- 14 Dec, 2023 7 commits
-
-
Jianbo Liu authored
Coverity Scan reports the following issue. But it's impossible that mlx5_get_dev_index returns 7 for PF, even if the index is calculated from PCI FUNC ID. So add the checking to make coverity slience. CID 610894 (#2 of 2): Out-of-bounds write (OVERRUN) Overrunning array esw->fdb_table.offloads.peer_miss_rules of 4 8-byte elements at element index 7 (byte offset 63) using index mlx5_get_dev_index(peer_dev) (which evaluates to 7). Fixes: 9bee385a ("net/mlx5: E-switch, refactor FDB miss rule add/remove") Signed-off-by:
Jianbo Liu <jianbol@nvidia.com> Reviewed-by:
Roi Dayan <roid@nvidia.com> Signed-off-by:
Saeed Mahameed <saeedm@nvidia.com>
-
Dinghao Liu authored
When kcalloc() for ft->g succeeds but kvzalloc() for in fails, fs_udp_create_groups() will free ft->g. However, its caller fs_udp_create_table() will free ft->g again through calling mlx5e_destroy_flow_table(), which will lead to a double-free. Fix this by setting ft->g to NULL in fs_udp_create_groups(). Fixes: 1c80bd68 ("net/mlx5e: Introduce Flow Steering UDP API") Signed-off-by:
Dinghao Liu <dinghao.liu@zju.edu.cn> Reviewed-by:
Tariq Toukan <tariqt@nvidia.com> Reviewed-by:
Simon Horman <horms@kernel.org> Signed-off-by:
-
Shifeng Li authored
Fix a cmd->ent use after free due to a race on command entry. Such race occurs when one of the commands releases its last refcount and frees its index and entry while another process running command flush flow takes refcount to this command entry. The process which handles commands flush may see this command as needed to be flushed if the other process allocated a ent->idx but didn't set ent to cmd->ent_arr in cmd_work_handler(). Fix it by moving the assignment of cmd->ent_arr into the spin lock. [70013.081955] BUG: KASAN: use-after-free in mlx5_cmd_trigger_completions+0x1e2/0x4c0 [mlx5_core] [70013.081967] Write of size 4 at addr ffff88880b1510b4 by task kworker/26:1/1433361 [70013.081968] [70013.082028] Workqueue: events aer_isr [70013.082053] Call Trace: [70013.082067] dump_stack+0x8b/0xbb [70013.082086] print_address_description+0x6a/0x270 [70013.082102] kasan_report+0x179/0x2c0 [70013.082173] mlx5_cmd_trigger_completions+0x1e2/0x4c0 [mlx5_core] [70013.082267] mlx5_cmd_flush+0x80/0x180 [mlx5_core] [70013.082304] mlx5_enter_error_state+0x106/0x1d0 [mlx5_core] [70013.082338] mlx5_try_fast_unload+0x2ea/0x4d0 [mlx5_core] [70013.082377] remove_one+0x200/0x2b0 [mlx5_core] [70013.082409] pci_device_remove+0xf3/0x280 [70013.082439] device_release_driver_internal+0x1c3/0x470 [70013.082453] pci_stop_bus_device+0x109/0x160 [70013.082468] pci_stop_and_remove_bus_device+0xe/0x20 [70013.082485] pcie_do_fatal_recovery+0x167/0x550 [70013.082493] aer_isr+0x7d2/0x960 [70013.082543] process_one_work+0x65f/0x12d0 [70013.082556] worker_thread+0x87/0xb50 [70013.082571] kthread+0x2e9/0x3a0 [70013.082592] ret_from_fork+0x1f/0x40 The logical relationship of this error is as follows: aer_recover_work | ent->work -------------------------------------------+------------------------------ aer_recover_work_func | |- pcie_do_recovery | |- report_error_detected | |- mlx5_pci_err_detected |cmd_work_handler |- mlx5_enter_error_state | |- cmd_alloc_index |- enter_error_state | |- lock cmd->alloc_lock |- mlx5_cmd_flush | |- clear_bit |- mlx5_cmd_trigger_completions| |- unlock cmd->alloc_lock |- lock cmd->alloc_lock | |- vector = ~dev->cmd.vars.bitmask |- for_each_set_bit | |- cmd_ent_get(cmd->ent_arr[i]) (UAF) |- unlock cmd->alloc_lock | |- cmd->ent_arr[ent->idx]=ent The cmd->ent_arr[ent->idx] assignment and the bit clearing are not protected by the cmd->alloc_lock in cmd_work_handler(). Fixes: 50b2412b ("net/mlx5: Avoid possible free of command entry while timeout comp handler") Reviewed-by:Moshe Shemesh <moshe@nvidia.com> Signed-off-by:
Shifeng Li <lishifeng@sangfor.com.cn> Signed-off-by:
-
Shifeng Li authored
Out_sz that the size of out buffer is calculated using query_nic_vport _context_in structure when driver query the MAC list. However query_nic _vport_context_in structure is smaller than query_nic_vport_context_out. When allowed_list_size is greater than 96, calling ether_addr_copy() will trigger an slab-out-of-bounds. [ 1170.055866] BUG: KASAN: slab-out-of-bounds in mlx5_query_nic_vport_mac_list+0x481/0x4d0 [mlx5_core] [ 1170.055869] Read of size 4 at addr ffff88bdbc57d912 by task kworker/u128:1/461 [ 1170.055870] [ 1170.055932] Workqueue: mlx5_esw_wq esw_vport_change_handler [mlx5_core] [ 1170.055936] Call Trace: [ 1170.055949] dump_stack+0x8b/0xbb [ 1170.055958] print_address_description+0x6a/0x270 [ 1170.055961] kasan_report+0x179/0x2c0 [ 1170.056061] mlx5_query_nic_vport_mac_list+0x481/0x4d0 [mlx5_core] [ 1170.056162] esw_update_vport_addr_list+0x2c5/0xcd0 [mlx5_core] [ 1170.056257] esw_vport_change_handle_locked+0xd08/0x1a20 [mlx5_core] [ 1170.056377] esw_vport_change_handler+0x6b/0x90 [mlx5_core] [ 1170.056381] process_one_work+0x65f/0x12d0 [ 1170.056383] worker_thread+0x87/0xb50 [ 1170.056390] kthread+0x2e9/0x3a0 [ 1170.056394] ret_from_fork+0x1f/0x40 Fixes: e16aea27 ("net/mlx5: Introduce access functions to modify/query vport mac lists") Cc: Ding Hui <dinghui@sangfor.com.cn> Signed-off-by:
