The task_fork_fair() callback already calls __set_task_cpu() and takes
rq->lock.
If we move the sched_class::task_fork callback in sched_fork() under
the existing p->pi_lock, right after its set_task_cpu() call, we can
avoid doing two such calls and omit the IRQ disabling on the rq->lock.
Change to __set_task_cpu() to skip the migration bits, this is a new
task, not a migration. Similarly, make wake_up_new_task() use
__set_task_cpu() for the same reason, the task hasn't actually
migrated as it hasn't ever ran.
This cures the problem of calling migrate_task_rq_fair(), which does
remove_entity_from_load_avg() on tasks that have never been added to
the load avg to begin with.
This bug would result in transiently messed up load_avg values, averaged
out after a few dozen milliseconds. This is probably the reason why
this bug was not found for such a long time.
Reported-by: Vincent Guittot <vincent.guittot@linaro.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
(cherry picked from commit e210bffd39)
Change-Id: Icbddbaa6e8c1071859673d8685bc3f38955cf144
Signed-off-by: Brendan Jackman <brendan.jackman@arm.com>
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
While set_task_rq_fair() is introduced in mainline by commit ad936d8658
("sched/fair: Make it possible to account fair load avg consistently"),
the function results to be introduced here by the backport of
commit 09a43ace1f ("sched/fair: Propagate load during synchronous
attach/detach"). The problem (apart from the confusion introduced by the
backport) is actually that set_task_rq_fair() is currently not called at
all.
Fix the problem by backporting again commit ad936d8658
("sched/fair: Make it possible to account fair load avg consistently").
Original change log:
The current code accounts for the time a task was absent from the fair
class (per ATTACH_AGE_LOAD). However it does not work correctly when a
task got migrated or moved to another cgroup while outside of the fair
class.
This patch tries to address that by aging on migration. We locklessly
read the 'last_update_time' stamp from both the old and new cfs_rq,
ages the load upto the old time, and sets it to the new time.
These timestamps should in general not be more than 1 tick apart from
one another, so there is a definite bound on things.
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
[ Changelog, a few edits and !SMP build fix ]
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/1445616981-29904-2-git-send-email-byungchul.park@lge.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
(cherry-picked from ad936d8658)
Signed-off-by: Juri Lelli <juri.lelli@arm.com>
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Change-Id: I17294ab0ada3901d35895014715fd60952949358
Signed-off-by: Brendan Jackman <brendan.jackman@arm.com>
When using schedfreq on cpus with max capacity significantly smaller than
1024, the tick update uses non-normalised capacities - this leads to
selecting an incorrect OPP as we were scaling the frequency as if the
max capacity achievable was 1024 rather than the max for that particular
cpu or group. This could result in a cpu being stuck at the lowest OPP
and unable to generate enough utilisation to climb out if the max
capacity is significantly smaller than 1024.
Instead, normalize the capacity to be in the range 0-1024 in the tick
so that when we later select a frequency, we get the correct one.
Also comments updated to be clearer about what is needed.
Change-Id: Id84391c7ac015311002ada21813a353ee13bee60
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Show the high watermark of the index into the alloc->pages
array, to facilitate sizing the buffer on a per-process
basis.
Change-Id: I2b40cd16628e0ee45216c51dc9b3c5b0c862032e
Signed-off-by: Martijn Coenen <maco@android.com>
This flag determines whether the thread should currently
process the work in the thread->todo worklist.
The prime usecase for this is improving the performance
of synchronous transactions: all synchronous transactions
post a BR_TRANSACTION_COMPLETE to the calling thread,
but there's no reason to return that command to userspace
right away - userspace anyway needs to wait for the reply.
Likewise, a synchronous transaction that contains a binder
object can cause a BC_ACQUIRE/BC_INCREFS to be returned to
userspace; since the caller must anyway hold a strong/weak
ref for the duration of the call, postponing these commands
until the reply comes in is not a problem.
Note that this flag is not used to determine whether a
thread can handle process work; a thread should never pick
up process work when thread work is still pending.
Before patch:
------------------------------------------------------------------
Benchmark Time CPU Iterations
------------------------------------------------------------------
BM_sendVec_binderize/4 45959 ns 20288 ns 34351
BM_sendVec_binderize/8 45603 ns 20080 ns 34909
BM_sendVec_binderize/16 45528 ns 20113 ns 34863
BM_sendVec_binderize/32 45551 ns 20122 ns 34881
BM_sendVec_binderize/64 45701 ns 20183 ns 34864
BM_sendVec_binderize/128 45824 ns 20250 ns 34576
BM_sendVec_binderize/256 45695 ns 20171 ns 34759
BM_sendVec_binderize/512 45743 ns 20211 ns 34489
BM_sendVec_binderize/1024 46169 ns 20430 ns 34081
After patch:
------------------------------------------------------------------
Benchmark Time CPU Iterations
------------------------------------------------------------------
BM_sendVec_binderize/4 42939 ns 17262 ns 40653
BM_sendVec_binderize/8 42823 ns 17243 ns 40671
BM_sendVec_binderize/16 42898 ns 17243 ns 40594
BM_sendVec_binderize/32 42838 ns 17267 ns 40527
BM_sendVec_binderize/64 42854 ns 17249 ns 40379
BM_sendVec_binderize/128 42881 ns 17288 ns 40427
BM_sendVec_binderize/256 42917 ns 17297 ns 40429
BM_sendVec_binderize/512 43184 ns 17395 ns 40411
BM_sendVec_binderize/1024 43119 ns 17357 ns 40432
Signed-off-by: Martijn Coenen <maco@android.com>
Change-Id: Ia70287066d62aba64e98ac44ff1214e37ca75693
(cherry picked from commit 82d24d114f)
Commit a1d5ebaf8c ("arm64: big-endian: don't treat code as data when
copying sigret code") moved the 32-bit sigreturn trampoline code from
the aarch32_sigret_code array to kuser32.S. The commit removed the
array definition from signal32.c, but not its declaration in
signal32.h. Remove the leftover declaration.
