In the CONFIG_AUDITSYSCALL fast-path for x86 64-bit system calls,
we can pass a bad return value and/or error indication for the
system call to audit_syscall_exit().  This happens when
TIF_NEED_RESCHED was set as the system call returned, so we went
out to schedule() and came back to the exit-audit fast-path.  The
fix is to reload the user return value register from the pt_regs
before using it for audit_syscall_exit().

Both the 32-bit kernel's fast path and the 64-bit kernel's 32-bit
system call fast paths work slightly differently, so that they
always leave the fast path entirely to reschedule and don't return
there, so they don't have the analogous bugs.
Reported-by: NAlexander Viro <aviro@redhat.com>
Signed-off-by: NRoland McGrath <roland@redhat.com>

Pointed out by Lucas who found the new one in a comment in
setup_percpu.c. And then I fixed the others that I grepped
for.
Reported-by: NLucas <canolucas@gmail.com>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Commit e534c7c5 ("numa: x86_64: use generic percpu var
numa_node_id() implementation") broke numa systems that don't have ram
on node0 when MEMORY_HOTPLUG is enabled, because cpu_up() will call
cpu_to_node() before per_cpu(numa_node) is setup for APs.

When Node0 doesn't have RAM, on x86, cpus already round it to nearest
node with RAM in x86_cpu_to_node_map.  and per_cpu(numa_node) is not set
up until in c_init for APs.

When later cpu_up() calling cpu_to_node() will get 0 again, and make it
online even there is no RAM on node0.  so later all APs can not booted up,
and later will have panic.

[    1.611101] On node 0 totalpages: 0
.........
[    2.608558] On node 0 totalpages: 0
[    2.612065] Brought up 1 CPUs
[    2.615199] Total of 1 processors activated (3990.31 BogoMIPS).
...
   93.225341] calling  loop_init+0x0/0x1a4 @ 1
[   93.229314] PERCPU: allocation failed, size=80 align=8, failed to populate
[   93.246539] Pid: 1, comm: swapper Tainted: G        W   2.6.35-rc4-tip-yh-04371-gd64e6c4-dirty #354
[   93.264621] Call Trace:
[   93.266533]  [<ffffffff81125e43>] pcpu_alloc+0x83a/0x8e7
[   93.270710]  [<ffffffff81125f15>] __alloc_percpu+0x10/0x12
[   93.285849]  [<ffffffff8140786c>] alloc_disk_node+0x94/0x16d
[   93.291811]  [<ffffffff81407956>] alloc_disk+0x11/0x13
[   93.306157]  [<ffffffff81503e51>] loop_alloc+0xa7/0x180
[   93.310538]  [<ffffffff8277ef48>] loop_init+0x9b/0x1a4
[   93.324909]  [<ffffffff8277eead>] ? loop_init+0x0/0x1a4
[   93.329650]  [<ffffffff810001f2>] do_one_initcall+0x57/0x136
[   93.345197]  [<ffffffff827486d0>] kernel_init+0x184/0x20e
[   93.348146]  [<ffffffff81034954>] kernel_thread_helper+0x4/0x10
[   93.365194]  [<ffffffff81c7cc3c>] ? restore_args+0x0/0x30
[   93.369305]  [<ffffffff8274854c>] ? kernel_init+0x0/0x20e
[   93.386011]  [<ffffffff81034950>] ? kernel_thread_helper+0x0/0x10
[   93.392047] loop: out of memory
...

Try to assign per_cpu(numa_node) early

[akpm@linux-foundation.org: tidy up code comment]
Signed-off-by: NYinghai <yinghai@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Tejun Heo <tj@kernel.org>
Cc: Denys Vlasenko <vda.linux@googlemail.com>
Acked-by: NLee Schermerhorn <lee.schermerhorn@hp.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

The current shrinker implementation requires the registered callback
to have global state to work from. This makes it difficult to shrink
caches that are not global (e.g. per-filesystem caches). Pass the shrinker
structure to the callback so that users can embed the shrinker structure
in the context the shrinker needs to operate on and get back to it in the
callback via container_of().
Signed-off-by: NDave Chinner <dchinner@redhat.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>

In commit f007ea26, the order of the %es and %ds segment registers
got accidentally swapped, so synthesized 'struct pt_regs' frames
have the two values inverted.  It's almost sure that these values
never matter, and that they also never differ.  But wrong is wrong.
Signed-off-by: NRoland McGrath <roland@redhat.com>

The fixed bar capability structure is searched in PCI extended
configuration space.  We need to make sure there is a valid capability
ID to begin with otherwise, the search code may stuck in a infinite
loop which results in boot hang.  This patch adds additional check for
cap ID 0, which is also invalid, and indicates end of chain.

