pneigh_enqueue() tries to obtain a random delay by mod
NEIGH_VAR(p, PROXY_DELAY). However, NEIGH_VAR(p, PROXY_DELAY)
migth be zero at that point because someone could write zero
to /proc/sys/net/ipv4/neigh/[device]/proxy_delay after the
callers check it.
This patch uses prandom_u32_max() to get a random delay instead
which avoids potential division by zero.
Signed-off-by: weichenchen <weichen.chen@linux.alibaba.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
RT_TOS() only clears one of the ECN bits. Therefore, when
fib_compute_spec_dst() resorts to a fib lookup, it can return
different results depending on the value of the second ECN bit.
For example, ECT(0) and ECT(1) packets could be treated differently.
$ ip netns add ns0
$ ip netns add ns1
$ ip link add name veth01 netns ns0 type veth peer name veth10 netns ns1
$ ip -netns ns0 link set dev lo up
$ ip -netns ns1 link set dev lo up
$ ip -netns ns0 link set dev veth01 up
$ ip -netns ns1 link set dev veth10 up
$ ip -netns ns0 address add 192.0.2.10/24 dev veth01
$ ip -netns ns1 address add 192.0.2.11/24 dev veth10
$ ip -netns ns1 address add 192.0.2.21/32 dev lo
$ ip -netns ns1 route add 192.0.2.10/32 tos 4 dev veth10 src 192.0.2.21
$ ip netns exec ns1 sysctl -wq net.ipv4.icmp_echo_ignore_broadcasts=0
With TOS 4 and ECT(1), ns1 replies using source address 192.0.2.21
(ping uses -Q to set all TOS and ECN bits):
$ ip netns exec ns0 ping -c 1 -b -Q 5 192.0.2.255
[...]
64 bytes from 192.0.2.21: icmp_seq=1 ttl=64 time=0.544 ms
But with TOS 4 and ECT(0), ns1 replies using source address 192.0.2.11
because the "tos 4" route isn't matched:
$ ip netns exec ns0 ping -c 1 -b -Q 6 192.0.2.255
[...]
64 bytes from 192.0.2.11: icmp_seq=1 ttl=64 time=0.597 ms
After this patch the ECN bits don't affect the result anymore:
$ ip netns exec ns0 ping -c 1 -b -Q 6 192.0.2.255
[...]
64 bytes from 192.0.2.21: icmp_seq=1 ttl=64 time=0.591 ms
Fixes: 35ebf65e85 ("ipv4: Create and use fib_compute_spec_dst() helper.")
Signed-off-by: Guillaume Nault <gnault@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The packet coalescing interrupt threshold has separated registers
for different aggregated/cpu (sw-thread). The required value should
be loaded for every thread but not only for 1 current cpu.
Fixes: 213f428f50 ("net: mvpp2: add support for TX interrupts and RX queue distribution modes")
Signed-off-by: Stefan Chulski <stefanc@marvell.com>
Link: https://lore.kernel.org/r/1608748521-11033-1-git-send-email-stefanc@marvell.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Alex Elder says:
====================
net: ipa: fix some new build warnings
I got a super friendly message from the Intel kernel test robot that
pointed out that two patches I posted last week caused new build
warnings. I already had these problems fixed in my own tree but
the fix was not included in what I sent out last week.
====================
Link: https://lore.kernel.org/r/20201226213737.338928-1-elder@linaro.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Callers of evt_ring_command() no longer care whether the command
times out, and don't use what evt_ring_command() returns. Redefine
that function to have void return type.
Reported-by: kernel test robot <lkp@intel.com>
Fixes: 428b448ee7 ("net: ipa: use state to determine event ring command success")
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Callers of gsi_channel_command() no longer care whether the command
times out, and don't use what gsi_channel_command() returns. Redefine
that function to have void return type.
Reported-by: kernel test robot <lkp@intel.com>
Fixes: 6ffddf3b3d ("net: ipa: use state to determine channel command success")
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
TQM rings are hardware resources that require host context memory
managed by the driver. The driver supports up to 9 TQM rings and
the number of rings to use is requested by firmware during run-time.
Cap this number to the maximum supported to prevent accessing beyond
the array. Future firmware may request more than 9 TQM rings. Define
macros to remove the magic number 9 from the C code.
Fixes: ac3158cb01 ("bnxt_en: Allocate TQM ring context memory according to fw specification.")
Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Reviewed-by: Vasundhara Volam <vasundhara-v.volam@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
A recent change skips sending firmware messages to the firmware when
pci_channel_offline() is true during fatal AER error. To make this
complete, we need to move the re-initialization sequence to
bnxt_io_resume(), otherwise the firmware messages to re-initialize
will all be skipped. In any case, it is more correct to re-initialize
in bnxt_io_resume().
Also, fix the reverse x-mas tree format when defining variables
in bnxt_io_slot_reset().
Fixes: b340dc680e ("bnxt_en: Avoid sending firmware messages when AER error is detected.")
Reviewed-by: Edwin Peer <edwin.peer@broadcom.com>
Signed-off-by: Vasundhara Volam <vasundhara-v.volam@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Tony Nguyen says:
====================
Intel Wired LAN Driver Updates 2020-12-23
Commit e086ba2fcc ("e1000e: disable s0ix entry and exit flows for ME
systems") disabled S0ix flows for systems that have various incarnations of
the i219-LM ethernet controller. This was done because of some regressions
caused by an earlier commit 632fbd5eb5 ("e1000e: fix S0ix flows for
cable connected case") with i219-LM controller.
Per discussion with Intel architecture team this direction should be
changed and allow S0ix flows to be used by default. This patch series
includes directional changes for their conclusions in
https://lkml.org/lkml/2020/12/13/15.
* '1GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/net-queue:
e1000e: Export S0ix flags to ethtool
Revert "e1000e: disable s0ix entry and exit flows for ME systems"
e1000e: bump up timeout to wait when ME un-configures ULP mode
e1000e: Only run S0ix flows if shutdown succeeded
====================
Link: https://lore.kernel.org/r/20201223233625.92519-1-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Currently the tun_napi_alloc_frags() function returns -ENOMEM when the
number of iovs exceeds MAX_SKB_FRAGS + 1. However this is inappropriate,
we should use -EMSGSIZE instead of -ENOMEM.
The following distinctions are matters:
1. the caller need to drop the bad packet when -EMSGSIZE is returned,
which means meeting a persistent failure.
2. the caller can try again when -ENOMEM is returned, which means
meeting a transient failure.
Fixes: 90e33d4594 ("tun: enable napi_gro_frags() for TUN/TAP driver")
Signed-off-by: Yunjian Wang <wangyunjian@huawei.com>
Acked-by: Willem de Bruijn <willemb@google.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Link: https://lore.kernel.org/r/1608864736-24332-1-git-send-email-wangyunjian@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The CPTS driver registers PTP PHC clock when first netif is going up and
unregister it when all netif are down. Now ethtool will show:
- PTP PHC clock index 0 after boot until first netif is up;
- the last assigned PTP PHC clock index even if PTP PHC clock is not
registered any more after all netifs are down.
This patch ensures that -1 is returned by ethtool when PTP PHC clock is not
registered any more.
Fixes: 8a2c9a5ab4 ("net: ethernet: ti: cpts: rework initialization/deinitialization")
Signed-off-by: Grygorii Strashko <grygorii.strashko@ti.com>
Acked-by: Richard Cochran <richardcochran@gmail.com>
Link: https://lore.kernel.org/r/20201224162405.28032-1-grygorii.strashko@ti.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Antoine Tenart says:
====================
net-sysfs: fix race conditions in the xps code
This series fixes race conditions in the xps code, where out of bound
accesses can occur when dev->num_tc is updated, triggering oops. The
root cause is linked to locking issues. An explanation is given in each
of the commit logs.
We had a discussion on the v1 of this series about using the xps_map
mutex instead of the rtnl lock. While that seemed a better compromise,
v2 showed the added complexity wasn't best for fixes. So we decided to
go back to v1 and use the rtnl lock.
Because of this, the only differences between v1 and v3 are improvements
in the commit messages.
====================
Link: https://lore.kernel.org/r/20201223212323.3603139-1-atenart@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Accesses to dev->xps_rxqs_map (when using dev->num_tc) should be
protected by the rtnl lock, like we do for netif_set_xps_queue. I didn't
see an actual bug being triggered, but let's be safe here and take the
rtnl lock while accessing the map in sysfs.
Fixes: 8af2c06ff4 ("net-sysfs: Add interface for Rx queue(s) map per Tx queue")
Signed-off-by: Antoine Tenart <atenart@kernel.org>
Reviewed-by: Alexander Duyck <alexanderduyck@fb.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Two race conditions can be triggered when storing xps rxqs, resulting in
various oops and invalid memory accesses:
1. Calling netdev_set_num_tc while netif_set_xps_queue:
- netif_set_xps_queue uses dev->tc_num as one of the parameters to
compute the size of new_dev_maps when allocating it. dev->tc_num is
also used to access the map, and the compiler may generate code to
retrieve this field multiple times in the function.
