In OOM situations where no skb can be allocated, can_change_state() may
be called with cf == NULL. As this function updates the state and error
statistics it's not an option to skip the call to can_change_state() in
OOM situations.

This patch makes can_change_state() robust, so that it can be called
with cf == NULL.
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>

Commit 1aa6c4f6 ("net: vrf: Add l3mdev rules on first device create")
adds the l3mdev FIB rule the first time a VRF device is created. However,
it only creates the rule once and only in the namespace the first device
is created - which may not be init_net. Fix by using the net_generic
capability to make the add_fib_rules flag per network namespace.

Fixes: 1aa6c4f6 ("net: vrf: Add l3mdev rules on first device create")
Reported-by: Petr Machata <petrm@mellanox.com>
Signed-off-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

I noticed that test_l4lb was failing in selftests:

  # ./test_progs
  test_pkt_access:PASS:ipv4 77 nsec
  test_pkt_access:PASS:ipv6 44 nsec
  test_xdp:PASS:ipv4 2933 nsec
  test_xdp:PASS:ipv6 1500 nsec
  test_l4lb:PASS:ipv4 377 nsec
  test_l4lb:PASS:ipv6 544 nsec
  test_l4lb:FAIL:stats 6297600000 200000
  test_tcp_estats:PASS: 0 nsec
  Summary: 7 PASSED, 1 FAILED

Tracking down the issue actually revealed that endianness selection
in bpf_endian.h is broken when compiled with clang with bpf target.
test_pkt_access.c, test_l4lb.c is compiled with __BYTE_ORDER as
__BIG_ENDIAN, test_xdp.c as __LITTLE_ENDIAN! test_l4lb noticeably
fails, because the test accounts bytes via bpf_ntohs(ip6h->payload_len)
and bpf_ntohs(iph->tot_len), and compares them against a defined
value and given a wrong endianness, the test outcome is different,
of course.

Turns out that there are actually two bugs: i) when we do __BYTE_ORDER
comparison with __LITTLE_ENDIAN/__BIG_ENDIAN, then depending on the
include order we see different outcomes. Reason is that __BYTE_ORDER
is undefined due to missing endian.h include. Before we include the
asm/byteorder.h (e.g. through linux/in.h), then __BYTE_ORDER equals
__LITTLE_ENDIAN since both are undefined, after the include which
correctly pulls in linux/byteorder/little_endian.h, __LITTLE_ENDIAN
is defined, but given __BYTE_ORDER is still undefined, we match on
__BYTE_ORDER equals to __BIG_ENDIAN since __BIG_ENDIAN is also
undefined at that point, sigh. ii) But even that would be wrong,
since when compiling the test cases with clang, one can select between
bpfeb and bpfel targets for cross compilation. Hence, we can also not
rely on what the system's endian.h provides, but we need to look at
the compiler's defined endianness. The compiler defines __BYTE_ORDER__,
and we can match __ORDER_LITTLE_ENDIAN__ and __ORDER_BIG_ENDIAN__,
which also reflects targets bpf (native), bpfel, bpfeb correctly,
thus really only rely on that. After patch:

  # ./test_progs
  test_pkt_access:PASS:ipv4 74 nsec
  test_pkt_access:PASS:ipv6 42 nsec
  test_xdp:PASS:ipv4 2340 nsec
  test_xdp:PASS:ipv6 1461 nsec
  test_l4lb:PASS:ipv4 400 nsec
  test_l4lb:PASS:ipv6 530 nsec
  test_tcp_estats:PASS: 0 nsec
  Summary: 7 PASSED, 0 FAILED

Fixes: 43bcf707 ("bpf: fix _htons occurences in test_progs")
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Each time a new speed is added, the bonding 802.3ad isn't updated. Add a
comment to remind the developer to update this driver.
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Acked-by: Andy Gospodarek <andy@greyhouse.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch adds 14 Gbps enum definition, and fixes
aggregated bandwidth calculation based on above slave links.

Fixes: 0d7e2d21 ("IB/ipoib: add get_link_ksettings in ethtool")
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Acked-by: Andy Gospodarek <andy@greyhouse.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch adds [5|50] Gbps enum definition, and fixes
aggregated bandwidth calculation based on above slave links.

Fixes: c9a70d43 ("net-next: ethtool: Added port speed macros.")
Signed-off-by: Thibaut Collet <thibaut.collet@6wind.com>
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Acked-by: Andy Gospodarek <andy@greyhouse.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

The first netlink attribute (value 0) must always be defined
as none/unspec.

