The page migration code employs try_to_unmap() to try and unmap the source
page. This is accomplished by using rmap_walk to find all vmas where the
page is mapped. This search stops when page mapcount is zero. For shared
PMD huge pages, the page map count is always 1 no matter the number of
mappings. Shared mappings are tracked via the reference count of the PMD
page. Therefore, try_to_unmap stops prematurely and does not completely
unmap all mappings of the source page.
This problem can result is data corruption as writes to the original
source page can happen after contents of the page are copied to the target
page. Hence, data is lost.
This problem was originally seen as DB corruption of shared global areas
after a huge page was soft offlined due to ECC memory errors. DB
developers noticed they could reproduce the issue by (hotplug) offlining
memory used to back huge pages. A simple testcase can reproduce the
problem by creating a shared PMD mapping (note that this must be at least
PUD_SIZE in size and PUD_SIZE aligned (1GB on x86)), and using
migrate_pages() to migrate process pages between nodes while continually
writing to the huge pages being migrated.
To fix, have the try_to_unmap_one routine check for huge PMD sharing by
calling huge_pmd_unshare for hugetlbfs huge pages. If it is a shared
mapping it will be 'unshared' which removes the page table entry and drops
the reference on the PMD page. After this, flush caches and TLB.
mmu notifiers are called before locking page tables, but we can not be
sure of PMD sharing until page tables are locked. Therefore, check for
the possibility of PMD sharing before locking so that notifiers can
prepare for the worst possible case.
Link: http://lkml.kernel.org/r/20180823205917.16297-2-mike.kravetz@oracle.com
[mike.kravetz@oracle.com: make _range_in_vma() a static inline]
Link: http://lkml.kernel.org/r/6063f215-a5c8-2f0c-465a-2c515ddc952d@oracle.com
Fixes: 39dde65c99 ("shared page table for hugetlb page")
Signed-off-by: Mike Kravetz <mike.kravetz@oracle.com>
Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Reviewed-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: Jerome Glisse <jglisse@redhat.com>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
4.19 merge window.
- Fix leak and dangling pointer in DM multipath's scsi_dh related code.
- A couple stable@ fixes for DM cache's resize support.
- A DM raid fix to remove "const" from decipher_sync_action()'s return
type.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
iQEcBAABAgAGBQJbt7RQAAoJEMUj8QotnQNac2MIAK0p2Ndy7ErItA+qIlmBodsO
juWGdnhBXTHQ9qoeIuamqA66g7LZ8oxnaiIQX8tqzt+GyzxgcjH+9C+Nn7qf8VLF
4SkbI8/8CxBKmsKv5G5OI2tCfdfTXop9rOTg+3tF6BcTxXzPGeY+5mImsMN4KI/S
1+hnq2xMtScdTDLyN0qKxs0+e8YdP7IJn2QAscLVNik0wdWn+7Hfp7Hg7tXcupyF
WQ1dkY8kaGCUid167jAkCmY4etXHIfAb81EWxFtGb/GNGlFmc57t7HGa5mVaO3qv
QXgp9u27q7v/wqzDHO68+mzAfiBHMFmJHa4uRg7gR1LrE9/0y3CDWC3zMYDL3Es=
=Av3x
-----END PGP SIGNATURE-----
Merge tag 'for-4.19/dm-fixes-2' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm
Mike writes:
"device mapper fixes
- Fix a DM thinp __udivdi3 undefined on 32-bit bug introduced during
4.19 merge window.
- Fix leak and dangling pointer in DM multipath's scsi_dh related code.
- A couple stable@ fixes for DM cache's resize support.
- A DM raid fix to remove "const" from decipher_sync_action()'s return
type."
* tag 'for-4.19/dm-fixes-2' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm:
dm cache: fix resize crash if user doesn't reload cache table
dm cache metadata: ignore hints array being too small during resize
dm raid: remove bogus const from decipher_sync_action() return type
dm mpath: fix attached_handler_name leak and dangling hw_handler_name pointer
dm thin metadata: fix __udivdi3 undefined on 32-bit
Free the last used descriptor, an off by one error.
This is tagged for stable as well.
