++++++++++++++++----- net/netfilter/nf_synproxy_core.c | 5 ++--- net/netfilter/xt_CT.c | 7 +------ 4 files changed, 34 insertions(+), 14 deletions(-)
include/net/netfilter/nf_conntrack.h +2 -0
@@ -293,6 +293,8 @@ extern unsigned int nf_conntrack_max;extern unsigned int nf_conntrack_hash_rnd;void init_nf_conntrack_hash_rnd (void ); +void nf_conntrack_tmpl_insert (struct net *net,struct nf_conn *tmpl); +#define NF_CT_STAT_INC (net, count) __this_cpu_inc((net)->ct.stat->count)#define NF_CT_STAT_INC_ATOMIC (net, count) this_cpu_inc((net)->ct.stat->count)
net/netfilter/nf_conntrack_core.c +29 -5
@@ -463,7 +463,9 @@ nf_conntrack_hash_check_insert(struct nf_conn *ct) goto out;add_timer (&ct->timeout); -nf_conntrack_get (&ct->ct_general); +smp_wmb (); +/* The caller holds a reference to this object */ +atomic_set (&ct->ct_general.use,2 );__nf_conntrack_hash_insert (ct, hash, repl_hash);NF_CT_STAT_INC (net, insert);spin_unlock_bh (&nf_conntrack_lock);
@@ -477,6 +479,21 @@ out: }EXPORT_SYMBOL_GPL (nf_conntrack_hash_check_insert); +/* deletion from this larval template list happens via nf_ct_put() */ +void nf_conntrack_tmpl_insert (struct net *net,struct nf_conn *tmpl) +{ +__set_bit (IPS_TEMPLATE_BIT , &tmpl->status); +__set_bit (IPS_CONFIRMED_BIT , &tmpl->status); +nf_conntrack_get (&tmpl->ct_general); + +spin_lock_bh (&nf_conntrack_lock); +/* Overload tuple linked list to put us in template list. */ +hlist_nulls_add_head_rcu (&tmpl->tuplehash[IP_CT_DIR_ORIGINAL ].hnnode, + &net->ct.tmpl); +spin_unlock_bh (&nf_conntrack_lock); +} +EXPORT_SYMBOL_GPL (nf_conntrack_tmpl_insert); +/* Confirm a connection given skb; places it in hash table */ int __nf_conntrack_confirm (struct sk_buff *skb)
@@ -748,11 +765,10 @@ __nf_conntrack_alloc(struct net *net, u16 zone, nf_ct_zone->id = zone; }#endif -/* -* changes to lookup keys must be done before setting refcnt to 1 +/* Because we use RCU lookups, we set ct_general.use to zero before +* this is inserted in any list. */ -smp_wmb (); -atomic_set (&ct->ct_general.use,1 ); +atomic_set (&ct->ct_general.use,0 );return ct;#ifdef CONFIG_NF_CONNTRACK_ZONES
@@ -776,6 +792,11 @@ void nf_conntrack_free(struct nf_conn *ct) {struct net *net =nf_ct_net (ct); +/* A freed object has refcnt == 0, that's +* the golden rule for SLAB_DESTROY_BY_RCU +*/ +NF_CT_ASSERT (atomic_read (&ct->ct_general.use) ==0 ); +nf_ct_ext_destroy (ct);atomic_dec (&net->ct.count);nf_ct_ext_free (ct);
@@ -870,6 +891,9 @@ init_conntrack(struct net *net, struct nf_conn *tmpl,NF_CT_STAT_INC (net, new); } +/* Now it is inserted into the unconfirmed list, bump refcount */ +nf_conntrack_get (&ct->ct_general); +/* Overload tuple linked list to put us in unconfirmed list. */ hlist_nulls_add_head_rcu (&ct->tuplehash[IP_CT_DIR_ORIGINAL ].hnnode, &net->ct.unconfirmed);
net/netfilter/nf_synproxy_core.c +2 -3
@@ -362,9 +362,8 @@ static int __net_init synproxy_net_init(struct net *net) goto err2;if (!nfct_synproxy_ext_add (ct)) goto err2; -__set_bit (IPS_TEMPLATE_BIT , &ct->status); -__set_bit (IPS_CONFIRMED_BIT , &ct->status); +nf_conntrack_tmpl_insert (net, ct); snet->tmpl = ct; snet->stats = alloc_percpu(struct synproxy_stats);
@@ -389,7 +388,7 @@ static void __net_exit synproxy_net_exit(struct net *net) {struct synproxy_net *snet =synproxy_pernet (net); -nf_conntrack_free (snet->tmpl); +nf_ct_put (snet->tmpl);synproxy_proc_exit (net);free_percpu (snet->stats); }
net/netfilter/xt_CT.c +1 -6
@@ -226,12 +226,7 @@ static int xt_ct_tg_check(const struct xt_tgchk_param *par, goto err3; } -__set_bit (IPS_TEMPLATE_BIT , &ct->status); -__set_bit (IPS_CONFIRMED_BIT , &ct->status); - -/* Overload tuple linked list to put us in template list. */ -hlist_nulls_add_head_rcu (&ct->tuplehash[IP_CT_DIR_ORIGINAL ].hnnode, - &par->net->ct.tmpl); +nf_conntrack_tmpl_insert (par->net, ct); out: info->ct = ct;return 0 ;
Signature · 1 line
2.5.2
PATCH 32/33 PCI: Add dev_flags bit to access VPD through function 0
In reply to Jiri SlabyFrom: Mark Rustad <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
commit 932c435caba8a2ce473a91753bad0173269ef334 upstream.
Add a dev_flags bit, PCI_DEV_FLAGS_VPD_REF_F0, to access VPD through function 0 to provide VPD access on other functions. This is for hardware devices that provide copies of the same VPD capability registers in multiple functions. Because the kernel expects that each function has its own registers, both the locking and the state tracking are affected by VPD accesses to different functions.
On such devices for example, if a VPD write is performed on function 0, *any* later attempt to read VPD from any other function of that device will hang. This has to do with how the kernel tracks the expected value of the F bit per function.
Concurrent accesses to different functions of the same device can not only hang but also corrupt both read and write VPD data.
When hangs occur, typically the error message:
vpd r/w failed. This is likely a firmware bug on this device.
will be seen.
Never set this bit on function 0 or there will be an infinite recursion.
