- Jan 27, 2022
-
-
Florian Westphal authored
Add a specific test for the reload issue fixed with commit 23c54263 ("netfilter: nft_set_pipapo: allocate pcpu scratch maps on clone"). Add to set, then flush set content + restore without other add/remove in the transaction. On kernels before the fix, this test case fails: net,mac with reload [FAIL] Signed-off-by:
Florian Westphal <fw@strlen.de> Reviewed-by:
Stefano Brivio <sbrivio@redhat.com> Signed-off-by:
Pablo Neira Ayuso <pablo@netfilter.org>
-
Pablo Neira Ayuso authored
Cancel tracking for byteorder operation, otherwise selector + byteorder operation is incorrectly reduced if source and destination registers are the same. Reported-by:
kernel test robot <oliver.sang@intel.com> Signed-off-by:
Pablo Neira Ayuso <pablo@netfilter.org>
-
Phil Sutter authored
Prior to commit fa538f7c ("netfilter: nf_reject: add reject skbuff creation helpers"), nft_reject_bridge did not assign to nskb->dev before passing nskb on to br_forward(). The shared skbuff creation helpers introduced in above commit do which seems to confuse br_forward() as reject statements in prerouting hook won't emit a packet anymore. Fix this by simply passing NULL instead of 'dev' to the helpers - they use the pointer for just that assignment, nothing else. Fixes: fa538f7c ("netfilter: nf_reject: add reject skbuff creation helpers") Signed-off-by:
Phil Sutter <phil@nwl.cc> Signed-off-by:
Pablo Neira Ayuso <pablo@netfilter.org>
-
Florian Westphal authored
Add a test that sends large udp packet (which is fragmented) via a stateless nft nat rule, i.e. 'ip saddr set 10.2.3.4' and check that the datagram is received by peer. On kernels without commit 4e1860a3 ("netfilter: nft_payload: do not update layer 4 checksum when mangling fragments")', this will fail with: cmp: EOF on /tmp/tmp.V1q0iXJyQF which is empty -rw------- 1 root root 4096 Jan 24 22:03 /tmp/tmp.Aaqnq4rBKS -rw------- 1 root root 0 Jan 24 22:03 /tmp/tmp.V1q0iXJyQF ERROR: in and output file mismatch when checking udp with stateless nat FAIL: nftables v1.0.0 (Fearless Fosdick #2) On patched kernels, this will show: PASS: IP statless for ns2-PFp89amx Signed-off-by:
Florian Westphal <fw@strlen.de> Signed-off-by:
Pablo Neira Ayuso <pablo@netfilter.org>
-
Florian Westphal authored
This selftests needs almost 3 minutes to complete, reduce the insertes zones to 1000. Test now completes in about 20 seconds. Signed-off-by:
Florian Westphal <fw@strlen.de> Signed-off-by:
Pablo Neira Ayuso <pablo@netfilter.org>
-
Florian Westphal authored
The conversion erroneously removed the refcount increment. In case we can use the percpu template, we need to increment the refcount, else it will be released when the skb gets freed. In case the slowpath is taken, the new template already has a refcount of 1. Fixes: 71977437 ("netfilter: conntrack: convert to refcount_t api") Reported-by:
kernel test robot <oliver.sang@intel.com> Signed-off-by:
Florian Westphal <fw@strlen.de> Signed-off-by:
Pablo Neira Ayuso <pablo@netfilter.org>
-
Geert Uytterhoeven authored
NF_FLOW_TABLE_IPV4 and NF_FLOW_TABLE_IPV6 are invisble, selected by nothing (so they can no longer be enabled), and their last real users have been removed (nf_flow_table_ipv6.c is empty). Clean up the leftovers. Fixes: c42ba429 ("netfilter: flowtable: remove ipv4/ipv6 modules") Signed-off-by:
Geert Uytterhoeven <geert@linux-m68k.org> Signed-off-by:
Pablo Neira Ayuso <pablo@netfilter.org>
-
- Jan 25, 2022
-
-
Yufeng Mo authored
Since some interrupt states may be cleared by hardware, the driver may receive an empty interrupt. Currently, the VF driver directly disables the vector0 interrupt in this case. As a result, the VF is unavailable. Therefore, the vector0 interrupt should be enabled in this case. Fixes: b90fcc5b ("net: hns3: add reset handling for VF when doing Core/Global/IMP reset") Signed-off-by:
