openwrt/staging/blogic.git
5 years agonet: hns3: prevent double free in hns3_put_ring_config()
Huazhong Tan [Thu, 25 Apr 2019 12:42:54 +0000 (20:42 +0800)]
net: hns3: prevent double free in hns3_put_ring_config()

This patch adds a check for the hns3_put_ring_config() to prevent
double free, and for more readable, move the NULL assignment of
priv->ring_data into the hns3_put_ring_config().

Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: hns3: extend the loopback state acquisition time
liuzhongzhu [Thu, 25 Apr 2019 12:42:53 +0000 (20:42 +0800)]
net: hns3: extend the loopback state acquisition time

The test results show that the maximum time of hardware return
to mac link state is 500MS.The software needs to set twice the
maximum time of hardware return state (1000MS).

If not modified, the loopback test returns probability failure.

Signed-off-by: liuzhongzhu <liuzhongzhu@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: hns3: fix pause configure fail problem
Huazhong Tan [Thu, 25 Apr 2019 12:42:52 +0000 (20:42 +0800)]
net: hns3: fix pause configure fail problem

When configure pause, current implementation returns directly
after setup PFC without setup BP, which is not sufficient.

So this patch fixes it, only return while setting PFC failed.

Fixes: 44e59e375bf7 ("net: hns3: do not return GE PFC setting err when initializing")
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: hns3: not reset TQP in the DOWN while VF resetting
Huazhong Tan [Thu, 25 Apr 2019 12:42:51 +0000 (20:42 +0800)]
net: hns3: not reset TQP in the DOWN while VF resetting

Since the hardware does not handle mailboxes and the hardware
reset include TQP reset, so it is unnecessary to reset TQP
in the hclgevf_ae_stop() while doing VF reset. Also it is
unnecessary to reset the remaining TQP when one reset fails.

Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: hns3: use a reserved byte to identify need_resp flag
Huazhong Tan [Thu, 25 Apr 2019 12:42:50 +0000 (20:42 +0800)]
net: hns3: use a reserved byte to identify need_resp flag

This patch uses a reserved byte in the hclge_mbx_vf_to_pf_cmd
to save the need_resp flag, so when PF received the mailbox,
it can use it to decise whether send a response to VF.

For hclge_set_vf_uc_mac_addr(), it should use mbx_need_resp flag
to decide whether send response to VF.

Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: hns3: use atomic_t replace u32 for arq's count
Huazhong Tan [Thu, 25 Apr 2019 12:42:49 +0000 (20:42 +0800)]
net: hns3: use atomic_t replace u32 for arq's count

Since irq handler and mailbox task will both update arq's count,
so arq's count should use atomic_t instead of u32, otherwise
its value may go wrong finally.

Fixes: 07a0556a3a73 ("net: hns3: Changes to support ARQ(Asynchronous Receive Queue)")
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: hns3: stop sending keep alive msg when VF command queue needs reinit
Huazhong Tan [Thu, 25 Apr 2019 12:42:48 +0000 (20:42 +0800)]
net: hns3: stop sending keep alive msg when VF command queue needs reinit

HCLGEVF_STATE_CMD_DISABLE is more suitable than
HCLGEVF_STATE_RST_HANDLING to stop sending keep alive msg,
since HCLGEVF_STATE_RST_HANDLING only be set when the reset
task is running.

Fixes: c59a85c07e77 ("net: hns3: stop sending keep alive msg to PF when VF is resetting")
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: hns3: handle the BD info on the last BD of the packet
Yunsheng Lin [Thu, 25 Apr 2019 12:42:47 +0000 (20:42 +0800)]
net: hns3: handle the BD info on the last BD of the packet

The bdinfo handled in hns3_handle_bdinfo is only valid on the
last BD of the current packet, currently the bd info may be handled
based on the first BD if the packet has more than two BDs, which
may cause rx error.

This patch fixes it by using the last BD of the current packet in
hns3_handle_bdinfo.

Also, hns3_set_rx_skb_rss_type has used RSS hash value from the last
BD of the current packet, so remove the same last BD calculation in
hns3_set_rx_skb_rss_type and call it from hns3_handle_bdinfo.

Fixes: e55970950556 ("net: hns3: Add handling of GRO Pkts not fully RX'ed in NAPI poll")
Signed-off-by: Yunsheng Lin <linyunsheng@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: hns3: fix for TX clean num when cleaning TX BD
Yunsheng Lin [Thu, 25 Apr 2019 12:42:46 +0000 (20:42 +0800)]
net: hns3: fix for TX clean num when cleaning TX BD

hns3_desc_unused() returns how many BD have been cleaned, but new
buffer has not been attached to them. The register of
HNS3_RING_RX_RING_FBDNUM_REG returns how many BD need allocating new
buffer to or need to cleaned. So the remaining BD need to be clean
is HNS3_RING_RX_RING_FBDNUM_REG - hns3_desc_unused().

Also, new buffer can not attach to the pending BD when the last BD is
not handled, because memcpy has not been done on the first pending BD.

This patch fixes by subtracting the pending BD num from unused_count
after 'HNS3_RING_RX_RING_FBDNUM_REG - unused_count' is used to calculate
the BD bum need to be clean.

Fixes: e55970950556 ("net: hns3: Add handling of GRO Pkts not fully RX'ed in NAPI poll")
Signed-off-by: Yunsheng Lin <linyunsheng@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: hns3: fix data race between ring->next_to_clean
Yunsheng Lin [Thu, 25 Apr 2019 12:42:45 +0000 (20:42 +0800)]
net: hns3: fix data race between ring->next_to_clean

hns3_clean_tx_ring calls hns3_nic_reclaim_one_desc to clean
buffers and set ring->next_to_clean, then hns3_nic_net_xmit
reuses the cleaned buffers. But there are no memory barriers
when buffers gets recycled, so the recycled buffers can be
corrupted.

This patch uses smp_store_release to update ring->next_to_clean
and smp_load_acquire to read ring->next_to_clean to properly
hand off buffers from hns3_clean_tx_ring to hns3_nic_net_xmit.

Fixes: 76ad4f0ee747 ("net: hns3: Add support of HNS3 Ethernet Driver for hip08 SoC")
Signed-off-by: Yunsheng Lin <linyunsheng@huawei.com>
Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonfp: implement PCI driver shutdown callback
Dirk van der Merwe [Thu, 25 Apr 2019 03:18:02 +0000 (20:18 -0700)]
nfp: implement PCI driver shutdown callback

Device may be shutdown without the hardware being reinitialized, in
which case we want to ensure we cleanup properly.

This is especially important for kexec with traffic flowing.

The shutdown procedures resembles the remove procedures, so we can reuse
those common tasks.

Signed-off-by: Dirk van der Merwe <dirk.vandermerwe@netronome.com>
Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoNFC: st95hf: remove set but not used variables 'dev, nfcddev'
YueHaibing [Thu, 25 Apr 2019 02:07:20 +0000 (02:07 +0000)]
NFC: st95hf: remove set but not used variables 'dev, nfcddev'

Fixes gcc '-Wunused-but-set-variable' warning:

drivers/nfc/st95hf/core.c: In function 'st95hf_irq_thread_handler':
drivers/nfc/st95hf/core.c:786:26: warning:
 variable 'nfcddev' set but not used [-Wunused-but-set-variable]

drivers/nfc/st95hf/core.c:784:17: warning:
 variable 'dev' set but not used [-Wunused-but-set-variable]

They are never used since introduction in
commit cab47333f0f7 ("NFC: Add STMicroelectronics ST95HF driver")

Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agotipc: remove rcu_read_unlock() left in tipc_udp_recv()
Eric Dumazet [Thu, 25 Apr 2019 00:21:40 +0000 (17:21 -0700)]
tipc: remove rcu_read_unlock() left in tipc_udp_recv()

I forgot to remove one rcu_read_unlock() before a return statement.

Joy of mixing goto and return styles in a function :)

Fixes: 4109a2c3b91e ("tipc: tipc_udp_recv() cleanup vs rcu verbs")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: kbuild test robot <lkp@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: phy: improve genphy_soft_reset
Heiner Kallweit [Wed, 24 Apr 2019 19:41:06 +0000 (21:41 +0200)]
net: phy: improve genphy_soft_reset

PHY's behave differently when being reset. Some reset registers to
defaults, some don't. Some trigger an autoneg restart, some don't.

So let's also set the autoneg restart bit when resetting. Then PHY
behavior should be more consistent. Clearing BMCR_ISOLATE serves the
same purpose and is borrowed from genphy_restart_aneg.

BMCR holds the speed / duplex settings in fixed mode. Therefore
we may have an issue if a soft reset resets BMCR to its default.
So better call genphy_setup_forced() afterwards in fixed mode.
We've seen no related complaint in the last >10 yrs, so let's
treat it as an improvement.

Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agousbnet: ipheth: Simplify device detection
Guenter Roeck [Wed, 24 Apr 2019 17:58:24 +0000 (10:58 -0700)]
usbnet: ipheth: Simplify device detection

All Apple products use the same protocol for tethering over USB.
To simplify the code and make it future proof, use
USB_VENDOR_AND_INTERFACE_INFO() instead of
USB_DEVICE_AND_INTERFACE_INFO() to automatically detect and support
all existing and future Apple products using the same interface.

Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoipv6: Initialize fib6_result in bpf_ipv6_fib_lookup
David Ahern [Wed, 24 Apr 2019 17:36:06 +0000 (10:36 -0700)]
ipv6: Initialize fib6_result in bpf_ipv6_fib_lookup

fib6_result is not initialized in bpf_ipv6_fib_lookup and potentially
passses garbage to the fib lookup which triggers a KASAN warning:

[  262.055450] ==================================================================
[  262.057640] BUG: KASAN: user-memory-access in fib6_rule_suppress+0x4b/0xce
[  262.059488] Read of size 8 at addr 00000a20000000b0 by task swapper/1/0
[  262.061238]
[  262.061673] CPU: 1 PID: 0 Comm: swapper/1 Not tainted 5.1.0-rc5+ #56
[  262.063493] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.11.1-1 04/01/2014
[  262.065593] Call Trace:
[  262.066277]  <IRQ>
[  262.066848]  dump_stack+0x7e/0xbb
[  262.067764]  kasan_report+0x18b/0x1b5
[  262.069921]  __asan_load8+0x7f/0x81
[  262.070879]  fib6_rule_suppress+0x4b/0xce
[  262.071980]  fib_rules_lookup+0x275/0x2cd
[  262.073090]  fib6_lookup+0x119/0x218
[  262.076457]  bpf_ipv6_fib_lookup+0x39d/0x664
...

Initialize fib6_result to 0.

Fixes: b1d40991506aa ("ipv6: Rename fib6_multipath_select and pass fib6_result")
Signed-off-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agocnic: Refactor code and mark expected switch fall-through
Gustavo A. R. Silva [Wed, 24 Apr 2019 16:37:32 +0000 (11:37 -0500)]
cnic: Refactor code and mark expected switch fall-through

In preparation to enabling -Wimplicit-fallthrough, refactor code a
bit and mark switch cases where we are expecting to fall through.