Because we cannot change an existing UAPI, I add a comment to point the
mistake and avoid to propagate it in a new ovs API in the future.
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

While discussing the possible merits of clang warning about unused initialized
functions, I found one function that was clearly meant to be called but
never actually is.

__ila_hash_secret_init() initializes the hash value for the ila locator,
apparently this is intended to prevent hash collision attacks, but this ends
up being a read-only zero constant since there is no caller. I could find
no indication of why it was never called, the earliest patch submission
for the module already was like this. If my interpretation is right, we
certainly want to backport the patch to stable kernels as well.

I considered adding it to the ila_xlat_init callback, but for best effect
the random data is read as late as possible, just before it is first used.
The underlying net_get_random_once() is already highly optimized to avoid
overhead when called frequently.

Fixes: 7f00feaf ("ila: Add generic ILA translation facility")
Cc: stable@vger.kernel.org
Link: https://www.spinics.net/lists/kernel/msg2527243.htmlSigned-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/staging/rtl8723bs/os_dep/ioctl_cfg80211.c: In function ‘rtw_cfg80211_add_monitor_if’:
drivers/staging/rtl8723bs/os_dep/ioctl_cfg80211.c:2670:10: error: ‘struct net_device’ has no member named ‘destructor’
  mon_ndev->destructor = rtw_ndev_destructor;
          ^
Signed-off-by: David S. Miller <davem@davemloft.net>

Stephen Hemminger says:

====================
netvsc: bug fixes

These are bugfixes for netvsc driver in 4.12.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

The work queue and handling of network filter parameters should
be in rndis_device. This gets rid of warning from RCU checks,
eliminates a race and cleans up code.
Signed-off-by: Stephen Hemminger <sthemmin@microsoft.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The ndo_poll_controller function needs to schedule NAPI to pick
up arriving packets and send completions. Otherwise no data
will ever be received. For simple case of netconsole, it also
will allow send completions to happen.  Without this netpoll
will eventually get stuck.
Signed-off-by: Stephen Hemminger <sthemmin@microsoft.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The ethtool info command calls the netvsc get_sset_count with RTNL
but not with RCU. Which causes warning:

drivers/net/hyperv/netvsc_drv.c:1010 suspicious rcu_dereference_check() usage!
Signed-off-by: Stephen Hemminger <sthemmin@microsoft.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Roopa reported attempts to delete a bond device that is referenced in a
multipath route is hanging:

$ ifdown bond2    # ifupdown2 command that deletes virtual devices
unregister_netdevice: waiting for bond2 to become free. Usage count = 2

Steps to reproduce:
    echo 1 > /proc/sys/net/ipv6/conf/all/ignore_routes_with_linkdown
    ip link add dev bond12 type bond
    ip link add dev bond13 type bond
    ip addr add 2001:db8:2::0/64 dev bond12
    ip addr add 2001:db8:3::0/64 dev bond13
    ip route add 2001:db8:33::0/64 nexthop via 2001:db8:2::2 nexthop via 2001:db8:3::2
    ip link del dev bond12
    ip link del dev bond13

The root cause is the recent change to keep routes on a linkdown. Update
the check to detect when the device is unregistering and release the
route for that case.

Fixes: a1a22c12 ("net: ipv6: Keep nexthop of multipath route on admin down")
Reported-by: Roopa Prabhu <roopa@cumulusnetworks.com>
Signed-off-by: David Ahern <dsahern@gmail.com>
Acked-by: Roopa Prabhu <roopa@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Some of the structure's fields are not initialized by the
rtnetlink. If driver doesn't set those in ndo_get_vf_config(),
they'd leak memory to user.
Signed-off-by: Yuval Mintz <Yuval.Mintz@cavium.com>
CC: Michal Schmidt <mschmidt@redhat.com>
Reviewed-by: Greg Rose <gvrose8192@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Verify that the length of the socket buffer is sufficient to cover the
nlmsghdr structure before accessing the nlh->nlmsg_len field for further
input sanitization. If the client only supplies 1-3 bytes of data in
sk_buff, then nlh->nlmsg_len remains partially uninitialized and
contains leftover memory from the corresponding kernel allocation.
Operating on such data may result in indeterminate evaluation of the
nlmsg_len < sizeof(*nlh) expression.

