1. 04 May, 2015 3 commits
    • shamir rabinovitch's avatar
      net/rds: fix unaligned memory access · c0adf54a
      shamir rabinovitch authored
      rdma_conn_param private data is copied using memcpy after headers such
      as cma_hdr (see cma_resolve_ib_udp as example). so the start of the
      private data is aligned to the end of the structure that come before. if
      this structure end with u32 the meaning is that the start of the private
      data will be 4 bytes aligned. structures that use u8/u16/u32/u64 are
      naturally aligned but in case the structure start is not 8 bytes aligned,
      all u64 members of this structure will not be aligned. to solve this issue
      we must use special macros that allow unaligned access to those
      unaligned members.
      
      Addresses the following kernel log seen when attempting to use RDMA:
      
      Kernel unaligned access at TPC[10507a88] rds_ib_cm_connect_complete+0x1bc/0x1e0 [rds_rdma]
      Acked-by: default avatarChien Yen <chien.yen@oracle.com>
      Signed-off-by: default avatarshamir rabinovitch <shamir.rabinovitch@oracle.com>
      [Minor tweaks for top of tree by:]
      Signed-off-by: default avatarDavid Ahern <david.ahern@oracle.com>
      Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
      c0adf54a
    • Herbert Xu's avatar
      netlink: Remove max_size setting · edac450d
      Herbert Xu authored
      We currently limit the hash table size to 64K which is very bad
      as even 10 years ago it was relatively easy to generate millions
      of sockets.
      
      Since the hash table is naturally limited by memory allocation
      failure, we don't really need an explicit limit so this patch
      removes it.
      Signed-off-by: default avatarHerbert Xu <herbert@gondor.apana.org.au>
      Acked-by: default avatarThomas Graf <tgraf@noironetworks.com>
      Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
      edac450d
    • Eric Dumazet's avatar
      codel: fix maxpacket/mtu confusion · a5d28090
      Eric Dumazet authored
      Under presence of TSO/GSO/GRO packets, codel at low rates can be quite
      useless. In following example, not a single packet was ever dropped,
      while average delay in codel queue is ~100 ms !
      
      qdisc codel 0: parent 1:12 limit 16000p target 5.0ms interval 100.0ms
       Sent 134376498 bytes 88797 pkt (dropped 0, overlimits 0 requeues 0)
       backlog 13626b 3p requeues 0
        count 0 lastcount 0 ldelay 96.9ms drop_next 0us
        maxpacket 9084 ecn_mark 0 drop_overlimit 0
      
      This comes from a confusion of what should be the minimal backlog. It is
      pretty clear it is not 64KB or whatever max GSO packet ever reached the
      qdisc.
      
      codel intent was to use MTU of the device.
      
      After the fix, we finally drop some packets, and rtt/cwnd of my single
      TCP flow are meeting our expectations.
      
      qdisc codel 0: parent 1:12 limit 16000p target 5.0ms interval 100.0ms
       Sent 102798497 bytes 67912 pkt (dropped 1365, overlimits 0 requeues 0)
       backlog 6056b 3p requeues 0
        count 1 lastcount 1 ldelay 36.3ms drop_next 0us
        maxpacket 10598 ecn_mark 0 drop_overlimit 0
      Signed-off-by: default avatarEric Dumazet <edumazet@google.com>
      Cc: Kathleen Nichols <nichols@pollere.com>
      Cc: Dave Taht <dave.taht@gmail.com>
      Cc: Van Jacobson <vanj@google.com>
      Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
      a5d28090
  2. 02 May, 2015 5 commits
  3. 01 May, 2015 3 commits
    • Ilya Dryomov's avatar
      rbd: end I/O the entire obj_request on error · 082a75da
      Ilya Dryomov authored
      When we end I/O struct request with error, we need to pass
      obj_request->length as @nr_bytes so that the entire obj_request worth
      of bytes is completed.  Otherwise block layer ends up confused and we
      trip on
      
          rbd_assert(more ^ (which == img_request->obj_request_count));
      
      in rbd_img_obj_callback() due to more being true no matter what.  We
      already do it in most cases but we are missing some, in particular
      those where we don't even get a chance to submit any obj_requests, due
      to an early -ENOMEM for example.
      
      A number of obj_request->xferred assignments seem to be redundant but
      I haven't touched any of obj_request->xferred stuff to keep this small
      and isolated.
      
      Cc: Alex Elder <elder@linaro.org>
      Cc: stable@vger.kernel.org # 3.10+
      Reported-by: default avatarShawn Edwards <lesser.evil@gmail.com>
      Reviewed-by: default avatarSage Weil <sage@redhat.com>
      Signed-off-by: default avatarIlya Dryomov <idryomov@gmail.com>
      082a75da
    • Linus Torvalds's avatar
      Merge branch 'for-linus-4.1' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs · 64887b68
      Linus Torvalds authored
      Pull btrfs fixes from Chris Mason:
       "A few more btrfs fixes.
      
        These range from corners Filipe found in the new free space cache
        writeback to a grab bag of fixes from the list"
      
      * 'for-linus-4.1' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs:
        Btrfs: btrfs_release_extent_buffer_page didn't free pages of dummy extent
        Btrfs: fill ->last_trans for delayed inode in btrfs_fill_inode.
        btrfs: unlock i_mutex after attempting to delete subvolume during send
        btrfs: check io_ctl_prepare_pages return in __btrfs_write_out_cache
        btrfs: fix race on ENOMEM in alloc_extent_buffer
        btrfs: handle ENOMEM in btrfs_alloc_tree_block
        Btrfs: fix find_free_dev_extent() malfunction in case device tree has hole
        Btrfs: don't check for delalloc_bytes in cache_save_setup
        Btrfs: fix deadlock when starting writeback of bg caches
        Btrfs: fix race between start dirty bg cache writeout and bg deletion
      64887b68
    • Linus Torvalds's avatar
      Merge tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux · 036f351e
      Linus Torvalds authored
      Pull arm64 fixes from Will Deacon:
       "Not too much here, but we've addressed a couple of nasty issues in the
        dma-mapping code as well as adding the halfword and byte variants of
        load_acquire/store_release following on from the CSD locking bug that
        you fixed in the core.
      
         - fix perf devicetree warnings at probe time
      
         - fix memory leak in __dma_free()
      
         - ensure DMA buffers are always zeroed
      
         - show IRQ trigger in /proc/interrupts (for parity with ARM)
      
         - implement byte and halfword access for smp_{load_acquire,store_release}"
      
      * tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux:
        arm64: perf: Fix the pmu node name in warning message
        arm64: perf: don't warn about missing interrupt-affinity property for PPIs
        arm64: add missing PAGE_ALIGN() to __dma_free()
        arm64: dma-mapping: always clear allocated buffers
        ARM64: Enable CONFIG_GENERIC_IRQ_SHOW_LEVEL
        arm64: add missing data types in smp_load_acquire/smp_store_release
      036f351e
  4. 30 Apr, 2015 19 commits
  5. 29 Apr, 2015 10 commits