1. 29 May, 2008 12 commits
    • Linus Torvalds's avatar
      Merge branch 'fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/davej/cpufreq · 4bd27972
      Linus Torvalds authored
      * 'fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/davej/cpufreq:
        [CPUFREQ] fix double unlock of cpu_policy_rwsem in drivers/cpufreq/cpufreq.c
      4bd27972
    • Linus Torvalds's avatar
      Merge branch 'sched-fixes-for-linus' of... · a7f75d3b
      Linus Torvalds authored
      Merge branch 'sched-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip
      
      * 'sched-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
        sched: re-tune NUMA topologies
        sched: stop wake_affine from causing serious imbalance
        sched: fix sched_clock_cpu()
        revert ("sched: fair-group: SMP-nice for group scheduling")
        sched: cleanup
        show_schedstat(): fix memleak
        sched: unite unlikely pairs in rt_policy() and schedule_debug()
        revert ("sched: fair: weight calculations")
      a7f75d3b
    • Lothar Waßmann's avatar
      [CPUFREQ] fix double unlock of cpu_policy_rwsem in drivers/cpufreq/cpufreq.c · dca02613
      Lothar Waßmann authored
      In drivers/cpufreq/cpufreq.c the function cpufreq_add_dev() takes the
      error exit 'err_out_unregister' from different places once with the
      'cpu_policy_rwsem' lock held, once with the lock released:
      |		if (ret)
      |			goto err_out_unregister;
      |	}
      |
      |	policy->governor = NULL; /* to assure that the starting sequence is
      |				  * run in cpufreq_set_policy */
      |
      |	/* set default policy */
      |	ret = __cpufreq_set_policy(policy, &new_policy);
      |	policy->user_policy.policy = policy->policy;
      |	policy->user_policy.governor = policy->governor;
      |
      |	unlock_policy_rwsem_write(cpu);
      |
      |	if (ret) {
      |		dprintk("setting policy failed\n");
      |		goto err_out_unregister;
      |	}
      
      This leads to the following error message in case of a failing
      __cpufreq_set_policy() call:
      =====================================
      [ BUG: bad unlock balance detected! ]
      -------------------------------------
      swapper/1 is trying to release lock (&per_cpu(cpu_policy_rwsem, cpu)) at:
      [<c01b4564>] unlock_policy_rwsem_write+0x30/0x40
      but there are no more locks to release!
      
      other info that might help us debug this:
      1 lock held by swapper/1:
       #0:  (sysdev_drivers_lock){--..}, at: [<c018fd18>] sysdev_driver_register+0x74/0x130
      
      stack backtrace:
      [<c002f588>] (dump_stack+0x0/0x14) from [<c00692fc>] (print_unlock_inbalance_bug+0xc8/0x104)
      [<c0069234>] (print_unlock_inbalance_bug+0x0/0x104) from [<c006b7ac>] (lock_release_non_nested+0xc4/0x19c)
       r6:00000028 r5:c3c1ab80 r4:c01b4564
      [<c006b6e8>] (lock_release_non_nested+0x0/0x19c) from [<c006b9e0>] (lock_release+0x15c/0x18c)
       r8:60000013 r7:00000001 r6:c01b4564 r5:c0541bb4 r4:c3c1ab80
      [<c006b884>] (lock_release+0x0/0x18c) from [<c0061ba0>] (up_write+0x24/0x30)
       r8:c0541b80 r7:00000000 r6:ffffffea r5:c3c34828 r4:c0541b8c
      [<c0061b7c>] (up_write+0x0/0x30) from [<c01b4564>] (unlock_policy_rwsem_write+0x30/0x40)
       r4:c3c34884
      [<c01b4534>] (unlock_policy_rwsem_write+0x0/0x40) from [<c01b4c40>] (cpufreq_add_dev+0x324/0x398)
      [<c01b491c>] (cpufreq_add_dev+0x0/0x398) from [<c018fd64>] (sysdev_driver_register+0xc0/0x130)
      [<c018fca4>] (sysdev_driver_register+0x0/0x130) from [<c01b3574>] (cpufreq_register_driver+0xbc/0x174)
      Signed-off-by: default avatarLothar Waßmann <LW@KARO-electronics.de>
      Signed-off-by: default avatarDave Jones <davej@redhat.com>
      dca02613
    • Ingo Molnar's avatar
      67159306
    • Ingo Molnar's avatar
      sched: re-tune NUMA topologies · ea3f01f8
      Ingo Molnar authored
      improve the sysbench ramp-up phase and its peak throughput on
      a 16way NUMA box, by turning on WAKE_AFFINE:
      
