openwrt/staging/blogic.git
8 years agonfp: bpf: zero extend 4 byte context loads
Jakub Kicinski [Wed, 28 Sep 2016 22:47:37 +0000 (23:47 +0100)]
nfp: bpf: zero extend 4 byte context loads

Set upper 32 bits of destination register to zeros after
load from the context structure.

Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agomlx4: remove unused fields
David Decotigny [Wed, 28 Sep 2016 18:00:04 +0000 (11:00 -0700)]
mlx4: remove unused fields

This also can address following UBSAN warnings:
[   36.640343] ================================================================================
[   36.648772] UBSAN: Undefined behaviour in drivers/net/ethernet/mellanox/mlx4/fw.c:857:26
[   36.656853] shift exponent 64 is too large for 32-bit type 'int'
[   36.663348] ================================================================================
[   36.671783] ================================================================================
[   36.680213] UBSAN: Undefined behaviour in drivers/net/ethernet/mellanox/mlx4/fw.c:861:27
[   36.688297] shift exponent 35 is too large for 32-bit type 'int'
[   36.694702] ================================================================================

Tested:
  reboot with UBSAN, no warning.

Signed-off-by: David Decotigny <decot@googlers.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoipv6 addrconf: implement RFC7559 router solicitation backoff
Maciej Żenczykowski [Wed, 28 Sep 2016 06:57:58 +0000 (23:57 -0700)]
ipv6 addrconf: implement RFC7559 router solicitation backoff

This implements:
  https://tools.ietf.org/html/rfc7559

Backoff is performed according to RFC3315 section 14:
  https://tools.ietf.org/html/rfc3315#section-14

We allow setting /proc/sys/net/ipv6/conf/*/router_solicitations
to a negative value meaning an unlimited number of retransmits,
and we make this the new default (inline with the RFC).

We also add a new setting:
  /proc/sys/net/ipv6/conf/*/router_solicitation_max_interval
defaulting to 1 hour (per RFC recommendation).

Signed-off-by: Maciej Żenczykowski <maze@google.com>
Acked-by: Erik Kline <ek@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge branch 'net_proc_perf'
David S. Miller [Fri, 30 Sep 2016 05:50:57 +0000 (01:50 -0400)]
Merge branch 'net_proc_perf'

Jia He says:

====================
Reduce cache miss for snmp_fold_field

In a PowerPc server with large cpu number(160), besides commit
a3a773726c9f ("net: Optimize snmp stat aggregation by walking all
the percpu data at once"), I watched several other snmp_fold_field
callsites which would cause high cache miss rate.

test source code:
================
My simple test case, which read from the procfs items endlessly:
/***********************************************************/
int main(int argc, char **argv)
{
        int i;
        int fd = -1 ;
        int rdsize = 0;
        char buf[LINELEN+1];

        buf[LINELEN] = 0;
        memset(buf,0,LINELEN);

        if(1 >= argc) {
                printf("file name empty\n");
                return -1;
        }

        fd = open(argv[1], O_RDWR, 0644);
        if(0 > fd){
                printf("open error\n");
                return -2;
        }

        for(i=0;i<0xffffffff;i++) {
                while(0 < (rdsize = read(fd,buf,LINELEN))){
                        //nothing here
                }

                lseek(fd, 0, SEEK_SET);
        }

        close(fd);
        return 0;
}
/**********************************************************/

compile and run:
================
gcc test.c -o test

perf stat -d -e cache-misses ./test /proc/net/snmp
perf stat -d -e cache-misses ./test /proc/net/snmp6
perf stat -d -e cache-misses ./test /proc/net/sctp/snmp
perf stat -d -e cache-misses ./test /proc/net/xfrm_stat

before the patch set:
====================
 Performance counter stats for 'system wide':

         355911097      cache-misses                                                 [40.08%]
        2356829300      L1-dcache-loads                                              [60.04%]
         355642645      L1-dcache-load-misses     #   15.09% of all L1-dcache hits   [60.02%]
         346544541      LLC-loads                                                    [59.97%]
            389763      LLC-load-misses           #    0.11% of all LL-cache hits    [40.02%]

       6.245162638 seconds time elapsed

After the patch set:
===================
 Performance counter stats for 'system wide':

         194992476      cache-misses                                                 [40.03%]
        6718051877      L1-dcache-loads                                              [60.07%]
         194871921      L1-dcache-load-misses     #    2.90% of all L1-dcache hits   [60.11%]
         187632232      LLC-loads                                                    [60.04%]
            464466      LLC-load-misses           #    0.25% of all LL-cache hits    [39.89%]

       6.868422769 seconds time elapsed
The cache-miss rate can be reduced from 15% to 2.9%

changelog
=========
v6:
- correct v5
v5:
- order local variables from longest to shortest line
v4:
- move memset into one block of if statement in snmp6_seq_show_item
- remove the changes in netstat_seq_show considerred the stack usage is too large
v3:
- introduce generic interface (suggested by Marcelo Ricardo Leitner)
- use max_t instead of self defined macro (suggested by David Miller)
v2:
- fix bug in udplite statistics.
- snmp_seq_show is split into 2 parts
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: Suppress the "Comparison to NULL could be written" warnings
Jia He [Fri, 30 Sep 2016 03:29:04 +0000 (11:29 +0800)]
net: Suppress the "Comparison to NULL could be written" warnings

This is to suppress the checkpatch.pl warning "Comparison to NULL
could be written". No functional changes here.

Signed-off-by: Jia He <hejianet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoipv6: Remove useless parameter in __snmp6_fill_statsdev
Jia He [Fri, 30 Sep 2016 03:29:03 +0000 (11:29 +0800)]
ipv6: Remove useless parameter in __snmp6_fill_statsdev

The parameter items(is always ICMP6_MIB_MAX) is useless for __snmp6_fill_statsdev

Signed-off-by: Jia He <hejianet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoproc: Reduce cache miss in xfrm_statistics_seq_show
Jia He [Fri, 30 Sep 2016 03:29:02 +0000 (11:29 +0800)]
proc: Reduce cache miss in xfrm_statistics_seq_show

This is to use the generic interfaces snmp_get_cpu_field{,64}_batch to
aggregate the data by going through all the items of each cpu sequentially.

Signed-off-by: Jia He <hejianet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoproc: Reduce cache miss in sctp_snmp_seq_show
Jia He [Fri, 30 Sep 2016 03:29:01 +0000 (11:29 +0800)]
proc: Reduce cache miss in sctp_snmp_seq_show

This is to use the generic interfaces snmp_get_cpu_field{,64}_batch to
aggregate the data by going through all the items of each cpu sequentially.

Signed-off-by: Jia He <hejianet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoproc: Reduce cache miss in snmp6_seq_show
Jia He [Fri, 30 Sep 2016 03:29:00 +0000 (11:29 +0800)]
proc: Reduce cache miss in snmp6_seq_show

This is to use the generic interfaces snmp_get_cpu_field{,64}_batch to
aggregate the data by going through all the items of each cpu sequentially.

Signed-off-by: Jia He <hejianet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoproc: Reduce cache miss in snmp_seq_show
Jia He [Fri, 30 Sep 2016 03:28:59 +0000 (11:28 +0800)]
proc: Reduce cache miss in snmp_seq_show

This is to use the generic interfaces snmp_get_cpu_field{,64}_batch to
aggregate the data by going through all the items of each cpu sequentially.
Then snmp_seq_show is split into 2 parts to avoid build warning "the frame
size" larger than 1024.

Signed-off-by: Jia He <hejianet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet:snmp: Introduce generic interfaces for snmp_get_cpu_field{, 64}
Jia He [Fri, 30 Sep 2016 03:28:58 +0000 (11:28 +0800)]
net:snmp: Introduce generic interfaces for snmp_get_cpu_field{, 64}

This is to introduce the generic interfaces for snmp_get_cpu_field{,64}.
It exchanges the two for-loops for collecting the percpu statistics data.
This can aggregate the data by going through all the items of each cpu
sequentially.

