1. 07 Jan, 2020 3 commits
    • Chuck Lever's avatar
      RDMA/core: Add trace points to follow MR allocation · 622db5b6
      Chuck Lever authored
      Track the lifetime of ib_mr objects. Here's sample output from a test run
      with NFS/RDMA:
      
                 <...>-361   [009] 79238.772782: mr_alloc:             pd.id=3 mr.id=11 type=MEM_REG max_num_sg=30 rc=0
                 <...>-361   [009] 79238.772812: mr_alloc:             pd.id=3 mr.id=12 type=MEM_REG max_num_sg=30 rc=0
                 <...>-361   [009] 79238.772839: mr_alloc:             pd.id=3 mr.id=13 type=MEM_REG max_num_sg=30 rc=0
                 <...>-361   [009] 79238.772866: mr_alloc:             pd.id=3 mr.id=14 type=MEM_REG max_num_sg=30 rc=0
                 <...>-361   [009] 79238.772893: mr_alloc:             pd.id=3 mr.id=15 type=MEM_REG max_num_sg=30 rc=0
                 <...>-361   [009] 79238.772921: mr_alloc:             pd.id=3 mr.id=16 type=MEM_REG max_num_sg=30 rc=0
                 <...>-361   [009] 79238.772947: mr_alloc:             pd.id=3 mr.id=17 type=MEM_REG max_num_sg=30 rc=0
                 <...>-361   [009] 79238.772974: mr_alloc:             pd.id=3 mr.id=18 type=MEM_REG max_num_sg=30 rc=0
                 <...>-361   [009] 79238.773001: mr_alloc:             pd.id=3 mr.id=19 type=MEM_REG max_num_sg=30 rc=0
                 <...>-361   [009] 79238.773028: mr_alloc:             pd.id=3 mr.id=20 type=MEM_REG max_num_sg=30 rc=0
                 <...>-361   [009] 79238.773055: mr_alloc:             pd.id=3 mr.id=21 type=MEM_REG max_num_sg=30 rc=0
                 <...>-361   [009] 79240.270942: mr_alloc:             pd.id=3 mr.id=22 type=MEM_REG max_num_sg=30 rc=0
                 <...>-361   [009] 79240.270975: mr_alloc:             pd.id=3 mr.id=23 type=MEM_REG max_num_sg=30 rc=0
                 <...>-361   [009] 79240.271007: mr_alloc:             pd.id=3 mr.id=24 type=MEM_REG max_num_sg=30 rc=0
                 <...>-361   [009] 79240.271036: mr_alloc:             pd.id=3 mr.id=25 type=MEM_REG max_num_sg=30 rc=0
                 <...>-361   [009] 79240.271067: mr_alloc:             pd.id=3 mr.id=26 type=MEM_REG max_num_sg=30 rc=0
                 <...>-361   [009] 79240.271095: mr_alloc:             pd.id=3 mr.id=27 type=MEM_REG max_num_sg=30 rc=0
                 <...>-361   [009] 79240.271121: mr_alloc:             pd.id=3 mr.id=28 type=MEM_REG max_num_sg=30 rc=0
                 <...>-361   [009] 79240.271153: mr_alloc:             pd.id=3 mr.id=29 type=MEM_REG max_num_sg=30 rc=0
                 <...>-361   [009] 79240.271181: mr_alloc:             pd.id=3 mr.id=30 type=MEM_REG max_num_sg=30 rc=0
                 <...>-361   [009] 79240.271208: mr_alloc:             pd.id=3 mr.id=31 type=MEM_REG max_num_sg=30 rc=0
                 <...>-361   [009] 79240.271236: mr_alloc:             pd.id=3 mr.id=32 type=MEM_REG max_num_sg=30 rc=0
                 <...>-4351  [001] 79242.299400: mr_dereg:             mr.id=32
                 <...>-4351  [001] 79242.299467: mr_dereg:             mr.id=31
                 <...>-4351  [001] 79242.299554: mr_dereg:             mr.id=30
                 <...>-4351  [001] 79242.299615: mr_dereg:             mr.id=29
                 <...>-4351  [001] 79242.299684: mr_dereg:             mr.id=28
                 <...>-4351  [001] 79242.299748: mr_dereg:             mr.id=27
                 <...>-4351  [001] 79242.299812: mr_dereg:             mr.id=26
                 <...>-4351  [001] 79242.299874: mr_dereg:             mr.id=25
                 <...>-4351  [001] 79242.299944: mr_dereg:             mr.id=24
                 <...>-4351  [001] 79242.300009: mr_dereg:             mr.id=23
                 <...>-4351  [001] 79242.300190: mr_dereg:             mr.id=22
                 <...>-4351  [001] 79242.300263: mr_dereg:             mr.id=21
                 <...>-4351  [001] 79242.300326: mr_dereg:             mr.id=20
                 <...>-4351  [001] 79242.300388: mr_dereg:             mr.id=19
                 <...>-4351  [001] 79242.300450: mr_dereg:             mr.id=18
                 <...>-4351  [001] 79242.300516: mr_dereg:             mr.id=17
                 <...>-4351  [001] 79242.300629: mr_dereg:             mr.id=16
                 <...>-4351  [001] 79242.300718: mr_dereg:             mr.id=15
                 <...>-4351  [001] 79242.300784: mr_dereg:             mr.id=14
                 <...>-4351  [001] 79242.300879: mr_dereg:             mr.id=13
                 <...>-4351  [001] 79242.300945: mr_dereg:             mr.id=12
                 <...>-4351  [001] 79242.301012: mr_dereg:             mr.id=11
      