--- drivers/pci/access.c | 61 +++++++++++++++++++++++++++++++++++++++++++++++++++- include/linux/pci.h | 2 ++ 2 files changed, 62 insertions(+), 1 deletion(-)
drivers/pci/access.c +60 -1
@@ -359,6 +359,56 @@ static const struct pci_vpd_ops pci_vpd_pci22_ops = { .release = pci_vpd_pci22_release, }; +static ssize_t pci_vpd_f0_read (struct pci_dev *dev,loff_t pos,size_t count, +void *arg) +{ +struct pci_dev *tdev =pci_get_slot (dev->bus,PCI_SLOT (dev->devfn)); +ssize_t ret; + +if (!tdev) +return -ENODEV ; + + ret =pci_read_vpd (tdev, pos, count, arg); +pci_dev_put (tdev); +return ret; +} + +static ssize_t pci_vpd_f0_write (struct pci_dev *dev,loff_t pos,size_t count, +const void *arg) +{ +struct pci_dev *tdev =pci_get_slot (dev->bus,PCI_SLOT (dev->devfn)); +ssize_t ret; + +if (!tdev) +return -ENODEV ; + + ret =pci_write_vpd (tdev, pos, count, arg); +pci_dev_put (tdev); +return ret; +} + +static const struct pci_vpd_ops pci_vpd_f0_ops = { + .read = pci_vpd_f0_read, + .write = pci_vpd_f0_write, + .release = pci_vpd_pci22_release, +}; + +static int pci_vpd_f0_dev_check (struct pci_dev *dev) +{ +struct pci_dev *tdev =pci_get_slot (dev->bus,PCI_SLOT (dev->devfn)); +int ret =0 ; + +if (!tdev) +return -ENODEV ; +if (!tdev->vpd || !tdev->multifunction || + dev->class != tdev->class || dev->vendor != tdev->vendor || + dev->device != tdev->device) + ret = -ENODEV ; + +pci_dev_put (tdev); +return ret; +} +int pci_vpd_pci22_init (struct pci_dev *dev) {struct pci_vpd_pci22 *vpd;
@@ -367,12 +417,21 @@ int pci_vpd_pci22_init(struct pci_dev *dev) cap =pci_find_capability (dev,PCI_CAP_ID_VPD );if (!cap)return -ENODEV ; +if (dev->dev_flags &PCI_DEV_FLAGS_VPD_REF_F0 ) { +int ret =pci_vpd_f0_dev_check (dev); + +if (ret) +return ret; + } vpd =kzalloc (sizeof (*vpd),GFP_ATOMIC );if (!vpd)return -ENOMEM ; vpd->base.len =PCI_VPD_PCI22_SIZE ; - vpd->base.ops = &pci_vpd_pci22_ops; +if (dev->dev_flags &PCI_DEV_FLAGS_VPD_REF_F0 ) + vpd->base.ops = &pci_vpd_f0_ops; +else + vpd->base.ops = &pci_vpd_pci22_ops;mutex_init (&vpd->lock); vpd->cap = cap; vpd->busy = false;
include/linux/pci.h +2 -0
@@ -170,6 +170,8 @@ enum pci_dev_flags {PCI_DEV_FLAGS_NO_D3 = (__force pci_dev_flags_t )2 ,/* Provide indication device is assigned by a Virtual Machine Manager */ PCI_DEV_FLAGS_ASSIGNED = (__force pci_dev_flags_t )4 , +/* Get VPD from function 0 VPD */ + PCI_DEV_FLAGS_VPD_REF_F0 = (__force pci_dev_flags_t) (1 << 8), }; enum pci_irq_reroute_variant {
Signature · 1 line
2.5.2
PATCH 33/33 PCI: Add VPD function 0 quirk for Intel Ethernet devices
In reply to Jiri SlabyFrom: Mark Rustad <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
commit 7aa6ca4d39edf01f997b9e02cf6d2fdeb224f351 upstream.
Set the PCI_DEV_FLAGS_VPD_REF_F0 flag on all Intel Ethernet device functions other than function 0, so that on multi-function devices, we will always read VPD from function 0 instead of from the other functions.
--- drivers/pci/quirks.c | 9 +++++++++ 1 file changed, 9 insertions(+)
drivers/pci/quirks.c +9 -0
@@ -1849,6 +1849,15 @@ static void quirk_netmos(struct pci_dev *dev)DECLARE_PCI_FIXUP_CLASS_HEADER (PCI_VENDOR_ID_NETMOS ,PCI_ANY_ID ,PCI_CLASS_COMMUNICATION_SERIAL ,8 , quirk_netmos); +static void quirk_f0_vpd_link (struct pci_dev *dev) +{ +if (!dev->multifunction || !PCI_FUNC (dev->devfn)) +return ; + dev->dev_flags |=PCI_DEV_FLAGS_VPD_REF_F0 ; +} +DECLARE_PCI_FIXUP_CLASS_EARLY (PCI_VENDOR_ID_INTEL ,PCI_ANY_ID , +PCI_CLASS_NETWORK_ETHERNET ,8 , quirk_f0_vpd_link); +static void quirk_e100_interrupt (struct pci_dev *dev) {u16 command, pmcsr;
Signature · 1 line
2.5.2
Hi Jiry,
Maybe you would want to consider this:
https://patchwork.ozlabs.org/patch/459088/
It has already found its ways in other stable kernels, despite not being cc'ed to stable.
Regards,
Nikolay
Quoted 143 lines · Jiri Slaby
This is the start of the stable review cycle for the 3.12.48 release. There are 33 patches in this series, all will be posted as a response to this one. If anyone has any issues with these being applied, please let me know.
Responses should be made by Thu Sep 17 16:20:25 CEST 2015. Anything received after that time might be too late.
The whole patch series can be found in one patch at:
http://kernel.org/pub/linux/kernel/people/jirislaby/stable-review/patch-3.12.48-rc1.xzand the diffstat can be found below.