Signed-off-by: Jia He <hejianet@gmail.com>
Suggested-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge tag 'rxrpc-rewrite-20160929' of git://git.kernel.org/pub/scm/linux/kernel/git...
David S. Miller [Fri, 30 Sep 2016 05:46:36 +0000 (01:46 -0400)]
Merge tag 'rxrpc-rewrite-20160929' of git://git./linux/kernel/git/dhowells/linux-fs

David Howells says:

====================
rxrpc: Fixes and adjustments

This set of patches contains some fixes and adjustments:

 (1) Connections for exclusive calls are being reused because the check to
     work out whether to set RXRPC_CONN_DONT_REUSE is checking where the
     parameters will be copied to (but haven't yet).

 (2) Make Tx loss-injection go through the normal return, so the state gets
     set correctly for what the code thinks it has done.

     Note lost Tx packets in the tx_data trace rather than the skb
     tracepoint.

 (3) Activate channels according to the current state from within the
     channel_lock to avoid someone changing it on us.

 (4) Reduce the local endpoint's services list to a single pointer as we
     don't allow service AF_RXRPC sockets to share UDP ports with other
     AF_RXRPC sockets, so there can't be more than one element in the list.

 (5) Request more ACKs in slow-start mode to help monitor the state driving
     the window configuration.

 (6) Display the serial number of the packet being ACK'd rather than the
     ACK packet's own serial number in the congestion trace as this can be
     related to other entries in the trace.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge tag 'wireless-drivers-next-for-davem-2016-09-29' of git://git.kernel.org/pub...
David S. Miller [Fri, 30 Sep 2016 05:31:51 +0000 (01:31 -0400)]
Merge tag 'wireless-drivers-next-for-davem-2016-09-29' of git://git./linux/kernel/git/kvalo/wireless-drivers-next

Kalle Valo says:

wireless-drivers-next patches for 4.9

Major changes:

iwlwifi

* work for new hardware support continues
* dynamic queue allocation stabilization
* improvements in the MSIx code
* multiqueue support work continues
* new firmware version support (API 26)
* add 8275 series support
* add 9560 series support
* add support for MU-MIMO sniffer
* add support for RRM by scan
* add support for "reverse" rx packet injection faking hw descriptors
* migrate to devm memory allocation handling
* Remove support for older firmwares (API older than -17 and -22)

wl12xx

* support booting the same rootfs with both wl12xx and wl18xx

hostap

* mark the driver as obsolete

ath9k

* disable RNG by default
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge branch 'dsa-global-cosmetics'
David S. Miller [Fri, 30 Sep 2016 05:26:04 +0000 (01:26 -0400)]
Merge branch 'dsa-global-cosmetics'

Vivien Didelot says:

====================
net: dsa: mv88e6xxx: Global (1) cosmetics

The Global (1) internal SMI device of Marvell switches is a set of
registers providing support to different units for MAC addresses (ATU),
VLANs (VTU), PHY polling (PPU), etc.

Chips (like 88E6060) may use a different address for it, or have
subtleties in the units (e.g. different number of databases, changing
how registers must be accessed), making it hard to maintain properly.

This patchset is a first step to polish the Global (1) support, with no
functional changes though. Here's basically what it does:

  - add helpers to access Global1 registers (same for Global2)
  - remove a few family checks (VTU/STU FID registers)
  - s/mv88e6xxx_vtu_stu_entry/mv88e6xxx_vtu_entry/
  - add a per-chip mv88e6xxx_ops structure of function pointers:

  struct mv88e6xxx_ops {
        int (*get_eeprom)(struct mv88e6xxx_chip *chip,
                          struct ethtool_eeprom *eeprom, u8 *data);
        int (*set_eeprom)(struct mv88e6xxx_chip *chip,
                          struct ethtool_eeprom *eeprom, u8 *data);

        int (*set_switch_mac)(struct mv88e6xxx_chip *chip, u8 *addr);

        int (*phy_read)(struct mv88e6xxx_chip *chip, int addr, int reg,
                        u16 *val);
        int (*phy_write)(struct mv88e6xxx_chip *chip, int addr, int reg,
                         u16 val);
  };

Future patchsets will add ATU/VTU ops, software reset, etc.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: dsa: mv88e6xxx: add eeprom ops
Vivien Didelot [Thu, 29 Sep 2016 16:22:02 +0000 (12:22 -0400)]
net: dsa: mv88e6xxx: add eeprom ops

Remove EEPROM flags in favor of new {get,set}_eeprom chip-wide
functions in the mv88e6xxx_ops structure.

Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: dsa: mv88e6xxx: add set_switch_mac to ops
Vivien Didelot [Thu, 29 Sep 2016 16:22:01 +0000 (12:22 -0400)]
net: dsa: mv88e6xxx: add set_switch_mac to ops

Add a set_switch_mac chip-wide function to mv88e6xxx_ops and remove
MV88E6XXX_FLAG_G2_SWITCH_MAC flags.

Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: dsa: mv88e6xxx: add chip-wide ops
Vivien Didelot [Thu, 29 Sep 2016 16:22:00 +0000 (12:22 -0400)]
net: dsa: mv88e6xxx: add chip-wide ops

Introduce a mv88e6xxx_ops structure to describe supported chip-wide
functions and assign the correct variant to the chip models.

For the moment, add only PHY access routines. This allows to get rid of
the PHY ops structures and the usage of PHY flags.

Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: dsa: mv88e6xxx: rename mv88e6xxx_ops
Vivien Didelot [Thu, 29 Sep 2016 16:21:59 +0000 (12:21 -0400)]
net: dsa: mv88e6xxx: rename mv88e6xxx_ops

The mv88e6xxx_ops is used to describe how to access the chip registers.
It can be through SMI (via an MDIO bus), or via another interface such
as crafted remote management frames.

The correct BUS operations structure is chosen at runtime, depending on
the chip address and connectivity.

We will need the mv88e6xxx_ops name for future chip-wide operation
structure, thus rename mv88e6xxx_ops to more explicit mv88e6xxx_bus_ops.

Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: dsa: mv88e6xxx: rename mv88e6xxx_vtu_stu_entry
Vivien Didelot [Thu, 29 Sep 2016 16:21:58 +0000 (12:21 -0400)]
net: dsa: mv88e6xxx: rename mv88e6xxx_vtu_stu_entry

The STU (if the switch has one) is abstracted and accessed through the
VTU operations and data registers.

Thus rename the mv88e6xxx_vtu_stu_entry struct to mv88e6xxx_vtu_entry.

Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: dsa: mv88e6xxx: add mv88e6xxx_num_ports helper
Vivien Didelot [Thu, 29 Sep 2016 16:21:57 +0000 (12:21 -0400)]
net: dsa: mv88e6xxx: add mv88e6xxx_num_ports helper

Add an mv88e6xxx_num_ports helper instead of digging in the chip info
structure.

Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: dsa: mv88e6xxx: expose mv88e6xxx_num_databases
Vivien Didelot [Thu, 29 Sep 2016 16:21:56 +0000 (12:21 -0400)]
net: dsa: mv88e6xxx: expose mv88e6xxx_num_databases

The mv88e6xxx_num_databases will be used by shared code, so move it
inline to the header file.

Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: dsa: mv88e6xxx: add flags for FID registers
Vivien Didelot [Thu, 29 Sep 2016 16:21:55 +0000 (12:21 -0400)]
net: dsa: mv88e6xxx: add flags for FID registers

Add flags to describe the presence of Global 1 ATU FID register (0x01)
and VTU FID register (0x02), instead of checking families.

Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: dsa: mv88e6xxx: abstract REG_GLOBAL2
Vivien Didelot [Thu, 29 Sep 2016 16:21:54 +0000 (12:21 -0400)]
net: dsa: mv88e6xxx: abstract REG_GLOBAL2

Similarly to the ports, phys, and Global SMI devices, abstract the SMI
device address of the Global 2 registers in a few g2 static helpers.

Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: dsa: mv88e6xxx: add global1 helpers
Vivien Didelot [Thu, 29 Sep 2016 16:21:53 +0000 (12:21 -0400)]
net: dsa: mv88e6xxx: add global1 helpers

The Global (1) internal SMI device is an extended set of registers
containing ATU, PPU, VTU, STU, etc.

It is present on every switches, usually at SMI address 0x1B. But old
models such as 88E6060 access it at address 0xF, thus using REG_GLOBAL
is erroneous.

Add a global1_addr info member used by mv88e6xxx_g1_{read,write} and
mv88e6xxx_g1_wait helpers in a new global1.c file.

This patch finally removes _mv88e6xxx_reg_{read,write}, in favor on the
appropriate helpers. No functional changes here.

Signed-off-by: Vivien Didelot <vivien.didelot@savoirfairelinux.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agorxrpc: Note serial number being ACK'd in the congestion management trace
David Howells [Thu, 29 Sep 2016 21:37:16 +0000 (22:37 +0100)]
rxrpc: Note serial number being ACK'd in the congestion management trace

Note the serial number of the packet being ACK'd in the congestion
management trace rather than the serial number of the ACK packet.  Whilst
the serial number of the ACK packet is useful for matching ACK packet in
the output of wireshark, the serial number that the ACK is in response to
is of more use in working out how different trace lines relate.

Signed-off-by: David Howells <dhowells@redhat.com>
8 years agorxrpc: Request more ACKs in slow-start mode
David Howells [Thu, 29 Sep 2016 21:37:16 +0000 (22:37 +0100)]
rxrpc: Request more ACKs in slow-start mode

Set the request-ACK on more DATA packets whilst we're in slow start mode so
that we get sufficient ACKs back to supply information to configure the
window.

Signed-off-by: David Howells <dhowells@redhat.com>
8 years agorxrpc: Reduce the rxrpc_local::services list to a pointer
David Howells [Thu, 29 Sep 2016 21:37:15 +0000 (22:37 +0100)]
rxrpc: Reduce the rxrpc_local::services list to a pointer

Reduce the rxrpc_local::services list to just a pointer as we don't permit
multiple service endpoints to bind to a single transport endpoints (this is
excluded by rxrpc_lookup_local()).

The reason we don't allow this is that if you send a request to an AFS
filesystem service, it will try to talk back to your cache manager on the
port you sent from (this is how file change notifications are handled).  To
prevent someone from stealing your CM callbacks, we don't let AF_RXRPC
sockets share a UDP socket if at least one of them has a service bound.

Signed-off-by: David Howells <dhowells@redhat.com>
8 years agorxrpc: When activating client conn channels, do state check inside lock
David Howells [Thu, 29 Sep 2016 21:37:15 +0000 (22:37 +0100)]
rxrpc: When activating client conn channels, do state check inside lock

In rxrpc_activate_channels(), the connection cache state is checked outside
of the lock, which means it can change whilst we're waking calls up,
thereby changing whether or not we're allowed to wake calls up.

Fix this by moving the check inside the locked region.  The check to see if
all the channels are currently busy can stay outside of the locked region.

Whilst we're at it:

 (1) Split the locked section out into its own function so that we can call
     it from other places in a later patch.

 (2) Determine the mask of channels dependent on the state as we're going
     to add another state in a later patch that will restrict the number of
     simultaneous calls to 1 on a connection.

Signed-off-by: David Howells <dhowells@redhat.com>
8 years agorxrpc: Make Tx loss-injection go through normal return and adjust tracing
David Howells [Thu, 29 Sep 2016 21:37:15 +0000 (22:37 +0100)]
rxrpc: Make Tx loss-injection go through normal return and adjust tracing

In rxrpc_send_data_packet() make the loss-injection path return through the
same code as the transmission path so that the RTT determination is
initiated and any future timer shuffling will be done, despite the packet
having been binned.

Whilst we're at it:

 (1) Add to the tx_data tracepoint an indication of whether or not we're
     retransmitting a data packet.

 (2) When we're deciding whether or not to request an ACK, rather than
     checking if we're in fast-retransmit mode check instead if we're
     retransmitting.

 (3) Don't invoke the lose_skb tracepoint when losing a Tx packet as we're
     not altering the sk_buff refcount nor are we just seeing it after
     getting it off the Tx list.

 (4) The rxrpc_skb_tx_lost note is then no longer used so remove it.

 (5) rxrpc_lose_skb() no longer needs to deal with rxrpc_skb_tx_lost.

Signed-off-by: David Howells <dhowells@redhat.com>
8 years agorxrpc: Fix exclusive client connections
David Howells [Thu, 29 Sep 2016 21:37:15 +0000 (22:37 +0100)]
rxrpc: Fix exclusive client connections

Exclusive connections are currently reusable (which they shouldn't be)
because rxrpc_alloc_client_connection() checks the exclusive flag in the
rxrpc_connection struct before it's initialised from the function
parameters.  This means that the DONT_REUSE flag doesn't get set.

Fix this by checking the function parameters for the exclusive flag.

Signed-off-by: David Howells <dhowells@redhat.com>
8 years agoMerge branch 'qcom-emac-acpi'
David S. Miller [Thu, 29 Sep 2016 05:50:20 +0000 (01:50 -0400)]
Merge branch 'qcom-emac-acpi'

Timur Tabi says:

====================
Add basic ACPI support to the Qualcomm Technologies EMAC driver

This patch series adds support to the EMAC driver for extracting addresses,
interrupts, and some _DSDs (properties) from ACPI.  The first two patches
clean up the code, and the third patch adds ACPI-specific functionality.

The first patch fixes a bug with handling the platform_device for the
internal PHY.  This phy is treated as a separate device in both DT and
ACPI, but since the platform is not released automatically when the
driver unloads, managed functions like devm_ioremap_resource cannot be
used.

The second patch replaces of_get_mac_address with its platform-independent
equivalent device_get_mac_address.

The third patch parses the ACPI tables to obtain the platform_device for
the primary EMAC node ("QCOM8070") and the internal phy node ("QCOM8071").
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: qcom/emac: initial ACPI support
Timur Tabi [Wed, 28 Sep 2016 16:58:44 +0000 (11:58 -0500)]
net: qcom/emac: initial ACPI support

Add support for reading addresses, interrupts, and _DSD properties
from ACPI tables, just like with device tree.  The HID for the
EMAC device itself is QCOM8070.  The internal PHY is represented
by a child node with a HID of QCOM8071.

The EMAC also has some complex clock initialization requirements
that are not represented by this patch.  This will be addressed
in a future patch.

Signed-off-by: Timur Tabi <timur@codeaurora.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: qcom/emac: use device_get_mac_address
Timur Tabi [Wed, 28 Sep 2016 16:58:43 +0000 (11:58 -0500)]
net: qcom/emac: use device_get_mac_address

Replace the DT-specific of_get_mac_address() function with
device_get_mac_address, which works on both DT and ACPI platforms.  This
change makes it easier to add ACPI support.

Signed-off-by: Timur Tabi <timur@codeaurora.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: qcom/emac: do not use devm on internal phy pdev
Timur Tabi [Wed, 28 Sep 2016 16:58:42 +0000 (11:58 -0500)]
net: qcom/emac: do not use devm on internal phy pdev

The platform_device returned by of_find_device_by_node() is not
automatically released when the driver unprobes.  Therefore,
managed calls like devm_ioremap_resource() should not be used.
Instead, we manually allocate the resources and then free them
on driver release.