                   tip/sched   tip/sched+wake-affine
      -------------------------------------------------
          1:             700              830    +15.65%
          2:            1465             1391    -5.28%
          4:            3017             3105    +2.81%
          8:            5100             6021    +15.30%
         16:           10725            10745    +0.19%
         32:           10135            10150    +0.16%
         64:            9338             9240    -1.06%
        128:            8599             8252    -4.21%
        256:            8475             8144    -4.07%
      -------------------------------------------------
        SUM:           57558            57882    +0.56%
      
      this change also improves lat_ctx from 6.69 usecs to 1.11 usec:
      
        $ ./lat_ctx -s 0 2
        "size=0k ovr=1.19
        2 1.11
      
        $ ./lat_ctx -s 0 2
        "size=0k ovr=1.22
        2 6.69
      
      in sysbench it's an overall win with some weakness at the lots-of-clients
      side. That happens because we now under-balance this workload
      a bit. To counter that effect, turn on NEWIDLE:
      
                    wake-idle          wake-idle+newidle
       -------------------------------------------------
           1:             830              834    +0.43%
           2:            1391             1401    +0.65%
           4:            3105             3091    -0.43%
           8:            6021             6046    +0.42%
          16:           10745            10736    -0.08%
          32:           10150            10206    +0.55%
          64:            9240             9533    +3.08%
         128:            8252             8355    +1.24%
         256:            8144             8384    +2.87%
       -------------------------------------------------
         SUM:           57882            58591    +1.21%
      
      as a bonus this not only improves the many-clients case but
      also improves the (more important) rampup phase.
      
      sysbench is a workload that quickly breaks down if the
      scheduler over-balances, so since it showed an improvement
      under NEWIDLE this change is definitely good.
      ea3f01f8
    • Mike Galbraith's avatar
      sched: stop wake_affine from causing serious imbalance · b3137bc8
      Mike Galbraith authored
      Prevent short-running wakers of short-running threads from overloading a single
      cpu via wakeup affinity, and wire up disconnected debug option.
      Signed-off-by: default avatarMike Galbraith <efault@gmx.de>
      Signed-off-by: default avatarPeter Zijlstra <peterz@infradead.org>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      b3137bc8
    • Peter Zijlstra's avatar
      sched: fix sched_clock_cpu() · a381759d
      Peter Zijlstra authored
      Make sched_clock_cpu() return 0 before it has been initialized and avoid
      corrupting its state due to doing so.
      
      This fixes the weird printk timestamp jump reported.
      Signed-off-by: default avatarPeter Zijlstra <a.p.zijlstra@chello.nl>
      a381759d
    • Ingo Molnar's avatar
      revert ("sched: fair-group: SMP-nice for group scheduling") · 6363ca57
      Ingo Molnar authored
      Yanmin Zhang reported:
      
      Comparing with 2.6.25, volanoMark has big regression with kernel 2.6.26-rc1.
      It's about 50% on my 8-core stoakley, 16-core tigerton, and Itanium Montecito.
      
      With bisect, I located the following patch:
      
      | 18d95a28 is first bad commit
      | commit 18d95a28
      | Author: Peter Zijlstra <a.p.zijlstra@chello.nl>
      | Date:   Sat Apr 19 19:45:00 2008 +0200
      |
      |     sched: fair-group: SMP-nice for group scheduling
      
      Revert it so that we get v2.6.25 behavior.
      Bisected-by: default avatarYanmin Zhang <yanmin_zhang@linux.intel.com>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      6363ca57
    • Ingo Molnar's avatar
      sched: cleanup · 4285f594
      Ingo Molnar authored
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      4285f594
    • Adrian Bunk's avatar
      show_schedstat(): fix memleak · c6fba545
      Adrian Bunk authored
      The Coverity checker spotted a memleak introduced by commit
      39106dcf (cpumask: use new cpus_scnprintf
      function).
      