thanks,
js===============
Akinobu Mita (1):
bio: fix argument of__bio_add_page()formax_sectors>0xffffAndrey Vagin (1):
netfilter:nf_conntrack: fix RCU race innf_conntrack_find_getAngga (1):
ipv6: Make MLD packets to only be processed locallyBenjamin LaHaise (1):
aio: fixreqs_availablehandlingDan Carpenter (1):
rds: fix an integer overflow test inrds_info_getsockopt()Daniel Borkmann (1):
rtnetlink: verifyIFLA_VF_INFOattributes before passing them to
driverDave Airlie (1):
drm/radeon: fix hotplug race at startupDavid Milburn (1):
mtip32xx: dynamically allocate buffer in debugfs functionsEdward Hyunkoo Jee (1):
inet: frags: fix defragmented packet's IP header foraf_packetEric Dumazet (2):
net: graceful exit fromnetif_alloc_netdev_queues()
ipv6: lock socket inip6_datagram_connect()Florian Westphal (1):
netlink: don't hold mutex in rcu callback when releasing mmapd ringHeinz Mauelshagen (1):
dm cache mq: fix memory allocation failure for large cache devicesHerbert Xu (3):
net: Clone skb before setting peeked flag
net: Fix skb csum races when peeking
net: Fixskb_set_peekeduse-after-free bugJack Morgenstein (1):
net/mlx4_core: Fix wrong index in propagating port change event to VFsJames Smart (1):
lpfc: Fix scsi prep dma buf error.Julian Anastasov (2):
net: do not process device backlog during unregistration
net: callrcu_read_lockearly inprocess_backlogMark Rustad (2):
PCI: Adddev_flagsbit to access VPD through function 0
PCI: Add VPD function 0 quirk for Intel Ethernet devicesMika Westerberg (1):
mfd:lpc_ich: Assign subdevice ids automaticallyNikolay Aleksandrov (3):
bridge: mdb: zero out the localbr_ipvariable before use
bridge: mdb: fix double add notification
bonding: fix destruction of bond with devices different from
arphrd_etherOleg Nesterov (1):
net: pktgen: fix race betweenpktgen_thread_worker()and
kthread_stop()Pablo Neira Ayuso (1):
netfilter:nf_conntrack: don't release a conntrack with non-zero
refcntShirish Pargaonkar (1):
cifs: Send a logoff request before removing a smb sessionStephen Smalley (1):
net/tipc: initialize security state for new connection socketTilman Schmidt (1):
isdn/gigaset: resettty->receive_roomwhen attachingser_gigasetTimo Teräs (1):
ip_tunnel: fix ipv4 pmtu check to honor inner ip header dfdingtianhong (1):
bonding: correct the MAC address for "follow"fail_over_macpolicydrivers/block/mtip32xx/mtip32xx.c | 47 +++++++++--- drivers/gpu/drm/radeon/radeon_irq_kms.c | 5 ++ drivers/isdn/gigaset/ser-gigaset.c | 11 ++- drivers/md/dm-cache-policy-mq.c | 4 +- drivers/mfd/lpc_ich.c | 8 +- drivers/net/bonding/bond_main.c | 20 +++++ drivers/net/ethernet/mellanox/mlx4/eq.c | 4 +- drivers/pci/access.c | 61 ++++++++++++++- drivers/pci/quirks.c | 9 +++ drivers/scsi/lpfc/lpfc_scsi.c | 2 +- fs/aio.c | 77 ++++++++++++++++++- fs/bio.c | 2 +- fs/cifs/connect.c | 25 +++++-- fs/cifs/smb2transport.c | 10 ++- fs/cifs/transport.c | 11 ++- include/linux/pci.h | 2 + include/net/ip.h | 1 + include/net/netfilter/nf_conntrack.h | 2 + net/bridge/br_mdb.c | 3 +- net/core/datagram.c | 45 ++++++++++- net/core/dev.c | 38 +++++----- net/core/pktgen.c | 4 +- net/core/rtnetlink.c | 128 ++++++++++++++++---------------- net/ipv4/datagram.c | 16 +++- net/ipv4/ip_fragment.c | 7 +- net/ipv4/ip_tunnel.c | 8 +- net/ipv6/datagram.c | 20 +++-- net/ipv6/ip6_input.c | 6 +- net/netfilter/nf_conntrack_core.c | 55 +++++++++++--- net/netfilter/nf_synproxy_core.c | 5 +- net/netfilter/xt_CT.c | 7 +- net/netlink/af_netlink.c | 79 ++++++++++++-------- net/rds/info.c | 2 +- net/tipc/socket.c | 1 + 34 files changed, 535 insertions(+), 190 deletions(-)
Quoted 8 lines · Jiri Slaby
This is the start of the stable review cycle for the 3.12.48 release. There are 33 patches in this series, all will be posted as a response to this one. If anyone has any issues with these being applied, please let me know.
Responses should be made by Thu Sep 17 16:20:25 CEST 2015. Anything received after that time might be too late.
Quoted 6 lines
The whole patch series can be found in one patch at:
http://kernel.org/pub/linux/kernel/people/jirislaby/stable-review/patch-3.12.48-rc1.xzand the diffstat can be found below.
thanks,
js
Jiri,
I am seeing problems during PCI scans with this patch. I had to boot it in recovery mode once and it is booting fine after that, however, this is a concern
Could these be the reason?
Mark Rustad (2):
PCI: Add dev_flags bit to access VPD through function 0
PCI: Add VPD function 0 quirk for Intel Ethernet devices
PCI-DMA: Intel(R) Virtualization Technology for Directed I/O
------------[ cut here ]------------ WARNING: CPU: 0 PID: 1 at drivers/pci/search.c:46 pci_find_upstream_pcie_bridge+ 0x65/0x90() Modules linked in: CPU: 0 PID: 1 Comm: swapper/0 Not tainted 3.12.48+ #29 Hardware name: System76, Inc. Wild Dog Performance/H87-PLUS, BIOS 0705
12/05/201
3
0000000000000009 ffff88040952bd28 ffffffff81718bd7 0000000000000000 ffff88040952bd60 ffffffff8106497d ffff88040801c098 ffff88040801c000 ffff88040801c098 0000000000000000 ffff88040f0058a0 ffff88040952bd70
Call Trace: [<ffffffff81718bd7>] dump_stack+0x45/0x56 [<ffffffff8106497d>] warn_slowpath_common+0x7d/0xa0 [<ffffffff81064a5a>] warn_slowpath_null+0x1a/0x20 [<ffffffff813b75d5>] pci_find_upstream_pcie_bridge+0x65/0x90 [<ffffffff815fb10d>] intel_iommu_add_device+0x4d/0x220 [<ffffffff815f2000>] ? bus_set_iommu+0x50/0x50 [<ffffffff815f202a>] add_iommu_group+0x2a/0x60 [<ffffffff81497aab>] bus_for_each_dev+0x6b/0xb0 [<ffffffff815f1ff8>] bus_set_iommu+0x48/0x50 [<ffffffff81d93ef3>] intel_iommu_init+0xa9d/0xb9c [<ffffffff81d4a730>] ? memblock_find_dma_reserve+0x124/0x124 [<ffffffff81d4a742>] pci_iommu_init+0x12/0x3c [<ffffffff810020d2>] do_one_initcall+0xd2/0x1a0 [<ffffffff81087700>] ? parse_args+0x160/0x490 [<ffffffff81d40ef6>] kernel_init_freeable+0x144/0x1cc [<ffffffff8170cfe0>] ? rest_init+0x80/0x80 [<ffffffff8170cfee>] kernel_init+0xe/0x190 [<ffffffff81728fd8>] ret_from_fork+0x58/0x90 [<ffffffff8170cfe0>] ? rest_init+0x80/0x80 ---[ end trace 6540c4180167ec85 ]---
Scanning for low memory corruption every 60 seconds Initialise module verification
audit: initializing netlink socket (disabled) type=2000 audit(1442332755.772:1): initialized bounce pool size: 64 pages
HugeTLB registered 2 MB page size, pre-allocated 0 pages
zbud: loaded
Signature · 5 lines
Shuah Khan Sr. Linux Kernel Developer Open Source Innovation Group Samsung Research America (Silicon Valley) [email protected] | (970) 217-8978
Quoted 8 lines · Jiri Slaby
This is the start of the stable review cycle for the 3.12.48 release. There are 33 patches in this series, all will be posted as a response to this one. If anyone has any issues with these being applied, please let me know.