- netdev_set_num_tc sets dev->tc_num.
If new_dev_maps is allocated using dev->tc_num and then dev->tc_num
is set to a higher value through netdev_set_num_tc, later accesses to
new_dev_maps in netif_set_xps_queue could lead to accessing memory
outside of new_dev_maps; triggering an oops.
2. Calling netif_set_xps_queue while netdev_set_num_tc is running:
2.1. netdev_set_num_tc starts by resetting the xps queues,
dev->tc_num isn't updated yet.
2.2. netif_set_xps_queue is called, setting up the map with the
*old* dev->num_tc.
2.3. netdev_set_num_tc updates dev->tc_num.
2.4. Later accesses to the map lead to out of bound accesses and
oops.
A similar issue can be found with netdev_reset_tc.
One way of triggering this is to set an iface up (for which the driver
uses netdev_set_num_tc in the open path, such as bnx2x) and writing to
xps_rxqs in a concurrent thread. With the right timing an oops is
triggered.
Both issues have the same fix: netif_set_xps_queue, netdev_set_num_tc
and netdev_reset_tc should be mutually exclusive. We do that by taking
the rtnl lock in xps_rxqs_store.
Fixes: 8af2c06ff4 ("net-sysfs: Add interface for Rx queue(s) map per Tx queue")
Signed-off-by: Antoine Tenart <atenart@kernel.org>
Reviewed-by: Alexander Duyck <alexanderduyck@fb.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Accesses to dev->xps_cpus_map (when using dev->num_tc) should be
protected by the rtnl lock, like we do for netif_set_xps_queue. I didn't
see an actual bug being triggered, but let's be safe here and take the
rtnl lock while accessing the map in sysfs.
Fixes: 184c449f91 ("net: Add support for XPS with QoS via traffic classes")
Signed-off-by: Antoine Tenart <atenart@kernel.org>
Reviewed-by: Alexander Duyck <alexanderduyck@fb.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Two race conditions can be triggered when storing xps cpus, resulting in
various oops and invalid memory accesses:
1. Calling netdev_set_num_tc while netif_set_xps_queue:
- netif_set_xps_queue uses dev->tc_num as one of the parameters to
compute the size of new_dev_maps when allocating it. dev->tc_num is
also used to access the map, and the compiler may generate code to
retrieve this field multiple times in the function.
- netdev_set_num_tc sets dev->tc_num.
If new_dev_maps is allocated using dev->tc_num and then dev->tc_num
is set to a higher value through netdev_set_num_tc, later accesses to
new_dev_maps in netif_set_xps_queue could lead to accessing memory
outside of new_dev_maps; triggering an oops.
2. Calling netif_set_xps_queue while netdev_set_num_tc is running:
2.1. netdev_set_num_tc starts by resetting the xps queues,
dev->tc_num isn't updated yet.
2.2. netif_set_xps_queue is called, setting up the map with the
*old* dev->num_tc.
2.3. netdev_set_num_tc updates dev->tc_num.
2.4. Later accesses to the map lead to out of bound accesses and
oops.
A similar issue can be found with netdev_reset_tc.
One way of triggering this is to set an iface up (for which the driver
uses netdev_set_num_tc in the open path, such as bnx2x) and writing to
xps_cpus in a concurrent thread. With the right timing an oops is
triggered.
Both issues have the same fix: netif_set_xps_queue, netdev_set_num_tc
and netdev_reset_tc should be mutually exclusive. We do that by taking
the rtnl lock in xps_cpus_store.
Fixes: 184c449f91 ("net: Add support for XPS with QoS via traffic classes")
Signed-off-by: Antoine Tenart <atenart@kernel.org>
Reviewed-by: Alexander Duyck <alexanderduyck@fb.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The cdc_ncm driver passes network connection notifications up to
usbnet_link_change(), which is the right place for any logging.
Remove the netdev_info() duplicating this from the driver itself.
This stops devices such as my "TRENDnet USB 10/100/1G/2.5G LAN"
(ID 20f4:e02b) adapter from spamming the kernel log with
cdc_ncm 2-2:2.0 enp0s2u2c2: network connection: connected
messages every 60 msec or so.
Signed-off-by: Roland Dreier <roland@kernel.org>
Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Link: https://lore.kernel.org/r/20201224032116.2453938-1-roland@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
This flag can be used by an end user to disable S0ix flows on a
buggy system or by an OEM for development purposes.
