- 22 Mar, 2018 24 commits
-
-
Stephen Hemminger authored
The receive processing may continue to happen while the internal network device state is in RCU grace period. The internal RNDIS structure is associated with the internal netvsc_device structure; both have the same RCU lifetime. Defer freeing all associated parts until after grace period. Fixes: 0cf73780 ("hv_netvsc: netvsc_teardown_gpadl() split") Signed-off-by:
Stephen Hemminger <sthemmin@microsoft.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Stephen Hemminger authored
This makes sure that no CPU is still process packets when the channel is closed. Fixes: 76bb5db5 ("netvsc: fix use after free on module removal") Signed-off-by:
-
David Ahern authored
For multipath routes the ONLINK flag can be specified per nexthop in rtnh_flags or globally in rtm_flags. Update ip6_route_multipath_add to consider the ONLINK setting coming from rtnh_flags. Each loop over nexthops the config for the sibling route is initialized to the global config and then per nexthop settings overlayed. The flag is 'or'ed into fib6_config to handle the ONLINK flag coming from either rtm_flags or rtnh_flags. Fixes: fc1e64e1 ("net/ipv6: Add support for onlink flag") Signed-off-by:
David Ahern <dsahern@gmail.com> Signed-off-by:
-
Guillaume Nault authored
We already detect situations where a PPP channel sends packets back to its upper PPP device. While this is enough to avoid deadlocking on xmit locks, this doesn't prevent packets from looping between the channel and the unit. The problem is that ppp_start_xmit() enqueues packets in ppp->file.xq before checking for xmit recursion. Therefore, __ppp_xmit_process() might dequeue a packet from ppp->file.xq and send it on the channel which, in turn, loops it back on the unit. Then ppp_start_xmit() queues the packet back to ppp->file.xq and __ppp_xmit_process() picks it up and sends it again through the channel. Therefore, the packet will loop between __ppp_xmit_process() and ppp_start_xmit() until some other part of the xmit path drops it. For L2TP, we rapidly fill the skb's headroom and pppol2tp_xmit() drops the packet after a few iterations. But PPTP reallocates the headroom if necessary, letting the loop run and exhaust the machine resources (as reported in https://bugzilla.kernel.org/show_bug.cgi?id=199109). Fix this by letting __ppp_xmit_process() enqueue the skb to ppp->file.xq, so that we can check for recursion before adding it to the queue. Now ppp_xmit_process() can drop the packet when recursion is detected. __ppp_channel_push() is a bit special. It calls __ppp_xmit_process() without having any actual packet to send. This is used by ppp_output_wakeup() to re-enable transmission on the parent unit (for implementations like ppp_async.c, where the .start_xmit() function might not consume the skb, leaving it in ppp->xmit_pending and disabling transmission). Therefore, __ppp_xmit_process() needs to handle the case where skb is NULL, dequeuing as many packets as possible from ppp->file.xq. Reported-by:
xu heng <xuheng333@zoho.com> Fixes: 55454a56 ("ppp: avoid dealock on recursive xmit") Signed-off-by:
Guillaume Nault <g.nault@alphalink.fr> Signed-off-by:
-
David Lebrun authored
When using seg6 in encap mode, we call ipv6_dev_get_saddr() to set the source address of the outer IPv6 header, in case none was specified. Using skb->dev can lead to BUG() when it is in an inconsistent state. This patch uses the net_device attached to the skb's dst instead. [940807.667429] BUG: unable to handle kernel NULL pointer dereference at 000000000000047c [940807.762427] IP: ipv6_dev_get_saddr+0x8b/0x1d0 [940807.815725] PGD 0 P4D 0 [940807.847173] Oops: 0000 [#1] SMP PTI [940807.890073] Modules linked in: [940807.927765] CPU: 6 PID: 0 Comm: swapper/6 Tainted: G W 4.16.0-rc1-seg6bpf+ #2 [940808.028988] Hardware name: HP ProLiant DL120 G6/ProLiant DL120 G6, BIOS O26 09/06/2010 [940808.128128] RIP: 0010:ipv6_dev_get_saddr+0x8b/0x1d0 [940808.187667] RSP: 