Signed-off-by: Kevin Brodsky <kevin.brodsky@arm.com>
Signed-off-by: Mark Salyzyn <salyzyn@android.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Bug: 20045882
Bug: 63737556
Change-Id: Ic8a5f0e367f0ecd5c5ddd9e3885d0285f91cf89e
The util returned from group_max_util is not capped at the max util
present in the group, so it can be larger than the capacity stored in
the array. Ensure that when this happens, we always use the last entry
in the array to fetch energy from.
Tested with synthetics on Juno board.
Bug: 38159576
Change-Id: I89fb52fb7e68fa3e682e308acc232596672d03f7
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
The code is to get the CPU util by accumulate different scheduling
classes and when the total util value is larger than CPU capacity
then it clamps util to CPU maximum capacity. So we can get correct util
value when use PELT signal but if with WALT signal it misses to clamp
util value.
On the other hand, WALT doesn't accumulate different class utilization
but it needs to applying boost margin for WALT signal the CPU util
value is possible to be larger than CPU capacity; so this patch is to
always clamp util to CPU maximum capacity.
Change-Id: I05481ddbf20246bb9be15b6bd21b6ec039015ea8
Signed-off-by: Leo Yan <leo.yan@linaro.org>
When we convert capacity into frequency, we used policy->max to get
the max freq of the cpu. Since this can be changed by userspace policy
or thermal events, we are potentially asking for a lower frequency
than the utilization demands.
Change over to using cpuinfo.max which is the max freq supported by
that cpu rather than the currently-chosen max. Frequency granted still
honours the max policy.
Tested by setting a userspace policy and observing the relevant vars
in a trace. In this instance, we ask for around 1ghz instead of 620MHz.
freq_new=1013512
unfixed_freq_new=624487
capacity=546
cpuinfo_max=1900800
policy_max=1171200
Change-Id: I8c5694db42243c6fb78bb9be9046b06ac81295e7
Signed-off-by: Chris Redpath <chris.redpath@arm.com>
Changes in 4.4.94
percpu: make this_cpu_generic_read() atomic w.r.t. interrupts
drm/dp/mst: save vcpi with payloads
MIPS: Fix minimum alignment requirement of IRQ stack
sctp: potential read out of bounds in sctp_ulpevent_type_enabled()
bpf/verifier: reject BPF_ALU64|BPF_END
udpv6: Fix the checksum computation when HW checksum does not apply
ip6_gre: skb_push ipv6hdr before packing the header in ip6gre_header
net: emac: Fix napi poll list corruption
packet: hold bind lock when rebinding to fanout hook
bpf: one perf event close won't free bpf program attached by another perf event
isdn/i4l: fetch the ppp_write buffer in one shot
vti: fix use after free in vti_tunnel_xmit/vti6_tnl_xmit
l2tp: Avoid schedule while atomic in exit_net
l2tp: fix race condition in l2tp_tunnel_delete
tun: bail out from tun_get_user() if the skb is empty
packet: in packet_do_bind, test fanout with bind_lock held
packet: only test po->has_vnet_hdr once in packet_snd
net: Set sk_prot_creator when cloning sockets to the right proto
tipc: use only positive error codes in messages
Revert "bsg-lib: don't free job in bsg_prepare_job"
locking/lockdep: Add nest_lock integrity test
watchdog: kempld: fix gcc-4.3 build
irqchip/crossbar: Fix incorrect type of local variables
mac80211_hwsim: check HWSIM_ATTR_RADIO_NAME length
mac80211: fix power saving clients handling in iwlwifi
net/mlx4_en: fix overflow in mlx4_en_init_timestamp()
netfilter: nf_ct_expect: Change __nf_ct_expect_check() return value.
iio: adc: xilinx: Fix error handling
Btrfs: send, fix failure to rename top level inode due to name collision
f2fs: do not wait for writeback in write_begin
md/linear: shutup lockdep warnning
sparc64: Migrate hvcons irq to panicked cpu
net/mlx4_core: Fix VF overwrite of module param which disables DMFS on new probed PFs
crypto: xts - Add ECB dependency
ocfs2/dlmglue: prepare tracking logic to avoid recursive cluster lock
slub: do not merge cache if slub_debug contains a never-merge flag
scsi: scsi_dh_emc: return success in clariion_std_inquiry()
net: mvpp2: release reference to txq_cpu[] entry after unmapping
i2c: at91: ensure state is restored after suspending
ceph: clean up unsafe d_parent accesses in build_dentry_path
uapi: fix linux/rds.h userspace compilation errors
uapi: fix linux/mroute6.h userspace compilation errors
target/iscsi: Fix unsolicited data seq_end_offset calculation
nfsd/callback: Cleanup callback cred on shutdown
cpufreq: CPPC: add ACPI_PROCESSOR dependency
Revert "tty: goldfish: Fix a parameter of a call to free_irq"
Linux 4.4.94
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
[ Upstream commit a578884fa0 ]
Without the Kconfig dependency, we can get this warning:
warning: ACPI_CPPC_CPUFREQ selects ACPI_CPPC_LIB which has unmet direct dependencies (ACPI && ACPI_PROCESSOR)
Fixes: 5477fb3bd1 (ACPI / CPPC: Add a CPUFreq driver for use with CPPC)
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Sasha Levin <alexander.levin@verizon.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 4d65491c26 ]
In case of unsolicited data for the first sequence
seq_end_offset must be set to minimum of total data length
and FirstBurstLength, so do not add cmd->write_data_done
to the min of total data length and FirstBurstLength.