Signed-off-by: Timur Tabi <timur@codeaurora.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agobpf: allow access into map value arrays
Josef Bacik [Wed, 28 Sep 2016 14:54:32 +0000 (10:54 -0400)]
bpf: allow access into map value arrays

Suppose you have a map array value that is something like this

struct foo {
unsigned iter;
int array[SOME_CONSTANT];
};

You can easily insert this into an array, but you cannot modify the contents of
foo->array[] after the fact.  This is because we have no way to verify we won't
go off the end of the array at verification time.  This patch provides a start
for this work.  We accomplish this by keeping track of a minimum and maximum
value a register could be while we're checking the code.  Then at the time we
try to do an access into a MAP_VALUE we verify that the maximum offset into that
region is a valid access into that memory region.  So in practice, code such as
this

unsigned index = 0;

if (foo->iter >= SOME_CONSTANT)
foo->iter = index;
else
index = foo->iter++;
foo->array[index] = bar;

would be allowed, as we can verify that index will always be between 0 and
SOME_CONSTANT-1.  If you wish to use signed values you'll have to have an extra
check to make sure the index isn't less than 0, or do something like index %=
SOME_CONSTANT.

Signed-off-by: Josef Bacik <jbacik@fb.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: do not export sk_stream_write_space
Eric Dumazet [Wed, 28 Sep 2016 15:41:16 +0000 (08:41 -0700)]
net: do not export sk_stream_write_space

Since commit 900f65d361d3 ("tcp: move duplicate code from
tcp_v4_init_sock()/tcp_v6_init_sock()") we no longer need
to export sk_stream_write_space()

From: Eric Dumazet <edumazet@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge ath-next from git://git.kernel.org/pub/scm/linux/kernel/git/kvalo/ath.git
Kalle Valo [Wed, 28 Sep 2016 13:37:33 +0000 (16:37 +0300)]
Merge ath-next from git://git./linux/kernel/git/kvalo/ath.git

ath.git patches for 4.9. Major changes:

ath9k

* disable RNG by default

8 years agotcp: Change txhash on every SYN and RTO retransmit
Lawrence Brakmo [Wed, 28 Sep 2016 02:03:37 +0000 (19:03 -0700)]
tcp: Change txhash on every SYN and RTO retransmit

The current code changes txhash (flowlables) on every retransmitted
SYN/ACK, but only after the 2nd retransmitted SYN and only after
tcp_retries1 RTO retransmits.

With this patch:
1) txhash is changed with every SYN retransmits
2) txhash is changed with every RTO.

The result is that we can start re-routing around failed (or very
congested paths) as soon as possible. Otherwise application health
checks may fail and the connection may be terminated before we start
to change txhash.

v4: Removed sysctl, txhash is changed for all RTOs
v3: Removed text saying default value of sysctl is 0 (it is 100)
v2: Added sysctl documentation and cleaned code

Tested with packetdrill tests

Signed-off-by: Lawrence Brakmo <brakmo@fb.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMAINTAINERS: hostap: Mark the Host AP driver obsolete
Jouni Malinen [Tue, 27 Sep 2016 15:42:58 +0000 (18:42 +0300)]
MAINTAINERS: hostap: Mark the Host AP driver obsolete

This is old code for old hardware and it is not really accurate to claim
this to be maintained anymore. Change the status to "Obsolete" to make
it clearer that minor cleanup and other unnecessary changes from
automated tools is not necessarily beneficial and has larger risk of
breaking something without being quickly noticed due to lack of testing.

In addition, remove the old mailing list that does not work anymore and
point the web-page to a more accurate URL.

Signed-off-by: Jouni Malinen <j@w1.fi>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
8 years agoath6kl: fix return value in ath6kl_wmi_set_pvb_cmd
Chaehyun Lim [Sun, 18 Sep 2016 06:30:24 +0000 (15:30 +0900)]
ath6kl: fix return value in ath6kl_wmi_set_pvb_cmd

When building with W=1, we got one warning as belows:
drivers/net/wireless/ath/ath6kl/wmi.c:3509:6: warning: variable ‘ret’
set but not used [-Wunused-but-set-variable]

At the end of ath6kl_wmi_set_pvb_cmd, it is returned by 0 regardless of
return value of ath6kl_wmi_cmd_send.
This patch fixes return value from 0 to ret that has result of
ath6kl_wmi_cmd_send execution.

Signed-off-by: Chaehyun Lim <chaehyun.lim@gmail.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
8 years agoath9k: disable RNG by default
Miaoqing Pan [Tue, 9 Aug 2016 07:02:27 +0000 (15:02 +0800)]
ath9k: disable RNG by default

ath9k RNG will dominates all the noise sources from the real HW
RNG, disable it by default. But we strongly recommand to enable
it if the system without HW RNG, especially on embedded systems.

Signed-off-by: Miaoqing Pan <miaoqing@codeaurora.org>
Acked-by: Stephan Mueller <smueller@chronox.de>
Acked-by: Stephan Mueller <smueller@chronox.de>
Reviewed-by: Jason Cooper <jason@lakedaemon.net>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
8 years agoath10k: fix copy engine 5 destination ring stuck
Rajkumar Manoharan [Wed, 21 Sep 2016 10:58:06 +0000 (16:28 +0530)]
ath10k: fix copy engine 5 destination ring stuck

Firmware is running watchdog timer for tracking copy engine ring index
and write index. Whenever both indices are stuck at same location for
given duration, watchdog will be trigger to assert target. While
updating copy engine destination ring write index, driver ensures that
write index will not be same as read index by finding delta between these
two indices (CE_RING_DELTA).

HTT target to host copy engine (CE5) is special case where ring buffers
will be reused and delta check is not applied while updating write index.
In rare scenario, whenever CE5 ring is full, both indices will be referring
same location and this is causing CE ring stuck issue as explained
above. This issue is originally reported on IPQ4019 during long hour stress
testing and during veriwave max clients testsuites. The same issue is
also observed in other chips as well. Fix this by ensuring that write
index is one less than read index which means that full ring is
available for receiving data.

Cc: stable@vger.kernel.org
Tested-by: Tamizh chelvam <c_traja@qti.qualcomm.com>
Signed-off-by: Rajkumar Manoharan <rmanohar@qti.qualcomm.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
8 years agoath10k: Ignore SWBA event for a vif if its marked for no beacon
Mohammed Shafi Shajakhan [Tue, 20 Sep 2016 08:22:07 +0000 (13:52 +0530)]
ath10k: Ignore SWBA event for a vif if its marked for no beacon

Ignore processing further in SWBA event scheduled for a vif, if mac80211
has marked the particular vif for stop beaconing and brought the vdev
down in 'ath10k_control_beaconing'. This should potentially avoid ath10k
warning/error messages while running continuous wifi down/up with max
number of vaps configured. Found this change during code walk through
and going through other beacon configuration related functions in ath10k

Signed-off-by: Mohammed Shafi Shajakhan <mohammed@qti.qualcomm.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
8 years agoath10k: fix error return code in ahb
Wei Yongjun [Fri, 16 Sep 2016 13:05:35 +0000 (13:05 +0000)]
ath10k: fix error return code in ahb

Fix to return a negative error code from the error handling case
instead of 0, as done elsewhere in function ath10k_ahb_probe() or
ath10k_ahb_resource_init().

Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
8 years agonet/sched: pkt_cls: change tc actions order to be as the user sets
Hadar Hen Zion [Tue, 27 Sep 2016 08:09:51 +0000 (11:09 +0300)]
net/sched: pkt_cls: change tc actions order to be as the user sets

Currently the created tc actions list is reversed against the order
set by the user.
Change the actions list order to be the same as was set by the user.

This patch doesn't affect dump actions behavior.
For dumping, action->order parameter is used so the list order doesn't
matter.

Signed-off-by: Hadar Hen Zion <hadarh@mellanox.com>
Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Acked-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agosh_eth: add R8A7743/5 support
Sergei Shtylyov [Mon, 26 Sep 2016 22:23:26 +0000 (01:23 +0300)]
sh_eth: add R8A7743/5 support

Add support for the first two members of the Renesas RZ/G family, RZ/G1M/E
(also known as  R8A7743/5). The Ether core is the same as in the R-Car gen2
SoCs, so will share the code/data with them...