If you need this flag to be persisted across reboots, it's suggested
to use a udev rule to call adjust it until the kernel could have your
configuration in a disallow list.
Signed-off-by: Mario Limonciello <mario.limonciello@dell.com>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Tested-by: Yijun Shen <Yijun.shen@dell.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
commit e086ba2fcc ("e1000e: disable s0ix entry and exit flows for ME
systems") disabled s0ix flows for systems that have various incarnations of
the i219-LM ethernet controller. This changed caused power consumption
regressions on the following shipping Dell Comet Lake based laptops:
* Latitude 5310
* Latitude 5410
* Latitude 5410
* Latitude 5510
* Precision 3550
* Latitude 5411
* Latitude 5511
* Precision 3551
* Precision 7550
* Precision 7750
This commit was introduced because of some regressions on certain Thinkpad
laptops. This comment was potentially caused by an earlier
commit 632fbd5eb5 ("e1000e: fix S0ix flows for cable connected case").
or it was possibly caused by a system not meeting platform architectural
requirements for low power consumption. Other changes made in the driver
with extended timeouts are expected to make the driver more impervious to
platform firmware behavior.
Fixes: e086ba2fcc ("e1000e: disable s0ix entry and exit flows for ME systems")
Reviewed-by: Alexander Duyck <alexander.duyck@gmail.com>
Signed-off-by: Mario Limonciello <mario.limonciello@dell.com>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Tested-by: Yijun Shen <Yijun.shen@dell.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Per guidance from Intel ethernet architecture team, it may take
up to 1 second for unconfiguring ULP mode.
However in practice this seems to be taking up to 2 seconds on
some Lenovo machines. Detect scenarios that take more than 1 second
but less than 2.5 seconds and emit a warning on resume for those
scenarios.
Suggested-by: Aaron Ma <aaron.ma@canonical.com>
Suggested-by: Sasha Netfin <sasha.neftin@intel.com>
Suggested-by: Hans de Goede <hdegoede@redhat.com>
CC: Mark Pearson <markpearson@lenovo.com>
Fixes: f15bb6dde7 ("e1000e: Add support for S0ix")
BugLink: https://bugs.launchpad.net/bugs/1865570
Link: https://patchwork.ozlabs.org/project/intel-wired-lan/patch/20200323191639.48826-1-aaron.ma@canonical.com/
Link: https://lkml.org/lkml/2020/12/13/15
Link: https://lkml.org/lkml/2020/12/14/708
Signed-off-by: Mario Limonciello <mario.limonciello@dell.com>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Tested-by: Yijun Shen <Yijun.shen@dell.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
If the shutdown failed, the part will be thawed and running
S0ix flows will put it into an undefined state.
Reported-by: Alexander Duyck <alexander.duyck@gmail.com>
Reviewed-by: Alexander Duyck <alexander.duyck@gmail.com>
Signed-off-by: Mario Limonciello <mario.limonciello@dell.com>
Tested-by: Yijun Shen <Yijun.shen@dell.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Commit f9c6cea0b3 ("ibmvnic: Skip fatal error reset after passive init")
says "If the passive
CRQ initialization occurs before the FATAL reset task is processed,
the FATAL error reset task would try to access a CRQ message queue
that was freed, causing an oops. The problem may be most likely to
occur during DLPAR add vNIC with a non-default MTU, because the DLPAR
process will automatically issue a change MTU request.
Fix this by not processing fatal error reset if CRQ is passively
initialized after client-driven CRQ initialization fails."
The original commit skips a specific reset condition, but that does
not fix the problem it claims to fix, and misses a reset condition.
The effective fix is commit 0e435befae ("ibmvnic: fix NULL pointer
dereference in ibmvic_reset_crq") and commit a0faaa27c7 ("ibmvnic:
fix NULL pointer dereference in reset_sub_crq_queues"). With above
two fixes, there are no more crashes seen as described even without
the original commit, so I would like to revert the original commit.
Fixes: f9c6cea0b3 ("ibmvnic: Skip fatal error reset after passive init")
Signed-off-by: Lijun Pan <ljp@linux.ibm.com>
Link: https://lore.kernel.org/r/20201223204904.12677-1-ljp@linux.ibm.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
When mdiobus_register() fails, priv->mdio allocated
by mdiobus_alloc() has not been freed, which leads
to memleak.
Fixes: e7f4dc3536 ("mdio: Move allocation of interrupts into core")
Signed-off-by: Dinghao Liu <dinghao.liu@zju.edu.cn>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://lore.kernel.org/r/20201223110615.31389-1-dinghao.liu@zju.edu.cn
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
When aggregating ncsi interfaces and dedicated interfaces to bond
interfaces, the ncsi response handler will use the wrong net device to
find ncsi_dev, so that the ncsi interface will not work properly.
