| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
Draining the health workqueue will ignore future health works including
the one that report hardware failure and thus we can't enter error state
Instead cancel the recovery flow and make sure only recovery flow won't
be scheduled.
Fixes: 5e44fca50470 ('net/mlx5: Only cancel recovery work when cleaning up device')
Signed-off-by: Mohamad Haj Yahia <mohamad@mellanox.com>
Signed-off-by: Moshe Shemesh <moshe@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
|
|
When wait for firmware init fails, previous code would mistakenly
return success and cause inconsistency in the driver state.
Fixes: 6c780a0267b8 ("net/mlx5: Wait for FW readiness before initializing command interface")
Signed-off-by: Gal Pressman <galp@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
|
|
We have to reset the sk->sk_rx_dst when we disconnect a TCP
connection, because otherwise when we re-connect it this
dst reference is simply overridden in tcp_finish_connect().
This fixes a dst leak which leads to a loopback dev refcnt
leak. It is a long-standing bug, Kevin reported a very similar
(if not same) bug before. Thanks to Andrei for providing such
a reliable reproducer which greatly narrows down the problem.
Fixes: 41063e9dd119 ("ipv4: Early TCP socket demux.")
Reported-by: Andrei Vagin <avagin@gmail.com>
Reported-by: Kevin Xu <kaiwen.xu@hulu.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
In __ip6_datagram_connect(), reset sk->sk_v6_daddr and inet->dport if
error occurs.
In udp_v6_early_demux(), check for sk_state to make sure it is in
TCP_ESTABLISHED state.
Together, it makes sure unconnected UDP socket won't be considered as a
valid candidate for early demux.
v3: add TCP_ESTABLISHED state check in udp_v6_early_demux()
v2: fix compilation error
Fixes: 5425077d73e0 ("net: ipv6: Add early demux handler for UDP unicast")
Signed-off-by: Wei Wang <weiwan@google.com>
Acked-by: Maciej Żenczykowski <maze@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
When mc configuration changes bnx2x_config_mcast() can return 0 for
success, negative for failure and positive for benign reason preventing
its immediate work, e.g., when the command awaits the completion of
a previously sent command.
When removing all configured macs on a 578xx adapter, if a positive
value would be returned driver would errneously log it as an error.
Fixes: c7b7b483ccc9 ("bnx2x: Don't flush multicast MACs")
Signed-off-by: Yuval Mintz <Yuval.Mintz@cavium.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
To handle netpoll properly, the driver must only handle TX packets
during NAPI. Handling RX events cause warnings and errors in
netpoll mode. The ndo_poll_controller() method should call
napi_schedule() directly so that a NAPI weight of zero will be used
during netpoll mode.
The bnxt_en driver supports 2 ring modes: combined, and separate rx/tx.
In separate rx/tx mode, the ndo_poll_controller() method will only
process the tx rings. In combined mode, the rx and tx completion
entries are mixed in the completion ring and we need to drop the rx
entries and recycle the rx buffers.
Add a function bnxt_force_rx_discard() to handle this in netpoll mode
when we see rx entries in combined ring mode.
Reported-by: Calvin Owens <calvinowens@fb.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
When we get a TPA_END completion to handle a completed LRO packet, it
is possible that hardware would indicate errors. The current code is
not checking for the error condition. Define the proper error bits and
the macro to check for this error and abort properly.
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
The function, skb_complete_tx_timestamp(), used to allow passing in a
NULL pointer for the time stamps, but that was changed in commit
62bccb8cdb69051b95a55ab0c489e3cab261c8ef ("net-timestamp: Make the
clone operation stand-alone from phy timestamping"), and the existing
call sites, all of which are in the dp83640 driver, were fixed up.
Even though the kernel-doc was subsequently updated in commit
7a76a021cd5a292be875fbc616daf03eab1e6996 ("net-timestamp: Update
skb_complete_tx_timestamp comment"), still a bug fix from Manfred
Rudigier came into the driver using the old semantics. Probably
Manfred derived that patch from an older kernel version.
This fix should be applied to the stable trees as well.
Fixes: 81e8f2e930fe ("net: dp83640: Fix tx timestamp overflow handling.")
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
The memory allocation size is controlled by user-space,
if it is too large just fail silently and return NULL,
not to mention there is a fallback allocation later.