Signed-off-by: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Acked-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge branch 'fib-offload-notifications'
David S. Miller [Wed, 28 Sep 2016 08:48:20 +0000 (04:48 -0400)]
Merge branch 'fib-offload-notifications'

Jiri Pirko says:

====================
fib offload: switch to notifier

The goal of this patchset is to allow driver to propagate all prefixes
configured in kernel down HW. This is necessary for routing to work
as expected. If we don't do that HW might forward prefixes known to kernel
incorrectly. Take an example when default route is set in switch HW and there
is an IP address set on a management (non-switch) port.

Currently, only FIB entries related to the switch port netdev are
offloaded using switchdev ops. This model is not extendable so the
first patch introduces a replacement: notifier to propagate FIB entry
additions and removals to whoever is interested.

The second patch introduces couple of helpers to deal with RTNH_F_OFFLOAD
flags. Currently it is set in switchdev core. There the assumption is
that only one offload device exists. But for FIB notifier, we assume
multiple offload devices. So the patch introduces a per FIB entry
reference counter and helpers use it in order to achieve this:
   0 means RTNH_F_OFFLOAD is not set, no device offloads this entry
   n means RTNH_F_OFFLOAD is set and the entry is offloaded by n devices

Patches 3 and 4 convert mlxsw and rocker to adopt this new way, registering
one notifier block for each asic instance. Both of these patches also
implement internal "abort" mechanism.

Using switchdev ops, "abort" is called by switchdev core whenever there is
an error during FIB entry add offload. This leads to removal of all
offloaded entries on system by fib_trie code.

Now the new notifier assumes the driver takes care of the abort action.
Here's why:
1) The fact that one HW cannot offload an entry does not mean that the
   others can't do it. So let only one entity to abort and leave the rest
   to work happily.
2) The driver knows what to in order to properly abort. For example,
   currently abort is broken for mlxsw, as for Spectrum there is a need
   to set 0.0.0.0/0 trap in RALUE register.

The fifth patch removes the old, no longer used FIB offload infrastructure.

The last patch reflects the changes into switchdev documentation file.

---
v2->v3:
 -patch 3/6
   -fixed offload inc/dec to be done in fib4_entry_init/fini and only
    in case !trap as suggested by Ido
v1->v2:
 -patch 3/6:
   -fixed lpm tree setup and binding for abort and pointed out by Ido
   -do nexthop checks as suggested by Ido
   -fix use after free during abort
 -patch 6/6:
   -fixed texts as suggested by Ido
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agodoc: update switchdev L3 section
Jiri Pirko [Mon, 26 Sep 2016 10:52:34 +0000 (12:52 +0200)]
doc: update switchdev L3 section

This is to reflect the change of FIB offload infrastructure from
switchdev objects to FIB notifier.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoswitchdev: remove FIB offload infrastructure
Jiri Pirko [Mon, 26 Sep 2016 10:52:33 +0000 (12:52 +0200)]
switchdev: remove FIB offload infrastructure

Since this is now taken care of by FIB notifier, remove the code, with
all unused dependencies.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agorocker: use FIB notifications instead of switchdev calls
Jiri Pirko [Mon, 26 Sep 2016 10:52:32 +0000 (12:52 +0200)]
rocker: use FIB notifications instead of switchdev calls

Until now, in order to offload a FIB entry to HW we use switchdev op.
Use the newly introduced FIB notifier infrasturucture to process
FIB entries to offload the in HW.
Abort mechanism is now handled within the rocker driver.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agomlxsw: spectrum_router: Use FIB notifications instead of switchdev calls
Jiri Pirko [Mon, 26 Sep 2016 10:52:31 +0000 (12:52 +0200)]
mlxsw: spectrum_router: Use FIB notifications instead of switchdev calls

Until now, in order to offload a FIB entry to HW we use switchdev op.
However that has limits. Mainly in case we need to make the HW aware of
all route prefixes configured in kernel. HW needs to know those in order
to properly trap appropriate packets and pass the to kernel to do
the forwarding. Abort mechanism is now handled within the mlxsw driver.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agofib: introduce FIB info offload flag helpers
Jiri Pirko [Mon, 26 Sep 2016 10:52:30 +0000 (12:52 +0200)]
fib: introduce FIB info offload flag helpers

These helpers are to be used in case someone offloads the FIB entry. The
result is that if the entry is offloaded to at least one device, the
offload flag is set.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Reviewed-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agofib: introduce FIB notification infrastructure
Jiri Pirko [Mon, 26 Sep 2016 10:52:29 +0000 (12:52 +0200)]
fib: introduce FIB notification infrastructure

This allows to pass information about added/deleted FIB entries/rules to
whoever is interested. This is done in a very similar way as devinet
notifies address additions/removals.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet/sched: cls_flower: Use a proper mask value for enc key id parameter
Hadar Hen Zion [Tue, 27 Sep 2016 08:21:18 +0000 (11:21 +0300)]
net/sched: cls_flower: Use a proper mask value for enc key id parameter

The current code use the encapsulation key id value as the mask of that
parameter which is wrong. Fix that by using a full mask.

Fixes: bc3103f1ed40 ('net/sched: cls_flower: Classify packet in ip tunnels')
Signed-off-by: Hadar Hen Zion <hadarh@mellanox.com>
Acked-by: Amir Vadai <amir@vadai.me>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge branch '1GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/next...
David S. Miller [Wed, 28 Sep 2016 03:22:27 +0000 (23:22 -0400)]
Merge branch '1GbE' of git://git./linux/kernel/git/jkirsher/next-queue

Jeff Kirsher says:

====================
1GbE Intel Wired LAN Driver Updates 2016-09-27

This series contains updates to igb and igbvf.

Wei Yongjun makes a function static to shut up sparse.

Todd bumps the igb and igbvf version, which is long overdue.

Jake fixes an issue where the PPS SYS_WRAP interrupt was not re-enabled
after a reset, which resulted in disabling of the PPS signaling.

v2: dropped patch 5 of the original series, since the PCI quirk patch
    needs to be reworked by Alex and Sasha to address issues that Bjorn
    Helgaas and Alex Williamson brought up.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agolib: clean up put_cpu_var usage
Shaohua Li [Tue, 27 Sep 2016 15:42:42 +0000 (08:42 -0700)]
lib: clean up put_cpu_var usage

put_cpu_var takes the percpu data, not the data returned from
get_cpu_var.

This doesn't change the behavior.

Cc: Tejun Heo <tj@kernel.org>
Signed-off-by: Shaohua Li <shli@fb.com>
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agobpf: clean up put_cpu_var usage
Shaohua Li [Tue, 27 Sep 2016 15:42:41 +0000 (08:42 -0700)]
bpf: clean up put_cpu_var usage

put_cpu_var takes the percpu data, not the data returned from
get_cpu_var.

This doesn't change the behavior.

Cc: Tejun Heo <tj@kernel.org>
Cc: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Shaohua Li <shli@fb.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoigb: restore PPS signal on igb_ptp_reset
Jacob Keller [Fri, 9 Sep 2016 16:10:51 +0000 (09:10 -0700)]
igb: restore PPS signal on igb_ptp_reset

When a reset occurs, the PPS SYS_WRAP interrupt was not re-enabled which
resulted in disabling of the PPS signaling. Fix this by recording when
the interrupt is on and ensuring that we re-enable it every time we
reset.

Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoigb: bump version to igb-5.4.0
Todd Fujinaka [Wed, 24 Aug 2016 15:40:23 +0000 (08:40 -0700)]
igb: bump version to igb-5.4.0

Bump igb version to match other igb drivers.

Signed-off-by: Todd Fujinaka <todd.fujinaka@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoigbvf: bump version to igbvf-2.4.0
Todd Fujinaka [Wed, 24 Aug 2016 15:38:46 +0000 (08:38 -0700)]
igbvf: bump version to igbvf-2.4.0

Bump version to match other igbvf drivers.