End of chain is supposed to have all fields zero, but that doesn't
seem to always be the case in the field.
Suggested-by: N"H. Peter Anvin" <hpa@zytor.com>
Signed-off-by: NJacob Pan <jacob.jun.pan@linux.intel.com>
Reviewed-by: NJesse Barnes <jbarnes@virtuousgeek.org>
LKML-Reference: <1279306706-27087-1-git-send-email-jacob.jun.pan@linux.intel.com>
Signed-off-by: NH. Peter Anvin <hpa@linux.intel.com>

Found one x2apic system kexec loop test failed
when CONFIG_NMI_WATCHDOG=y (old) or CONFIG_LOCKUP_DETECTOR=y (current tip)

first kernel can kexec second kernel, but second kernel can not kexec third one.

it can be duplicated on another system with BIOS preenabled x2apic.
First kernel can not kexec second kernel.

It turns out, when kernel boot with pre-enabled x2apic, it will not execute
disable_local_APIC on shutdown path.

when init_apic_mappings() is called in setup_arch, it will skip setting of
apic_phys when x2apic_mode is set. ( x2apic_mode is much early check_x2apic())
Then later, disable_local_APIC() will bail out early because !apic_phys.

So check !x2apic_mode in x2apic_mode in disable_local_APIC with !apic_phys.

another solution could be updating init_apic_mappings() to set apic_phys even
for preenabled x2apic system. Actually even for x2apic system, that lapic
address is mapped already in early stage.

BTW: is there any x2apic preenabled system with apicid of boot cpu > 255?
Signed-off-by: NYinghai Lu <yinghai@kernel.org>
LKML-Reference: <4C3EB22B.3000701@kernel.org>
Acked-by: NSuresh Siddha <suresh.b.siddha@intel.com>
Cc: stable@kernel.org
Signed-off-by: NH. Peter Anvin <hpa@linux.intel.com>

If we fail to assign resources to a PCI BAR, this patch makes us try the
original address from BIOS rather than leaving it disabled.

Linux tries to make sure all PCI device BARs are inside the upstream
PCI host bridge or P2P bridge apertures, reassigning BARs if necessary.
Windows does similar reassignment.

Before this patch, if we could not move a BAR into an aperture, we left
the resource unassigned, i.e., at address zero.  Windows leaves such BARs
at the original BIOS addresses, and this patch makes Linux do the same.

This is a bit ugly because we disable the resource long before we try to
reassign it, so we have to keep track of the BIOS BAR address somewhere.
For lack of a better place, I put it in the struct pci_dev.

I think it would be cleaner to attempt the assignment immediately when the
claim fails, so we could easily remember the original address.  But we
currently claim motherboard resources in the middle, after attempting to
claim PCI resources and before assigning new PCI resources, and changing
that is a fairly big job.

Addresses https://bugzilla.kernel.org/show_bug.cgi?id=16263Reported-by: NAndrew <nitr0@seti.kr.ua>
Tested-by: NAndrew <nitr0@seti.kr.ua>
Signed-off-by: NBjorn Helgaas <bjorn.helgaas@hp.com>
Signed-off-by: NJesse Barnes <jbarnes@virtuousgeek.org>

commit 30a564be (x86, hpet: Restrict read back to affected ATI
chipset) restricted the workaround for the HPET bug to SMX00
chipsets. This was reasonable as those were the only ones against
which we ever got a bug report.

Stephan Wolf reported now that this patch breaks his IXP400 based
machine. Though it's confirmed to work on other IXP400 based systems.