-
Vlad Buslov authored
Cited commit introduced potential double free since encap_header can be destroyed twice in some cases - once by error cleanup sequence in mlx5e_tc_tun_{create|update}_header_ipv{4|6}(), once by generic mlx5e_encap_put() that user calls as a result of getting an error from tunnel create|update. At the same time the point where e->encap_header is assigned can't be delayed because the function can still return non-error code 0 as a result of checking for NUD_VALID flag, which will cause neighbor update to dereference NULL encap_header. Fix the issue by: - Nulling local encap_header variables in mlx5e_tc_tun_{create|update}_header_ipv{4|6}() to make kfree(encap_header) call in error cleanup sequence noop after that point. - Assigning reformat_params.data from e->encap_header instead of local variable encap_header that was set to NULL pointer by previous step. Also assign reformat_params.size from e->encap_size for uniformity and in order to make the code less error-prone in the future. Fixes: d589e785 ("net/mlx5e: Allow concurrent creation of encap entries") Reported-by:Dust Li <dust.li@linux.alibaba.com> Reported-by:
Cruz Zhao <cruzzhao@linux.alibaba.com> Reported-by:
Tianchen Ding <dtcccc@linux.alibaba.com> Signed-off-by:
Vlad Buslov <vladbu@nvidia.com> Signed-off-by:
-
Vlad Buslov authored
This reverts commit 6f9b1a07. This patch is causing a null ptr issue, the proper fix is in the next patch. Fixes: 6f9b1a07 ("net/mlx5e: fix double free of encap_header") Signed-off-by:
-
Vlad Buslov authored
This reverts commit 3a4aa3cb. This patch is causing a null ptr issue, the proper fix is in the next patch. Fixes: 3a4aa3cb ("net/mlx5e: fix double free of encap_header in update funcs") Signed-off-by:
-
- 13 Dec, 2023 12 commits
-
-
Jakub Kicinski authored
This reverts commit f3f32a35. Paolo reports that the change disables autocorking even after the userspace sets TCP_CORK. Fixes: f3f32a35 ("tcp: disable tcp_autocorking for socket when TCP_NODELAY flag is set") Link: https://lore.kernel.org/r/0d30d5a41d3ac990573016308aaeacb40a9dc79f.camel@redhat.comSigned-off-by:
Jakub Kicinski <kuba@kernel.org>
-
David S. Miller authored
Yanteng Si says: ==================== stmmac: Some bug fixes * Put Krzysztof's patch into my thread, pick Conor's Reviewed-by tag and Jiaxun's Acked-by tag.(prev version is RFC patch) * I fixed an Oops related to mdio, mainly to ensure that mdio is initialized before use, because it will be used in a series of patches I am working on. see <https://lore.kernel.org/loongarch/cover.1699533745.git.siyanteng@loongson.cn/T/#t> ==================== Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Krzysztof Kozlowski authored
Device binds to proper PCI ID (LOONGSON, 0x7a03), already listed in DTS, so checking for some other compatible does not make sense. It cannot be bound to unsupported platform. Drop useless, incorrect (space in between) and undocumented compatible. Signed-off-by:
Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> Signed-off-by:
Yanteng Si <siyanteng@loongson.cn> Reviewed-by:
Conor Dooley <conor.dooley@microchip.com> Acked-by:
Jiaxun Yang <jiaxun.yang@flygoat.com> Signed-off-by:
-
Krzysztof Kozlowski authored
Device binds to proper PCI ID (LOONGSON, 0x7a03), already listed in DTS, so checking for some other compatible does not make sense. It cannot be bound to unsupported platform. Drop useless, incorrect (space in between) and undocumented compatible. Signed-off-by:
-
Yanteng Si authored
Generic code will use mdio. If it is not initialized before use, the kernel will Oops. Fixes: 30bba69d ("stmmac: pci: Add dwmac support for Loongson") Signed-off-by:
Feiyang Chen <chenfeiyang@loongson.cn> Reviewed-by:
Andrew Lunn <andrew@lunn.ch> Signed-off-by:
-
Salvatore Dipietro authored