      Some features of the output:
      - The lifetime and owner PD of each MR is clearly visible.
      - The type of MR is captured, as is the SGE array size.
      - Failing MR allocation can be recorded.
      
      Link: https://lore.kernel.org/r/20191218201820.30584.34636.stgit@manet.1015granger.netSigned-off-by: default avatarChuck Lever <chuck.lever@oracle.com>
      Signed-off-by: default avatarJason Gunthorpe <jgg@mellanox.com>
      622db5b6
    • Chuck Lever's avatar
      RDMA/core: Trace points for diagnosing completion queue issues · 3e5901cb
      Chuck Lever authored
      Sample trace events:
      
         kworker/u29:0-300   [007]   120.042217: cq_alloc:             cq.id=4 nr_cqe=161 comp_vector=2 poll_ctx=WORKQUEUE
                <idle>-0     [002]   120.056292: cq_schedule:          cq.id=4
          kworker/2:1H-482   [002]   120.056402: cq_process:           cq.id=4 wake-up took 109 [us] from interrupt
          kworker/2:1H-482   [002]   120.056407: cq_poll:              cq.id=4 requested 16, returned 1
                <idle>-0     [002]   120.067503: cq_schedule:          cq.id=4
          kworker/2:1H-482   [002]   120.067537: cq_process:           cq.id=4 wake-up took 34 [us] from interrupt
          kworker/2:1H-482   [002]   120.067541: cq_poll:              cq.id=4 requested 16, returned 1
                <idle>-0     [002]   120.067657: cq_schedule:          cq.id=4
          kworker/2:1H-482   [002]   120.067672: cq_process:           cq.id=4 wake-up took 15 [us] from interrupt
          kworker/2:1H-482   [002]   120.067674: cq_poll:              cq.id=4 requested 16, returned 1
      
       ...
      
               systemd-1     [002]   122.392653: cq_schedule:          cq.id=4
          kworker/2:1H-482   [002]   122.392688: cq_process:           cq.id=4 wake-up took 35 [us] from interrupt
          kworker/2:1H-482   [002]   122.392693: cq_poll:              cq.id=4 requested 16, returned 16
          kworker/2:1H-482   [002]   122.392836: cq_poll:              cq.id=4 requested 16, returned 16
          kworker/2:1H-482   [002]   122.392970: cq_poll:              cq.id=4 requested 16, returned 16
          kworker/2:1H-482   [002]   122.393083: cq_poll:              cq.id=4 requested 16, returned 16
          kworker/2:1H-482   [002]   122.393195: cq_poll:              cq.id=4 requested 16, returned 3
      
      Several features to note in this output:
       - The WCE count and context type are reported at allocation time
       - The CPU and kworker for each CQ is evident
       - The CQ's restracker ID is tagged on each trace event
       - CQ poll scheduling latency is measured
       - Details about how often single completions occur versus multiple
         completions are evident
       - The cost of the ULP's completion handler is recorded
      
      Link: https://lore.kernel.org/r/20191218201815.30584.3481.stgit@manet.1015granger.netSigned-off-by: default avatarChuck Lever <chuck.lever@oracle.com>
      Reviewed-by: default avatarParav Pandit <parav@mellanox.com>
      Signed-off-by: default avatarJason Gunthorpe <jgg@mellanox.com>
      3e5901cb
    • Chuck Lever's avatar
      RDMA/cma: Add trace points in RDMA Connection Manager · ed999f82
      Chuck Lever authored
      Record state transitions as each connection is established. The IP address
      of both peers and the Type of Service is reported. These trace points are
      not in performance hot paths.
      
