Commit Graph

814531 Commits

Author SHA1 Message Date
Dedy Lansky
7b834639c4 wil6210: use OEM MAC address from OTP
In addition to existing MAC address field in OTP, new field added for
OEM MAC address.
wil6210 gives precedence to the new OEM MAC address and will use it if
its valid.

Signed-off-by: Dedy Lansky <dlansky@codeaurora.org>
Signed-off-by: Maya Erez <merez@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
2019-04-03 15:34:57 +03:00
Ahmad Masri
e4a29bdd8f wil6210: check mid is valid
Check that the mid is valid and that it does not exceed the memory
size allocated to vifs array.

Signed-off-by: Ahmad Masri <amasri@codeaurora.org>
Signed-off-by: Maya Erez <merez@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
2019-04-03 15:34:47 +03:00
Maya Erez
044974fbea wil6210: update WIL_MCS_MAX to 15
Update max MCS to 15, which is supported by Talyn-MB.
This will allow collecting statistics on number of RX packets
in higher MCS.

Signed-off-by: Maya Erez <merez@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
2019-04-03 15:34:36 +03:00
Maya Erez
f6194f769d wil6210: do not set BIT_USER_SUPPORT_T_POWER_ON_0 in Talyn-MB
In Sparrow, FW might sleep long time due to T_Power_On calculation
in slow clock, so T_Power_On was set to zero to shorten the L1SS
wake-up time.
In Talyn-MB the L1SS wake-up procedure is handled by the PMU (HW),
hence T_Power_On calculation is accurate and should not be forced
to zero.

Signed-off-by: Maya Erez <merez@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
2019-04-03 15:34:26 +03:00
Maya Erez
5793fe9d4f wil6210: increase PCP stop command timeout
In case there are connected stations, FW needs to disconnect
them before handling PCP stop. This flow can take several
seconds.
Increasing PCP stop timeout to 5 seconds to allow that.

Signed-off-by: Maya Erez <merez@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
2019-04-03 15:34:15 +03:00
Ahmad Masri
a061894587 wil6210: prevent device memory access while in reset or suspend
Accessing some of the memory of the device while the device is
resetting or suspending may cause unexpected error as the HW is still
not in a stable state. Prevent this access to guarantee successful
read/write memory operations.

Signed-off-by: Ahmad Masri <amasri@codeaurora.org>
Signed-off-by: Maya Erez <merez@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
2019-04-03 15:34:04 +03:00
Alexei Avshalom Lazar
73a7d1e34d wil6210: align to latest auto generated wmi.h
Align to latest version of the auto generated wmi file
describing the interface with FW.

Signed-off-by: Alexei Avshalom Lazar <ailizaro@codeaurora.org>
Signed-off-by: Maya Erez <merez@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
2019-04-03 15:33:53 +03:00
Alexei Avshalom Lazar
de77a53c2d wil6210: check null pointer in _wil_cfg80211_merge_extra_ies
ies1 or ies2 might be null when code inside
_wil_cfg80211_merge_extra_ies access them.
Add explicit check for null and make sure ies1/ies2 are not
accessed in such a case.

spos might be null and be accessed inside
_wil_cfg80211_merge_extra_ies.
Add explicit check for null in the while condition statement
and make sure spos is not accessed in such a case.

Signed-off-by: Alexei Avshalom Lazar <ailizaro@codeaurora.org>
Signed-off-by: Maya Erez <merez@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
2019-02-28 11:25:09 +02:00
Maya Erez
979c9d8d01 wil6210: ignore HALP ICR if already handled
HALP ICR is set as long as the FW should stay awake.
To prevent its multiple handling the driver masks this IRQ bit.
However, if there is a different MISC ICR before the driver clears
this bit, there is a risk of race condition between HALP mask and
unmask. This race leads to HALP timeout, in case it is mistakenly
masked.
Add an atomic flag to indicate if HALP ICR should be handled.

Signed-off-by: Maya Erez <merez@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
2019-02-28 11:24:59 +02:00
Dedy Lansky
a380eb5736 wil6210: fix invalid sta statistics update
Upon status ring handling, in case there are both unicast and
multicast (cid == max) status messages to handle, wrong sta statistics
might get updated.
Fix this by setting stats to NULL upon invalid cid
(e.g. == max_assoc_sta).