To error out on the safe side, we force the HPET readback workaround
for all ATI SMbus class chipsets.
Reported-by: NStephan Wolf <stephan@letzte-bankreihe.de>
LKML-Reference: <alpine.LFD.2.00.1007142134140.3321@localhost.localdomain>
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Tested-by: NStephan Wolf <stephan@letzte-bankreihe.de>
Acked-by: NBorislav Petkov <borislav.petkov@amd.com>

After remove a rmap, we should flush all vcpu's tlb
Signed-off-by: NXiao Guangrong <xiaoguangrong@cn.fujitsu.com>
Signed-off-by: NMarcelo Tosatti <mtosatti@redhat.com>

It will just return 0 as there is no i8042 controller
Signed-off-by: NFeng Tang <feng.tang@intel.com>
LKML-Reference: <1278342202-10973-3-git-send-email-feng.tang@intel.com>
Acked-by: NDmitry Torokhov <dtor@mail.ru>
Signed-off-by: NH. Peter Anvin <hpa@linux.intel.com>

Some x86 platforms like Intel MID platforms don't have i8042 controllers,
and i8042 driver's probe to some legacy IO ports may hang the MID
processor. With this hook, i8042 driver can runtime check and skip the
probe when the pretection fail which also saves some probe time

[ hpa note: this is currently a compile-time check, which breaks the
  i386 allyesconfig build.  This patch series thus does fix a regression. ]
Signed-off-by: NFeng Tang <feng.tang@intel.com>
LKML-Reference: <1278342202-10973-2-git-send-email-feng.tang@intel.com>
Acked-by: NDmitry Torokhov <dtor@mail.ru>
Signed-off-by: NH. Peter Anvin <hpa@linux.intel.com>

Export x86_platform to modules in preparation of using it for i8042
discovery control.
Signed-off-by: NH. Peter Anvin <hpa@linux.intel.com>
LKML-Reference: <1278342202-10973-1-git-send-email-feng.tang@intel.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Feng Tang <feng.tang@intel.com>
Cc: Dmitry Torokhov <dtor@mail.ru>

enter_lmode() and exit_lmode() modify the guest's EFER.LMA before calling
vmx_set_efer().  However, the latter function depends on the value of EFER.LMA
to determine whether MSR_KERNEL_GS_BASE needs reloading, via
vmx_load_host_state().  With EFER.LMA changing under its feet, it took the
wrong choice and corrupted userspace's %gs.

This causes 32-on-64 host userspace to fault.

Fix not touching EFER.LMA; instead ask vmx_set_efer() to change it.
Signed-off-by: NAvi Kivity <avi@redhat.com>

Reimplement augmented RB-trees without sprinkling extra branches
all over the RB-tree code (which lives in the scheduler hot path).

This approach is 'borrowed' from Fabio's BFQ implementation and
relies on traversing the rebalance path after the RB-tree-op to
correct the heap property for insertion/removal and make up for
the damage done by the tree rotations.

For insertion the rebalance path is trivially that from the new
node upwards to the root, for removal it is that from the deepest
node in the path from the to be removed node that will still
be around after the removal.

[ This patch also fixes a video driver regression reported by
  Ali Gholami Rudi - the memtype->subtree_max_end was updated
  incorrectly. ]
Acked-by: NSuresh Siddha <suresh.b.siddha@intel.com>
Acked-by: NVenkatesh Pallipadi <venki@google.com>
Signed-off-by: NPeter Zijlstra <a.p.zijlstra@chello.nl>
Tested-by: NAli Gholami Rudi <ali@rudi.ir>
Cc: Fabio Checconi <fabio@gandalf.sssup.it>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
LKML-Reference: <1275414172.27810.27961.camel@twins>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

While doing some performance counter validation tests on some
assembly language programs I noticed that the "branches:u"
count was very wrong on AMD machines.

It looks like the wrong event was selected.
Signed-off-by: NVince Weaver <vweaver1@eecs.utk.edu>
Acked-by: NPeter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Borislav Petkov <borislav.petkov@amd.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: <stable@kernel.org>
LKML-Reference: <alpine.DEB.2.00.1007011526010.23160@cl320.eecs.utk.edu>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

The x3950 family can have as many as 256 PCI buses in a single system, so
change the limits to the maximum.  Since there can only be 256 PCI buses in one
domain, we no longer need the BUG_ON check.
Signed-off-by: NDarrick J. Wong <djwong@us.ibm.com>
LKML-Reference: <20100701004519.GQ15515@tux1.beaverton.ibm.com>
Signed-off-by: NH. Peter Anvin <hpa@zytor.com>

Before we had a generic breakpoint layer, x86 used to send a
sigtrap for any debug event that happened in userspace,
except if it was caused by lazy dr7 switches.

Currently we only send such signal for single step or breakpoint
events.

