- 19 Oct, 2019 5 commits
-
-
Jakub Kicinski authored
If packet corruption failed we jump to finish_segs and return NET_XMIT_SUCCESS. Seeing success will make the parent qdisc increment its backlog, that's incorrect - we need to return NET_XMIT_DROP. Fixes: 6071bd1a ("netem: Segment GSO packets on enqueue") Signed-off-by:
Jakub Kicinski <jakub.kicinski@netronome.com> Reviewed-by:
Simon Horman <simon.horman@netronome.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Jakub Kicinski authored
To corrupt a GSO frame we first perform segmentation. We then proceed using the first segment instead of the full GSO skb and requeue the rest of the segments as separate packets. If there are any issues with processing the first segment we still want to process the rest, therefore we jump to the finish_segs label. Commit 177b8007 ("net: netem: fix backlog accounting for corrupted GSO frames") started using the pointer to the first segment in the "rest of segments processing", but as mentioned above the first segment may had already been freed at this point. Backlog corrections for parent qdiscs have to be adjusted. Fixes: 177b8007 ("net: netem: fix backlog accounting for corrupted GSO frames") Reported-by:
kbuild test robot <lkp@intel.com> Reported-by:
Dan Carpenter <dan.carpenter@oracle.com> Reported-by:
Ben Hutchings <ben@decadent.org.uk> Signed-off-by:
-
Michael Tretter authored
The tx_clk, rx_clk, and tsu_clk are optional. Currently the macb driver marks clock as not available if it receives an error when trying to get a clock. This is wrong, because a clock controller might return -EPROBE_DEFER if a clock is not available, but will eventually become available. In these cases, the driver would probe successfully but will never be able to adjust the clocks, because the clocks were not available during probe, but became available later. For example, the clock controller for the ZynqMP is implemented in the PMU firmware and the clocks are only available after the firmware driver has been probed. Use devm_clk_get_optional() in instead of devm_clk_get() to get the optional clock and propagate all errors to the calling function. Signed-off-by:
Michael Tretter <m.tretter@pengutronix.de> Acked-by:
Nicolas Ferre <nicolas.ferre@microchip.com> Tested-by:
-
Juergen Gross authored
xenvif_connect_data() calls module_put() in case of error. This is wrong as there is no related module_get(). Remove the superfluous module_put(). Fixes: 279f438e ("xen-netback: Don't destroy the netdev until the vif is shut down") Cc: <stable@vger.kernel.org> # 3.12 Signed-off-by:
Juergen Gross <jgross@suse.com> Reviewed-by:
Paul Durrant <paul@xen.org> Reviewed-by:
Wei Liu <wei.liu@kernel.org> Signed-off-by:
-
Yonglong Liu authored
Currently, the num_msi_left means the vector numbers of NIC, but if the PF supported RoCE, it contains the vector numbers of NIC and RoCE(Not expected). This may cause interrupts lost in some case, because of the NIC module used the vector resources which belongs to RoCE. This patch adds a new variable num_nic_msi to store the vector numbers of NIC, and adjust the default TQP numbers and rss_size according to the value of num_nic_msi. Fixes: 46a3df9f ("net: hns3: Add HNS3 Acceleration Engine & Compatibility Layer Support") Signed-off-by:
Yonglong Liu <liuyonglong@huawei.com> Signed-off-by:
Huazhong Tan <tanhuazhong@huawei.com> Signed-off-by:
-
- 18 Oct, 2019 12 commits
-
-
Andrew Lunn authored
As soon as the netdev is registers, the kernel can start using the interface. If the driver connects the MAC to the PHY after the netdev is registered, there is a race condition where the interface can be opened without having the PHY connected. Change the order to close this race condition. Fixes: 92571a1a ("lan78xx: Connect phy early") Reported-by:
Daniel Wagner <dwagner@suse.de> Signed-off-by:
Andrew Lunn <andrew@lunn.ch> Tested-by:
-
David S. Miller authored
Stefano Garzarella says: ==================== vsock/virtio: make the credit mechanism more robust This series makes the credit mechanism implemented in the virtio-vsock devices more robust. Patch 1 sends an update to the remote peer when the buf_alloc change. Patch 2 prevents a malicious peer (especially the guest) can consume all the memory of the other peer, discarding packets when the credit available is not respected. ==================== Signed-off-by: -
Stefano Garzarella authored
If the remote peer doesn't respect the credit information (buf_alloc, fwd_cnt), sending more data than it can send, we should drop the packets to prevent a malicious peer from using all of our memory. This is patch follows the VIRTIO spec: "VIRTIO_VSOCK_OP_RW data packets MUST only be transmitted when the peer has sufficient free buffer space for the payload" Signed-off-by:
Stefano Garzarella <sgarzare@redhat.com> Signed-off-by:
-
Stefano Garzarella authored
When the user application set a new buffer size value, we should update the remote peer about this change, since it uses this information to calculate the credit available. Signed-off-by:
-
Ido Schimmel authored
devlink maintains packets and bytes statistics for each trap. Since eth_type_trans() was called to set the skb's protocol, the data pointer no longer points to the start of the packet and the bytes accounting is off by 14 bytes. Fix this by pushing the skb's data pointer to the start of the packet. Fixes: b5ce611f ("mlxsw: spectrum: Add devlink-trap support") Reported-by:
Alex Kushnarov <alexanderk@mellanox.com> Tested-by:
Jiri Pirko <jiri@mellanox.com> Signed-off-by:
Ido Schimmel <idosch@mellanox.com> Signed-off-by:
-
Eric Dumazet authored
Thomas found that some forwarded packets would be stuck in FQ packet scheduler because their skb->tstamp contained timestamps far in the future. We thought we addressed this point in commit 8203e2d8 ("net: clear skb->tstamp in forwarding paths") but there is still an issue when/if a packet needs to be fragmented. In order to meet EDT requirements, we have to make sure all fragments get the original skb->tstamp. Note that this original skb->tstamp should be zero in forwarding path, but might have a non zero value in output path if user decided so. Fixes: fb420d5d ("tcp/fq: move back to CLOCK_MONOTONIC") Signed-off-by:
Eric Dumazet <edumazet@google.com> Reported-by:
Thomas Bartschies <Thomas.Bartschies@cvk.de> Signed-off-by:
-
David S. Miller authored
Doug Berger says: ==================== net: bcmgenet: restore internal EPHY support I managed to get my hands on an old BCM97435SVMB board to do some testing with the latest kernel and uncovered a number of things that managed to get broken over the years (some by me ;). This commit set attempts to correct the errors I observed in my testing. The first commit applies to all internal PHYs to restore proper reporting of link status when a link comes up. The second commit restores the soft reset to the initialization of the older internal EPHYs used by 40nm Set-Top Box devices. The third corrects a bug I introduced when removing excessive soft resets by altering the initialization sequence in a way that keeps the GENETv3 MAC interface happy. Finally, I observed a number of issues when manually configuring the network interface of the older EPHYs that appear to be resolved by the fourth commit. ==================== Signed-off-by: -
Doug Berger authored