Signed-off-by: Dedy Lansky <dlansky@codeaurora.org>
Signed-off-by: Maya Erez <merez@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
2019-02-28 11:24:50 +02:00
Ahmad Masri
6d1ba32c80 wil6210: accessing 802.3 addresses via utility functions
Rearrange the code by having functions to access 802.3 header
members, source and destination addresses.

Signed-off-by: Ahmad Masri <amasri@codeaurora.org>
Signed-off-by: Maya Erez <merez@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
2019-02-28 11:24:39 +02:00
Ahmad Masri
bf0353a674 wil6210: support up to 20 stations in AP mode
New FW added support for upto 20 clients in AP mode. Change the driver
to support this as well. FW reports it's max supported associations in
WMI_READY_EVENT. Some WMI commands/events use cidxtid field which is
limited to 16 cids. Use new cid/tid fields instead.

For Rx packets cid from rx descriptor is limited to 3 bits (0..7),
to find the real cid, compare transmitter address with the stored
stations mac address in the driver sta array.

EDMA FW still supports 8 stations. Extending the support to 20
stations will come later.

Signed-off-by: Ahmad Masri <amasri@codeaurora.org>
Signed-off-by: Maya Erez <merez@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
2019-02-28 11:24:29 +02:00
Dedy Lansky
0439a5e035 wil6210: add option to drop Tx packets when Tx ring is full
In AP mode with multiple clients, driver stops net queue
(netif_tx_stop_queue) upon first ring (serving specific client)
becoming full. This can have negative effect on transmission to
other clients which may still have room in their corresponding rings.

Implement new policy in which stop/wake net queue are not used. In
case there is no room in the ring for a transmitted packet, drop the
packet.

New policy can be helpful to debug performance issues, to guarantee
maximum utilization of net queues.
New policy is disabled by default and can be enabled by debugfs:
echo 1 > drop_if_ring_full

Signed-off-by: Dedy Lansky <dlansky@codeaurora.org>
Signed-off-by: Maya Erez <merez@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
2019-02-28 11:24:19 +02:00
Maya Erez
387f3794b8 wil6210: remove rtap_include_phy_info module param
Due to a HW issue in PHY info collection rtap_include_phy_info is not
in use, hence can be removed.

Signed-off-by: Maya Erez <merez@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
2019-02-28 11:24:09 +02:00
Surabhi Vishnoi
f40a307eb9 ath10k: Fill rx duration for each peer in fw_stats for WCN3990
Currently, rx_duration for each peer is not getting populated in
fw_stats debugfs entry for WCN3990.

WCN3990 firmware sends rx duration for each peer as part of
peer_extd_stats in WMI_UPDATE_STATS_EVENT. To enable peer_extd_stats,
firmware expects host to send fw_stats_req_mask with flag
WMI_TLV_PEER_STATS_EXTD set in WMI_REQUEST_STATS_CMD.

Send fw_stats_req_mask with flag WMI_TLV_PEER_STATS_EXTD set in
WMI_REQUEST_STATS_CMD and parse the peer_extd_stats in
WMI_UPDATE_STATS_EVENT to populate the rx_duration of each peer
in fw_stats debugfs entry.

Currently the driver handles 32-bit rx_duration, but the rx_duration
for WCN3990 can be upto 63 bit. The firmware sends rx_duration split
into two 32-bit fields, with the upper 32-bits being valid only if its
MSB is set. This change handles the 63-bit rx_duration obtained from
WCN3990 and maintain the backward compatibility.

To get the rx_duration of each connected peer :
cat /sys/kernel/debug/ieee80211/phyX/ath10k/fw_stats

Tested HW: WCN3990
Tested FW: WLAN.HL.3.1-00784-QCAHLSWMTPLZ-1

Signed-off-by: Surabhi Vishnoi <svishnoi@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
2019-02-28 10:48:19 +02:00
Surabhi Vishnoi
d23c2cdaa0 ath10k: Fix the wrong calculation ht_idx and idx of rate table for tx_stats
ht_idx (ht rate index) and idx (rate table index) are calculated based on
mcs index. This mcs index used in the above calculation should be 0-9 for
getting the correct ht_idx and idx.

Currently the mcs index used for the above calculations is mcs index which
can be 0-31 (in case of HT), leading to incorrect rate index and ht index values.

Fix the issue by obtaining mcs value from the ratecode reported by firmware
and use it for calculating ht_idx and idx (rate-table index).