However, there are three other kind of debug exceptions:

- debug register access detected: trigger an exception if the
  next instruction touches the debug registers. We don't use
  it.
- task switch, but we don't use tss.
- icebp/int01 trap. This instruction (0xf1) is undocumented and
  generates an int 1 exception. Unlike single step through TF
  flag, it doesn't set the single step origin of the exception
  in dr6.

icebp then used to be reported in userspace using trap signals
but this have been incidentally broken with the new breakpoint
code. Reenable this. Since this is the only debug event that
doesn't set anything in dr6, this is all we have to check.

This fixes a regression in Wine where World Of Warcraft got broken
as it uses this for software protection checks purposes. And
probably other apps do.
Reported-and-tested-by: NAlexandre Julliard <julliard@winehq.org>
Signed-off-by: NFrederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Prasad <prasad@linux.vnet.ibm.com>
Cc: 2.6.33.x 2.6.34.x <stable@kernel.org>

Newer systems (x3950M2) can have 48 PHBs per chassis and 8
chassis, so bump the limits up and provide an explanation
of the requirements for each class.
Signed-off-by: NDarrick J. Wong <djwong@us.ibm.com>
Acked-by: NMuli Ben-Yehuda <muli@il.ibm.com>
Cc: Corinna Schultz <cschultz@linux.vnet.ibm.com>
Cc: <stable@kernel.org>
LKML-Reference: <20100624212647.GI15515@tux1.beaverton.ibm.com>
[ v2: Fixed build bug, added back PHBS_PER_CALGARY == 4 ]
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Dell Precision WorkStation T7400 freezes on reboot unless
reboot=b is used.

Reference: https://qa.mandriva.com/show_bug.cgi?id=58017Signed-off-by: NThomas Backlund <tmb@mandriva.org>
LKML-Reference: <4C1CC6E9.6000701@mandriva.org>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

This fixes the -Os breaks with gcc 4.5 bug.  rdtsc_barrier needs to be
force inlined, otherwise user space will jump into kernel space and
kill init.

This also addresses http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44129
I believe.
Signed-off-by: NAndi Kleen <ak@linux.intel.com>
LKML-Reference: <20100618210859.GA10913@basil.fritz.box>
Signed-off-by: NH. Peter Anvin <hpa@linux.intel.com>
Cc: <stable@kernel.org>

subtree_max_end that was recently added to struct memtype was not getting
properly initialized resulting in

WARNING: kmemcheck: Caught 64-bit read from uninitialized memory
in memtype_rb_augment_cb()
reported here
https://bugzilla.kernel.org/show_bug.cgi?id=16092

This change fixes the problem.
Reported-by: NChristian Casteyde <casteyde.christian@free.fr>
Tested-by: NChristian Casteyde <casteyde.christian@free.fr>
Signed-off-by: NVenkatesh Pallipadi <venki@google.com>
LKML-Reference: <1276217101-11515-1-git-send-email-venki@google.com>
Signed-off-by: NH. Peter Anvin <hpa@linux.intel.com>
Cc: Suresh Siddha <suresh.b.siddha@intel.com>

Yannick found that video does not work with 2.6.34.  The cause of this
bug was that the BIOS had assigned the wrong range to the PCI bridge
above the video device.  Before 2.6.34 the kernel would have shrunk
the size of the bridge window, but since
  d65245c3 PCI: don't shrink bridge resources
the kernel will avoid shrinking BIOS ranges.

So zero out the old range if we fail to claim it at boot time; this will
cause us to allocate a new range at startup, restoring the 2.6.34
behavior.

Fixes regression https://bugzilla.kernel.org/show_bug.cgi?id=16009.
Reported-by: NYannick <yannick.roehlly@free.fr>
Acked-by: NBjorn Helgaas <bjorn.helgaas@hp.com>
Signed-off-by: NYinghai Lu <yinghai@kernel.org>
Signed-off-by: NJesse Barnes <jbarnes@virtuousgeek.org>

Catch missing conversion to the register structure "glove box" scheme.