Based on the tcp man page, if TCP_NODELAY is set, it disables Nagle's algorithm and packets are sent as soon as possible. However in the `tcp_push` function where autocorking is evaluated the `nonagle` value set by TCP_NODELAY is not considered which can trigger unexpected corking of packets and induce delays. For example, if two packets are generated as part of a server's reply, if the first one is not transmitted on the wire quickly enough, the second packet can trigger the autocorking in `tcp_push` and be delayed instead of sent as soon as possible. It will either wait for additional packets to be coalesced or an ACK from the client before transmitting the corked packet. This can interact badly if the receiver has tcp delayed acks enabled, introducing 40ms extra delay in completion times. It is not always possible to control who has delayed acks set, but it is possible to adjust when and how autocorking is triggered. Patch prevents autocorking if the TCP_NODELAY flag is set on the socket. Patch has been tested using an AWS c7g.2xlarge instance with Ubuntu 22.04 and Apache Tomcat 9.0.83 running the basic servlet below: import java.io.IOException; import java.io.OutputStreamWriter; import java.io.PrintWriter; import javax.servlet.ServletException; import javax.servlet.http.HttpServlet; import javax.servlet.http.HttpServletRequest; import javax.servlet.http.HttpServletResponse; public class HelloWorldServlet extends HttpServlet { @Override protected void doGet(HttpServletRequest request, HttpServletResponse response) throws ServletException, IOException { response.setContentType("text/html;charset=utf-8"); OutputStreamWriter osw = new OutputStreamWriter(response.getOutputStream(),"UTF-8"); String s = "a".repeat(3096); osw.write(s,0,s.length()); osw.flush(); } } Load was applied using wrk2 (https://github.com/kinvolk/wrk2) from an AWS c6i.8xlarge instance. With the current auto-corking behavior and TCP_NODELAY set an additional 40ms latency from P99.99+ values are observed. With the patch applied we see no occurrences of 40ms latencies. The patch has also been tested with iperf and uperf benchmarks and no regression was observed. # No patch with tcp_autocorking=1 and TCP_NODELAY set on all sockets ./wrk -t32 -c128 -d40s --latency -R10000 http://172.31.49.177:8080/hello/hello' ... 50.000% 0.91ms 75.000% 1.12ms 90.000% 1.46ms 99.000% 1.73ms 99.900% 1.96ms 99.990% 43.62ms <<< 40+ ms extra latency 99.999% 48.32ms 100.000% 49.34ms # With patch ./wrk -t32 -c128 -d40s --latency -R10000 http://172.31.49.177:8080/hello/hello' ... 50.000% 0.89ms 75.000% 1.13ms 90.000% 1.44ms 99.000% 1.67ms 99.900% 1.78ms 99.990% 2.27ms <<< no 40+ ms extra latency 99.999% 3.71ms 100.000% 4.57ms Fixes: f54b3111 ("tcp: auto corking") Signed-off-by:Salvatore Dipietro <dipiets@amazon.com> Reviewed-by:
Eric Dumazet <edumazet@google.com> Signed-off-by:
-
Jiri Pirko authored
User may not pass DPLL_A_PIN_STATE attribute in the pin set operation message. Sanitize that by checking if the attr pointer is not null and process the passed state attribute value only in that case. Reported-by:
Xingyuan Mo <hdthky0@gmail.com> Fixes: 9d71b54b ("dpll: netlink: Add DPLL framework base functions") Signed-off-by:
Jiri Pirko <jiri@nvidia.com> Acked-by:
Vadim Fedorenko <vadim.fedorenko@linux.dev> Link: https://lore.kernel.org/r/20231211083758.1082853-1-jiri@resnulli.usSigned-off-by:
-
Jakub Kicinski authored
David Arinzon says: ==================== ENA driver XDP bug fixes This patchset contains multiple XDP-related bug fixes in the ENA driver. ==================== Link: https://lore.kernel.org/r/20231211062801.27891-1-darinzon@amazon.comSigned-off-by:
-
David Arinzon authored
When sending TX packets, the meta descriptor can be all zeroes as no meta information is required (as in XDP). This patch removes the validity check, as when `disable_meta_caching` is enabled, such TX packets will be dropped otherwise. Fixes: 0e3a3f6d ("net: ena: support new LLQ acceleration mode") Signed-off-by:
Shay Agroskin <shayagr@amazon.com> Signed-off-by:
David Arinzon <darinzon@amazon.com> Link: https://lore.kernel.org/r/20231211062801.27891-5-darinzon@amazon.comSigned-off-by:
-
David Arinzon authored