Tested HW: WCN3990
Tested FW: WLAN.HL.3.1-00784-QCAHLSWMTPLZ-1

Fixes: e88975ca37 ("ath10k: dump tx stats in rate table format")
Signed-off-by: Surabhi Vishnoi <svishnoi@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
2019-02-28 10:46:26 +02:00
Surabhi Vishnoi
8e55fdaa8e ath10k: Fix the wrong updation of SGI in tx_stats debugfs
The SGI is updated wrongly in tx stats table in debugfs per sta
entry. To know whether the packets/bytes are sent with SHORT GI,
test whether the SGI bit(ATH10K_RATE_INFO_FLAGS_SGI_BIT) is set or
not in the txrate flags.

Tested HW: WCN3990
Tested FW: WLAN.HL.3.1-00784-QCAHLSWMTPLZ-1

Fixes: a904417fc8 ("ath10k: add extended per sta tx statistics support")
Signed-off-by: Surabhi Vishnoi <svishnoi@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
2019-02-28 10:46:16 +02:00
Surabhi Vishnoi
ef9051c72a ath10k: Fix the wrong updation of BW in tx_stats debugfs entry
Currently, the bandwidth is updated wrongly in BW table in tx_stats
debugfs per sta as there is difference in number of bandwidth type
in mac80211 and driver stats table. This leads to bandwidth getting
updated at wrong index in bandwidth table in tx_stats.

Fix this index mismatch between mac80211 and driver stats table (BW table)
by making the number of bandwidth type in driver compatible with mac80211.

Tested HW: WCN3990
Tested FW: WLAN.HL.3.1-00784-QCAHLSWMTPLZ-1

Fixes: a904417fc8 ("ath10k: add extended per sta tx statistics support")
Signed-off-by: Surabhi Vishnoi <svishnoi@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
2019-02-28 10:46:06 +02:00
Surabhi Vishnoi
3a08ac3e79 ath10k: Fix the incorrect updation of NSS data in tx stats
The NSS data is updated incorrectly in the tx stats as the array
indexing starts from zero.

Fix the incorrect updation of NSS data in tx_stats by taking into
consideration the array index starting from zero.

Tested HW: WCN3990
Tested FW: WLAN.HL.3.1-00784-QCAHLSWMTPLZ-1

Fixes: a904417fc8 ("ath10k: add extended per sta tx statistics support")
Signed-off-by: Surabhi Vishnoi <svishnoi@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
2019-02-28 10:45:56 +02:00
YueHaibing
03af21d6ba ath9k: remove set but not used variable 'acq'
Fixes gcc '-Wunused-but-set-variable' warning:

drivers/net/wireless/ath/ath9k/recv.c: In function 'ath_rx_count_airtime':
drivers/net/wireless/ath/ath9k/recv.c:1010:18: warning:
 variable 'acq' set but not used [-Wunused-but-set-variable]

It's not used after 89cea7493a ("ath9k: Switch to mac80211 TXQ scheduling
and airtime APIs"). Also remove related variables.

Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Acked-by: Toke Høiland-Jørgensen <toke@redhat.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
2019-02-26 15:09:00 +02:00
Andrea Greco
d0480d4326 ath9k: debugfs: Fix SPUR-DOWN field
SPUR DOWN field returns spurup instead of spurdown.

Signed-off-by: Andrea Greco <a.greco@4sigma.it>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
2019-02-26 15:08:16 +02:00
Toke Høiland-Jørgensen
cc591d77ab ath9k: Make sure to zero status.tx_time before reporting TX status
Since ath9k reports airtime usage directly using the
ieee80211_report_airtime() callback, it shouldn't also report it using the
tx_time in status. Make sure the field is zeroed before TX status is
reported to avoid spurious airtime being accounted by bits being left over
from earlier uses of the cb.

Fixes: 89cea7493a ("ath9k: Switch to mac80211 TXQ scheduling and airtime APIs")
Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
2019-02-26 15:07:28 +02:00
Abhishek Ambure
15493239ea ath10k: update the max num of peers supported for WCN3990
WCN3990 firmware versions WLAN.HL.2.0-01617-QCAHLSWMTPLZ-1 & onwards
supports maximum 33 peers including self peer. To support maximum peers,
send updated peer param to firmware during initialization.

