'perf cqm' never worked due to the incompatibility between perf
infrastructure and cqm hardware support. The hardware uses RMIDs to
track the llc occupancy of tasks and these RMIDs are per package. This
makes monitoring a hierarchy like cgroup along with monitoring of tasks
separately difficult and several patches sent to lkml to fix them were
NACKed. Further more, the following issues in the current perf cqm make
it almost unusable:
1. No support to monitor the same group of tasks for which we do
allocation using resctrl.
2. It gives random and inaccurate data (mostly 0s) once we run out
of RMIDs due to issues in Recycling.
3. Recycling results in inaccuracy of data because we cannot
guarantee that the RMID was stolen from a task when it was not
pulling data into cache or even when it pulled the least data. Also
for monitoring llc_occupancy, if we stop using an RMID_x and then
start using an RMID_y after we reclaim an RMID from an other event,
we miss accounting all the occupancy that was tagged to RMID_x at a
later perf_count.
2. Recycling code makes the monitoring code complex including
scheduling because the event can lose RMID any time. Since MBM
counters count bandwidth for a period of time by taking snap shot of
total bytes at two different times, recycling complicates the way we
count MBM in a hierarchy. Also we need a spin lock while we do the
processing to account for MBM counter overflow. We also currently
use a spin lock in scheduling to prevent the RMID from being taken
away.
4. Lack of support when we run different kind of event like task,
system-wide and cgroup events together. Data mostly prints 0s. This
is also because we can have only one RMID tied to a cpu as defined
by the cqm hardware but a perf can at the same time tie multiple
events during one sched_in.
5. No support of monitoring a group of tasks. There is partial support
for cgroup but it does not work once there is a hierarchy of cgroups
or if we want to monitor a task in a cgroup and the cgroup itself.
6. No support for monitoring tasks for the lifetime without perf
overhead.
7. It reported the aggregate cache occupancy or memory bandwidth over
all sockets. But most cloud and VMM based use cases want to know the
individual per-socket usage.
Signed-off-by: Vikas Shivappa <vikas.shivappa@linux.intel.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: ravi.v.shankar@intel.com
Cc: tony.luck@intel.com
Cc: fenghua.yu@intel.com
Cc: peterz@infradead.org
Cc: eranian@google.com
Cc: vikas.shivappa@intel.com
Cc: ak@linux.intel.com
Cc: davidcc@google.com
Cc: reinette.chatre@intel.com
Link: http://lkml.kernel.org/r/1501017287-28083-2-git-send-email-vikas.shivappa@linux.intel.com
We are using the same vector for nested/non-nested posted
interrupts delivery, this may cause interrupts latency in
L1 since we can't kick the L2 vcpu out of vmx-nonroot mode.
This patch introduces a new vector which is only for nested
posted interrupts to solve the problems above.
Signed-off-by: Wincy Van <fanwenyi0529@gmail.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Add the new ORC unwinder which is enabled by CONFIG_ORC_UNWINDER=y.
It plugs into the existing x86 unwinder framework.
It relies on objtool to generate the needed .orc_unwind and
.orc_unwind_ip sections.
For more details on why ORC is used instead of DWARF, see
Documentation/x86/orc-unwinder.txt - but the short version is
that it's a simplified, fundamentally more robust debugninfo
data structure, which also allows up to two orders of magnitude
faster lookups than the DWARF unwinder - which matters to
profiling workloads like perf.
Thanks to Andy Lutomirski for the performance improvement ideas:
splitting the ORC unwind table into two parallel arrays and creating a
fast lookup table to search a subset of the unwind table.
Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Denys Vlasenko <dvlasenk@redhat.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Jiri Slaby <jslaby@suse.cz>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: live-patching@vger.kernel.org
Link: http://lkml.kernel.org/r/0a6cbfb40f8da99b7a45a1a8302dc6aef16ec812.1500938583.git.jpoimboe@redhat.com
[ Extended the changelog. ]
Signed-off-by: Ingo Molnar <mingo@kernel.org>
PCID is a "process context ID" -- it's what other architectures call
an address space ID. Every non-global TLB entry is tagged with a
PCID, only TLB entries that match the currently selected PCID are
used, and we can switch PGDs without flushing the TLB. x86's
PCID is 12 bits.
This is an unorthodox approach to using PCID. x86's PCID is far too
short to uniquely identify a process, and we can't even really
uniquely identify a running process because there are monster
systems with over 4096 CPUs. To make matters worse, past attempts
to use all 12 PCID bits have resulted in slowdowns instead of
speedups.
This patch uses PCID differently. We use a PCID to identify a
recently-used mm on a per-cpu basis. An mm has no fixed PCID
binding at all; instead, we give it a fresh PCID each time it's
loaded except in cases where we want to preserve the TLB, in which
case we reuse a recent value.
Here are some benchmark results, done on a Skylake laptop at 2.3 GHz
(turbo off, intel_pstate requesting max performance) under KVM with
the guest using idle=poll (to avoid artifacts when bouncing between
CPUs). I haven't done any real statistics here -- I just ran them
in a loop and picked the fastest results that didn't look like
outliers. Unpatched means commit a4eb8b9935, so all the
bookkeeping overhead is gone.
ping-pong between two mms on the same CPU using eventfd:
patched: 1.22µs
patched, nopcid: 1.33µs
unpatched: 1.34µs
Same ping-pong, but now touch 512 pages (all zero-page to minimize
cache misses) each iteration. dTLB misses are measured by
dtlb_load_misses.miss_causes_a_walk:
patched: 1.8µs 11M dTLB misses
patched, nopcid: 6.2µs, 207M dTLB misses
unpatched: 6.1µs, 190M dTLB misses
Signed-off-by: Andy Lutomirski <luto@kernel.org>
Reviewed-by: Nadav Amit <nadav.amit@gmail.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Arjan van de Ven <arjan@linux.intel.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Dave Hansen <dave.hansen@intel.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Rik van Riel <riel@redhat.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-mm@kvack.org
Link: http://lkml.kernel.org/r/9ee75f17a81770feed616358e6860d98a2a5b1e7.1500957502.git.luto@kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Pull x86 fixes from Ingo Molnar:
"Half of the fixes are for various build time warnings triggered by
randconfig builds. Most (but not all...) were harmless.
There's also:
- ACPI boundary condition fixes
- UV platform fixes
- defconfig updates
- an AMD K6 CPU init fix
- a %pOF printk format related preparatory change
- .. and a warning fix related to the tlb/PCID changes"
* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/devicetree: Convert to using %pOF instead of ->full_name
x86/platform/uv/BAU: Disable BAU on single hub configurations
x86/platform/intel-mid: Fix a format string overflow warning
x86/platform: Add PCI dependency for PUNIT_ATOM_DEBUG
x86/build: Silence the build with "make -s"
x86/io: Add "memory" clobber to insb/insw/insl/outsb/outsw/outsl
x86/fpu/math-emu: Avoid bogus -Wint-in-bool-context warning
x86/fpu/math-emu: Fix possible uninitialized variable use
perf/x86: Shut up false-positive -Wmaybe-uninitialized warning
x86/defconfig: Remove stale, old Kconfig options
x86/ioapic: Pass the correct data to unmask_ioapic_irq()
x86/acpi: Prevent out of bound access caused by broken ACPI tables
x86/mm, KVM: Fix warning when !CONFIG_PREEMPT_COUNT
x86/platform/uv/BAU: Fix congested_response_us not taking effect
x86/cpu: Use indirect call to measure performance in init_amd_k6()
Pull core fixes from Ingo Molnar:
"A fix to WARN_ON_ONCE() done by modules, plus a MAINTAINERS update"
* 'core-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
debug: Fix WARN_ON_ONCE() for modules
MAINTAINERS: Update the PTRACE entry
On x86, 5-level paging enables 56-bit userspace virtual address space.