      It seems the kfree() got lost between v2 and v3 of this patch...
      Signed-off-by: default avatarAdrian Bunk <bunk@kernel.org>
      Cc: Mike Travis <travis@sgi.com>
      Signed-off-by: default avatarAndrew Morton <akpm@linux-foundation.org>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      c6fba545
    • Roel Kluin's avatar
      sched: unite unlikely pairs in rt_policy() and schedule_debug() · 3f33a7ce
      Roel Kluin authored
      Removes obfuscation and may improve assembly.
      Signed-off-by: default avatarRoel Kluin <roel.kluin@gmail.com>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      3f33a7ce
    • Ingo Molnar's avatar
      revert ("sched: fair: weight calculations") · f9305d4a
      Ingo Molnar authored
      Yanmin Zhang reported:
      
      Comparing with kernel 2.6.25, sysbench+mysql(oltp, readonly) has many
      regressions with 2.6.26-rc1:
      
       1) 8-core stoakley: 28%;
       2) 16-core tigerton: 20%;
       3) Itanium Montvale: 50%.
      
      Bisect located this patch:
      
      | 8f1bc385 is first bad commit
      | commit 8f1bc385
      | Author: Peter Zijlstra <a.p.zijlstra@chello.nl>
      | Date:   Sat Apr 19 19:45:00 2008 +0200
      |
      |     sched: fair: weight calculations
      
      Revert it to the 2.6.25 state.
      Bisected-by: default avatarYanmin Zhang <yanmin_zhang@linux.intel.com>
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      f9305d4a
  2. 28 May, 2008 14 commits
  3. 27 May, 2008 11 commits
    • Kenji Kaneshige's avatar
      pciehp: add message about pciehp_slot_with_bus option · 9e4f2e8d
      Kenji Kaneshige authored
      Some (broken?) platform assign the same slot name to multiple hotplug
      slots. On such system, slot initialization would fail because of name
      collision. The pciehp driver already have a "slot_with_bus" module
      option which adds the bus number into the slot name. This patch adds
      the message about this module option that will be displayed when slot
      name collision is detected.
      Signed-off-by: default avatarKenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
      Signed-off-by: default avatarKristen Carlson Accardi <kristen.c.accardi@intel.com>
      Signed-off-by: default avatarJesse Barnes <jbarnes@virtuousgeek.org>
      9e4f2e8d
    • Kenji Kaneshige's avatar
      pci hotplug core: add check of duplicate slot name · a86161b3
      Kenji Kaneshige authored
      Fix the following errors reported by Jan C. Nordholz in
      http://bugzilla.kernel.org/show_bug.cgi?id=10751.
      
      kobject_add_internal failed for 2 with -EEXIST, don't try to register things with the same name in the same directory.
      Pid: 1, comm: swapper Tainted: G        W 2.6.26-rc3 #1
       [<c0266980>] kobject_add_internal+0x140/0x190
       [<c0266afd>] kobject_init_and_add+0x2d/0x40
       [<c027bc91>] pci_hp_register+0x81/0x2f0
       [<c027fd07>] pciehp_probe+0x1a7/0x470
       [<c01b3b84>] sysfs_add_one+0x44/0xa0
       [<c01b3c1f>] sysfs_addrm_start+0x3f/0xb0
       [<c01b497a>] sysfs_create_link+0x8a/0xf0
       [<c0279570>] pcie_port_probe_service+0x50/0x80
       [<c02e0545>] driver_sysfs_add+0x55/0x70
       [<c02e0662>] driver_probe_device+0x82/0x180
       [<c02e07cc>] __driver_attach+0x6c/0x70
       [<c02dfe0a>] bus_for_each_dev+0x3a/0x60
       [<c05db2d0>] pcied_init+0x0/0x80
       [<c02e04e6>] driver_attach+0x16/0x20
       [<c02e0760>] __driver_attach+0x0/0x70
       [<c02e0341>] bus_add_driver+0x1a1/0x220
       [<c05db2d0>] pcied_init+0x0/0x80
       [<c02e09cd>] driver_register+0x4d/0x120
       [<c05db050>] ibm_acpiphp_init+0x0/0x190
       [<c0125aab>] printk+0x1b/0x20
       [<c05db2d0>] pcied_init+0x0/0x80
       [<c05db2de>] pcied_init+0xe/0x80
       [<c05c751a>] kernel_init+0x10a/0x300
       [<c0120138>] schedule_tail+0x18/0x50
       [<c0103b9a>] ret_from_fork+0x6/0x1c
       [<c05c7410>] kernel_init+0x0/0x300
       [<c05c7410>] kernel_init+0x0/0x300
       [<c010485b>] kernel_thread_helper+0x7/0x1c
       =======================
      pci_hotplug: Unable to register kobject '2'<3>pciehp: pci_hp_register failed with error -22
      