The EPHY integrated into the 40nm Set-Top Box devices can falsely detect energy when connected to a disabled peer interface. When the peer interface is enabled the EPHY will detect and report the link as active, but on occasion may get into a state where it is not able to exchange data with the connected GENET MAC. This issue has not been observed when the link parameters are auto-negotiated; however, it has been observed with a manually configured link. It has been empirically determined that issuing a soft reset to the EPHY when energy is detected prevents it from getting into this bad state. Fixes: 1c1008c7 ("net: bcmgenet: add main driver file") Signed-off-by:
Doug Berger <opendmb@gmail.com> Acked-by:
Florian Fainelli <f.fainelli@gmail.com> Signed-off-by:
-
Doug Berger authored
It turns out that the "Workaround for putting the PHY in IDDQ mode" used by the internal EPHYs on 40nm Set-Top Box chips when powering down puts the interface to the GENET MAC in a state that can cause subsequent MAC resets to be incomplete. Rather than restore the forced soft reset when powering up internal PHYs, this commit moves the invocation of phy_init_hw earlier in the MAC initialization sequence to just before the MAC reset in the open and resume functions. This allows the interface to be stable and allows the MAC resets to be successful. The bcmgenet_mii_probe() function is split in two to accommodate this. The new function bcmgenet_mii_connect() handles the first half of the functionality before the MAC initialization, and the bcmgenet_mii_config() function is extended to provide the remaining PHY configuration following the MAC initialization. Fixes: 484bfa15 ("Revert "net: bcmgenet: Software reset EPHY after power on"") Signed-off-by:
-
Doug Berger authored
The internal 40nm EPHYs use a "Workaround for putting the PHY in IDDQ mode." These PHYs require a soft reset to restore functionality after they are powered back up. This commit defines the soft_reset function to use genphy_soft_reset during phy_init_hw to accommodate this. Fixes: 6e2d85ec ("net: phy: Stop with excessive soft reset") Signed-off-by:
-
Doug Berger authored
When commit 28b2e0d2 ("net: phy: remove parameter new_link from phy_mac_interrupt()") removed the new_link parameter it set the phydev->link state from the MAC before invoking phy_mac_interrupt(). However, once commit 88d6272a ("net: phy: avoid unneeded MDIO reads in genphy_read_status") was added this initialization prevents the proper determination of the connection parameters by the function genphy_read_status(). This commit removes that initialization to restore the proper functionality. Fixes: 88d6272a ("net: phy: avoid unneeded MDIO reads in genphy_read_status") Signed-off-by:
-
Sean Wang authored
Update maintainers for MediaTek ethernet driver with Mark Lee. He is familiar with MediaTek mt762x series ethernet devices and will keep following maintenance from the vendor side. Signed-off-by:
Sean Wang <sean.wang@mediatek.com> Signed-off-by:
Mark Lee <Mark-MC.Lee@mediatek.com> Signed-off-by:
-
- 17 Oct, 2019 11 commits
-
-
Wei Wang authored
Jesse and Ido reported the following race condition: <CPU A, t0> - Received packet A is forwarded and cached dst entry is taken from the nexthop ('nhc->nhc_rth_input'). Calls skb_dst_set() <t1> - Given Jesse has busy routers ("ingesting full BGP routing tables from multiple ISPs"), route is added / deleted and rt_cache_flush() is called <CPU B, t2> - Received packet B tries to use the same cached dst entry from t0, but rt_cache_valid() is no longer true and it is replaced in rt_cache_route() by the newer one. This calls dst_dev_put() on the original dst entry which assigns the blackhole netdev to 'dst->dev' <CPU A, t3> - dst_input(skb) is called on packet A and it is dropped due to 'dst->dev' being the blackhole netdev There are 2 issues in the v4 routing code: 1. A per-netns counter is used to do the validation of the route. That means whenever a route is changed in the netns, users of all routes in the netns needs to redo lookup. v6 has an implementation of only updating fn_sernum for routes that are affected. 