Here, we use the original net device to fix it.
Fixes: 138635cc27 ("net/ncsi: NCSI response packet handler")
Signed-off-by: John Wang <wangzhiqiang.bj@bytedance.com>
Link: https://lore.kernel.org/r/20201223055523.2069-1-wangzhiqiang.bj@bytedance.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
DCB uses the same handler function for both RTM_GETDCB and RTM_SETDCB
messages. dcb_doit() bounces RTM_SETDCB mesasges if the user does not have
the CAP_NET_ADMIN capability.
However, the operation to be performed is not decided from the DCB message
type, but from the DCB command. Thus DCB_CMD_*_GET commands are used for
reading DCB objects, the corresponding SET and DEL commands are used for
manipulation.
The assumption is that set-like commands will be sent via an RTM_SETDCB
message, and get-like ones via RTM_GETDCB. However, this assumption is not
enforced.
It is therefore possible to manipulate DCB objects without CAP_NET_ADMIN
capability by sending the corresponding command in an RTM_GETDCB message.
That is a bug. Fix it by validating the type of the request message against
the type used for the response.
Fixes: 2f90b8657e ("ixgbe: this patch adds support for DCB to the kernel and ixgbe driver")
Signed-off-by: Petr Machata <me@pmachata.org>
Link: https://lore.kernel.org/r/a2a9b88418f3a58ef211b718f2970128ef9e3793.1608673640.git.me@pmachata.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Alex Elder says:
====================
net: ipa: GSI interrupt handling fixes
This series implements fixes for some issues related to handling
interrupts when GSI channel and event ring commands complete.
The first issue is that the completion condition for an event ring
or channel command could occur while the associated interrupt is
disabled. This would cause the interrupt to fire when it is
subsequently enabled, even if the condition it signals had already
been handled. The fix is to clear any pending interrupt conditions
before re-enabling the interrupt.
The second and third patches change how the success of an event ring
or channel command is determined. These commands change the state
of an event ring or channel. Previously the receipt of a completion
interrupt was required to consider a command successful. Instead, a
command is successful if it changes the state of the target event
ring or channel in the way expected. This way the command can
succeed even if the completion interrupt did not arrive while it was
enabled.
====================
Link: https://lore.kernel.org/r/20201222180012.22489-1-elder@linaro.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
This patch implements the same basic fix for event rings as the
previous one does for channels.
The result of issuing an event ring control command should be that
the event ring changes state. If enabled, a completion interrupt
signals that the event ring state has changed. This interrupt is
enabled by gsi_evt_ring_command() and disabled again after the
command has completed (or we time out).
There is a window of time during which the command could complete
successfully without interrupting. This would cause the event ring
to transition to the desired new state.
So whether a event ring command ends via completion interrupt or
timeout, we can consider the command successful if the event ring
has entered the desired state (and a failure if it has not,
regardless of the cause).
Fixes: b4175f8731 ("net: ipa: only enable GSI event control IRQs when needed")
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The result of issuing a channel control command should be that the
channel changes state. If enabled, a completion interrupt signals
that the channel state has changed. This interrupt is enabled by
gsi_channel_command() and disabled again after the command has
completed (or we time out).
There is a window of time--after the completion interrupt is disabled
but before the channel state is read--during which the command could
complete successfully without interrupting. This would cause the
channel to transition to the desired new state.
So whether a channel command ends via completion interrupt or
timeout, we can consider the command successful if the channel
has entered the desired state (and a failure if it has not,
regardless of the cause).
Fixes: d6c9e3f506 ("net: ipa: only enable generic command completion IRQ when needed");
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
We enable the completion interrupt for channel or event ring
commands only when we issue them. The interrupt is disabled after
the interrupt has fired, or after we have timed out waiting for it.
If we time out, the command could complete after the interrupt has
been disabled, causing a state change in the channel or event ring.
The interrupt associated with that state change would be delivered
the next time the completion interrupt is enabled.
To avoid previous command completions interfering with new commands,
clear all pending completion interrupts before re-enabling them for
a new command.
Fixes: b4175f8731 ("net: ipa: only enable GSI event control IRQs when needed")
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Add TGL-H PCI info and PCI IDs for the new TSN Controller to the list
of supported devices.