-----BEGIN PGP SIGNATURE-----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=PKfb
-----END PGP SIGNATURE-----
Merge tag 'gpio-v4.19-3' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-gpio
Linus writes:
"A single GPIO fix:
Free the last used descriptor, an off by one error.
This is tagged for stable as well."
* tag 'gpio-v4.19-3' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-gpio:
gpiolib: Free the last requested descriptor
Fix a bug that may cause runtime PM to misbehave for some devices
after a failing or aborted system suspend which is nasty enough for
an -rc7 time frame fix.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2
iQIcBAABCAAGBQJbtzD9AAoJEILEb/54YlRxRCEP/jBooTFH6SKEY0EIvzpda6yV
Pi3q8sMkcGJMm8/HRqRqpg6N03eUK/2sdoqQf2fRDoJfWa/NcZ3Vcq9gElN8JSgp
vTYUYVANsD+je8NbWyV6WrwGRedlebw4nkeS+/uAbpEPwKNJv94H0YZ71JVZa/lE
Mx6C+gClepKz3So4JKcfn0NNyLlAl6iCJOB/8WRHZIj17WJBsMSV+I4mfjexlbQD
AloPeVYtsGSSsi/Y0aKbjcpE1i49zlFoS6Vqj4rfkXqbwGEaFEoQutIviO+hYsH6
3DDUtoCF3w0haJJNaeAbB2XB+uVeIzxp4N5eq9mEWwZhKwkm37NsDWmV6D1oCQXl
2nhDmj2gCypMRQr3/QvAareW3FxHrJqyEEcgnSdM7EX0Sc3EmXZF5T8teGUtZP2s
w8byBKnTbvER9X6THZajlYym0owR5LybjXAtvb2Y9s5KCauw9Dw6n14dlPnZAHQo
O5eqcDnQaQhfznU+ngYFa8RBdKKJPbORrma0yKRhs5S8CvdQci2nx4cwkaHw/pPR
GSSV6yrL5ogEPLISXvITrg0iglHizxPonbzcEcfKljEd1+2Ai/oJZpM90BveaOvV
t1IBY3eHFefW4tZnToS4fRy/WtamPIqOXFuIrnbU4KKeQWv7Yreb2NbPmM/okzr6
SxKdPMUUmeyI2jxA03r2
=wko2
-----END PGP SIGNATURE-----
Merge tag 'pm-4.19-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm
Rafael writes:
"Power management fix for 4.19-rc7
Fix a bug that may cause runtime PM to misbehave for some devices
after a failing or aborted system suspend which is nasty enough for
an -rc7 time frame fix."
* tag 'pm-4.19-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
PM / core: Clear the direct_complete flag on errors
Ingo writes:
"perf fixes:
- fix a CPU#0 hot unplug bug and a PCI enumeration bug in the x86 Intel uncore PMU driver
- fix a CPU event enumeration bug in the x86 AMD PMU driver
- fix a perf ring-buffer corruption bug when using tracepoints
- fix a PMU unregister locking bug"
* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
perf/x86/amd/uncore: Set ThreadMask and SliceMask for L3 Cache perf events
perf/x86/intel/uncore: Fix PCI BDF address of M3UPI on SKX
perf/ring_buffer: Prevent concurent ring buffer access
perf/x86/intel/uncore: Use boot_cpu_data.phys_proc_id instead of hardcorded physical package ID 0
perf/core: Fix perf_pmu_unregister() locking
Ingo writes:
"scheduler fixes:
These fixes address a rather involved performance regression between
v4.17->v4.19 in the sched/numa auto-balancing code. Since distros
really need this fix we accelerated it to sched/urgent for a faster
upstream merge.
NUMA scheduling and balancing performance is now largely back to
v4.17 levels, without reintroducing the NUMA placement bugs that
v4.18 and v4.19 fixed.
Many thanks to Srikar Dronamraju, Mel Gorman and Jirka Hladky, for
reporting, testing, re-testing and solving this rather complex set of
bugs."