This patch avoids that with ImmediateData=Yes, InitialR2T=No,
MaxXmitDataSegmentLength < FirstBurstLength that a WRITE command
with IO size above FirstBurstLength triggers sequence error
messages, for example
Set following parameters on target (linux-4.8.12)
ImmediateData = Yes
InitialR2T = No
MaxXmitDataSegmentLength = 8k
FirstBurstLength = 64k
Log in from Open iSCSI initiator and execute
dd if=/dev/zero of=/dev/sdb bs=128k count=1 oflag=direct
Error messages on target
Command ITT: 0x00000035 with Offset: 65536, Length: 8192 outside
of Sequence 73728:131072 while DataSequenceInOrder=Yes.
Command ITT: 0x00000035, received DataSN: 0x00000001 higher than
expected 0x00000000.
Unable to perform within-command recovery while ERL=0.
Signed-off-by: Varun Prakash <varun@chelsio.com>
[ bvanassche: Use min() instead of open-coding it / edited patch description ]
Signed-off-by: Bart Van Assche <bart.vanassche@sandisk.com>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
Signed-off-by: Sasha Levin <alexander.levin@verizon.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 72aa107df6 ]
Include <linux/in6.h> to fix the following linux/mroute6.h userspace
compilation errors:
/usr/include/linux/mroute6.h:80:22: error: field 'mf6cc_origin' has incomplete type
struct sockaddr_in6 mf6cc_origin; /* Origin of mcast */
/usr/include/linux/mroute6.h:81:22: error: field 'mf6cc_mcastgrp' has incomplete type
struct sockaddr_in6 mf6cc_mcastgrp; /* Group in question */
/usr/include/linux/mroute6.h:91:22: error: field 'src' has incomplete type
struct sockaddr_in6 src;
/usr/include/linux/mroute6.h:92:22: error: field 'grp' has incomplete type
struct sockaddr_in6 grp;
/usr/include/linux/mroute6.h:132:18: error: field 'im6_src' has incomplete type
struct in6_addr im6_src, im6_dst;
/usr/include/linux/mroute6.h:132:27: error: field 'im6_dst' has incomplete type
struct in6_addr im6_src, im6_dst;
Signed-off-by: Dmitry V. Levin <ldv@altlinux.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <alexander.levin@verizon.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit feb0869d90 ]
Consistently use types from linux/types.h to fix the following
linux/rds.h userspace compilation errors:
/usr/include/linux/rds.h:106:2: error: unknown type name 'uint8_t'
uint8_t name[32];
/usr/include/linux/rds.h:107:2: error: unknown type name 'uint64_t'
uint64_t value;
/usr/include/linux/rds.h:117:2: error: unknown type name 'uint64_t'
uint64_t next_tx_seq;
/usr/include/linux/rds.h:118:2: error: unknown type name 'uint64_t'
uint64_t next_rx_seq;
/usr/include/linux/rds.h:121:2: error: unknown type name 'uint8_t'
uint8_t transport[TRANSNAMSIZ]; /* null term ascii */
/usr/include/linux/rds.h:122:2: error: unknown type name 'uint8_t'
uint8_t flags;
/usr/include/linux/rds.h:129:2: error: unknown type name 'uint64_t'
uint64_t seq;
/usr/include/linux/rds.h:130:2: error: unknown type name 'uint32_t'
uint32_t len;
/usr/include/linux/rds.h:135:2: error: unknown type name 'uint8_t'
uint8_t flags;
/usr/include/linux/rds.h:139:2: error: unknown type name 'uint32_t'
uint32_t sndbuf;
/usr/include/linux/rds.h:144:2: error: unknown type name 'uint32_t'
uint32_t rcvbuf;
/usr/include/linux/rds.h:145:2: error: unknown type name 'uint64_t'
uint64_t inum;
/usr/include/linux/rds.h:153:2: error: unknown type name 'uint64_t'
uint64_t hdr_rem;
/usr/include/linux/rds.h:154:2: error: unknown type name 'uint64_t'
uint64_t data_rem;
/usr/include/linux/rds.h:155:2: error: unknown type name 'uint32_t'
uint32_t last_sent_nxt;
/usr/include/linux/rds.h:156:2: error: unknown type name 'uint32_t'
uint32_t last_expected_una;
/usr/include/linux/rds.h:157:2: error: unknown type name 'uint32_t'
uint32_t last_seen_una;
/usr/include/linux/rds.h:164:2: error: unknown type name 'uint8_t'
uint8_t src_gid[RDS_IB_GID_LEN];
/usr/include/linux/rds.h:165:2: error: unknown type name 'uint8_t'
uint8_t dst_gid[RDS_IB_GID_LEN];
/usr/include/linux/rds.h:167:2: error: unknown type name 'uint32_t'
uint32_t max_send_wr;
/usr/include/linux/rds.h:168:2: error: unknown type name 'uint32_t'
uint32_t max_recv_wr;
/usr/include/linux/rds.h:169:2: error: unknown type name 'uint32_t'
uint32_t max_send_sge;
/usr/include/linux/rds.h:170:2: error: unknown type name 'uint32_t'
uint32_t rdma_mr_max;
/usr/include/linux/rds.h:171:2: error: unknown type name 'uint32_t'
uint32_t rdma_mr_size;
/usr/include/linux/rds.h:212:9: error: unknown type name 'uint64_t'
typedef uint64_t rds_rdma_cookie_t;
/usr/include/linux/rds.h:215:2: error: unknown type name 'uint64_t'
uint64_t addr;
/usr/include/linux/rds.h:216:2: error: unknown type name 'uint64_t'
uint64_t bytes;
/usr/include/linux/rds.h:221:2: error: unknown type name 'uint64_t'
uint64_t cookie_addr;
/usr/include/linux/rds.h:222:2: error: unknown type name 'uint64_t'
uint64_t flags;
/usr/include/linux/rds.h:228:2: error: unknown type name 'uint64_t'
