Pull scheduler updates from Ingo Molnar:
"The main changes in this cycle are:
- 'Nested Sleep Debugging', activated when CONFIG_DEBUG_ATOMIC_SLEEP=y.
This instruments might_sleep() checks to catch places that nest
blocking primitives - such as mutex usage in a wait loop. Such
bugs can result in hard to debug races/hangs.
Another category of invalid nesting that this facility will detect
is the calling of blocking functions from within schedule() ->
sched_submit_work() -> blk_schedule_flush_plug().
There's some potential for false positives (if secondary blocking
primitives themselves are not ready yet for this facility), but the
kernel will warn once about such bugs per bootup, so the warning
isn't much of a nuisance.
This feature comes with a number of fixes, for problems uncovered
with it, so no messages are expected normally.
- Another round of sched/numa optimizations and refinements, for
CONFIG_NUMA_BALANCING=y.
- Another round of sched/dl fixes and refinements.
Plus various smaller fixes and cleanups"
* 'sched-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (54 commits)
sched: Add missing rcu protection to wake_up_all_idle_cpus
sched/deadline: Introduce start_hrtick_dl() for !CONFIG_SCHED_HRTICK
sched/numa: Init numa balancing fields of init_task
sched/deadline: Remove unnecessary definitions in cpudeadline.h
sched/cpupri: Remove unnecessary definitions in cpupri.h
sched/deadline: Fix rq->dl.pushable_tasks bug in push_dl_task()
sched/fair: Fix stale overloaded status in the busiest group finding logic
sched: Move p->nr_cpus_allowed check to select_task_rq()
sched/completion: Document when to use wait_for_completion_io_*()
sched: Update comments about CLONE_NEWUTS and CLONE_NEWIPC
sched/fair: Kill task_struct::numa_entry and numa_group::task_list
sched: Refactor task_struct to use numa_faults instead of numa_* pointers
sched/deadline: Don't check CONFIG_SMP in switched_from_dl()
sched/deadline: Reschedule from switched_from_dl() after a successful pull
sched/deadline: Push task away if the deadline is equal to curr during wakeup
sched/deadline: Add deadline rq status print
sched/deadline: Fix artificial overrun introduced by yield_task_dl()
sched/rt: Clean up check_preempt_equal_prio()
sched/core: Use dl_bw_of() under rcu_read_lock_sched()
sched: Check if we got a shallowest_idle_cpu before searching for least_loaded_cpu
...
Pull perf events update from Ingo Molnar:
"On the kernel side there's few changes, the one that stands out is
PEBS machine state sampling support on x86, by Stephane Eranian.
On the tooling side:
User visible tooling changes:
- Don't open the DWARF info multiple times, keeping instead a dwfl
handle in struct dso, greatly speeding up 'perf report' on powerpc.
(Sukadev Bhattiprolu)
- Introduce PARSE_OPT_DISABLED option flag and use it to avoid
showing undersired options in tools that provides frontends to
'perf record', like sched, kvm, etc (Namhyung Kim)
- Fallback to kallsyms when using the minimal 'ELF' loader (Arnaldo
Carvalho de Melo)
- Fix annotation with kcore (Adrian Hunter)
- Support source line numbers in annotate using a hotkey (Andi Kleen)
- Callchain improvements including:
* Enable printing the srcline in the history
* Make get_srcline fall back to sym+offset (Andi Kleen)
- TUI hist_entry browser fixes, including showing missing overhead
value for first level callchain. Detected comparing the output of
--stdio/--gui (that matched) with --tui, that had this problem.
(Namhyung Kim)
- Support handling complete branch stacks as histograms (Andi Kleen)
Tooling infrastructure changes:
- Prep work for supporting per-pkg and snapshot counters in 'perf
stat' (Jiri Olsa)
- 'perf stat' refactorings, moving stuff from it to evsel.c to use in
per-pkg/snapshot format changes (Jiri Olsa)
- Add per-pkg format file parsing (Matt Fleming)
- Clean up libelf feature support code (Namhyung Kim)
- Add gzip decompression support for kernel modules (Namhyung Kim)
- More prep patches for Intel PT, including a a thread stack and more
stuff made available via the database export mechanism (Adrian
Hunter)
- More Intel PT work, including a facility to export sample data
(comms, threads, symbol names, etc) in a database friendly way,
with an script to use this to create a postgresql database.