This patch fixes the following warning:

drivers/net/ethernet/broadcom/cnic.c: In function ‘cnic_cm_process_kcqe’:
drivers/net/ethernet/broadcom/cnic.c:4044:11: warning: this statement may fall through [-Wimplicit-fallthrough=]
    opcode = L4_KCQE_OPCODE_VALUE_CLOSE_COMP;
drivers/net/ethernet/broadcom/cnic.c:4050:2: note: here
  case L4_KCQE_OPCODE_VALUE_RESET_RECEIVED:
  ^~~~

Warning level 3 was used: -Wimplicit-fallthrough=3

Notice that, in this particular case, the code comment is modified
in accordance with what GCC is expecting to find.

This patch is part of the ongoing efforts to enable
-Wimplicit-fallthrough.

Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agocxgb4/cxgb4vf_main: Mark expected switch fall-through
Gustavo A. R. Silva [Wed, 24 Apr 2019 16:27:42 +0000 (11:27 -0500)]
cxgb4/cxgb4vf_main: Mark expected switch fall-through

In preparation to enabling -Wimplicit-fallthrough, mark switch
cases where we are expecting to fall through.

This patch fixes the following warning:

drivers/net/ethernet/chelsio/cxgb4vf/cxgb4vf_main.c: In function ‘fwevtq_handler’:
drivers/net/ethernet/chelsio/cxgb4vf/cxgb4vf_main.c:520:7: warning: this statement may fall through [-Wimplicit-fallthrough=]
   cpl = (void *)p;
   ~~~~^~~~~~~~~~~
drivers/net/ethernet/chelsio/cxgb4vf/cxgb4vf_main.c:524:2: note: here
  case CPL_SGE_EGR_UPDATE: {
  ^~~~

Warning level 3 was used: -Wimplicit-fallthrough=3

Notice that, in this particular case, the code comment is modified
in accordance with what GCC is expecting to find.

This patch is part of the ongoing efforts to enable
-Wimplicit-fallthrough.

Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agowimax/i2400m/control: Mark expected switch fall-through
Gustavo A. R. Silva [Wed, 24 Apr 2019 16:20:16 +0000 (11:20 -0500)]
wimax/i2400m/control: Mark expected switch fall-through

In preparation to enabling -Wimplicit-fallthrough, mark switch
cases where we are expecting to fall through.

This patch fixes the following warning:

In file included from drivers/net/wimax/i2400m/debug-levels.h:30,
                 from drivers/net/wimax/i2400m/control.c:86:
drivers/net/wimax/i2400m/control.c: In function ‘i2400m_report_tlv_system_state’:
./include/linux/wimax/debug.h:200:4: warning: this statement may fall through [-Wimplicit-fallthrough=]
 do {         \
    ^
./include/linux/wimax/debug.h:404:36: note: in expansion of macro ‘_d_printf’
 #define d_printf(l, _dev, f, a...) _d_printf(l, "", _dev, f, ## a)
                                    ^~~~~~~~~
drivers/net/wimax/i2400m/control.c:354:3: note: in expansion of macro ‘d_printf’
   d_printf(1, dev, "entering BS-negotiated idle mode\n");
   ^~~~~~~~
drivers/net/wimax/i2400m/control.c:355:2: note: here
  case I2400M_SS_DISCONNECTING:
  ^~~~

Warning level 3 was used: -Wimplicit-fallthrough=3

This patch is part of the ongoing efforts to enable
-Wimplicit-fallthrough.

Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoamd-xgbe: Mark expected switch fall-throughs
Gustavo A. R. Silva [Wed, 24 Apr 2019 16:08:24 +0000 (11:08 -0500)]
amd-xgbe: Mark expected switch fall-throughs

In preparation to enabling -Wimplicit-fallthrough, mark switch
cases where we are expecting to fall through.

This patch fixes the following warnings:

In file included from drivers/net/ethernet/amd/xgbe/xgbe-drv.c:129:
drivers/net/ethernet/amd/xgbe/xgbe-drv.c: In function ‘xgbe_set_hwtstamp_settings’:
drivers/net/ethernet/amd/xgbe/xgbe-common.h:1392:9: warning: this statement may fall through [-Wimplicit-fallthrough=]
  (_var) |= (((_val) & ((0x1 << (_width)) - 1)) << (_index)); \
  ~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
drivers/net/ethernet/amd/xgbe/xgbe-common.h:1419:2: note: in expansion of macro ‘SET_BITS’
  SET_BITS((_var),      \
  ^~~~~~~~
drivers/net/ethernet/amd/xgbe/xgbe-drv.c:1614:3: note: in expansion of macro ‘XGMAC_SET_BITS’
   XGMAC_SET_BITS(mac_tscr, MAC_TSCR, TSVER2ENA, 1);
   ^~~~~~~~~~~~~~
drivers/net/ethernet/amd/xgbe/xgbe-drv.c:1616:2: note: here
  case HWTSTAMP_FILTER_PTP_V1_L4_EVENT:
  ^~~~
In file included from drivers/net/ethernet/amd/xgbe/xgbe-drv.c:129:
drivers/net/ethernet/amd/xgbe/xgbe-common.h:1392:9: warning: this statement may fall through [-Wimplicit-fallthrough=]
  (_var) |= (((_val) & ((0x1 << (_width)) - 1)) << (_index)); \
  ~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
drivers/net/ethernet/amd/xgbe/xgbe-common.h:1419:2: note: in expansion of macro ‘SET_BITS’
  SET_BITS((_var),      \
  ^~~~~~~~
drivers/net/ethernet/amd/xgbe/xgbe-drv.c:1625:3: note: in expansion of macro ‘XGMAC_SET_BITS’
   XGMAC_SET_BITS(mac_tscr, MAC_TSCR, TSVER2ENA, 1);
   ^~~~~~~~~~~~~~
drivers/net/ethernet/amd/xgbe/xgbe-drv.c:1627:2: note: here
  case HWTSTAMP_FILTER_PTP_V1_L4_SYNC:
  ^~~~
In file included from drivers/net/ethernet/amd/xgbe/xgbe-drv.c:129:
drivers/net/ethernet/amd/xgbe/xgbe-common.h:1392:9: warning: this statement may fall through [-Wimplicit-fallthrough=]
  (_var) |= (((_val) & ((0x1 << (_width)) - 1)) << (_index)); \
  ~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
drivers/net/ethernet/amd/xgbe/xgbe-common.h:1419:2: note: in expansion of macro ‘SET_BITS’
  SET_BITS((_var),      \
  ^~~~~~~~
drivers/net/ethernet/amd/xgbe/xgbe-drv.c:1636:3: note: in expansion of macro ‘XGMAC_SET_BITS’
   XGMAC_SET_BITS(mac_tscr, MAC_TSCR, TSVER2ENA, 1);
   ^~~~~~~~~~~~~~
drivers/net/ethernet/amd/xgbe/xgbe-drv.c:1638:2: note: here
  case HWTSTAMP_FILTER_PTP_V1_L4_DELAY_REQ:
  ^~~~

Warning level 3 was used: -Wimplicit-fallthrough=3

Notice that, in this particular case, the code comments are modified
in accordance with what GCC is expecting to find.

This patch is part of the ongoing efforts to enable
-Wimplicit-fallthrough.

Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: socket: Fix missing break in switch statement
Gustavo A. R. Silva [Wed, 24 Apr 2019 15:31:24 +0000 (10:31 -0500)]
net: socket: Fix missing break in switch statement

Add missing break statement in order to prevent the code from falling
through to cases SIOCGSTAMP_NEW and SIOCGSTAMPNS_NEW.

This bug was found thanks to the ongoing efforts to enable
-Wimplicit-fallthrough.

Fixes: 0768e17073dc ("net: socket: implement 64-bit timestamps")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoMerge branch 's390-qeth-cleanups'
David S. Miller [Fri, 26 Apr 2019 15:14:06 +0000 (11:14 -0400)]
Merge branch 's390-qeth-cleanups'

Julian Wiedmann says:

====================
s390/qeth: updates 2019-04-25

please apply one more patch series for qeth to net-next. Nothing special,
just a bunch of cleanups.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agos390/qeth: trust non-IP cast type in qeth_l3_fill_header()
Julian Wiedmann [Thu, 25 Apr 2019 16:26:01 +0000 (18:26 +0200)]
s390/qeth: trust non-IP cast type in qeth_l3_fill_header()

When building the L3 HW header for non-IP packets, trust the cast type
that was passed as parameter. qeth_l3_get_cast_type() has most likely
also used h_dest to determine the cast type, so we get the same
result, and can remove that duplicated code.
In the unlikely case that we would get a _different_ cast type, then
that's based off a route lookup and should be considered authoritative.

Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agos390/qeth: extract helper to determine L2 cast type
Julian Wiedmann [Thu, 25 Apr 2019 16:26:00 +0000 (18:26 +0200)]
s390/qeth: extract helper to determine L2 cast type

This de-duplicates the L2 and L3 cast-type code, and makes the L2 code
a bit more robust by removing the fragile assumption that skb->data
always points to the Ethernet Header. This would break in code paths
where we pushed the HW header onto the skb.

Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agos390/qeth: cache max number of available buffer elements
Julian Wiedmann [Thu, 25 Apr 2019 16:25:59 +0000 (18:25 +0200)]
s390/qeth: cache max number of available buffer elements

The QETH_MAX_BUFFER_ELEMENTS() macro effectively returns a constant
value. To avoid some redundant pointer chasing and computations in the
xmit hot path, cache this value in the queue struct.

Take this as opportunity to shrink some of the queue struct's fields to
their appropriate value range, slightly reducing its total size.

Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agos390/qeth: don't clear Output buffers on every queue init
Julian Wiedmann [Thu, 25 Apr 2019 16:25:58 +0000 (18:25 +0200)]
s390/qeth: don't clear Output buffers on every queue init

On the first initialization of a queue, its Output Buffers are in a
clean state with no attached resources. On every subsequent
initialization, qeth_l?_stop_card() has previously put them in a clean
state via qeth_drain_output_queues(). So the call to
qeth_clear_output_buffer() is redundant and can be removed.

While at it, move the initialization of the queue's card pointer into
the queue allocation. It never changes afterwards.

Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agos390/qeth: use IS_* helpers for checking device type
Julian Wiedmann [Thu, 25 Apr 2019 16:25:57 +0000 (18:25 +0200)]
s390/qeth: use IS_* helpers for checking device type

We have helper macros for all possible device types, replace all
remaining open-coded accesses to the type fields.

Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agos390/qeth: clean up stale buffer state documentation
Julian Wiedmann [Thu, 25 Apr 2019 16:25:56 +0000 (18:25 +0200)]
s390/qeth: clean up stale buffer state documentation

We don't keep track of Input Buffer states, so remove the comments that
make it sound like the qeth_qdio_buffer_states enum applies to
Input Buffers.

Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agos390/qeth: remove RX seqno in skb->cb
Julian Wiedmann [Thu, 25 Apr 2019 16:25:55 +0000 (18:25 +0200)]
s390/qeth: remove RX seqno in skb->cb

It's unclear what exact purpose this seqno may have served in the past.
But it's certainly no longer used anymore, as the following