Yufeng Mo <moyufeng@huawei.com> Signed-off-by:
Guangbin Huang <huangguangbin2@huawei.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Jakub Kicinski authored
Recent changes made netdev->dev_addr const, and it's passed directly to mpc52xx_fec_set_paddr(). Similar problem exists on the probe patch, the driver needs to call eth_hw_addr_set(). Reported-by:
Geert Uytterhoeven <geert@linux-m68k.org> Fixes: adeef3e3 ("net: constify netdev->dev_addr") Signed-off-by:
Jakub Kicinski <kuba@kernel.org> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Toke Høiland-Jørgensen authored
The cpsw driver didn't properly initialise the struct page_pool_params before calling page_pool_create(), which leads to crashes after the struct has been expanded with new parameters. The second Fixes tag below is where the buggy code was introduced, but because the code was moved around this patch will only apply on top of the commit in the first Fixes tag. Fixes: c5013ac1 ("net: ethernet: ti: cpsw: move set of common functions in cpsw_priv") Fixes: 9ed4050c ("net: ethernet: ti: cpsw: add XDP support") Reported-by:
Colin Foster <colin.foster@in-advantage.com> Signed-off-by:
Toke Høiland-Jørgensen <toke@redhat.com> Tested-by:
Colin Foster <colin.foster@in-advantage.com> Acked-by:
Jesper Dangaard Brouer <brouer@redhat.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Hangyu Hua authored
ym needs to be free when ym->cmd != SIOCYAMSMCS. Fixes: 0781168e ("yam: fix a missing-check bug") Signed-off-by:
Hangyu Hua <hbh25y@gmail.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Jisheng Zhang authored
Currently, on EEE capable platforms, if EEE SW timer is used, the SW timer cause 1 wakeup/s even if the TX has successfully entered EEE. Remove this unnecessary wakeup by only calling mod_timer() if we haven't successfully entered EEE. Signed-off-by:
Jisheng Zhang <jszhang@kernel.org> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Jakub Kicinski authored
Merge tag 'linux-can-fixes-for-5.17-20220124' of git://git.kernel.org/pub/scm/linux/kernel/git/mkl/linux-can Marc Kleine-Budde says: ==================== pull-request: can 2022-01-24 The first patch updates the email address of Brian Silverman from his former employer to his private address. The next patch fixes DT bindings information for the tcan4x5x SPI CAN driver. The following patch targets the m_can driver and fixes the introduction of FIFO bulk read support. Another patch for the tcan4x5x driver, which fixes the max register value for the regmap config. The last patch for the flexcan driver marks the RX mailbox support for the MCF5441X as support. * tag 'linux-can-fixes-for-5.17-20220124' of git://git.kernel.org/pub/scm/linux/kernel/git/mkl/linux-can: can: flexcan: mark RX via mailboxes as supported on MCF5441X can: tcan4x5x: regmap: fix max register value can: m_can: m_can_fifo_{read,write}: don't read or write from/to FIFO if length is 0 dt-bindings: can: tcan4x5x: fix mram-cfg RX FIFO config mailmap: update email address of Brian Silverman ==================== Link: https://lore.kernel.org/r/20220124175955.3464134-1-mkl@pengutronix.de Signed-off-by:
Jakub Kicinski <kuba@kernel.org>
-
Marc Kleine-Budde authored
Most flexcan IP cores support 2 RX modes: - FIFO - mailbox The flexcan IP core on the MCF5441X cannot receive CAN RTR messages via mailboxes. However the mailbox mode is more performant. The commit | 1c45f577 ("can: flexcan: add ethtool support to change rx-rtr setting during runtime") added support to switch from FIFO to mailbox mode on these cores. After testing the mailbox mode on the MCF5441X by Angelo Dureghello, this patch marks it (without RTR capability) as supported. Further the IP core overview table is updated, that RTR reception via mailboxes is not supported. Link: https://lore.kernel.org/all/20220121084425.3141218-1-mkl@pengutronix.de Tested-by:
Angelo Dureghello <angelo@kernel-space.org> Signed-off-by:
Marc Kleine-Budde <mkl@pengutronix.de>
-
Marc Kleine-Budde authored
The MRAM of the tcan4x5x has a size of 2K and starts at 0x8000. There are no further registers in the tcan4x5x making 0x87fc the biggest addressable register. This patch fixes the max register value of the regmap config from 0x8ffc to 0x87fc. Fixes: 6e1caaf8 ("can: tcan4x5x: fix max register value") Link: https://lore.kernel.org/all/20220119064011.2943292-1-mkl@pengutronix.de Signed-off-by:
Marc Kleine-Budde <mkl@pengutronix.de>
-
Marc Kleine-Budde authored
In order to optimize FIFO access, especially on m_can cores attached to slow busses like SPI, in patch | e3938177 ("can: m_can: Disable IRQs on FIFO bus errors") bulk read/write support has been added to the m_can_fifo_{read,write} functions. That change leads to the tcan driver to call regmap_bulk_{read,write}() with a length of 0 (for CAN frames with 0 data length). regmap treats this as an error: | tcan4x5x spi1.0 tcan4x5x0: FIFO write returned -22 This patch fixes the problem by not calling the cdev->ops->{read,write)_fifo() in case of a 0 length read/write. Fixes: e3938177 ("can: m_can: Disable IRQs on FIFO bus errors") Link: https://lore.kernel.org/all/20220114155751.2651888-1-mkl@pengutronix.de Cc: stable@vger.kernel.org Cc: Matt Kline <matt@bitbashing.io> Cc: Chandrasekar Ramakrishnan <rcsekar@samsung.com> Reported-by:
Michael Anochin <anochin@photo-meter.com> Signed-off-by:
Marc Kleine-Budde <mkl@pengutronix.de>
-
Marc Kleine-Budde authored
This tcan4x5x only comes with 2K of MRAM, a RX FIFO with a dept of 32 doesn't fit into the MRAM. Use a depth of 16 instead. Fixes: 4edd396a ("dt-bindings: can: tcan4x5x: Add DT bindings for TCAN4x5X driver") Link: https://lore.kernel.org/all/20220119062951.2939851-1-mkl@pengutronix.de Signed-off-by:
Marc Kleine-Budde <mkl@pengutronix.de>
-
Marc Kleine-Budde authored
Brian Silverman's address at bluerivertech.com is not valid anymore, use Brian's private email address instead. Link: https://lore.kernel.org/all/20220110082359.2019735-1-mkl@pengutronix.de Cc: Brian Silverman <bsilver16384@gmail.com> Signed-off-by:
Marc Kleine-Budde <mkl@pengutronix.de>
-
- Jan 24, 2022
-
-
Jisheng Zhang authored
The tx_coalesce and mii_irq are not used at all now, so remove them. Signed-off-by:
Jisheng Zhang <jszhang@kernel.org> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Christophe JAILLET authored
Simplify code by using bitmap_weight() and bitmap_zero() instead of hand-writing these functions. Signed-off-by:
Christophe JAILLET <christophe.jaillet@wanadoo.fr> Reviewed-by:
Igor Russkikh <irusskikh@marvell.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Xin Long authored
When 'ping' changes to use PING socket instead of RAW socket by: # sysctl -w net.ipv4.ping_group_range="0 100" the selftests 'router_broadcast.sh' will fail, as such command # ip vrf exec vrf-h1 ping -I veth0 198.51.100.255 -b can't receive the response skb by the PING socket. It's caused by mismatch of sk_bound_dev_if and dif in ping_rcv() when looking up the PING socket, as dif is vrf-h1 if dif's master was set to vrf-h1. This patch is to fix this regression by also checking the sk_bound_dev_if against sdif so that the packets can stil be received even if the socket is not bound to the vrf device but to the real iif. Fixes: c319b4d7 ("net: ipv4: add IPPROTO_ICMP socket kind") Reported-by:
Hangbin Liu <liuhangbin@gmail.com> Signed-off-by:
Xin Long <lucien.xin@gmail.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Wen Gu authored
We encountered a crash in smc_setsockopt() and it is caused by accessing smc->clcsock after clcsock was released. BUG: kernel NULL pointer dereference, address: 0000000000000020 #PF: supervisor read access in kernel mode #PF: error_code(0x0000) - not-present page PGD 0 P4D 0 Oops: 0000 [#1] PREEMPT SMP PTI CPU: 1 PID: 50309 Comm: nginx Kdump: loaded Tainted: G E 5.16.0-rc4+ #53 RIP: 0010:smc_setsockopt+0x59/0x280 [smc] Call Trace: <TASK> __sys_setsockopt+0xfc/0x190 __x64_sys_setsockopt+0x20/0x30 do_syscall_64+0x34/0x90 entry_SYSCALL_64_after_hwframe+0x44/0xae RIP: 0033:0x7f16ba83918e </TASK> This patch tries to fix it by holding clcsock_release_lock and checking whether clcsock has already been released before access. In case that a crash of the same reason happens in smc_getsockopt() or smc_switch_to_fallback(), this patch also checkes smc->clcsock in them too. And the caller of smc_switch_to_fallback() will identify whether fallback succeeds according to the return value. Fixes: fd57770d ("net/smc: wait for pending work before clcsock release_sock") Link: https://lore.kernel.org/lkml/5dd7ffd1-28e2-24cc-9442-1defec27375e@linux.ibm.com/T/ Signed-off-by:
Wen Gu <guwen@linux.alibaba.com> Acked-by:
Karsten Graul <kgraul@linux.ibm.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Sukadev Bhattiprolu authored
With previous bug fix, ->wait_capability flag is no longer needed and can be removed. Fixes: 249168ad ("ibmvnic: Make CRQ interrupt tasklet wait for all capabilities crqs") Signed-off-by:
Sukadev Bhattiprolu <sukadev@linux.ibm.com> Reviewed-by:
Dany Madden <drt@linux.ibm.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Sukadev Bhattiprolu authored
ibmvnic_tasklet() continuously spins waiting for responses to all capability requests. It does this to avoid encountering an error during initialization of the vnic. However if there is a bug in the VIOS and we do not receive a response to one or more queries the tasklet ends up spinning continuously leading to hard lock ups. If we fail to receive a message from the VIOS it is reasonable to timeout the login attempt rather than spin indefinitely in the tasklet. Fixes: 249168ad ("ibmvnic: Make CRQ interrupt tasklet wait for all capabilities crqs") Signed-off-by:
Sukadev Bhattiprolu <sukadev@linux.ibm.com> Reviewed-by:
Dany Madden <drt@linux.ibm.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Sukadev Bhattiprolu authored
We use ->running_cap_crqs to determine when the ibmvnic_tasklet() should send out the next protocol message type. i.e when we get back responses to all our QUERY_CAPABILITY CRQs we send out REQUEST_CAPABILITY crqs. Similiary, when we get responses to all the REQUEST_CAPABILITY crqs, we send out the QUERY_IP_OFFLOAD CRQ. We currently increment ->running_cap_crqs as we send out each CRQ and have the ibmvnic_tasklet() send out the next message type, when this running_cap_crqs count drops to 0. This assumes that all the CRQs of the current type were sent out before the count drops to 0. However it is possible that we send out say 6 CRQs, get preempted and receive all the 6 responses before we send out the remaining CRQs. This can result in ->running_cap_crqs count dropping to zero before all messages of the current type were sent and we end up sending the next protocol message too early. Instead initialize the ->running_cap_crqs upfront so the tasklet will only send the next protocol message after all responses are received. Use the cap_reqs local variable to also detect any discrepancy (either now or in future) in