(Adrian Hunter)
- Make sure that thread->mg->machine points to the machine where the
thread exists (it was being set only for the kmaps kernel modules
case, do it as well for the mmaps) and use it to shorten function
signatures (Arnaldo Carvalho de Melo)
... and lots of other fixes and smaller improvements"
* 'perf-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (91 commits)
perf report: In branch stack mode use address history sorting
perf report: Add --branch-history option
perf callchain: Support handling complete branch stacks as histograms
perf stat: Add support for snapshot counters
perf stat: Add support for per-pkg counters
perf tools: Remove perf_evsel__read interface
perf stat: Use read_counter in read_counter_aggr
perf stat: Make read_counter work over the thread dimension
perf stat: Use perf_evsel__read_cb in read_counter
perf tools: Add snapshot format file parsing
perf tools: Add per-pkg format file parsing
perf evsel: Introduce perf_evsel__read_cb function
perf evsel: Introduce perf_counts_values__scale function
perf evsel: Introduce perf_evsel__compute_deltas function
perf tools: Allow to force redirect pr_debug to stderr.
perf tools: Fix segfault due to invalid kernel dso access
perf callchain: Make get_srcline fall back to sym+offset
perf symbols: Move bfd_demangle stubbing to its only user
perf callchain: Enable printing the srcline in the history
perf tools: Collapse first level callchain entry if it has sibling
...
Pull locking tree changes from Ingo Molnar:
"Two changes: a documentation update and a ticket locks live lock fix"
* 'core-locking-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/ticketlock: Fix spin_unlock_wait() livelock
locking/lglocks: Add documentation of current lglocks implementation
Pull asm-generic asm/io.h rewrite from Arnd Bergmann:
"While there normally is no reason to have a pull request for
asm-generic but have all changes get merged through whichever tree
needs them, I do have a series for 3.19.
There are two sets of patches that change significant portions of
asm/io.h, and this branch contains both in order to resolve the
conflicts:
- Will Deacon has done a set of patches to ensure that all
architectures define {read,write}{b,w,l,q}_relaxed() functions or
get them by including asm-generic/io.h.
These functions are commonly used on ARM specific drivers to avoid
expensive L2 cache synchronization implied by the normal
{read,write}{b,w,l,q}, but we need to define them on all
architectures in order to share the drivers across architectures
and to enable CONFIG_COMPILE_TEST configurations for them
- Thierry Reding has done an unrelated set of patches that extends
the asm-generic/io.h file to the degree necessary to make it useful
on ARM64 and potentially other architectures"
* tag 'asm-generic-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/arnd/asm-generic: (29 commits)
ARM64: use GENERIC_PCI_IOMAP
sparc: io: remove duplicate relaxed accessors on sparc32
ARM: sa11x0: Use void __iomem * in MMIO accessors
arm64: Use include/asm-generic/io.h
ARM: Use include/asm-generic/io.h
asm-generic/io.h: Implement generic {read,write}s*()
asm-generic/io.h: Reconcile I/O accessor overrides
/dev/mem: Use more consistent data types
Change xlate_dev_{kmem,mem}_ptr() prototypes
ARM: ixp4xx: Properly override I/O accessors
ARM: ixp4xx: Fix build with IXP4XX_INDIRECT_PCI
ARM: ebsa110: Properly override I/O accessors
ARC: Remove redundant PCI_IOBASE declaration
documentation: memory-barriers: clarify relaxed io accessor semantics
x86: io: implement dummy relaxed accessor macros for writes
tile: io: implement dummy relaxed accessor macros for writes
sparc: io: implement dummy relaxed accessor macros for writes
powerpc: io: implement dummy relaxed accessor macros for writes
parisc: io: implement dummy relaxed accessor macros for writes
mn10300: io: implement dummy relaxed accessor macros for writes
...
Introduce two helper functions to safely read and write unsigned long
values from or to memory when the access may fault because the mapping
is non-present or read-only.
These helpers can be used instead of open coded uses of __get_user()
and __put_user() avoiding the need to do casts to fix sparse warnings.
Use the helpers in page.h and p2m.c. This will fix the sparse
warnings when doing "make C=1".
Signed-off-by: Juergen Gross <jgross@suse.com>
Signed-off-by: David Vrabel <david.vrabel@citrix.com>
Pull AMD range breakpoints support from Frederic Weisbecker:
" - Extend breakpoint tools and core to support address range through perf
event with initial backend support for AMD extended breakpoints.