      Slot with the same name can be registered multiple times if shpchp or
      pciehp driver is loaded after acpiphp is loaded because ACPI based
      hotplug driver and Native OS hotplug driver trying to handle the same
      physical slot. In this case, current pci_hotplug core will call
      kobject_init_and_add() muliple time with the same name. This is the
      cause of this problem. To fix this problem, this patch adds the check
      into pci_hp_register() to see if the slot with the same name.
      Signed-off-by: default avatarKenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
      Signed-off-by: default avatarKristen Carlson Accardi <kristen.c.accardi@intel.com>
      Signed-off-by: default avatarJesse Barnes <jbarnes@virtuousgeek.org>
      a86161b3
    • Kenji Kaneshige's avatar
      pciehp: move msleep after power off · 0711c70e
      Kenji Kaneshige authored
      According to the PCI Express specification, we must wait for at least
      1 second after turning power off before taking any action that relies
      on power having been removed from the slot/adapter. For this, current
      pciehp wait for 1 second after issuing the power off command in
      hpc_power_off_slot() function. But waiting for 1 second in
      hpc_power_off_slot() can make pciehp probing slow-down because pciehp
      probe code calls hpc_power_off_slot() if the slot is not occupied just
      in case. We don't need to wait for 1 second at the pciehp probe time
      because there is no action on that empty slot. So move 1 second wait
      from hpc_power_off_slot() to the caller of hpc_power_off_slot().
      Signed-off-by: default avatarKenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
      Signed-off-by: default avatarKristen Carlson Accardi <kristen.c.accardi@intel.com>
      Signed-off-by: default avatarJesse Barnes <jbarnes@virtuousgeek.org>
      0711c70e
    • Kenji Kaneshige's avatar
      pciehp: poll cmd completion if hotplug interrupt is disabled · 6592e02a
      Kenji Kaneshige authored
      Fix improper long wait for command completion in pciehp probing.
      
      As described in PCI Express specification, software notification is
      not generated if the command that occurs as a result of a write to the
      Slot Control register that disables software notification of command
      completed events. Since pciehp driver doesn't take it into account,
      such command is issued in pciehp probing, and it causes improper long
      wait for command completion.
      
      This patch changes the pciehp driver to take such command into
      account.
      Signed-off-by: default avatarKenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
      Signed-off-by: default avatarKristen Carlson Accardi <kristen.c.accardi@intel.com>
      Signed-off-by: default avatarJesse Barnes <jbarnes@virtuousgeek.org>
      6592e02a
    • Kenji Kaneshige's avatar
      pciehp: fix slow probing · 5808639b
      Kenji Kaneshige authored
      Fix the "pciehp probing slow" problem reported from Jan C. Nordholz in
      http://bugzilla.kernel.org/show_bug.cgi?id=10751.
      
      The command completed bit in Slot Status register applies only to
      commands issued to control the attention indicator, power indicator,
      power controller, or electromechanical interlock. However, writes to
      other parts of the Slot Control register would end up writing to the
      control fields. Hence, any write to Slot Control register is
      considered as a command. However, if the controller doesn't support
      any of attention indicator, power indicator, power controller and
      electromechanical interlock, command completed bit would not set in
      writing to Slot Control register. In this case, we should not wait for
      command completed bit set, otherwise all commands would be considered
      not completed in timeout seconds (1 sec.).
      
      The cause of the problem is pciehp driver didn't take this situation
      into account. This patch changes pciehp to take it into account. This
      patch also add the check for "No Command Completed Support" bit in
      Slot Capability register. If it is set, we should not wait for command
      completed bit set as well.
      
      This problem seems to be revealed by the commit
      c27fb883 that fixed the bug that
      pciehp did not wait for command completed properly (pciehp just
      ignored the command completion event).
      Signed-off-by: default avatarKenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
      Signed-off-by: default avatarKristen Carlson Accardi <kristen.c.accardi@intel.com>
      Signed-off-by: default avatarJesse Barnes <jbarnes@virtuousgeek.org>
      5808639b
    • Kenji Kaneshige's avatar
      pciehp: fix NULL dereference in interrupt handler · dbd79aed
      Kenji Kaneshige authored
      Fix the following NULL dereference problem reported from Pierre Ossman
      and Ingo Molnar.
      