Reported-by: Andrey Konovalov <andreyknvl@google.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Tested-by: Andrey Konovalov <andreyknvl@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Our customer encountered stuck NFS writes for blocks starting at specific
offsets w.r.t. page boundary caused by networking stack sending packets via
UFO enabled device with wrong checksum. The problem can be reproduced by
composing a long UDP datagram from multiple parts using MSG_MORE flag:
sendto(sd, buff, 1000, MSG_MORE, ...);
sendto(sd, buff, 1000, MSG_MORE, ...);
sendto(sd, buff, 3000, 0, ...);
Assume this packet is to be routed via a device with MTU 1500 and
NETIF_F_UFO enabled. When second sendto() gets into __ip_append_data(),
this condition is tested (among others) to decide whether to call
ip_ufo_append_data():
((length + fragheaderlen) > mtu) || (skb && skb_is_gso(skb))
At the moment, we already have skb with 1028 bytes of data which is not
marked for GSO so that the test is false (fragheaderlen is usually 20).
Thus we append second 1000 bytes to this skb without invoking UFO. Third
sendto(), however, has sufficient length to trigger the UFO path so that we
end up with non-UFO skb followed by a UFO one. Later on, udp_send_skb()
uses udp_csum() to calculate the checksum but that assumes all fragments
have correct checksum in skb->csum which is not true for UFO fragments.
When checking against MTU, we need to add skb->len to length of new segment
if we already have a partially filled skb and fragheaderlen only if there
isn't one.
In the IPv6 case, skb can only be null if this is the first segment so that
we have to use headersize (length of the first IPv6 header) rather than
fragheaderlen (length of IPv6 header of further fragments) for skb == NULL.
Fixes: e89e9cf539a2 ("[IPv4/IPv6]: UFO Scatter-gather approach")
Fixes: e4c5e13aa45c ("ipv6: Should use consistent conditional judgement for
ip6 fragment between __ip6_append_data and ip6_finish_output")
Signed-off-by: Michal Kubecek <mkubecek@suse.cz>
Acked-by: Vlad Yasevich <vyasevic@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
The 8000 series adapters uses catch-all filters for encapsulated traffic
to support filtering VXLAN, NVGRE and GENEVE traffic.
This new filter functionality requires a longer MCDI command.
This patch increases the size of buffers on stack that were missed, which
fixes a kernel panic from the stack protector.
Fixes: 9b41080125176 ("sfc: insert catch-all filters for encapsulated traffic")
Signed-off-by: Martin Habets <mhabets@solarflare.com>
Acked-by: Edward Cree <ecree@solarflare.com>
Acked-by: Bert Kenward bkenward@solarflare.com
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
This structure member is hidden behind CONFIG_SYSFS, and we
get a build error when that is disabled:
drivers/net/hyperv/netvsc_drv.c: In function 'netvsc_set_channels':
drivers/net/hyperv/netvsc_drv.c:754:49: error: 'struct net_device' has no member named 'num_rx_queues'; did you mean 'num_tx_queues'?
drivers/net/hyperv/netvsc_drv.c: In function 'netvsc_set_rxfh':
drivers/net/hyperv/netvsc_drv.c:1181:25: error: 'struct net_device' has no member named 'num_rx_queues'; did you mean 'num_tx_queues'?
As the value is only set once to the argument of alloc_netdev_mq(),
we can compare against that constant directly.
Fixes: ff4a44199012 ("netvsc: allow get/set of RSS indirection table")
Fixes: 2b01888d1b45 ("netvsc: allow more flexible setting of number of channels")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Haiyang Zhang <haiyangz@microsoft.com>
Signed-off-by: Stephen Hemminger <sthemmin@microsoft.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
The per netns loopback_dev->ip6_ptr is unregistered and set to
NULL when its mtu is set to smaller than IPV6_MIN_MTU, this
leads to that we could set rt->rt6i_idev NULL after a
rt6_uncached_list_flush_dev() and then crash after another
call.
In this case we should just bring its inet6_dev down, rather
than unregistering it, at least prior to commit 176c39af29bc
("netns: fix addrconf_ifdown kernel panic") we always
override the case for loopback.
Thanks a lot to Andrey for finding a reliable reproducer.
Fixes: 176c39af29bc ("netns: fix addrconf_ifdown kernel panic")
Reported-by: Andrey Konovalov <andreyknvl@google.com>
Cc: Andrey Konovalov <andreyknvl@google.com>
Cc: Daniel Lezcano <dlezcano@fr.ibm.com>
Cc: David Ahern <dsahern@gmail.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Acked-by: David Ahern <dsahern@gmail.com>
Tested-by: Andrey Konovalov <andreyknvl@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Passthru macvlans directly change the mac address of the lower
level device. That's OK, but after the macvlan is deleted,
the lower device is left with changed address and one needs to
reboot to bring back the origina HW addresses.
This scenario is actually quite common with passthru macvtap devices.
This patch attempts to solve this, by storing the mac address
of the lower device in macvlan_port structure and keeping track of
it through the changes.
After this patch, any changes to the lower device mac address
done trough the macvlan device, will be reverted back. Any
changs done directly to the lower device mac address will be kept.