napi_gro_receive() will straight away clear this part of the cb again.

Suggested-by: Karsten Graul <kgraul@linux.ibm.com>
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agos390: qeth: address type mismatch warning
Arnd Bergmann [Thu, 25 Apr 2019 16:25:54 +0000 (18:25 +0200)]
s390: qeth: address type mismatch warning

clang produces a harmless warning for each use for the qeth_adp_supported
macro:

drivers/s390/net/qeth_l2_main.c:559:31: warning: implicit conversion from enumeration type 'enum qeth_ipa_setadp_cmd' to
      different enumeration type 'enum qeth_ipa_funcs' [-Wenum-conversion]
        if (qeth_adp_supported(card, IPA_SETADP_SET_PROMISC_MODE))
            ~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~
drivers/s390/net/qeth_core.h:179:41: note: expanded from macro 'qeth_adp_supported'
        qeth_is_ipa_supported(&c->options.adp, f)
        ~~~~~~~~~~~~~~~~~~~~~                  ^

Add a version of this macro that uses the correct types, and
remove the unused qeth_adp_enabled() macro that has the same
problem.

Reviewed-by: Nathan Chancellor <natechancellor@gmail.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoMerge branch 'netdevsim-implement-proper-device-model'
David S. Miller [Fri, 26 Apr 2019 05:52:03 +0000 (01:52 -0400)]
Merge branch 'netdevsim-implement-proper-device-model'

Jiri Pirko says:

====================
netdevsim: implement proper device model

Currently the model of netdevsim is a bit odd in multiple ways.
1) devlink instance is not in any way related with actual netdevsim
   netdevices. Instead, it is created per-namespace.
2) multi-port netdevsim device is done using "link" attribute.
3) netdevsim bus is there only to have something to bind the netdev to,
   it really does not act as a bus.
4) netdevsim instances are created by "ip link add" which is great for
   soft devices with no hw backend. The rtnl core allocates netdev and
   calls into driver holding rtnl mutex. For hw-backed devices, this
   flow is wrong as it breaks order in which things are done.

This patchset adjust netdevsim to fix all above.

In order to support proper devlink and devlink port instances and to be
able to emulate real devices, there is need to implement bus probe and
instantiate everything from there. User can specify device id and port
count to be instantianted. For example:

$ echo "10 4" > /sys/bus/netdevsim/new_device

Then devlink shows this:

$ devlink dev
netdevsim/netdevsim10

$ devlink port
netdevsim/netdevsim10/0: type eth netdev eni0np1 flavour physical
netdevsim/netdevsim10/1: type eth netdev eni0np2 flavour physical
netdevsim/netdevsim10/2: type eth netdev eni0np3 flavour physical
netdevsim/netdevsim10/3: type eth netdev eni0np4 flavour physical

There is possible to add and delete ports using their indexes
during netdevsim device lifetime like this:

Then devlink shows this:

$ devlink port
netdevsim/netdevsim10/1: type eth netdev eni10np2 flavour physical
netdevsim/netdevsim10/2: type eth netdev eni10np3 flavour physical
netdevsim/netdevsim10/3: type eth netdev eni10np4 flavour physical
netdevsim/netdevsim10/43: type eth netdev eni10np44 flavour physical

Debugfs topology is also adjusted a bit. The rest stays the same as
before.

Udev bits are merged un systemd upstream git:
https://github.com/systemd/systemd/commit/eaa9d507d85509c8bf727356e3884ec54b0fc646

See individual patches for changelog.
====================

Acked-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonetdevsim: implement ndo_get_devlink_port
Jiri Pirko [Thu, 25 Apr 2019 13:59:56 +0000 (15:59 +0200)]
netdevsim: implement ndo_get_devlink_port

Implement ndo_get_devlink_port and allow switch_id and port_name to be
handled by devlink.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonetdevsim: move netdev creation/destruction to dev probe
Jiri Pirko [Thu, 25 Apr 2019 13:59:55 +0000 (15:59 +0200)]
netdevsim: move netdev creation/destruction to dev probe

Remove the existing way to create netdevsim over rtnetlink and move the
netdev creation/destruction to dev probe, so for every probed port,
a netdevsim-netdev instance is created.

Adjust selftests to work with new interface.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonetdevsim: extend device attrs to support port addition and deletion
Jiri Pirko [Thu, 25 Apr 2019 13:59:54 +0000 (15:59 +0200)]
netdevsim: extend device attrs to support port addition and deletion

In order to test flows in core, it is beneficial to maintain previously
supported possibility to add and delete ports during netdevsim lifetime.
Do it by extending device sysfs attrs by "new_port" and "del_port".

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonetdevsim: implement dev probe/remove skeleton with port initialization
Jiri Pirko [Thu, 25 Apr 2019 13:59:53 +0000 (15:59 +0200)]
netdevsim: implement dev probe/remove skeleton with port initialization

Implement netdevsim bus probing of netdevsim devices. For every probed
device create a devlink instance. According to the user-passed value,
create a number of ports represented by devlink port instances.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonetdevsim: change debugfs tree topology
Jiri Pirko [Thu, 25 Apr 2019 13:59:52 +0000 (15:59 +0200)]
netdevsim: change debugfs tree topology

With the model where dev is represented by devlink and ports are
represented by devlink ports, make debugfs file names independent
on netdev names. Change the topology to the one illustrated
by the following example:

$ ls /sys/kernel/debug/netdevsim/
netdevsim1
$ ls /sys/kernel/debug/netdevsim/netdevsim1/
bpf_bind_accept  bpf_bind_verifier_delay  bpf_bound_progs  ports
$ ls /sys/kernel/debug/netdevsim/netdevsim1/ports/
0  1
$ ls /sys/kernel/debug/netdevsim/netdevsim1/ports/0/
bpf_map_accept  bpf_offloaded_id  bpf_tc_accept  bpf_tc_non_bound_accept  bpf_xdpdrv_accept  bpf_xdpoffload_accept  dev  ipsec
$ ls /sys/kernel/debug/netdevsim/netdevsim1/ports/0/dev -l
lrwxrwxrwx 1 root root 0 Apr 13 15:58 /sys/kernel/debug/netdevsim/netdevsim1/ports/0/dev -> ../../../netdevsim1

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonetdevsim: generate random switch id instead of using dev id
Jiri Pirko [Thu, 25 Apr 2019 13:59:51 +0000 (15:59 +0200)]
netdevsim: generate random switch id instead of using dev id

Current implementation of parent_id/switch_id does not follow the
original idea of being unique. The values are "0", "1", etc. Instead of
that, generate 32 random bytes.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonetdevsim: merge sdev into dev
Jiri Pirko [Thu, 25 Apr 2019 13:59:50 +0000 (15:59 +0200)]
netdevsim: merge sdev into dev

As previously introduce dev which is mapped 1:1 to a bus device covers
the purpose of the original shared device, merge the sdev code into dev.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonetdevsim: rename dev_init/exit() functions and make them independent on ns
Jiri Pirko [Thu, 25 Apr 2019 13:59:49 +0000 (15:59 +0200)]
netdevsim: rename dev_init/exit() functions and make them independent on ns

These functions are going to be called from bus probe/release(),
therefore make them independent on ns struct and rename accordingly.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonetdevsim: add bus attributes to add new and delete devices
Jiri Pirko [Thu, 25 Apr 2019 13:59:48 +0000 (15:59 +0200)]
netdevsim: add bus attributes to add new and delete devices

Add a way to add new netdevsim device on netdevsim bus and also to
delete existing netdevsim device from the bus. Track the bus devices
in using a list.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonetdevsim: use ida for bus device ids
Jiri Pirko [Thu, 25 Apr 2019 13:59:47 +0000 (15:59 +0200)]
netdevsim: use ida for bus device ids

Instead of increments of u32 value, use ida to manage bus device ids.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonetdevsim: add stub netdevsim driver implementation
Jiri Pirko [Thu, 25 Apr 2019 13:59:46 +0000 (15:59 +0200)]
netdevsim: add stub netdevsim driver implementation

In order to bus probing to work correctly, register a simple  netdevsim
driver implementation.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonetdevsim: move device registration and related code to bus.c
Jiri Pirko [Thu, 25 Apr 2019 13:59:45 +0000 (15:59 +0200)]
netdevsim: move device registration and related code to bus.c

Move netdevsim device registration into bus.c and alongside with that
the related sysfs attributes. Introduce new struct nsim_bus_dev to
represent a netdevsim device on netdevsim bus.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonetdevsim: put netdevsim bus code into separate file
Jiri Pirko [Thu, 25 Apr 2019 13:59:44 +0000 (15:59 +0200)]
netdevsim: put netdevsim bus code into separate file

As the code related to netdevsim bus is going to get bigger, move the
existing code to a separate file.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonetdevsim: rename devlink.c to dev.c to contain per-dev(asic) items
Jiri Pirko [Thu, 25 Apr 2019 13:59:43 +0000 (15:59 +0200)]
netdevsim: rename devlink.c to dev.c to contain per-dev(asic) items

The existing devlink.c code is going to be extended to represent asic
device on a bus. As this is about more than just devlink,
rename the file. Do appropriate prefix renaming alongside with that.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonetdevsim: create devlink instance per netdevsim instance
Jiri Pirko [Thu, 25 Apr 2019 13:59:42 +0000 (15:59 +0200)]
netdevsim: create devlink instance per netdevsim instance

Currently there is one devlink instance created per network namespace.
That is quite odd considering the fact that devlink instance should
represent an ASIC. The following patches are going to move the devlink
instance even more down to a bus device, but until then, have one
devlink instance per netdevsim instance. Struct nsim_devlink is
introduced to hold fib setting.

The changes in the fib code are only related to holding the