0018:ffff88043fd836b0 EFLAGS: 00010206 [940808.251366] RAX: 0000000000000005 RBX: ffff88042cb1c860 RCX: 00000000000000fe [940808.338025] RDX: 00000000000002c0 RSI: ffff88042cb1c860 RDI: 0000000000004500 [940808.424683] RBP: ffff88043fd83740 R08: 0000000000000000 R09: ffffffffffffffff [940808.511342] R10: 0000000000000040 R11: 0000000000000000 R12: ffff88042cb1c850 [940808.598012] R13: ffffffff8208e380 R14: ffff88042ac8da00 R15: 0000000000000002 [940808.684675] FS: 0000000000000000(0000) GS:ffff88043fd80000(0000) knlGS:0000000000000000 [940808.783036] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [940808.852975] CR2: 000000000000047c CR3: 00000004255fe000 CR4: 00000000000006e0 [940808.939634] Call Trace: [940808.970041] <IRQ> [940808.995250] ? ip6t_do_table+0x265/0x640 [940809.043341] seg6_do_srh_encap+0x28f/0x300 [940809.093516] ? seg6_do_srh+0x1a0/0x210 [940809.139528] seg6_do_srh+0x1a0/0x210 [940809.183462] seg6_output+0x28/0x1e0 [940809.226358] lwtunnel_output+0x3f/0x70 [940809.272370] ip6_xmit+0x2b8/0x530 [940809.313185] ? ac6_proc_exit+0x20/0x20 [940809.359197] inet6_csk_xmit+0x7d/0xc0 [940809.404173] tcp_transmit_skb+0x548/0x9a0 [940809.453304] __tcp_retransmit_skb+0x1a8/0x7a0 [940809.506603] ? ip6_default_advmss+0x40/0x40 [940809.557824] ? tcp_current_mss+0x24/0x90 [940809.605925] tcp_retransmit_skb+0xd/0x80 [940809.654016] tcp_xmit_retransmit_queue.part.17+0xf9/0x210 [940809.719797] tcp_ack+0xa47/0x1110 [940809.760612] tcp_rcv_established+0x13c/0x570 [940809.812865] tcp_v6_do_rcv+0x151/0x3d0 [940809.858879] tcp_v6_rcv+0xa5c/0xb10 [940809.901770] ? seg6_output+0xdd/0x1e0 [940809.946745] ip6_input_finish+0xbb/0x460 [940809.994837] ip6_input+0x74/0x80 [940810.034612] ? ip6_rcv_finish+0xb0/0xb0 [940810.081663] ipv6_rcv+0x31c/0x4c0 ... Fixes: 6c8702c6 ("ipv6: sr: add support for SRH encapsulation and injection with lwtunnels") Reported-by:
Tom Herbert <tom@quantonium.net> Signed-off-by:
David Lebrun <dlebrun@google.com> Signed-off-by:
-
David Lebrun authored
The seg6_build_state() function is called with RCU read lock held, so we cannot use GFP_KERNEL. This patch uses GFP_ATOMIC instead. [ 92.770271] ============================= [ 92.770628] WARNING: suspicious RCU usage [ 92.770921] 4.16.0-rc4+ #12 Not tainted [ 92.771277] ----------------------------- [ 92.771585] ./include/linux/rcupdate.h:302 Illegal context switch in RCU read-side critical section! [ 92.772279] [ 92.772279] other info that might help us debug this: [ 92.772279] [ 92.773067] [ 92.773067] rcu_scheduler_active = 2, debug_locks = 1 [ 92.773514] 2 locks held by ip/2413: [ 92.773765] #0: (rtnl_mutex){+.+.}, at: [<00000000e5461720>] rtnetlink_rcv_msg+0x441/0x4d0 [ 92.774377] #1: (rcu_read_lock){....}, at: [<00000000df4f161e>] lwtunnel_build_state+0x59/0x210 [ 92.775065] [ 92.775065] stack backtrace: [ 92.775371] CPU: 0 PID: 2413 Comm: ip Not tainted 4.16.0-rc4+ #12 [ 92.775791] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.10.2-1.fc27 04/01/2014 [ 92.776608] Call Trace: [ 92.776852] dump_stack+0x7d/0xbc [ 92.777130] __schedule+0x133/0xf00 [ 92.777393] ? unwind_get_return_address_ptr+0x50/0x50 [ 92.777783] ? __sched_text_start+0x8/0x8 [ 92.778073] ? rcu_is_watching+0x19/0x30 [ 92.778383] ? kernel_text_address+0x49/0x60 [ 92.778800] ? __kernel_text_address+0x9/0x30 [ 92.779241] ? unwind_get_return_address+0x29/0x40 [ 92.779727] ? pcpu_alloc+0x102/0x8f0 [ 92.780101] _cond_resched+0x23/0x50 [ 92.780459] __mutex_lock+0xbd/0xad0 [ 92.780818] ? pcpu_alloc+0x102/0x8f0 [ 92.781194] ? seg6_build_state+0x11d/0x240 [ 92.781611] ? save_stack+0x9b/0xb0 [ 92.781965] ? __ww_mutex_wakeup_for_backoff+0xf0/0xf0 [ 92.782480] ? seg6_build_state+0x11d/0x240 [ 92.782925] ? lwtunnel_build_state+0x1bd/0x210 [ 92.783393] ? ip6_route_info_create+0x687/0x1640 [ 92.783846] ? ip6_route_add+0x74/0x110 [ 92.784236] ? inet6_rtm_newroute+0x8a/0xd0 Fixes: 6c8702c6 ("ipv6: sr: add support for SRH encapsulation and injection with lwtunnels") Signed-off-by: -
David S. Miller authored