Signed-off-by: Vladislav Yasevich <vyasevic@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Convert the port passthru boolean into flags with accesor functions.
Signed-off-by: Vladislav Yasevich <vyasevic@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
When a lower device of the passthru macvlan changes it's address,
passthru macvlan is supposed to change it's own address as well.
However, that doesn't happen correctly because the check in
macvlan_addr_busy() will catch the fact that the lower level
(port) mac address is the same as the address we are trying to
assign to the macvlan, and return an error. As a reasult,
the address of the passthru macvlan device is never changed.
The same thing happens when the user attempts to change the
mac address of the passthru macvlan.
The simple solution appers to be to not check against
the lower device in case of passthru macvlan device, since
the 2 addresses are _supposed_ to be the same.
Signed-off-by: Vladislav Yasevich <vyasevic@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
The user currently gets an EBUSY error when attempting to set
the mac address on a macvlan device to the same value.
This should really be a no-op as nothing changes. Catch
the condition and return early.
Signed-off-by: Vladislav Yasevich <vyasevic@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Add a flag to indicate if a queue is rate-limited. Test the flag in
NAPI poll handler and avoid rescheduling the queue if true, otherwise
we risk locking up the host. The rescheduling will be done in the
timer callback function.
Reported-by: Jean-Louis Dupond <jean-louis@dupond.be>
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
Tested-by: Jean-Louis Dupond <jean-louis@dupond.be>
Reviewed-by: Paul Durrant <paul.durrant@citrix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
There are number of problems with configuration peer
network device in absence of IFLA_VETH_PEER attributes
where attributes for main network device shared with
peer.
First it is not feasible to configure both network
devices with same MAC address since this makes
communication in such configuration problematic.
This case can be reproduced with following sequence:
# ip link add address 02:11:22:33:44:55 type veth
# ip li sh
...
26: veth0@veth1: <BROADCAST,MULTICAST,M-DOWN> mtu 1500 qdisc \
noop state DOWN mode DEFAULT qlen 1000
link/ether 00:11:22:33:44:55 brd ff:ff:ff:ff:ff:ff
27: veth1@veth0: <BROADCAST,MULTICAST,M-DOWN> mtu 1500 qdisc \
noop state DOWN mode DEFAULT qlen 1000
link/ether 00:11:22:33:44:55 brd ff:ff:ff:ff:ff:ff
Second it is not possible to register both main and
peer network devices with same name, that happens
when name for main interface is given with IFLA_IFNAME
and same attribute reused for peer.
This case can be reproduced with following sequence:
# ip link add dev veth1a type veth
RTNETLINK answers: File exists
To fix both of the cases check if corresponding netlink
attributes are taken from peer_tb when valid or
name based on rtnl ops kind and random address is used.
Signed-off-by: Serhey Popovych <serhe.popovych@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
cpsw driver tries to get macid for am43xx SoCs using the compatible
ti,am4372. But not all variants of am43x uses this complatible like
epos evm uses ti,am438x. So use a generic compatible ti,am43 to get
macid for all am43 based platforms.
Reviewed-by: Dave Gerlach <d-gerlach@ti.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
In commit 242d3a49a2a1 ("ipv6: reorder ip6_route_dev_notifier after ipv6_dev_notf")
I assumed NETDEV_REGISTER and NETDEV_UNREGISTER are paired,
unfortunately, as reported by jeffy, netdev_wait_allrefs()
could rebroadcast NETDEV_UNREGISTER event until all refs are
gone.
We have to add an additional check to avoid this corner case.
For netdev_wait_allrefs() dev->reg_state is NETREG_UNREGISTERED,
for dev_change_net_namespace(), dev->reg_state is
NETREG_REGISTERED. So check for dev->reg_state != NETREG_UNREGISTERED.
Fixes: 242d3a49a2a1 ("ipv6: reorder ip6_route_dev_notifier after ipv6_dev_notf")
Reported-by: jeffy <jeffy.chen@rock-chips.com>
Cc: David Ahern <dsahern@gmail.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Acked-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
The commit ("net/phy: micrel: Add workaround for bad autoneg") fixes an
autoneg failure case by resetting the hardware. This turns off
intterupts. Things will work themselves out if the phy polls, as it will
figure out it's state during a poll. However if the phy uses only
intterupts, the phy will stall, since interrupts are off. This patch
fixes the issue by calling config_intr after resetting the phy.
Fixes: d2fd719bcb0e ("net/phy: micrel: Add workaround for bad autoneg ")
Signed-off-by: Zach Brown <zach.brown@ni.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
IP6CB(skb)->nhoff is the offset of the nexthdr field in an IPv6
header, unless there are extension headers present, in which case
nhoff points to the nexthdr field of the last extension header.