Responses should be made by Thu Sep 17 16:20:25 CEST 2015. Anything received after that time might be too late.
Build results:
total: 123 pass: 123 fail: 0 Qemu test results: total: 76 pass: 76 fail: 0
Details are available at http://server.roeck-us.net:8010/builders.
Guenter
Re: [PATCH 3.12 16/33] isdn/gigaset: reset tty->receive_room when attaching ser_gigaset
In reply to a message not in the archiveAm 16.09.2015 um 01:08 schrieb Peter Hurley:
Quoted 19 lines
On Tue, Sep 15, 2015 at 10:22 AM, Jiri Slaby <[email protected] <mailto:[email protected]>> wrote:
From: Tilman Schmidt <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
[ Upstream commit
fd98e9419d8d622a4de91f76b306af6aa627aa9c]Commit
79901317ce80("n_tty: Don't flush buffer when closing ldisc"), first merged in kernel release 3.10, caused the following regression in the Gigaset M101 driver:Again, I'll just note my objection to this commit log.
Quoted 7 lines
This driver was always broken because it never initialized
tty->receive_room,
but rather relied on common but not guaranteed circumstances to function.The commit noted simply made the underlying bug more evident, but the root cause was from the original merge commit of this driver.
I must admit I still don't understand that objection. The meaning of the term "regression" is simply that something which previously worked stopped working. It doesn't imply any statement about the root cause.
The ser-gigaset driver worked before the introduction of commit 79901317ce80. It didn't work anymore after the introduction of that commit. So it is correct, and does not contradict your statements above in any way, to state that commit introduced the described regression.
Signature · 4 lines
Tilman Schmidt E-Mail: [email protected] Bonn, Germany Diese Nachricht besteht zu 100% aus wiederverwerteten Bits. Ungeöffnet mindestens haltbar bis: (siehe Rückseite)
Re: [PATCH 3.12 16/33] isdn/gigaset: reset tty->receive_room when attaching ser_gigaset
Quoted 28 lines · Tilman Schmidt
Am 16.09.2015 um 01:08 schrieb Peter Hurley:
Quoted 19 lines
On Tue, Sep 15, 2015 at 10:22 AM, Jiri Slaby <[email protected] <mailto:[email protected]>> wrote:
From: Tilman Schmidt <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
[ Upstream commit
fd98e9419d8d622a4de91f76b306af6aa627aa9c]Commit
79901317ce80("n_tty: Don't flush buffer when closing ldisc"), first merged in kernel release 3.10, caused the following regression in the Gigaset M101 driver:Again, I'll just note my objection to this commit log.
Quoted 7 lines
This driver was always broken because it never initialized
tty->receive_room,
but rather relied on common but not guaranteed circumstances to function.The commit noted simply made the underlying bug more evident, but the root cause was from the original merge commit of this driver.
Quoted 8 lines
I must admit I still don't understand that objection. The meaning of the term "regression" is simply that something which previously worked stopped working. It doesn't imply any statement about the root cause.
The ser-gigaset driver worked before the introduction of commit
79901317ce80. It didn't work anymore after the introduction of that commit. So it is correct, and does not contradict your statements above in any way, to state that commit introduced the described regression.
By asserting that commit 79901317ce80 caused the regression, you're claiming that this fix is unnecessary for kernel versions prior to 3.10
Are you certain that no other sequence of state leads to the same condition (and thus requiring the same fix) in earlier kernel versions?
Regards,
Peter Hurley
Quoted 22 lines · Shuah Khan
Quoted 8 lines · Jiri Slaby
This is the start of the stable review cycle for the 3.12.48 release. There are 33 patches in this series, all will be posted as a response to this one. If anyone has any issues with these being applied, please let me know.
Responses should be made by Thu Sep 17 16:20:25 CEST 2015. Anything received after that time might be too late.
Quoted 6 lines
The whole patch series can be found in one patch at:
http://kernel.org/pub/linux/kernel/people/jirislaby/stable-review/patch-3.12.48-rc1.xzand the diffstat can be found below.
thanks,
jsJiri,
I am seeing problems during PCI scans with this patch. I had to boot it in recovery mode once and it is booting fine after that, however, this is a concern
Quoted 5 lines
Could these be the reason?
Mark Rustad (2):
PCI: Adddev_flagsbit to access VPD through function 0
PCI: Add VPD function 0 quirk for Intel Ethernet devices
Hi, maybe. Could you either revert those patches or test commit a7775d15b11a277f8af from this tree:git://git.kernel.org/pub/scm/linux/kernel/git/jirislaby/linux-stable.git
to verify?
thanks,
Signature · 2 lines
js suse labs
Re: [PATCH 3.12 16/33] isdn/gigaset: reset tty->receive_room when attaching ser_gigaset
In reply to Peter HurleyAm 16.09.2015 um 03:18 schrieb Peter Hurley:
Quoted 32 lines · Tilman Schmidt
Quoted 31 lines
Am 16.09.2015 um 01:08 schrieb Peter Hurley:
Quoted 18 lines · Jiri Slaby
From: Tilman Schmidt <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
[ Upstream commit
fd98e9419d8d622a4de91f76b306af6aa627aa9c]Commit
79901317ce80("n_tty: Don't flush buffer when closing ldisc"), first merged in kernel release 3.10, caused the following regression in the Gigaset M101 driver:Again, I'll just note my objection to this commit log.
Quoted 7 lines
This driver was always broken because it never initialized
tty->receive_room,
but rather relied on common but not guaranteed circumstances to function.The commit noted simply made the underlying bug more evident, but the root cause was from the original merge commit of this driver.