* 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
sched/numa: Migrate pages to local nodes quicker early in the lifetime of a task
mm, sched/numa: Remove rate-limiting of automatic NUMA balancing migration
sched/numa: Avoid task migration for small NUMA improvement
mm/migrate: Use spin_trylock() while resetting rate limit
sched/numa: Limit the conditions where scan period is reset
sched/numa: Reset scan rate whenever task moves across nodes
sched/numa: Pass destination CPU as a parameter to migrate_task_rq
sched/numa: Stop multiple tasks from moving to the CPU at the same time
Ingo writes:
"locking fixes:
A fix in the ww_mutex self-test that produces a scary splat, plus an
updates to the maintained-filed patters in MAINTAINER."
* 'locking-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
locking/ww_mutex: Fix runtime warning in the WW mutex selftest
MAINTAINERS: Remove dead path from LOCKING PRIMITIVES entry
Just two small fixes for HD-audio: one is for a typo in completion
timeout, and another a fixup for Dell machines as usual
-----BEGIN PGP SIGNATURE-----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=fjND
-----END PGP SIGNATURE-----
Merge tag 'sound-4.19-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound
Takashi writes:
"sound fixes for 4.19-rc7
Just two small fixes for HD-audio: one is for a typo in completion
timeout, and another a fixup for Dell machines as usual"
* tag 'sound-4.19-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound:
ALSA: hda/realtek - Cannot adjust speaker's volume on Dell XPS 27 7760
ALSA: hda: Fix the audio-component completion timeout
This patch adds OEM commands and response handling. It also defines OEM
command and response structure as per NCSI specification along with its
handlers.
ncsi_cmd_handler_oem: This is a generic command request handler for OEM
commands
ncsi_rsp_handler_oem: This is a generic response handler for OEM commands
Signed-off-by: Vijay Khemka <vijaykhemka@fb.com>
Reviewed-by: Justin Lee <justin.lee1@dell.com>
Reviewed-by: Samuel Mendoza-Jonas <sam@mendozajonas.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Address compiler warning:
ip_defrag.c: In function 'send_udp_frags':
ip_defrag.c:206:16: warning: unused variable 'udphdr' [-Wunused-variable]
struct udphdr udphdr;
^~~~~~
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When offloading the L3 and L4 csum computation on TX, we need to extract
the l3_proto from the ethtype, independently of the presence of a vlan
tag.
The actual driver uses skb->protocol as-is, resulting in packets with
the wrong L4 checksum being sent when there's a vlan tag in the packet
header and checksum offloading is enabled.
This commit makes use of vlan_protocol_get() to get the correct ethtype
regardless the presence of a vlan tag.
Fixes: 3f518509de ("ethernet: Add new driver for Marvell Armada 375 network unit")
Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
use a TDC plugin, instead of building eBPF programs in the 'setup' stage.
'-B' argument can be used to build eBPF programs in $EBPFDIR directory,
in the 'pre-suite' stage. Binaries are then cleaned in 'post-suite' stage.
Signed-off-by: Davide Caratti <dcaratti@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
rely on uAPI headers in the current kernel tree, rather than requiring the
correct version installed on the test system. While at it, group all
sections in a single binary and test the 'section' parameter.
Reported-by: Lucas Bates <lucasb@mojatatu.com>
Signed-off-by: Davide Caratti <dcaratti@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Quentin Schulz says:
====================
mscc: ocelot: add support for SerDes muxing configuration
The Ocelot switch has currently an hardcoded SerDes muxing that suits only
a particular use case. Any other board setup will fail to work.
To prepare for upcoming boards' support that do not have the same muxing,
create a PHY driver that will handle all possible cases.
A SerDes can work in SGMII, QSGMII or PCIe and is also muxed to use a
given port depending on the selected mode or board design.
The SerDes configuration is in the middle of an address space (HSIO) that
is used to configure some parts in the MAC controller driver, that is why
we need to use a syscon so that we can write to the same address space from
different drivers safely using regmap.
This breaks backward compatibility but it's fine because there's only one
board at the moment that is using what's modified in this patch series.
This will break git bisect.