The bug was discovered by a runtime instrumentation designed to detect
use of uninitialized memory in the kernel. The patch prevents this and
other similar tools (e.g. KMSAN) from flagging this behavior in the future.
Signed-off-by: Mateusz Jurczyk <mjurczyk@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This reverts commit 85eac2ba.

There is an updated version of this fix which we should
use instead.
Signed-off-by: David S. Miller <davem@davemloft.net>

emac_mdio_read_link() was not copying the requested phy settings
back into the emac driver's own phy api. This has caused a link
speed mismatch issue for the AR8035 as the emac driver kept
trying to connect with 10/100MBps on a 1GBit/s link.

This patch also unifies shared code between emac_setup_aneg()
and emac_mdio_setup_forced(). And furthermore it removes
a chunk of emac_mdio_init_phy(), that was copying the same
data into itself.
Signed-off-by: Christian Lamparter <chunkeey@googlemail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch fixes a problem where the AR8035 PHY can't be
detected on an Cisco Meraki MR24, if the ethernet cable is
not connected on boot.

Russell Senior provided steps to reproduce the issue:
|Disconnect ethernet cable, apply power, wait until device has booted,
|plug in ethernet, check for interfaces, no eth0 is listed.
|
|This appears to be a problem during probing of the AR8035 Phy chip.
|When ethernet has no link, the phy detection fails, and eth0 is not
|created. Plugging ethernet later has no effect, because there is no
|interface as far as the kernel is concerned. The relevant part of
|the boot log looks like this:
|this is the failing case:
|
|[    0.876611] /plb/opb/emac-rgmii@ef601500: input 0 in RGMII mode
|[    0.882532] /plb/opb/ethernet@ef600c00: reset timeout
|[    0.888546] /plb/opb/ethernet@ef600c00: can't find PHY!
|and the succeeding case:
|
|[    0.876672] /plb/opb/emac-rgmii@ef601500: input 0 in RGMII mode
|[    0.883952] eth0: EMAC-0 /plb/opb/ethernet@ef600c00, MAC 00:01:..
|[    0.890822] eth0: found Atheros 8035 Gigabit Ethernet PHY (0x01)

Based on the comment and the commit message of
commit 23fbb5a8 ("emac: Fix EMAC soft reset on 460EX/GT").
This is because the AR8035 PHY doesn't provide the TX Clock,
if the ethernet cable is not attached. This causes the reset
to timeout and the PHY detection code in emac_init_phy() is
unable to detect the AR8035 PHY. As a result, the emac driver
bails out early and the user left with no ethernet.

In order to stay compatible with existing configurations, the driver
tries the current reset approach at first. Only if the first attempt
timed out, it does perform one more retry with the clock temporarily
switched to the internal source for just the duration of the reset.

LEDE-Bug: #687 <https://bugs.lede-project.org/index.php?do=details&task_id=687>

Cc: Chris Blake <chrisrblake93@gmail.com>
Reported-by: Russell Senior <russell@personaltelco.net>
Fixes: 23fbb5a8 ("emac: Fix EMAC soft reset on 460EX/GT")
Signed-off-by: Christian Lamparter <chunkeey@googlemail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

Verify that the length of the socket buffer is sufficient to cover the
entire nlh->nlmsg_len field before accessing that field for further
input sanitization. If the client only supplies 1-3 bytes of data in
sk_buff, then nlh->nlmsg_len remains partially uninitialized and
contains leftover memory from the corresponding kernel allocation.
Operating on such data may result in indeterminate evaluation of the
nlmsg_len < sizeof(*nlh) expression.

The bug was discovered by a runtime instrumentation designed to detect
use of uninitialized memory in the kernel. The patch prevents this and
other similar tools (e.g. KMSAN) from flagging this behavior in the future.
Signed-off-by: Mateusz Jurczyk <mjurczyk@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

> ../drivers/hsi/clients/ssi_protocol.c:1069:5: error: 'struct net_device' has no member named 'destructor'
Reported-by: Mark Brown <broonie@kernel.org>
Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>

Network devices can allocate reasources and private memory using
netdev_ops->ndo_init().  However, the release of these resources
can occur in one of two different places.

Either netdev_ops->ndo_uninit() or netdev->destructor().