configuration per devlink instance instead of network namespace.

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonetdevsim: move device registration on bus to be done earlier in init
Jiri Pirko [Thu, 25 Apr 2019 13:59:41 +0000 (15:59 +0200)]
netdevsim: move device registration on bus to be done earlier in init

As a dependency of the subsequent patch, mode device registration to be
done earlier, directly in nsim_newlink().

Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoMerge branch 'for-upstream' of git://git.kernel.org/pub/scm/linux/kernel/git/bluetoot...
David S. Miller [Fri, 26 Apr 2019 05:39:38 +0000 (01:39 -0400)]
Merge branch 'for-upstream' of git://git./linux/kernel/git/bluetooth/bluetooth-next

Johan Hedberg says:

====================
pull request: bluetooth-next 2019-04-25

Here's the main bluetooth-next pull request for the 5.2 kernel.

 - Added support for Mediatek SDIO controllers
 - Added support for Broadcom BCM2076B1 UART controller
 - Added support for Marvel SD8987 chipset
 - Fix buffer overflow bug in hidp protocol
 - Various other smaller fixes & improvements

Please let me know if there are any issues pulling. Thanks.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
David S. Miller [Fri, 26 Apr 2019 03:52:29 +0000 (23:52 -0400)]
Merge git://git./linux/kernel/git/davem/net

Two easy cases of overlapping changes.

Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoBluetooth: Align minimum encryption key size for LE and BR/EDR connections
Marcel Holtmann [Wed, 24 Apr 2019 20:19:17 +0000 (22:19 +0200)]
Bluetooth: Align minimum encryption key size for LE and BR/EDR connections

The minimum encryption key size for LE connections is 56 bits and to
align LE with BR/EDR, enforce 56 bits of minimum encryption key size for
BR/EDR connections as well.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Cc: stable@vger.kernel.org
5 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
Linus Torvalds [Wed, 24 Apr 2019 23:18:59 +0000 (16:18 -0700)]
Merge git://git./linux/kernel/git/davem/net

Pull networking fixes from David Miller:
 "Just the usual assortment of small'ish fixes:

   1) Conntrack timeout is sometimes not initialized properly, from
      Alexander Potapenko.

   2) Add a reasonable range limit to tcp_min_rtt_wlen to avoid
      undefined behavior. From ZhangXiaoxu.

   3) des1 field of descriptor in stmmac driver is initialized with the
      wrong variable. From Yue Haibing.

   4) Increase mlxsw pci sw reset timeout a little bit more, from Ido
      Schimmel.

   5) Match IOT2000 stmmac devices more accurately, from Su Bao Cheng.

   6) Fallback refcount fix in TLS code, from Jakub Kicinski.

   7) Fix max MTU check when using XDP in mlx5, from Maxim Mikityanskiy.

   8) Fix recursive locking in team driver, from Hangbin Liu.

   9) Fix tls_set_device_offload_Rx() deadlock, from Jakub Kicinski.

  10) Don't use napi_alloc_frag() outside of softiq context of socionext
      driver, from Ilias Apalodimas.

  11) MAC address increment overflow in ncsi, from Tao Ren.

  12) Fix a regression in 8K/1M pool switching of RDS, from Zhu Yanjun.

  13) ipv4_link_failure has to validate the headers that are actually
      there because RAW sockets can pass in arbitrary garbage, from Eric
      Dumazet"

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (43 commits)
  ipv4: add sanity checks in ipv4_link_failure()
  net/rose: fix unbound loop in rose_loopback_timer()
  rxrpc: fix race condition in rxrpc_input_packet()
  net: rds: exchange of 8K and 1M pool
  net: vrf: Fix operation not supported when set vrf mac
  net/ncsi: handle overflow when incrementing mac address
  net: socionext: replace napi_alloc_frag with the netdev variant on init
  net: atheros: fix spelling mistake "underun" -> "underrun"
  spi: ST ST95HF NFC: declare missing of table
  spi: Micrel eth switch: declare missing of table
  net: stmmac: move stmmac_check_ether_addr() to driver probe
  netfilter: fix nf_l4proto_log_invalid to log invalid packets
  netfilter: never get/set skb->tstamp
  netfilter: ebtables: CONFIG_COMPAT: drop a bogus WARN_ON
  Documentation: decnet: remove reference to CONFIG_DECNET_ROUTE_FWMARK
  dt-bindings: add an explanation for internal phy-mode
  net/tls: don't leak IV and record seq when offload fails
  net/tls: avoid potential deadlock in tls_set_device_offload_rx()
  selftests/net: correct the return value for run_afpackettests
  team: fix possible recursive locking when add slaves
  ...

5 years agoMerge tag 'leds-for-5.1-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/j.anasz...
Linus Torvalds [Wed, 24 Apr 2019 23:15:38 +0000 (16:15 -0700)]
Merge tag 'leds-for-5.1-rc7' of git://git./linux/kernel/git/j.anaszewski/linux-leds

Pull LED update from Jacek Anaszewski:
 "A single change to MAINTAINERS:

  We announce a new LED reviewer - Dan Murphy"

* tag 'leds-for-5.1-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/j.anaszewski/linux-leds:
  MAINTAINERS: LEDs: Add designated reviewer for LED subsystem

5 years agoipv4: add sanity checks in ipv4_link_failure()
Eric Dumazet [Wed, 24 Apr 2019 15:04:05 +0000 (08:04 -0700)]
ipv4: add sanity checks in ipv4_link_failure()

Before calling __ip_options_compile(), we need to ensure the network
header is a an IPv4 one, and that it is already pulled in skb->head.

RAW sockets going through a tunnel can end up calling ipv4_link_failure()
with total garbage in the skb, or arbitrary lengthes.

syzbot report :

BUG: KASAN: stack-out-of-bounds in memcpy include/linux/string.h:355 [inline]
BUG: KASAN: stack-out-of-bounds in __ip_options_echo+0x294/0x1120 net/ipv4/ip_options.c:123
Write of size 69 at addr ffff888096abf068 by task syz-executor.4/9204

CPU: 0 PID: 9204 Comm: syz-executor.4 Not tainted 5.1.0-rc5+ #77
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
Call Trace:
 __dump_stack lib/dump_stack.c:77 [inline]
 dump_stack+0x172/0x1f0 lib/dump_stack.c:113
 print_address_description.cold+0x7c/0x20d mm/kasan/report.c:187
 kasan_report.cold+0x1b/0x40 mm/kasan/report.c:317
 check_memory_region_inline mm/kasan/generic.c:185 [inline]
 check_memory_region+0x123/0x190 mm/kasan/generic.c:191
 memcpy+0x38/0x50 mm/kasan/common.c:133
 memcpy include/linux/string.h:355 [inline]
 __ip_options_echo+0x294/0x1120 net/ipv4/ip_options.c:123
 __icmp_send+0x725/0x1400 net/ipv4/icmp.c:695
 ipv4_link_failure+0x29f/0x550 net/ipv4/route.c:1204
 dst_link_failure include/net/dst.h:427 [inline]
 vti6_xmit net/ipv6/ip6_vti.c:514 [inline]
 vti6_tnl_xmit+0x10d4/0x1c0c net/ipv6/ip6_vti.c:553
 __netdev_start_xmit include/linux/netdevice.h:4414 [inline]
 netdev_start_xmit include/linux/netdevice.h:4423 [inline]
 xmit_one net/core/dev.c:3292 [inline]
 dev_hard_start_xmit+0x1b2/0x980 net/core/dev.c:3308
 __dev_queue_xmit+0x271d/0x3060 net/core/dev.c:3878
 dev_queue_xmit+0x18/0x20 net/core/dev.c:3911
 neigh_direct_output+0x16/0x20 net/core/neighbour.c:1527
 neigh_output include/net/neighbour.h:508 [inline]
 ip_finish_output2+0x949/0x1740 net/ipv4/ip_output.c:229
 ip_finish_output+0x73c/0xd50 net/ipv4/ip_output.c:317
 NF_HOOK_COND include/linux/netfilter.h:278 [inline]
 ip_output+0x21f/0x670 net/ipv4/ip_output.c:405
 dst_output include/net/dst.h:444 [inline]
 NF_HOOK include/linux/netfilter.h:289 [inline]
 raw_send_hdrinc net/ipv4/raw.c:432 [inline]
 raw_sendmsg+0x1d2b/0x2f20 net/ipv4/raw.c:663
 inet_sendmsg+0x147/0x5d0 net/ipv4/af_inet.c:798
 sock_sendmsg_nosec net/socket.c:651 [inline]
 sock_sendmsg+0xdd/0x130 net/socket.c:661
 sock_write_iter+0x27c/0x3e0 net/socket.c:988
 call_write_iter include/linux/fs.h:1866 [inline]
 new_sync_write+0x4c7/0x760 fs/read_write.c:474
 __vfs_write+0xe4/0x110 fs/read_write.c:487
 vfs_write+0x20c/0x580 fs/read_write.c:549
 ksys_write+0x14f/0x2d0 fs/read_write.c:599
 __do_sys_write fs/read_write.c:611 [inline]
 __se_sys_write fs/read_write.c:608 [inline]
 __x64_sys_write+0x73/0xb0 fs/read_write.c:608
 do_syscall_64+0x103/0x610 arch/x86/entry/common.c:290
 entry_SYSCALL_64_after_hwframe+0x49/0xbe
RIP: 0033:0x458c29
Code: ad b8 fb ff c3 66 2e 0f 1f 84 00 00 00 00 00 66 90 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 0f 83 7b b8 fb ff c3 66 2e 0f 1f 84 00 00 00 00
RSP: 002b:00007f293b44bc78 EFLAGS: 00000246 ORIG_RAX: 0000000000000001
RAX: ffffffffffffffda RBX: 0000000000000003 RCX: 0000000000458c29
RDX: 0000000000000014 RSI: 00000000200002c0 RDI: 0000000000000003
RBP: 000000000073bf00 R08: 0000000000000000 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000246 R12: 00007f293b44c6d4
R13: 00000000004c8623 R14: 00000000004ded68 R15: 00000000ffffffff

The buggy address belongs to the page:
page:ffffea00025aafc0 count:0 mapcount:0 mapping:0000000000000000 index:0x0
flags: 0x1fffc0000000000()
raw: 01fffc0000000000 0000000000000000 ffffffff025a0101 0000000000000000
raw: 0000000000000000 0000000000000000 00000000ffffffff 0000000000000000
page dumped because: kasan: bad access detected

Memory state around the buggy address:
 ffff888096abef80: 00 00 00 f2 f2 f2 f2 f2 00 00 00 00 00 00 00 f2
 ffff888096abf000: f2 f2 f2 f2 00 00 00 00 00 00 00 00 00 00 00 00
>ffff888096abf080: 00 00 f3 f3 f3 f3 00 00 00 00 00 00 00 00 00 00
                         ^
 ffff888096abf100: 00 00 00 00 f1 f1 f1 f1 00 00 f3 f3 00 00 00 00
 ffff888096abf180: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00

Fixes: ed0de45a1008 ("ipv4: recompile ip options in ipv4_link_failure")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Stephen Suryaputra <ssuryaextr@gmail.com>
Acked-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet/rose: fix unbound loop in rose_loopback_timer()
Eric Dumazet [Wed, 24 Apr 2019 12:35:00 +0000 (05:35 -0700)]
net/rose: fix unbound loop in rose_loopback_timer()

This patch adds a limit on the number of skbs that fuzzers can queue
into loopback_queue. 1000 packets for rose loopback seems more than enough.

Then, since we now have multiple cpus in most linux hosts,