Even though this patch series is about SerDes __muxing__ configuration, the
DT node is named serdes for the simple reason that I couldn't find any
mention to SerDes anywhere else from the address space handled by this
driver.
v4:
- add reviewed-by,
- format the patch series with -M for identifying renamed files,
- add parent info in DT binding of the SerDes IP,
- move to macros SERDES[16]G(X) instead of multiple SERDES[16]G_[012345]
constants,
- move to SERDES[16]G_MAX being the last VALID macro of a type, so
migrate to <= conditions instead of < when iterating,
- create a SERDES_MUX_SGMII and SERDES_MUX_QSGMII macro so the muxing
configurations are a tad more readable,
- use a bunch of unsigned int instead of int,
- return -EOPNOTSUPP for SERDES6G/PCIe until it's supported,
- simplify condition when there is an error code returned by
devm_of_phy_get,
v3:
- add Paul Burton's Acked-By on MIPS patches so that the patch series can
be merged in the net tree in its entirety,
v2:
- use a switch case for setting the phy_mode in the SerDes driver as
suggested by Andrew,
- stop replacing the value of the error pointer in the SerDes driver,
- use a dev_dbg for the deferring of the probe in the SerDes driver,
- use constants in the Device Tree to select the SerDes macro in use with
a port,
- adapt the SerDes driver to use those constants,
- add a header file in include/dt-bindings for the constants,
- fix space/tab issue,
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Previously, the SerDes muxing was hardcoded to a given mode in the MAC
controller driver. Now, the SerDes muxing is configured within the
Device Tree and is enforced in the MAC controller driver so we can have
a lot of different SerDes configurations.
Make use of the SerDes PHYs in the MAC controller to set up the SerDes
according to the SerDes<->switch port mapping and the communication mode
with the Ethernet PHY.
Signed-off-by: Quentin Schulz <quentin.schulz@bootlin.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The Microsemi Ocelot can mux SerDes lanes (aka macros) to different
switch ports or even make it act as a PCIe interface.
This adds support for the muxing of the SerDes.
Signed-off-by: Quentin Schulz <quentin.schulz@bootlin.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The Microsemi Ocelot has multiple SerDes and requires that the SerDes be
muxed accordingly to the hardware representation.
Let's add a constant for each SerDes available in the Microsemi Ocelot.
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Quentin Schulz <quentin.schulz@bootlin.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The Microsemi Ocelot has a set of register for SerDes/switch port muxing
as well as PCIe muxing for a specific SerDes, so let's add the device
and all SerDes in the Device Tree.
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Acked-by: Paul Burton <paul.burton@mips.com>
Signed-off-by: Quentin Schulz <quentin.schulz@bootlin.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Prepare for upcoming phys that'll handle QSGMII or PCIe.
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: Quentin Schulz <quentin.schulz@bootlin.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since HSIO address space can be accessed by different drivers, let's
simplify the register address definitions so that it can be easily used
by all drivers and put the register address definition in the
include/soc/mscc/ocelot_hsio.h header file.
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Acked-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Quentin Schulz <quentin.schulz@bootlin.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since HSIO address space can be used by different drivers (PLL, SerDes
muxing, temperature sensor), let's move it somewhere it can be included
by all drivers.
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Acked-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Quentin Schulz <quentin.schulz@bootlin.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
HSIO address space was moved to a syscon, hence we need to get the
regmap of this address space from there and no more from the device
node.
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Acked-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Quentin Schulz <quentin.schulz@bootlin.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
HSIO register address space should be handled outside of the MAC
controller as there are some registers for PLL5 configuring,
SerDes/switch port muxing and a thermal sensor IP, so let's remove it.
Acked-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Quentin Schulz <quentin.schulz@bootlin.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
HSIO contains registers for PLL5 configuration, SerDes/switch port
muxing and a thermal sensor, hence we can't keep it in the switch DT
node.
Acked-by: Paul Burton <paul.burton@mips.com>
Acked-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Quentin Schulz <quentin.schulz@bootlin.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
move_addr_to_kernel() returns only negative values on error, or zero on
success. Rewrite the error check to an idiomatic form to avoid confusing
the reader.
Signed-off-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
ns_poll() used spin_is_locked() + spin_lock() to get achieve the same
thing as a spin_trylock(), so simplify it by using that instead. This is
also a step towards possibly removing spin_is_locked().
Signed-off-by: Lance Roy <ldr709@gmail.com>
Cc: Chas Williams <3chas3@gmail.com>
Cc: <linux-atm-general@lists.sourceforge.net>
Cc: <netdev@vger.kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
A number of TC attributes are processed without proper validation
(e.g., length checks). Add a tca policy for all input attributes and use
when invoking nlmsg_parse.
