FreeBSD source tree https://cgit-dev.freebsd.org/src/atom?h=release%2F13.2.0 2023-02-02T07:18:51Z 2023-02-02T07:18:51Z Navdeep Parhar np@FreeBSD.org 2022-09-09T22:16:22Z urn:sha1:5adfcaeacafa73522cf98403428e7785a7aecc82 hw.cxgbe.cong_drop=2 will generate backpressure *and* drop frames for queues that are congested. Sponsored by: Chelsio Communications (cherry picked from commit df275ae5090048fb7a8c082d509fae57e891957a) 2023-02-02T07:18:45Z Navdeep Parhar np@FreeBSD.org 2022-09-08T18:47:38Z urn:sha1:af0f5fe24c9b16341ee9c981d572150b74c41fa9 The firmware takes the type into account when setting up the PCIe channel for the queue. Sponsored by: Chelsio Communications (cherry picked from commit c387ff0045d620d60ed8250360d4196a4c962a13) 2023-01-24T05:11:47Z John Baldwin jhb@FreeBSD.org 2022-11-15T20:03:57Z urn:sha1:b0cb849d7f629408dd6ed532e26db3f4cf076131 Mostly trailing whitespace and spaces before tabs. Sponsored by: Chelsio Communications Differential Revision: https://reviews.freebsd.org/D37350 (cherry picked from commit 21186bdb2d6a14dbfff7b4366425285b8612cc22) 2021-10-20T13:55:38Z Navdeep Parhar np@FreeBSD.org 2021-04-26T21:03:06Z urn:sha1:8ec8f56ce25b0bcb95bba5ae52d284bcbd88617e The driver uses both software resources (locks, callouts, memory for descriptors and for bookkeeping, sysctls, etc.) and hardware resources (VIs, DMA queues, TCAM entries, etc.) to operate the NIC. This commit splits the single *_ALLOCATED flag used to track all these resources into separate *_SW_ALLOCATED and *_HW_ALLOCATED flags. This is the simplified pseudocode that now applies to most queues (foo can be ctrlq/txq/rxq/ofld_txq/ofld_rxq): /* Idempotent */ alloc_foo { if (!SW_ALLOCATED) init_iq/init_eq/init_fl no-fail sw init alloc_iq_fl/alloc_eq/alloc_wrq may-fail sw alloc add_foo_sysctls, etc. no-fail post-alloc items if (!HW_ALLOCATED) alloc_iq_fl_hwq/alloc_eq_hwq hw resource allocation } /* Idempotent */ free_foo { if (!HW_ALLOCATED) free_iq_fl_hwq/free_eq_hwq release hw resources if (!SW_ALLOCATED) free_iq_fl/free_eq/free_wrq release sw resources } The routines that take the driver to FULL_INIT_DONE and VI_INIT_DONE and back are now all idempotent. The quiesce routines pay attention to the HW_ALLOCATED flag and will not wait on the hardware for pidx/cidx updates and other completions if this flag is not set. Sponsored by: Chelsio Communications (cherry picked from commit 43bbae19483fbde0a91e61acad8a6e71e334c8b8) 2021-05-16T03:06:35Z Navdeep Parhar np@FreeBSD.org 2021-03-19T17:56:24Z urn:sha1:564d1f6aef5e72a94e53f70b727c675481086622 These routines to handle 8b and 16b types were added in r289773 5+ years ago. Sponsored by: Chelsio Communications (cherry picked from commit 473f6163e310b773dfd7e500e255d01d7328dd16) 2020-12-03T08:30:29Z Navdeep Parhar np@FreeBSD.org 2020-12-03T08:30:29Z urn:sha1:8eba75ed68854188c01a2058fe1f826b3428dbfa It is common for freelists to be starving when a netmap application stops. Mailbox commands to free queues can hang in such a situation. Avoid that by not freeing the queues when netmap is switched off. Instead, use an alternate method to stop the queues without releasing the context ids. If netmap is enabled again later then the same queue is reinitialized for use. Move alloc_nm_rxq and txq to t4_netmap.c while here. MFC after: 1 week Sponsored by: Chelsio Communications 2020-12-02T20:54:03Z Navdeep Parhar np@FreeBSD.org 2020-12-02T20:54:03Z urn:sha1:f42f3b2955e5afa770275fd76b4f754fbd67bbeb r367917 fixed the backpressure on the netmap rxq being stopped but that doesn't help if some other netmap rxq is starved (because it is stopping too although the driver doesn't know this yet) and blocks the pipeline. An alternate fix that works in all cases will be checked in instead. Sponsored by: Chelsio Communications 2020-11-21T03:27:32Z Navdeep Parhar np@FreeBSD.org 2020-11-21T03:27:32Z urn:sha1:b3718e2d7ebfe7fc0bb57bd0906aacc28dd755b2 The netmap application using the driver is responsible for replenishing the receive freelists and they may be totally depleted when the application exits. Packets in flight, if any, might block the pipeline in case there aren't enough buffers left in the freelist. Avoid this by filling up the freelists with a driver allocated buffer. MFC after: 1 week Sponsored by: Chelsio Communications 2020-10-22T08:40:25Z Navdeep Parhar np@FreeBSD.org 2020-10-22T08:40:25Z urn:sha1:b20b25e744778ac54d54f5486035e70d6570f4d9 The firmware can allocate ingress and egress context ids anywhere from its configured range. Size the iq/eq maps to match the entire range instead of assuming that the firmware always allocates the first available context id. Reported by: Baptiste Wicht @ Verisign MFC after: 1 week Sponsored by: Chelsio Communications 2020-09-29T09:25:52Z Navdeep Parhar np@FreeBSD.org 2020-09-29T09:25:52Z urn:sha1:822967e7e5b24d5fa980f69cab7b2982e1b6fa42 Bind the netmap tx queues to a special '0xff' scheduling class which makes the firmware skip some processing related to rate limiting on the outgoing traffic. Future firmwares will do this automatically. MFC after: 1 week Sponsored by: Chelsio Communications