Not all user space is ready to handle wide addresses. It's known that
at least some JIT compilers use higher bits in pointers to encode their
information. It collides with valid pointers with 5-level paging and
leads to crashes.
To mitigate this, we are not going to allocate virtual address space
above 47-bit by default.
But userspace can ask for allocation from full address space by
specifying hint address (with or without MAP_FIXED) above 47-bits.
If hint address set above 47-bit, but MAP_FIXED is not specified, we try
to look for unmapped area by specified address. If it's already
occupied, we look for unmapped area in *full* address space, rather than
from 47-bit window.
A high hint address would only affect the allocation in question, but not
any future mmap()s.
Specifying high hint address on older kernel or on machine without 5-level
paging support is safe. The hint will be ignored and kernel will fall back
to allocation from 47-bit address space.
This approach helps to easily make application's memory allocator aware
about large address space without manually tracking allocated virtual
address space.
The patch puts all machinery in place, but not yet allows userspace to have
mappings above 47-bit -- TASK_SIZE_MAX has to be raised to get the effect.
Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Dave Hansen <dave.hansen@intel.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-arch@vger.kernel.org
Cc: linux-mm@kvack.org
Link: http://lkml.kernel.org/r/20170716225954.74185-7-kirill.shutemov@linux.intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
They really are, and the "take the address of a single character" makes
the string fortification code unhappy (it believes that you can now only
acccess one byte, rather than a byte range, and then raises errors for
the memory copies going on in there).
We could now remove a few 'addressof' operators (since arrays naturally
degrade to pointers), but this is the minimal patch that just changes
the C prototypes of those template arrays (the templates themselves are
defined in inline asm).
Reported-by: kernel test robot <xiaolong.ye@intel.com>
Acked-and-tested-by: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Daniel Micay <danielmicay@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Mike Galbraith reported a situation where a WARN_ON_ONCE() call in DRM
code turned into an oops. As it turns out, WARN_ON_ONCE() seems to be
completely broken when called from a module.
The bug was introduced with the following commit:
19d436268d ("debug: Add _ONCE() logic to report_bug()")
That commit changed WARN_ON_ONCE() to move its 'once' logic into the bug
trap handler. It requires a writable bug table so that the BUGFLAG_DONE
bit can be written to the flags to indicate the first warning has
occurred.
The bug table was made writable for vmlinux, which relies on
vmlinux.lds.S and vmlinux.lds.h for laying out the sections. However,
it wasn't made writable for modules, which rely on the ELF section
header flags.
Reported-by: Mike Galbraith <efault@gmx.de>
Tested-by: Masami Hiramatsu <mhiramat@kernel.org>
Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Fixes: 19d436268d ("debug: Add _ONCE() logic to report_bug()")
Link: http://lkml.kernel.org/r/a53b04235a65478dd9afc51f5b329fdc65c84364.1500095401.git.jpoimboe@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
The x86 version of insb/insw/insl uses an inline assembly that does
not have the target buffer listed as an output. This can confuse
the compiler, leading it to think that a subsequent access of the
buffer is uninitialized:
drivers/net/wireless/wl3501_cs.c: In function ‘wl3501_mgmt_scan_confirm’:
drivers/net/wireless/wl3501_cs.c:665:9: error: ‘sig.status’ is used uninitialized in this function [-Werror=uninitialized]
drivers/net/wireless/wl3501_cs.c:668:12: error: ‘sig.cap_info’ may be used uninitialized in this function [-Werror=maybe-uninitialized]
drivers/net/sb1000.c: In function 'sb1000_rx':
drivers/net/sb1000.c:775:9: error: 'st[0]' is used uninitialized in this function [-Werror=uninitialized]
drivers/net/sb1000.c:776:10: error: 'st[1]' may be used uninitialized in this function [-Werror=maybe-uninitialized]
drivers/net/sb1000.c:784:11: error: 'st[1]' may be used uninitialized in this function [-Werror=maybe-uninitialized]
I tried to mark the exact input buffer as an output here, but couldn't
figure it out. As suggested by Linus, marking all memory as clobbered
however is good enough too. For the outs operations, I also add the
memory clobber, to force the input to be written to local variables.