we also need to limit the number of skbs rose_loopback_timer()
can dequeue at each round.

rose_loopback_queue() can be drop-monitor friendly, calling
consume_skb() or kfree_skb() appropriately.

Finally, use mod_timer() instead of del_timer() + add_timer()

syzbot report was :

rcu: INFO: rcu_preempt self-detected stall on CPU
rcu:    0-...!: (10499 ticks this GP) idle=536/1/0x4000000000000002 softirq=103291/103291 fqs=34
rcu:     (t=10500 jiffies g=140321 q=323)
rcu: rcu_preempt kthread starved for 10426 jiffies! g140321 f0x0 RCU_GP_WAIT_FQS(5) ->state=0x402 ->cpu=1
rcu: RCU grace-period kthread stack dump:
rcu_preempt     I29168    10      2 0x80000000
Call Trace:
 context_switch kernel/sched/core.c:2877 [inline]
 __schedule+0x813/0x1cc0 kernel/sched/core.c:3518
 schedule+0x92/0x180 kernel/sched/core.c:3562
 schedule_timeout+0x4db/0xfd0 kernel/time/timer.c:1803
 rcu_gp_fqs_loop kernel/rcu/tree.c:1971 [inline]
 rcu_gp_kthread+0x962/0x17b0 kernel/rcu/tree.c:2128
 kthread+0x357/0x430 kernel/kthread.c:253
 ret_from_fork+0x3a/0x50 arch/x86/entry/entry_64.S:352
NMI backtrace for cpu 0
CPU: 0 PID: 7632 Comm: kworker/0:4 Not tainted 5.1.0-rc5+ #172
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
Workqueue: events iterate_cleanup_work
Call Trace:
 <IRQ>
 __dump_stack lib/dump_stack.c:77 [inline]
 dump_stack+0x172/0x1f0 lib/dump_stack.c:113
 nmi_cpu_backtrace.cold+0x63/0xa4 lib/nmi_backtrace.c:101
 nmi_trigger_cpumask_backtrace+0x1be/0x236 lib/nmi_backtrace.c:62
 arch_trigger_cpumask_backtrace+0x14/0x20 arch/x86/kernel/apic/hw_nmi.c:38
 trigger_single_cpu_backtrace include/linux/nmi.h:164 [inline]
 rcu_dump_cpu_stacks+0x183/0x1cf kernel/rcu/tree.c:1223
 print_cpu_stall kernel/rcu/tree.c:1360 [inline]
 check_cpu_stall kernel/rcu/tree.c:1434 [inline]
 rcu_pending kernel/rcu/tree.c:3103 [inline]
 rcu_sched_clock_irq.cold+0x500/0xa4a kernel/rcu/tree.c:2544
 update_process_times+0x32/0x80 kernel/time/timer.c:1635
 tick_sched_handle+0xa2/0x190 kernel/time/tick-sched.c:161
 tick_sched_timer+0x47/0x130 kernel/time/tick-sched.c:1271
 __run_hrtimer kernel/time/hrtimer.c:1389 [inline]
 __hrtimer_run_queues+0x33e/0xde0 kernel/time/hrtimer.c:1451
 hrtimer_interrupt+0x314/0x770 kernel/time/hrtimer.c:1509
 local_apic_timer_interrupt arch/x86/kernel/apic/apic.c:1035 [inline]
 smp_apic_timer_interrupt+0x120/0x570 arch/x86/kernel/apic/apic.c:1060
 apic_timer_interrupt+0xf/0x20 arch/x86/entry/entry_64.S:807
RIP: 0010:__sanitizer_cov_trace_pc+0x0/0x50 kernel/kcov.c:95
Code: 89 25 b4 6e ec 08 41 bc f4 ff ff ff e8 cd 5d ea ff 48 c7 05 9e 6e ec 08 00 00 00 00 e9 a4 e9 ff ff 90 90 90 90 90 90 90 90 90 <55> 48 89 e5 48 8b 75 08 65 48 8b 04 25 00 ee 01 00 65 8b 15 c8 60
RSP: 0018:ffff8880ae807ce0 EFLAGS: 00000286 ORIG_RAX: ffffffffffffff13
RAX: ffff88806fd40640 RBX: dffffc0000000000 RCX: ffffffff863fbc56
RDX: 0000000000000100 RSI: ffffffff863fbc1d RDI: ffff88808cf94228
RBP: ffff8880ae807d10 R08: ffff88806fd40640 R09: ffffed1015d00f8b
R10: ffffed1015d00f8a R11: 0000000000000003 R12: ffff88808cf941c0
R13: 00000000fffff034 R14: ffff8882166cd840 R15: 0000000000000000
 rose_loopback_timer+0x30d/0x3f0 net/rose/rose_loopback.c:91
 call_timer_fn+0x190/0x720 kernel/time/timer.c:1325
 expire_timers kernel/time/timer.c:1362 [inline]
 __run_timers kernel/time/timer.c:1681 [inline]
 __run_timers kernel/time/timer.c:1649 [inline]
 run_timer_softirq+0x652/0x1700 kernel/time/timer.c:1694
 __do_softirq+0x266/0x95a kernel/softirq.c:293
 do_softirq_own_stack+0x2a/0x40 arch/x86/entry/entry_64.S:1027

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: syzbot <syzkaller@googlegroups.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agorxrpc: fix race condition in rxrpc_input_packet()
Eric Dumazet [Wed, 24 Apr 2019 16:44:11 +0000 (09:44 -0700)]
rxrpc: fix race condition in rxrpc_input_packet()

After commit 5271953cad31 ("rxrpc: Use the UDP encap_rcv hook"),
rxrpc_input_packet() is directly called from lockless UDP receive
path, under rcu_read_lock() protection.

It must therefore use RCU rules :

- udp_sk->sk_user_data can be cleared at any point in this function.
  rcu_dereference_sk_user_data() is what we need here.

- Also, since sk_user_data might have been set in rxrpc_open_socket()
  we must observe a proper RCU grace period before kfree(local) in
  rxrpc_lookup_local()

v4: @local can be NULL in xrpc_lookup_local() as reported by kbuild test robot <lkp@intel.com>
        and Julia Lawall <julia.lawall@lip6.fr>, thanks !

v3,v2 : addressed David Howells feedback, thanks !

syzbot reported :

kasan: CONFIG_KASAN_INLINE enabled
kasan: GPF could be caused by NULL-ptr deref or user memory access
general protection fault: 0000 [#1] PREEMPT SMP KASAN
CPU: 0 PID: 19236 Comm: syz-executor703 Not tainted 5.1.0-rc6 #79
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
RIP: 0010:__lock_acquire+0xbef/0x3fb0 kernel/locking/lockdep.c:3573
Code: 00 0f 85 a5 1f 00 00 48 81 c4 10 01 00 00 5b 41 5c 41 5d 41 5e 41 5f 5d c3 48 b8 00 00 00 00 00 fc ff df 4c 89 ea 48 c1 ea 03 <80> 3c 02 00 0f 85 4a 21 00 00 49 81 7d 00 20 54 9c 89 0f 84 cf f4
RSP: 0018:ffff88809d7aef58 EFLAGS: 00010002
RAX: dffffc0000000000 RBX: 0000000000000000 RCX: 0000000000000000
RDX: 0000000000000026 RSI: 0000000000000000 RDI: 0000000000000001
RBP: ffff88809d7af090 R08: 0000000000000001 R09: 0000000000000001
R10: ffffed1015d05bc7 R11: ffff888089428600 R12: 0000000000000000
R13: 0000000000000130 R14: 0000000000000001 R15: 0000000000000001
FS:  00007f059044d700(0000) GS:ffff8880ae800000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00000000004b6040 CR3: 00000000955ca000 CR4: 00000000001406f0
Call Trace:
 lock_acquire+0x16f/0x3f0 kernel/locking/lockdep.c:4211
 __raw_spin_lock_irqsave include/linux/spinlock_api_smp.h:110 [inline]
 _raw_spin_lock_irqsave+0x95/0xcd kernel/locking/spinlock.c:152
 skb_queue_tail+0x26/0x150 net/core/skbuff.c:2972
 rxrpc_reject_packet net/rxrpc/input.c:1126 [inline]
 rxrpc_input_packet+0x4a0/0x5536 net/rxrpc/input.c:1414
 udp_queue_rcv_one_skb+0xaf2/0x1780 net/ipv4/udp.c:2011
 udp_queue_rcv_skb+0x128/0x730 net/ipv4/udp.c:2085
 udp_unicast_rcv_skb.isra.0+0xb9/0x360 net/ipv4/udp.c:2245
 __udp4_lib_rcv+0x701/0x2ca0 net/ipv4/udp.c:2301
 udp_rcv+0x22/0x30 net/ipv4/udp.c:2482
 ip_protocol_deliver_rcu+0x60/0x8f0 net/ipv4/ip_input.c:208
 ip_local_deliver_finish+0x23b/0x390 net/ipv4/ip_input.c:234
 NF_HOOK include/linux/netfilter.h:289 [inline]
 NF_HOOK include/linux/netfilter.h:283 [inline]
 ip_local_deliver+0x1e9/0x520 net/ipv4/ip_input.c:255
 dst_input include/net/dst.h:450 [inline]
 ip_rcv_finish+0x1e1/0x300 net/ipv4/ip_input.c:413
 NF_HOOK include/linux/netfilter.h:289 [inline]
 NF_HOOK include/linux/netfilter.h:283 [inline]
 ip_rcv+0xe8/0x3f0 net/ipv4/ip_input.c:523
 __netif_receive_skb_one_core+0x115/0x1a0 net/core/dev.c:4987
 __netif_receive_skb+0x2c/0x1c0 net/core/dev.c:5099
 netif_receive_skb_internal+0x117/0x660 net/core/dev.c:5202
 napi_frags_finish net/core/dev.c:5769 [inline]
 napi_gro_frags+0xade/0xd10 net/core/dev.c:5843
 tun_get_user+0x2f24/0x3fb0 drivers/net/tun.c:1981
 tun_chr_write_iter+0xbd/0x156 drivers/net/tun.c:2027
 call_write_iter include/linux/fs.h:1866 [inline]
 do_iter_readv_writev+0x5e1/0x8e0 fs/read_write.c:681
 do_iter_write fs/read_write.c:957 [inline]
 do_iter_write+0x184/0x610 fs/read_write.c:938
 vfs_writev+0x1b3/0x2f0 fs/read_write.c:1002
 do_writev+0x15e/0x370 fs/read_write.c:1037
 __do_sys_writev fs/read_write.c:1110 [inline]
 __se_sys_writev fs/read_write.c:1107 [inline]
 __x64_sys_writev+0x75/0xb0 fs/read_write.c:1107
 do_syscall_64+0x103/0x610 arch/x86/entry/common.c:290
 entry_SYSCALL_64_after_hwframe+0x49/0xbe

Fixes: 5271953cad31 ("rxrpc: Use the UDP encap_rcv hook")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: syzbot <syzkaller@googlegroups.com>
Acked-by: David Howells <dhowells@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoMAINTAINERS: LEDs: Add designated reviewer for LED subsystem
Dan Murphy [Tue, 23 Apr 2019 20:00:24 +0000 (15:00 -0500)]
MAINTAINERS: LEDs: Add designated reviewer for LED subsystem

Add a designated reviewer for the LED subsystem as there
are already two maintainers assigned.

Signed-off-by: Dan Murphy <dmurphy@ti.com>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Jacek Anaszewski <jacek.anaszewski@gmail.com>
5 years agonet: rds: exchange of 8K and 1M pool
Zhu Yanjun [Wed, 24 Apr 2019 06:56:42 +0000 (02:56 -0400)]
net: rds: exchange of 8K and 1M pool

Before the commit 490ea5967b0d ("RDS: IB: move FMR code to its own file"),
when the dirty_count is greater than 9/10 of max_items of 8K pool,
1M pool is used, Vice versa. After the commit 490ea5967b0d ("RDS: IB: move
FMR code to its own file"), the above is removed. When we make the
following tests.

Server:
  rds-stress -r 1.1.1.16 -D 1M

Client:
  rds-stress -r 1.1.1.14 -s 1.1.1.16 -D 1M

The following will appear.
"
connecting to 1.1.1.16:4000
negotiated options, tasks will start in 2 seconds
Starting up..header from 1.1.1.166:4001 to id 4001 bogus
..
tsks  tx/s  rx/s tx+rx K/s  mbi K/s  mbo K/s tx us/c  rtt us
cpu %
   1    0    0     0.00     0.00     0.00    0.00 0.00 -1.00
   1    0    0     0.00     0.00     0.00    0.00 0.00 -1.00
   1    0    0     0.00     0.00     0.00    0.00 0.00 -1.00
   1    0    0     0.00     0.00     0.00    0.00 0.00 -1.00
   1    0    0     0.00     0.00     0.00    0.00 0.00 -1.00
...
"
So this exchange between 8K and 1M pool is added back.

Fixes: commit 490ea5967b0d ("RDS: IB: move FMR code to its own file")
Signed-off-by: Zhu Yanjun <yanjun.zhu@oracle.com>
Acked-by: Santosh Shilimkar <santosh.shilimkar@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: sched: flower: refactor reoffload for concurrent access
Vlad Buslov [Wed, 24 Apr 2019 06:53:31 +0000 (09:53 +0300)]
net: sched: flower: refactor reoffload for concurrent access

Recent changes that introduced unlocked flower did not properly account for
case when reoffload is initiated concurrently with filter updates. To fix