the number of capability requests we actually send. Currently only send_query_cap() is affected by this behavior (of sending next message early) since it is called from the worker thread (during reset) and from application thread (during ->ndo_open()) and they can be preempted. send_request_cap() is only called from the tasklet which processes CRQ responses sequentially, is not be affected. But to maintain the existing symmtery with send_query_capability() we update send_request_capability() also. Fixes: 249168ad ("ibmvnic: Make CRQ interrupt tasklet wait for all capabilities crqs") Signed-off-by:
Sukadev Bhattiprolu <sukadev@linux.ibm.com> Reviewed-by:
Dany Madden <drt@linux.ibm.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Sukadev Bhattiprolu authored
If auto-priority-failover (APF) is enabled and there are at least two backing devices of different priorities, some resets like fail-over, change-param etc can cause at least two back to back failovers. (Failover from high priority backing device to lower priority one and then back to the higher priority one if that is still functional). Depending on the timimg of the two failovers it is possible to trigger a "hard" reset and for the hard reset to fail due to failovers. When this occurs, the driver assumes that the network is unstable and disables the VNIC for a 60-second "settling time". This in turn can cause the ethtool command to fail with "No such device" while the vnic automatically recovers a little while later. Given that it's possible to have two back to back failures, allow for extra failures before disabling the vnic for the settling time. Fixes: f15fde9d ("ibmvnic: delay next reset if hard reset fails") Signed-off-by:
Sukadev Bhattiprolu <sukadev@linux.ibm.com> Reviewed-by:
Dany Madden <drt@linux.ibm.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Jakub Kicinski authored
During IP fragmentation we sanitize IP options. This means overwriting options which should not be copied with NOPs. Only the first fragment has the original, full options. ip_fraglist_prepare() copies the IP header and options from previous fragment to the next one. Commit 19c3401a ("net: ipv4: place control buffer handling away from fragmentation iterators") moved sanitizing options before ip_fraglist_prepare() which means options are sanitized and then overwritten again with the old values. Fixing this is not enough, however, nor did the sanitization work prior to aforementioned commit. ip_options_fragment() (which does the sanitization) uses ipcb->opt.optlen for the length of the options. ipcb->opt of fragments is not populated (it's 0), only the head skb has the state properly built. So even when called at the right time ip_options_fragment() does nothing. This seems to date back all the way to v2.5.44 when the fast path for pre-fragmented skbs had been introduced. Prior to that ip_options_build() would have been called for every fragment (in fact ever since v2.5.44 the fragmentation handing in ip_options_build() has been dead code, I'll clean it up in -next). In the original patch (see Link) caixf mentions fixing the handling for fragments other than the second one, but I'm not sure how _any_ fragment could have had their options sanitized with the code as it stood. Tested with python (MTU on lo lowered to 1000 to force fragmentation): import socket s = socket.socket(socket.AF_INET, socket.SOCK_DGRAM) s.setsockopt(socket.IPPROTO_IP, socket.IP_OPTIONS, bytearray([7,4,5,192, 20|0x80,4,1,0])) s.sendto(b'1'*2000, ('127.0.0.1', 1234)) Before: IP (tos 0x0, ttl 64, id 1053, offset 0, flags [+], proto UDP (17), length 996, options (RR [bad length 4] [bad ptr 5] 192.148.4.1,,RA value 256)) localhost.36500 > localhost.search-agent: UDP, length 2000 IP (tos 0x0, ttl 64, id 1053, offset 968, flags [+], proto UDP (17), length 996, options (RR [bad length 4] [bad ptr 5] 192.148.4.1,,RA value 256)) localhost > localhost: udp IP (tos 0x0, ttl 64, id 1053, offset 1936, flags [none], proto UDP (17), length 100, options (RR [bad length 4] [bad ptr 5] 192.148.4.1,,RA value 256)) localhost > localhost: udp After: IP (tos 0x0, ttl 96, id 42549, offset 0, flags [+], proto UDP (17), length 996, options (RR [bad length 4] [bad ptr 5] 192.148.4.1,,RA value 256)) localhost.51607 > localhost.search-agent: UDP, bad length 2000 > 960 IP (tos 0x0, ttl 96, id 42549, offset 968, flags [+], proto UDP (17), length 996, options (NOP,NOP,NOP,NOP,RA value 256)) localhost > localhost: udp IP (tos 0x0, ttl 96, id 42549, offset 1936, flags [none], proto UDP (17), length 100, options (NOP,NOP,NOP,NOP,RA value 256)) localhost > localhost: udp RA (20 | 0x80) is now copied as expected, RR (7) is "NOPed out". Link: https://lore.kernel.org/netdev/20220107080559.122713-1-ooppublic@163.com/ Fixes: 19c3401a ("net: ipv4: place control buffer handling away from fragmentation iterators") Fixes: 1da177e4 ("Linux-2.6.12-rc2") Signed-off-by:
caixf <ooppublic@163.com> Signed-off-by:
Jakub Kicinski <kuba@kernel.org> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Jianguo Wu authored
After commit:7866a621 ("dev: add per net_device packet type chains"), we can not get packet types that are bound to a specified net device by /proc/net/ptype, this patch fix the regression. Run "tcpdump -i ens192 udp -nns0" Before and after apply this patch: Before: [root@localhost ~]# cat /proc/net/ptype Type Device Function 0800 ip_rcv 0806 arp_rcv 86dd ipv6_rcv After: [root@localhost ~]# cat /proc/net/ptype Type Device Function ALL ens192 tpacket_rcv 0800 ip_rcv 0806 arp_rcv 86dd ipv6_rcv v1 -> v2: - fix the regression rather than adding new /proc API as suggested by Stephen Hemminger. Fixes: 7866a621 ("dev: add per net_device packet type chains") Signed-off-by:
Jianguo Wu <wujianguo@chinatelecom.cn> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Hangbin Liu authored
bond_option_active_slave_get_rcu() should not be used in rtnl_mutex as it use rcu_dereference(). Replace to rcu_dereference_rtnl() so we also can use this function in rtnl protected context. With this update, we can rmeove the rcu_read_lock/unlock in bonding .ndo_eth_ioctl and .get_ts_info. Reported-by:
Vladimir Oltean <vladimir.oltean@nxp.com> Fixes: 94dd016a ("bond: pass get_ts_info and SIOC[SG]HWTSTAMP ioctl to active device") Signed-off-by:
Hangbin Liu <liuhangbin@gmail.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Marek Behún authored
Commit ce0aa27f ("sfp: add sfp-bus to bridge between network devices and sfp cages") added code which finds SFP bus DT node even if the node is disabled with status = "disabled". Because of this, when phylink is created, it ends with non-null .sfp_bus member, even though the SFP module is not probed (because the node is disabled). We need to ignore disabled SFP bus node. Fixes: ce0aa27f ("sfp: add sfp-bus to bridge between network devices and sfp cages") Signed-off-by:
Marek Behún <kabel@kernel.org> Cc: stable@vger.kernel.org # 2203cbf2 ("net: sfp: move fwnode parsing into sfp-bus layer") Signed-off-by:
David S. Miller <davem@davemloft.net>
-
- Jan 22, 2022
-
-
Justin Iurman authored
The IOAM queue-depth data field was added a few weeks ago, but the test unit was not updated accordingly. Reported-by:
kernel test robot <oliver.sang@intel.com> Fixes: b63c5478 ("ipv6: ioam: Support for Queue depth data field") Signed-off-by:
Justin Iurman <justin.iurman@uliege.be> Link: https://lore.kernel.org/r/20220121173449.26918-1-justin.iurman@uliege.be Signed-off-by:
Jakub Kicinski <kuba@kernel.org>
-
Kees Cook authored
In preparation for FORTIFY_SOURCE performing compile-time and run-time field bounds checking for memcpy(), memmove(), and memset(), avoid intentionally writing across neighboring fields. Use struct_group() to capture the fields to be reset, so that memset() can be appropriately bounds-checked by the compiler. Cc: Matthieu Baerts <matthieu.baerts@tessares.net> Cc: mptcp@lists.linux.dev Signed-off-by:
Kees Cook <keescook@chromium.org> Reviewed-by:
Mat Martineau <mathew.j.martineau@linux.intel.com> Link: https://lore.kernel.org/r/20220121073935.1154263-1-keescook@chromium.org Signed-off-by:
Jakub Kicinski <kuba@kernel.org>
-
David Howells authored
Improve retransmission backoff by only backing off when we retransmit data packets rather than when we set the lost ack timer. To this end: (1) In rxrpc_resend(), use rxrpc_get_rto_backoff() when setting the retransmission timer and only tell it that we are retransmitting if we actually have things to retransmit. Note that it's possible for the retransmission algorithm to race with the processing of a received ACK, so we may see no packets needing retransmission. (2) In rxrpc_send_data_packet(), don't bump the backoff when setting the ack_lost_at timer, as it may then get bumped twice. With this, when looking at one particular packet, the retransmission intervals were seen to be 1.5ms, 2ms, 3ms, 5ms, 9ms, 17ms, 33ms, 71ms, 136ms, 264ms, 544ms, 1.088s, 2.1s, 4.2s and 8.3s. Fixes: c410bf01 ("rxrpc: Fix the excessive initial retransmission timeout") Suggested-by:
Marc Dionne <marc.dionne@auristor.com> Signed-off-by:
David Howells <dhowells@redhat.com> Reviewed-by:
Marc Dionne <marc.dionne@auristor.com> Tested-by:
Marc Dionne <marc.dionne@auristor.com> cc: linux-afs@lists.infradead.org Link: https://lore.kernel.org/r/164138117069.2023386.17446904856843997127.stgit@warthog.procyon.org.uk/ Signed-off-by:
David S. Miller <davem@davemloft.net>
-
- Jan 21, 2022
-
-
David S. Miller authored
Subbaraya Sundeep says: ==================== octeontx-af2: Fixes for CN10K and CN9xxx platforms This patchset has consolidated fixes in Octeontx2 driver handling CN10K and CN9xxx platforms. When testing the new CN10K hardware some issues resurfaced like accessing wrong register for CN10K and enabling loopback on not supported interfaces. Some fixes are needed for CN9xxx platforms as well. Below is the description of patches Patch 1: AF sets RX RSS action for all the VFs when a VF is brought up. But when a PF sets RX action for its VF like Drop/Direct to a queue in ntuple filter it is not retained because of AF fixup. This patch skips modifying VF RX RSS action if PF has already set its action. Patch 2: When configuring backpressure wrong register is being read for LBKs hence fixed it. Patch 3: Some RVU blocks may take longer time to reset but are guaranteed to complete the reset. Hence wait till reset is complete. Patch 4: For enabling LMAC CN10K needs another register compared to CN9xxx platforms. Hence changed it. Patch 5: Adds missing barrier before submitting memory pointer to the aura hardware. Patch 6: Increase polling time while link credit restore and also return proper error code when timeout occurs. Patch 7: Internal loopback not supported on LPCS interfaces like SGMII/QSGMII so do not enable it. Patch 8: When there is a error in message processing, AF sets the error response and replies back to requestor. PF forwards a invalid message to VF back if AF reply has error in it. This way VF lacks the actual error set by AF for its message. This is changed such that PF simply forwards the actual reply and let VF handle the error. Patch 9: ntuple filter with "flow-type ether proto 0x8842 vlan 0x92e" was not working since ethertype 0x8842 is NGIO protocol. Hardware parser explicitly parses such NGIO packets and sets the packet as NGIO and do not set it as tagged packet. Fix this by changing parser such that it sets the packet as both NGIO and tagged by using separate layer types. ==================== Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Kiran Kumar K authored