The decision of which operation frees the resources depends upon
whether it is necessary for all netdev refs to be released before it
is safe to perform the freeing.

netdev_ops->ndo_uninit() presumably can occur right after the
NETDEV_UNREGISTER notifier completes and the unicast and multicast
address lists are flushed.

netdev->destructor(), on the other hand, does not run until the
netdev references all go away.

Further complicating the situation is that netdev->destructor()
almost universally does also a free_netdev().

This creates a problem for the logic in register_netdevice().
Because all callers of register_netdevice() manage the freeing
of the netdev, and invoke free_netdev(dev) if register_netdevice()
fails.

If netdev_ops->ndo_init() succeeds, but something else fails inside
of register_netdevice(), it does call ndo_ops->ndo_uninit().  But
it is not able to invoke netdev->destructor().

This is because netdev->destructor() will do a free_netdev() and
then the caller of register_netdevice() will do the same.

However, this means that the resources that would normally be released
by netdev->destructor() will not be.

Over the years drivers have added local hacks to deal with this, by
invoking their destructor parts by hand when register_netdevice()
fails.

Many drivers do not try to deal with this, and instead we have leaks.

Let's close this hole by formalizing the distinction between what
private things need to be freed up by netdev->destructor() and whether
the driver needs unregister_netdevice() to perform the free_netdev().

netdev->priv_destructor() performs all actions to free up the private
resources that used to be freed by netdev->destructor(), except for
free_netdev().

netdev->needs_free_netdev is a boolean that indicates whether
free_netdev() should be done at the end of unregister_netdevice().

Now, register_netdevice() can sanely release all resources after
ndo_ops->ndo_init() succeeds, by invoking both ndo_ops->ndo_uninit()
and netdev->priv_destructor().

And at the end of unregister_netdevice(), we invoke
netdev->priv_destructor() and optionally call free_netdev().
Signed-off-by: David S. Miller <davem@davemloft.net>

Will reported that in BPF_XADD we must use a different register in stxr
instruction for the status flag due to otherwise CONSTRAINED UNPREDICTABLE
behavior per architecture. Reference manual says [1]:

  If s == t, then one of the following behaviors must occur:

   * The instruction is UNDEFINED.
   * The instruction executes as a NOP.
   * The instruction performs the store to the specified address, but
     the value stored is UNKNOWN.

Thus, use a different temporary register for the status flag to fix it.

Disassembly extract from test 226/STX_XADD_DW from test_bpf.ko:

  [...]
  0000003c:  c85f7d4b  ldxr x11, [x10]
  00000040:  8b07016b  add x11, x11, x7
  00000044:  c80c7d4b  stxr w12, x11, [x10]
  00000048:  35ffffac  cbnz w12, 0x0000003c
  [...]

  [1] https://static.docs.arm.com/ddi0487/b/DDI0487B_a_armv8_arm.pdf, p.6132

Fixes: 85f68fe8 ("bpf, arm64: implement jiting of BPF_XADD")
Reported-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The mvpp22_port_mii_set() function was added by 26975821, but the
function directly returns without doing anything. This return was used
when debugging and wasn't removed before sending the patch. Fix this.

Fixes: 26975821 ("net: mvpp2: handle misc PPv2.1/PPv2.2 differences")
Signed-off-by: Antoine Tenart <antoine.tenart@free-electrons.com>
Acked-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Changing the mtu is currently not supported in the ibmvnic driver.

Implement .ndo_change_mtu in the driver so that attempting to use ifconfig
to change the mtu will fail and present the user with an error message.
Signed-off-by: John Allen <jallen@linux.vnet.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

commit 61b905da ("net: Rename skb->rxhash to skb->hash")
didn't update the documentation, fix this up.

Cc: Tom Herbert <therbert@google.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

When freeing VF's DMA mappings, an already NULLed pointer was checked
again due to an apparent copy&paste error. Consequently, the pf2vf
bulletin DMA mapping was not freed.
Signed-off-by: Michal Schmidt <mschmidt@redhat.com>
Acked-by: Yuval Mintz <Yuval.Mintz@cavium.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

KMSAN reported a use of uninitialized memory in dev_set_alias(),
which was caused by calling strlcpy() (which in turn called strlen())
on the user-supplied non-terminated string.
Signed-off-by: Alexander Potapenko <glider@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Pull networking fixes from David Miller:

 1) Made TCP congestion control documentation match current reality,
    from Anmol Sarma.

 2) Various build warning and failure fixes from Arnd Bergmann.

 3) Fix SKB list leak in ipv6_gso_segment().