uint64_t cookie_addr;
/usr/include/linux/rds.h:229:2: error: unknown type name 'uint64_t'
uint64_t flags;
/usr/include/linux/rds.h:234:2: error: unknown type name 'uint64_t'
uint64_t flags;
/usr/include/linux/rds.h:240:2: error: unknown type name 'uint64_t'
uint64_t local_vec_addr;
/usr/include/linux/rds.h:241:2: error: unknown type name 'uint64_t'
uint64_t nr_local;
/usr/include/linux/rds.h:242:2: error: unknown type name 'uint64_t'
uint64_t flags;
/usr/include/linux/rds.h:243:2: error: unknown type name 'uint64_t'
uint64_t user_token;
/usr/include/linux/rds.h:248:2: error: unknown type name 'uint64_t'
uint64_t local_addr;
/usr/include/linux/rds.h:249:2: error: unknown type name 'uint64_t'
uint64_t remote_addr;
/usr/include/linux/rds.h:252:4: error: unknown type name 'uint64_t'
uint64_t compare;
/usr/include/linux/rds.h:253:4: error: unknown type name 'uint64_t'
uint64_t swap;
/usr/include/linux/rds.h:256:4: error: unknown type name 'uint64_t'
uint64_t add;
/usr/include/linux/rds.h:259:4: error: unknown type name 'uint64_t'
uint64_t compare;
/usr/include/linux/rds.h:260:4: error: unknown type name 'uint64_t'
uint64_t swap;
/usr/include/linux/rds.h:261:4: error: unknown type name 'uint64_t'
uint64_t compare_mask;
/usr/include/linux/rds.h:262:4: error: unknown type name 'uint64_t'
uint64_t swap_mask;
/usr/include/linux/rds.h:265:4: error: unknown type name 'uint64_t'
uint64_t add;
/usr/include/linux/rds.h:266:4: error: unknown type name 'uint64_t'
uint64_t nocarry_mask;
/usr/include/linux/rds.h:269:2: error: unknown type name 'uint64_t'
uint64_t flags;
/usr/include/linux/rds.h:270:2: error: unknown type name 'uint64_t'
uint64_t user_token;
/usr/include/linux/rds.h:274:2: error: unknown type name 'uint64_t'
uint64_t user_token;
/usr/include/linux/rds.h:275:2: error: unknown type name 'int32_t'
int32_t status;
Signed-off-by: Dmitry V. Levin <ldv@altlinux.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <alexander.levin@verizon.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit c6b0b656ca ]
While we hold a reference to the dentry when build_dentry_path is
called, we could end up racing with a rename that changes d_parent.
Handle that situation correctly, by using the rcu_read_lock to
ensure that the parent dentry and inode stick around long enough
to safely check ceph_snap and ceph_ino.
Link: http://tracker.ceph.com/issues/18148
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Reviewed-by: Yan, Zheng <zyan@redhat.com>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Signed-off-by: Sasha Levin <alexander.levin@verizon.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 36fb7435b6 ]
The mvpp2_txq_bufs_free() function is called upon TX completion to DMA
unmap TX buffers, and free the corresponding SKBs. It gets the
references to the SKB to free and the DMA buffer to unmap from a per-CPU
txq_pcpu data structure.
However, the code currently increments the pointer to the next entry
before doing the DMA unmap and freeing the SKB. It does not cause any
visible problem because for a given SKB the TX completion is guaranteed
to take place on the CPU where the TX was started. However, it is much
more logical to increment the pointer to the next entry once the current
entry has been completely unmapped/released.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: Russell King <rmk+kernel@armlinux.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <alexander.levin@verizon.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 439a36b8ef ]
We are in the situation that we have to avoid recursive cluster locking,
but there is no way to check if a cluster lock has been taken by a precess
already.
Mostly, we can avoid recursive locking by writing code carefully.
However, we found that it's very hard to handle the routines that are
invoked directly by vfs code. For instance:
const struct inode_operations ocfs2_file_iops = {
.permission = ocfs2_permission,
.get_acl = ocfs2_iop_get_acl,
.set_acl = ocfs2_iop_set_acl,
};
Both ocfs2_permission() and ocfs2_iop_get_acl() call ocfs2_inode_lock(PR):
do_sys_open
may_open
inode_permission
ocfs2_permission
ocfs2_inode_lock() <=== first time
generic_permission
get_acl
ocfs2_iop_get_acl
ocfs2_inode_lock() <=== recursive one
A deadlock will occur if a remote EX request comes in between two of
ocfs2_inode_lock(). Briefly describe how the deadlock is formed:
On one hand, OCFS2_LOCK_BLOCKED flag of this lockres is set in
BAST(ocfs2_generic_handle_bast) when downconvert is started on behalf of
the remote EX lock request. Another hand, the recursive cluster lock
(the second one) will be blocked in in __ocfs2_cluster_lock() because of
OCFS2_LOCK_BLOCKED. But, the downconvert never complete, why? because
there is no chance for the first cluster lock on this node to be
unlocked - we block ourselves in the code path.