the issue, extend flower with 'hw_filters' list that is used to store
filters that don't have 'skip_hw' flag set. Filter is added to the list
when it is inserted to hardware and only removed from it after being
unoffloaded from all drivers that parent block is attached to. This ensures
that concurrent reoffload can still access filter that is being deleted and
prevents race condition when driver callback can be removed when filter is
no longer accessible trough idr, but is still present in hardware.

Refactor fl_change() to respect new filter reference counter and to release
filter reference with __fl_put() in case of error, instead of directly
deallocating filter memory. This allows for concurrent access to filter
from fl_reoffload() and protects it with reference counting. Refactor
fl_reoffload() to iterate over hw_filters list instead of idr. Implement
fl_get_next_hw_filter() helper function that is used to iterate over
hw_filters list with reference counting and skips filters that are being
concurrently deleted.

Fixes: 92149190067d ("net: sched: flower: set unlocked flag for flower proto ops")
Signed-off-by: Vlad Buslov <vladbu@mellanox.com>
Reviewed-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: mvneta: Switch to using devm_alloc_etherdev_mqs
Rosen Penev [Tue, 23 Apr 2019 20:46:03 +0000 (13:46 -0700)]
net: mvneta: Switch to using devm_alloc_etherdev_mqs

It allows some of the code to be simplified.

Tested on Turris Omnia.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agotipc: tipc_udp_recv() cleanup vs rcu verbs
Eric Dumazet [Tue, 23 Apr 2019 16:24:46 +0000 (09:24 -0700)]
tipc: tipc_udp_recv() cleanup vs rcu verbs

First thing tipc_udp_recv() does is to use rcu_dereference_sk_user_data(),
and this is really hinting we already own rcu_read_lock() from the caller
(UDP stack).

No need to add another rcu_read_lock()/rcu_read_unlock() pair.

Also use rcu_dereference() instead of rcu_dereference_rtnl()
in the data path.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Jon Maloy <jon.maloy@ericsson.com>
Cc: Ying Xue <ying.xue@windriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: vrf: Fix operation not supported when set vrf mac
Miaohe Lin [Sat, 20 Apr 2019 04:09:39 +0000 (12:09 +0800)]
net: vrf: Fix operation not supported when set vrf mac

Vrf device is not able to change mac address now because lack of
ndo_set_mac_address. Complete this in case some apps need to do
this.

Reported-by: Hui Wang <wanghui104@huawei.com>
Signed-off-by: Miaohe Lin <linmiaohe@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoBluetooth: hci_core: Don't stop BT if the BD address missing in dts
Balakrishna Godavarthi [Thu, 18 Apr 2019 13:21:23 +0000 (18:51 +0530)]
Bluetooth: hci_core: Don't stop BT if the BD address missing in dts

When flag HCI_QUIRK_USE_BDADDR_PROPERTY is set, we will read the
bluetooth address from dts. If the bluetooth address node is missing
from the dts we will enable it controller UNCONFIGURED state.
This patch enables the normal flow even if the BD address is missing
from the dts tree.

Signed-off-by: Balakrishna Godavarthi <bgodavar@codeaurora.org>
Tested-by: Harish Bandi <c-hbandi@codeaurora.org>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
5 years agoipv6: Use result arg in fib_lookup_arg consistently
David Ahern [Wed, 24 Apr 2019 01:05:33 +0000 (18:05 -0700)]
ipv6: Use result arg in fib_lookup_arg consistently

arg.result is sometimes used as fib6_result and sometimes used to
hold the rt6_info. Add rt6_info to fib6_result and make the use
of arg.result consistent through ipv6 rules.

The rt6 entry is filled in for lookups returning a dst_entry, but not
for direct fib_lookups that just want a fib6_info.

Fixes: effda4dd97e8 ("ipv6: Pass fib6_result to fib lookups")
Signed-off-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoipv6: fib6_rule_action_alt needs to return -EAGAIN
David Ahern [Wed, 24 Apr 2019 01:06:30 +0000 (18:06 -0700)]
ipv6: fib6_rule_action_alt needs to return -EAGAIN

fib rule actions should return -EAGAIN for the rules to continue to the
next one. A recent change overwrote err with the lookup always returning
0 (future change will make it more like IPv4) which means the rules
stopped at the first (e.g., local table lookup only). Catch and reset err
to -EAGAIN.

Fixes: effda4dd97e87 ("ipv6: Pass fib6_result to fib lookups")
Signed-off-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet/ncsi: handle overflow when incrementing mac address
Tao Ren [Wed, 24 Apr 2019 01:43:32 +0000 (01:43 +0000)]
net/ncsi: handle overflow when incrementing mac address

Previously BMC's MAC address is calculated by simply adding 1 to the
last byte of network controller's MAC address, and it produces incorrect
result when network controller's MAC address ends with 0xFF.

The problem can be fixed by calling eth_addr_inc() function to increment
MAC address; besides, the MAC address is also validated before assigning
to BMC.

Fixes: cb10c7c0dfd9 ("net/ncsi: Add NCSI Broadcom OEM command")
Signed-off-by: Tao Ren <taoren@fb.com>
Acked-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Acked-by: Samuel Mendoza-Jonas <sam@mendozajonas.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoMerge tag 'drm-fixes-2019-04-24' of git://anongit.freedesktop.org/drm/drm
Linus Torvalds [Wed, 24 Apr 2019 04:08:52 +0000 (21:08 -0700)]
Merge tag 'drm-fixes-2019-04-24' of git://anongit.freedesktop.org/drm/drm

Pull drm regression fixes from Dave Airlie:
 "We interrupt your regularly scheduled drm fixes for a regression
  special.

  The first is for a fix in i915 that had unexpected side effects
  fallout in the userspace X.org modesetting driver where X would no
  longer start. I got tired of the nitpicking and issued a large hammer
  on it. The X.org driver is buggy, but blackscreen regressions are
  worse.

  The second was an oversight that myself and Gerd should have noticed
  better, Gerd is trying to fix this properly, but the regression is too
  large to leave, even if the original behaviour is bad in some cases,
  it's clearly bad to break a bunch of working use cases.

  I'll likely have a regular fixes pull later, but I really wanted to
  highlight these"

* tag 'drm-fixes-2019-04-24' of git://anongit.freedesktop.org/drm/drm:
  Revert "drm/virtio: drop prime import/export callbacks"
  Revert "drm/i915/fbdev: Actually configure untiled displays"

5 years agomlxsw: spectrum_router: Prevent ipv6 gateway with v4 route via replace and append
David Ahern [Tue, 23 Apr 2019 23:03:29 +0000 (16:03 -0700)]
mlxsw: spectrum_router: Prevent ipv6 gateway with v4 route via replace and append

mlxsw currently does not support v6 gateways with v4 routes. Commit
19a9d136f198 ("ipv4: Flag fib_info with a fib_nh using IPv6 gateway")
prevents a route from being added, but nothing stops the replace or
append. Add a catch for them too.
    $ ip  ro add 172.16.2.0/24 via 10.99.1.2
    $ ip  ro replace 172.16.2.0/24 via inet6 fe80::202:ff:fe00:b dev swp1s0
    Error: mlxsw_spectrum: IPv6 gateway with IPv4 route is not supported.
    $ ip  ro append 172.16.2.0/24 via inet6 fe80::202:ff:fe00:b dev swp1s0
    Error: mlxsw_spectrum: IPv6 gateway with IPv4 route is not supported.

Signed-off-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoMerge branch 'Taprio-qdisc-fixes'
David S. Miller [Wed, 24 Apr 2019 02:52:33 +0000 (19:52 -0700)]
Merge branch 'Taprio-qdisc-fixes'

Andre Guedes says:

====================
Taprio qdisc fixes

I'm re-sending this series, now with the "net-next" prefix in the subject.

The only change from the previous version is in patch 3. As suggested by Cong
Wang, it removes the !entry check within should_restart_cycle() since it is
already checked by the caller. As a side effect, that function becomes a dummy
wrapper on list_is_last() so we simply remove it and call list_is_last()
instead.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: sched: taprio: Fix taprio_dequeue()
Andre Guedes [Tue, 23 Apr 2019 19:44:24 +0000 (12:44 -0700)]
net: sched: taprio: Fix taprio_dequeue()

In case we don't have 'guard' or 'budget' to transmit the skb, we should
continue traversing the qdisc list since the remaining guard/budget
might be enough to transmit a skb from other children qdiscs.

Fixes: 5a781ccbd19e (“tc: Add support for configuring the taprio scheduler”)
Signed-off-by: Andre Guedes <andre.guedes@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: sched: taprio: Fix taprio_peek()
Andre Guedes [Tue, 23 Apr 2019 19:44:23 +0000 (12:44 -0700)]
net: sched: taprio: Fix taprio_peek()

While traversing taprio's children qdisc list, if the gate is closed for
a given traffic class, we should continue traversing the list since the
remaining qdiscs may have skb ready for transmission.

This patch also takes this opportunity and changes the function to use
the TAPRIO_ALL_GATES_OPEN macro instead of the magic number '-1'.

Fixes: 5a781ccbd19e (“tc: Add support for configuring the taprio scheduler”)
Signed-off-by: Andre Guedes <andre.guedes@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: sched: taprio: Remove should_restart_cycle()
Andre Guedes [Tue, 23 Apr 2019 19:44:22 +0000 (12:44 -0700)]
net: sched: taprio: Remove should_restart_cycle()

The 'entry' argument from should_restart_cycle() cannot be NULL since it
is already checked by the caller so the WARN_ON() within should_
restart_cycle() could be removed.  By doing that, that function becomes
a dummy wrapper on list_is_last() so this patch simply gets rid of it
and call list_is_last() within advance_sched() instead.

Signed-off-by: Andre Guedes <andre.guedes@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: sched: taprio: Refactor taprio_get_start_time()
Andre Guedes [Tue, 23 Apr 2019 19:44:21 +0000 (12:44 -0700)]
net: sched: taprio: Refactor taprio_get_start_time()

This patch does a code refactoring to taprio_get_start_time() function
to improve readability and report error properly.