I must admit I still don't understand that objection. The meaning of the term "regression" is simply that something which previously worked stopped working. It doesn't imply any statement about the root cause.
Quoted 7 lines
Quoted 4 lines
The ser-gigaset driver worked before the introduction of commit
79901317ce80. It didn't work anymore after the introduction of that commit. So it is correct, and does not contradict your statements above in any way, to state that commit introduced the described regression.By asserting that commit
79901317ce80caused the regression, you're claiming that this fix is unnecessary for kernel versions prior to 3.10
Correct.
Quoted 2 lines
Are you certain that no other sequence of state leads to the same condition (and thus requiring the same fix) in earlier kernel versions?
Reasonably certain, yes, for three reasons:
- There where no reports of that problem before 3.10.
- My own tests did never encounter that condition, and even after being made aware of it I was not able to come up with a test that would provoke it with a kernel version before 3.10.
- The requirement for line disciplines to set receive_room wasn't (and btw still isn't) documented anywhere, so it's unlikely anything actively relied on it.
But if you want to propose that patch for inclusion in earlier stable releases I won't oppose it.
Signature · 4 lines
Tilman Schmidt E-Mail: [email protected] Bonn, Germany Diese Nachricht besteht zu 100% aus wiederverwerteten Bits. Ungeöffnet mindestens haltbar bis: (siehe Rückseite)
Quoted 5 lines · Nikolay Borisov
Maybe you would want to consider this:
https://patchwork.ozlabs.org/patch/459088/It has already found its ways in other stable kernels, despite not being cc'ed to stable.
Hi, it does not apply to 3.12 and is not in 3.14 where I source patches from too. So you would have to provide a backport...
thanks,
Signature · 2 lines
js suse labs
Re: [PATCH 3.12 16/33] isdn/gigaset: reset tty->receive_room when attaching ser_gigaset
In reply to Tilman SchmidtQuoted 46 lines · Tilman Schmidt
Am 16.09.2015 um 03:18 schrieb Peter Hurley:
Quoted 32 lines · Tilman Schmidt
Quoted 31 lines
Am 16.09.2015 um 01:08 schrieb Peter Hurley:
Quoted 18 lines · Jiri Slaby
From: Tilman Schmidt <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
[ Upstream commit
fd98e9419d8d622a4de91f76b306af6aa627aa9c]Commit
79901317ce80("n_tty: Don't flush buffer when closing ldisc"), first merged in kernel release 3.10, caused the following regression in the Gigaset M101 driver:Again, I'll just note my objection to this commit log.
Quoted 7 lines
This driver was always broken because it never initialized
tty->receive_room,
but rather relied on common but not guaranteed circumstances to function.The commit noted simply made the underlying bug more evident, but the root cause was from the original merge commit of this driver.
I must admit I still don't understand that objection. The meaning of the term "regression" is simply that something which previously worked stopped working. It doesn't imply any statement about the root cause.
Quoted 7 lines
Quoted 4 lines
The ser-gigaset driver worked before the introduction of commit
79901317ce80. It didn't work anymore after the introduction of that commit. So it is correct, and does not contradict your statements above in any way, to state that commit introduced the described regression.By asserting that commit
79901317ce80caused the regression, you're claiming that this fix is unnecessary for kernel versions prior to 3.10Correct.
Quoted 2 lines
Are you certain that no other sequence of state leads to the same condition (and thus requiring the same fix) in earlier kernel versions?
Quoted 8 lines
Reasonably certain, yes, for three reasons:
- There where no reports of that problem before 3.10.- My own tests did never encounter that condition, and even after being made aware of it I was not able to come up with a test that would provoke it with a kernel version before 3.10.
Do any of your tests switch to this line discipline from any other than N_TTY? Because if so, you would realize that whatever _that_ line discipline sets tty->receive_room to when it initializes, is what your line discipline will use without this fix.
So for example, if you manually set N_PPP (as if by user error) and then set this line discipline, tty->receive_room will be 64K, not 4K.
Quoted 3 lines
- The requirement for line disciplines to set
receive_roomwasn't (and btw still isn't) documented anywhere, so it's unlikely anything actively relied on it.
Nevertheless, that is the requirement, and what every other in-tree line discipline does.
Regards,
Peter Hurley
Quoted 22 lines · Shuah Khan
Quoted 8 lines · Jiri Slaby
This is the start of the stable review cycle for the 3.12.48 release. There are 33 patches in this series, all will be posted as a response to this one. If anyone has any issues with these being applied, please let me know.
Responses should be made by Thu Sep 17 16:20:25 CEST 2015. Anything received after that time might be too late.
Quoted 6 lines
The whole patch series can be found in one patch at:
http://kernel.org/pub/linux/kernel/people/jirislaby/stable-review/patch-3.12.48-rc1.xzand the diffstat can be found below.
thanks,
jsJiri,
I am seeing problems during PCI scans with this patch. I had to boot it in recovery mode once and it is booting fine after that, however, this is a concern
Quoted 5 lines
Could these be the reason?
Mark Rustad (2):
PCI: Adddev_flagsbit to access VPD through function 0
PCI: Add VPD function 0 quirk for Intel Ethernet devices
Ok, so I released 3.12.48 without these two, but left them in the queue for the next release as they don't look as they caused the issue. If you can confirm whether they introduce the problem or not, it would be great.
thanks,
Signature · 2 lines
js suse labs
Quoted 6 lines · Guenter Roeck
Build results:
total: 123 pass: 123 fail: 0 Qemu test results: total: 76 pass: 76 fail: 0Details are available at http://server.roeck-us.net:8010/builders.
Thanks.
Signature · 2 lines
js suse labs
Re: [PATCH 3.12 16/33] isdn/gigaset: reset tty->receive_room when attaching ser_gigaset
In reply to Peter HurleyAm 17.09.2015 um 20:13 schrieb Peter Hurley:
Quoted 52 lines · Tilman Schmidt
Quoted 48 lines
Am 16.09.2015 um 03:18 schrieb Peter Hurley:
Quoted 32 lines · Tilman Schmidt
Quoted 31 lines
Am 16.09.2015 um 01:08 schrieb Peter Hurley:
Quoted 18 lines · Jiri Slaby
From: Tilman Schmidt <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
[ Upstream commit
fd98e9419d8d622a4de91f76b306af6aa627aa9c]Commit
79901317ce80("n_tty: Don't flush buffer when closing ldisc"), first merged in kernel release 3.10, caused the following regression in the Gigaset M101 driver:Again, I'll just note my objection to this commit log.