      Also, record each cm_event_handler call to ULPs. This eliminates the need
      for each ULP to add its own similar trace point in its CM event handler
      function.
      
      These new trace points appear in a new trace subsystem called "rdma_cma".
      
      Sample events:
      
                 <...>-220   [004]   121.430733: cm_id_create:         cm.id=0
                 <...>-472   [003]   121.430991: cm_event_handler:     cm.id=0 src=192.168.2.51:35090 dst=192.168.2.55:20049 tos=0 ADDR_RESOLVED (0/0)
                 <...>-472   [003]   121.430995: cm_event_done:        cm.id=0 src=192.168.2.51:35090 dst=192.168.2.55:20049 tos=0 result=0
                 <...>-472   [003]   121.431172: cm_event_handler:     cm.id=0 src=192.168.2.51:35090 dst=192.168.2.55:20049 tos=0 ROUTE_RESOLVED (2/0)
                 <...>-472   [003]   121.431174: cm_event_done:        cm.id=0 src=192.168.2.51:35090 dst=192.168.2.55:20049 tos=0 result=0
                 <...>-220   [004]   121.433480: cm_qp_create:         cm.id=0 src=192.168.2.51:35090 dst=192.168.2.55:20049 tos=0 pd.id=2 qp_type=RC send_wr=4091 recv_wr=256 qp_num=521 rc=0
                 <...>-220   [004]   121.433577: cm_send_req:          cm.id=0 src=192.168.2.51:35090 dst=192.168.2.55:20049 tos=0 qp_num=521
           kworker/1:2-973   [001]   121.436190: cm_send_mra:          cm.id=0 src=192.168.2.51:35090 dst=192.168.2.55:20049 tos=0
           kworker/1:2-973   [001]   121.436340: cm_send_rtu:          cm.id=0 src=192.168.2.51:35090 dst=192.168.2.55:20049 tos=0
           kworker/1:2-973   [001]   121.436359: cm_event_handler:     cm.id=0 src=192.168.2.51:35090 dst=192.168.2.55:20049 tos=0 ESTABLISHED (9/0)
           kworker/1:2-973   [001]   121.436365: cm_event_done:        cm.id=0 src=192.168.2.51:35090 dst=192.168.2.55:20049 tos=0 result=0
                 <...>-1975  [005]   123.161954: cm_disconnect:        cm.id=0 src=192.168.2.51:35090 dst=192.168.2.55:20049 tos=0
                 <...>-1975  [005]   123.161974: cm_sent_dreq:         cm.id=0 src=192.168.2.51:35090 dst=192.168.2.55:20049 tos=0
                 <...>-220   [004]   123.162102: cm_disconnect:        cm.id=0 src=192.168.2.51:35090 dst=192.168.2.55:20049 tos=0
           kworker/0:1-13    [000]   123.162391: cm_event_handler:     cm.id=0 src=192.168.2.51:35090 dst=192.168.2.55:20049 tos=0 DISCONNECTED (10/0)
           kworker/0:1-13    [000]   123.162393: cm_event_done:        cm.id=0 src=192.168.2.51:35090 dst=192.168.2.55:20049 tos=0 result=0
                 <...>-220   [004]   123.164456: cm_qp_destroy:        cm.id=0 src=192.168.2.51:35090 dst=192.168.2.55:20049 tos=0 qp_num=521
                 <...>-220   [004]   123.165290: cm_id_destroy:        cm.id=0 src=192.168.2.51:35090 dst=192.168.2.55:20049 tos=0
      
      Some features to note:
      - restracker ID of the rdma_cm_id is tagged on each trace event
      - The source and destination IP addresses and TOS are reported
      - CM event upcalls are shown with decoded event and status
      - CM state transitions are reported
      - rdma_cm_id lifetime events are captured
      - The latency of ULP CM event handlers is reported
      - Lifetime events of associated QPs are reported
      - Device removal and insertion is reported
      
      This patch is based on previous work by:
      
      Saeed Mahameed <saeedm@mellanox.com>
      Mukesh Kacker <mukesh.kacker@oracle.com>
      Ajaykumar Hotchandani <ajaykumar.hotchandani@oracle.com>
      Aron Silverton <aron.silverton@oracle.com>
      Avinash Repaka <avinash.repaka@oracle.com>
      Somasundaram Krishnasamy <somasundaram.krishnasamy@oracle.com>
      