Tested HW: WCN3990
Tested FW: WLAN.HL.3.1-00784-QCAHLSWMTPLZ-1

Signed-off-by: Abhishek Ambure <aambure@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
2019-02-26 15:06:38 +02:00
Wen Gong
6566abea0b ath10k: remove the calibration data fetch for sdio
The calibration data fetch will trigger sdio error, then sdio will
become fail untill reboot system.

If happens when run ifconfig wlan down, then ifconfig wlan up will
fail untill reboot system.Remove it fix the ifconfig wlan issue.

Tested with QCA6174 SDIO with firmware
WLAN.RMH.4.4.1-00005-QCARMSWP-1.

Signed-off-by: Wen Gong <wgong@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
2019-02-26 15:05:26 +02:00
Surabhi Vishnoi
761156ff57 ath10k: Fix length of wmi tlv command for protected mgmt frames
The length of wmi tlv command for management tx send is calculated
incorrectly in case of protected management frames as there is addition
of IEEE80211_CCMP_MIC_LEN twice. This leads to improper behaviour of
firmware as the wmi tlv mgmt tx send command for protected mgmt frames
is formed wrongly.

Fix the length calculation of wmi tlv command for mgmt tx send in case
of protected management frames by adding the IEEE80211_CCMP_MIC_LEN only
once.

Tested HW: WCN3990
Tested FW: WLAN.HL.3.1-00784-QCAHLSWMTPLZ-1

Fixes: 1807da4973 "ath10k: wmi: add management tx by reference support over wmi"
Signed-off-by: Surabhi Vishnoi <svishnoi@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
2019-02-26 15:03:54 +02:00
Alagu Sankar
7d44452230 ath10k: don't report unset rssi values to mac80211
The SDIO firmware does not provide RSSI value to the host, it's only set to
zero. In that case don't report the value to mac80211. One risk here is that
value zero might be a valid value with other firmware, currently there's no way
to detect that.

Without the fix, the rssi value indicated by iw changes between the actual
value and -95.

Tested with QCA6174 SDIO with firmware WLAN.RMH.4.4.1-00005-QCARMSWP-1.

Co-developed-by: Wen Gong <wgong@codeaurora.org>
Signed-off-by: Alagu Sankar <alagusankar@silex-india.com>
Signed-off-by: Wen Gong <wgong@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
2019-02-26 15:02:18 +02:00
Wen Gong
55545b0870 ath10k: sdio: reset chip on power_down()
The target device needs to be reset during power_down(), otherwise only the
first power_up() will work. And as ath10k calls power_up() during driver
initialisation the driver would be otherwise unusable.

Tested with QCA6174 SDIO with firmware WLAN.RMH.4.4.1-00005-QCARMSWP-1.

Signed-off-by: Wen Gong <wgong@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
2019-02-26 15:02:08 +02:00
Alagu Sankar
6cd70c6564 ath10k: sdio: disable fwlog prints
The SDIO firmware may turn it on based on scratch registers so disable the
firmware log to avoid that.

Co-developed-by: Wen Gong <wgong@codeaurora.org>
Signed-off-by: Alagu Sankar <alagusankar@silex-india.com>
Signed-off-by: Wen Gong <wgong@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
2019-02-26 15:01:58 +02:00
Alagu Sankar
bf1f0a1a4d ath10k: sdio: set hi_acs_flags
The SDIO firmware does not allow transmitting packets with the
reduced tx completion HI_ACS option. SDIO firmware uses 1544 as
alternate credit size, which is not big enough for the maximum sized
mac80211 frames. Disable both these HI_ACS flags for SDIO.

Co-developed-by: Wen Gong <wgong@codeaurora.org>
Signed-off-by: Alagu Sankar <alagusankar@silex-india.com>
Signed-off-by: Wen Gong <wgong@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
2019-02-26 15:01:49 +02:00
Yu Wang
d961284df2 ath10k: correct the format of host memory chunks in wmi init command
This is a theoretical fix, the issue is found in code review.
When adding the host memory chunks into wmi-tlv init command,
there is no separate tlv header for each host memory chunk
in the struct array, which breaches the convention between
host and firmware, will result in mismatch between the two.

To fix this issue, add separate tlv headers for the host
memory chunks in wmi-tlv init command.

Signed-off-by: Yu Wang <yyuwang@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
2019-02-26 15:00:08 +02:00
Rakesh Pillai
1c136e41fb ath10k: enhance logging for vdev pdev & peer set param
Currently after enabling the WMI debug logging,
there is no detail printed about the param id
and the param value for the pdev, vdev and
peer params which are set.