If 'base' time is later than 'now', the start time is equal to 'base'
and taprio_get_start_time() is done. That's the natural case so we move
that code to the beginning of the function. Also, if 'cycle' calculation
is zero, something went really wrong with taprio and we should log that
internal error properly.

Signed-off-by: Andre Guedes <andre.guedes@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: sched: taprio: Remove pointless variable assigment
Andre Guedes [Tue, 23 Apr 2019 19:44:20 +0000 (12:44 -0700)]
net: sched: taprio: Remove pointless variable assigment

This patch removes a pointless variable assigment in taprio_change().
The 'err' variable is not used from this assignment to the next one so
this patch removes it.

Signed-off-by: Andre Guedes <andre.guedes@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: Change nhc_flags to unsigned char
David Ahern [Tue, 23 Apr 2019 15:48:09 +0000 (08:48 -0700)]
net: Change nhc_flags to unsigned char

nhc_flags holds the RTNH_F flags for a given nexthop (fib{6}_nh).
All of the RTNH_F_ flags fit in an unsigned char, and since the API to
userspace (rtnh_flags and lower byte of rtm_flags) is 1 byte it can not
grow. Make nhc_flags in fib_nh_common an unsigned char and shrink the
size of the struct by 8, from 56 to 48 bytes.

Update the flags arguments for up netdevice events and fib_nexthop_info
which determines the RTNH_F flags to return on a dump/event. The RTNH_F
flags are passed in the lower byte of rtm_flags which is an unsigned int
so use a temp variable for the flags to fib_nexthop_info and combine
with rtm_flags in the caller.

Signed-off-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agolwtunnel: Pass encap and encap type attributes to lwtunnel_fill_encap
David Ahern [Tue, 23 Apr 2019 15:23:41 +0000 (08:23 -0700)]
lwtunnel: Pass encap and encap type attributes to lwtunnel_fill_encap

Currently, lwtunnel_fill_encap hardcodes the encap and encap type
attributes as RTA_ENCAP and RTA_ENCAP_TYPE, respectively. The nexthop
objects want to re-use this code but the encap attributes passed to
userspace as NHA_ENCAP and NHA_ENCAP_TYPE. Since that is the only
difference, change lwtunnel_fill_encap to take the attribute type as
an input.

Signed-off-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: socionext: replace napi_alloc_frag with the netdev variant on init
Ilias Apalodimas [Tue, 23 Apr 2019 06:01:41 +0000 (09:01 +0300)]
net: socionext: replace napi_alloc_frag with the netdev variant on init

The netdev variant is usable on any context since it disables interrupts.
The napi variant of the call should only be used within softirq context.
Replace napi_alloc_frag on driver init with the correct netdev_alloc_frag
call

Changes since v1:
- Adjusted commit message

Acked-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Acked-by: Jassi Brar <jaswinder.singh@linaro.org>
Fixes: 4acb20b46214 ("net: socionext: different approach on DMA")
Signed-off-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: atheros: fix spelling mistake "underun" -> "underrun"
Colin Ian King [Tue, 23 Apr 2019 14:30:07 +0000 (15:30 +0100)]
net: atheros: fix spelling mistake "underun" -> "underrun"

There are spelling mistakes in structure elements, fix these.

Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoravb: Avoid unsupported internal delay mode for R-Car E3/D3
Simon Horman [Tue, 23 Apr 2019 13:01:53 +0000 (15:01 +0200)]
ravb: Avoid unsupported internal delay mode for R-Car E3/D3

According to the R-Car Gen3 Hardware Manual Rev 1.50 of Nov 30, 2018, the
TX clock internal delay mode isn't supported on R-Car E3 (r8a77990) or D3
(r8a77995). And by extension it is also not supported by RZ/G2E (r9a774c0).

This matches all ES versions of the affected SoCs as it is
not clear if this problem will be resolved in newer chips.
This can be revisited, as necessary.

This patch does not error-out if PHY_INTERFACE_MODE_RGMII_ID or
PHY_INTERFACE_MODE_RGMII_TXID are used on SoCs where TX clock delay
mode is not supported as there is a risk of introducing a regression
when used in conjunction with older DT blobs present in the field.
Rather, a warning is logged in such cases.

Based on work by Kazuya Mizuguchi.

Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agonet: fix sparc64 compilation of sock_gettstamp
Stephen Rothwell [Tue, 23 Apr 2019 07:25:24 +0000 (17:25 +1000)]
net: fix sparc64 compilation of sock_gettstamp

net/core/sock.c: In function 'sock_gettstamp':
net/core/sock.c:3007:23: error: expected '}' before ';' token
    .tv_sec = ts.tv_sec;
                       ^
net/core/sock.c:3011:4: error: expected ')' before 'return'
    return -EFAULT;
    ^~~~~~
net/core/sock.c:3013:2: error: expected expression before '}' token
  }
  ^

Fixes: c7cbdbf29f48 ("net: rework SIOCGSTAMP ioctl handling")
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoisdn:mISDN: fix misuse of %x in hfcpci.c
Fuqian Huang [Tue, 23 Apr 2019 02:56:23 +0000 (10:56 +0800)]
isdn:mISDN: fix misuse of %x in hfcpci.c

Pointers should be printed with %p or %px rather than
cast to (u_long) and printed with %lx.
Change %lx to %p to print the pointer.
Change %lx to %pad to print the dma_addr_t.

Signed-off-by: Fuqian Huang <huangfq.daxian@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoisdn: hisax: Fix misuse of %x in config.c
Fuqian Huang [Tue, 23 Apr 2019 02:25:28 +0000 (10:25 +0800)]
isdn: hisax: Fix misuse of %x in config.c

Pointers should be printed with %p or %px rather than
cast to (u_long) type and printed with %lX.
As the function seems to be for debug purpose.
Change %lX to %px to print the pointer value.

Signed-off-by: Fuqian Huang <huangfq.daxian@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoRevert "drm/virtio: drop prime import/export callbacks"
Dave Airlie [Wed, 24 Apr 2019 00:52:20 +0000 (10:52 +1000)]
Revert "drm/virtio: drop prime import/export callbacks"

This patch does more harm than good, as it breaks both Xwayland and
gnome-shell with X11.

Xwayland requires DRI3 & DRI3 requires PRIME.

X11 crash for obscure double-free reason which are hard to debug
(starting X11 by hand doesn't trigger the crash).

I don't see an apparent problem implementing those stub prime
functions, they may return an error at run-time, and it seems to be
handled fine by GNOME at least.

This reverts commit b318e3ff7ca065d6b107e424c85a63d7a6798a69.
[airlied:
This broke userspace for virtio-gpus, and regressed things from DRI3 to DRI2.

This brings back the original problem, but it's better than regressions.]

Fixes: b318e3ff7ca065d6b107e424c85a63d7a6798a ("drm/virtio: drop prime import/export callbacks")
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
5 years agoRevert "drm/i915/fbdev: Actually configure untiled displays"
Dave Airlie [Wed, 24 Apr 2019 00:47:56 +0000 (10:47 +1000)]
Revert "drm/i915/fbdev: Actually configure untiled displays"

This reverts commit d179b88deb3bf6fed4991a31fd6f0f2cad21fab5.

This commit is documented to break userspace X.org modesetting driver in certain configurations.

The X.org modesetting userspace driver is broken. No fixes are available yet. In order for this patch to be applied it either needs a config option or a workaround developed.

This has been reported a few times, saying it's a userspace problem is clearly against the regression rules.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=109806
Signed-off-by: Dave Airlie <airlied@redhat.com>
Cc: <stable@vger.kernel.org> # v3.19+
5 years agoMerge branch 'ipv6-fib6_ref-conversion-to-refcount_t'
David S. Miller [Wed, 24 Apr 2019 00:19:48 +0000 (17:19 -0700)]
Merge branch 'ipv6-fib6_ref-conversion-to-refcount_t'

Eric Dumazet says:

====================
ipv6: fib6_ref conversion to refcount_t

We are chasing use-after-free in IPv6 that could have their origin
in fib6_ref 0 -> 1 transitions.

This patch series should help finding the root causes if these
illegal transitions ever happen.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoipv6: convert fib6_ref to refcount_t
Eric Dumazet [Tue, 23 Apr 2019 01:35:03 +0000 (18:35 -0700)]
ipv6: convert fib6_ref to refcount_t

We suspect some issues involving fib6_ref 0 -> 1 transitions might
cause strange syzbot reports.

Lets convert fib6_ref to refcount_t to catch them earlier.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Wei Wang <weiwan@google.com>
Acked-by: Wei Wang <weiwan@google.com>
Reviewed-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoipv6: broadly use fib6_info_hold() helper
Eric Dumazet [Tue, 23 Apr 2019 01:35:02 +0000 (18:35 -0700)]
ipv6: broadly use fib6_info_hold() helper

Instead of using atomic_inc(), prefer fib6_info_hold()
so that upcoming refcount_t conversion is simpler.

Only fib6_info_alloc() is using atomic_set() since we
just allocated a new object.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Wei Wang <weiwan@google.com>
Acked-by: Wei Wang <weiwan@google.com>
Reviewed-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoipv6: fib6_info_destroy_rcu() cleanup
Eric Dumazet [Tue, 23 Apr 2019 01:35:01 +0000 (18:35 -0700)]
ipv6: fib6_info_destroy_rcu() cleanup

We do not need to clear f6i->rt6i_exception_bucket right before
freeing f6i.

Note that f6i->rt6i_exception_bucket is properly protected by
f6i->exception_bucket_flushed being set to one in rt6_flush_exceptions()
under the protection of rt6_exception_lock.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Wei Wang <weiwan@google.com>
Acked-by: Wei Wang <weiwan@google.com>
Reviewed-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoMerge tag 'mlx5-updates-2019-04-22' of git://git.kernel.org/pub/scm/linux/kernel...
David S. Miller [Wed, 24 Apr 2019 00:03:40 +0000 (17:03 -0700)]
Merge tag 'mlx5-updates-2019-04-22' of git://git./linux/kernel/git/saeed/linux

Saeed Mahameed says:

====================
mlx5-updates-2019-04-22

This series includes updates to mlx5e driver RX data path and some
significant XDP RX/TX improvements to overcome/mitigate HW and PCIE