Quoted 7 lines
This driver was always broken because it never initialized
tty->receive_room,
but rather relied on common but not guaranteed circumstances to function.The commit noted simply made the underlying bug more evident, but the root cause was from the original merge commit of this driver.
I must admit I still don't understand that objection. The meaning of the term "regression" is simply that something which previously worked stopped working. It doesn't imply any statement about the root cause.
Quoted 7 lines
Quoted 4 lines
The ser-gigaset driver worked before the introduction of commit
79901317ce80. It didn't work anymore after the introduction of that commit. So it is correct, and does not contradict your statements above in any way, to state that commit introduced the described regression.By asserting that commit
79901317ce80caused the regression, you're claiming that this fix is unnecessary for kernel versions prior to 3.10Correct.
Quoted 2 lines
Are you certain that no other sequence of state leads to the same condition (and thus requiring the same fix) in earlier kernel versions?
Reasonably certain, yes, for three reasons:
- There where no reports of that problem before 3.10.
Quoted 5 lines
Quoted 3 lines
- My own tests did never encounter that condition, and even after being made aware of it I was not able to come up with a test that would provoke it with a kernel version before 3.10.
Do any of your tests switch to this line discipline from any other than
N_TTY?
Of course not. That wouldn't make any sense.
Quoted 1 line
So for example, if you manually set
N_PPP(as if by user error)
User error wouldn't suffice, as the LD would get reset to N_TTY when the serial device is closed. You would have to write a program that deliberately switched the LD first to N_PPP and then to N_GIGASET_M101 without closing the device in between.
Quoted 1 line
and then set this line discipline,
tty->receive_roomwill be 64K, not 4K.
That wouldn't affect the operation of ser_gigaset, so even if I had set up such a contrived test scenario it wouldn't have exposed any problem. Only setting tty->receive_room to 0 causes the problem, and N_TTY with commit 79901317ce80 is the only LD which does that.
Quoted 6 lines
Quoted 3 lines
- The requirement for line disciplines to set
receive_roomwasn't (and btw still isn't) documented anywhere, so it's unlikely anything actively relied on it.Nevertheless, that is the requirement, and what every other in-tree line discipline does.
Your word for it. Still I don't understand the curious resistance to documenting it. If it is the requirement, why keep it secret?
Regards,
Tilman
Signature · 4 lines
Tilman Schmidt E-Mail: [email protected] Bonn, Germany Diese Nachricht besteht zu 100% aus wiederverwerteten Bits. Ungeöffnet mindestens haltbar bis: (siehe Rückseite)
Quoted 13 lines · Jiri Slaby
Quoted 10 lines · Shuah Khan
Jiri,
I am seeing problems during PCI scans with this patch. I had to boot it in recovery mode once and it is booting fine after that, however, this is a concern
Could these be the reason?
Quoted 7 lines
Quoted 3 lines
Mark Rustad (2):
PCI: Adddev_flagsbit to access VPD through function 0
PCI: Add VPD function 0 quirk for Intel Ethernet devicesOk, so I released 3.12.48 without these two, but left them in the queue for the next release as they don't look as they caused the issue. If you can confirm whether they introduce the problem or not, it would be great.
I haven't been able to see a way for the patches to be involved, but I have no objection to deferring these. My only concern is what the cause actually was.
--
Mark Rustad, Networking Division, Intel Corporation
Re: [PATCH 3.12 16/33] isdn/gigaset: reset tty->receive_room when attaching ser_gigaset
In reply to Tilman SchmidtQuoted 59 lines · Tilman Schmidt
Am 17.09.2015 um 20:13 schrieb Peter Hurley:
Quoted 52 lines · Tilman Schmidt
Quoted 48 lines
Am 16.09.2015 um 03:18 schrieb Peter Hurley:
Quoted 32 lines · Tilman Schmidt
Quoted 31 lines
Am 16.09.2015 um 01:08 schrieb Peter Hurley:
Quoted 18 lines · Jiri Slaby
From: Tilman Schmidt <[email protected]>
3.12-stable review patch. If anyone has any objections, please let me know.
===============
[ Upstream commit
fd98e9419d8d622a4de91f76b306af6aa627aa9c]Commit
79901317ce80("n_tty: Don't flush buffer when closing ldisc"), first merged in kernel release 3.10, caused the following regression in the Gigaset M101 driver:Again, I'll just note my objection to this commit log.
Quoted 7 lines
This driver was always broken because it never initialized
tty->receive_room,
but rather relied on common but not guaranteed circumstances to function.The commit noted simply made the underlying bug more evident, but the root cause was from the original merge commit of this driver.
I must admit I still don't understand that objection. The meaning of the term "regression" is simply that something which previously worked stopped working. It doesn't imply any statement about the root cause.
Quoted 7 lines
Quoted 4 lines
The ser-gigaset driver worked before the introduction of commit
79901317ce80. It didn't work anymore after the introduction of that commit. So it is correct, and does not contradict your statements above in any way, to state that commit introduced the described regression.By asserting that commit
79901317ce80caused the regression, you're claiming that this fix is unnecessary for kernel versions prior to 3.10Correct.
Quoted 2 lines
Are you certain that no other sequence of state leads to the same condition (and thus requiring the same fix) in earlier kernel versions?
Reasonably certain, yes, for three reasons:
- There where no reports of that problem before 3.10.Quoted 5 lines
Quoted 3 lines
- My own tests did never encounter that condition, and even after being made aware of it I was not able to come up with a test that would provoke it with a kernel version before 3.10.
Do any of your tests switch to this line discipline from any other than
N_TTY?
Quoted 8 lines
Of course not. That wouldn't make any sense.
Quoted 1 line
So for example, if you manually set
N_PPP(as if by user error)User error wouldn't suffice, as the LD would get reset to
N_TTYwhen the serial device is closed. You would have to write a program that deliberately switched the LD first toN_PPPand then toN_GIGASET_M101without closing the device in between.
???
The tool you authored will do it from the command line
$ ldattach PPP /dev/ttyS1 $ ldattach GIGASET_M101 /dev/ttyS1
Note that nothing here closes the serial device 'in between', and the tty core has switched directly from PPP to GIGASET_M101. n_tty->receive_room is now 64K.
Please add switching from line disciplines other than N_TTY to your regression testing.
Quoted 3 lines
Quoted 1 line
and then set this line discipline,
tty->receive_roomwill be 64K, not 4K.That wouldn't affect the operation of
ser_gigaset,
I've explained this before to you, but here it is again:
tty->receive_room announces the maximum amt of data the line discipline can accept from tty core with each call to its receive_buf() method (for line disciplines that don't provide flow control).