Found by gcc 4.6's new warnings.
Signed-off-by: NAndi Kleen <ak@linux.intel.com>
LKML-Reference: <20100610111040.F1781B1A2B@basil.firstfloor.org>
Signed-off-by: NH. Peter Anvin <hpa@zytor.com>

Avoid hundreds of warnings with a gcc 4.6 -Wall build.
Signed-off-by: NAndi Kleen <ak@linux.intel.com>
Acked-by: NTejun Heo <tj@kernel.org>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

Saving platform non-volatile state may be required for suspend to RAM as
well as hibernation. Move it to more generic code.
Signed-off-by: NMatthew Garrett <mjg@redhat.com>
Acked-by: NRafael J. Wysocki <rjw@sisk.pl>
Tested-by: NMaxim Levitsky <maximlevitsky@gmail.com>
Signed-off-by: NLen Brown <len.brown@intel.com>

Based on Intel Vol3b (March 2010), the event
SNOOPQ_REQUEST_OUTSTANDING is restricted to counters 0,1 so
update the event table for Intel Westmere accordingly.
Signed-off-by: NStephane Eranian <eranian@google.com>
Cc: peterz@infradead.org
Cc: paulus@samba.org
Cc: davem@davemloft.net
Cc: fweisbec@gmail.com
Cc: perfmon2-devel@lists.sf.net
Cc: eranian@gmail.com
Cc: <stable@kernel.org> # .34.x
LKML-Reference: <4c10cb56.5120e30a.2eb4.ffffc3de@mx.google.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

When I introduced the global variable gsi_end I thought gsi_end on
io_apics was one past the end of the gsi range for the io_apic.  After
it was pointed out the the range on io_apics was inclusive I changed
my global variable to match.  That was a big mistake.  Inclusive
semantics without a range start cannot describe the case when no gsi's
are allocated.  Describing the case where no gsi's are allocated is
important in sfi.c and mpparse.c so that we can assign gsi numbers
instead of blindly copying the gsi assignments the BIOS has done as we
do in the acpi case.

To keep from getting the global variable confused with the gsi range
end rename it gsi_top.

To allow describing the case where no gsi's are allocated have gsi_top
be one place the highest gsi number seen in the system.

This fixes an off by one bug in sfi.c:
Reported-by: Njacob pan <jacob.jun.pan@linux.intel.com>

This fixes the same off by one bug in mpparse.c:

This fixes an off unreachable by one bug in acpi/boot.c:irq_to_gsi
Reported-by: NYinghai <yinghai.lu@oracle.com>
Signed-off-by: NEric W. Biederman <ebiederm@xmission.com>
LKML-Reference: <m17hm9jre7.fsf_-_@fess.ebiederm.org>
Signed-off-by: NH. Peter Anvin <hpa@linux.intel.com>

If cr0.wp=0, we have to allow the guest kernel access to a page with pte.w=0.
We do that by setting spte.w=1, since the host cr0.wp must remain set so the
host can write protect pages.  Once we allow write access, we must remove
user access otherwise we mistakenly allow the user to write the page.
Reviewed-by: NXiao Guangrong <xiaoguangrong@cn.fujitsu.com>
Signed-off-by: NAvi Kivity <avi@redhat.com>

Always invalidate spte and flush TLBs when changing page size, to make
sure different sized translations for the same address are never cached
in a CPU's TLB.

Currently the only case where this occurs is when a non-leaf spte pointer is
overwritten by a leaf, large spte entry. This can happen after dirty
logging is disabled on a memslot, for example.

Noticed by Andrea.

KVM-Stable-Tag
Signed-off-by: NMarcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: NAvi Kivity <avi@redhat.com>

This patch implements a workaround for AMD erratum 383 into
KVM. Without this erratum fix it is possible for a guest to
kill the host machine. This patch implements the suggested
workaround for hypervisors which will be published by the
next revision guide update.

[jan: fix overflow warning on i386]
[xiao: fix unused variable warning]

Cc: stable@kernel.org
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>
Signed-off-by: NJan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: NXiao Guangrong <xiaoguangrong@cn.fujitsu.com>
Signed-off-by: NAvi Kivity <avi@redhat.com>

This patch moves handling of the MC vmexits to an earlier
point in the vmexit. The handle_exit function is too late
because the vcpu might alreadry have changed its physical
cpu.

Cc: stable@kernel.org
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>
Signed-off-by: NAvi Kivity <avi@redhat.com>

As Ingo pointed out in a separate patch, we should be using __ASSEMBLY__.
Make that the case in pgtable headers.
Signed-off-by: NAndres Salomon <dilinger@queued.net>
LKML-Reference: <20100605114042.35ac69c1@dev.queued.net>
Signed-off-by: NH. Peter Anvin <hpa@linux.intel.com>

Save/restore MISC_ENABLE register on suspend/resume.
This fixes OOPS (invalid opcode) on resume from STR on Asus P4P800-VM,
which wakes up with MWAIT disabled.