In non-GRO code path, nhoff is set by ipv6_rcv before any XFRM code
is executed. Conversely, in GRO code path (when esp6_offload is loaded),
nhoff is not set. The following functions fail to read the correct value
and eventually the packet is dropped:
xfrm6_transport_finish
xfrm6_tunnel_input
xfrm6_rcv_tnl
Set nhoff to the proper offset of nexthdr in esp6_gro_receive.
Fixes: 7785bba299a8 ("esp: Add a software GRO codepath")
Signed-off-by: Yossi Kuperman <yossiku@mellanox.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
|
|
IPv6 payload length indicates the size of the payload, including any
extension headers.
In xfrm6_transport_finish, ipv6_hdr(skb)->payload_len is set to the
payload size only, regardless of the presence of any extension headers.
After ESP GRO transport mode decapsulation, ipv6_rcv trims the packet
according to the wrong payload_len, thus corrupting the packet.
Set payload_len to account for extension headers as well.
Fixes: 7785bba299a8 ("esp: Add a software GRO codepath")
Signed-off-by: Yossi Kuperman <yossiku@mellanox.com>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
|
|
Fix expand_upwards() on architectures with an upward-growing stack (parisc,
metag and partly IA-64) to allow the stack to reliably grow exactly up to
the address space limit given by TASK_SIZE.
Signed-off-by: Helge Deller <deller@gmx.de>
Acked-by: Hugh Dickins <hughd@google.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
Trinity gets kernel BUG at mm/mmap.c:1963! in about 3 minutes of
mmap testing. That's the VM_BUG_ON(gap_end < gap_start) at the
end of unmapped_area_topdown(). Linus points out how MAP_FIXED
(which does not have to respect our stack guard gap intentions)
could result in gap_end below gap_start there. Fix that, and
the similar case in its alternative, unmapped_area().
Cc: stable@vger.kernel.org
Fixes: 1be7107fbe18 ("mm: larger stack guard gap, between vmas")
Reported-by: Dave Jones <davej@codemonkey.org.uk>
Debugged-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Hugh Dickins <hughd@google.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
Andrey reported a lockdep warning on non-initialized
spinlock:
INFO: trying to register non-static key.
the code is fine but needs lockdep annotation.
turning off the locking correctness validator.
CPU: 1 PID: 4099 Comm: a.out Not tainted 4.12.0-rc6+ #9
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
Call Trace:
__dump_stack lib/dump_stack.c:16
dump_stack+0x292/0x395 lib/dump_stack.c:52
register_lock_class+0x717/0x1aa0 kernel/locking/lockdep.c:755
? 0xffffffffa0000000
__lock_acquire+0x269/0x3690 kernel/locking/lockdep.c:3255
lock_acquire+0x22d/0x560 kernel/locking/lockdep.c:3855
__raw_spin_lock_bh ./include/linux/spinlock_api_smp.h:135
_raw_spin_lock_bh+0x36/0x50 kernel/locking/spinlock.c:175
spin_lock_bh ./include/linux/spinlock.h:304
ip_mc_clear_src+0x27/0x1e0 net/ipv4/igmp.c:2076
igmpv3_clear_delrec+0xee/0x4f0 net/ipv4/igmp.c:1194
ip_mc_destroy_dev+0x4e/0x190 net/ipv4/igmp.c:1736
We miss a spin_lock_init() in igmpv3_add_delrec(), probably
because previously we never use it on this code path. Since
we already unlink it from the global mc_tomb list, it is
probably safe not to acquire this spinlock here. It does not
harm to have it although, to avoid conditional locking.
Fixes: c38b7d327aaf ("igmp: acquire pmc lock for ip_mc_clear_src()")
Reported-by: Andrey Konovalov <andreyknvl@google.com>
Signed-off-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
When having the skb pointer in the first descriptor, stmmac_tx_clean
can get called at a moment where the IP has only cleared the own bit
of the first descriptor, thus freeing the skb, even though there can
be several descriptors whose buffers point into the same skb.
By simply moving the skb pointer from the first descriptor to the last
descriptor, a skb will get freed only when the IP has cleared the
own bit of all the descriptors that are using that skb.
Signed-off-by: Niklas Cassel <niklas.cassel@axis.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Somehow two copies of the line 'up_write(&vf->efx->filter_sem);' got into
efx_ef10_sriov_set_vf_vlan(). This would put the mutex in a bad state and
cause all subsequent down attempts to hang.
Fixes: 671b53eec2ed ("sfc: Ensure down_write(&filter_sem) and up_write() are matched before calling efx_net_open()")
Signed-off-by: Edward Cree <ecree@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Network interface groups support added while ago, however
there is no IFLA_GROUP attribute description in policy
and netlink message size calculations until now.