Signed-off-by: Noor Azura Ahmad Tarmizi <noor.azura.ahmad.tarmizi@intel.com>
Signed-off-by: Voon Weifeng <weifeng.voon@intel.com>
Signed-off-by: Muhammad Husaini Zulkifli <muhammad.husaini.zulkifli@intel.com>
Link: https://lore.kernel.org/r/20201222160337.30870-1-muhammad.husaini.zulkifli@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
When the core clock rate and interconnect bandwidth specifications
were moved into configuration data, a copy/paste bug was introduced,
causing the memory interconnect bandwidth to be set three times
rather than enabling the three different interconnects.
Fix this bug.
Fixes: 91d02f9551 ("net: ipa: use config data for clocking")
Signed-off-by: Alex Elder <elder@linaro.org>
Reviewed-by: Georgi Djakov <georgi.djakov@linaro.org>
Link: https://lore.kernel.org/r/20201222151613.5730-1-elder@linaro.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
virtnet_set_channels can recursively call cpus_read_lock if CONFIG_XPS
and CONFIG_HOTPLUG are enabled.
The path is:
virtnet_set_channels - calls get_online_cpus(), which is a trivial
wrapper around cpus_read_lock()
netif_set_real_num_tx_queues
netif_reset_xps_queues_gt
netif_reset_xps_queues - calls cpus_read_lock()
This call chain and potential deadlock happens when the number of TX
queues is reduced.
This commit the removes netif_set_real_num_[tr]x_queues calls from
inside the get/put_online_cpus section, as they don't require that it
be held.
Fixes: 47be24796c ("virtio-net: fix the set affinity bug when CPU IDs are not consecutive")
Signed-off-by: Jeff Dike <jdike@akamai.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Link: https://lore.kernel.org/r/20201223025421.671-1-jdike@akamai.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
First set of fixes for v5.11, more fixes than usual this time. For
ath11k we have several fixes for QCA6390 PCI support and mt76 has
several. Also one build fix for mt76.
mt76
* fix two NULL pointer dereference
* fix build error when CONFIG_MAC80211_MESH is disabled
rtlwifi
* fix use-after-free in firmware handling code
ath11k
* error handling fixes
* fix crash found during connect and disconnect test
* handle HT disable better
* avoid printing qmi memory failure during firmware bootup
* disable ASPM during firmware bootup
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
iQEcBAABAgAGBQJf4h9uAAoJEG4XJFUm622byVMIAIwMhqVmDL7jqqOwd2xZa4jf
/QzAvx7tize/tBF6FubOpF+Bxzn4XaJzwFE97B0wwO6/rnxD2a56WUARuYF8ztnT
St18uYBU23n+QKfvL1qJI5Az0JXxia/cLDwvr8YPK3rmh2jmSOVyJSCUtHSJ+yQ2
hnJA3Af8ves3VpCZ2jqF7VcHM/xYfmoJZy3I7P6LpT5m+6FOqqdN8Mx6ei2aHEzY
+HHN26QoW35N2DCna+iTtEMpepWU8eoJmIOU+t881bRfZBoczc+WH99ZlSaVDUug
hl+Hy5bwDzN+HdGziSgWP6U5a+XNjRu6Y5dUpEm4YmL+uJeU4NPc0Q35nyx4RBM=
=fjHi
-----END PGP SIGNATURE-----
Merge tag 'wireless-drivers-2020-12-22' of git://git.kernel.org/pub/scm/linux/kernel/git/kvalo/wireless-drivers
Kalle Valo says:
====================
wireless-drivers fixes for v5.11
First set of fixes for v5.11, more fixes than usual this time. For
ath11k we have several fixes for QCA6390 PCI support and mt76 has
several. Also one build fix for mt76.
mt76
* fix two NULL pointer dereference
* fix build error when CONFIG_MAC80211_MESH is disabled
rtlwifi
* fix use-after-free in firmware handling code
ath11k
* error handling fixes
* fix crash found during connect and disconnect test
* handle HT disable better
* avoid printing qmi memory failure during firmware bootup
* disable ASPM during firmware bootup
* tag 'wireless-drivers-2020-12-22' of git://git.kernel.org/pub/scm/linux/kernel/git/kvalo/wireless-drivers:
MAINTAINERS: switch to different email address
mt76: mt7915: fix MESH ifdef block
mt76: mt76s: fix NULL pointer dereference in mt76s_process_tx_queue
mt76: sdio: remove wake logic in mt76s_process_tx_queue
mt76: usb: remove wake logic in mt76u_status_worker
ath11k: pci: disable ASPM L0sLs before downloading firmware
ath11k: qmi: try to allocate a big block of DMA memory first
rtlwifi: rise completion at the last step of firmware callback
mt76: mt76u: fix NULL pointer dereference in mt76u_status_worker
ath11k: Fix ath11k_pci_fix_l1ss()
ath11k: Fix error code in ath11k_core_suspend()
ath11k: start vdev if a bss peer is already created
ath11k: fix crash caused by NULL rx_channel
ath11k: add missing null check on allocated skb
====================
Link: https://lore.kernel.org/r/20201222163727.D4336C433C6@smtp.codeaurora.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
IPIP tunnels packets are unknown to device,
hence these packets are incorrectly parsed and
caused the packet corruption, so disable offlods
for such packets at run time.