Signed-off-by: Todd Fujinaka <todd.fujinaka@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agoigb: fix non static symbol warning
Wei Yongjun [Tue, 23 Aug 2016 15:08:09 +0000 (15:08 +0000)]
igb: fix non static symbol warning

Fixes the following sparse warning:

drivers/net/ethernet/intel/igb/igb_ethtool.c:2707:5: warning:
 symbol 'igb_rxnfc_write_vlan_prio_filter' was not declared. Should it be static?

Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
8 years agobrcmfmac: use correct skb freeing helper when deleting flowring
Rafał Miłecki [Tue, 27 Sep 2016 12:11:04 +0000 (14:11 +0200)]
brcmfmac: use correct skb freeing helper when deleting flowring

Flowrings contain skbs waiting for transmission that were passed to us
by netif. It means we checked every one of them looking for 802.1x
Ethernet type. When deleting flowring we have to use freeing function
that will check for 802.1x type as well.

Freeing skbs without a proper check was leading to counter not being
properly decreased. This was triggering a WARNING every time
brcmf_netdev_wait_pend8021x was called.

Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Acked-by: Arend van Spriel <arend@broadcom.com>
Cc: stable@vger.kernel.org # 4.5+
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
8 years agobrcmfmac: replace WARNING on timeout with a simple error message
Rafał Miłecki [Tue, 27 Sep 2016 10:12:24 +0000 (12:12 +0200)]
brcmfmac: replace WARNING on timeout with a simple error message

Even with timeout increased to 950 ms we get WARNINGs from time to time.
It mostly happens on A-MPDU stalls (e.g. when station goes out of
range). It may take up to 5-10 secods for the firmware to recover and
for that time it doesn't process packets.

It's still useful to have a message on time out as it may indicate some
firmware problem and incorrect key update. Raising a WARNING however
wasn't really that necessary, it doesn't point to any driver bug anymore
and backtrace wasn't much useful.

Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Acked-by: Arend van Spriel <arend.vanspriel@broadcom.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
8 years agortlwifi: Add explicit values to hw_variables enum
Larry Finger [Sat, 24 Sep 2016 16:57:19 +0000 (11:57 -0500)]
rtlwifi: Add explicit values to hw_variables enum

The entries in this enum may be referenced in debug output. Adding explicit
values simplifies the lookup.

Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
8 years agortlwifi: Add HAL_DEF_WOWLAN case to *_get_hw() routines
Larry Finger [Sat, 24 Sep 2016 16:57:18 +0000 (11:57 -0500)]
rtlwifi: Add HAL_DEF_WOWLAN case to *_get_hw() routines

Only rtl8821ae implements WOWLAN; however, the other drivers may receive
a call requesting information about this mode. The other drivers need to
ignore the request rather than logging that the default branch of the
switch statement has been reached.

Reported by: Jean Delvare <jdelvare@suse.de>
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
8 years agortlwifi: Add switch variable to 'switch case not processed' messages
Joe Perches [Fri, 23 Sep 2016 18:27:19 +0000 (11:27 -0700)]
rtlwifi: Add switch variable to 'switch case not processed' messages

Help along debugging by showing what switch/case variable is not
being processed in these messages.

Signed-off-by: Joe Perches <joe@perches.com>
Acked-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
8 years agoMerge tag 'iwlwifi-next-for-kalle-2015-09-26' of git://git.kernel.org/pub/scm/linux...
Kalle Valo [Tue, 27 Sep 2016 15:32:24 +0000 (18:32 +0300)]
Merge tag 'iwlwifi-next-for-kalle-2015-09-26' of git://git./linux/kernel/git/iwlwifi/iwlwifi-next

* Some new HW IDs;
* Small fix in CTDP error handling;
* Remove support for older firmwares;
* Clean-ups and a few simple fixes here and there.

8 years agobnx2x: free the mac filter group list before freeing the cmd
jbaron@akamai.com [Mon, 26 Sep 2016 15:00:44 +0000 (11:00 -0400)]
bnx2x: free the mac filter group list before freeing the cmd

The group list must be freed prior to freeing the command otherwise
we have a use-after-free.

Signed-off-by: Jason Baron <jbaron@akamai.com>
Cc: Yuval Mintz <Yuval.Mintz@qlogic.com>
Cc: Ariel Elior <Ariel.Elior@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoMerge branch 'mediatek-pdam-lro-fixes'
David S. Miller [Tue, 27 Sep 2016 13:41:38 +0000 (09:41 -0400)]
Merge branch 'mediatek-pdam-lro-fixes'

Nelson Chang says:

====================
net: ethernet: mediatek: some bug fixes for PDAM and HW LRO

1) Add to stop PDMA while stopping the frame engine
2) Modify the register settings for LRO relinquishments
3) Jump out from the waiting loop while LRO relinquishments are done
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: ethernet: mediatek: bug fix to disable HW LRO
Nelson Chang [Mon, 26 Sep 2016 06:33:50 +0000 (14:33 +0800)]
net: ethernet: mediatek: bug fix to disable HW LRO

(1) Modify the register settings for LRO relinquishments
(2) Jump out from the waiting loop while LRO relinquishments are done

Signed-off-by: Nelson Chang <nelson.chang@mediatek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: ethernet: mediatek: add to stop PDMA while stopping the frame engine
Nelson Chang [Mon, 26 Sep 2016 06:33:49 +0000 (14:33 +0800)]
net: ethernet: mediatek: add to stop PDMA while stopping the frame engine

Stop PDMA while the frame engine is going to stop.

Signed-off-by: Nelson Chang <nelson.chang@mediatek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoath10k: support up to 64 vdevs
Ben Greear [Mon, 26 Sep 2016 18:56:26 +0000 (21:56 +0300)]
ath10k: support up to 64 vdevs

The (1 << x) - 1 trick won't work when you
are trying to fill up all 64 bits, so add special
case for that.

Signed-off-by: Ben Greear <greearb@candelatech.com>
[kvalo@qca.qualcomm.com: remove the sentence about moving limits]
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
8 years agoath10k: document cycle count related counters
Ben Greear [Mon, 26 Sep 2016 18:56:26 +0000 (21:56 +0300)]
ath10k: document cycle count related counters

They are not necessarily named in an intuitive manner,
so at least add some comments to help the next person.

Signed-off-by: Ben Greear <greearb@candelatech.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
8 years agoath10k: fix typo in logging message
Ben Greear [Mon, 26 Sep 2016 18:56:25 +0000 (21:56 +0300)]
ath10k: fix typo in logging message

Signed-off-by: Ben Greear <greearb@candelatech.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
8 years agoath10k: fix rfc1042 header retrieval in QCA4019 with eth decap mode
Vasanthakumar Thiagarajan [Mon, 26 Sep 2016 18:56:24 +0000 (21:56 +0300)]
ath10k: fix rfc1042 header retrieval in QCA4019 with eth decap mode

Chipset from QCA99X0 onwards (QCA99X0, QCA9984, QCA4019 & future)
rx_hdr_status is not padded to align in 4-byte boundary. Define a
new hw_params field to handle different alignment behaviour between
different hw. This patch fixes improper retrieval of rfc1042 header
with QCA4019. This patch along with "ath10k: Properly remove padding
from the start of rx payload" will fix traffic failure in ethernet
decap mode for QCA4019.