      Link: https://lore.kernel.org/r/20191218201810.30584.3052.stgit@manet.1015granger.netSigned-off-by: default avatarChuck Lever <chuck.lever@oracle.com>
      Signed-off-by: default avatarJason Gunthorpe <jgg@mellanox.com>
      ed999f82
  2. 04 Jan, 2020 2 commits
  3. 03 Jan, 2020 22 commits
  4. 29 Dec, 2019 5 commits
  5. 28 Dec, 2019 4 commits
  6. 27 Dec, 2019 4 commits
    • Linus Torvalds's avatar
      Merge tag 'drm-fixes-2019-12-28' of git://anongit.freedesktop.org/drm/drm · 48a8dd17
      Linus Torvalds authored
      Pull drm fixes from Dave Airlie:
       "Post-xmas food coma recovery fixes. Only three fixes for i915 since I
        expect most people are holidaying.
      
        i915:
         - power management rc6 fix
         - framebuffer tracking fix
         - display power management ratelimit fix"
      
      * tag 'drm-fixes-2019-12-28' of git://anongit.freedesktop.org/drm/drm:
        drm/i915: Hold reference to intel_frontbuffer as we track activity
        drm/i915/gt: Ratelimit display power w/a
        drm/i915/pmu: Ensure monotonic rc6
      48a8dd17
    • Linus Torvalds's avatar
      Merge tag 'linux-kselftest-5.5-rc4' of... · f4b39746
      Linus Torvalds authored
      Merge tag 'linux-kselftest-5.5-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest
      
      Pull Kselftest fixes from Shuah Khan:
      
       - rseq build failures fixes related to glibc 2.30 compatibility from
         Mathieu Desnoyers
      
       - Kunit fixes and cleanups from SeongJae Park
      
       - Fixes to filesystems/epoll, firmware, and livepatch build failures
         and skip handling.
      
      * tag 'linux-kselftest-5.5-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest:
        rseq/selftests: Clarify rseq_prepare_unload() helper requirements
        rseq/selftests: Fix: Namespace gettid() for compatibility with glibc 2.30
        rseq/selftests: Turn off timeout setting
        kunit/kunit_tool_test: Test '--build_dir' option run
        kunit: Rename 'kunitconfig' to '.kunitconfig'
        kunit: Place 'test.log' under the 'build_dir'
        kunit: Create default config in '--build_dir'
        kunit: Remove duplicated defconfig creation
        docs/kunit/start: Use in-tree 'kunit_defconfig'
        selftests: livepatch: Fix it to do root uid check and skip
        selftests: firmware: Fix it to do root uid check and skip
        selftests: filesystems/epoll: fix build error
      f4b39746
    • Linus Torvalds's avatar
      Merge tag 'pm-5.5-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm · 1413c361
      Linus Torvalds authored
      Pull power management fixes from Rafael Wysocki:
       "Fix compile test of the Tegra devfreq driver (Arnd Bergmann) and
        remove redundant Kconfig dependencies from multiple devfreq drivers
        (Leonard Crestez)"
      
      * tag 'pm-5.5-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
        PM / devfreq: tegra: Add COMMON_CLK dependency
        PM / devfreq: Drop explicit selection of PM_OPP
      1413c361
    • Linus Torvalds's avatar
      Merge tag 'io_uring-5.5-20191226' of git://git.kernel.dk/linux-block · 534121d2
      Linus Torvalds authored
      Pull io_uring fixes from Jens Axboe:
      
       - Removal of now unused busy wqe list (Hillf)
      
       - Add cond_resched() to io-wq work processing (Hillf)
      
       - And then the series that I hinted at from last week, which removes
         the sqe from the io_kiocb and keeps all sqe handling on the prep
         side. This guarantees that an opcode can't do the wrong thing and
         read the sqe more than once. This is unchanged from last week, no
         issues have been observed with this in testing. Hence I really think
         we should fold this into 5.5.
      
      * tag 'io_uring-5.5-20191226' of git://git.kernel.dk/linux-block:
        io-wq: add cond_resched() to worker thread
        io-wq: remove unused busy list from io_sqe
        io_uring: pass in 'sqe' to the prep handlers
        io_uring: standardize the prep methods
        io_uring: read 'count' for IORING_OP_TIMEOUT in prep handler
        io_uring: move all prep state for IORING_OP_{SEND,RECV}_MGS to prep handler
        io_uring: move all prep state for IORING_OP_CONNECT to prep handler
        io_uring: add and use struct io_rw for read/writes
        io_uring: use u64_to_user_ptr() consistently
      534121d2