The 2 Fixes tags below cover the latest additions. The other attributes
are a string (KIND), nested attribute (OPTIONS which does seem to have
validation in most cases), for dumps only or a flag.
Fixes: 5bc1701881 ("net: sched: introduce multichain support for filters")
Fixes: d47a6b0e7c ("net: sched: introduce ingress/egress block index attributes for qdisc")
Signed-off-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently, rtnl_fdb_dump() assumes the family header is 'struct ifinfomsg',
which is not always true -- 'struct ndmsg' is used by iproute2 ('ip neigh').
The problem is, the function bails out early if nlmsg_parse() fails, which
does occur for iproute2 usage of 'struct ndmsg' because the payload length
is shorter than the family header alone (as 'struct ifinfomsg' is assumed).
This breaks backward compatibility with userspace -- nothing is sent back.
Some examples with iproute2 and netlink library for go [1]:
1) $ bridge fdb show
33:33:00:00:00:01 dev ens3 self permanent
01:00:5e:00:00:01 dev ens3 self permanent
33:33:ff:15:98:30 dev ens3 self permanent
This one works, as it uses 'struct ifinfomsg'.
fdb_show() @ iproute2/bridge/fdb.c
"""
.n.nlmsg_len = NLMSG_LENGTH(sizeof(struct ifinfomsg)),
...
if (rtnl_dump_request(&rth, RTM_GETNEIGH, [...]
"""
2) $ ip --family bridge neigh
RTNETLINK answers: Invalid argument
Dump terminated
This one fails, as it uses 'struct ndmsg'.
do_show_or_flush() @ iproute2/ip/ipneigh.c
"""
.n.nlmsg_type = RTM_GETNEIGH,
.n.nlmsg_len = NLMSG_LENGTH(sizeof(struct ndmsg)),
"""
3) $ ./neighlist
< no output >
This one fails, as it uses 'struct ndmsg'-based.
neighList() @ netlink/neigh_linux.go
"""
req := h.newNetlinkRequest(unix.RTM_GETNEIGH, [...]
msg := Ndmsg{
"""
The actual breakage was introduced by commit 0ff50e83b5 ("net: rtnetlink:
bail out from rtnl_fdb_dump() on parse error"), because nlmsg_parse() fails
if the payload length (with the _actual_ family header) is less than the
family header length alone (which is assumed, in parameter 'hdrlen').
This is true in the examples above with struct ndmsg, with size and payload
length shorter than struct ifinfomsg.
However, that commit just intends to fix something under the assumption the
family header is indeed an 'struct ifinfomsg' - by preventing access to the
payload as such (via 'ifm' pointer) if the payload length is not sufficient
to actually contain it.
The assumption was introduced by commit 5e6d243587 ("bridge: netlink dump
interface at par with brctl"), to support iproute2's 'bridge fdb' command
(not 'ip neigh') which indeed uses 'struct ifinfomsg', thus is not broken.
So, in order to unbreak the 'struct ndmsg' family headers and still allow
'struct ifinfomsg' to continue to work, check for the known message sizes
used with 'struct ndmsg' in iproute2 (with zero or one attribute which is
not used in this function anyway) then do not parse the data as ifinfomsg.
Same examples with this patch applied (or revert/before the original fix):
$ bridge fdb show
33:33:00:00:00:01 dev ens3 self permanent
01:00:5e:00:00:01 dev ens3 self permanent
33:33:ff:15:98:30 dev ens3 self permanent
$ ip --family bridge neigh
dev ens3 lladdr 33:33:00:00:00:01 PERMANENT
dev ens3 lladdr 01:00:5e:00:00:01 PERMANENT
dev ens3 lladdr 33:33:ff:15:98:30 PERMANENT
$ ./neighlist
netlink.Neigh{LinkIndex:2, Family:7, State:128, Type:0, Flags:2, IP:net.IP(nil), HardwareAddr:net.HardwareAddr{0x33, 0x33, 0x0, 0x0, 0x0, 0x1}, LLIPAddr:net.IP(nil), Vlan:0, VNI:0}
netlink.Neigh{LinkIndex:2, Family:7, State:128, Type:0, Flags:2, IP:net.IP(nil), HardwareAddr:net.HardwareAddr{0x1, 0x0, 0x5e, 0x0, 0x0, 0x1}, LLIPAddr:net.IP(nil), Vlan:0, VNI:0}
netlink.Neigh{LinkIndex:2, Family:7, State:128, Type:0, Flags:2, IP:net.IP(nil), HardwareAddr:net.HardwareAddr{0x33, 0x33, 0xff, 0x15, 0x98, 0x30}, LLIPAddr:net.IP(nil), Vlan:0, VNI:0}
Tested on mainline (v4.19-rc6) and net-next (3bd09b05b0).