Igor Russkikh says: ==================== Aquantia atlantic hot fixes 03-2018 This is a set of atlantic driver hot fixes for various areas: Some issues with hardware reset covered, Fixed napi_poll flood happening on some traffic conditions, Allow system to change MAC address on live device, Add pci shutdown handler. patch v2: - reverse christmas tree - remove driver private parameter, replacing it with define. ==================== Signed-off-by: -
Igor Russkikh authored
Signed-off-by:
Igor Russkikh <igor.russkikh@aquantia.com> Signed-off-by:
-
Igor Russkikh authored
We should close link and all NIC operations during shutdown. On some systems graceful reboot never closes NIC interface on its own, but only indicates pci device shutdown. Without explicit handler, NIC rx rings continued to transfer DMA data into prepared buffers while CPU rebooted already. That caused memory corruptions on soft reboot. Signed-off-by:
-
Igor Russkikh authored
There is nothing prevents us from changing MAC on the running interface. Allow this with ndev priv flag. Signed-off-by:
-
Igor Russkikh authored
We should report to napi full budget only when we have more job to do. Before this fix, on any tx queue cleanup we forced napi to do poll again. Thats a waste of cpu resources and caused storming with napi polls when there was at least one tx on each interrupt. With this fix we report full budget only when there is more job on TX to do. Or, as before, when rx budget was fully consumed. Signed-off-by:
-
Igor Russkikh authored
B1 hardware changes behavior of mailbox interface, it has busy bit always raised. Data ready condition should be detected by increment of address register. Old code has empty `for` loop, and that caused cpu overloads on B1 hardware. aq_nic_service_timer_cb consumed ~100ms because of that. Signed-off-by:
-
Igor Russkikh authored
FW 1.5.58 and below needs a fixed delay even after 0x18 register is filled. Otherwise, setting MPI_INIT state too fast causes traffic hang. Signed-off-by:
-
Igor Russkikh authored
Under some circumstances (notably using thunderbolt interface) SPI on chip reset may be in active transaction. Here we forcibly cleanup SPI to prevent possible hangups. Signed-off-by:
-
David S. Miller authored
Julian Wiedmann says: ==================== s390/qeth: fixes 2018-03-20 Please apply one final set of qeth patches for 4.16. All of these fix long-standing bugs, so please queue them up for -stable as well. ==================== Signed-off-by: -
Julian Wiedmann authored
When the IRQ handler determines that one of the cmd IO channels has failed and schedules recovery, block any further cmd requests from being submitted. The request would inevitably stall, and prevent the recovery from making progress until the request times out. This sort of error was observed after Live Guest Relocation, where the pending IO on the READ channel intentionally gets terminated to kick-start recovery. Simultaneously the guest executed SIOCETHTOOL, triggering qeth to issue a QUERY CARD INFO command. The command then stalled in the inoperabel WRITE channel. Signed-off-by:
Julian Wiedmann <jwi@linux.vnet.ibm.com> Signed-off-by:
-
Julian Wiedmann authored
For calling ccw_device_start(), issue_next_read() needs to hold the device's ccwlock. This is satisfied for the IRQ handler path (where qeth_irq() gets called under the ccwlock), but we need explicit locking for the initial call by the MPC initialization. Signed-off-by:
-
Julian Wiedmann authored
qeth_wait_for_threads() is potentially called by multiple users, make sure to notify all of them after qeth_clear_thread_running_bit() adjusted the thread_running_mask. With no timeout, callers would otherwise stall. Signed-off-by:
-
Julian Wiedmann authored