Syntax is:
perf record -e mem:addr/len:type
For example set write breakpoint from 0x1000 to 0x1200 (0x1000 + 512)
perf record -e mem:0x1000/512:w
- Clean up a bit breakpoint code validation
It has been acked by Jiri and Oleg. "
Signed-off-by: Ingo Molnar <mingo@kernel.org>
arch_spin_unlock_wait() looks very suboptimal, to the point I
think this is just wrong and can lead to livelock: if the lock
is heavily contended we can never see head == tail.
But we do not need to wait for arch_spin_is_locked() == F. If it
is locked we only need to wait until the current owner drops
this lock. So we could simply spin until old_head !=
lock->tickets.head in this case, but .head can overflow and thus
we can't check "unlocked" only once before the main loop.
Also, the "unlocked" check can ignore TICKET_SLOWPATH_FLAG bit.
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Acked-by: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Jeremy Fitzhardinge <jeremy@goop.org>
Cc: Paul E.McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Waiman Long <Waiman.Long@hp.com>
Link: http://lkml.kernel.org/r/20141201213417.GA5842@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Normally, we do reapply microcode on resume. However, in the cases where
that microcode comes from the early loader and the late loader hasn't
been utilized yet, there's no easy way for us to go and apply the patch
applied during boot by the early loader.
Thus, reuse the patch stashed by the early loader for the BSP.
Signed-off-by: Borislav Petkov <bp@suse.de>
Expose the XSAVES feature to the guest if the kvm_x86_ops say it is
available.
Signed-off-by: Wanpeng Li <wanpeng.li@linux.intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
At start of the day the Xen hypervisor presents a contiguous mfn list
to a pv-domain. In order to support sparse memory this mfn list is
accessed via a three level p2m tree built early in the boot process.
Whenever the system needs the mfn associated with a pfn this tree is
used to find the mfn.
Instead of using a software walked tree for accessing a specific mfn
list entry this patch is creating a virtual address area for the
entire possible mfn list including memory holes. The holes are
covered by mapping a pre-defined page consisting only of "invalid
mfn" entries. Access to a mfn entry is possible by just using the
virtual base address of the mfn list and the pfn as index into that
list. This speeds up the (hot) path of determining the mfn of a
pfn.
Kernel build on a Dell Latitude E6440 (2 cores, HT) in 64 bit Dom0
showed following improvements:
Elapsed time: 32:50 -> 32:35
System: 18:07 -> 17:47
User: 104:00 -> 103:30
Tested with following configurations:
- 64 bit dom0, 8GB RAM
- 64 bit dom0, 128 GB RAM, PCI-area above 4 GB
- 32 bit domU, 512 MB, 8 GB, 43 GB (more wouldn't work even without
the patch)
- 32 bit domU, ballooning up and down
- 32 bit domU, save and restore
- 32 bit domU with PCI passthrough
- 64 bit domU, 8 GB, 2049 MB, 5000 MB
- 64 bit domU, ballooning up and down
- 64 bit domU, save and restore
- 64 bit domU with PCI passthrough
Signed-off-by: Juergen Gross <jgross@suse.com>
Signed-off-by: David Vrabel <david.vrabel@citrix.com>
Today get_phys_to_machine() is always called when the mfn for a pfn
is to be obtained. Add a wrapper __pfn_to_mfn() as inline function
to be able to avoid calling get_phys_to_machine() when possible as
soon as the switch to a linear mapped p2m list has been done.
Signed-off-by: Juergen Gross <jgross@suse.com>
Reviewed-by: David Vrabel <david.vrabel@citrix.com>
Signed-off-by: David Vrabel <david.vrabel@citrix.com>
Introduces lookup_pmd_address() to get the address of the pmd entry
related to a virtual address in the current address space. This
function is needed for support of a virtual mapped sparse p2m list
in xen pv domains, as we need the address of the pmd entry, not the
one of the pte in that case.
Signed-off-by: Juergen Gross <jgross@suse.com>
Signed-off-by: David Vrabel <david.vrabel@citrix.com>
When the physical memory configuration is initialized the p2m entries
for not pouplated memory pages are set to "invalid". As those pages
are beyond the hypervisor built p2m list the p2m tree has to be
extended.