The idea to fix this issue is mostly taken from gfs2 code.
1. introduce a new field: struct ocfs2_lock_res.l_holders, to keep track
of the processes' pid who has taken the cluster lock of this lock
resource;
2. introduce a new flag for ocfs2_inode_lock_full:
OCFS2_META_LOCK_GETBH; it means just getting back disk inode bh for
us if we've got cluster lock.
3. export a helper: ocfs2_is_locked_by_me() is used to check if we have
got the cluster lock in the upper code path.
The tracking logic should be used by some of the ocfs2 vfs's callbacks,
to solve the recursive locking issue cuased by the fact that vfs
routines can call into each other.
The performance penalty of processing the holder list should only be
seen at a few cases where the tracking logic is used, such as get/set
acl.
You may ask what if the first time we got a PR lock, and the second time
we want a EX lock? fortunately, this case never happens in the real
world, as far as I can see, including permission check,
(get|set)_(acl|attr), and the gfs2 code also do so.
[sfr@canb.auug.org.au remove some inlines]
Link: http://lkml.kernel.org/r/20170117100948.11657-2-zren@suse.com
Signed-off-by: Eric Ren <zren@suse.com>
Reviewed-by: Junxiao Bi <junxiao.bi@oracle.com>
Reviewed-by: Joseph Qi <jiangqi903@gmail.com>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>
Cc: Mark Fasheh <mfasheh@versity.com>
Cc: Joel Becker <jlbec@evilplan.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Sasha Levin <alexander.levin@verizon.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 95f1ba9a24 ]
In the VF driver, module parameter mlx4_log_num_mgm_entry_size was
mistakenly overwritten -- and in a manner which overrode the
device-managed flow steering option encoded in the parameter.
log_num_mgm_entry_size is a global module parameter which
affects all ConnectX-3 PFs installed on that host.
If a VF changes log_num_mgm_entry_size, this will affect all PFs
which are probed subsequent to the change (by disabling DMFS for
those PFs).
Fixes: 3c439b5586 ("mlx4_core: Allow choosing flow steering mode")
Signed-off-by: Majd Dibbiny <majd@mellanox.com>
Reviewed-by: Jack Morgenstein <jackm@dev.mellanox.co.il>
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <alexander.levin@verizon.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 4dd9920d99 ]
Under certain situations, an incremental send operation can fail due to a
premature attempt to create a new top level inode (a direct child of the
subvolume/snapshot root) whose name collides with another inode that was
removed from the send snapshot.
Consider the following example scenario.
Parent snapshot:
. (ino 256, gen 8)
|---- a1/ (ino 257, gen 9)
|---- a2/ (ino 258, gen 9)
Send snapshot:
. (ino 256, gen 3)
|---- a2/ (ino 257, gen 7)
In this scenario, when receiving the incremental send stream, the btrfs
receive command fails like this (ran in verbose mode, -vv argument):
rmdir a1
mkfile o257-7-0
rename o257-7-0 -> a2
ERROR: rename o257-7-0 -> a2 failed: Is a directory
What happens when computing the incremental send stream is:
1) An operation to remove the directory with inode number 257 and
generation 9 is issued.
2) An operation to create the inode with number 257 and generation 7 is
issued. This creates the inode with an orphanized name of "o257-7-0".
3) An operation rename the new inode 257 to its final name, "a2", is
issued. This is incorrect because inode 258, which has the same name
and it's a child of the same parent (root inode 256), was not yet
processed and therefore no rmdir operation for it was yet issued.
The rename operation is issued because we fail to detect that the
name of the new inode 257 collides with inode 258, because their
parent, a subvolume/snapshot root (inode 256) has a different
generation in both snapshots.
So fix this by ignoring the generation value of a parent directory that
matches a root inode (number 256) when we are checking if the name of the
inode currently being processed collides with the name of some other
inode that was not yet processed.
We can achieve this scenario of different inodes with the same number but
different generation values either by mounting a filesystem with the inode
cache option (-o inode_cache) or by creating and sending snapshots across
different filesystems, like in the following example:
$ mkfs.btrfs -f /dev/sdb
$ mount /dev/sdb /mnt
$ mkdir /mnt/a1
$ mkdir /mnt/a2
$ btrfs subvolume snapshot -r /mnt /mnt/snap1
$ btrfs send /mnt/snap1 -f /tmp/1.snap
$ umount /mnt
$ mkfs.btrfs -f /dev/sdc
$ mount /dev/sdc /mnt
$ touch /mnt/a2
$ btrfs subvolume snapshot -r /mnt /mnt/snap2
$ btrfs receive /mnt -f /tmp/1.snap
# Take note that once the filesystem is created, its current
# generation has value 7 so the inode from the second snapshot has
# a generation value of 7. And after receiving the first snapshot
# the filesystem is at a generation value of 10, because the call to
# create the second snapshot bumps the generation to 8 (the snapshot
# creation ioctl does a transaction commit), the receive command calls
# the snapshot creation ioctl to create the first snapshot, which bumps
# the filesystem's generation to 9, and finally when the receive
# operation finishes it calls an ioctl to transition the first snapshot
# (snap1) from RW mode to RO mode, which does another transaction commit
# and bumps the filesystem's generation to 10.
$ rm -f /tmp/1.snap
$ btrfs send /mnt/snap1 -f /tmp/1.snap
$ btrfs send -p /mnt/snap1 /mnt/snap2 -f /tmp/2.snap
$ umount /mnt
$ mkfs.btrfs -f /dev/sdd
$ mount /dev/sdd /mnt
$ btrfs receive /mnt /tmp/1.snap
# Receive of snapshot snap2 used to fail.
$ btrfs receive /mnt /tmp/2.snap
Signed-off-by: Robbie Ko <robbieko@synology.com>
Reviewed-by: Filipe Manana <fdmanana@suse.com>
[Rewrote changelog to be more precise and clear]
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Sasha Levin <alexander.levin@verizon.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 4b86c459c7 ]
Commit 4dee62b1b9 ("netfilter: nf_ct_expect: nf_ct_expect_insert()
returns void") inadvertently changed the successful return value of
nf_ct_expect_related_report() from 0 to 1 due to
__nf_ct_expect_check() returning 1 on success. Prevent this
regression in the future by changing the return value of
__nf_ct_expect_check() to 0 on success.
Signed-off-by: Jarno Rajahalme <jarno@ovn.org>
Acked-by: Joe Stringer <joe@ovn.org>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Sasha Levin <alexander.levin@verizon.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 47d3a07528 ]
The cited commit makes a great job of finding optimal shift/multiplier
values assuming a 10 seconds wrap around, but forgot to change the
overflow_period computation.
It overflows in cyclecounter_cyc2ns(), and the final result is 804 ms,
which is silly.
Lets simply use 5 seconds, no need to recompute this, given how it is
supposed to work.
Later, we will use a timer instead of a work queue, since the new RX
allocation schem will no longer need mlx4_en_recover_from_oom() and the
service_task firing every 250 ms.
Fixes: 31c128b66e ("net/mlx4_en: Choose time-stamping shift value according to HW frequency")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Tariq Toukan <tariqt@mellanox.com>
Cc: Eugenia Emantayev <eugenia@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <alexander.levin@verizon.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit d98937f4ea ]
iwlwifi now supports RSS and can't let mac80211 track the
PS state based on the Rx frames since they can come out of
order. iwlwifi is now advertising AP_LINK_PS, and uses
explicit notifications to teach mac80211 about the PS state
of the stations and the PS poll / uAPSD trigger frames
coming our way from the peers.
Because of that, the TIM stopped being maintained in
mac80211. I tried to fix this in commit c68df2e7be
("mac80211: allow using AP_LINK_PS with mac80211-generated TIM IE")
but that was later reverted by Felix in commit 6c18a6b4e7
("Revert "mac80211: allow using AP_LINK_PS with mac80211-generated TIM IE")
since it broke drivers that do not implement set_tim.
Since none of the drivers that set AP_LINK_PS have the
set_tim() handler set besides iwlwifi, I can bail out in
__sta_info_recalc_tim if AP_LINK_PS AND .set_tim is not
implemented.
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Sasha Levin <alexander.levin@verizon.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit ff4dd73dd2 ]
Unfortunately, the nla policy was defined to have HWSIM_ATTR_RADIO_NAME
as an NLA_STRING, rather than NLA_NUL_STRING, so we can't use it as a
NUL-terminated string in the kernel.
Rather than break the API, kasprintf() the string to a new buffer to
guarantee NUL termination.
Reported-by: Andrew Zaborowski <andrew.zaborowski@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Sasha Levin <alexander.levin@verizon.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit b28ace1266 ]
The max and entry variables are unsigned according to the dt-bindings.
Fix following 3 sparse issues (-Wtypesign):
drivers/irqchip/irq-crossbar.c:222:52: warning: incorrect type in argument 3 (different signedness)
drivers/irqchip/irq-crossbar.c:222:52: expected unsigned int [usertype] *out_value
drivers/irqchip/irq-crossbar.c:222:52: got int *<noident>
drivers/irqchip/irq-crossbar.c:245:56: warning: incorrect type in argument 4 (different signedness)
drivers/irqchip/irq-crossbar.c:245:56: expected unsigned int [usertype] *out_value
drivers/irqchip/irq-crossbar.c:245:56: got int *<noident>
drivers/irqchip/irq-crossbar.c:263:56: warning: incorrect type in argument 4 (different signedness)
drivers/irqchip/irq-crossbar.c:263:56: expected unsigned int [usertype] *out_value
drivers/irqchip/irq-crossbar.c:263:56: got int *<noident>
Signed-off-by: Franck Demathieu <fdemathieu@gmail.com>
Cc: marc.zyngier@arm.com
Cc: jason@lakedaemon.net
Link: http://lkml.kernel.org/r/20170223094855.6546-1-fdemathieu@gmail.com
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Sasha Levin <alexander.levin@verizon.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 3736d4eb6a ]
gcc-4.3 can't decide whether the constant value in
kempld_prescaler[PRESCALER_21] is built-time constant or
not, and gets confused by the logic in do_div():
drivers/watchdog/kempld_wdt.o: In function `kempld_wdt_set_stage_timeout':
kempld_wdt.c:(.text.kempld_wdt_set_stage_timeout+0x130): undefined reference to `__aeabi_uldivmod'
This adds a call to ACCESS_ONCE() to force it to not consider
it to be constant, and leaves the more efficient normal case
in place for modern compilers, using an #ifdef to annotate
why we do this hack.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Sasha Levin <alexander.levin@verizon.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This reverts commit 668cee82cd which was
commit f507b54dcc upstream.
Ben reports:
That function doesn't exist here (it was introduced in 4.13).