On removal, a qeth card's netdevice is currently not properly freed because the call chain looks as follows: qeth_core_remove_device(card) lx_remove_device(card) unregister_netdev(card->dev) card->dev = NULL !!! qeth_core_free_card(card) if (card->dev) !!! free_netdev(card->dev) Fix it by free'ing the netdev straight after unregistering. This also fixes the sysfs-driven layer switch case (qeth_dev_layer2_store()), where the need to free the current netdevice was not considered at all. Note that free_netdev() takes care of the netif_napi_del() for us too. Fixes: 4a71df50 ("qeth: new qeth device driver") Signed-off-by:
Ursula Braun <ubraun@linux.vnet.ibm.com> Signed-off-by:
-
David S. Miller authored
Kevin Hao says: ==================== net: phy: Add general dummy stubs for MMD register access v2: As suggested by Andrew: - Add general dummy stubs - Also use that for the micrel phy This patch series fix the Ethernet broken on the mpc8315erdb board introduced by commit b6b5e8a6 ("gianfar: Disable EEE autoneg by default"). ==================== Signed-off-by:
-
Kevin Hao authored
The new general dummy stubs for MMD register access were introduced. Use that for the codes reuse. Signed-off-by:
Kevin Hao <haokexin@gmail.com> Signed-off-by:
-
Kevin Hao authored
The Ethernet on mpc8315erdb is broken since commit b6b5e8a6 ("gianfar: Disable EEE autoneg by default"). The reason is that even though the rtl8211b doesn't support the MMD extended registers access, it does return some random values if we trying to access the MMD register via indirect method. This makes it seem that the EEE is supported by this phy device. And the subsequent writing to the MMD registers does cause the phy malfunction. So use the dummy stubs for the MMD register access to fix this issue. Fixes: b6b5e8a6 ("gianfar: Disable EEE autoneg by default") Signed-off-by:
-
Kevin Hao authored
For some phy devices, even though they don't support the MMD extended register access, it does have some side effect if we are trying to read/write the MMD registers via indirect method. So introduce general dummy stubs for MMD register access which these devices can use to avoid such side effect. Fixes: b6b5e8a6 ("gianfar: Disable EEE autoneg by default") Signed-off-by:
-
git://git.open-mesh.org/linux-mergeDavid S. Miller authored
Simon Wunderlich says: ==================== Here are some batman-adv bugfixes: - fix possible IPv6 packet loss when multicast extension is used, by Linus Luessing - fix SKB handling issues for TTVN and DAT, by Matthias Schiffer (two patches) - fix include for eventpoll, by Sven Eckelmann - fix skb checksum for ttvn reroutes, by Sven Eckelmann ==================== Signed-off-by:
-
- 21 Mar, 2018 12 commits
-
-
David S. Miller authored
Davide Caratti says: ==================== fix idr leak in actions This series fixes situations where a temporary failure to install a TC action results in the permanent impossibility to reuse the configured value of 'index'. Thanks to Cong Wang for the initial review. v2: fix build error in act_ipt.c, reported by kbuild test robot ==================== Acked-by:
Cong Wang <xiyou.wangcong@gmail.com> Signed-off-by:
-
Davide Caratti authored
tcf_skbmod_init() can fail after the idr has been successfully reserved. When this happens, every subsequent attempt to configure skbmod rules using the same idr value will systematically fail with -ENOSPC, unless the first attempt was done using the 'replace' keyword: # tc action add action skbmod swap mac index 100 RTNETLINK answers: Cannot allocate memory We have an error talking to the kernel # tc action add action skbmod swap mac index 100 RTNETLINK answers: No space left on device We have an error talking to the kernel # tc action add action skbmod swap mac index 100 RTNETLINK answers: No space left on device We have an error talking to the kernel ... Fix this in tcf_skbmod_init(), ensuring that tcf_idr_release() is called on the error path when the idr has been reserved, but not yet inserted. Also, don't test 'ovr' in the error path, to avoid a 'replace' failure implicitly become a 'delete' that leaks refcount in act_skbmod module: # rmmod act_skbmod; modprobe act_skbmod # tc action add action skbmod swap mac index 100 # tc action add action skbmod swap mac continue index 100 RTNETLINK answers: File exists We have an error talking to the kernel # tc action replace action skbmod swap mac continue index 100 RTNETLINK answers: Cannot allocate memory We have an error talking to the kernel # tc action list action skbmod # # rmmod act_skbmod rmmod: ERROR: Module act_skbmod is in use Fixes: 65a206c0 ("net/sched: Change act_api and act_xxx modules to use IDR") Acked-by:
Jamal Hadi Salim <jhs@mojatatu.com> Signed-off-by:
Davide Caratti <dcaratti@redhat.com> Signed-off-by:
-
Davide Caratti authored
tcf_vlan_init() can fail after the idr has been successfully reserved. When this happens, every subsequent attempt to configure vlan rules using the same idr value will systematically fail with -ENOSPC, unless the first attempt was done using the 'replace' keyword. # tc action add action vlan pop index 100 RTNETLINK answers: Cannot allocate memory We have an error talking to the kernel # tc action add action vlan pop index 100 RTNETLINK answers: No space left on device We have an error talking to the kernel # tc action add action vlan pop index 100 RTNETLINK answers: No space left on device We have an error talking to the kernel ... Fix this in tcf_vlan_init(), ensuring that tcf_idr_release() is called on the error path when the idr has been reserved, but not yet inserted. Also, don't test 'ovr' in the error path, to avoid a 'replace' failure implicitly become a 'delete' that leaks refcount in act_vlan module: # rmmod act_vlan; modprobe act_vlan # tc action add action vlan push id 5 index 100 # tc action replace action vlan push id 7 index 100 RTNETLINK answers: Cannot allocate memory We have an error talking to the kernel # tc action list action vlan # # rmmod act_vlan rmmod: ERROR: Module act_vlan is in use Fixes: 4c5b9d96 ("act_vlan: VLAN action rewrite to use RCU lock/unlock and update") Fixes: 65a206c0 ("net/sched: Change act_api and act_xxx modules to use IDR") Acked-by:
-
Davide Caratti authored
__tcf_ipt_init() can fail after the idr has been successfully reserved. When this happens, subsequent attempts to configure xt/ipt rules using the same idr value systematically fail with -ENOSPC: # tc action add action xt -j LOG --log-prefix test1 index 100 tablename: mangle hook: NF_IP_POST_ROUTING target: LOG level warning prefix "test1" index 100 RTNETLINK answers: Cannot allocate memory We have an error talking to the kernel Command "(null)" is unknown, try "tc actions help". # tc action add action xt -j LOG --log-prefix test1 index 100 tablename: mangle hook: NF_IP_POST_ROUTING target: LOG level warning prefix "test1" index 100 RTNETLINK answers: No space left on device We have an error talking to the kernel Command "(null)" is unknown, try "tc actions help". # tc action add action xt -j LOG --log-prefix test1 index 100 tablename: mangle hook: NF_IP_POST_ROUTING target: LOG level warning prefix "test1" index 100 RTNETLINK answers: No space left on device We have an error talking to the kernel ... Fix this in the error path of __tcf_ipt_init(), calling tcf_idr_release() in place of tcf_idr_cleanup(). Since tcf_ipt_release() can now be called when tcfi_t is NULL, we also need to protect calls to ipt_destroy_target() to avoid NULL pointer dereference. Fixes: 65a206c0 ("net/sched: Change act_api and act_xxx modules to use IDR") Acked-by: -
Davide Caratti authored
tcf_pedit_init() can fail to allocate 'keys' after the idr has been successfully reserved. When this happens, subsequent attempts to configure a pedit rule using the same idr value systematically fail with -ENOSPC: # tc action add action pedit munge ip ttl set 63 index 100 RTNETLINK answers: Cannot allocate memory We have an error talking to the kernel # tc action add action pedit munge ip ttl set 63 index 100 RTNETLINK answers: No space left on device We have an error talking to the kernel # tc action add action pedit munge ip ttl set 63 index 100 RTNETLINK answers: No space left on device We have an error talking to the kernel ... Fix this in the error path of tcf_act_pedit_init(), calling tcf_idr_release() in place of tcf_idr_cleanup(). Fixes: 65a206c0 ("net/sched: Change act_api and act_xxx modules to use IDR") Acked-by:
-
Davide Caratti authored
tcf_act_police_init() can fail after the idr has been successfully reserved (e.g., qdisc_get_rtab() may return NULL). When this happens, subsequent attempts to configure a police rule using the same idr value systematiclly fail with -ENOSPC: # tc action add action police rate 1000 burst 1000 drop index 100 RTNETLINK answers: Cannot allocate memory We have an error talking to the kernel # tc action add action police rate 1000 burst 1000 drop index 100 RTNETLINK answers: No space left on device We have an error talking to the kernel # tc action add action police rate 1000 burst 1000 drop index 100 RTNETLINK answers: No space left on device ... Fix this in the error path of tcf_act_police_init(), calling tcf_idr_release() in place of tcf_idr_cleanup(). Fixes: 65a206c0 ("net/sched: Change act_api and act_xxx modules to use IDR") Acked-by:
-
Davide Caratti authored
if the kernel fails to duplicate 'sdata', creation of a new action fails with -ENOMEM. However, subsequent attempts to install the same action using the same value of 'index' systematically fail with -ENOSPC, and that value of 'index' will no more be usable by act_simple, until rmmod / insmod of act_simple.ko is done: # tc actions add action simple sdata hello index 100 # tc actions list action simple action order 0: Simple <hello> index 100 ref 1 bind 0 # tc actions flush action simple # tc actions add action simple sdata hello index 100 RTNETLINK answers: Cannot allocate memory We have an error talking to the kernel # tc actions flush action simple # tc actions add action simple sdata hello index 100 RTNETLINK answers: No space left on device We have an error talking to the kernel # tc actions add action simple sdata hello index 100 RTNETLINK answers: No space left on device We have an error talking to the kernel ... Fix this in the error path of tcf_simp_init(), calling tcf_idr_release() in place of tcf_idr_cleanup(). Fixes: 65a206c0 ("net/sched: Change act_api and act_xxx modules to use IDR") Suggested-by: -
Davide Caratti authored
when the following command sequence is entered # tc action add action bpf bytecode '4,40 0 0 12,31 0 1 2048,6 0 0 262144,6 0 0 0' index 100 RTNETLINK answers: Invalid argument We have an error talking to the kernel # tc action add action bpf bytecode '4,40 0 0 12,21 0 1 2048,6 0 0 262144,6 0 0 0' index 100 RTNETLINK answers: No space left on device We have an error talking to the kernel act_bpf correctly refuses to install the first TC rule, because 31 is not a valid instruction. However, it refuses to install the second TC rule, even if the BPF code is correct. Furthermore, it's no more possible to install any other rule having the same value of 'index' until act_bpf module is unloaded/inserted again. After the idr has been reserved, call tcf_idr_release() instead of tcf_idr_cleanup(), to fix this issue. Fixes: 65a206c0 ("net/sched: Change act_api and act_xxx modules to use IDR") Acked-by:
-
Colin Ian King authored
Trivial fix to spelling mistakes in DP_ERR error message text and comments Signed-off-by:
Colin Ian King <colin.king@canonical.com> Signed-off-by:
-
Colin Ian King authored
Trivial fix to spelling mistake in BNX2X_ERR error message text Signed-off-by:
Sudarsana Kalluru <Sudarsana.Kalluru@cavium.com> Signed-off-by:
-
git://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpfDavid S. Miller authored