Add IFLA_GROUP attribute to the policy.
Fixes: cbda10fa97d7 ("net_device: add support for network device groups")
Signed-off-by: Serhey Popovych <serhe.popovych@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
While commit 73ba57bfae4a ("ipv6: fix backtracking for throw routes")
does good job on error propagation to the fib_rules_lookup()
in fib rules core framework that also corrects throw routes
handling, it does not solve route reference leakage problem
happened when we return -EAGAIN to the fib_rules_lookup()
and leave routing table entry referenced in arg->result.
If rule with matched throw route isn't last matched in the
list we overwrite arg->result losing reference on throw
route stored previously forever.
We also partially revert commit ab997ad40839 ("ipv6: fix the
incorrect return value of throw route") since we never return
routing table entry with dst.error == -EAGAIN when
CONFIG_IPV6_MULTIPLE_TABLES is on. Also there is no point
to check for RTF_REJECT flag since it is always set throw
route.
Fixes: 73ba57bfae4a ("ipv6: fix backtracking for throw routes")
Signed-off-by: Serhey Popovych <serhe.popovych@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
The lan911x family of devices require supplying from 3.3 V power
supplies (connected to VDD_IO, VDD_A and VREG_3.3 pins). The existing
driver however obtains only VDD_IO and VDD_A regulators in an optional
way so document this in bindings.
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Remove the use of arch_setup_dma_ops() that was not exported
and was breaking loadable module compilation.
Signed-off-by: Madalin Bucur <madalin.bucur@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Make sure dma_ops are set, to be later used by the Ethernet driver.
Signed-off-by: Madalin Bucur <madalin.bucur@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Since commit 217f69743681 ("net: busy-poll: allow preemption in
sk_busy_loop()") there is an explicit do_softirq() invocation after
local_bh_enable() has been invoked.
I don't understand why we need this because local_bh_enable() will
invoke do_softirq() once the softirq counter reached zero and we have
softirq-related work pending.
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
We should avoid marking goto rules unresolved when their
target is actually reachable after rule deletion.
Consolder following sample scenario:
# ip -4 ru sh
0: from all lookup local
32000: from all goto 32100
32100: from all lookup main
32100: from all lookup default
32766: from all lookup main
32767: from all lookup default
# ip -4 ru del pref 32100 table main
# ip -4 ru sh
0: from all lookup local
32000: from all goto 32100 [unresolved]
32100: from all lookup default
32766: from all lookup main
32767: from all lookup default
After removal of first rule with preference 32100 we
mark all goto rules as unreachable, even when rule with
same preference as removed one still present.
Check if next rule with same preference is available
and make all rules with goto action pointing to it.
Signed-off-by: Serhey Popovych <serhe.popovych@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
rcu_read_(un)lock(), list_*_rcu(), and synchronize_rcu() are used for a secure
access and manipulation of the list of patches that modify the same function.
In particular, it is the variable func_stack that is accessible from the ftrace
handler via struct ftrace_ops and klp_ops.
Of course, it synchronizes also some states of the patch on the top of the
stack, e.g. func->transition in klp_ftrace_handler.
At the same time, this mechanism guards also the manipulation of
task->patch_state. It is modified according to the state of the transition and
the state of the process.
Now, all this works well as long as RCU works well. Sadly livepatching might
get into some corner cases when this is not true. For example, RCU is not
watching when rcu_read_lock() is taken in idle threads. It is because they
might sleep and prevent reaching the grace period for too long.
There are ways how to make RCU watching even in idle threads, see
rcu_irq_enter(). But there is a small location inside RCU infrastructure when
even this does not work.
This small problematic location can be detected either before calling
rcu_irq_enter() by rcu_irq_enter_disabled() or later by rcu_is_watching().
Sadly, there is no safe way how to handle it. Once we detect that RCU was not
watching, we might see inconsistent state of the function stack and the related
variables in klp_ftrace_handler(). Then we could do a wrong decision, use an
incompatible implementation of the function and break the consistency of the
system. We could warn but we could not avoid the damage.
Fortunately, ftrace has similar problems and they seem to be solved well there.
It uses a heavy weight implementation of some RCU operations. In particular, it
replaces:
+ rcu_read_lock() with preempt_disable_notrace()
+ rcu_read_unlock() with preempt_enable_notrace()
+ synchronize_rcu() with schedule_on_each_cpu(sync_work)
My understanding is that this is RCU implementation from a stone age. It meets
the core RCU requirements but it is rather ineffective. Especially, it does not
allow to batch or speed up the synchronize calls.