If the line discipline sets ->receive_room to 64K but can only handle 8K (as in the case of GIGASET_M101), then data loss should be the expected result.
Quoted 9 lines
so even if I had set
up such a contrived test scenario it wouldn't have exposed any problem. Only settingtty->receive_roomto 0 causes the problem, andN_TTYwith commit79901317ce80is the only LD which does that.Quoted 4 lines
Quoted 3 lines
- The requirement for line disciplines to set
receive_roomwasn't (and btw still isn't) documented anywhere, so it's unlikely anything actively relied on it.
Quoted 5 lines
Quoted 2 lines
Nevertheless, that is the requirement, and what every other in-tree line discipline does.
Your word for it. Still I don't understand the curious resistance to documenting it. If it is the requirement, why keep it secret?
Nothing sinister here :)
Feel free to submit documentation patches.
Regards,
Peter Hurley
Re: [PATCH 3.12 16/33] isdn/gigaset: reset tty->receive_room when attaching ser_gigaset
Am 21.09.2015 um 15:13 schrieb Peter Hurley:
Quoted 3 lines · Tilman Schmidt
Quoted 2 lines
Am 17.09.2015 um 20:13 schrieb Peter Hurley:
Quoted 1 line · Tilman Schmidt
[...]
Quoted 7 lines
Quoted 7 lines
Quoted 6 lines
Quoted 3 lines
- The requirement for line disciplines to set
receive_roomwasn't (and btw still isn't) documented anywhere, so it's unlikely anything actively relied on it.Nevertheless, that is the requirement, and what every other in-tree line discipline does.
Quoted 6 lines
Quoted 2 lines
Your word for it. Still I don't understand the curious resistance to documenting it. If it is the requirement, why keep it secret?
Nothing sinister here :)
Feel free to submit documentation patches.
I already did. For some unknown reason nobody wants to merge them.
Signature · 4 lines
Tilman Schmidt E-Mail: [email protected] Bonn, Germany Diese Nachricht besteht zu 100% aus wiederverwerteten Bits. Ungeöffnet mindestens haltbar bis: (siehe Rückseite)
Re: [PATCH 3.12 16/33] isdn/gigaset: reset tty->receive_room when attaching ser_gigaset
In reply to Peter HurleyAm 21.09.2015 um 15:13 schrieb Peter Hurley:
Quoted 2 lines · Tilman Schmidt
Quoted 1 line
Am 17.09.2015 um 20:13 schrieb Peter Hurley:
[...]
Quoted 9 lines
Quoted 6 lines
Quoted 1 line
So for example, if you manually set
N_PPP(as if by user error)User error wouldn't suffice, as the LD would get reset to
N_TTYwhen the serial device is closed. You would have to write a program that deliberately switched the LD first toN_PPPand then toN_GIGASET_M101without closing the device in between.???
Quoted 8 lines
The tool you authored will do it from the command line
$ ldattach PPP /dev/ttyS1 $ ldattach GIGASET_M101 /dev/ttyS1Note that nothing here closes the serial device 'in between', and the tty core has switched directly from PPP to
GIGASET_M101.n_tty->receive_roomis now 64K.
Indeed it does. I stand corrected. The possibility of running ldattach a second time without terminating the first instance didn't occur to me.
Quoted 2 lines
Please add switching from line disciplines other than
N_TTYto your regression testing.
I don't do regression tests for the driver anymore since I stepped down as a maintainer, so that would be up to the present maintainer of ser_gigaset. But I see no reason for that. As I already explained, N_TTY is the only problematic case.
Quoted 6 lines
Quoted 3 lines
Quoted 1 line
and then set this line discipline,
tty->receive_roomwill be 64K, not 4K.That wouldn't affect the operation of
ser_gigaset,I've explained this before to you, but here it is again:
Quoted 7 lines
tty->receive_roomannounces the maximum amt of data the line discipline can accept from tty core with each call to itsreceive_buf()method (for line disciplines that don't provide flow control).If the line discipline sets ->receive_room to 64K but can only handle 8K (as in the case of
GIGASET_M101), then data loss should be the expected result.
If you'd care to look at the actual code you'd notice that it truly won't make any difference. The receive_buf() method of ser_gigaset is prepared to drop data and log an error when its receive buffer overflows, no matter how big the block of data passed from tty core is. The only difference a smaller ->receive_room value might possibly make is to distribute the overflowing data to more receive_buf() calls.
(Note that the Gigaset M101 device operates at 115200 bits/sec max. so it takes at least 700 msecs to transmit 8k bytes. If we ever get into a situation where tty core actually accumulates more than that amount of data before forwarding it to GIGASET_M101 then we have a more serious problem anyway.)
Again, I won't oppose applying this patch to stable releases before 3.10. I just don't see the need, so it would be up to you to advocate such a request.
Signature · 4 lines
Tilman Schmidt E-Mail: [email protected] Bonn, Germany Diese Nachricht besteht zu 100% aus wiederverwerteten Bits. Ungeöffnet mindestens haltbar bis: (siehe Rückseite)
Re: [PATCH 3.12 16/33] isdn/gigaset: reset tty->receive_room when attaching ser_gigaset
In reply to Tilman SchmidtQuoted 12 lines · Tilman Schmidt
Am 21.09.2015 um 15:13 schrieb Peter Hurley:
Quoted 3 lines · Tilman Schmidt
Quoted 2 lines
Am 17.09.2015 um 20:13 schrieb Peter Hurley:
Quoted 1 line · Tilman Schmidt
[...]
Quoted 7 lines
Quoted 7 lines
Quoted 6 lines
Quoted 3 lines
- The requirement for line disciplines to set
receive_roomwasn't (and btw still isn't) documented anywhere, so it's unlikely anything actively relied on it.Nevertheless, that is the requirement, and what every other in-tree line discipline does.
Quoted 8 lines
Quoted 6 lines
Quoted 2 lines
Your word for it. Still I don't understand the curious resistance to documenting it. If it is the requirement, why keep it secret?
Nothing sinister here :)
Feel free to submit documentation patches.
I already did. For some unknown reason nobody wants to merge them.
I vaguely recall that. A quick search reminded me there were unaddressed comments wrt that patch: https://lkml.org/lkml/2015/7/14/608
Regards,
Peter Hurley
Re: [PATCH 3.12 16/33] isdn/gigaset: reset tty->receive_room when attaching ser_gigaset
Am 21.09.2015 um 18:54 schrieb Peter Hurley:
Quoted 16 lines · Tilman Schmidt
Quoted 15 lines
Am 21.09.2015 um 15:13 schrieb Peter Hurley:
Quoted 3 lines · Tilman Schmidt
Quoted 2 lines
Am 17.09.2015 um 20:13 schrieb Peter Hurley:
Quoted 1 line · Tilman Schmidt
[...]
