1. 16 Mar, 2012 5 commits
  2. 15 Mar, 2012 19 commits
  3. 14 Mar, 2012 11 commits
  4. 13 Mar, 2012 5 commits
    • Nicholas Bellinger's avatar
      iscsi-target: Fix reservation conflict -EBUSY response handling bug · 00fdc6bb
      Nicholas Bellinger authored
      This patch addresses a iscsi-target specific bug related to reservation conflict
      handling in iscsit_handle_scsi_cmd() that has been causing reservation conflicts
      to complete and not fail as expected due to incorrect errno checking.  The problem
      occured with the change to return -EBUSY from transport_generic_cmd_sequencer() ->
      transport_generic_allocate_tasks() failures, that broke iscsit_handle_scsi_cmd()
      checking for -EINVAL in order to invoke a non GOOD status response.
      This was manifesting itself as data corruption with legacy SPC-2 reservations,
      but also effects iscsi-target LUNs with SPC-3 persistent reservations.
      This bug was originally introduced in lio-core commit:
      commit 03e98c9e
      Author: Nicholas Bellinger <nab@linux-iscsi.org>
      Date:   Fri Nov 4 02:36:16 2011 -0700
          target: Address legacy PYX_TRANSPORT_* return code breakage
      Reported-by: default avatarMartin Svec <martin.svec@zoner.cz>
      Cc: Martin Svec <martin.svec@zoner.cz>
      Cc: Christoph Hellwig <hch@lst.de>
      Cc: stable@vger.kernel.org
      Signed-off-by: default avatarNicholas Bellinger <nab@linux-iscsi.org>
    • Nicholas Bellinger's avatar
      target: Fix compatible reservation handling (CRH=1) with legacy RESERVE/RELEASE · 087a03b3
      Nicholas Bellinger authored
      This patch addresses a bug with target_check_scsi2_reservation_conflict()
      return checking in target_scsi2_reservation_[reserve,release]() that was
      preventing CRH=1 operation from silently succeeding in the two special
      cases defined by SPC-3, and not failing with reservation conflict status
      when dealing with legacy RESERVE/RELEASE + active SPC-3 PR logic.
      Also explictly set cmd->scsi_status = SAM_STAT_RESERVATION_CONFLICT during
      the early non reservation holder failure from pr_ops->t10_seq_non_holder()
      check in transport_generic_cmd_sequencer() for fabrics that already expect
      it to be set.
      This bug was originally introduced in mainline commit:
      commit eacac00c
      Author: Christoph Hellwig <hch@infradead.org>
      Date:   Thu Nov 3 17:50:40 2011 -0400
          target: split core_scsi2_emulate_crh
      Reported-by: default avatarMartin Svec <martin.svec@zoner.cz>
      Cc: Martin Svec <martin.svec@zoner.cz>
      Cc: Christoph Hellwig <hch@lst.de>
      Cc: stable@vger.kernel.org
      Signed-off-by: default avatarNicholas Bellinger <nab@linux-iscsi.org>
    • Linus Torvalds's avatar
      Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc · 762ad8a5
      Linus Torvalds authored
      Pull sparc updates from David Miller:
       "Please pull to get this fix for the sparc32 build when using a more
        recent binutils."
      * git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc:
        sparc32: Add -Av8 to assembler command line.
    • David S. Miller's avatar
      sparc32: Add -Av8 to assembler command line. · e0adb990
      David S. Miller authored
      Newer version of binutils are more strict about specifying the
      correct options to enable certain classes of instructions.
      The sparc32 build is done for v7 in order to support sun4c systems
      which lack hardware integer multiply and divide instructions.
      So we have to pass -Av8 when building the assembler routines that
      use these instructions and get patched into the kernel when we find
      out that we have a v8 capable cpu.
      Reported-by: default avatarPaul Gortmaker <paul.gortmaker@windriver.com>
      Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
    • Linus Torvalds's avatar
      Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net · b8fa7d41
      Linus Torvalds authored
      Pull networking changes from David Miller:
       "The most important bit here is the TCP syncookies issue, which seems
        to have been busted for some time.  That fix has been verified in
        production by the reporter.
        1) Persistent TUN devices erroneously hold on to the network namespace
           in such a way that it cannot be shutdown.  Fix from Stanislav
           Kinsbursky with help from Eric Dumazet.
        2) TCP SYN cookies have been broken for a while due to how the route
           lookup flow key is managed, connections can be delayed by as much
           as 20 seconds due to this bug.  Fix from Eric Dumazet.
        3) Missing jiffies.h include in lib/dynamic_queue_limits.c can break
           the build, from Tom Herbert.
        4) Add USB device ID for Sitecom LN-031, from Joerg Neikes.
        5) Fix OOPS in delayed workqueue in iwlegacy, from Stanislaw Gruszka.
        6) rt2x00 TX queue can be disabled forever due to races, fix by
           synchronizing pause/unpause with a lock.  Also from Stanislaw
        7) Statistics and endian fix in bnx2x driver from Yuval Mintz, Eilon
           Greenstein, and Ariel Elior."
      * git://git.kernel.org/pub/scm/linux/kernel/git/davem/net:
        tun: don't hold network namespace by tun sockets
        bnx2x: FCoE statistics id fixed
        bnx2x: dcb bit indices flags used as bits
        bnx2x: added cpu_to_le16 when preparing ramrod's data
        bnx2x: pfc statistics counts pfc events twice
        rt2x00: fix random stalls
        iwl3945: fix possible il->txq NULL pointer dereference in delayed works
        dql: Fix undefined jiffies
        tcp: fix syncookie regression
        usb: asix: Patch for Sitecom LN-031