References:
[1] netlink library for go (test-case)
https://github.com/vishvananda/netlink
$ cat ~/go/src/neighlist/main.go
package main
import ("fmt"; "syscall"; "github.com/vishvananda/netlink")
func main() {
neighs, _ := netlink.NeighList(0, syscall.AF_BRIDGE)
for _, neigh := range neighs { fmt.Printf("%#v\n", neigh) }
}
$ export GOPATH=~/go
$ go get github.com/vishvananda/netlink
$ go build neighlist
$ ~/go/src/neighlist/neighlist
Thanks to David Ahern for suggestions to improve this patch.
Fixes: 0ff50e83b5 ("net: rtnetlink: bail out from rtnl_fdb_dump() on parse error")
Fixes: 5e6d243587 ("bridge: netlink dump interface at par with brctl")
Reported-by: Aidan Obley <aobley@pivotal.io>
Signed-off-by: Mauricio Faria de Oliveira <mfo@canonical.com>
Reviewed-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Salil Mehta says:
====================
Fixes, minor changes & cleanups for the Unicast MAC VLAN table
This patch-set presents necessary modifications, fixes & cleanups related
to Unicast MAC Vlan Table to support revision 0x21 hardware.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
In revision 0x20, we use vlan id != 0 to check whether a vlan tag
has been offloaded, so vlan id 0 is not supported.
In revision 0x21, rx buffer descriptor adds two bits to indicate
whether one or more vlan tags have been offloaded, so vlan id 0
is valid now.
This patch seperates the handle for vlan id 0, add vlan id 0 support
for revision 0x21.
Fixes: 5b5455a9ed ("net: hns3: Add STRP_TAGP field support for hardware revision 0x21")
Signed-off-by: Jian Shen <shenjian15@huawei.com>
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In revision 0x21, hw supports both ingress and egress vlan filter.
This patch adds support for it.
Signed-off-by: Zhongzhu Liu <liuzhongzhu@huawei.com>
Signed-off-by: Jian Shen <shenjian15@huawei.com>
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
For mta table support has been dropped, remove the code for mta table.
Signed-off-by: Zhongzhu Liu <liuzhongzhu@huawei.com>
Signed-off-by: Jian Shen <shenjian15@huawei.com>
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In previously implement for unicast mac vlan table, the space is
shared by all the functions, driver does nothing when the space is
exhausted. This patch preallocates the space of unicast mac vlan
table for each function by software. Each function can only use its
private space and available shared space, avoiding single function
exhausts too much space, and other functions are unable to add
unicast mac address.
Signed-off-by: Jian Shen <shenjian15@huawei.com>
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In original codes, the mac vlan table entries are not cleared when
unload hns3 driver. The dirty mac vlan table entries will make the
result of looking up mac vlan table being unexpected.
When doing core reset or global reset, the firmware will clear all
the tables for driver, and driver shouldn't send any commands to
firmware during reset. But when doing function reset, the driver
needs to clear the tables itself.
This patch clears the mac vlan table entries for each client when
unload driver or reset.
Signed-off-by: Jian Shen <shenjian15@huawei.com>
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The default mask configuration has been done by firmware, so the driver
doesn't need to do it any more.
Signed-off-by: Zhongzhu Liu <liuzhongzhu@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Jian Shen <shenjian15@huawei.com>
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add ipv4 and ipv6 test cases with an invalid metrics option causing
ip_metrics_convert to fail. Tests clean up path during route add.