This patch fixes two issues: Issue 1 ------- Description ``````````` Current code does not call dma_sync_single_for_cpu() to sync data from the device side memory to the CPU side memory before the XDP code path uses the CPU side data. This causes the XDP code path to read the unset garbage data in the CPU side memory, resulting in incorrect handling of the packet by XDP. Solution ```````` 1. Add a call to dma_sync_single_for_cpu() before the XDP code starts to use the data in the CPU side memory. 2. The XDP code verdict can be XDP_PASS, in which case there is a fallback to the non-XDP code, which also calls dma_sync_single_for_cpu(). To avoid calling dma_sync_single_for_cpu() twice: 2.1. Put the dma_sync_single_for_cpu() in the code in such a place where it happens before XDP and non-XDP code. 2.2. Remove the calls to dma_sync_single_for_cpu() in the non-XDP code for the first buffer only (rx_copybreak and non-rx_copybreak cases), since the new call that was added covers these cases. The call to dma_sync_single_for_cpu() for the second buffer and on stays because only the first buffer is handled by the newly added dma_sync_single_for_cpu(). And there is no need for special handling of the second buffer and on for the XDP path since currently the driver supports only single buffer packets. Issue 2 ------- Description ``````````` In case the XDP code forwarded the packet (ENA_XDP_FORWARDED), ena_unmap_rx_buff_attrs() is called with attrs set to 0. This means that before unmapping the buffer, the internal function dma_unmap_page_attrs() will also call dma_sync_single_for_cpu() on the whole buffer (not only on the data part of it). This sync is both wasteful (since a sync was already explicitly called before) and also causes a bug, which will be explained using the below diagram. The following diagram shows the flow of events causing the bug. The order of events is (1)-(4) as shown in the diagram. CPU side memory area (3)convert_to_xdp_frame() initializes the headroom with xdpf metadata || \/ ___________________________________ | | 0 | V 4K --------------------------------------------------------------------- | xdpf->data | other xdpf | < data > | tailroom ||...| | | fields | | GARBAGE || | --------------------------------------------------------------------- /\ /\ || || (4)ena_unmap_rx_buff_attrs() calls (2)dma_sync_single_for_cpu() dma_sync_single_for_cpu() on the copies data from device whole buffer page, overwriting side to CPU side memory the xdpf->data with GARBAGE. || 0 4K --------------------------------------------------------------------- | headroom | < data > | tailroom ||...| | GARBAGE | | GARBAGE || | --------------------------------------------------------------------- Device side memory area /\ || (1) device writes RX packet data After the call to ena_unmap_rx_buff_attrs() in (4), the xdpf->data becomes corrupted, and so when it is later accessed in ena_clean_xdp_irq()->xdp_return_frame(), it causes a page fault, crashing the kernel. Solution ```````` Explicitly tell ena_unmap_rx_buff_attrs() not to call dma_sync_single_for_cpu() by passing it the ENA_DMA_ATTR_SKIP_CPU_SYNC flag. Fixes: f7d625ad ("net: ena: Add dynamic recycling mechanism for rx buffers") Signed-off-by:Arthur Kiyanovski <akiyano@amazon.com> Signed-off-by:
-
David Arinzon authored
Current xdp code drops packets larger than ENA_XDP_MAX_MTU. This is an incorrect condition since the problem is not the size of the packet, rather the number of buffers it contains. This commit: 1. Identifies and drops XDP multi-buffer packets at the beginning of the function. 2. Increases the xdp drop statistic when this drop occurs. 3. Adds a one-time print that such drops are happening to give better indication to the user. Fixes: 838c93dc ("net: ena: implement XDP drop support") Signed-off-by:
-
David Arinzon authored
The ena_setup_and_create_all_xdp_queues() function freed all the resources upon failure, after creating only xdp_num_queues queues, instead of freeing just the created ones. In this patch, the only resources that are freed, are the ones allocated right before the failure occurs. Fixes: 548c4940 ("net: ena: Implement XDP_TX action") Signed-off-by:
Shahar Itzko <itzko@amazon.com> Signed-off-by:
-
- 12 Dec, 2023 4 commits
-
-
Dong Chenchen authored
syzkaller report: kernel BUG at net/core/skbuff.c:3452! invalid opcode: 0000 [#1] PREEMPT SMP KASAN PTI CPU: 0 PID: 0 Comm: swapper/0 Not tainted 6.7.0-rc4-00009-gbee0e776-dirty #135 RIP: 0010:skb_copy_and_csum_bits (net/core/skbuff.c:3452) Call Trace: icmp_glue_bits (net/ipv4/icmp.c:357) __ip_append_data.isra.0 (net/ipv4/ip_output.c:1165) ip_append_data (net/ipv4/ip_output.c:1362 net/ipv4/ip_output.c:1341) icmp_push_reply (net/ipv4/icmp.c:370) __icmp_send (./include/net/route.h:252 net/ipv4/icmp.c:772) ip_fragment.constprop.0 (./include/linux/skbuff.h:1234 net/ipv4/ip_output.c:592 net/ipv4/ip_output.c:577) __ip_finish_output (net/ipv4/ip_output.c:311 net/ipv4/ip_output.c:295) ip_output (net/ipv4/ip_output.c:427) __ip_queue_xmit (net/ipv4/ip_output.c:535) __tcp_transmit_skb (net/ipv4/tcp_output.c:1462) __tcp_retransmit_skb (net/ipv4/tcp_output.c:3387) tcp_retransmit_skb (net/ipv4/tcp_output.c:3404) tcp_retransmit_timer (net/ipv4/tcp_timer.c:604) tcp_write_timer (./include/linux/spinlock.h:391 net/ipv4/tcp_timer.c:716) The panic issue was trigered by tcp simultaneous initiation. The initiation process is as follows: TCP A TCP B 1. CLOSED CLOSED 2. SYN-SENT --> <SEQ=100><CTL=SYN> ... 