 4) Use after free in ravb driver, from Eugeniu Rosca.

 5) Don't use udp_poll() in ping protocol driver, from Eric Dumazet.

 6) Don't crash in PCI error recovery of cxgb4 driver, from Guilherme
    Piccoli.

 7) _SRC_NAT_DONE_BIT needs to be cleared using atomics, from Liping
    Zhang.

 8) Use after free in vxlan deletion, from Mark Bloch.

 9) Fix ordering of NAPI poll enabled in ethoc driver, from Max
    Filippov.

10) Fix stmmac hangs with TSO, from Niklas Cassel.

11) Fix crash in CALIPSO ipv6, from Richard Haines.

12) Clear nh_flags properly on mpls link up. From Roopa Prabhu.

13) Fix regression in sk_err socket error queue handling, noticed by
    ping applications. From Soheil Hassas Yeganeh.

14) Update mlx4/mlx5 MAINTAINERS information.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (78 commits)
  net: stmmac: fix a broken u32 less than zero check
  net: stmmac: fix completely hung TX when using TSO
  net: ethoc: enable NAPI before poll may be scheduled
  net: bridge: fix a null pointer dereference in br_afspec
  ravb: Fix use-after-free on `ifconfig eth0 down`
  net/ipv6: Fix CALIPSO causing GPF with datagram support
  net: stmmac: ensure jumbo_frm error return is correctly checked for -ve value
  Revert "sit: reload iphdr in ipip6_rcv"
  i40e/i40evf: proper update of the page_offset field
  i40e: Fix state flags for bit set and clean operations of PF
  iwlwifi: fix host command memory leaks
  iwlwifi: fix min API version for 7265D, 3168, 8000 and 8265
  iwlwifi: mvm: clear new beacon command template struct
  iwlwifi: mvm: don't fail when removing a key from an inexisting sta
  iwlwifi: pcie: only use d0i3 in suspend/resume if system_pm is set to d0i3
  iwlwifi: mvm: fix firmware debug restart recording
  iwlwifi: tt: move ucode_loaded check under mutex
  iwlwifi: mvm: support ibss in dqa mode
  iwlwifi: mvm: Fix command queue number on d0i3 flow
  iwlwifi: mvm: rs: start using LQ command color
  ...

Pull sparc fixes from David Miller:

 1) Fix TLB context wrap races, from Pavel Tatashin.

 2) Cure some gcc-7 build issues.

 3) Handle invalid setup_hugepagesz command line values properly, from
    Liam R Howlett.

 4) Copy TSB using the correct address shift for the huge TSB, from Mike
    Kravetz.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc:
  sparc64: delete old wrap code
  sparc64: new context wrap
  sparc64: add per-cpu mm of secondary contexts
  sparc64: redefine first version
  sparc64: combine activate_mm and switch_mm
  sparc64: reset mm cpumask after wrap
  sparc/mm/hugepages: Fix setup_hugepagesz for invalid values.
  sparc: Machine description indices can vary
  sparc64: mm: fix copy_tsb to correctly copy huge page TSBs
  arch/sparc: support NR_CPUS = 4096
  sparc64: Add __multi3 for gcc 7.x and later.
  sparc64: Fix build warnings with gcc 7.
  arch/sparc: increase CONFIG_NODES_SHIFT on SPARC64 to 5

GCC explicitly does not warn for unused static inline functions for
-Wunused-function.  The manual states:

	Warn whenever a static function is declared but not defined or
	a non-inline static function is unused.

Clang does warn for static inline functions that are unused.

It turns out that suppressing the warnings avoids potentially complex
#ifdef directives, which also reduces LOC.

Suppress the warning for clang.
Signed-off-by: David Rientjes <rientjes@google.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Pavel Tatashin says:

====================
sparc64: context wrap fixes

This patch series contains fixes for context wrap: when we are out of
context ids, and need to get a new version.