Fixes https://bugzilla.kernel.org/show_bug.cgi?id=15385Signed-off-by: NOndrej Zary <linux@rainbow-software.org>
Tested-by: NAlan Stern <stern@rowland.harvard.edu>
Acked-by: NH. Peter Anvin <hpa@linux.intel.com>
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>

The core suspend/resume code is run from stop_machine on CPU0 but
parts of the suspend/resume machinery (including xen_arch_resume) are
run on whichever CPU happened to schedule the xenwatch kernel thread.

As part of the non-core resume code xen_arch_resume is called in order
to restart the timer tick on non-boot processors. The boot processor
itself is taken care of by core timekeeping code.

xen_arch_resume uses smp_call_function which does not call the given
function on the current processor. This means that we can end up with
one CPU not receiving timer ticks if the xenwatch thread happened to
be scheduled on CPU > 0.

Use on_each_cpu instead of smp_call_function to ensure the timer tick
is resumed everywhere.
Signed-off-by: NIan Campbell <ian.campbell@citrix.com>
Acked-by: NJeremy Fitzhardinge <jeremy@goop.org>
Cc: Stable Kernel <stable@kernel.org> # .32.x

Percpu initialization happens now after booting the cores on the
machine and this causes them all to be displayed as belonging to
node 0:

Jun  8 05:57:21 kepek kernel: [    0.106999] Booting Node   0,
Processors  #1 #2 #3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15 #16 #17 #18 #19 #20 #21 #22 #23 Ok.

Use early_cpu_to_node() to get the correct node of each core
instead.
Signed-off-by: NBorislav Petkov <borislav.petkov@amd.com>
Cc: Mike Travis <travis@sgi.com>
LKML-Reference: <20100601190455.GA14237@aftab>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

This patch implements a fallback to the GART IOMMU if this
is possible and the AMD IOMMU initialization failed.
Otherwise the fallback would be nommu which is very
problematic on machines with more than 4GB of memory or
swiotlb which hurts io-performance.

Cc: stable@kernel.org
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

When request_mem_region fails the error path tries to
disable the IOMMUs. This accesses the mmio-region which was
not allocated leading to a kernel crash. This patch fixes
the issue.

Cc: stable@kernel.org
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

DBG() macro for CONFIG_DEBUG_PER_CPU_MAPS is unused.
Signed-off-by: NAkinobu Mita <akinobu.mita@gmail.com>
LKML-Reference: <1274706291-13554-1-git-send-email-akinobu.mita@gmail.com>
Signed-off-by: NIngo Molnar <mingo@elte.hu>

The transactional API patch between the generic and model-specific
code introduced several important bugs with event scheduling, at
least on X86. If you had pinned events, e.g., watchdog,  and were
over-committing the PMU, you would get bogus counts. The bug was
showing up on Intel CPU because events would move around more
often that on AMD. But the problem also existed on AMD, though
harder to expose.

The issues were:

 - group_sched_in() was missing a cancel_txn() in the error path

 - cpuc->n_added was not properly maintained, leading to missing
   actions in hw_perf_enable(), i.e., n_running being 0. You cannot
   update n_added until you know the transaction has succeeded. In
   case of failed transaction n_added was not adjusted back.

 - in case of failed transactions, event_sched_out() was called
   and eventually invoked x86_disable_event() to touch the HW reg.
   But with transactions, on X86, event_sched_in() does not touch
   HW registers, it simply collects events into a list. Thus, you
   could end up calling x86_disable_event() on a counter which
   did not correspond to the current event when idx != -1.

The patch modifies the generic and X86 code to avoid all those problems.

First, we keep track of the number of events added last. In case the
transaction fails, we substract them from n_added. This approach is
necessary (as opposed to delaying updates to n_added) because not all
event updates use the transaction API, e.g., single events.

Second, we encapsulate the event_sched_in() and event_sched_out() in
group_sched_in() inside the transaction. That makes the operations
symmetrical and you can also detect that you are inside a transaction
and skip the HW reg access by checking cpuc->group_flag.

With this patch, you can now overcommit the PMU even with pinned
system-wide events present and still get valid counts.
Signed-off-by: NStephane Eranian <eranian@google.com>
Signed-off-by: NPeter Zijlstra <a.p.zijlstra@chello.nl>
LKML-Reference: <1274796225.5882.1389.camel@twins>
Signed-off-by: NIngo Molnar <mingo@elte.hu>