      pciehp: HPC vendor_id 8086 device_id 27d0 ss_vid 0 ss_did 0
      pciehp: pciehp_find_slot: slot (device=0x0) not found
      BUG: unable to handle kernel NULL pointer dereference at 0000000000000070
      IP: [<ffffffff80494a8b>] pciehp_handle_presence_change+0x7e/0x113
      PGD 0
      Oops: 0000 [1]
      CPU 0
      Modules linked in:
      Pid: 1, comm: swapper Tainted: G        W 2.6.26-rc3-sched-devel.git-00001-g2b99b26-dirty #170
      RIP: 0010:[<ffffffff80494a8b>]  [<ffffffff80494a8b>] pciehp_handle_presence_change+0x7e/0x113
      RSP: 0000:ffff81003f83fbb0  EFLAGS: 00010046
      RAX: 0000000000000039 RBX: 0000000000000000 RCX: 0000000000000000
      RDX: 0000000000000000 RSI: 0000000000000001 RDI: 0000000000000046
      RBP: ffff81003f83fbd0 R08: 0000000000000001 R09: ffffffff80245103
      R10: 0000000000000020 R11: 0000000000000000 R12: ffff81003ea53a30
      R13: 0000000000000000 R14: 0000000000000011 R15: ffffffff80495926
      FS:  0000000000000000(0000) GS:ffffffff80be7400(0000) knlGS:0000000000000000
      CS:  0010 DS: 0018 ES: 0018 CR0: 000000008005003b
      CR2: 0000000000000070 CR3: 0000000000201000 CR4: 00000000000006a0
      DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
      DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
      Process swapper (pid: 1, threadinfo ffff81003f83e000, task ffff81003f840000)
      Stack:  0000000000000008 ffff81003f83fbf6 ffff81003ea53a30 0000000000000008
       ffff81003f83fc10 ffffffff80495ab4 0000000000000011 0000000000000002
       0000000000000202 0000000000000202 00000000fffffff4 ffff81003ea53a30
      Call Trace:
       [<ffffffff80495ab4>] pcie_isr+0x18e/0x1bc
       [<ffffffff80260831>] request_irq+0x106/0x12f
       [<ffffffff80495fb6>] pcie_init+0x15e/0x6cc
       [<ffffffff804933a3>] pciehp_probe+0x64/0x541
       [<ffffffff8048f4e7>] pcie_port_probe_service+0x4c/0x76
       [<ffffffff8054af70>] driver_probe_device+0xd4/0x1f0
       [<ffffffff8054b108>] __driver_attach+0x7c/0x7e
       [<ffffffff8054b08c>] ? __driver_attach+0x0/0x7e
       [<ffffffff8054a4b6>] bus_for_each_dev+0x53/0x7d
       [<ffffffff8054ad3c>] driver_attach+0x1c/0x1e
       [<ffffffff8054a9c2>] bus_add_driver+0xdd/0x25b
       [<ffffffff80c09d3d>] ? pcied_init+0x0/0x8b
       [<ffffffff8054b288>] driver_register+0x5f/0x13e
       [<ffffffff80c09d3d>] ? pcied_init+0x0/0x8b
       [<ffffffff8048f441>] pcie_port_service_register+0x47/0x49
       [<ffffffff80c09d52>] pcied_init+0x15/0x8b
       [<ffffffff80bf3938>] kernel_init+0x75/0x243
       [<ffffffff808639d2>] ? _spin_unlock_irq+0x2b/0x3a
       [<ffffffff80228d1f>] ? finish_task_switch+0x57/0x9a
       [<ffffffff8020c258>] child_rip+0xa/0x12
       [<ffffffff8020bcec>] ? restore_args+0x0/0x30
       [<ffffffff80bf38c3>] ? kernel_init+0x0/0x243
       [<ffffffff8020c24e>] ? child_rip+0x0/0x12
      
      Code: 83 80 00 00 00 48 39 f0 75 e1 0f b6 c9 48 c7 c2 00 0e 8d 80 48 c7 c6 8a 60 a6 80 48 c7 c7 10 db a8 80 31 c0 e8 3f 8d d9 ff 31 db <48> 8b 43 70 48 8d 75 ef 48 89 df ff 50 30 80 7d ef 00 74 37 48
      RIP  [<ffffffff80494a8b>] pciehp_handle_presence_change+0x7e/0x113
       RSP <ffff81003f83fbb0>
      CR2: 0000000000000070
      Kernel panic - not syncing: Fatal exception
      
      The situation under which it occurs is hw and timing related: it appears
      to happen on a system that has PCI hotplug hardware but with no active
      hotplug cards, and another interrupt in the same (shared) IRQ line
      arrives too early, before the hotplug-slot entry has been set up - as
      triggered by CONFIG_DEBUG_SHIRQ=y:
      
      This patch contains the following two fixes.
      