Signed-off-by: Manish Chopra <manishc@marvell.com>
Signed-off-by: Sudarsana Kalluru <skalluru@marvell.com>
Signed-off-by: Igor Russkikh <irusskikh@marvell.com>
Link: https://lore.kernel.org/r/20201221145530.7771-1-manishc@marvell.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Join adjacent questions to a single question line. This fixes the
formatting of questions that were not part of the heading.
Also, drop Q: and A: prefixes. We don't need them now that questions and
answers are visually separated.
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Link: https://lore.kernel.org/r/f76078ba5547744f2ec178984c32fbc7dcd29a2b.1608454187.git.baruch@tkos.co.il
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
When mvneta_port_power_up() fails, we should execute
cleanup functions after label err_netdev to avoid memleak.
Fixes: 41c2b6b4f0 ("net: ethernet: mvneta: Add back interface mode validation")
Signed-off-by: Dinghao Liu <dinghao.liu@zju.edu.cn>
Link: https://lore.kernel.org/r/20201220082930.21623-1-dinghao.liu@zju.edu.cn
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Commit 34f0f4e3f4 ("ibmvnic: Fix login buffer memory leaks") frees
login_rsp_buffer in release_resources() and send_login()
because handle_login_rsp() does not free it.
Commit f3ae59c0c0 ("ibmvnic: store RX and TX subCRQ handle array in
ibmvnic_adapter struct") frees login_rsp_buffer in handle_login_rsp().
It seems unnecessary to free it in release_resources() and send_login().
There are chances that handle_login_rsp returns earlier without freeing
buffers. Double-checking the buffer is harmless since
release_login_buffer and release_login_rsp_buffer will
do nothing if buffer is already freed.
Fixes: f3ae59c0c0 ("ibmvnic: store RX and TX subCRQ handle array in ibmvnic_adapter struct")
Fixes: 34f0f4e3f4 ("ibmvnic: Fix login buffer memory leaks")
Signed-off-by: Lijun Pan <ljp@linux.ibm.com>
Link: https://lore.kernel.org/r/20201219213919.21045-1-ljp@linux.ibm.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
When searching for inactive maintainers it's useful to filter
out mailing list addresses. Such "maintainers" will obviously
never feature in a "From:" line of an email or a review tag.
Since "L:" entries only provide the address of a mailing list
without a fancy name extend this pattern to "M:" entries.
Alternatively we could reserve M: entries for humans only
and move the fake "maintainers" to L:. While I'd personally
prefer to reserve M: for humans only, I'm not 100% that's
a great choice either, given most L: entries are in fact
open mailing lists with public archives.
Link: https://lore.kernel.org/r/20201219185538.750076-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The dwmac glue registers on Amlogic Meson8b and newer SoCs has two clock
inputs:
- Meson8b and Meson8m2: MPLL2 and MPLL2 (the same parent is wired to
both inputs)
- GXBB, GXL, GXM, AXG, G12A, G12B, SM1: FCLK_DIV2 and MPLL2
All known vendor kernels and u-boots are using the first input only. We
let the common clock framework automatically choose the "right" parent.
For some boards this causes a problem though, specificially with G12A and
newer SoCs. The clock input is used for generating the 125MHz RGMII TX
clock. For the two input clocks this means on G12A:
- FCLK_DIV2: 999999985Hz / 8 = 124999998.125Hz
- MPLL2: 499999993Hz / 4 = 124999998.25Hz
In theory MPLL2 is the "better" clock input because it's gets us 0.125Hz
closer to the requested frequency than FCLK_DIV2. In reality however
there is a resource conflict because MPLL2 is needed to generate some of
the audio clocks. dwmac-meson8b probes first and sets up the clock tree
with MPLL2. This works fine until the audio driver comes and "steals"
the MPLL2 clocks and configures it with it's own rate (294909637Hz). The
common clock framework happily changes the MPLL2 rate but does not
reconfigure our RGMII TX clock tree, which then ends up at 73727409Hz,
which is more than 40% off the requested 125MHz.