Enhance the WMI logging to print the param id
and the param value for pdev, vdev and peer set
param wmi commands.

Tested HW: WCN3990
Tested FW: WLAN.HL.2.0-01387-QCAHLSWMTPLZ-1
		WLAN.HL.3.1-00784-QCAHLSWMTPLZ-1

Signed-off-by: Rakesh Pillai <pillair@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
2019-02-26 14:59:15 +02:00
Abhishek Ambure
6ddc3860a5 ath10k: add support for ack rssi value of data tx packets
In WCN3990, WMI_TLV_SERVICE_TX_DATA_MGMT_ACK_RSSI service Indicates that
the firmware has the capability to send the RSSI value of the ACK for all
data and management packets transmitted.

If WMI_RSRC_CFG_FLAG_TX_ACK_RSSI is set in host capability then firmware
sends RSSI value in "data" tx completion event. Host extracts ack rssi
values of data packets from their tx completion event.

Tested HW: WCN3990
Tested FW: WLAN.HL.2.0-01617-QCAHLSWMTPLZ-1

Signed-off-by: Abhishek Ambure <aambure@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
2019-02-26 14:58:06 +02:00
Abhishek Ambure
4b816f170b ath10k: add support for ack rssi value of management tx packets
In WCN3990, WMI_TLV_SERVICE_TX_DATA_MGMT_ACK_RSSI service Indicates that
the firmware has the capability to send the RSSI value of the ACK for all
data and management packets transmitted.

If WMI_RSRC_CFG_FLAG_TX_ACK_RSSI is set in host capability then firmware
sends RSSI value in "management" tx completion event. Host extracts ack
rssi values of management packets from their tx completion event.

Tested HW: WCN3990
Tested FW: WLAN.HL.2.0-01617-QCAHLSWMTPLZ-1

Signed-off-by: Abhishek Ambure <aambure@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
2019-02-26 14:57:56 +02:00
Rakesh Pillai
02f73d3a9b ath10k: fix descriptor size in ce tx completion for WCN3990
When the driver receives the tx completion of the
descriptor over ce, it clears the nbytes configured
for that particular descriptor. WCN3990 uses ce
descriptors with 64-bit address.

Currently during handling the tx completion of the
descriptors, the nbytes are accessed from the descriptors
using ce_desc for 32-bit targets. This will lead to clearing
of memory at incorrect offset if DMA MASK is set to greater
than 32 bits.

Attach different ce tx copy completed handler for targets
using address above 32-bit address.

Tested HW: WCN3990
Tested FW: WLAN.HL.2.0-01387-QCAHLSWMTPLZ-1

Signed-off-by: Rakesh Pillai <pillair@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
2019-02-26 14:56:19 +02:00
Peter Oskolkov
bd16693f35 net: fix double-free in bpf_lwt_xmit_reroute
dst_output() frees skb when it fails (see, for example,
ip_finish_output2), so it must not be freed in this case.

Fixes: 3bd0b15281 ("bpf: add handling of BPF_LWT_REROUTE to lwt_bpf.c")
Signed-off-by: Peter Oskolkov <posk@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2019-02-24 22:24:50 -08:00
wenxu
186d93669f ip_tunnel: Add ip tunnel tun_info type dst_cache in ip_tunnel_xmit
ip l add dev tun type gretap key 1000

Non-tunnel-dst ip tunnel device can send packet through lwtunnel
This patch provide the tun_inf dst cache support for this mode.

Signed-off-by: wenxu <wenxu@ucloud.cn>
Signed-off-by: David S. Miller <davem@davemloft.net>
2019-02-24 22:22:33 -08:00
David S. Miller
169431ed16 Merge branch 'dsa-mv88e6xxx-lockdep'
Andrew Lunn says:

====================
mv88e6xxx: Avoid false positive Lockdep splats

When acquiring the GPIO interrupt line for the switch, it is possible
to trigger lockdep splats. These are false positives, the mutex is in
a different IRQ descriptor. But fix it anyway, since it could mask
real locking issues.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2019-02-24 22:21:23 -08:00
Andrew Lunn
342a0ee70a net: dsa: mv88e6xxx: Release lock while requesting IRQ
There is no need to hold the register lock while requesting the GPIO
interrupt. By not holding it we can also avoid a false positive
lockdep splat.