On the other hand, it is very trivial. It allows to safely trace and/or
livepatch even the RCU core infrastructure. And the effectiveness is a not a
big issue because using ftrace or livepatches on productive systems is a rare
operation. The safety is much more important than a negligible extra load.
Note that the alternative implementation follows the RCU principles. Therefore,
we could and actually must use list_*_rcu() variants when manipulating the
func_stack. These functions allow to access the pointers in the right
order and with the right barriers. But they do not use any other
information that would be set only by rcu_read_lock().
Also note that there are actually two problems solved in ftrace:
First, it cares about the consistency of RCU read sections. It is being solved
the way as described and used in this patch.
Second, ftrace needs to make sure that nobody is inside the dynamic trampoline
when it is being freed. For this, it also calls synchronize_rcu_tasks() in
preemptive kernel in ftrace_shutdown().
Livepatch has similar problem but it is solved by ftrace for free.
klp_ftrace_handler() is a good guy and never sleeps. In addition, it is
registered with FTRACE_OPS_FL_DYNAMIC. It causes that
unregister_ftrace_function() calls:
* schedule_on_each_cpu(ftrace_sync) - always
* synchronize_rcu_tasks() - in preemptive kernel
The effect is that nobody is neither inside the dynamic trampoline nor inside
the ftrace handler after unregister_ftrace_function() returns.
[jkosina@suse.cz: reformat changelog, fix comment]
Signed-off-by: Petr Mladek <pmladek@suse.com>
Acked-by: Josh Poimboeuf <jpoimboe@redhat.com>
Acked-by: Miroslav Benes <mbenes@suse.cz>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
|
|
Setting these bits causes libinput to fail to initialize the device;
setting BTN_TOUCH and BTN_TOOL_FINGER causes it to treat the mouse as a
touchpad, and it then refuses to continue when it discovers ABS_X is not
set.
This breaks all known Wayland compositors, as well as Xorg when the
libinput driver is being used.
This reverts commit f4b65b9563216b3e01a5cc844c3ba68901d9b195.
Signed-off-by: Daniel Stone <daniels@collabora.com>
Cc: Che-Liang Chiou <clchiou@chromium.org>
Cc: Thierry Escande <thierry.escande@collabora.com>
Cc: Jiri Kosina <jkosina@suse.cz>
Cc: Benjamin Tissoires <benjamin.tissoires@redhat.com>
Acked-by: Benjamin Tissoires <benjamin.tissoires@redhat.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
|
|
Now before dumping a sock in sctp_diag, it only holds the sock while
the ep may be already destroyed. It can cause a use-after-free panic
when accessing ep->asocs.
This patch is to set sctp_sk(sk)->ep NULL in sctp_endpoint_destroy,
and check if this ep is already destroyed before dumping this ep.
Suggested-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Acked-by: Neil Horman <nhorman@tuxdrver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Do not sleep in ntb_async_tx_submit, which could deadlock.
This reverts commit "8c874cc140d667f84ae4642bb5b5e0d6396d2ca4"
Fixes: 8c874cc140d6 ("NTB: Address out of DMA descriptor issue with NTB")
Reported-by: Jia-Ju Bai <baijiaju1990@163.com>
Signed-off-by: Allen Hubbe <Allen.Hubbe@dell.com>
Acked-by: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Jon Mason <jdmason@kudzu.us>
|
|
Fixing doorbell register length to 32bits per spec. On Skylake NTB, the
doorbell registers are 32bit write only registers. The source for the
doorbell is a 64bit register that shows the interrupt bits.
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Fixes: 783dfa6cc41b ("ntb: Adding Skylake Xeon NTB support")
Acked-by: Allen Hubbe <Allen.Hubbe@dell.com>
Signed-off-by: Jon Mason <jdmason@kudzu.us>
|
|
A divide by zero error occurs if qp_count is less than mw_count because
num_qps_mw is calculated to be zero. The calculation appears to be
incorrect.
The requirement is for num_qps_mw to be set to qp_count / mw_count
with any remainder divided among the earlier mws.
For example, if mw_count is 5 and qp_count is 12 then mws 0 and 1
will have 3 qps per window and mws 2 through 4 will have 2 qps per window.
Thus, when mw_num < qp_count % mw_count, num_qps_mw is 1 higher
than when mw_num >= qp_count.
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Fixes: e26a5843f7f5 ("NTB: Split ntb_hw_intel and ntb_transport drivers")
Acked-by: Allen Hubbe <Allen.Hubbe@dell.com>
Signed-off-by: Jon Mason <jdmason@kudzu.us>
|
|
In cases where there are more mw's than spads/2-2, the mw count gets
reduced to match the limitation. ntb_transport also tries to ensure that
there are fewer qps than mws but uses the full mw count instead of
the reduced one. When this happens, the math in
'ntb_transport_setup_qp_mw' will get confused and result in a kernel
paging request bug.