Signed-off-by: Vasanthakumar Thiagarajan <vthiagar@qti.qualcomm.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
8 years agoath10k: do not check if reset is NULL
Masahiro Yamada [Mon, 26 Sep 2016 18:56:24 +0000 (21:56 +0300)]
ath10k: do not check if reset is NULL

Since reset_control_get() never returns NULL, we can use IS_ERR()
instead of IS_ERR_OR_NULL().  The return statements can be simpler
as well.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
8 years agoath10k: use devm_reset_control_get() instead of reset_control_get()
Masahiro Yamada [Mon, 26 Sep 2016 18:56:23 +0000 (21:56 +0300)]
ath10k: use devm_reset_control_get() instead of reset_control_get()

Use the managed variant of reset_control_get() to simplify the
failure path and the .remove callback.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
8 years agoath10k: use devm_clk_get() instead of clk_get()
Masahiro Yamada [Mon, 26 Sep 2016 18:56:22 +0000 (21:56 +0300)]
ath10k: use devm_clk_get() instead of clk_get()

Use the managed variant of clk_get() to simplify the failure path
and the .remove callback.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
8 years agoath10k: spelling and miscellaneous neatening
Joe Perches [Mon, 26 Sep 2016 18:56:21 +0000 (21:56 +0300)]
ath10k: spelling and miscellaneous neatening

Correct some trivial comment typos.
Remove unnecessary parentheses in a long line.

Signed-off-by: Joe Perches <joe@perches.com>
[kvalo@qca.qualcomm.com: drop the change for return]
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
8 years agoMerge branch 'bcmgenet-phydev-revert'
David S. Miller [Tue, 27 Sep 2016 11:43:13 +0000 (07:43 -0400)]
Merge branch 'bcmgenet-phydev-revert'

Philippe Reynes says:

====================
net: bcmgenet: only use new api ethtool_{get|set}_link_ksettings

Some times ago, a serie of patches were committed :
- commit 62469c76007e ("net: ethernet: bcmgenet: use phydev from struct net_device")
- commit 6b352ebccbcf ("net: ethernet: broadcom: bcmgenet: use new api ethtool_{get|set}_link_ksettings")
The first patch add a regression on this driver, so it should be reverted.
As the second patch depend on the former, it should be reverted too.

The first patch is buggy because there is a "trick" in this driver.
The structure phydev is kept in the private data when the interface
go down, and used when the interface go up to enable the phy before
the function phy_connect is called.

I don't have this hardware, neither the datasheet. So I won't
update the driver to avoid this trick.

But the real goal of the first serie was to move to the new api
ethtool_{get|set}_link_ksettings. So I provide a new version of
the patch without the "cleaning" of driver to use the phydev
store in the net_device structure.

Changelog:
v3:
- use priv instead of dev (so all the code use the same phydev)
v2:
- use Florian Fainelli patches for the revert instead of Jaedon Shin
- simply use net: bcmgenet: for the prefix of the patch
====================

Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: bcmgenet: use new api ethtool_{get|set}_link_ksettings
Philippe Reynes [Mon, 26 Sep 2016 20:31:57 +0000 (22:31 +0200)]
net: bcmgenet: use new api ethtool_{get|set}_link_ksettings

The ethtool api {get|set}_settings is deprecated.
We move this driver to new api {get|set}_link_ksettings.

Signed-off-by: Philippe Reynes <tremyfr@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoRevert "net: ethernet: bcmgenet: use phydev from struct net_device"
Florian Fainelli [Mon, 26 Sep 2016 20:31:56 +0000 (22:31 +0200)]
Revert "net: ethernet: bcmgenet: use phydev from struct net_device"

This reverts commit 62469c76007e ("net: ethernet: bcmgenet: use phydev
from struct net_device") because it causes GENETv1/2/3 adapters to
expose the following behavior after an ifconfig down/up sequence:

PING fainelli-linux (10.112.156.244): 56 data bytes
64 bytes from 10.112.156.244: seq=1 ttl=61 time=1.352 ms
64 bytes from 10.112.156.244: seq=1 ttl=61 time=1.472 ms (DUP!)
64 bytes from 10.112.156.244: seq=1 ttl=61 time=1.496 ms (DUP!)
64 bytes from 10.112.156.244: seq=1 ttl=61 time=1.517 ms (DUP!)
64 bytes from 10.112.156.244: seq=1 ttl=61 time=1.536 ms (DUP!)
64 bytes from 10.112.156.244: seq=1 ttl=61 time=1.557 ms (DUP!)
64 bytes from 10.112.156.244: seq=1 ttl=61 time=752.448 ms (DUP!)

This was previously fixed by commit 5dbebbb44a6a ("net: bcmgenet:
Software reset EPHY after power on") but the commit we are reverting was
essentially making this previous commit void, here is why.

Without commit 62469c76007e we would have the following scenario after
an ifconfig down then up sequence:

- bcmgenet_open() calls bcmgenet_power_up() to make sure the PHY is
  initialized *before* we get to initialize the UniMAC, this is
  critical to ensure the PHY is in a correct state, priv->phydev is
  valid, this code executes fine

- second time from bcmgenet_mii_probe(), through the normal
  phy_init_hw() call (which arguably could be optimized out)

Everything is fine in that case. With commit 62469c76007e, we would have
the following scenario to happen after an ifconfig down then up
sequence:

- bcmgenet_close() calls phy_disonnect() which makes dev->phydev become
  NULL

- when bcmgenet_open() executes again and calls bcmgenet_mii_reset() from
  bcmgenet_power_up() to initialize the internal PHY, the NULL check
  becomes true, so we do not reset the PHY, yet we keep going on and
  initialize the UniMAC, causing MAC activity to occur

- we call bcmgenet_mii_reset() from bcmgenet_mii_probe(), but this is
  too late, the PHY is botched, and causes the above bogus pings/packets
  transmission/reception to occur

Reported-by: Jaedon Shin <jaedon.shin@gmail.com>
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: Philippe Reynes <tremyfr@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoRevert "net: ethernet: bcmgenet: use new api ethtool_{get|set}_link_ksettings"
Philippe Reynes [Mon, 26 Sep 2016 20:31:55 +0000 (22:31 +0200)]
Revert "net: ethernet: bcmgenet: use new api ethtool_{get|set}_link_ksettings"

This reverts commit 6b352ebccbcf ("net: ethernet: broadcom: bcmgenet:
use new api ethtool_{get|set}_link_ksettings").

We needs to revert the commit 62469c76007e ("net: ethernet: bcmgenet:
use phydev from struct net_device"), because this commit add a
regression. As the commit 6b352ebccbcf ("net: ethernet: broadcom:
bcmgenet: use new api ethtool_{get|set}_link_ksettings") depend
on the first one, we also need to revert it first.

Signed-off-by: Philippe Reynes <tremyfr@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agobpf: Set register type according to is_valid_access()
Mickaël Salaün [Sat, 24 Sep 2016 18:01:50 +0000 (20:01 +0200)]
bpf: Set register type according to is_valid_access()

This prevent future potential pointer leaks when an unprivileged eBPF
program will read a pointer value from its context. Even if
is_valid_access() returns a pointer type, the eBPF verifier replace it
with UNKNOWN_VALUE. The register value that contains a kernel address is
then allowed to leak. Moreover, this fix allows unprivileged eBPF
programs to use functions with (legitimate) pointer arguments.

Not an issue currently since reg_type is only set for PTR_TO_PACKET or
PTR_TO_PACKET_END in XDP and TC programs that can only be loaded as
privileged. For now, the only unprivileged eBPF program allowed is for
socket filtering and all the types from its context are UNKNOWN_VALUE.
However, this fix is important for future unprivileged eBPF programs
which could use pointers in their context.

Signed-off-by: Mickaël Salaün <mic@digikod.net>
Cc: Alexei Starovoitov <ast@kernel.org>
Cc: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agobpf samples: update tracex5 sample to use __seccomp_filter
Naveen N. Rao [Fri, 23 Sep 2016 20:40:05 +0000 (02:10 +0530)]
bpf samples: update tracex5 sample to use __seccomp_filter

seccomp_phase1() does not exist anymore. Instead, update sample to use
__seccomp_filter(). While at it, set max locked memory to unlimited.

Signed-off-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agobpf samples: fix compiler errors with sockex2 and sockex3
Naveen N. Rao [Fri, 23 Sep 2016 20:40:04 +0000 (02:10 +0530)]
bpf samples: fix compiler errors with sockex2 and sockex3

These samples fail to compile as 'struct flow_keys' conflicts with
definition in net/flow_dissector.h. Fix the same by renaming the
structure used in the sample.