This patch delays processing the extra memory related p2m entries
during the boot process until some more basic memory management
functions are callable. This removes the need to create new p2m
entries until virtual memory management is available.
Signed-off-by: Juergen Gross <jgross@suse.com>
Reviewed-by: David Vrabel <david.vrabel@citrix.com>
Signed-off-by: David Vrabel <david.vrabel@citrix.com>
Early in the boot process the memory layout of a pv-domain is changed
to match the E820 map (either the host one for Dom0 or the Xen one)
regarding placement of RAM and PCI holes. This requires removing memory
pages initially located at positions not suitable for RAM and adding
them later at higher addresses where no restrictions apply.
To be able to operate on the hypervisor supported p2m list until a
virtual mapped linear p2m list can be constructed, remapping must
be delayed until virtual memory management is initialized, as the
initial p2m list can't be extended unlimited at physical memory
initialization time due to it's fixed structure.
A further advantage is the reduction in complexity and code volume as
we don't have to be careful regarding memory restrictions during p2m
updates.
Signed-off-by: Juergen Gross <jgross@suse.com>
Reviewed-by: David Vrabel <david.vrabel@citrix.com>
Signed-off-by: David Vrabel <david.vrabel@citrix.com>
Some functions in arch/x86/xen/p2m.c are used locally only. Make them
static. Rearrange the functions in p2m.c to avoid forward declarations.
Signed-off-by: Juergen Gross <jgross@suse.com>
Signed-off-by: David Vrabel <david.vrabel@citrix.com>
When hardware supports APIC/x2APIC virtualization we don't need to use
pirqs for MSI handling and instead use APIC since most APIC accesses
(MMIO or MSR) will now be processed without VMEXITs.
As an example, netperf on the original code produces this profile
(collected wih 'xentrace -e 0x0008ffff -T 5'):
342 cpu_change
260 CPUID
34638 HLT
64067 INJ_VIRQ
28374 INTR
82733 INTR_WINDOW
10 NPF
24337 TRAP
370610 vlapic_accept_pic_intr
307528 VMENTRY
307527 VMEXIT
140998 VMMCALL
127 wrap_buffer
After applying this patch the same test shows
230 cpu_change
260 CPUID
36542 HLT
174 INJ_VIRQ
27250 INTR
222 INTR_WINDOW
20 NPF
24999 TRAP
381812 vlapic_accept_pic_intr
166480 VMENTRY
166479 VMEXIT
77208 VMMCALL
81 wrap_buffer
ApacheBench results (ab -n 10000 -c 200) improve by about 10%
Signed-off-by: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
Signed-off-by: David Vrabel <david.vrabel@citrix.com>
Introduce an arch specific function to find out whether a particular dma
mapping operation needs to bounce on the swiotlb buffer.
On ARM and ARM64, if the page involved is a foreign page and the device
is not coherent, we need to bounce because at unmap time we cannot
execute any required cache maintenance operations (we don't know how to
find the pfn from the mfn).
No change of behaviour for x86.
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Reviewed-by: David Vrabel <david.vrabel@citrix.com>
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Acked-by: Ian Campbell <ian.campbell@citrix.com>
Acked-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
dev_addr is the machine address of the page.
The new parameter can be used by the ARM and ARM64 implementations of
xen_dma_map_page to find out if the page is a local page (pfn == mfn) or
a foreign page (pfn != mfn).
dev_addr could be retrieved again from the physical address, using
pfn_to_mfn, but it requires accessing an rbtree. Since we already have
the dev_addr in our hands at the call site there is no need to get the
mfn twice.
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Acked-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Hand down the cpu number instead, otherwise lockdep screams when doing
echo 1 > /sys/devices/system/cpu/microcode/reload.
BUG: using smp_processor_id() in preemptible [00000000] code: amd64-microcode/2470
caller is debug_smp_processor_id+0x12/0x20
CPU: 1 PID: 2470 Comm: amd64-microcode Not tainted 3.18.0-rc6+ #26
...
Signed-off-by: Borislav Petkov <bp@suse.de>
Link: http://lkml.kernel.org/r/1417428741-4501-1-git-send-email-bp@alien8.de
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
cs.base is declared as a __u64 variable and vector is a u32 so this
causes a static checker warning. The user indeed can set "sipi_vector"
to any u32 value in kvm_vcpu_ioctl_x86_set_vcpu_events(), but the
value should really have 8-bit precision only.