Daniel Borkmann says: ==================== pull-request: bpf 2018-03-21 The following pull-request contains BPF updates for your *net* tree. The main changes are: 1) Follow-up fix to the fault injection framework to prevent jump optimization on the kprobe by installing a dummy post-handler, from Masami. 2) Drop bpf_perf_prog_read_value helper from tracepoint type programs which was mistakenly added there and would otherwise crash due to wrong input context, from Yonghong. 3) Fix a crash in BPF fs when compiled with clang. Code appears to be fine just that clang tries to overly aggressive optimize in non C conform ways, therefore fix the kernel's Makefile to generally prevent such issues, from Daniel. 4) Skip unnecessary capability checks in bpf syscall, which is otherwise triggering unnecessary security hooks on capability checking and causing false alarms on unprivileged processes trying to access CAP_SYS_ADMIN restricted infra, from Chenbo. 5) Fix the test_bpf.ko module when CONFIG_BPF_JIT_ALWAYS_ON is set with regards to a test case that is really just supposed to fail on x8_64 JIT but not others, from Thadeu. ==================== Signed-off-by: -
Daniel Borkmann authored
Prasad reported that he has seen crashes in BPF subsystem with netd on Android with arm64 in the form of (note, the taint is unrelated): [ 4134.721483] Unable to handle kernel paging request at virtual address 800000001 [ 4134.820925] Mem abort info: [ 4134.901283] Exception class = DABT (current EL), IL = 32 bits [ 4135.016736] SET = 0, FnV = 0 [ 4135.119820] EA = 0, S1PTW = 0 [ 4135.201431] Data abort info: [ 4135.301388] ISV = 0, ISS = 0x00000021 [ 4135.359599] CM = 0, WnR = 0 [ 4135.470873] user pgtable: 4k pages, 39-bit VAs, pgd = ffffffe39b946000 [ 4135.499757] [0000000800000001] *pgd=0000000000000000, *pud=0000000000000000 [ 4135.660725] Internal error: Oops: 96000021 [#1] PREEMPT SMP [ 4135.674610] Modules linked in: [ 4135.682883] CPU: 5 PID: 1260 Comm: netd Tainted: G S W 4.14.19+ #1 [ 4135.716188] task: ffffffe39f4aa380 task.stack: ffffff801d4e0000 [ 4135.731599] PC is at bpf_prog_add+0x20/0x68 [ 4135.741746] LR is at bpf_prog_inc+0x20/0x2c [ 4135.751788] pc : [<ffffff94ab7ad584>] lr : [<ffffff94ab7ad638>] pstate: 60400145 [ 4135.769062] sp : ffffff801d4e3ce0 [...] [ 4136.258315] Process netd (pid: 1260, stack limit = 0xffffff801d4e0000) [ 4136.273746] Call trace: [...] [ 4136.442494] 3ca0: ffffff94ab7ad584 0000000060400145 ffffffe3a01bf8f8 0000000000000006 [ 4136.460936] 3cc0: 0000008000000000 ffffff94ab844204 ffffff801d4e3cf0 ffffff94ab7ad584 [ 4136.479241] [<ffffff94ab7ad584>] bpf_prog_add+0x20/0x68 [ 4136.491767] [<ffffff94ab7ad638>] bpf_prog_inc+0x20/0x2c [ 4136.504536] [<ffffff94ab7b5d08>] bpf_obj_get_user+0x204/0x22c [ 4136.518746] [<ffffff94ab7ade68>] SyS_bpf+0x5a8/0x1a88 Android's netd was basically pinning the uid cookie BPF map in BPF fs (/sys/fs/bpf/traffic_cookie_uid_map) and later on retrieving it again resulting in above panic. Issue is that the map was wrongly identified as a prog! Above kernel was compiled with clang 4.0, and it turns out that clang decided to merge the bpf_prog_iops and bpf_map_iops into a single memory location, such that the two i_ops could then not be distinguished anymore. Reason for this miscompilation is that clang has the more aggressive -fmerge-all-constants enabled by default. In fact, clang source code has a comment about it in lib/AST/ExprConstant.cpp on why it is okay to do so: Pointers with different bases cannot represent the same object. (Note that clang defaults to -fmerge-all-constants, which can lead to inconsistent results for comparisons involving the address of a constant; this generally doesn't matter in practice.) The issue never appeared with gcc however, since gcc does not enable -fmerge-all-constants by default and even *explicitly* states in it's option description that using this flag results in non-conforming behavior, quote from man gcc: Languages like C or C++ require each variable, including multiple instances of the same variable in recursive calls, to have distinct locations, so using this option results in non-conforming behavior. There are also various clang bug reports open on that matter [1], where clang developers acknowledge the non-conforming behavior, and refer to disabling it with -fno-merge-all-constants. But even