Don't use the second clock input for now to force the common clock
framework to always select the first parent. This mimics the behavior
from the vendor driver and fixes the clock resource conflict with the
audio driver on G12A boards. Once the common clock framework can handle
this situation this change can be reverted again.
Fixes: 566e825162 ("net: stmmac: add a glue driver for the Amlogic Meson 8b / GXBB DWMAC")
Reported-by: Thomas Graichen <thomas.graichen@gmail.com>
Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
Tested-by: thomas graichen <thomas.graichen@gmail.com>
Link: https://lore.kernel.org/r/20201219135036.3216017-1-martin.blumenstingl@googlemail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
During GoP port 2 Networking Complex Control mode of operation configurations,
also GoP port 3 mode of operation was wrongly set.
Patch removes these configurations.
Fixes: f84bf386f3 ("net: mvpp2: initialize the GoP")
Acked-by: Marcin Wojtas <mw@semihalf.com>
Signed-off-by: Stefan Chulski <stefanc@marvell.com>
Link: https://lore.kernel.org/r/1608462149-1702-1-git-send-email-stefanc@marvell.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
This error path needs to disable the pci device before returning.
Fixes: ede58ef28e ("atm: remove deprecated use of pci api")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Link: https://lore.kernel.org/r/X93dmC4NX0vbTpGp@mwanda
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Let the FW know we have enough receive buffer space for the
vlan tag if it isn't stripped.
Fixes: 0f3154e6bc ("ionic: Add Tx and Rx handling")
Signed-off-by: Shannon Nelson <snelson@pensando.io>
Link: https://lore.kernel.org/r/20201218215001.64696-1-snelson@pensando.io
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
ugeth is the netdiv_priv() part of the netdevice. Accessing the memory
pointed to by ugeth (such as done by ucc_geth_memclean() and the two
of_node_puts) after free_netdev() is thus use-after-free.
Fixes: 80a9fad8e8 ("ucc_geth: fix module removal")
Signed-off-by: Rasmus Villemoes <rasmus.villemoes@prevas.dk>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Table 8-53 in the QUICC Engine Reference manual shows definitions of
fields up to a size of 192 bytes, not just 128. But in table 8-111,
one does find the text
Base Address of the Global Transmitter Parameter RAM Page. [...]
The user needs to allocate 128 bytes for this page. The address must
be aligned to the page size.
I've checked both rev. 7 (11/2015) and rev. 9 (05/2018) of the manual;
they both have this inconsistency (and the table numbers are the
same).
Adding a bit of debug printing, on my board the struct
ucc_geth_tx_global_pram is allocated at offset 0x880, while
the (opaque) ucc_geth_thread_data_tx gets allocated immediately
afterwards, at 0x900. So whatever the engine writes into the thread
data overlaps with the tail of the global tx pram (and devmem says
that something does get written during a simple ping).
I haven't observed any failure that could be attributed to this, but
it seems to be the kind of thing that would be extremely hard to
debug. So extend the struct definition so that we do allocate 192
bytes.
Signed-off-by: Rasmus Villemoes <rasmus.villemoes@prevas.dk>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
All the buffers and registers are already set up appropriately for an
MTU slightly above 1500, so we just need to expose this to the
networking stack. AFAICT, there's no need to implement .ndo_change_mtu
when the receive buffers are always set up to support the max_mtu.
This fixes several warnings during boot on our mpc8309-board with an
embedded mv88e6250 switch:
mv88e6085 mdio@e0102120:10: nonfatal error -34 setting MTU 1500 on port 0
...
mv88e6085 mdio@e0102120:10: nonfatal error -34 setting MTU 1500 on port 4
ucc_geth e0102000.ethernet eth1: error -22 setting MTU to 1504 to include DSA overhead
The last line explains what the DSA stack tries to do: achieving an MTU
of 1500 on-the-wire requires that the master netdevice connected to
the CPU port supports an MTU of 1500+the tagging overhead.
Fixes: bfcb813203 ("net: dsa: configure the MTU for switch ports")
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Rasmus Villemoes <rasmus.villemoes@prevas.dk>
Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The driver is already allocating receive buffers of 2KiB and the
Ethernet MAC is configured to accept frames up to UMAC_MAX_MTU_SIZE.
Fixes: bfcb813203 ("net: dsa: configure the MTU for switch ports")
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Vladimir Oltean <olteanv@gmail.com>
Link: https://lore.kernel.org/r/20201218173843.141046-1-f.fainelli@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
weichenchen