Quoted 10 lines
Quoted 9 lines
Quoted 6 lines
Quoted 3 lines
- The requirement for line disciplines to set
receive_roomwasn't (and btw still isn't) documented anywhere, so it's unlikely anything actively relied on it.Nevertheless, that is the requirement, and what every other in-tree line discipline does.
Your word for it. Still I don't understand the curious resistance to documenting it. If it is the requirement, why keep it secret?
Quoted 8 lines
Quoted 5 lines
Quoted 3 lines
Nothing sinister here :)
Feel free to submit documentation patches.
I already did. For some unknown reason nobody wants to merge them.
I vaguely recall that. A quick search reminded me there were unaddressed comments wrt that patch: https://lkml.org/lkml/2015/7/14/608
Ah, so that's the blocking condition? How can I address that comment in order to unblock that patch?
Signature · 4 lines
Tilman Schmidt E-Mail: [email protected] Bonn, Germany Diese Nachricht besteht zu 100% aus wiederverwerteten Bits. Ungeöffnet mindestens haltbar bis: (siehe Rückseite)
Re: [PATCH 3.12 16/33] isdn/gigaset: reset tty->receive_room when attaching ser_gigaset
In reply to Tilman SchmidtQuoted 8 lines · Tilman Schmidt
Am 21.09.2015 um 15:13 schrieb Peter Hurley:
Quoted 7 lines
???
The tool you authored will do it from the command line
$ ldattach PPP /dev/ttyS1 $ ldattach GIGASET_M101 /dev/ttyS1
Quoted 6 lines
Quoted 3 lines
Note that nothing here closes the serial device 'in between', and the tty core has switched directly from PPP to
GIGASET_M101.n_tty->receive_roomis now 64K.Indeed it does. I stand corrected. The possibility of running ldattach a second time without terminating the first instance didn't occur to me.
Naive question: when would running ldattach a second time make sense?
Quoted 7 lines
Quoted 2 lines
Please add switching from line disciplines other than
N_TTYto your regression testing.I don't do regression tests for the driver anymore since I stepped down as a maintainer, so that would be up to the present maintainer of
ser_gigaset. But I see no reason for that. As I already explained,N_TTYis the only problematic case.
I'm the current maintainer. (By virtue of having a stash of gigaset hardware _and_ an ISDN capable line _and_ an ISP that still answers when I do ISDN dial up. That ISP also does ADSL, which I actually use on a day to day basis.) I try to test whether ISDN still has a pulse, every RC, but I can't promise much.
(Off topic: does anyone actually _care_ about ISDN? It seems the other two ISDN maintainers all have effectively left mainline. Now I'm happy to (lightly!) test mainline ISDN support as long as that's feasible, but I'd _really_ like to hear from people using ISDN on machines running currently supported kernels. Are there any left?)
Quoted 3 lines
Again, I won't oppose applying this patch to stable releases before 3.10. I just don't see the need, so it would be up to you to advocate such a request.
That would be fixing a theoretical problem, as apparently no one managed to hit this problem before v3.10 _in practice_, so that's not something I see a need for either.
Thanks,
Paul Bolle
Re: [PATCH 3.12 16/33] isdn/gigaset: reset tty->receive_room when attaching ser_gigaset
Paul,
Am 06.10.2015 um 23:00 schrieb Paul Bolle:
Quoted 9 lines · Tilman Schmidt
Quoted 8 lines
Am 21.09.2015 um 15:13 schrieb Peter Hurley:
Quoted 7 lines
???
The tool you authored will do it from the command line
$ ldattach PPP /dev/ttyS1 $ ldattach GIGASET_M101 /dev/ttyS1
Quoted 8 lines
Quoted 6 lines
Quoted 3 lines
Note that nothing here closes the serial device 'in between', and the tty core has switched directly from PPP to
GIGASET_M101.n_tty->receive_roomis now 64K.Indeed it does. I stand corrected. The possibility of running ldattach a second time without terminating the first instance didn't occur to me.
Naive question: when would running ldattach a second time make sense?
Peter's argument wasn't about making sense, but about operator error. While it doesn't make any sense indeed to run two instances of ldattach in parallel on one and the same serial port, it is entirely conceivable that someone might do so inadvertently, by not being aware that one is running already.
Best Regards,
Tilman
Signature · 4 lines
Tilman Schmidt E-Mail: [email protected] Bonn, Germany Diese Nachricht besteht zu 100% aus wiederverwerteten Bits. Ungeöffnet mindestens haltbar bis: (siehe Rückseite)
Re: [PATCH 3.12 16/33] isdn/gigaset: reset tty->receive_room when attaching ser_gigaset
[Dropped stable from Cc:. I can't see how this is still relevant for that list.]
Hi Tilman,
Quoted 5 lines · Tilman Schmidt
While it doesn't make any sense indeed to run two instances of ldattach
in parallel on one and the same serial port, it is entirely conceivable that someone might do so inadvertently, by not being aware that one is running already.
I'm wandering off topic a bit, but doesn't that imply that ldattach should bail out with an error if someone tries to do that?
Paul Bolle
Re: [PATCH 3.12 16/33] isdn/gigaset: reset tty->receive_room when attaching ser_gigaset
Hi Paul,
Am 19.10.2015 um 11:09 schrieb Paul Bolle:
Quoted 9 lines · Tilman Schmidt
Quoted 5 lines
While it doesn't make any sense indeed to run two instances of ldattach
in parallel on one and the same serial port, it is entirely conceivable that someone might do so inadvertently, by not being aware that one is running already.I'm wandering off topic a bit, but doesn't that imply that ldattach should bail out with an error if someone tries to do that?
I'm of two minds about this. On the pro side, it might prevent some surprises. But on the other hand, nothing actually breaks if someone does, and I'm not 100% sure there are really no legitimate scenarios. Add to this that I don't have a clear idea how to actually implement such a bailout, and that I'm really short of time. So I'm reluctant to tackle this topic.
Perhaps the best way forward would be someone (not me) submitting a patch to ldattach, thereby triggering a discussion of the pros and cons that would ideally include all (or most) ldattach users and consider all line disciplines it is used with.
Signature · 4 lines
Tilman Schmidt E-Mail: [email protected] Bonn, Germany Diese Nachricht besteht zu 100% aus wiederverwerteten Bits. Ungeöffnet mindestens haltbar bis: (siehe Rückseite)