2. When rt_cache_valid() returns false, rt_cache_route() is called to throw away the current cache, and create a new one. This seems unnecessary because as long as this route does not change, the route cache does not need to be recreated. To fully solve the above 2 issues, it probably needs quite some code changes and requires careful testing, and does not suite for net branch. So this patch only tries to add the deleted cached rt into the uncached list, so user could still be able to use it to receive packets until it's done. Fixes: 95c47f9c ("ipv4: call dst_dev_put() properly") Signed-off-by:Wei Wang <weiwan@google.com> Reported-by:
Ido Schimmel <idosch@idosch.org> Reported-by:
Jesse Hathaway <jesse@mbuki-mvuki.org> Tested-by:
Martin KaFai Lau <kafai@fb.com> Cc: David Ahern <dsahern@gmail.com> Reviewed-by:
-
Stefano Brivio authored
...instead of -EINVAL. An issue was found with older kernel versions while unplugging a NFS client with pending RPCs, and the wrong error code here prevented it from recovering once link is back up with a configured address. Incidentally, this is not an issue anymore since commit 4f8943f8 ("SUNRPC: Replace direct task wakeups from softirq context"), included in 5.2-rc7, had the effect of decoupling the forwarding of this error by using SO_ERROR in xs_wake_error(), as pointed out by Benjamin Coddington. To the best of my knowledge, this isn't currently causing any further issue, but the error code doesn't look appropriate anyway, and we might hit this in other paths as well. In detail, as analysed by Gonzalo Siero, once the route is deleted because the interface is down, and can't be resolved and we return -EINVAL here, this ends up, courtesy of inet_sk_rebuild_header(), as the socket error seen by tcp_write_err(), called by tcp_retransmit_timer(). In turn, tcp_write_err() indirectly calls xs_error_report(), which wakes up the RPC pending tasks with a status of -EINVAL. This is then seen by call_status() in the SUN RPC implementation, which aborts the RPC call calling rpc_exit(), instead of handling this as a potentially temporary condition, i.e. as a timeout. Return -EINVAL only if the input parameters passed to ip_route_output_key_hash_rcu() are actually invalid (this is the case if the specified source address is multicast, limited broadcast or all zeroes), but return -ENETUNREACH in all cases where, at the given moment, the given source address doesn't allow resolving the route. While at it, drop the initialisation of err to -ENETUNREACH, which was added to __ip_route_output_key() back then by commit 0315e382 ("net: Fix behaviour of unreachable, blackhole and prohibit routes"), but actually had no effect, as it was, and is, overwritten by the fib_lookup() return code assignment, and anyway ignored in all other branches, including the if (fl4->saddr) one: I find this rather confusing, as it would look like -ENETUNREACH is the "default" error, while that statement has no effect. Also note that after commit fc75fc83 ("ipv4: dont create routes on down devices"), we would get -ENETUNREACH if the device is down, but -EINVAL if the source address is specified and we can't resolve the route, and this appears to be rather inconsistent. Reported-by:
Stefan Walter <walteste@inf.ethz.ch> Analysed-by:
Benjamin Coddington <bcodding@redhat.com> Analysed-by:
Gonzalo Siero <gsierohu@redhat.com> Signed-off-by:
Stefano Brivio <sbrivio@redhat.com> Signed-off-by:
-
Marek Vasut authored
The KSZ8795 PHY ID is in fact used by KSZ8794/KSZ8795/KSZ8765 switches. Update the PHY ID and name to reflect that, as this family of switches is commonly refered to as KSZ87xx Signed-off-by:
Marek Vasut <marex@denx.de> Cc: Andrew Lunn <andrew@lunn.ch> Cc: David S. Miller <davem@davemloft.net> Cc: Florian Fainelli <f.fainelli@gmail.com> Cc: George McCollister <george.mccollister@gmail.com> Cc: Heiner Kallweit <hkallweit1@gmail.com> Cc: Sean Nyekjaer <sean.nyekjaer@prevas.dk> Cc: Tristram Ha <Tristram.Ha@microchip.com> Cc: Woojung Huh <woojung.huh@microchip.com> Signed-off-by:
-
Marek Vasut authored
The KSZ8051 PHY and the KSZ8794/KSZ8795/KSZ8765 switch share exactly the same PHY ID. Since KSZ8051 is higher in the ksphy_driver[] list of PHYs in the micrel PHY driver, it is used even with the KSZ87xx switch. This is wrong, since the KSZ8051 configures registers of the PHY which are not present on the simplified KSZ87xx switch PHYs and misconfigures other registers of the KSZ87xx switch PHYs. Fortunatelly, it is possible to tell apart the KSZ8051 PHY from the KSZ87xx switch by checking the Basic Status register Bit 0, which is read-only and indicates presence of the Extended Capability Registers. The KSZ8051 PHY has those registers while the KSZ87xx switch does not. This patch implements simple check for the presence of this bit for both the KSZ8051 PHY and KSZ87xx switch, to let both use the correct PHY driver instance. Fixes: 9d162ed6 ("net: phy: micrel: add support for KSZ8795") Signed-off-by:
-
Marek Vasut authored
The KSZ driver uses one regmap per register width (8/16/32), each with it's own lock, but accessing the same set of registers. In theory, it is possible to create a race condition between these regmaps, although the underlying bus (SPI or I2C) locking should assure nothing bad will really happen and the accesses would be correct. To make the driver do the right thing, add one single shared mutex for all the regmaps used by the driver instead. This assures that even if some future hardware is on a bus which does not serialize the accesses the same way SPI or I2C does, nothing bad will happen. Note that the status_mutex was unused and only initied, hence it was renamed and repurposed as the regmap mutex. Signed-off-by:
-
Marek Vasut authored
The KSZ87xx driver calls mutex_init() on mutexes already inited in ksz_common.c ksz_switch_register(). Do not do it twice, drop the reinitialization. Signed-off-by:
-
Ben Dooks (Codethink) authored
The stmmac_pcs_ctrl_ane() expects a register address as argument 1, but for some reason the mac_device_info is being passed. Fix the warning (and possible bug) from sparse: drivers/net/ethernet/stmicro/stmmac/stmmac_main.c:2613:17: warning: incorrect type in argument 1 (different address spaces) drivers/net/ethernet/stmicro/stmmac/stmmac_main.c:2613:17: expected void [noderef] <asn:2> *ioaddr drivers/net/ethernet/stmicro/stmmac/stmmac_main.c:2613:17: got struct mac_device_info *hw Signed-off-by:
Ben Dooks <ben.dooks@codethink.co.uk> Signed-off-by:
-
David S. Miller authored
Ioana Ciornei says: ==================== dpaa2-eth: misc fixes This patch set adds a couple of fixes around updating configuration on MAC change. Depending on when MC connects the DPNI to a MAC, both the MAC address and TX FQIDs should be updated everytime there is a change in configuration. Changes in v2: - used reverse christmas tree ordering in patch 2/2 Changes in v3: - add a missing new line - go back to FQ based enqueueing after a transient error ==================== Signed-off-by: -
Ioana Radulescu authored
Depending on when MC connects the DPNI to a MAC, Tx FQIDs may not be available during probe time. Read the FQIDs each time the link goes up to avoid using invalid values. In case an error occurs or an invalid value is retrieved, fall back to QDID-based enqueueing. Fixes: 1fa0f68c ("dpaa2-eth: Use FQ-based DPIO enqueue API") Signed-off-by:
Ioana Radulescu <ruxandra.radulescu@nxp.com> Signed-off-by:
Ioana Ciornei <ioana.ciornei@nxp.com> Signed-off-by:
-
Florin Chiculita authored
Add IRQ for the DPNI endpoint change event, resolving the issue when a dynamically created DPNI gets a randomly generated hw address when the endpoint is a DPMAC object. Signed-off-by:
Florin Chiculita <florinlaurentiu.chiculita@nxp.com> Signed-off-by:
-
Oliver Neukum authored
The serial state information must not be embedded into another data structure, as this interferes with cache handling for DMA on architectures without cache coherence.. That would result in data corruption on some architectures Allocating it separately. v2: fix syntax error Signed-off-by:
Oliver Neukum <oneukum@suse.com> Signed-off-by:
-
- 16 Oct, 2019 12 commits
-
-
Biao Huang authored
disable ptp_ref_clk in suspend flow, and enable it in resume flow. Fixes: f573c0b9 ("stmmac: move stmmac_clk, pclk, clk_ptp_ref and stmmac_rst to platform structure") Signed-off-by:
Biao Huang <biao.huang@mediatek.com> Signed-off-by:
-
Yonglong Liu authored
Some drivers just call phy_ethtool_ksettings_set() to set the links, for those phy drivers that use genphy_read_status(), if autoneg is on, and the link is up, than execute "ethtool -s ethx autoneg on" will cause "link partner" information disappear. The call trace is phy_ethtool_ksettings_set()->phy_start_aneg() ->linkmode_zero(phydev->lp_advertising)->genphy_read_status(), the link didn't change, so genphy_read_status() just return, and phydev->lp_advertising is zero now. This patch moves the clear operation of lp_advertising from phy_start_aneg() to genphy_read_lpa()/genphy_c45_read_lpa(), and if autoneg on and autoneg not complete, just clear what the generic functions care about. Fixes: 88d6272a ("net: phy: avoid unneeded MDIO reads in genphy_read_status") Signed-off-by:
Heiner Kallweit <hkallweit1@gmail.com> Signed-off-by:
-
Eric Dumazet authored
We need to extend the rcu_read_lock() section in rxrpc_error_report() and use rcu_dereference_sk_user_data() instead of plain access to sk->sk_user_data to make sure all rules are respected. The compiler wont reload sk->sk_user_data at will, and RCU rules prevent memory beeing freed too soon. Fixes: f0308fb0 ("rxrpc: Fix possible NULL pointer access in ICMP handling") Fixes: 17926a79 ("[AF_RXRPC]: Provide secure RxRPC sockets for use by userspace and kernel both") Signed-off-by:
-
Mahesh Bandewar authored
This reverts commit b0818f80. Started seeing weird behavior after this patch especially in the IPv6 code path. Haven't root caused it, but since this was applied to net branch, taking a precautionary measure to revert it and look / analyze those failures Revert this now and I'll send a better fix after analysing / fixing the weirdness observed. CC: Eric Dumazet <edumazet@google.com> CC: Wei Wang <weiwan@google.com> CC: David S. Miller <davem@davemloft.net> Signed-off-by:
Mahesh Bandewar <maheshb@google.com> Signed-off-by:
-
Valentin Vidic authored
Make sure res does not contain random value if the call to sr_read_cmd fails for some reason. Reported-by: syzbot+f1842130bbcfb335bac1@syzkaller.appspotmail.com Signed-off-by:
Valentin Vidic <vvidic@valentin-vidic.from.hr> Signed-off-by:
-
Florian Fainelli authored
The RGMII_MODE_EN bit value was 0 for GENET versions 1 through 3, and became 6 for GENET v4 and above, account for that difference. Fixes: aa09677c ("net: bcmgenet: add MDIO routines") Signed-off-by:
-
Ben Dooks (Codethink) authored
The tc_flow_parsers is not used outside of the driver, so make it static to avoid the following sparse warning: drivers/net/ethernet/stmicro/stmmac/stmmac_tc.c:516:3: warning: symbol 'tc_flow_parsers' was not declared. Should it be static? Signed-off-by:
-
Ben Dooks (Codethink) authored
The cpdma_chan_split_pool() function is not used outside of the driver, so make it static to avoid the following sparse warning: drivers/net/ethernet/ti/davinci_cpdma.c:725:5: warning: symbol 'cpdma_chan_split_pool' was not declared. Should it be static? Signed-off-by:
-
Thomas Bogendoerfer authored
Commit 7f683b92 ("i825xx: switch to switch to dma_alloc_attrs") switched dma allocation over to dma_alloc_attr, but didn't convert the SNI part to request consistent DMA memory. This broke sni_82596 since driver doesn't do dma_cache_sync for performance reasons. Fix this by using different DMA_ATTRs for lasi_82596 and sni_82596. Fixes: 7f683b92 ("i825xx: switch to switch to dma_alloc_attrs") Signed-off-by:
Thomas Bogendoerfer <tbogendoerfer@suse.de> Signed-off-by:
-
Xin Long authored
syzbot reported a memory leak: BUG: memory leak, unreferenced object 0xffff888120b3d380 (size 64): backtrace: [...] slab_alloc mm/slab.c:3319 [inline] [...] kmem_cache_alloc+0x13f/0x2c0 mm/slab.c:3483 [...] sctp_bucket_create net/sctp/socket.c:8523 [inline] [...] sctp_get_port_local+0x189/0x5a0 net/sctp/socket.c:8270 [...] sctp_do_bind+0xcc/0x200 net/sctp/socket.c:402 [...] sctp_bindx_add+0x4b/0xd0 net/sctp/socket.c:497 [...] sctp_setsockopt_bindx+0x156/0x1b0 net/sctp/socket.c:1022 [...] sctp_setsockopt net/sctp/socket.c:4641 [inline] [...] sctp_setsockopt+0xaea/0x2dc0 net/sctp/socket.c:4611 [...] sock_common_setsockopt+0x38/0x50 net/core/sock.c:3147 [...] __sys_setsockopt+0x10f/0x220 net/socket.c:2084 [...] __do_sys_setsockopt net/socket.c:2100 [inline] It was caused by when sending msgs without binding a port, in the path: inet_sendmsg() -> inet_send_prepare() -> inet_autobind() -> .get_port/sctp_get_port(), sp->bind_hash will be set while bp->port is not. Later when binding another port by sctp_setsockopt_bindx(), a new bucket will be created as bp->port is not set. sctp's autobind is supposed to call sctp_autobind() where it does all things including setting bp->port. Since sctp_autobind() is called in sctp_sendmsg() if the sk is not yet bound, it should have skipped the auto bind. THis patch is to avoid calling inet_autobind() in inet_send_prepare() by changing sctp_prot .no_autobind with true, also remove the unused .get_port. Reported-by: syzbot+d44f7bbebdea49dbc84a@syzkaller.appspotmail.com Signed-off-by:Xin Long <lucien.xin@gmail.com> Acked-by:
Marcelo Ricardo Leitner <marcelo.leitner@gmail.com> Signed-off-by:
-
Vinicius Costa Gomes authored
When a application sends many packets with the same txtime, they may be transmitted out of order (different from the order in which they were enqueued). This happens because when inserting elements into the tree, when the txtime of two packets are the same, the new packet is inserted at the left side of the tree, causing the reordering. The only effect of this change should be that packets with the same txtime will be transmitted in the order they are enqueued. The application in question (the AVTP GStreamer plugin, still in development) is sending video traffic, in which each video frame have a single presentation time, the problem is that when packetizing, multiple packets end up with the same txtime. The receiving side was rejecting packets because they were being received out of order. Fixes: 25db26a9 ("net/sched: Introduce the ETF Qdisc") Reported-by:
Ederson de Souza <ederson.desouza@intel.com> Signed-off-by:
Vinicius Costa Gomes <vinicius.gomes@intel.com> Signed-off-by:
-
Eric Dumazet authored