Instead, this backport has modified bsg_create_job(), creating a
leak. Please revert this on the 3.18, 4.4 and 4.9 stable
branches.
So I'm dropping it from here.
Reported-by: Ben Hutchings <ben.hutchings@codethink.co.uk>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Ming Lei <ming.lei@redhat.com>
Cc: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org
[ Upstream commit aad06212d3 ]
In commit e3a77561e7 ("tipc: split up function tipc_msg_eval()"),
we have updated the function tipc_msg_lookup_dest() to set the error
codes to negative values at destination lookup failures. Thus when
the function sets the error code to -TIPC_ERR_NO_NAME, its inserted
into the 4 bit error field of the message header as 0xf instead of
TIPC_ERR_NO_NAME (1). The value 0xf is an unknown error code.
In this commit, we set only positive error code.
Fixes: e3a77561e7 ("tipc: split up function tipc_msg_eval()")
Signed-off-by: Parthasarathy Bhuvaragan <parthasarathy.bhuvaragan@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 9d538fa60b ]
sk->sk_prot and sk->sk_prot_creator can differ when the app uses
IPV6_ADDRFORM (transforming an IPv6-socket to an IPv4-one).
Which is why sk_prot_creator is there to make sure that sk_prot_free()
does the kmem_cache_free() on the right kmem_cache slab.
Now, if such a socket gets transformed back to a listening socket (using
connect() with AF_UNSPEC) we will allocate an IPv4 tcp_sock through
sk_clone_lock() when a new connection comes in. But sk_prot_creator will
still point to the IPv6 kmem_cache (as everything got copied in
sk_clone_lock()). When freeing, we will thus put this
memory back into the IPv6 kmem_cache although it was allocated in the
IPv4 cache. I have seen memory corruption happening because of this.
With slub-debugging and MEMCG_KMEM enabled this gives the warning
"cache_from_obj: Wrong slab cache. TCPv6 but object is from TCP"
A C-program to trigger this:
void main(void)
{
int fd = socket(AF_INET6, SOCK_STREAM, IPPROTO_TCP);
int new_fd, newest_fd, client_fd;
struct sockaddr_in6 bind_addr;
struct sockaddr_in bind_addr4, client_addr1, client_addr2;
struct sockaddr unsp;
int val;
memset(&bind_addr, 0, sizeof(bind_addr));
bind_addr.sin6_family = AF_INET6;
bind_addr.sin6_port = ntohs(42424);
memset(&client_addr1, 0, sizeof(client_addr1));
client_addr1.sin_family = AF_INET;
client_addr1.sin_port = ntohs(42424);
client_addr1.sin_addr.s_addr = inet_addr("127.0.0.1");
memset(&client_addr2, 0, sizeof(client_addr2));
client_addr2.sin_family = AF_INET;
client_addr2.sin_port = ntohs(42421);
client_addr2.sin_addr.s_addr = inet_addr("127.0.0.1");
memset(&unsp, 0, sizeof(unsp));
unsp.sa_family = AF_UNSPEC;
bind(fd, (struct sockaddr *)&bind_addr, sizeof(bind_addr));
listen(fd, 5);
client_fd = socket(AF_INET, SOCK_STREAM, IPPROTO_TCP);
connect(client_fd, (struct sockaddr *)&client_addr1, sizeof(client_addr1));
new_fd = accept(fd, NULL, NULL);
close(fd);
val = AF_INET;
setsockopt(new_fd, SOL_IPV6, IPV6_ADDRFORM, &val, sizeof(val));
connect(new_fd, &unsp, sizeof(unsp));
memset(&bind_addr4, 0, sizeof(bind_addr4));
bind_addr4.sin_family = AF_INET;
bind_addr4.sin_port = ntohs(42421);
bind(new_fd, (struct sockaddr *)&bind_addr4, sizeof(bind_addr4));
listen(new_fd, 5);
client_fd = socket(AF_INET, SOCK_STREAM, IPPROTO_TCP);
connect(client_fd, (struct sockaddr *)&client_addr2, sizeof(client_addr2));
newest_fd = accept(new_fd, NULL, NULL);
close(new_fd);
close(client_fd);
close(new_fd);
}
As far as I can see, this bug has been there since the beginning of the
git-days.
Signed-off-by: Christoph Paasch <cpaasch@apple.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit da7c956101 ]
Packet socket option po->has_vnet_hdr can be updated concurrently with
other operations if no ring is attached.
Do not test the option twice in packet_snd, as the value may change in
between calls. A race on setsockopt disable may cause a packet > mtu
to be sent without having GSO options set.
Fixes: bfd5f4a3d6 ("packet: Add GSO/csum offload support.")
Signed-off-by: Willem de Bruijn <willemb@google.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 4971613c16 ]
Once a socket has po->fanout set, it remains a member of the group
until it is destroyed. The prot_hook must be constant and identical
across sockets in the group.
If fanout_add races with packet_do_bind between the test of po->fanout
and taking the lock, the bind call may make type or dev inconsistent
with that of the fanout group.
Hold po->bind_lock when testing po->fanout to avoid this race.
I had to introduce artificial delay (local_bh_enable) to actually
observe the race.
Fixes: dc99f60069 ("packet: Add fanout support.")
Signed-off-by: Willem de Bruijn <willemb@google.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 62b982eeb4 ]
If we try to delete the same tunnel twice, the first delete operation
does a lookup (l2tp_tunnel_get), finds the tunnel, calls
l2tp_tunnel_delete, which queues it for deletion by
l2tp_tunnel_del_work.