bottlenecks.

From Tariq:
1) Some Enhancements in rq->flags
2) Stabilize RX packet rate (on Striding RQ) with
multiple outstanding UMR posts
In this patch, we add support for multiple outstanding UMR posts,
 to allow faster gap closure between consuming MPWQEs and reposting
them back into the WQ.

Performance test:
As expected, huge improvement in large-scale (48 cores).

xdp_redirect_map, 64B UDP multi-stream.
Redirect from ConnectX-5 100Gbps to ConnectX-6 100Gbps.
CPU: Intel(R) Xeon(R) CPU E5-2680 v3 @ 2.50GHz.

Before: Unstable, 7 to 30 Mpps
After:  Stable,   at 70.5 Mpps

From Shay:
3) XDP, Inline small packets into the TX MPWQE in XDP xmit flow

Upon high packet rate with multiple CPUs TX workloads, much of the HCA's
resources are spent on prefetching TX descriptors, thus affecting
transmission rates.
This patch comes to mitigate this problem by moving some workload to the
CPU and reducing the HW data prefetch overhead for small packets (<= 256B).

When forwarding packets with XDP, a packet that is smaller
than a certain size (set to ~256 bytes) would be sent inline within
its WQE TX descrptor (mem-copied), when the hardware tx queue is congested
beyond a pre-defined water-mark.

Performance:
    Tested packet rate for UDP 64Byte multi-stream
    over two dual port ConnectX-5 100Gbps NICs.
    CPU: Intel(R) Xeon(R) CPU E5-2680 v3 @ 2.50GHz

    * Tested with hyper-threading disabled

    XDP_TX:

    |          | before | after   |       |
    | 24 rings | 51Mpps | 116Mpps | +126% |
    | 1 ring   | 12Mpps | 12Mpps  | same  |

    XDP_REDIRECT:

    ** Below is the transmit rate, not the redirection rate
    which might be larger, and is not affected by this patch.

    |          | before  | after   |      |
    | 32 rings | 64Mpps  | 92Mpps  | +43% |
    | 1 ring   | 6.4Mpps | 6.4Mpps | same |

As we can see, feature significantly improves scaling, without
hurting single ring performance.

From Maxim:
4) Some trivial refactoring and code improvements prior to a larger series
to support AF_XDP.
====================

Acked-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
5 years agoMerge tag 'nfsd-5.1-1' of git://linux-nfs.org/~bfields/linux
Linus Torvalds [Tue, 23 Apr 2019 20:40:55 +0000 (13:40 -0700)]
Merge tag 'nfsd-5.1-1' of git://linux-nfs.org/~bfields/linux

Pull nfsd bugfixes from Bruce Fields:
 "Fix miscellaneous nfsd bugs, in NFSv4.1 callbacks, NFSv4.1
  lock-notification callbacks, NFSv3 readdir encoding, and the
  cache/upcall code"

* tag 'nfsd-5.1-1' of git://linux-nfs.org/~bfields/linux:
  nfsd: wake blocked file lock waiters before sending callback
  nfsd: wake waiters blocked on file_lock before deleting it
  nfsd: Don't release the callback slot unless it was actually held
  nfsd/nfsd3_proc_readdir: fix buffer count and page pointers
  sunrpc: don't mark uninitialised items as VALID.

5 years agoMerge tag 'syscalls-5.1' of git://git.kernel.org/pub/scm/linux/kernel/git/arnd/asm...
Linus Torvalds [Tue, 23 Apr 2019 20:34:17 +0000 (13:34 -0700)]
Merge tag 'syscalls-5.1' of git://git./linux/kernel/git/arnd/asm-generic

Pull syscall numbering updates from Arnd Bergmann:
 "arch: add pidfd and io_uring syscalls everywhere

  This comes a bit late, but should be in 5.1 anyway: we want the newly
  added system calls to be synchronized across all architectures in the
  release.

  I hope that in the future, any newly added system calls can be added
  to all architectures at the same time, and tested there while they are
  in linux-next, avoiding dependencies between the architecture
  maintainer trees and the tree that contains the new system call"

* tag 'syscalls-5.1' of git://git.kernel.org/pub/scm/linux/kernel/git/arnd/asm-generic:
  arch: add pidfd and io_uring syscalls everywhere

5 years agoBluetooth: hci_qca: Fix crash with non-serdev devices
Matthias Kaehlcke [Tue, 23 Apr 2019 18:16:52 +0000 (11:16 -0700)]
Bluetooth: hci_qca: Fix crash with non-serdev devices

qca_set_baudrate() calls serdev_device_wait_until_sent() assuming that
the HCI is always associated with a serdev device. This isn't true for
ROME controllers instantiated through ldisc, where the call causes a
crash due to a NULL pointer dereferentiation. Only call the function
when we have a serdev device. The timeout for ROME devices at the end
of qca_set_baudrate() is long enough to be reasonably sure that the
command was sent.

Fixes: fa9ad876b8e0 ("Bluetooth: hci_qca: Add support for Qualcomm Bluetooth chip wcn3990")
Reported-by: Balakrishna Godavarthi <bgodavar@codeaurora.org>
Reported-by: Rocky Liao <rjliao@codeaurora.org>
Signed-off-by: Matthias Kaehlcke <mka@chromium.org>
Reviewed-by: Rocky Liao <rjliao@codeaurora.org>
Tested-by: Rocky Liao <rjliao@codeaurora.org>
Reviewed-by: Balakrishna Godavarthi <bgodavar@codeaurora.org>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
5 years agonet/mlx5e: Use #define for the WQE wait timeout constant
Maxim Mikityanskiy [Tue, 5 Mar 2019 12:46:16 +0000 (14:46 +0200)]
net/mlx5e: Use #define for the WQE wait timeout constant

Create a #define for the timeout of mlx5e_wait_for_min_rx_wqes to
clarify the meaning of a magic number.

Signed-off-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
5 years agonet/mlx5e: Remove unused rx_page_reuse stat
Maxim Mikityanskiy [Thu, 21 Mar 2019 13:22:57 +0000 (15:22 +0200)]
net/mlx5e: Remove unused rx_page_reuse stat

Remove the no longer used page_reuse stat of RQs.

Signed-off-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
5 years agonet/mlx5e: Take HW interrupt trigger into a function
Maxim Mikityanskiy [Fri, 1 Mar 2019 10:05:21 +0000 (12:05 +0200)]
net/mlx5e: Take HW interrupt trigger into a function

mlx5e_trigger_irq posts a NOP to the ICO SQ just to trigger an IRQ and
enter the NAPI poll on the right CPU according to the affinity. Use it
in mlx5e_activate_rq.

Signed-off-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
5 years agonet/mlx5e: Remove unused parameter
Maxim Mikityanskiy [Wed, 27 Mar 2019 11:46:50 +0000 (13:46 +0200)]
net/mlx5e: Remove unused parameter

mdev is unused in mlx5e_rx_is_linear_skb.

Signed-off-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
5 years agonet/mlx5e: Add an underflow warning comment
Maxim Mikityanskiy [Wed, 27 Mar 2019 11:39:21 +0000 (13:39 +0200)]
net/mlx5e: Add an underflow warning comment

mlx5e_mpwqe_get_log_rq_size calculates the number of WQEs (N) based on
the requested number of frames in the RQ (F) and the number of packets
per WQE (P). It ensures that N is not less than the minimum number of
WQEs in an RQ (N_min). Arithmetically, it means that F / P >= N_min
should be true. This function deals with logarithms, so it should check
that log(F) - log(P) >= log(N_min). However, if F < P, this expression
will cause an unsigned underflow. Check log(F) >= log(P) + log(N_min)
instead.

Signed-off-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
5 years agonet/mlx5e: Move parameter calculation functions to en/params.c
Maxim Mikityanskiy [Wed, 27 Mar 2019 11:09:27 +0000 (13:09 +0200)]
net/mlx5e: Move parameter calculation functions to en/params.c

This commit moves the parameter calculation functions to a separate file
for better modularity and code sharing with future features.

Signed-off-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
5 years agonet/mlx5e: Report mlx5e_xdp_set errors
Maxim Mikityanskiy [Tue, 19 Mar 2019 16:32:37 +0000 (18:32 +0200)]
net/mlx5e: Report mlx5e_xdp_set errors

If the channels fail to reopen after setting an XDP program, return the
error code instead of 0. A proper fix is still needed, as now any error
while reopening the channels brings the interface down. This patch only
adds error reporting.

Signed-off-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
5 years agonet/mlx5e: Remove unused parameter
Maxim Mikityanskiy [Thu, 7 Mar 2019 17:30:30 +0000 (19:30 +0200)]
net/mlx5e: Remove unused parameter

params is unused in mlx5e_init_di_list.

Signed-off-by: Maxim Mikityanskiy <maximmi@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
5 years agonet/mlx5e: XDP, Inline small packets into the TX MPWQE in XDP xmit flow
Shay Agroskin [Thu, 14 Mar 2019 12:54:07 +0000 (14:54 +0200)]
net/mlx5e: XDP, Inline small packets into the TX MPWQE in XDP xmit flow

Upon high packet rate with multiple CPUs TX workloads, much of the HCA's
resources are spent on prefetching TX descriptors, thus affecting
transmission rates.
This patch comes to mitigate this problem by moving some workload to the
CPU and reducing the HW data prefetch overhead for small packets (<= 256B).

When forwarding packets with XDP, a packet that is smaller
than a certain size (set to ~256 bytes) would be sent inline within
its WQE TX descrptor (mem-copied), when the hardware tx queue is congested
beyond a pre-defined water-mark.

This is added to better utilize the HW resources (which now makes
one less packet data prefetch) and allow better scalability, on the
account of CPU usage (which now 'memcpy's the packet into the WQE).

To load balance between HW and CPU and get max packet rate, we use
watermarks to detect how much the HW is congested and move the work
loads back and forth between HW and CPU.

Performance:
Tested packet rate for UDP 64Byte multi-stream
over two dual port ConnectX-5 100Gbps NICs.
CPU: Intel(R) Xeon(R) CPU E5-2680 v3 @ 2.50GHz

* Tested with hyper-threading disabled

XDP_TX:

|          | before | after   |       |
| 24 rings | 51Mpps | 116Mpps | +126% |
| 1 ring   | 12Mpps | 12Mpps  | same  |

XDP_REDIRECT:

** Below is the transmit rate, not the redirection rate
which might be larger, and is not affected by this patch.

|          | before  | after   |      |
| 32 rings | 64Mpps  | 92Mpps  | +43% |
| 1 ring   | 6.4Mpps | 6.4Mpps | same |

As we can see, feature significantly improves scaling, without
hurting single ring performance.

Signed-off-by: Shay Agroskin <shayag@mellanox.com>
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>