Signed-off-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: tg3: use new api ethtool_{get|set}_link_ksettings
Philippe Reynes [Sun, 25 Sep 2016 21:31:24 +0000 (23:31 +0200)]
net: tg3: use new api ethtool_{get|set}_link_ksettings

The ethtool api {get|set}_settings is deprecated.
We move this driver to new api {get|set}_link_ksettings.

Signed-off-by: Philippe Reynes <tremyfr@gmail.com>
Acked-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoiwlwifi: mvm: initialise ADD_STA before sending it to the firmware
Emmanuel Grumbach [Tue, 20 Sep 2016 10:40:33 +0000 (13:40 +0300)]
iwlwifi: mvm: initialise ADD_STA before sending it to the firmware

When we unshare a queue, the ADD_STA was not properly
initialised.

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
8 years agoiwlwifi: allow error table address new range
Sara Sharon [Thu, 8 Sep 2016 07:44:38 +0000 (10:44 +0300)]
iwlwifi: allow error table address new range

The firmware has a new smart linker, and this table can now be
in ICCM or in SMEM. It is not hardcoded, but depends on code
size. Allow the full range.

Signed-off-by: Sara Sharon <sara.sharon@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
8 years agoiwlwifi: mvm: fix typo in TC_CMD_SEC_KEY_FROM_TABLE
Emmanuel Grumbach [Tue, 13 Sep 2016 19:59:27 +0000 (22:59 +0300)]
iwlwifi: mvm: fix typo in TC_CMD_SEC_KEY_FROM_TABLE

This define should really be TX_CMD_SEC_KEY_FROM_TABLE

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
8 years agoiwlwifi: remove support for fw older than -17 and -22
Luca Coelho [Mon, 12 Sep 2016 13:03:30 +0000 (16:03 +0300)]
iwlwifi: remove support for fw older than -17 and -22

FW versions older than -17 for 3160 and 7260 and older than -22 for
newer NICs are not supported anymore.  Don't load these versions
and remove code that handles them.

Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
8 years agobrcmfmac: drop unused fields from struct brcmf_pub
Rafał Miłecki [Fri, 23 Sep 2016 13:27:46 +0000 (15:27 +0200)]
brcmfmac: drop unused fields from struct brcmf_pub

They seem to be there from the first day. We calculate these values but
never use them.

Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
8 years agobrcmfmac: fix memory leak in brcmf_fill_bss_param
Rafał Miłecki [Wed, 21 Sep 2016 06:23:24 +0000 (08:23 +0200)]
brcmfmac: fix memory leak in brcmf_fill_bss_param

This function is called from get_station callback which means that every
time user space was getting/dumping station(s) we were leaking 2 KiB.

Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Fixes: 1f0dc59a6de ("brcmfmac: rework .get_station() callback")
Cc: stable@vger.kernel.org # 4.2+
Acked-by: Arend van Spriel <arend.vanspriel@broadcom.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
8 years agortl8xxxu: Stop log spam from each successful interrupt
Larry Finger [Wed, 21 Sep 2016 01:19:29 +0000 (21:19 -0400)]
rtl8xxxu: Stop log spam from each successful interrupt

As soon as debugging is turned on, the logs are filled with messages
reporting the interrupt status. As this quantity is usually zero, this
output is not needed. In fact, there will be a report if the status is
not zero, thus the debug line in question could probably be deleted.
Rather than taking that action, I have changed it to only be printed
when the newly added RTL8XXXU_DEBUG_INTERRUPT bit is set in the debug
mask.

Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
8 years agortl8xxxu: Use a struct rtl8xxxu_fileops * in rtl8xxxu_init_device()
Jes Sorensen [Wed, 21 Sep 2016 01:19:28 +0000 (21:19 -0400)]
rtl8xxxu: Use a struct rtl8xxxu_fileops * in rtl8xxxu_init_device()

This saves some 217, or about, derefences of priv->fops.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
8 years agortl8xxxu: Clean up llt_init() API
Jes Sorensen [Wed, 21 Sep 2016 01:19:27 +0000 (21:19 -0400)]
rtl8xxxu: Clean up llt_init() API

Remove last_tx_page argument from the llt_init() function. The
rtl8xxxu_fileops structure contains the correct TX_TOTAL_PAGE_NUM
value for the device, and rtl8xxxu_auto_llt_table() doesn't need to
know the value in the first place.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
8 years agortl8xxxu: Fix off by one error calculating pubq
Jes Sorensen [Wed, 21 Sep 2016 01:19:26 +0000 (21:19 -0400)]
rtl8xxxu: Fix off by one error calculating pubq

This was detected tracing the 8188eu driver, but doesn't seem to make
any difference when using it.

Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
8 years agomwifiex: code rearrangement in mwifiex_usb_host_to_card()
Amitkumar Karwar [Tue, 20 Sep 2016 15:19:04 +0000 (20:49 +0530)]
mwifiex: code rearrangement in mwifiex_usb_host_to_card()

This patch helps get rid of goto statement and improves readability.

Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Cathy Luo <cluo@marvell.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
8 years agomwifiex: fix race condition causing tx timeout
Cathy Luo [Tue, 20 Sep 2016 15:19:03 +0000 (20:49 +0530)]
mwifiex: fix race condition causing tx timeout

It's been observed that in a corner case mwifiex_usb_tx_complete()
gets called before we exit from mwifiex_usb_host_to_card() after
submitting the urb. 'data_sent' flag remains set in this case. It
blocks further Tx packets and triggers watchdog timeout.

The problem is fixed by setting data_sent and port_block flag at
correct place.

Signed-off-by: Cathy Luo <cluo@marvell.com>
Signed-off-by: Shengzhen Li <szli@marvell.com>
Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
8 years agomwifiex: fix kernel crash for USB chipsets
Cathy Luo [Tue, 20 Sep 2016 15:19:02 +0000 (20:49 +0530)]
mwifiex: fix kernel crash for USB chipsets

Following crash issue is observed during TCP traffic stress
test

[ 2253.625439] NMI watchdog: BUG: soft lockup - CPU#3 stuck for 22s!
[kworker/u17:1:5191]
[ 2253.625520] Call Trace:
[ 2253.625527]  [<ffffffffc0b47030>] ? moal_spin_lock+0x30/0x30
[usb8xxx]
[ 2253.625533]  [<ffffffffc0ac3ceb>] ? wlan_wmm_lists_empty+0xb/0xf0
[mlan]
[ 2253.625537]  [<ffffffffc0ab0ea3>] mlan_main_process+0x1b3/0x720
[mlan]
[ 2253.625540]  [<ffffffffc0b337f5>] woal_main_work_queue+0x45/0x80
[usb8xxx]
[ 2253.625543]  [<ffffffff8108aaf0>] process_one_work+0x150/0x3f0
[ 2253.625545]  [<ffffffff8108b1e1>] worker_thread+0x121/0x520
[ 2253.625547]  [<ffffffff8108b0c0>] ? rescuer_thread+0x330/0x330
[ 2253.625549]  [<ffffffff81090222>] kthread+0xd2/0xf0
[ 2253.625551]  [<ffffffff81090150>] ?
kthread_create_on_node+0x1c0/0x1c0
[ 2253.625553]  [<ffffffff8179423c>] ret_from_fork+0x7c/0xb0
[ 2253.625555]  [<ffffffff81090150>] ?
kthread_create_on_node+0x1c0/0x1c0

In mwifiex_usb_tx_complete(), we are updating port->block_status first
and then freeing the skb attached to that URB. We may end up attaching
new skb to URB in a corner case and same will be freed. This results in
the kernel crash. The problem is solved by changing the sequence.

Signed-off-by: Cathy Luo <cluo@marvell.com>
Signed-off-by: Shengzhen Li <szli@marvell.com>
Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>