Reported-by: Russell King <rmk+kernel@armlinux.org.uk>
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
2019-02-24 22:21:23 -08:00
Andrew Lunn
f6d9758b12 net: dsa: mv88e6xxx: Add lockdep classes to fix false positive splat
The following false positive lockdep splat has been observed.

======================================================
WARNING: possible circular locking dependency detected
4.20.0+ #302 Not tainted
------------------------------------------------------
systemd-udevd/160 is trying to acquire lock:
edea6080 (&chip->reg_lock){+.+.}, at: __setup_irq+0x640/0x704

but task is already holding lock:
edff0340 (&desc->request_mutex){+.+.}, at: __setup_irq+0xa0/0x704

which lock already depends on the new lock.

the existing dependency chain (in reverse order) is:

-> #1 (&desc->request_mutex){+.+.}:
       mutex_lock_nested+0x1c/0x24
       __setup_irq+0xa0/0x704
       request_threaded_irq+0xd0/0x150
       mv88e6xxx_probe+0x41c/0x694 [mv88e6xxx]
       mdio_probe+0x2c/0x54
       really_probe+0x200/0x2c4
       driver_probe_device+0x5c/0x174
       __driver_attach+0xd8/0xdc
       bus_for_each_dev+0x58/0x7c
       bus_add_driver+0xe4/0x1f0
       driver_register+0x7c/0x110
       mdio_driver_register+0x24/0x58
       do_one_initcall+0x74/0x2e8
       do_init_module+0x60/0x1d0
       load_module+0x1968/0x1ff4
       sys_finit_module+0x8c/0x98
       ret_fast_syscall+0x0/0x28
       0xbedf2ae8

-> #0 (&chip->reg_lock){+.+.}:
       __mutex_lock+0x50/0x8b8
       mutex_lock_nested+0x1c/0x24
       __setup_irq+0x640/0x704
       request_threaded_irq+0xd0/0x150
       mv88e6xxx_g2_irq_setup+0xcc/0x1b4 [mv88e6xxx]
       mv88e6xxx_probe+0x44c/0x694 [mv88e6xxx]
       mdio_probe+0x2c/0x54
       really_probe+0x200/0x2c4
       driver_probe_device+0x5c/0x174
       __driver_attach+0xd8/0xdc
       bus_for_each_dev+0x58/0x7c
       bus_add_driver+0xe4/0x1f0
       driver_register+0x7c/0x110
       mdio_driver_register+0x24/0x58
       do_one_initcall+0x74/0x2e8
       do_init_module+0x60/0x1d0
       load_module+0x1968/0x1ff4
       sys_finit_module+0x8c/0x98
       ret_fast_syscall+0x0/0x28
       0xbedf2ae8

other info that might help us debug this:

 Possible unsafe locking scenario:

       CPU0                    CPU1
       ----                    ----
  lock(&desc->request_mutex);
                               lock(&chip->reg_lock);
                               lock(&desc->request_mutex);
  lock(&chip->reg_lock);

&desc->request_mutex refer to two different mutex. #1 is the GPIO for
the chip interrupt. #2 is the chained interrupt between global 1 and
global 2.

Add lockdep classes to the GPIO interrupt to avoid this.

Reported-by: Russell King <rmk+kernel@armlinux.org.uk>
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
2019-02-24 22:21:23 -08:00
wenxu
3d25eabbbf ip_tunnel: Add dst_cache support in lwtunnel_state of ip tunnel
The lwtunnel_state is not init the dst_cache Which make the
ip_md_tunnel_xmit can't use the dst_cache. It will lookup
route table every packets.

Signed-off-by: wenxu <wenxu@ucloud.cn>
Signed-off-by: David S. Miller <davem@davemloft.net>
2019-02-24 22:13:49 -08:00
Vakul Garg
2b794c4098 tls: Return type of non-data records retrieved using MSG_PEEK in recvmsg
The patch enables returning 'type' in msghdr for records that are
retrieved with MSG_PEEK in recvmsg. Further it prevents records peeked
from socket from getting clubbed with any other record of different
type when records are subsequently dequeued from strparser.