Also, add nodad to to ipv6 address add. When running ipv6_route_metrics
directly seeing an occasional failure on the "Using route with mtu metric"
test case.
Signed-off-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In case ip_fib_metrics_init() returns an error, we better
rewrite rt->fib6_metrics with &dst_default_metrics so that
we do not crash later in ip_fib_metrics_put()
Fixes: 767a221753 ("net: common metrics init helper for FIB entries")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: syzbot <syzkaller@googlegroups.com>
Reviewed-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In yam_ioctl(), the concrete ioctl command is firstly copied from the
user-space buffer 'ifr->ifr_data' to 'ioctl_cmd' and checked through the
following switch statement. If the command is not as expected, an error
code EINVAL is returned. In the following execution the buffer
'ifr->ifr_data' is copied again in the cases of the switch statement to
specific data structures according to what kind of ioctl command is
requested. However, after the second copy, no re-check is enforced on the
newly-copied command. Given that the buffer 'ifr->ifr_data' is in the user
space, a malicious user can race to change the command between the two
copies. This way, the attacker can inject inconsistent data and cause
undefined behavior.
This patch adds a re-check in each case of the switch statement if there is
a second copy in that case, to re-check whether the command obtained in the
second copy is the same as the one in the first copy. If not, an error code
EINVAL will be returned.
Signed-off-by: Wenwen Wang <wang6495@umn.edu>
Signed-off-by: David S. Miller <davem@davemloft.net>
Avoid the socket lookup cost in udp_gro_receive if no socket has a
udp tunnel callback configured.
udp_sk(sk)->gro_receive requires a registration with
setup_udp_tunnel_sock, which enables the static key.
Signed-off-by: Willem de Bruijn <willemb@google.com>
Acked-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Fixes the following Sparse warnings:
net/bpfilter/bpfilter_kern.c:62:21: warning: cast removes address space
of expression
net/bpfilter/bpfilter_kern.c:101:49: warning: Using plain integer as
NULL pointer
Signed-off-by: Shanthosh RK <shanthosh.rk@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In cxgb_extension_ioctl(), the command of the ioctl is firstly copied from
the user-space buffer 'useraddr' to 'cmd' and checked through the
switch statement. If the command is not as expected, an error code
EOPNOTSUPP is returned. In the following execution, i.e., the cases of the
switch statement, the whole buffer of 'useraddr' is copied again to a
specific data structure, according to what kind of command is requested.
However, after the second copy, there is no re-check on the newly-copied
command. Given that the buffer 'useraddr' is in the user space, a malicious
user can race to change the command between the two copies. By doing so,
the attacker can supply malicious data to the kernel and cause undefined
behavior.
This patch adds a re-check in each case of the switch statement if there is
a second copy in that case, to re-check whether the command obtained in the
second copy is the same as the one in the first copy. If not, an error code
EINVAL is returned.
Signed-off-by: Wenwen Wang <wang6495@umn.edu>
Signed-off-by: David S. Miller <davem@davemloft.net>
Move mt76x02_irq_handler handler in mt76x02_mmio.c in order to be
reused in mt76x0 driver. Move mt76x02_rx_poll_complete routine in
mt76x02-lib module. Moreover remove pci_core.c and mt76x2/trace.{c,h}
since are empty files
Signed-off-by: Lorenzo Bianconi <lorenzo.bianconi@redhat.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Move tx_tasklet management in mt76x02_mmio.c in order to
be reused by mt76x0 driver and remove duplicated code
Signed-off-by: Lorenzo Bianconi <lorenzo.bianconi@redhat.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Use mt76x02_dev data structure as reference in mt76x02_mmio.c
instead of mt76_dev
Signed-off-by: Lorenzo Bianconi <lorenzo.bianconi@redhat.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Move mt76x02_tx_complete mt76x02-lib module in order to
be reused by mt76x0 drivers for irq unification.
Signed-off-by: Lorenzo Bianconi <lorenzo.bianconi@redhat.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Move mt76x02_mac_poll_tx_status in mt76x02_mac.c in order to
be reused by mt76x0 drivers for irq unification.
Moreover introduce mt76x02_trace source file to define mt76x02
trace points
Signed-off-by: Lorenzo Bianconi <lorenzo.bianconi@redhat.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