This patch fixes the bug by reducing qp_count to the reduced mw count
instead of the full mw count.
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Fixes: e26a5843f7f5 ("NTB: Split ntb_hw_intel and ntb_transport drivers")
Acked-by: Allen Hubbe <Allen.Hubbe@dell.com>
Signed-off-by: Jon Mason <jdmason@kudzu.us>
|
|
The code mistakenly prints the local perf results for the remote test
so the script reports identical results for both directions. Fix this
by ensuring we print the remote result.
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Fixes: a9c59ef77458 ("ntb_test: Add a selftest script for the NTB subsystem")
Acked-by: Allen Hubbe <Allen.Hubbe@dell.com>
Signed-off-by: Jon Mason <jdmason@kudzu.us>
|
|
The order parameters are powers of 2; adjust the usage information
to use correct mathematical representations.
Signed-off-by: Gary R Hook <gary.hook@amd.com>
Fixes: 8a7b6a778a85 ("ntb: ntb perf tool")
Acked-by: Dave Jiang <dave.jiang@intel.com>
Signed-off-by: Jon Mason <jdmason@kudzu.us>
|
|
This patch fixes the phy loopback self_test failed issue. when
Marvell Phy Module is loaded, it will powerdown fiber when doing
phy loopback self test, which cause phy loopback self_test fail.
Signed-off-by: Lin Yun Sheng <linyunsheng@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
The register_vlan_device would invoke free_netdev directly, when
register_vlan_dev failed. It would trigger the BUG_ON in free_netdev
if the dev was already registered. In this case, the netdev would be
freed in netdev_run_todo later.
So add one condition check now. Only when dev is not registered, then
free it directly.
The following is the part coredump when netdev_upper_dev_link failed
in register_vlan_dev. I removed the lines which are too long.
[ 411.237457] ------------[ cut here ]------------
[ 411.237458] kernel BUG at net/core/dev.c:7998!
[ 411.237484] invalid opcode: 0000 [#1] SMP
[ 411.237705] [last unloaded: 8021q]
[ 411.237718] CPU: 1 PID: 12845 Comm: vconfig Tainted: G E 4.12.0-rc5+ #6
[ 411.237737] Hardware name: VMware, Inc. VMware Virtual Platform/440BX Desktop Reference Platform, BIOS 6.00 07/02/2015
[ 411.237764] task: ffff9cbeb6685580 task.stack: ffffa7d2807d8000
[ 411.237782] RIP: 0010:free_netdev+0x116/0x120
[ 411.237794] RSP: 0018:ffffa7d2807dbdb0 EFLAGS: 00010297
[ 411.237808] RAX: 0000000000000002 RBX: ffff9cbeb6ba8fd8 RCX: 0000000000001878
[ 411.237826] RDX: 0000000000000001 RSI: 0000000000000282 RDI: 0000000000000000
[ 411.237844] RBP: ffffa7d2807dbdc8 R08: 0002986100029841 R09: 0002982100029801
[ 411.237861] R10: 0004000100029980 R11: 0004000100029980 R12: ffff9cbeb6ba9000
[ 411.238761] R13: ffff9cbeb6ba9060 R14: ffff9cbe60f1a000 R15: ffff9cbeb6ba9000
[ 411.239518] FS: 00007fb690d81700(0000) GS:ffff9cbebb640000(0000) knlGS:0000000000000000
[ 411.239949] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 411.240454] CR2: 00007f7115624000 CR3: 0000000077cdf000 CR4: 00000000003406e0
[ 411.240936] Call Trace:
[ 411.241462] vlan_ioctl_handler+0x3f1/0x400 [8021q]
[ 411.241910] sock_ioctl+0x18b/0x2c0
[ 411.242394] do_vfs_ioctl+0xa1/0x5d0
[ 411.242853] ? sock_alloc_file+0xa6/0x130
[ 411.243465] SyS_ioctl+0x79/0x90
[ 411.243900] entry_SYSCALL_64_fastpath+0x1e/0xa9
[ 411.244425] RIP: 0033:0x7fb69089a357
[ 411.244863] RSP: 002b:00007ffcd04e0fc8 EFLAGS: 00000202 ORIG_RAX: 0000000000000010
[ 411.245445] RAX: ffffffffffffffda RBX: 00007ffcd04e2884 RCX: 00007fb69089a357
[ 411.245903] RDX: 00007ffcd04e0fd0 RSI: 0000000000008983 RDI: 0000000000000003
[ 411.246527] RBP: 00007ffcd04e0fd0 R08: 0000000000000000 R09: 1999999999999999
[ 411.246976] R10: 000000000000053f R11: 0000000000000202 R12: 0000000000000004
[ 411.247414] R13: 00007ffcd04e1128 R14: 00007ffcd04e2888 R15: 0000000000000001
[ 411.249129] RIP: free_netdev+0x116/0x120 RSP: ffffa7d2807dbdb0
Signed-off-by: Gao Feng <gfree.wind@vip.163.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
During the module initialisation there is a possible race
(basically race between uld and lld) where neither the uld
nor lld notifies the uP about where to route the ctrl queue
completions. LLD skips notifying uP as the rdma queues were
not created by then (will leave it to ULD to notify the uP).