With current KPU profile NGIO is being parsed along with CTAG as a single layer. Because of this MCAM/ntuple rules installed with ethertype as 0x8842 are not being hit. Adding KPU profile changes to parse NGIO in separate ltype and CTAG in separate ltype. Fixes: f9c49be9 ("octeontx2-af: Update the default KPU profile and fixes") Signed-off-by:
Kiran Kumar K <kirankumark@marvell.com> Signed-off-by:
Subbaraya Sundeep <sbhatta@marvell.com> Signed-off-by:
Sunil Goutham <sgoutham@marvell.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Subbaraya Sundeep authored
PF forwards its VF messages to AF and corresponding replies from AF to VF. AF sets proper error code in the replies after processing message requests. Currently PF checks the error codes in replies and sends invalid message to VF. This way VF lacks the information of error code set by AF for its messages. This patch changes that such that PF simply forwards AF replies so that VF can handle error codes. Fixes: d424b6c0 ("octeontx2-pf: Enable SRIOV and added VF mbox handling") Signed-off-by:
Subbaraya Sundeep <sbhatta@marvell.com> Signed-off-by:
Sunil Goutham <sgoutham@marvell.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Geetha sowjanya authored
Internal looback is not supported to low rate LPCS interface like SGMII/QSGMII. Hence don't allow to enable for such interfaces. Fixes: 3ad3f8f9 ("octeontx2-af: cn10k: MAC internal loopback support") Signed-off-by:
Geetha sowjanya <gakula@marvell.com> Signed-off-by:
Subbaraya Sundeep <sbhatta@marvell.com> Signed-off-by:
Sunil Goutham <sgoutham@marvell.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Geetha sowjanya authored
It's been observed that sometimes link credit restore takes a lot of time than the current timeout. This patch increases the default timeout value and return the proper error value on failure. Fixes: 1c74b891 ("octeontx2-af: Wait for TX link idle for credits change") Signed-off-by:
Geetha sowjanya <gakula@marvell.com> Signed-off-by:
Subbaraya Sundeep <sbhatta@marvell.com> Signed-off-by:
Sunil Goutham <sgoutham@marvell.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Geetha sowjanya authored
While freeing SQB pointers to aura, driver first memcpy to target address and then triggers lmtst operation to free pointer to the aura. We need to ensure(by adding dmb barrier)that memcpy is finished before pointers are freed to the aura. This patch also adds the missing sq context structure entry in debugfs. Fixes: ef6c8da7 ("octeontx2-pf: cn10K: Reserve LMTST lines per core") Signed-off-by:
Geetha sowjanya <gakula@marvell.com> Signed-off-by:
Subbaraya Sundeep <sbhatta@marvell.com> Signed-off-by:
Sunil Goutham <sgoutham@marvell.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Geetha sowjanya authored
CN10K platforms uses RPM(0..2)_MTI_MAC100(0..3)_COMMAND_CONFIG register for lmac TX/RX enable whereas CN9xxx platforms use CGX_CMRX_CONFIG register. This config change was missed when adding support for CN10K RPM. Fixes: 91c6945e ("octeontx2-af: cn10k: Add RPM MAC support") Signed-off-by:
Geetha sowjanya <gakula@marvell.com> Signed-off-by:
Subbaraya Sundeep <sbhatta@marvell.com> Signed-off-by:
Sunil Goutham <sgoutham@marvell.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-