      (1) Clear all events bits in Slot Status register to prevent the pciehp
          driver from detecting the spurious events that would have been occur
          before pciehp loading.
      
      (2) Add check whether slot initialization had been already done.
      
      This is short term fix. We need more structural fixes to install
      interrupt handler after slot initialization is done.
      Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
      Signed-off-by: default avatarKenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
      Signed-off-by: default avatarKristen Carlson Accardi <kristen.c.accardi@intel.com>
      Signed-off-by: default avatarJesse Barnes <jbarnes@virtuousgeek.org>
      dbd79aed
    • Kenji Kaneshige's avatar
      shpchp: add message about shpchp_slot_with_bus option · b3bd307c
      Kenji Kaneshige authored
      Some (broken?) platform assign the same slot name to multiple hotplug
      slots. On such system, slot initialization would fail because of name
      collision. The shpchp driver already have a "slot_with_bus" module
      option which adds the bus number into the slot name. This patch adds
      the message about this module option that will be displayed when slot
      name collision is detected.
      Signed-off-by: default avatarKenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>
      Signed-off-by: default avatarKristen Carlson Accardi <kristen.c.accardi@intel.com>
      Signed-off-by: default avatarJesse Barnes <jbarnes@virtuousgeek.org>
      b3bd307c
    • Tony Luck's avatar
      [IA64] Workaround for RSE issue · 4dcc29e1
      Tony Luck authored
      Problem: An application violating the architectural rules regarding
      operation dependencies and having specific Register Stack Engine (RSE)
      state at the time of the violation, may result in an illegal operation
      fault and invalid RSE state.  Such faults may initiate a cascade of
      repeated illegal operation faults within OS interruption handlers.
      The specific behavior is OS dependent.
      
      Implication: An application causing an illegal operation fault with
      specific RSE state may result in a series of illegal operation faults
      and an eventual OS stack overflow condition.
      
      Workaround: OS interruption handlers that switch to kernel backing
      store implement a check for invalid RSE state to avoid the series
      of illegal operation faults.
      
      The core of the workaround is the RSE_WORKAROUND code sequence
      inserted into each invocation of the SAVE_MIN_WITH_COVER and
      SAVE_MIN_WITH_COVER_R19 macros.  This sequence includes hard-coded
      constants that depend on the number of stacked physical registers
      being 96.  The rest of this patch consists of code to disable this
      workaround should this not be the case (with the presumption that
      if a future Itanium processor increases the number of registers, it
      would also remove the need for this patch).
      
      Move the start of the RBS up to a mod32 boundary to avoid some
      corner cases.
      
      The dispatch_illegal_op_fault code outgrew the spot it was
      squatting in when built with this patch and CONFIG_VIRT_CPU_ACCOUNTING=y
      Move it out to the end of the ivt.
      Signed-off-by: default avatarTony Luck <tony.luck@intel.com>
      4dcc29e1
    • Linus Torvalds's avatar
      Merge git://git.kernel.org/pub/scm/linux/kernel/git/hskinnemoen/avr32-2.6 · 3dbfd080
      Linus Torvalds authored
      * git://git.kernel.org/pub/scm/linux/kernel/git/hskinnemoen/avr32-2.6:
        avr32: Fix cpufreq oops when ondemand governor is default
        avr32: Update defconfigs
        avr32: export strnlen_user
        avr32: export copy_page
      3dbfd080
    • David Woodhouse's avatar
      ck804rom: fix driver_data in probe table. · edb2301f
      David Woodhouse authored
      There's a reason why using C99 initialisers even in the supposedly
      trivial structs is a good idea.
      Signed-off-by: default avatarCarl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
      Signed-off-by: default avatarDavid Woodhouse <dwmw2@infradead.org>
      Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
      edb2301f
    • Haavard Skinnemoen's avatar
      avr32: Fix cpufreq oops when ondemand governor is default · f04d264a
      Haavard Skinnemoen authored
      Move the AP7 cpufreq init to late_initcall() so that we don't try to
      bring up cpufreq until the governor is ready. x86 also uses
      late_initcall() for this.
      Signed-off-by: default avatarHaavard Skinnemoen <haavard.skinnemoen@atmel.com>
      f04d264a
  4. 26 May, 2008 3 commits