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Create a new header, and hide the device assignment functions there.
Move struct kvm_assigned_dev_kernel to assigned-dev.c by modifying
arch/x86/kvm/iommu.c to take a PCI device struct.
Based on a patch by Radim Krcmar <rkrcmark@redhat.com>.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
x86 call do_notify_resume on paranoid returns if TIF_UPROBE is set but
not on non-paranoid returns. I suspect that this is a mistake and that
the code only works because int3 is paranoid.
Setting _TIF_NOTIFY_RESUME in the uprobe code was probably a workaround
for the x86 bug. With that bug fixed, we can remove _TIF_NOTIFY_RESUME
from the uprobes code.
Reported-by: Oleg Nesterov <oleg@redhat.com>
Acked-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Acked-by: Borislav Petkov <bp@suse.de>
Signed-off-by: Andy Lutomirski <luto@amacapital.net>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
On a 32-bit kernel, this has no effect, since there are no IST stacks.
On a 64-bit kernel, #SS can only happen in user code, on a failed iret
to user space, a canonical violation on access via RSP or RBP, or a
genuine stack segment violation in 32-bit kernel code. The first two
cases don't need IST, and the latter two cases are unlikely fatal bugs,
and promoting them to double faults would be fine.
This fixes a bug in which the espfix64 code mishandles a stack segment
violation.
This saves 4k of memory per CPU and a tiny bit of code.
Signed-off-by: Andy Lutomirski <luto@amacapital.net>
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Cc: stable@vger.kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Now that ia64 is gone, we can hide deprecated device assignment in x86.
Notable changes:
- kvm_vm_ioctl_assigned_device() was moved to x86/kvm_arch_vm_ioctl()
The easy parts were removed from generic kvm code, remaining
- kvm_iommu_(un)map_pages() would require new code to be moved
- struct kvm_assigned_dev_kernel depends on struct kvm_irq_ack_notifier
Signed-off-by: Radim Krčmář <rkrcmar@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
ia64 does not need them anymore. Ack notifiers become x86-specific
too.
Suggested-by: Gleb Natapov <gleb@kernel.org>
Reviewed-by: Radim Krcmar <rkrcmar@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Until now, the mce_severity mechanism can only identify the severity
of UCNA error as MCE_KEEP_SEVERITY. Meanwhile, it is not able to filter
out DEFERRED error for AMD platform.
This patch extends the mce_severity mechanism for handling
UCNA/DEFERRED error. In order to do this, the patch introduces a new
severity level - MCE_UCNA/DEFERRED_SEVERITY.
In addition, mce_severity is specific to machine check exception,
and it will check MCIP/EIPV/RIPV bits. In order to use mce_severity
mechanism in non-exception context, the patch also introduces a new
argument (is_excp) for mce_severity. `is_excp' is used to explicitly
specify the calling context of mce_severity.
Reviewed-by: Aravind Gopalakrishnan <Aravind.Gopalakrishnan@amd.com>
Signed-off-by: Chen Yucong <slaoub@gmail.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
asm-generic/mm_hooks.h provides some generic fillers for the 90%
of architectures that do not need to hook some mmap-manipulation
functions. A comment inside says:
> Define generic no-op hooks for arch_dup_mmap and
> arch_exit_mmap, to be included in asm-FOO/mmu_context.h
> for any arch FOO which doesn't need to hook these.
So, does x86 need to hook these? It depends on CONFIG_PARAVIRT.
We *conditionally* include this generic header if we have
CONFIG_PARAVIRT=n. That's madness.
With this patch, x86 stops using asm-generic/mmu_hooks.h entirely.
We use our own copies of the functions. The paravirt code
provides some stubs if it is disabled, and we always call those
stubs in our x86-private versions of arch_exit_mmap() and
arch_dup_mmap().
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Dave Hansen <dave@sr71.net>
Cc: x86@kernel.org
Link: http://lkml.kernel.org/r/20141118182349.14567FA5@viggo.jf.intel.com
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
The previous patch allocates bounds tables on-demand. As noted in
an earlier description, these can add up to *HUGE* amounts of
memory. This has caused OOMs in practice when running tests.
This patch adds support for freeing bounds tables when they are no
longer in use.
There are two types of mappings in play when unmapping tables:
1. The mapping with the actual data, which userspace is
munmap()ing or brk()ing away, etc...