It fixes memory corruption issues which happen when more than number of
context ids (currently set to 8K) number of processes are started
simultaneously, and processes can get a wrong context.

sparc64: new context wrap:
- contains explanation of new wrap method, and also explanation of races
  that it solves
sparc64: reset mm cpumask after wrap
- explains issue of not reseting cpu mask on a wrap
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

The old method that is using xcall and softint to get new context id is
deleted, as it is replaced by a method of using per_cpu_secondary_mm
without xcall to perform the context wrap.
Signed-off-by: Pavel Tatashin <pasha.tatashin@oracle.com>
Reviewed-by: Bob Picco <bob.picco@oracle.com>
Reviewed-by: Steven Sistare <steven.sistare@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The current wrap implementation has a race issue: it is called outside of
the ctx_alloc_lock, and also does not wait for all CPUs to complete the
wrap.  This means that a thread can get a new context with a new version
and another thread might still be running with the same context. The
problem is especially severe on CPUs with shared TLBs, like sun4v. I used
the following test to very quickly reproduce the problem:
- start over 8K processes (must be more than context IDs)
- write and read values at a  memory location in every process.

Very quickly memory corruptions start happening, and what we read back
does not equal what we wrote.

Several approaches were explored before settling on this one:

Approach 1:
Move smp_new_mmu_context_version() inside ctx_alloc_lock, and wait for
every process to complete the wrap. (Note: every CPU must WAIT before
leaving smp_new_mmu_context_version_client() until every one arrives).

This approach ends up with deadlocks, as some threads own locks which other
threads are waiting for, and they never receive softint until these threads
exit smp_new_mmu_context_version_client(). Since we do not allow the exit,
deadlock happens.

Approach 2:
Handle wrap right during mondo interrupt. Use etrap/rtrap to enter into
into C code, and issue new versions to every CPU.
This approach adds some overhead to runtime: in switch_mm() we must add
some checks to make sure that versions have not changed due to wrap while
we were loading the new secondary context. (could be protected by PSTATE_IE
but that degrades performance as on M7 and older CPUs as it takes 50 cycles
for each access). Also, we still need a global per-cpu array of MMs to know
where we need to load new contexts, otherwise we can change context to a
thread that is going way (if we received mondo between switch_mm() and
switch_to() time). Finally, there are some issues with window registers in
rtrap() when context IDs are changed during CPU mondo time.

The approach in this patch is the simplest and has almost no impact on
runtime.  We use the array with mm's where last secondary contexts were
loaded onto CPUs and bump their versions to the new generation without
changing context IDs. If a new process comes in to get a context ID, it
will go through get_new_mmu_context() because of version mismatch. But the
running processes do not need to be interrupted. And wrap is quicker as we
do not need to xcall and wait for everyone to receive and complete wrap.
Signed-off-by: Pavel Tatashin <pasha.tatashin@oracle.com>
Reviewed-by: Bob Picco <bob.picco@oracle.com>
Reviewed-by: Steven Sistare <steven.sistare@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The new wrap is going to use information from this array to figure out
mm's that currently have valid secondary contexts setup.
Signed-off-by: Pavel Tatashin <pasha.tatashin@oracle.com>
Reviewed-by: Bob Picco <bob.picco@oracle.com>
Reviewed-by: Steven Sistare <steven.sistare@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

CTX_FIRST_VERSION defines the first context version, but also it defines
first context. This patch redefines it to only include the first context
version.
Signed-off-by: Pavel Tatashin <pasha.tatashin@oracle.com>
Reviewed-by: Bob Picco <bob.picco@oracle.com>
Reviewed-by: Steven Sistare <steven.sistare@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The only difference between these two functions is that in activate_mm we
unconditionally flush context. However, there is no need to keep this
difference after fixing a bug where cpumask was not reset on a wrap. So, in
this patch we combine these.
Signed-off-by: Pavel Tatashin <pasha.tatashin@oracle.com>
Reviewed-by: Bob Picco <bob.picco@oracle.com>
Reviewed-by: Steven Sistare <steven.sistare@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

After a wrap (getting a new context version) a process must get a new
context id, which means that we would need to flush the context id from
the TLB before running for the first time with this ID on every CPU. But,
we use mm_cpumask to determine if this process has been running on this CPU
before, and this mask is not reset after a wrap. So, there are two possible
fixes for this issue:

1. Clear mm cpumask whenever mm gets a new context id
2. Unconditionally flush context every time process is running on a CPU

This patch implements the first solution
Signed-off-by: Pavel Tatashin <pasha.tatashin@oracle.com>
Reviewed-by: Bob Picco <bob.picco@oracle.com>
Reviewed-by: Steven Sistare <steven.sistare@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

hugetlb_bad_size needs to be called on invalid values.  Also change the
pr_warn to a pr_err to better align with other platforms.
Signed-off-by: Liam R. Howlett <Liam.Howlett@Oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>