tc_ctl_action() has the ability to loop forever if tcf_action_add() returns -EAGAIN. This special case has been done in case a module needed to be loaded, but it turns out that tcf_add_notify() could also return -EAGAIN if the socket sk_rcvbuf limit is hit. We need to separate the two cases, and only loop for the module loading case. While we are at it, add a limit of 10 attempts since unbounded loops are always scary. syzbot repro was something like : socket(PF_NETLINK, SOCK_RAW|SOCK_NONBLOCK, NETLINK_ROUTE) = 3 write(3, ..., 38) = 38 setsockopt(3, SOL_SOCKET, SO_RCVBUF, [0], 4) = 0 sendmsg(3, {msg_name(0)=NULL, msg_iov(1)=[{..., 388}], msg_controllen=0, msg_flags=0x10}, ...) NMI backtrace for cpu 0 CPU: 0 PID: 1054 Comm: khungtaskd Not tainted 5.4.0-rc1+ #0 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 nmi_cpu_backtrace.cold+0x70/0xb2 lib/nmi_backtrace.c:101 nmi_trigger_cpumask_backtrace+0x23b/0x28b lib/nmi_backtrace.c:62 arch_trigger_cpumask_backtrace+0x14/0x20 arch/x86/kernel/apic/hw_nmi.c:38 trigger_all_cpu_backtrace include/linux/nmi.h:146 [inline] check_hung_uninterruptible_tasks kernel/hung_task.c:205 [inline] watchdog+0x9d0/0xef0 kernel/hung_task.c:289 kthread+0x361/0x430 kernel/kthread.c:255 ret_from_fork+0x24/0x30 arch/x86/entry/entry_64.S:352 Sending NMI from CPU 0 to CPUs 1: NMI backtrace for cpu 1 CPU: 1 PID: 8859 Comm: syz-executor910 Not tainted 5.4.0-rc1+ #0 Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011 RIP: 0010:arch_local_save_flags arch/x86/include/asm/paravirt.h:751 [inline] RIP: 0010:lockdep_hardirqs_off+0x1df/0x2e0 kernel/locking/lockdep.c:3453 Code: 5c 08 00 00 5b 41 5c 41 5d 5d c3 48 c7 c0 58 1d f3 88 48 ba 00 00 00 00 00 fc ff df 48 c1 e8 03 80 3c 10 00 0f 85 d3 00 00 00 <48> 83 3d 21 9e 99 07 00 0f 84 b9 00 00 00 9c 58 0f 1f 44 00 00 f6 RSP: 0018:ffff8880a6f3f1b8 EFLAGS: 00000046 RAX: 1ffffffff11e63ab RBX: ffff88808c9c6080 RCX: 0000000000000000 RDX: dffffc0000000000 RSI: 0000000000000000 RDI: ffff88808c9c6914 RBP: ffff8880a6f3f1d0 R08: ffff88808c9c6080 R09: fffffbfff16be5d1 R10: fffffbfff16be5d0 R11: 0000000000000003 R12: ffffffff8746591f R13: ffff88808c9c6080 R14: ffffffff8746591f R15: 0000000000000003 FS: 00000000011e4880(0000) GS:ffff8880ae900000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: ffffffffff600400 CR3: 00000000a8920000 CR4: 00000000001406e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 Call Trace: trace_hardirqs_off+0x62/0x240 kernel/trace/trace_preemptirq.c:45 __raw_spin_lock_irqsave include/linux/spinlock_api_smp.h:108 [inline] _raw_spin_lock_irqsave+0x6f/0xcd kernel/locking/spinlock.c:159 __wake_up_common_lock+0xc8/0x150 kernel/sched/wait.c:122 __wake_up+0xe/0x10 kernel/sched/wait.c:142 netlink_unlock_table net/netlink/af_netlink.c:466 [inline] netlink_unlock_table net/netlink/af_netlink.c:463 [inline] netlink_broadcast_filtered+0x705/0xb80 net/netlink/af_netlink.c:1514 netlink_broadcast+0x3a/0x50 net/netlink/af_netlink.c:1534 rtnetlink_send+0xdd/0x110 net/core/rtnetlink.c:714 tcf_add_notify net/sched/act_api.c:1343 [inline] tcf_action_add+0x243/0x370 net/sched/act_api.c:1362 tc_ctl_action+0x3b5/0x4bc net/sched/act_api.c:1410 rtnetlink_rcv_msg+0x463/0xb00 net/core/rtnetlink.c:5386 netlink_rcv_skb+0x177/0x450 net/netlink/af_netlink.c:2477 rtnetlink_rcv+0x1d/0x30 net/core/rtnetlink.c:5404 netlink_unicast_kernel net/netlink/af_netlink.c:1302 [inline] netlink_unicast+0x531/0x710 net/netlink/af_netlink.c:1328 netlink_sendmsg+0x8a5/0xd60 net/netlink/af_netlink.c:1917 sock_sendmsg_nosec net/socket.c:637 [inline] sock_sendmsg+0xd7/0x130 net/socket.c:657 ___sys_sendmsg+0x803/0x920 net/socket.c:2311 __sys_sendmsg+0x105/0x1d0 net/socket.c:2356 __do_sys_sendmsg net/socket.c:2365 [inline] __se_sys_sendmsg net/socket.c:2363 [inline] __x64_sys_sendmsg+0x78/0xb0 net/socket.c:2363 do_syscall_64+0xfa/0x760 arch/x86/entry/common.c:290 entry_SYSCALL_64_after_hwframe+0x49/0xbe RIP: 0033:0x440939 Fixes: 1da177e4 ("Linux-2.6.12-rc2") Signed-off-by:
-