2. The mapping for the bounds table *backing* the data
(is tagged with VM_MPX, see the patch "add MPX specific
mmap interface").
If userspace use the prctl() indroduced earlier in this patchset
to enable the management of bounds tables in kernel, when it
unmaps the first type of mapping with the actual data, the kernel
needs to free the mapping for the bounds table backing the data.
This patch hooks in at the very end of do_unmap() to do so.
We look at the addresses being unmapped and find the bounds
directory entries and tables which cover those addresses. If
an entire table is unused, we clear associated directory entry
and free the table.
Once we unmap the bounds table, we would have a bounds directory
entry pointing at empty address space. That address space might
now be allocated for some other (random) use, and the MPX
hardware might now try to walk it as if it were a bounds table.
That would be bad. So any unmapping of an enture bounds table
has to be accompanied by a corresponding write to the bounds
directory entry to invalidate it. That write to the bounds
directory can fault, which causes the following problem:
Since we are doing the freeing from munmap() (and other paths
like it), we hold mmap_sem for write. If we fault, the page
fault handler will attempt to acquire mmap_sem for read and
we will deadlock. To avoid the deadlock, we pagefault_disable()
when touching the bounds directory entry and use a
get_user_pages() to resolve the fault.
The unmapping of bounds tables happends under vm_munmap(). We
also (indirectly) call vm_munmap() to _do_ the unmapping of the
bounds tables. We avoid unbounded recursion by disallowing
freeing of bounds tables *for* bounds tables. This would not
occur normally, so should not have any practical impact. Being
strict about it here helps ensure that we do not have an
exploitable stack overflow.
Based-on-patch-by: Qiaowei Ren <qiaowei.ren@intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Cc: linux-mm@kvack.org
Cc: linux-mips@linux-mips.org
Cc: Dave Hansen <dave@sr71.net>
Link: http://lkml.kernel.org/r/20141114151831.E4531C4A@viggo.jf.intel.com
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
This is really the meat of the MPX patch set. If there is one patch to
review in the entire series, this is the one. There is a new ABI here
and this kernel code also interacts with userspace memory in a
relatively unusual manner. (small FAQ below).
Long Description:
This patch adds two prctl() commands to provide enable or disable the
management of bounds tables in kernel, including on-demand kernel
allocation (See the patch "on-demand kernel allocation of bounds tables")
and cleanup (See the patch "cleanup unused bound tables"). Applications
do not strictly need the kernel to manage bounds tables and we expect
some applications to use MPX without taking advantage of this kernel
support. This means the kernel can not simply infer whether an application
needs bounds table management from the MPX registers. The prctl() is an
explicit signal from userspace.
PR_MPX_ENABLE_MANAGEMENT is meant to be a signal from userspace to
require kernel's help in managing bounds tables.
PR_MPX_DISABLE_MANAGEMENT is the opposite, meaning that userspace don't
want kernel's help any more. With PR_MPX_DISABLE_MANAGEMENT, the kernel
won't allocate and free bounds tables even if the CPU supports MPX.
PR_MPX_ENABLE_MANAGEMENT will fetch the base address of the bounds
directory out of a userspace register (bndcfgu) and then cache it into
a new field (->bd_addr) in the 'mm_struct'. PR_MPX_DISABLE_MANAGEMENT
will set "bd_addr" to an invalid address. Using this scheme, we can
use "bd_addr" to determine whether the management of bounds tables in
kernel is enabled.
Also, the only way to access that bndcfgu register is via an xsaves,
which can be expensive. Caching "bd_addr" like this also helps reduce
the cost of those xsaves when doing table cleanup at munmap() time.
Unfortunately, we can not apply this optimization to #BR fault time
because we need an xsave to get the value of BNDSTATUS.
==== Why does the hardware even have these Bounds Tables? ====
MPX only has 4 hardware registers for storing bounds information.
If MPX-enabled code needs more than these 4 registers, it needs to
spill them somewhere. It has two special instructions for this
which allow the bounds to be moved between the bounds registers
and some new "bounds tables".
They are similar conceptually to a page fault and will be raised by
the MPX hardware during both bounds violations or when the tables
are not present. This patch handles those #BR exceptions for
not-present tables by carving the space out of the normal processes
address space (essentially calling the new mmap() interface indroduced
earlier in this patch set.) and then pointing the bounds-directory
over to it.