For each record, we now retain its type in sk_buff's control buffer
cb[]. Inside control buffer, record's full length and offset are already
stored by strparser in 'struct strp_msg'. We store record type after
'struct strp_msg' inside 'struct tls_msg'. For tls1.2, the type is
stored just after record dequeue. For tls1.3, the type is stored after
record has been decrypted.

Inside process_rx_list(), before processing a non-data record, we check
that we must be able to return back the record type to the user
application. If not, the decrypted records in tls context's rx_list is
left there without consuming any data.

Fixes: 692d7b5d1f ("tls: Fix recvmsg() to be able to peek across multiple records")
Signed-off-by: Vakul Garg <vakul.garg@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2019-02-24 21:58:38 -08:00
David S. Miller
2bdeb8e5bd Merge branch 'ipv4-v6-icmp-small-cleanup-and-update'
Kefeng Wang says:

====================
ipv4/v6: icmp: small cleanup and update

v2:
- Add cover letter and user proper patch subject-prefix suggested-by Eric Dumazet

This patch series contains some small cleanup and update,
1) use icmp/v6_sk_exit when icmp_sk_init fails instead of open-code
2) use new percpu allocation interface for the ipv6.icmp_sk
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2019-02-24 21:57:26 -08:00
Kefeng Wang
75efc250d2 ipv6: icmp: use percpu allocation
Use percpu allocation for the ipv6.icmp_sk.

Signed-off-by: Kefeng Wang <wangkefeng.wang@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2019-02-24 21:57:26 -08:00
Kefeng Wang
3232a1ef0f ipv6: icmp: use icmpv6_sk_exit()
Simply use icmpv6_sk_exit() when inet_ctl_sock_create() fail
in icmpv6_sk_init().

Signed-off-by: Kefeng Wang <wangkefeng.wang@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2019-02-24 21:57:26 -08:00
Kefeng Wang
e9128c14bf ipv4: icmp: use icmp_sk_exit()
Simply use icmp_sk_exit() when inet_ctl_sock_create() fail in icmp_sk_init().

Signed-off-by: Kefeng Wang <wangkefeng.wang@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2019-02-24 21:57:26 -08:00
Herbert Xu
4ef595cbb3 ila: Fix uninitialised return value in ila_xlat_nl_cmd_flush
This patch fixes an uninitialised return value error in
ila_xlat_nl_cmd_flush.

Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Fixes: 6c4128f658 ("rhashtable: Remove obsolete...")
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
2019-02-24 21:54:44 -08:00
wenxu
41411e2fd6 net/sched: act_tunnel_key: Add dst_cache support
The metadata_dst is not init the dst_cache which make the
ip_md_tunnel_xmit can't use the dst_cache. It will lookup
route table every packets.

Signed-off-by: wenxu <wenxu@ucloud.cn>
Signed-off-by: David S. Miller <davem@davemloft.net>
2019-02-24 21:51:55 -08:00
David S. Miller
caf337bdef Merge branch 'code-optimizations-and-bugfixes-for-HNS3-driver'
Huazhong Tan says:

====================
code optimizations & bugfixes for HNS3 driver

This patchset includes bugfixes and code optimizations for
the HNS3 ethernet controller driver.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2019-02-24 20:27:51 -08:00
Huazhong Tan
186551284e net: hns3: fix improper error handling for hns3_client_start
If hns3_client_start() failed in the hns3_client_init(),
register_dev() should be undo in its error handling.

Fixes: a6d818e31d ("net: hns3: Add vport alive state checking support")
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>
2019-02-24 20:27:50 -08:00
Shiju Jose
eb4c2ccbad net: hns3: fix setting of the hns reset_type for rdma hw errors
Presently the hns reset_type for the roce errors is set
in the hclge_log_and_clear_rocee_ras_error function.
This function is also called to detect and clear roce errors
while enabling the rdma error interrupts. However there is no hns
reset requested for this case. This can cause issue of wrong
reset_type used with subsequent hns reset as the
reset_type set in the above case was not cleared.

This patch moves setting of hns reset_type for the roce errors from
hclge_log_and_clear_rocee_ras_error function
to hclge_handle_rocee_ras_error.

Fixes: 630ba007f4 ("net: hns3: add handling of RDMA RAS errors")
Reported-by: Huazhong Tan <tanhuazhong@huawei.com>
Reported-by: Xiaofei Tan <tanxiaofei@huawei.com>
Signed-off-by: Shiju Jose <shiju.jose@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>
2019-02-24 20:27:50 -08:00