- 07 Nov, 2019 11 commits
-
-
Xin Long authored
Without options copied to the dst tun_info in iptunnel_metadata_reply() called by arp_process for handling arp_request, the generated arp_reply packet may be dropped or sent out with wrong options for some tunnels like erspan and vxlan, and the traffic will break. Fixes: 63d008a4 ("ipv4: send arp replies to the correct tunnel") Signed-off-by:
Xin Long <lucien.xin@gmail.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Hoang Le authored
With huge cluster (e.g >200nodes), the amount of that flow: gap -> retransmit packet -> acked will take time in case of STATE_MSG dropped/delayed because a lot of traffic. This lead to 1.5 sec tolerance value criteria made link easy failure around 2nd, 3rd of failed retransmission attempts. Instead of re-introduced criteria of 99 faled retransmissions to fix the issue, we increase failure detection timer to ten times tolerance value. Fixes: 77cf8edb ("tipc: simplify stale link failure criteria") Acked-by:
Jon Maloy <jon.maloy@ericsson.com> Signed-off-by:
Hoang Le <hoang.h.le@dektech.com.au> Acked-by: Jon Signed-off-by:
-
Hoang Le authored
There are two improvements when re-calculate cluster capabilities: - When deleting a specific down node, need to re-calculate. - In tipc_node_cleanup(), do not need to re-calculate if node is still existing in cluster. Acked-by:
-
Francesco Ruggeri authored
Added the following traceroute tests. IPV6: Verify that in this scenario ------------------------ N2 | | ------ ------ N3 ---- | R1 | | R2 |------|H2| ------ ------ ---- | | ------------------------ N1 | ---- |H1| ---- where H1's default route goes through R1 and R1's default route goes through R2 over N2, traceroute6 from H1 to H2 reports R2's address on N2 and not N1. IPV4: Verify that traceroute from H1 to H2 shows 1.0.1.1 in this scenario 1.0.3.1/24 ---- 1.0.1.3/24 1.0.1.1/24 ---- 1.0.2.1/24 1.0.2.4/24 ---- |H1|--------------------------|R1|--------------------------|H2| ---- N1 ---- N2 ---- where net.ipv4.icmp_errors_use_inbound_ifaddr is set on R1 and 1.0.3.1/24 and 1.0.1.1/24 are respectively R1's primary and secondary address on N1. v2: fixed some typos, and have bridge in R1 instead of R2 in IPV6 test. Signed-off-by:Francesco Ruggeri <fruggeri@arista.com> Reviewed-by:
David Ahern <dsahern@gmail.com> Signed-off-by:
-
David S. Miller authored
Eric Dumazet says: ==================== net: various KCSAN inspired fixes This is a series of minor fixes, mostly dealing with lockless accesses to socket 'sk_ack_backlog', 'sk_max_ack_backlog' ane neighbour 'confirmed' fields. ==================== Signed-off-by: -
Eric Dumazet authored
sk->sk_max_ack_backlog can be read without any lock being held at least in TCP/DCCP cases. We need to use READ_ONCE()/WRITE_ONCE() to avoid load/store tearing and/or potential KCSAN warnings. Signed-off-by:
Eric Dumazet <edumazet@google.com> Signed-off-by:
-
Eric Dumazet authored
sk->sk_ack_backlog can be read without any lock being held. We need to use READ_ONCE()/WRITE_ONCE() to avoid load/store tearing and/or potential KCSAN warnings. Signed-off-by:
-
Eric Dumazet authored
Writers are holding a lock, but many readers do not. Following patch will add appropriate barriers in sk_acceptq_removed() and sk_acceptq_added(). Signed-off-by:
-
Eric Dumazet authored
There are common instances of the following construct : if (n->confirmed != now) n->confirmed = now; A C compiler could legally remove the conditional. Use READ_ONCE()/WRITE_ONCE() to avoid this problem. Signed-off-by:
-
Eric Dumazet authored
Use jiffies_delta_to_msecs() to avoid reporting 'infinite' timeouts and to cleanup code. Signed-off-by:
-
Eric Dumazet authored
A difference of two unsigned long needs long storage. Fixes: c7fb64db ("[NETLINK]: Neighbour table configuration and statistics via rtnetlink") Signed-off-by:
-
- 06 Nov, 2019 29 commits
-
-
Roman Mashak authored
Added tests for u8/u32 clear value, u8/16 retain value, u16/32 invert value, u8/u16/u32 preserve value and test for negative offsets. Signed-off-by:
Roman Mashak <mrv@mojatatu.com> Signed-off-by:
-
Jakub Kicinski authored
The netdevsim object is reused by all the tests, but the resource tests puts it into a broken state (failed reload in a different namespace). Make sure it's fixed up at the end of that test otherwise subsequent tests fail. Fixes: b74c37fd ("selftests: netdevsim: add tests for devlink reload with resources") Signed-off-by:
Jakub Kicinski <jakub.kicinski@netronome.com> Acked-by:
Jiri Pirko <jiri@mellanox.com> Signed-off-by:
-
Claudiu Manoil authored
Until now the size of a Rx buffer was artificially limited to 1536B (which happens to be the default, after reset, hardware value for a Rx buffer). This approach however leaves unused memory space for Rx packets, since the driver uses a paged allocation scheme that reserves half a page for each Rx skb. There's also the inconvenience that frames around 1536 bytes can get scattered if the limit is slightly exceeded. This limit can be exceeded even for standard MTU of 1500B traffic, for common cases like stacked VLANs, or DSA tags. To address these issues, let's just compute the buffer size starting from the upper limit of 2KB (half a page) and subtract the skb overhead and alignment restrictions. Signed-off-by:
Claudiu Manoil <claudiu.manoil@nxp.com> Signed-off-by:
-
Yunfeng Ye authored
The function page_shift() is supported after the commit 94ad9338 ("mm: introduce page_shift()"). So replace with page_shift() in ehea_is_hugepage() for readability. Signed-off-by:
Yunfeng Ye <yeyunfeng@huawei.com> Signed-off-by:
-
Zhu Yanjun authored
This change adds support for xmit_more based on the igb commit 6f19e12f ("igb: flush when in xmit_more mode and under descriptor pressure") and commit 6b16f9ee ("net: move skb->xmit_more hint to softnet data") that were made to igb to support this feature. The function netif_xmit_stopped is called to check whether transmit queue on device is currently unable to send to determine whether we must write the tail because we can add no further buffers. When normal packets and/or xmit_more packets fill up tx_desc, it is necessary to trigger NIC tx reg. Following the advice from David Miller and Jakub Kicinski, after the xmit_more feature is added, the following scenario will occur. | xmit_more packets | DMA_MAPPING | DMA_MAPPING error check | xmit_more packets already in HW xmit queue | In the above scenario, if DMA_MAPPING error occurrs, the xmit_more packets already in HW xmit queue will also be dropped. This is different from the behavior before xmit_more feature. So it is necessary to trigger NIC HW tx reg in the above scenario. To the non-xmit_more packets, the above scenario will not occur. Tested: - pktgen (xmit_more packets) SMP x86_64 -> Test command: ./pktgen_sample03_burst_single_flow.sh ... -b 8 -n 1000000 Test results: Params: ... burst: 8 ... Result: OK: 12194004(c12188996+d5007) usec, 1000001 (1500byte,0frags) 82007pps 984Mb/sec (984084000bps) errors: 0 - iperf (normal packets) SMP x86_64 -> Test command: Server: iperf -s Client: iperf -c serverip Result: TCP window size: 85.0 KByte (default) ------------------------------------------------------------ [ ID] Interval Transfer Bandwidth [ 3] 0.0-10.0 sec 1.10 GBytes 942 Mbits/sec CC: Joe Jin <joe.jin@oracle.com> CC: JUNXIAO_BI <junxiao.bi@oracle.com> Reported-and-tested-by:
Nan san <nan.1986san@gmail.com> Signed-off-by:
Zhu Yanjun <yanjun.zhu@oracle.com> Acked-by:
-
David S. Miller authored
Jakub Kicinski says: ==================== netdevsim: fix tests and netdevsim The first patch fixes a merge which brought back some dead code. Next a tiny re-write of the main test using netdevsim aims to ease debugging. ==================== Signed-off-by: -
Jakub Kicinski authored
Recent changes to netdevsim moved creating and destroying devices from netlink to sysfs. The sysfs writes have been implemented as direct writes, without shelling out. This is faster, but leaves no trace in the logs. Add explicit logs to make debugging possible. Signed-off-by:
Daniel Borkmann <daniel@iogearbox.net> Signed-off-by:
-
Jakub Kicinski authored
Looks like the port adding loop makes a re-appearance on net-next after net was merged back into it (even though it doesn't feature in the merge diff). The ports are already added in nsim_dev_create() so when we try to add them again get EEXIST, and see: netdevsim: probe of netdevsim0 failed with error -17 in the logs. When we remove the loop again the nsim_dev_probe() and nsim_dev_remove() become a wrapper of nsim_dev_create() and nsim_dev_destroy(). Remove this layer of indirection. Fixes: d31e9558 ("Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net") Signed-off-by:
-
David S. Miller authored
Merge tag 'wireless-drivers-next-2019-11-05' of git://git.kernel.org/pub/scm/linux/kernel/git/kvalo/wireless-drivers-next Kalle Valo says: ==================== wireless-drivers-next patches for 5.5 First set of patches for 5.5. The most active driver here clearly is rtw88, lots of patches for it. More quiet on other drivers, smaller fixes and cleanups all over. This pull request also has a trivial conflict, the report and example resolution here: https://lkml.kernel.org/r/20191031111242.50ab1eca@canb.auug.org.au Major changes: rtw88 * add deep power save support * add mac80211 software tx queue (wake_tx_queue) support * enable hardware rate control * add TX-AMSDU support * add NL80211_EXT_FEATURE_CAN_REPLACE_PTK0 support * add power tracking support * add 802.11ac beamformee support * add set_bitrate_mask support * add phy_info debugfs to show Tx/Rx physical status * add RFE type 3 support for 8822b ath10k * add support for hardware rfkill on devices where firmware supports it rtl8xxxu * add bluetooth co-existence support for single antenna iwlwifi * Revamp the debugging infrastructure ==================== Signed-off-by:
-
git://git.open-mesh.org/linux-mergeDavid S. Miller authored
Simon Wunderlich says: ==================== This feature/cleanup patchset includes the following patches: - bump version strings, by Simon Wunderlich - Simplify batadv_v_ogm_aggr_list_free using skb_queue_purge, by Christophe Jaillet - Replace aggr_list_lock with lock free skb handlers, by Christophe Jaillet - explicitly mark fallthrough cases, by Sven Eckelmann - Drop lockdep.h include from soft-interface.c, by Sven Eckelmann ==================== Signed-off-by: -
Kai-Heng Feng authored
ThinkPad Thunderbolt 3 Dock Gen 2 is another docking station that uses RTL8153 based USB ethernet. The device supports macpassthru, but it failed to pass the test of -AD, -BND and -BD. Simply bypass these tests since the device supports this feature just fine. Also the ACPI objects have some differences between Dell's and Lenovo's, so make those ACPI infos no longer hardcoded. BugLink: https://bugs.launchpad.net/bugs/1827961Signed-off-by:
Kai-Heng Feng <kai.heng.feng@canonical.com> Acked-by:
Hayes Wang <hayeswang@realtek.com> Signed-off-by:
-
Vishal Kulkarni authored
This patch implements reset_prepare and reset_done, which are used for handling FLR. Signed-off-by:
Vishal Kulkarni <vishal@chelsio.com> Signed-off-by:
-
David S. Miller authored
Sudarsana Reddy Kalluru says: ==================== bnx2x/cnic: Enable Multi-Cos. The patch series enables Multi-cos feature in the driver. This require the use of new firmware 7.13.15.0. Patch (1) adds driver changes to use new FW. Patches (2) - (3) enables multi-cos functionality in bnx2x driver. Patch (4) adds cnic driver change as required by new FW. ==================== Signed-off-by: -
Manish Rangankar authored
The new FW has added extra validation for HSI version to make FW backward compatible with older VF drivers. Hence set fp_hsi_ver to Fast Path HSI version of the FW in use. Signed-off-by:
Manish Rangankar <mrangankar@marvell.com> Signed-off-by:
Nilesh Javali <njavali@marvell.com> Signed-off-by:
Manish Chopra <manishc@marvell.com> Signed-off-by:
-
Manish Chopra authored
PF driver doesn't enable tx-switching for all cos queues/clients, which causes packets drop from PF to VF. Fix this by enabling tx-switching on all cos queues/clients. Signed-off-by:
-
Sudarsana Reddy Kalluru authored
FW version 7.13.15 addresses the issue in Multi-cos implementation. This patch re-enables the Multi-Cos support in the driver. Fixes: d1f0b5dc ("bnx2x: Disable multi-cos feature.") Signed-off-by:
Sudarsana Reddy Kalluru <skalluru@marvell.com> Signed-off-by:
Ariel Elior <aelior@marvell.com> Signed-off-by:
-
Sudarsana Reddy Kalluru authored
Commit 97a27d6d6e8d "bnx2x: Add FW 7.13.15.0" added said .bin FW to linux-firmware tree. This FW addresses few important issues in the earlier FW release. This patch incorporates FW 7.13.15.0 in the bnx2x driver. Signed-off-by:
-
Andrew Lunn authored
Pass a phy_interface_t to of_get_phy_mode(), by changing the type of phy_mode in the device structure. This then requires that zmii_attach() is also changes, since it takes a pointer to phy_mode. Fixes: 0c65b2b9 ("net: of_get_phy_mode: Change API to solve int/unit warnings") Reported-by:
kbuild test robot <lkp@intel.com> Reported-by:
Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by:
Andrew Lunn <andrew@lunn.ch> Signed-off-by:
-
David S. Miller authored
Eric Dumazet says: ==================== net_sched: convert packet counters to 64bit This small patch series add 64bit support for packet counts. Fact that the counters were still 32bit has been quite painful. tc -s -d qd sh dev eth0 | head -3 qdisc mq 1: root Sent 665706335338 bytes 6526520373 pkt (dropped 2441, overlimits 0 requeues 91) backlog 0b 0p requeues 91 ==================== Signed-off-by: -
Eric Dumazet authored
Now the kernel uses 64bit packet counters in scheduler layer, we want to export these counters to user space. Instead risking breaking user space by adding fields to struct gnet_stats_basic, add a new TCA_STATS_PKT64. Signed-off-by:
-
Eric Dumazet authored
After this change, qdisc packet counter is no longer a 32bit quantity. We still export 32bit values to user. Signed-off-by:
-
Eric Dumazet authored
gnet_stats_basic_packed was really meant to be private kernel structure. If this proves to be a problem, we will have to rename the in-kernel version. Signed-off-by:
-
David S. Miller authored
Andrew Lunn says: ==================== mv88e6xxx ATU occupancy as devlink resource This patchset add generic support to DSA for devlink resources. The Marvell switch Address Translation Unit occupancy is then exported as a resource. In order to do this, the number of ATU entries is added to the per switch info structure. Helpers are added, and then the resource itself is then added. ==================== Signed-off-by: -
Andrew Lunn authored
The ATU can report how many entries it contains. It does this per bin, there being 4 bins in total. Export the ATU as a devlink resource, and provide a method the needed callback to get the resource occupancy. Signed-off-by:
-
Andrew Lunn authored
When retrieving the ATU statistics, and ATU get next has to be performed to trigger the ATU to collect the statistics. Export a helper from global1_atu to perform this. Signed-off-by:
-
Andrew Lunn authored
Add helpers to set/get the ATU statistics register. Signed-off-by:
-
Andrew Lunn authored
For each supported switch, add an entry to the info structure for the number of MACs which can be stored in the ATU. This will later be used to export the ATU as a devlink resource, and indicate its occupancy, how full the ATU is. Signed-off-by:
-
Andrew Lunn authored
Add wrappers around the devlink resource API, so that DSA drivers can register and unregister devlink resources. Signed-off-by:
-
David S. Miller authored
Florian Fainelli says: ==================== net: dsa: bcm_sf2: Add support for optional reset controller line This patch series definest the optional reset controller line for the BCM7445/BCM7278 integrated Ethernet switches and updates the driver to drive that reset line in lieu of the internal watchdog based reset since it does not work on BCM7278. Changes in v2: - make the reset_control_assert() conditional to BCM7278 in the remove function as well ==================== Signed-off-by:
-