The tables *need* to be accessed and controlled by userspace because
the instructions for moving bounds in and out of them are extremely
frequent. They potentially happen every time a register pointing to
memory is dereferenced. Any direct kernel involvement (like a syscall)
to access the tables would obviously destroy performance.
==== Why not do this in userspace? ====
This patch is obviously doing this allocation in the kernel.
However, MPX does not strictly *require* anything in the kernel.
It can theoretically be done completely from userspace. Here are
a few ways this *could* be done. I don't think any of them are
practical in the real-world, but here they are.
Q: Can virtual space simply be reserved for the bounds tables so
that we never have to allocate them?
A: As noted earlier, these tables are *HUGE*. An X-GB virtual
area needs 4*X GB of virtual space, plus 2GB for the bounds
directory. If we were to preallocate them for the 128TB of
user virtual address space, we would need to reserve 512TB+2GB,
which is larger than the entire virtual address space today.
This means they can not be reserved ahead of time. Also, a
single process's pre-popualated bounds directory consumes 2GB
of virtual *AND* physical memory. IOW, it's completely
infeasible to prepopulate bounds directories.
Q: Can we preallocate bounds table space at the same time memory
is allocated which might contain pointers that might eventually
need bounds tables?
A: This would work if we could hook the site of each and every
memory allocation syscall. This can be done for small,
constrained applications. But, it isn't practical at a larger
scale since a given app has no way of controlling how all the
parts of the app might allocate memory (think libraries). The
kernel is really the only place to intercept these calls.
Q: Could a bounds fault be handed to userspace and the tables
allocated there in a signal handler instead of in the kernel?
A: (thanks to tglx) mmap() is not on the list of safe async
handler functions and even if mmap() would work it still
requires locking or nasty tricks to keep track of the
allocation state there.
Having ruled out all of the userspace-only approaches for managing
bounds tables that we could think of, we create them on demand in
the kernel.
Based-on-patch-by: Qiaowei Ren <qiaowei.ren@intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Cc: linux-mm@kvack.org
Cc: linux-mips@linux-mips.org
Cc: Dave Hansen <dave@sr71.net>
Link: http://lkml.kernel.org/r/20141114151829.AD4310DE@viggo.jf.intel.com
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
This patch sets bound violation fields of siginfo struct in #BR
exception handler by decoding the user instruction and constructing
the faulting pointer.
We have to be very careful when decoding these instructions. They
are completely controlled by userspace and may be changed at any
time up to and including the point where we try to copy them in to
the kernel. They may or may not be MPX instructions and could be
completely invalid for all we know.
Note: This code is based on Qiaowei Ren's specialized MPX
decoder, but uses the generic decoder whenever possible. It was
tested for robustness by generating a completely random data
stream and trying to decode that stream. I also unmapped random
pages inside the stream to test the "partial instruction" short
read code.
We kzalloc() the siginfo instead of stack allocating it because
we need to memset() it anyway, and doing this makes it much more
clear when it got initialized by the MPX instruction decoder.
Changes from the old decoder:
* Use the generic decoder instead of custom functions. Saved
~70 lines of code overall.
* Remove insn->addr_bytes code (never used??)
* Make sure never to possibly overflow the regoff[] array, plus
check the register range correctly in 32 and 64-bit modes.
* Allow get_reg() to return an error and have mpx_get_addr_ref()
handle when it sees errors.
* Only call insn_get_*() near where we actually use the values
instead if trying to call them all at once.
* Handle short reads from copy_from_user() and check the actual
number of read bytes against what we expect from
insn_get_length(). If a read stops in the middle of an
instruction, we error out.
* Actually check the opcodes intead of ignoring them.
* Dynamically kzalloc() siginfo_t so we don't leak any stack
data.
* Detect and handle decoder failures instead of ignoring them.
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Based-on-patch-by: Qiaowei Ren <qiaowei.ren@intel.com>
Cc: linux-mm@kvack.org
Cc: linux-mips@linux-mips.org
Cc: Dave Hansen <dave@sr71.net>
Link: http://lkml.kernel.org/r/20141114151828.5BDD0915@viggo.jf.intel.com
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
We have chosen to perform the allocation of bounds tables in
kernel (See the patch "on-demand kernel allocation of bounds
tables") and to mark these VMAs with VM_MPX.
However, there is currently no suitable interface to actually do
this. Existing interfaces, like do_mmap_pgoff(), have no way to
set a modified ->vm_ops or ->vm_flags and don't hold mmap_sem
long enough to let a caller do it.