As the ULD comes up, it also skips notifying the uP as the
flag FULL_INIT_DONE is not set yet (ULD assumes that the
interface is not up yet).
Consequently, this race between uld and lld leaves uP
unnotified about where to send the ctrl queue completions
to, leading to iwarp RI_RES WR failure.
Here is the race:
CPU 0 CPU1
- allocates nic rx queus
- t4_sge_alloc_ctrl_txq()
(if rdma rsp queues exists,
tell uP to route ctrl queue
compl to rdma rspq)
- acquires the mutex_lock
- allocates rdma response queues
- if FULL_INIT_DONE set,
tell uP to route ctrl queue compl
to rdma rspq
- relinquishes mutex_lock
- acquires the mutex_lock
- enable_rx()
- set FULL_INIT_DONE
- relinquishes mutex_lock
This patch fixes the above issue.
Fixes: e7519f9926f1('cxgb4: avoid enabling napi twice to the same queue')
Signed-off-by: Raju Rangoju <rajur@chelsio.com>
Acked-by: Steve Wise <swise@opengridcomputing.com>
CC: Stable <stable@vger.kernel.org> # 4.9+
Signed-off-by: Ganesh Goudar <ganeshgr@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
|
|
Stack guard page is a useful feature to reduce a risk of stack smashing
into a different mapping. We have been using a single page gap which
is sufficient to prevent having stack adjacent to a different mapping.
But this seems to be insufficient in the light of the stack usage in
userspace. E.g. glibc uses as large as 64kB alloca() in many commonly
used functions. Others use constructs liks gid_t buffer[NGROUPS_MAX]
which is 256kB or stack strings with MAX_ARG_STRLEN.
This will become especially dangerous for suid binaries and the default
no limit for the stack size limit because those applications can be
tricked to consume a large portion of the stack and a single glibc call
could jump over the guard page. These attacks are not theoretical,
unfortunatelly.
Make those attacks less probable by increasing the stack guard gap
to 1MB (on systems with 4k pages; but make it depend on the page size
because systems with larger base pages might cap stack allocations in
the PAGE_SIZE units) which should cover larger alloca() and VLA stack
allocations. It is obviously not a full fix because the problem is
somehow inherent, but it should reduce attack space a lot.
One could argue that the gap size should be configurable from userspace,
but that can be done later when somebody finds that the new 1MB is wrong
for some special case applications. For now, add a kernel command line
option (stack_guard_gap) to specify the stack gap size (in page units).
Implementation wise, first delete all the old code for stack guard page:
because although we could get away with accounting one extra page in a
stack vma, accounting a larger gap can break userspace - case in point,
a program run with "ulimit -S -v 20000" failed when the 1MB gap was
counted for RLIMIT_AS; similar problems could come with RLIMIT_MLOCK
and strict non-overcommit mode.
Instead of keeping gap inside the stack vma, maintain the stack guard
gap as a gap between vmas: using vm_start_gap() in place of vm_start
(or vm_end_gap() in place of vm_end if VM_GROWSUP) in just those few
places which need to respect the gap - mainly arch_get_unmapped_area(),
and and the vma tree's subtree_gap support for that.
Original-patch-by: Oleg Nesterov <oleg@redhat.com>
Original-patch-by: Michal Hocko <mhocko@suse.com>
Signed-off-by: Hugh Dickins <hughd@google.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Tested-by: Helge Deller <deller@gmx.de> # parisc
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
Mohamad Haj Yahia
Gal Pressman
WANG Cong
Wei Wang
Mintz, Yuval
Michael Chan
Richard Cochran
Michal Kubeček
Martin Habets
Arnd Bergmann
Vlad Yasevich
Wei Liu
Serhey Popovych
Lokesh Vutla
Zach Brown
Yossi Kuperman
Helge Deller
Hugh Dickins
Niklas Cassel
Edward Cree
Krzysztof Kozlowski
Madalin Bucur
Sebastian Siewior
Petr Mladek
Daniel Stone
Xin Long
Allen Hubbe
Dave Jiang
Logan Gunthorpe
Gary R Hook
Lin Yun Sheng
Gao Feng
Raju Rangoju
Linus Torvalds