3. SYN-RECEIVED <-- <SEQ=300><CTL=SYN> <-- SYN-SENT 4. ... <SEQ=100><CTL=SYN> --> SYN-RECEIVED 5. SYN-RECEIVED --> <SEQ=100><ACK=301><CTL=SYN,ACK> ... // TCP B: not send challenge ack for ack limit or packet loss // TCP A: close tcp_close tcp_send_fin if (!tskb && tcp_under_memory_pressure(sk)) tskb = skb_rb_last(&sk->tcp_rtx_queue); //pick SYN_ACK packet TCP_SKB_CB(tskb)->tcp_flags |= TCPHDR_FIN; // set FIN flag 6. FIN_WAIT_1 --> <SEQ=100><ACK=301><END_SEQ=102><CTL=SYN,FIN,ACK> ... // TCP B: send challenge ack to SYN_FIN_ACK 7. ... <SEQ=301><ACK=101><CTL=ACK> <-- SYN-RECEIVED //challenge ack // TCP A: <SND.UNA=101> 8. FIN_WAIT_1 --> <SEQ=101><ACK=301><END_SEQ=102><CTL=SYN,FIN,ACK> ... // retransmit panic __tcp_retransmit_skb //skb->len=0 tcp_trim_head len = tp->snd_una - TCP_SKB_CB(skb)->seq // len=101-100 __pskb_trim_head skb->data_len -= len // skb->len=-1, wrap around ... ... ip_fragment icmp_glue_bits //BUG_ON If we use tcp_trim_head() to remove acked SYN from packet that contains data or other flags, skb->len will be incorrectly decremented. We can remove SYN flag that has been acked from rtx_queue earlier than tcp_trim_head(), which can fix the problem mentioned above. Fixes: 1da177e4 ("Linux-2.6.12-rc2") Co-developed-by:
Dong Chenchen <dongchenchen2@huawei.com> Link: https://lore.kernel.org/r/20231210020200.1539875-1-dongchenchen2@huawei.comSigned-off-by:
-
Dinghao Liu authored
qed_ilt_shadow_alloc() will call qed_ilt_shadow_free() to free p_hwfn->p_cxt_mngr->ilt_shadow on error. However, qed_cxt_tables_alloc() accesses the freed pointer on failure of qed_ilt_shadow_alloc() through calling qed_cxt_mngr_free(), which may lead to use-after-free. Fix this issue by setting p_mngr->ilt_shadow to NULL in qed_ilt_shadow_free(). Fixes: fe56b9e6 ("qed: Add module with basic common support") Reviewed-by:
Przemek Kitszel <przemyslaw.kitszel@intel.com> Signed-off-by:
-
Hyunwoo Kim authored
Because rose_ioctl() accesses sk->sk_receive_queue without holding a sk->sk_receive_queue.lock, it can cause a race with rose_accept(). A use-after-free for skb occurs with the following flow. ``` rose_ioctl() -> skb_peek() rose_accept() -> skb_dequeue() -> kfree_skb() ``` Add sk->sk_receive_queue.lock to rose_ioctl() to fix this issue. Fixes: 1da177e4 ("Linux-2.6.12-rc2") Signed-off-by:
Hyunwoo Kim <v4bel@theori.io> Link: https://lore.kernel.org/r/20231209100538.GA407321@v4bel-B760M-AORUS-ELITE-AXSigned-off-by:
Paolo Abeni <pabeni@redhat.com>
-
Hyunwoo Kim authored
Because do_vcc_ioctl() accesses sk->sk_receive_queue without holding a sk->sk_receive_queue.lock, it can cause a race with vcc_recvmsg(). A use-after-free for skb occurs with the following flow. ``` do_vcc_ioctl() -> skb_peek() vcc_recvmsg() -> skb_recv_datagram() -> skb_free_datagram() ``` Add sk->sk_receive_queue.lock to do_vcc_ioctl() to fix this issue. Fixes: 1da177e4 ("Linux-2.6.12-rc2") Signed-off-by:
-
- 11 Dec, 2023 6 commits
-
-
Hariprasad Kelam authored
The current implementation's default Pause Forward setting is causing unnecessary network traffic. This patch disables Pause Forward to address this issue. Fixes: 1121f6b0 ("octeontx2-af: Priority flow control configuration support") Signed-off-by:
Hariprasad Kelam <hkelam@marvell.com> Signed-off-by:
Sunil Kovvuri Goutham <sgoutham@marvell.com> Signed-off-by:
-
David S. Miller authored
Hariprasad Kelam says: ==================== octeontx2: Fix issues with promisc/allmulti mode When interface is configured in promisc/all multi mode, low network performance observed. This series patches address the same. Patch1: Change the promisc/all multi mcam entry action to unicast if there are no trusted vfs associated with PF. Patch2: Configures RSS flow algorithm in promisc/all multi mcam entries to address flow distribution issues. ==================== Signed-off-by: -
Hariprasad Kelam authored
The RSS flow algorithm is not set up correctly for promiscuous or all multi MCAM entries. This has an impact on flow distribution. This patch fixes the issue by updating flow algorithm index in above mentioned MCAM entries. Fixes: 967db352 ("octeontx2-af: add support for multicast/promisc packet replication feature") Signed-off-by:
-
Hariprasad Kelam authored
Current implementation is such that, promisc mcam entry action is set as multicast even when there are no trusted VFs. multicast action causes the hardware to copy packet data, which reduces the performance. This patch fixes this issue by setting the promisc mcam entry action to unicast instead of multicast when there are no trusted VFs. The same change is made for the 'allmulti' mcam entry action. Fixes: ffd2f89a ("octeontx2-pf: Enable promisc/allmulti match MCAM entries.") Signed-off-by:
-
Shinas Rasheed authored
The firmware ready value is 1, and get firmware ready status function should explicitly test for that value. The firmware ready value read will be 2 after driver load, and on unbind till firmware rewrites the firmware ready back to 0, the value seen by driver will be 2, which should be regarded as not ready. Fixes: 10c073e4 ("octeon_ep: defer probe if firmware not ready") Signed-off-by:
Shinas Rasheed <srasheed@marvell.com> Reviewed-by:
-
Vlad Buslov authored
The referenced change added custom cleanup code to act_ct to delete any callbacks registered on the parent block when deleting the tcf_ct_flow_table instance. However, the underlying issue is that the drivers don't obtain the reference to the tcf_ct_flow_table instance when registering callbacks which means that not only driver callbacks may still be on the table when deleting it but also that the driver can still have pointers to its internal nf_flowtable and can use it concurrently which results either warning in netfilter[0] or use-after-free. Fix the issue by taking a reference to the underlying struct tcf_ct_flow_table instance when registering the callback and release the reference when unregistering. Expose new API required for such reference counting by adding two new callbacks to nf_flowtable_type and implementing them for act_ct flowtable_ct type. This fixes the issue by extending the lifetime of nf_flowtable until all users have unregistered. [0]: [106170.938634] ------------[ cut here ]------------ [106170.939111] WARNING: CPU: 21 PID: 3688 at include/net/netfilter/nf_flow_table.h:262 mlx5_tc_ct_del_ft_cb+0x267/0x2b0 [mlx5_core] [106170.940108] Modules linked in: act_ct nf_flow_table act_mirred act_skbedit act_tunnel_key vxlan cls_matchall nfnetlink_cttimeout act_gact cls_flower sch_ingress mlx5_vdpa vringh vhost_iotlb vdpa bonding openvswitch nsh rpcrdma rdma_ucm ib_iser libiscsi scsi_transport_iscsi ib_umad rdma_cm ib_ipoib iw_cm ib_cm mlx5_ib ib_uverbs ib_core xt_MASQUERADE nf_conntrack_netlink nfnetlink iptable_nat xt_addrtype xt_conntrack nf_nat br_netfilter rpcsec_gss_krb5 auth_rpcgss oid_regis try overlay mlx5_core [106170.943496] CPU: 21 PID: 3688 Comm: kworker/u48:0 Not tainted 6.6.0-rc7_for_upstream_min_debug_2023_11_01_13_02 #1 [106170.944361] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS rel-1.13.0-0-gf21b5a4aeb02-prebuilt.qemu.org 04/01/2014 [106170.945292] Workqueue: mlx5e mlx5e_rep_neigh_update [mlx5_core] [106170.945846] RIP: 0010:mlx5_tc_ct_del_ft_cb+0x267/0x2b0 [mlx5_core] [106170.946413] Code: 89 ef 48 83 05 71 a4 14 00 01 e8 f4 06 04 e1 48 83 05 6c a4 14 00 01 48 83 c4 28 5b 5d 41 5c 41 5d c3 48 83 05 d1 8b 14 00 01 <0f> 0b 48 83 05 d7 8b 14 00 01 e9 96 fe ff ff 48 83 05 a2 90 14 00 [106170.947924] RSP: 0018:ffff88813ff0fcb8 EFLAGS: 00010202 [106170.948397] RAX: 0000000000000000 RBX: ffff88811eabac40 RCX: ffff88811eabad48 [106170.949040] RDX: ffff88811eab8000 RSI: ffffffffa02cd560 RDI: 0000000000000000 [106170.949679] RBP: ffff88811eab8000 R08: 0000000000000001 R09: ffffffffa0229700 [106170.950317] R10: ffff888103538fc0 R11: 0000000000000001 R12: ffff88811eabad58 [106170.950969] R13: ffff888110c01c00 R14: ffff888106b40000 R15: 0000000000000000 [106170.951616] FS: 0000000000000000(0000) GS:ffff88885fd40000(0000) knlGS:0000000000000000 [106170.952329] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [106170.952834] CR2: 00007f1cefd28cb0 CR3: 000000012181b006 CR4: 0000000000370ea0 [106170.953482] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 [106170.954121] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 [106170.954766] Call Trace: [106170.955057] <TASK> [106170.955315] ? __warn+0x79/0x120 [106170.955648] ? mlx5_tc_ct_del_ft_cb+0x267/0x2b0 [mlx5_core] [106170.956172] ? report_bug+0x17c/0x190 [106170.956537] ? handle_bug+0x3c/0x60 [106170.956891] ? exc_invalid_op+0x14/0x70 [106170.957264] ? asm_exc_invalid_op+0x16/0x20 [106170.957666] ? mlx5_del_flow_rules+0x10/0x310 [mlx5_core] [106170.958172] ? mlx5_tc_ct_block_flow_offload_add+0x1240/0x1240 [mlx5_core] [106170.958788] ? mlx5_tc_ct_del_ft_cb+0x267/0x2b0 [mlx5_core] [106170.959339] ? mlx5_tc_ct_del_ft_cb+0xc6/0x2b0 [mlx5_core] [106170.959854] ? mapping_remove+0x154/0x1d0 [mlx5_core] [106170.960342] ? mlx5e_tc_action_miss_mapping_put+0x4f/0x80 [mlx5_core] [106170.960927] mlx5_tc_ct_delete_flow+0x76/0xc0 [mlx5_core] [106170.961441] mlx5_free_flow_attr_actions+0x13b/0x220 [mlx5_core] [106170.962001] mlx5e_tc_del_fdb_flow+0x22c/0x3b0 [mlx5_core] [106170.962524] mlx5e_tc_del_flow+0x95/0x3c0 [mlx5_core] [106170.963034] mlx5e_flow_put+0x73/0xe0 [mlx5_core] [106170.963506] mlx5e_put_flow_list+0x38/0x70 [mlx5_core] [106170.964002] mlx5e_rep_update_flows+0xec/0x290 [mlx5_core] [106170.964525] mlx5e_rep_neigh_update+0x1da/0x310 [mlx5_core] [106170.965056] process_one_work+0x13a/0x2c0 [106170.965443] worker_thread+0x2e5/0x3f0 [106170.965808] ? rescuer_thread+0x410/0x410 [106170.966192] kthread+0xc6/0xf0 [106170.966515] ? kthread_complete_and_exit+0x20/0x20 [106170.966970] ret_from_fork+0x2d/0x50 [106170.967332] ? kthread_complete_and_exit+0x20/0x20 [106170.967774] ret_from_fork_asm+0x11/0x20 [106170.970466] </TASK> [106170.970726] ---[ end trace 0000000000000000 ]--- Fixes: 77ac5e40 ("net/sched: act_ct: remove and free nf_table callbacks") Signed-off-by:
Paul Blakey <paulb@nvidia.com> Acked-by:
Pablo Neira Ayuso <pablo@netfilter.org> Signed-off-by:
-
- 10 Dec, 2023 2 commits
-
-
Zhipeng Lu authored
The rvu_dl will be freed in rvu_nix_health_reporters_destroy(rvu_dl) after the create_workqueue fails, and after that free, the rvu_dl will be translate back through the following call chain: rvu_nix_health_reporters_destroy |-> rvu_nix_health_reporters_create |-> rvu_health_reporters_create |-> rvu_register_dl (label err_dl_health) Finally. in the err_dl_health label, rvu_dl being freed again in rvu_health_reporters_destroy(rvu) by rvu_nix_health_reporters_destroy. In the second calls of rvu_nix_health_reporters_destroy, however, it uses rvu_dl->rvu_nix_health_reporter, which is already freed at the end of rvu_nix_health_reporters_destroy in the first call. So this patch prevents the first destroy by instantly returning -ENONMEN when create_workqueue fails. In addition, since the failure of create_workqueue is the only entrence of label err, it has been integrated into the error-handling path of create_workqueue. Fixes: 5ed66306 ("octeontx2-af: Add devlink health reporters for NIX") Signed-off-by:Zhipeng Lu <alexious@zju.edu.cn> Signed-off-by:
-
Radu Bulie authored
The old implementation extracted VLAN TCI info from the payload before the VLAN tag has been pushed in the payload. Another problem was that the VLAN TCI was extracted even if the packet did not have VLAN protocol header. This resulted in invalid VLAN TCI and as a consequence a random queue was computed. This patch fixes the above issues and use the VLAN TCI from the skb if it is present or VLAN TCI from payload if present. If no VLAN header is present queue 0 is selected. Fixes: 52c4a1a8 ("net: fec: add ndo_select_queue to fix TX bandwidth fluctuations") Signed-off-by:
Radu Bulie <radu-andrei.bulie@nxp.com> Signed-off-by:
Wei Fang <wei.fang@nxp.com> Signed-off-by:
-
- 09 Dec, 2023 9 commits
-
-
Pavel Begunkov authored
File reference cycles have caused lots of problems for io_uring in the past, and it still doesn't work exactly right and races with unix_stream_read_generic(). The safest fix would be to completely disallow sending io_uring files via sockets via SCM_RIGHT, so there are no possible cycles invloving registered files and thus rendering SCM accounting on the io_uring side unnecessary. Cc: stable@vger.kernel.org Fixes: 0091bfc8 ("io_uring/af_unix: defer registered files gc to io_uring release") Reported-and-suggested-by:
Jann Horn <jannh@google.com> Signed-off-by:
Pavel Begunkov <asml.silence@gmail.com> Acked-by:
-
Chengfeng Ye authored
As &card->tx_queue_lock is acquired under softirq context along the following call chain from solos_bh(), other acquisition of the same lock inside process context should disable at least bh to avoid double lock. <deadlock #2> pclose() --> spin_lock(&card->tx_queue_lock) <interrupt> --> solos_bh() --> fpga_tx() --> spin_lock(&card->tx_queue_lock) This flaw was found by an experimental static analysis tool I am developing for irq-related deadlock. To prevent the potential deadlock, the patch uses spin_lock_bh() on &card->tx_queue_lock under process context code consistently to prevent the possible deadlock scenario. Fixes: 213e85d3 ("solos-pci: clean up pclose() function") Signed-off-by:
Chengfeng Ye <dg573847474@gmail.com> Signed-off-by:
-
Chengfeng Ye authored
As &card->cli_queue_lock is acquired under softirq context along the following call chain from solos_bh(), other acquisition of the same lock inside process context should disable at least bh to avoid double lock. <deadlock #1> console_show() --> spin_lock(&card->cli_queue_lock) <interrupt> --> solos_bh() --> spin_lock(&card->cli_queue_lock) This flaw was found by an experimental static analysis tool I am developing for irq-related deadlock. To prevent the potential deadlock, the patch uses spin_lock_bh() on the card->cli_queue_lock under process context code consistently to prevent the possible deadlock scenario. Fixes: 9c54004e ("atm: Driver for Solos PCI ADSL2+ card.") Signed-off-by:
-
Jakub Kicinski authored
Michael Chan says: ==================== bnxt_en: Misc. fixes 4 miscellaneous driver fixes covering PM resume, SKB recycling, wrong return value check, and PTP HWTSTAMP_FILTER_ALL. ==================== Link: https://lore.kernel.org/r/20231208001658.14230-1-michael.chan@broadcom.comSigned-off-by:
-
Michael Chan authored
When the chip is configured to timestamp all receive packets, the timestamp in the RX completion is only valid if the metadata present flag is not set for packets received on the wire. In addition, internal loopback packets will never have a valid timestamp and the timestamp field will always be zero. We must exclude any 0 value in the timestamp field because there is no way to determine if it is a loopback packet or not. Add a new function bnxt_rx_ts_valid() to check for all timestamp valid conditions. Fixes: 66ed81dc ("bnxt_en: Enable packet timestamping for all RX packets") Reviewed-by:
Andy Gospodarek <andrew.gospodarek@broadcom.com> Reviewed-by:
Pavan Chebbi <pavan.chebbi@broadcom.com> Signed-off-by:
Michael Chan <michael.chan@broadcom.com> Link: https://lore.kernel.org/r/20231208001658.14230-5-michael.chan@broadcom.comSigned-off-by:
-
Kalesh AP authored
The wait_event_interruptible_timeout() function returns 0 if the timeout elapsed, -ERESTARTSYS if it was interrupted by a signal, and the remaining jiffies otherwise if the condition evaluated to true before the timeout elapsed. Driver should have checked for zero return value instead of a positive value. MChan: Print a warning for -ERESTARTSYS. The close operation will proceed anyway when wait_event_interruptible_timeout() returns for any reason. Since we do the close no matter what, we should not return this error code to the caller. Change bnxt_close_nic() to a void function and remove all error handling from some of the callers. Fixes: c0c050c5 ("bnxt_en: New Broadcom ethernet driver.") Reviewed-by:
Vikas Gupta <vikas.gupta@broadcom.com> Reviewed-by:
Somnath Kotur <somnath.kotur@broadcom.com> Signed-off-by:
Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by:
-
Sreekanth Reddy authored
Receive SKBs can go through the VF-rep path or the normal path. skb_mark_for_recycle() is only called for the normal path. Fix it to do it for both paths to fix possible stalled page pool shutdown errors. Fixes: 86b05508 ("bnxt_en: Use the unified RX page pool buffers for XDP and non-XDP") Reviewed-by:
Sreekanth Reddy <sreekanth.reddy@broadcom.com> Signed-off-by:
-
Somnath Kotur authored
We are issuing HWRM_FUNC_RESET cmd to reset the device including all reserved resources, but not clearing the reservations within the driver struct. As a result, when the driver re-initializes as part of resume, it believes that there is no need to do any resource reservation and goes ahead and tries to allocate rings which will eventually fail beyond a certain number pre-reserved by the firmware. Fixes: 674f50a5 ("bnxt_en: Implement new method to reserve rings.") Reviewed-by:
Ajit Khaparde <ajit.khaparde@broadcom.com> Reviewed-by:
-
Eric Dumazet authored
After commit 93946301 ("tcp: change data receiver flowlabel after one dup") we noticed an increase of TCPACKSkippedPAWS events. Neal Cardwell tracked the issue to tcp_disordered_ack() assumption about remote peer TS clock. RFC 1323 & 7323 are suggesting the following: "timestamp clock frequency in the range 1 ms to 1 sec per tick between 1ms and 1sec." This has to be adjusted for 1 MHz clock frequency. This hints at reorders of SACK packets on send side, this might deserve a future patch. (skb->ooo_okay is always set for pure ACK packets) Fixes: 614e8316 ("tcp: add support for usec resolution in TCP TS values") Co-developed-by:
Neal Cardwell <ncardwell@google.com> Signed-off-by:
-