This is probably already guaranteed by the "asm volatile", but it can't
hurt to do this for symmetry.
Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Borislav Petkov <bp@suse.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tom Lendacky <thomas.lendacky@amd.com>
Link: http://lkml.kernel.org/r/20170719125310.2487451-5-arnd@arndb.de
Link: https://lkml.org/lkml/2017/7/12/605
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Pull structure randomization updates from Kees Cook:
"Now that IPC and other changes have landed, enable manual markings for
randstruct plugin, including the task_struct.
This is the rest of what was staged in -next for the gcc-plugins, and
comes in three patches, largest first:
- mark "easy" structs with __randomize_layout
- mark task_struct with an optional anonymous struct to isolate the
__randomize_layout section
- mark structs to opt _out_ of automated marking (which will come
later)
And, FWIW, this continues to pass allmodconfig (normal and patched to
enable gcc-plugins) builds of x86_64, i386, arm64, arm, powerpc, and
s390 for me"
* tag 'gcc-plugins-v4.13-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux:
randstruct: opt-out externally exposed function pointer structs
task_struct: Allow randomized layout
randstruct: Mark various structs for randomization
Changes to the existing page table macros will allow the SME support to
be enabled in a simple fashion with minimal changes to files that use these
macros. Since the memory encryption mask will now be part of the regular
pagetable macros, we introduce two new macros (_PAGE_TABLE_NOENC and
_KERNPG_TABLE_NOENC) to allow for early pagetable creation/initialization
without the encryption mask before SME becomes active. Two new pgprot()
macros are defined to allow setting or clearing the page encryption mask.
The FIXMAP_PAGE_NOCACHE define is introduced for use with MMIO. SME does
not support encryption for MMIO areas so this define removes the encryption
mask from the page attribute.
Two new macros are introduced (__sme_pa() / __sme_pa_nodebug()) to allow
creating a physical address with the encryption mask. These are used when
working with the cr3 register so that the PGD can be encrypted. The current
__va() macro is updated so that the virtual address is generated based off
of the physical address without the encryption mask thus allowing the same
virtual address to be generated regardless of whether encryption is enabled
for that physical location or not.
Also, an early initialization function is added for SME. If SME is active,
this function:
- Updates the early_pmd_flags so that early page faults create mappings
with the encryption mask.
- Updates the __supported_pte_mask to include the encryption mask.
- Updates the protection_map entries to include the encryption mask so
that user-space allocations will automatically have the encryption mask
applied.
Signed-off-by: Tom Lendacky <thomas.lendacky@amd.com>
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Borislav Petkov <bp@suse.de>
Cc: Alexander Potapenko <glider@google.com>
Cc: Andrey Ryabinin <aryabinin@virtuozzo.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Brijesh Singh <brijesh.singh@amd.com>
Cc: Dave Young <dyoung@redhat.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: Larry Woodman <lwoodman@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Matt Fleming <matt@codeblueprint.co.uk>
Cc: Michael S. Tsirkin <mst@redhat.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Radim Krčmář <rkrcmar@redhat.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Toshimitsu Kani <toshi.kani@hpe.com>
Cc: kasan-dev@googlegroups.com
Cc: kvm@vger.kernel.org
Cc: linux-arch@vger.kernel.org
Cc: linux-doc@vger.kernel.org
Cc: linux-efi@vger.kernel.org
Cc: linux-mm@kvack.org
Link: http://lkml.kernel.org/r/b36e952c4c39767ae7f0a41cf5345adf27438480.1500319216.git.thomas.lendacky@amd.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>