if this gets fixed in clang today, there are already users out there that triggered this. Thus, fix this issue by explicitly adding -fno-merge-all-constants to the kernel's Makefile to generically disable this optimization, since potentially other places in the kernel could subtly break as well. Note, there is also a flag called -fmerge-constants (not supported by clang), which is more conservative and only applies to strings and it's enabled in gcc's -O/-O2/-O3/-Os optimization levels. In gcc's code, the two flags -fmerge-{all-,}constants share the same variable internally, so when disabling it via -fno-merge-all-constants, then we really don't merge any const data (e.g. strings), and text size increases with gcc (14,927,214 -> 14,942,646 for vmlinux.o). $ gcc -fverbose-asm -O2 foo.c -S -o foo.S -> foo.S lists -fmerge-constants under options enabled $ gcc -fverbose-asm -O2 -fno-merge-all-constants foo.c -S -o foo.S -> foo.S doesn't list -fmerge-constants under options enabled $ gcc -fverbose-asm -O2 -fno-merge-all-constants -fmerge-constants foo.c -S -o foo.S -> foo.S lists -fmerge-constants under options enabled Thus, as a workaround we need to set both -fno-merge-all-constants *and* -fmerge-constants in the Makefile in order for text size to stay as is. [1] https://bugs.llvm.org/show_bug.cgi?id=18538Reported-by:Prasad Sodagudi <psodagud@codeaurora.org> Signed-off-by:
Daniel Borkmann <daniel@iogearbox.net> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Chenbo Feng <fengc@google.com> Cc: Richard Smith <richard-llvm@metafoo.co.uk> Cc: Chandler Carruth <chandlerc@gmail.com> Cc: linux-kernel@vger.kernel.org Tested-by:
Alexei Starovoitov <ast@kernel.org> Signed-off-by:
-
- 20 Mar, 2018 4 commits
-
-
Chenbo Feng authored
The current check statement in BPF syscall will do a capability check for CAP_SYS_ADMIN before checking sysctl_unprivileged_bpf_disabled. This code path will trigger unnecessary security hooks on capability checking and cause false alarms on unprivileged process trying to get CAP_SYS_ADMIN access. This can be resolved by simply switch the order of the statement and CAP_SYS_ADMIN is not required anyway if unprivileged bpf syscall is allowed. Signed-off-by:
Chenbo Feng <fengc@google.com> Acked-by:
Lorenzo Colitti <lorenzo@google.com> Signed-off-by:
-
Yonghong Song authored
Commit 4bebdc7a ("bpf: add helper bpf_perf_prog_read_value") added helper bpf_perf_prog_read_value so that perf_event type program can read event counter and enabled/running time. This commit, however, introduced a bug which allows this helper for tracepoint type programs. This is incorrect as bpf_perf_prog_read_value needs to access perf_event through its bpf_perf_event_data_kern type context, which is not available for tracepoint type program. This patch fixed the issue by separating bpf_func_proto between tracepoint and perf_event type programs and removed bpf_perf_prog_read_value from tracepoint func prototype. Fixes: 4bebdc7a ("bpf: add helper bpf_perf_prog_read_value") Reported-by:
Yonghong Song <yhs@fb.com> Signed-off-by:
-
Thadeu Lima de Souza Cascardo authored
Function bpf_fill_maxinsns11 is designed to not be able to be JITed on x86_64. So, it fails when CONFIG_BPF_JIT_ALWAYS_ON=y, and commit 09584b40 ("bpf: fix selftests/bpf test_kmod.sh failure when CONFIG_BPF_JIT_ALWAYS_ON=y") makes sure that failure is detected on that case. However, it does not fail on other architectures, which have a different JIT compiler design. So, test_bpf has started to fail to load on those. After this fix, test_bpf loads fine on both x86_64 and ppc64el. Fixes: 09584b40 ("bpf: fix selftests/bpf test_kmod.sh failure when CONFIG_BPF_JIT_ALWAYS_ON=y") Signed-off-by:
Thadeu Lima de Souza Cascardo <cascardo@canonical.com> Reviewed-by:
-
Stefano Brivio authored
Fixes: 2f987a76 ("net: ipv6: keep sk status consistent after datagram connect failure") Signed-off-by:
Stefano Brivio <sbrivio@redhat.com> Acked-by:
Paolo Abeni <pabeni@redhat.com> Acked-by:
-