The second delete operation also finds the tunnel and calls
l2tp_tunnel_delete. If the workqueue has already fired and started
running l2tp_tunnel_del_work, then l2tp_tunnel_delete will queue the
same tunnel a second time, and try to free the socket again.
Add a dead flag to prevent firing the workqueue twice. Then we can
remove the check of queue_work's result that was meant to prevent that
race but doesn't.
Reproducer:
ip l2tp add tunnel tunnel_id 3000 peer_tunnel_id 4000 local 192.168.0.2 remote 192.168.0.1 encap udp udp_sport 5000 udp_dport 6000
ip l2tp add session name l2tp1 tunnel_id 3000 session_id 1000 peer_session_id 2000
ip link set l2tp1 up
ip l2tp del tunnel tunnel_id 3000
ip l2tp del tunnel tunnel_id 3000
Fixes: f8ccac0e44 ("l2tp: put tunnel socket release on a workqueue")
Reported-by: Jianlin Shi <jishi@redhat.com>
Signed-off-by: Sabrina Dubroca <sd@queasysnail.net>
Acked-by: Guillaume Nault <g.nault@alphalink.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 12d656af4e ]
While destroying a network namespace that contains a L2TP tunnel a
"BUG: scheduling while atomic" can be observed.
Enabling lockdep shows that this is happening because l2tp_exit_net()
is calling l2tp_tunnel_closeall() (via l2tp_tunnel_delete()) from
within an RCU critical section.
l2tp_exit_net() takes rcu_read_lock_bh()
<< list_for_each_entry_rcu() >>
l2tp_tunnel_delete()
l2tp_tunnel_closeall()
__l2tp_session_unhash()
synchronize_rcu() << Illegal inside RCU critical section >>
BUG: sleeping function called from invalid context
in_atomic(): 1, irqs_disabled(): 0, pid: 86, name: kworker/u16:2
INFO: lockdep is turned off.
CPU: 2 PID: 86 Comm: kworker/u16:2 Tainted: G W O 4.4.6-at1 #2
Hardware name: Xen HVM domU, BIOS 4.6.1-xs125300 05/09/2016
Workqueue: netns cleanup_net
0000000000000000 ffff880202417b90 ffffffff812b0013 ffff880202410ac0
ffffffff81870de8 ffff880202417bb8 ffffffff8107aee8 ffffffff81870de8
0000000000000c51 0000000000000000 ffff880202417be0 ffffffff8107b024
Call Trace:
[<ffffffff812b0013>] dump_stack+0x85/0xc2
[<ffffffff8107aee8>] ___might_sleep+0x148/0x240
[<ffffffff8107b024>] __might_sleep+0x44/0x80
[<ffffffff810b21bd>] synchronize_sched+0x2d/0xe0
[<ffffffff8109be6d>] ? trace_hardirqs_on+0xd/0x10
[<ffffffff8105c7bb>] ? __local_bh_enable_ip+0x6b/0xc0
[<ffffffff816a1b00>] ? _raw_spin_unlock_bh+0x30/0x40
[<ffffffff81667482>] __l2tp_session_unhash+0x172/0x220
[<ffffffff81667397>] ? __l2tp_session_unhash+0x87/0x220
[<ffffffff8166888b>] l2tp_tunnel_closeall+0x9b/0x140
[<ffffffff81668c74>] l2tp_tunnel_delete+0x14/0x60
[<ffffffff81668dd0>] l2tp_exit_net+0x110/0x270
[<ffffffff81668d5c>] ? l2tp_exit_net+0x9c/0x270
[<ffffffff815001c3>] ops_exit_list.isra.6+0x33/0x60
[<ffffffff81501166>] cleanup_net+0x1b6/0x280
...
This bug can easily be reproduced with a few steps:
$ sudo unshare -n bash # Create a shell in a new namespace
# ip link set lo up
# ip addr add 127.0.0.1 dev lo
# ip l2tp add tunnel remote 127.0.0.1 local 127.0.0.1 tunnel_id 1 \
peer_tunnel_id 1 udp_sport 50000 udp_dport 50000
# ip l2tp add session name foo tunnel_id 1 session_id 1 \
peer_session_id 1
# ip link set foo up
# exit # Exit the shell, in turn exiting the namespace
$ dmesg
...
[942121.089216] BUG: scheduling while atomic: kworker/u16:3/13872/0x00000200
...
To fix this, move the call to l2tp_tunnel_closeall() out of the RCU
critical section, and instead call it from l2tp_tunnel_del_work(), which
is running from the l2tp_wq workqueue.
Fixes: 2b551c6e7d ("l2tp: close sessions before initiating tunnel delete")
Signed-off-by: Ridge Kennedy <ridge.kennedy@alliedtelesis.co.nz>
Acked-by: Guillaume Nault <g.nault@alphalink.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 02388bf87f ]
In isdn_ppp_write(), the header (i.e., protobuf) of the buffer is
fetched twice from userspace. The first fetch is used to peek at the
protocol of the message and reset the huptimer if necessary; while the
second fetch copies in the whole buffer. However, given that buf resides
in userspace memory, a user process can race to change its memory content
across fetches. By doing so, we can either avoid resetting the huptimer
for any type of packets (by first setting proto to PPP_LCP and later
change to the actual type) or force resetting the huptimer for LCP
packets.
This patch changes this double-fetch behavior into two single fetches
decided by condition (lp->isdn_device < 0 || lp->isdn_channel <0).
A more detailed discussion can be found at
https://marc.info/?l=linux-kernel&m=150586376926123&w=2
Signed-off-by: Meng Xu <mengxu.gatech@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