This patch wraps mmap_region() and hold mmap_sem long enough to
make the modifications to the VMA which we need.
Also note the 32/64-bit #ifdef in the header. We actually need
to do this at runtime eventually. But, for now, we don't support
running 32-bit binaries on 64-bit kernels. Support for this will
come in later patches.
Signed-off-by: Qiaowei Ren <qiaowei.ren@intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Cc: linux-mm@kvack.org
Cc: linux-mips@linux-mips.org
Cc: Dave Hansen <dave@sr71.net>
Link: http://lkml.kernel.org/r/20141114151827.CE440F67@viggo.jf.intel.com
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Consider the bndX MPX registers. There 4 registers each
containing a 64-bit lower and a 64-bit upper bound. That's 8*64
bits and we declare it thusly:
struct bndregs_struct {
u64 bndregs[8];
}
Let's say you want to read the upper bound from the MPX register
bnd2 out of the xsave buf. You do:
bndregno = 2;
upper_bound = xsave_buf->bndregs.bndregs[2*bndregno+1];
That kinda sucks. Every time you access it, you need to know:
1. Each bndX register is two entries wide in "bndregs"
2. The lower comes first followed by upper. We do the +1 to get
upper vs. lower.
This replaces the old definition. You can now access them
indexed by the register number directly, and with a meaningful
name for the lower and upper bound:
bndregno = 2;
xsave_buf->bndreg[bndregno].upper_bound;
It's now *VERY* clear that there are 4 registers. The programmer
now doesn't have to care what order the lower and upper bounds
are in, and it's harder to get it wrong.
[ tglx: Changed ub/lb to upper_bound/lower_bound and renamed struct
bndreg_struct to struct bndreg ]
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Cc: x86@kernel.org
Cc: "H. Peter Anvin" <hpa@linux.intel.com>
Cc: Qiaowei Ren <qiaowei.ren@intel.com>
Cc: "Yu, Fenghua" <fenghua.yu@intel.com>
Cc: Dave Hansen <dave@sr71.net>
Link: http://lkml.kernel.org/r/20141031215820.5EA5E0EC@viggo.jf.intel.com
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
The current x86 instruction decoder steps along through the
instruction stream but always ensures that it never steps farther
than the largest possible instruction size (MAX_INSN_SIZE).
The MPX code is now going to be doing some decoding of userspace
instructions. We copy those from userspace in to the kernel and
they're obviously completely untrusted coming from userspace. In
addition to the constraint that instructions can only be so long,
we also have to be aware of how long the buffer is that came in
from userspace. This _looks_ to be similar to what the perf and
kprobes is doing, but it's unclear to me whether they are
affected.
The whole reason we need this is that it is perfectly valid to be
executing an instruction within MAX_INSN_SIZE bytes of an
unreadable page. We should be able to gracefully handle short
reads in those cases.
This adds support to the decoder to record how long the buffer
being decoded is and to refuse to "validate" the instruction if
we would have gone over the end of the buffer to decode it.
The kprobes code probably needs to be looked at here a bit more
carefully. This patch still respects the MAX_INSN_SIZE limit
there but the kprobes code does look like it might be able to
be a bit more strict than it currently is.
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Acked-by: Jim Keniston <jkenisto@us.ibm.com>
Acked-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: x86@kernel.org
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com>
Cc: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>
Cc: "David S. Miller" <davem@davemloft.net>
Link: http://lkml.kernel.org/r/20141114153957.E6B01535@viggo.jf.intel.com
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Commit e00c8cc93c "x86: Use new cache mode type in memtype related
functions" broke the ARCH=um build.
arch/x86/include/asm/cacheflush.h:67:36: error: return type is an incomplete type
static inline enum page_cache_mode get_page_memtype(struct page *pg)
The reason is simple. get_page_memtype() and set_page_memtype()
require enum page_cache_mode now, which is defined in
asm/pgtable_types.h. UM does not include that file for obvious reasons.
The simple solution is to move that functions to arch/x86/mm/pat.c
where the only callsites of this are located. They should have been
there in the first place.
Fixes: e00c8cc93c "x86: Use new cache mode type in memtype related functions"
Reported-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Juergen Gross <jgross@suse.com>
Cc: Richard Weinberger <richard@nod.at>
